Hypoglycaemia in Clinical Diabetes

  • No tags were found...

Hypoglycaemia in Clinical Diabetes

Hypoglycaemia in Clinical DiabetesSecond Edition

Other titles in the Wiley Diabetes in Practice SeriesExercise and Sport in Diabetes Second EditionEdited by Dinesh Nagi047002206XComplementary Therapies and the Management of Diabetes and Vascular DiseaseEdited by Patricia Dunning047001458XDiabetes in Clinical Practice: Questions and Answers from Case StudiesN. Katsilambros, E. Diakoumopoulou, I. Ioannidis, S. Liatis, K. Makrilakis, N. Tentolouris and P. Tsapogas978 0470 035221Obesity and DiabetesEdited by Anthony Barnett and Sudhesh Kumar0470848987Prevention of Type 2 DiabetesEdited by Manfred Ganz0470857331Diabetes – Chronic Complications Second EditionEdited by Kenneth Shaw and Michael Cummings0470865972The Metabolic SyndromeEdited by Christopher Byrne and Sarah Wild0470025115Psychology in Diabetes Care Second EditionEdited by Frank J. Snoek and T. Chas Skinner0470023848The Foot in Diabetes Fourth EditionEdited by Andrew J.M. Boulton, Peter R. Cavanagh and Gerry Rayman0470015047Gastrointestinal Function in Diabetes MellitusEdited by Michael Horowitz and Melvin Samson047189916XDiabetic NephropathyEdited by Christoph Hasslacher0471489921Nutritional Management of Diabetes MellitusEdited by Gary Frost, Anne Dornhorst and Robert Moses0471497517Diabetes in Pregnancy: An International Approach to Diagnosis and ManagementEdited by Anne Dornhorst and David R. Hadden047196204X

Hypoglycaemia in Clinical DiabetesSecond EditionEdited byBrian M. FrierThe Royal Infirmary of Edinburgh, Scotland, UKMiles FisherGlasgow Royal Infirmary, Scotland, UK

Copyright © 2007John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester,West Sussex PO19 8SQ, EnglandTelephone (+44) 1243 779777Email (for orders and customer service enquiries): cs-books@wiley.co.ukVisit our Home Page on www.wiley.comAll Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted inany form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except underthe terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the CopyrightLicensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission in writing ofthe Publisher. Requests to the Publisher should be addressed to the Permissions Department, John Wiley & SonsLtd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed topermreq@wiley.co.uk, or faxed to (+44) 1243 770620.Designations used by companies to distinguish their products are often claimed as trademarks. All brand namesand product names used in this book are trade names, service marks, trademarks or registered trademarks of theirrespective owners. The Publisher is not associated with any product or vendor mentioned in this book.This publication is designed to provide accurate and authoritative information in regard to the subject mattercovered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. Ifprofessional advice or other expert assistance is required, the services of a competent professional should besought.Other Wiley Editorial OfficesJohn Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USAJossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USAWiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, GermanyJohn Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, AustraliaJohn Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809John Wiley & Sons Canada Ltd, 6045 Freemont Blvd, Mississauga, Ontario, L5R 4J3Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not beavailable in electronic books.Anniversary Logo Design: Richard J. PacificoLibrary of Congress Cataloging in Publication DataHypoglycaemia in clinical diabetes / edited by Brian M. Frier and Miles Fisher. — 2nd ed.p. ; cm.Includes bibliographical references and index.ISBN 978-0-470-01844-6 (cloth : alk. paper)1. Hypoglycemia. 2. Diabetes—Treatment—Complications. 3. Hypoglycemic agents—Sideeffects. I. Frier, Brian M. II. Fisher, Miles. III. Title: Hypoglycemia in clinical diabetes.[DNLM: 1. Hypoglycemia—complications. 2. Hypoglycemia—physiopathology. 3. DiabetesComplications. 4. Insulin—adverse effects. WK 880 H9963 2007]RC662.2.H965 2007616.4 ′ 66—dc222007012095British Library Cataloguing in Publication DataA catalogue record for this book is available from the British LibraryISBN 978-0-470-01844-6Typeset in 10/12pt Times by Integra Software Services Pvt. Ltd, Pondicherry, IndiaPrinted and bound in Great Britain by Antony Rowe Ltd, Chippenham, WiltshireThis book is printed on acid-free paper responsibly manufactured from sustainable forestry in which at least twotrees are planted for each one used for paper production.

ToEmily, Ben and Marc

ContentsPrefaceContributorsixxi1 Normal Glucose Metabolism and Responses to Hypoglycaemia 1Ian A. Macdonald and Paromita King2 Symptoms of Hypoglycaemia and Effects on Mental Performanceand Emotions 25Ian J. Deary3 Frequency, Causes and Risk Factors for Hypoglycaemia in Type 1Diabetes 49Mark W.J. Strachan4 Nocturnal Hypoglycaemia 83Simon R. Heller5 Moderators, Monitoring and Management of Hypoglycaemia 101Tristan Richardson and David Kerr6 Counterregulatory Deficiencies in Diabetes 121David Kerr and Tristan Richardson7 Impaired Awareness of Hypoglycaemia 141Brian M. Frier8 Risks of Strict Glycaemic Control 171Stephanie A. Amiel9 Hypoglycaemia in Children with Diabetes 191Krystyna A. Matyka10 Hypoglycaemia in Pregnancy 217Ann E. Gold and Donald W.M. Pearson11 Hypoglycaemia in Type 2 Diabetes and in Elderly People 239Nicola N. Zammitt and Brian M. Frier12 Mortality, Cardiovascular Morbidity and Possible Effects ofHypoglycaemia on Diabetic Complications 265Miles Fisher and Simon R. Heller

viiiCONTENTS13 Long-term Effects of Hypoglycaemia on Cognitive Function and theBrain in Diabetes 285Petros Perros and Ian J. Deary14 Living with Hypoglycaemia 309Brian M. FrierIndex 333

PrefaceIn the second edition of this book, we have continued to emphasise the clinical significanceof hypoglycaemia to the person who has diabetes, particularly when receiving treatmentwith insulin. Since the first edition of the book was published in 1999, new therapies haveemerged, including new insulin analogues and inhaled insulin, and monitoring systems arenow available that can provide continuous recording of blood glucose. However, far fromminimising the risk of hypoglycaemia in clinical practice, the newer treatments have beenshown to be as liable to cause hypoglycaemia as before, while continuous blood glucosemonitoring has revealed that this side-effect of insulin therapy is even more common thanwas believed previously. The frequency of severe hypoglycaemia in vulnerable groups suchas children and elderly people receiving insulin therapy is unacceptably high, and presentspotentially serious risks to health as well as diminishing their quality of life. Much scientificresearch in recent years has focused on the effects of hypoglycaemia on the brain, providinga greater understanding of the protean effects of this metabolic abnormality. New data andconcepts have been incorporated in this edition, particularly where these are of importanceto clinical practice.In updating and revising this book about hypoglycaemia, particular emphasis has beengiven to the risk factors for hypoglycaemia and how these may be reduced or avoided. Newchapters have been included to discuss recognised moderators of hypoglycaemia and the roleof new glucose monitoring systems, to address the increasing problem of hypoglycaemia inpeople with type 2 diabetes and the elderly person, and to acknowledge the major importanceof nocturnal hypoglycaemia, which is frequently not identified in clinical practice but canhave serious consequences, not only in its immediate morbidity, but also in promoting thedevelopment of the acquired syndromes of hypoglycaemia.We are grateful for the expert assistance and support of the colleagues who havecontributed chapters, some of whom are new as authors for this edition. All have skillfullyhighlighted the relevance of the enhancement of scientific knowledge in this field to theeveryday management of diabetes, which we hope will assist all members of the diabetesteam in their efforts to prevent and manage the extremely common but unwanted scourgethat is hypoglycaemia.Brian M. FrierMiles Fisher

ContributorsProfessor Stephanie A. Amiel, R.D. Lawrence Professor of Diabetes, Departmentof Medicine, King’s College Hospital, Bessemer Road, London, SE5 9PJ (e-mail:stephanie.amiel@kcl.ac.uk)Professor Ian J. Deary, Department of Psychology, University of Edinburgh, 7 GeorgeSquare, Edinburgh, EH8 9JZ (e-mail: ian.deary@ed.ac.uk)Dr Miles Fisher, Consultant Physician, Glasgow Royal Infirmary, Glasgow, G4 0SF (e-mail:miles.fisher@northglasgow.scot.nhs.uk)Professor Brian M. Frier, Consultant Physician, Department of Diabetes, Royal Infirmary,51 Little France Crescent, Edinburgh, EH16 4SA (e-mail: brian.frier@luht.scot.nhs.uk)Dr Ann E. Gold, Consultant Physician, Wards 27 and 28, Aberdeen Royal Infirmary,Foresterhill, Aberdeen, AB25 2ZN (e-mail: ann.gold@nhs.net)Professor Simon R. Heller, Professor of Clinical Diabetes, Clinical Sciences Centre, Departmentof Diabetes & Endocrinology, Northern General Hospital, Herries Road, Sheffield, S57AU (e-mail: s.heller@sheffield.ac.uk)Dr David Kerr, Consultant Physician, The Royal Bournemouth Hospital, Castle Lane East,Bournemouth, BH7 7DW (e-mail: David.Kerr@rbch.nhs.uk)Dr Paromita King, Consultant Physician, Jenny O’Neill Diabetes Centre, Derbyshire RoyalInfirmary, London Road, Derby, DE1 2QY (e-mail: Paru.King@derbyhospitals.nhs.uk)Professor Ian A. Macdonald, Professor of Metabolic Physiology, Department of Physiologyand Pharmacology, Medical School, Queen’s Medical Centre, Nottingham, NG7 2UH(e-mail: Ian.Macdonald@nottingham.ac.uk)Dr Krystyna A. Matyka, Senior Lecturer in Paediatrics, University of Warwick MedicalSchool, Division of Clinical Sciences, CSB Research Wing, UHCW Trust, Clifford BridgeRoad, Coventry, CV2 2DX (e-mail: K.A.Matyka@warwick.ac.uk)Dr Donald W.M. Pearson, Consultant Physician, Aberdeen Royal Infirmary, Foresterhill,Aberdeen, AB25 2ZN (e-mail: Dwm.Pearson@arh.grampian.scot.nhs.uk)Dr Petros Perros, Consultant Endocrinologist, Ward 15, Freeman Hospital, Freeman Road,Newcastle-Upon-Tyne, NE7 7DN (e-mail: Petros.Perros@ncl.ac.uk)Dr Tristan Richardson, Consultant Physician, Bournemouth Diabetes and EndocrineCentre, The Royal Bournemouth Hospital, Castle Lane East, Bournemouth, BH7 7DW(e-mail: Tristan.Richardson@rbch.nhs.uk)

xiiCONTRIBUTORSDr Mark W.J. Strachan, Consultant Physician, Metabolic Unit, Western General Hospital,Crewe Road, Edinburgh, EH4 2XU (e-mail: mark.strachan@luht.scot.nhs.uk)Dr Nicola N. Zammitt, Specialist Registrar, Department of Diabetes, Royal Infirmaryof Edinburgh, 51 Little France Crescent, Edinburgh, EH16 4SA (e-mail: nicolazammitt@hotmail.com)

1 Normal Glucose Metabolismand Responses toHypoglycaemiaIan A. Macdonald and Paromita KingINTRODUCTIONControl of blood glucose is a fundamental feature of homeostasis, i.e., the process by whichthe internal environment of the body is maintained stable allowing optimal function. Bloodglucose concentrations are regulated within a narrow range (which in humans is known asnormoglycaemia or euglycaemia) despite wide variability in carbohydrate intake and physicalactivity. Teleologically, the upper limit is defended because high glucose concentrationscause microvascular complications, and the lower limit, because the brain cannot functionwithout an adequate supply of glucose. In this chapter the mechanisms that protect againsthypoglycaemia in healthy individuals and the physiological consequences of low glucoseconcentrations are discussed.NORMAL GLUCOSE HOMEOSTASISHumans evolved as hunter-gatherers and, unlike people today, did not consume regularmeals. Mechanisms therefore evolved for the body to store food when it was in abundance,and to use these stores to provide an adequate supply of energy, in particular in the form ofglucose when food was scarce. Cahill (1971) originally described the ‘rules of the metabolicgame’ which humans had to follow to ensure their survival. These rules were modified byTattersall (personal communication) and are as follows:1. Maintain glucose within very narrow limits.2. Maintain an emergency energy source (glycogen) which can be tapped quickly for fleeingor fighting.3. Waste not want not, i.e., store (fat and protein) in times of plenty.4. Use every trick in the book to maintain protein reserves.Hypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

2 NORMAL GLUCOSE METABOLISM AND RESPONSESInsulin and glucagon are the two hormones controlling glucose homeostasis, and thereforethe mechanisms enabling the ‘rules’ to be followed. The most important processes governedby these hormones are:• Glycogen synthesis and breakdown (glycogenolysis): Glycogen, a carbohydrate, is anenergy source stored in the liver and skeletal muscle. Liver glycogen is broken downto provide glucose for all tissues, whereas the breakdown of muscle glycogen results inlactate formation.• Gluconeogenesis: This is the production of glucose in the liver from precursors: glycerol,lactate and amino acids (in particular alanine). The process can also occur in the kidneys,but this site is not important under physiological conditions.• Glucose uptake and metabolism (glycolysis) by skeletal muscle and adipose tissue.The actions of insulin and glucagon are summarised in Boxes 1.1 and 1.2. Insulin is ananabolic hormone, reducing glucose output by the liver (hepatic glucose output), increasingthe uptake of glucose by muscle and adipose tissue (increasing peripheral uptake) andincreasing protein and fat formation. Glucagon opposes the actions of insulin in the liver.Thus insulin tends to reduce, and glucagon to increase, blood glucose concentrations.Box 1.1Actions of insulinLiver↑ Glycogen synthesis (↑ glycogen synthetase activity)↑ Glycolysis↑ Lipid formation↑ Protein formation↓ Glycogenolysis (↓ phosphorylase activity)↓ Gluconeogenesis↓ Ketone formationMuscle↑ Uptake of glucoseamino acidsketonepotassium↑ Glycolysis↑ Synthesis of glycogenprotein↓ Protein catabolism↓ Release of amino acidsAdipose tissue↑ Uptake glucosepotassiumStorage of triglyceride

NORMAL GLUCOSE HOMEOSTASIS 3Box 1.2Actions of glucagonLiver↑ Glycogenolysis↑ Gluconeogenesis↑ Extraction of alanine↑ KetogenesisNo significant peripheral actionThe metabolic effects of insulin and glucagon and their relationship to glucose homeostasisare best considered in relationship to fasting and the postprandial state (Siegal and Kreisberg,1975). In both these situations it is the relative and not absolute concentrations of thesehormones that are important.Fasting (Figure 1.1a)During fasting, insulin concentrations are reduced and glucagon increased, which maintainsblood glucose concentrations in accordance with rule 1 above. The net effect is toreduce peripheral glucose utilisation, to increase hepatic glucose production and to providenon-glucose fuels for tissues not entirely dependent on glucose. After a short (for exampleovernight) fast, glucose production needs to be 5–6 g/h to maintain blood glucose concentrations,with the brain using 80% of this. Glycogenolysis provides 60–80% and gluconeogenesis20–40% of the required glucose. In prolonged fasts, glycogen becomes depletedand glucose production is primarily from gluconeogenesis, with an increasing proportionfrom the kidney compared to the liver. In extreme situations renal gluconeogenesis cancontribute as much as 45% of glucose production. Thus glycogen is the short term or ‘emergency’fuel source (rule 2), with gluconeogenesis predominating during more prolongedfasts. The following metabolic alterations enable this increase in glucose productionto occur:• Muscle: Glucose uptake and oxidative metabolism are reduced and fatty acid oxidationincreased. Amino acids are released.• Adipose tissue: There are reductions in glucose uptake and triglyceride storage. Theincrease in the activity of the enzyme hormone-sensitive lipase results in hydrolysisof triglyceride to glycerol (a gluconeogenic precursor) and fatty acids, which can bemetabolised.• Liver: Increased cAMP concentrations result in increased glycogenolysis and gluconeogenesisthus increasing hepatic glucose output. The uptake of gluconeogenic precursors(i.e. amino acids, glycerol, lactate and pyruvate) is also increased. Ketone bodies areproduced in the liver from fatty acids. This process is normally inhibited by insulin andstimulated by glucagon, thus the hormonal changes during fasting lead to an increase inketone production. Fatty acids are also a metabolic fuel used by the liver and provide asource of energy for the reactions involved in gluconeogenesis.

4 NORMAL GLUCOSE METABOLISM AND RESPONSESglucagon, insulin Insulin, glucagon(a)FASTING(b)POST PRANDIALlactateGlycogenGlucoseGlycogenAmino acidsProteinMuscleAmino acidsProteinTGGlucoseFAGlycerolphosphateFAGlycerolphosphateFree FAGlycerolFree FATriglycerideTriglycerideAdipose TissueAmino acidslactateGlycogenKetonesKetonesGlycogenGlucoseFAGlycerolFFA, TG,lipoproteinGlucoseFAGlycerolHepatic glucoseoutputLiverFFA, TG,Hepatic glucoseoutputFigure 1.1 Metabolic pathways for glucose homeostasis in muscle, adipose tissue and liver duringfasting (left) and postprandially (right). FA = fatty acids; TG = triglyceride (associated CO 2 productionexcluded for clarity)The reduced insulin : glucagon ratio favours a catabolic state, but the effect on fatmetabolism is greater than protein, and thus muscle is relatively preserved (rule 4). Theseadaptations meant that not only did hunter-gatherers have sufficient muscle power topursue their next meal, but also that brain function was optimally maintained to help themdo this.

Fed state (Figure 1.1b)EFFECTS OF GLUCOSE DEPRIVATION 5In the fed state, in accordance with the rules of the metabolic game, excess food is stored asglycogen, protein and fat (rule 3). The rise in glucose concentrations results in an increasein insulin and reduction in glucagon secretion. This balance favours glucose utilisation,reduction of glucose production and increases glycogen, triglyceride and protein formation.The following changes enable these processes to occur:• Muscle: Insulin increases glucose transport, oxidative metabolism and glycogen synthesis.Amino acid release is inhibited and protein synthesis is increased.• Adipose tissue: In the fat cells, glucose transport is increased, while lipolysis is inhibited.At the same time the enzyme lipoprotein lipase, located in the capillaries, is activatedand causes triglyceride to be broken down to fatty acids and glycerol. The fatty acidsare taken up into the fat cells and re-esterified to triglyceride (using glycerol phosphatederived from glucose) before being stored.• Liver: Glucose uptake is increased in proportion to plasma glucose, a process which doesnot need insulin. However, insulin does decrease cAMP concentrations, which results in anincrease in glycogen synthesis and the inhibition of glycogenolysis and gluconeogenesis.These effects ‘retain’ glucose in the liver and reduce hepatic glucose output.This complex interplay between insulin and glucagon maintains euglycaemia and enablesthe rules of the metabolic game to be followed, ensuring not only the survival of thehunter-gatherer, but also of modern humans.EFFECTS OF GLUCOSE DEPRIVATION ON CENTRAL NERVOUSSYSTEM METABOLISMThe brain constitutes only 2% of body weight, but consumes 20% of the body’s oxygenand receives 15% of its cardiac output (Sokaloff, 1989). It is almost totally dependenton carbohydrate as a fuel and since it cannot store or synthesise glucose, depends ona continuous supply from circulating blood. The brain contains the enzymes needed tometabolise fuels other than glucose such as lactate, ketones and amino acids, but underphysiological conditions their use is limited by insufficient quantities in the blood or slowrates of transport across the blood-brain barrier. When arterial blood glucose falls below3 mmol/l, cerebral metabolism and function decline.Metabolism of glucose by the brain releases energy, and also generates neurotransmitterssuch as gamma amino butyric acid (GABA) and acetylcholine, together with phospholipidsneeded for cell membrane synthesis. When blood glucose concentration falls, changes inthe synthesis of these products may occur within minutes because of reduced glucosemetabolism, which can alter cerebral function. This is likely to be a factor in producingthe subtle changes in cerebral function detectable at blood glucose concentrations as highas 3 mmol/l, which is not sufficiently low to cause a major depletion in ATP or creatinephosphate, the brain’s two main sources of energy (McCall, 1993).

6 NORMAL GLUCOSE METABOLISM AND RESPONSESIsotope techniques and Positron Emission Tomography (PET) allow the study ofmetabolism in different parts of the brain and show regional variations in metabolismduring hypoglycaemia. The neocortex, hippocampus, hypothalamus and cerebellum are mostsensitive to hypoglycaemia, whereas metabolism is relatively preserved in the thalamusand brainstem. Changes in cerebral function are initially reversible, but during prolongedsevere hypoglycaemia, general energy failure (due to the depletion of ATP and creatinephosphate) can cause permanent cerebral damage. Pathologically this is caused by selectiveneuronal necrosis most likely due to ‘excitotoxin’ damage. Local energy failure induces theintrasynaptic release of glutamate or aspartate, and failure of reuptake of the neurotransmittersincreases their concentrations. This leads to the activation of N-methyl-D-aspartate(NMDA) receptors causing cerebral damage. One study in rats has shown that an experimentalcompound called AP7, which blocks the NMDA receptor, can prevent 90% of thecerebral damage associated with severe hypoglycaemia (Wieloch, 1985). In humans withfatal hypoglycaemia, protracted neuroglycopenia causes laminar necrosis in the cerebralcortex and diffuse demyelination. Regional differences in neuronal necrosis are seen, withthe basal ganglia and hippocampus being sensitive, but the hypothalamus and cerebellumbeing relatively spared (Auer and Siesjö, 1988; Sieber and Traysman, 1992).The brain is very sensitive to acute hypoglycaemia, but can adapt to chronic fuel deprivation.For example, during starvation, it can metabolise ketones for up to 60% of itsenergy requirements (Owen et al., 1967). Glucose transport can also be increased in theface of hypoglycaemia. Normally, glucose is transported into tissues using proteins calledglucose transporters (GLUT) (Bell et al., 1990). This transport occurs down a concentrationgradient faster than it would by simple diffusion and does not require energy (facilitateddiffusion). There are several of these transporters, with GLUT 1 being responsible fortransporting glucose across the blood-brain barrier and GLUT 3 for transporting glucoseinto neurones (Figure 1.2). Chronic hypoglycaemia in animals (McCall et al., 1986) andin humans (Boyle et al., 1995) increases cerebral glucose uptake, which is thought to bepromoted by an increase in the production and action of GLUT 1 protein. It has not beenFigure 1.2Transport of glucose into the brain across the blood–brain barrier

COUNTERREGULATION DURING HYPOGLYCAEMIA 7established whether this adaptation is of major benefit in protecting brain function duringhypoglycaemia.COUNTERREGULATION DURING HYPOGLYCAEMIAThe potentially serious effects of hypoglycaemia on cerebral function mean that not onlyare stable blood glucose concentrations maintained under physiological conditions, but alsoif hypoglycaemia occurs, mechanisms have developed to combat it. In clinical practice, theprincipal causes of hypoglycaemia are iatrogenic (as side-effects of insulin and sulphonylureasused to treat diabetes) and excessive alcohol consumption. Insulin secreting tumours(such as insulinoma) are rare. The mechanisms that correct hypoglycaemia are called counterregulation,because the hormones involved oppose the action of insulin and therefore arethe counterregulatory hormones. The processes of counterregulation were identified in themid 1970s and early 1980s, using either a bolus injection or continuous infusion of insulinto induce hypoglycaemia (Cryer, 1981; Gerich, 1988). The response to the bolus injectionof 0.1 U/kg insulin in a normal subject is shown in Figure 1.3. Blood glucose concentrationsdecline within minutes of the administration of insulin and reach a nadir after 20–30 minutes,then gradually rise to near normal by two hours after the insulin was administered. The factFigure 1.3 (a) Glucose and (b) insulin concentrations after intravenous injection of insulin 0.1 U/kgat time 0. Reproduced from Garber et al. (1976) by permission of the Journal of Clinical Investigation

8 NORMAL GLUCOSE METABOLISM AND RESPONSESthat blood glucose starts to rise when plasma insulin concentrations are still ten times thebaseline values means that it is not simply the reduction in insulin that reverses hypoglycaemia,but active counterregulation must also occur. Many hormones are released whenblood glucose is lowered (see below), but glucagon, the catecholamines, growth hormoneand cortisol are regarded as being the most important.Several studies have determined the relative importance of these hormones by producingisolated deficiencies of each hormone (by blocking its release or action) and assessingthe subsequent response to administration of insulin. These studies are exemplified inFigure 1.4 which assesses the relative importance of glucagon, adrenaline (epinephrine) andgrowth hormone in the counterregulation of short term hypoglycaemia. Somatostatin infusionblocks glucagon and growth hormone secretion and significantly impairs glucose recovery(Figure 1.4a). If growth hormone is replaced in the same model to produce isolated glucagondeficiency (Figure 1.4b), and glucagon replaced to produce isolated growth hormone deficiency(Figure 1.4c), it is clear that it is glucagon and not growth hormone that is responsiblefor acute counterregulation. Combined alpha and beta adrenoceptor blockade using phentolamineand propranolol infusions or adrenalectomy (Figure 1.4d), can be used to evaluatethe role of the catecholamines. These and other studies demonstrate that glucagon is themost important counterregulatory hormone whereas catecholamines provide a backup ifglucagon is deficient (for example in type 1 diabetes, see Chapters 6 and 7). Cortisol andFigure 1.4 Glucose recovery from acute hypoglycaemia. Glucose concentration following an intravenousinjection of insulin of 0.05 U/kg at time 0; after (a) saline infusion (continuous line) andsomatostatin, (b) somatostatin and growth hormone (GH), (c) somatostatin and glucagon, (d) combinedalpha and beta blockade with phentolamine and propranolol infusions or adrenalectomy, (e) somatostatinwith alpha and beta blockade, and (f) somatostatin in adrenalectomised patients. Saline infusion= continuous lines; experimental study = broken lines. Reproduced from Cryer (1981) courtesyof the American Diabetes Association (epinephrine = adrenaline)

COUNTERREGULATION DURING HYPOGLYCAEMIA 9growth hormone are important only in prolonged hypoglycaemia. Therefore if glucagonand catecholamines are both deficient, as in longstanding type 1 diabetes, counterregulationis seriously compromised, and the individual is defenceless against acute hypoglycaemia(Cryer, 1981).Glucagon and catecholamines increase glycogenolysis and stimulate gluconeogenesis.Catecholamines also reduce glucose utilisation peripherally and inhibit insulin secretion.Cortisol and growth hormone increase gluconeogenesis and reduce glucose utilisation. Therole of the other hormones (see below) in counterregulation is unclear, but they are unlikelyto make a significant contribution. Finally, there is evidence that during profound hypoglycaemia(blood glucose below 1.7 mmol/l), hepatic glucose output is stimulated directly,although the mechanism is unknown. This is termed hepatic autoregulation.The depth, as well as the duration, of hypoglycaemia is important in determining themagnitude of the counterregulatory hormone response. Studies using ‘hyperinsulinaemicclamps’ show a hierarchical response of hormone production. In this technique, insulin isinfused at a constant rate and a glucose infusion rate varied to maintain blood glucoseconcentrations within ±02 mmol/l of target concentrations. This permits the controlledevaluation of the counterregulatory hormone response at varying degrees of hypoglycaemia.It also demonstrates that glucagon, catecholamines and growth hormone start to be secretedat a blood glucose concentration of 3.5–3.7 mmol/l, with cortisol produced at a lower glucoseof 3.0 mmol/l (Mitrakou et al., 1991). The counterregulatory response is initiated beforeimpairment in cerebral function commences, usually at a blood glucose concentration ofapproximately 3.0 mmol/l (Heller and Macdonald, 1996).The magnitude of the hormonal response also depends on the length of the hypoglycaemicepisode. The counterregulatory hormonal response commences up to 20 minutesafter hypoglycaemia is achieved and continues to rise for 60 minutes (Kerr et al., 1989).In contrast, this response is attenuated as a result of a previous episode of hypoglycaemia(within a few days) (reviewed by Heller and Macdonald, 1996) and even by prolongedexercise the day before hypoglycaemia is induced. Galassetti et al. (2001) showed that innon-diabetic subjects three hours of moderate intensity exercise the previous day markedlydecreased the counterregulatory response to hypoglycaemia induced by the infusion ofinsulin, and that the reduced counterregulatory response was more marked in men thanin women.Although the primary role of the counterregulatory hormones is on glucose metabolism,any effects on fatty acid utilisation can have an indirect effect on blood glucose. Thus,the increase in plasma epinephrine (adrenaline) (and activation of the sympathetic nervoussystem) that is seen in hypoglycaemia can stimulate lipolysis of triglyceride in adipose tissueand muscle and release fatty acids which can be used as an alternative fuel to glucose, makingmore glucose available for the CNS. Enoksson et al. (2003) demonstrated that patients withtype 1 diabetes, who had lower plasma epinephrine responses to hypoglycaemia than nondiabeticcontrols, also had reduced rates of lipolysis in adipose tissue and skeletal muscle,making them more dependent on glucose as a fuel and therefore at risk of developing a moresevere hypoglycaemia.The complex counterregulatory and homeostatic mechanisms described above are thoughtto be mostly under the control of the central nervous system. Evidence for this comesfrom studies in dogs, where glucose was infused into the carotid and vertebral arteries tomaintain euglycaemia in the brain. Despite peripheral hypoglycaemia, glucagon did notincrease and responses of the other counterregulatory hormones were blunted. This, and

10 NORMAL GLUCOSE METABOLISM AND RESPONSESother studies in rats, led to the hypothesis that the ventromedial nucleus of the hypothalamus(VMH), which does not have a blood–brain barrier, acts as a glucose-sensor and co-ordinatescounterregulation (Borg et al., 1997). However, evidence exists that other parts of the brainmay also be involved in mediating counterregulation.It is now clear that glucose-sensing neurones can involve either glucokinase or ATPsensitiveK + channels (Levin et al., 2004). In rats, the VMH has ATP-sensitive K + channelswhich seem to be involved in the counterregulatory responses to hypoglycaemia, as injectionof the sulphonylurea, glibenclamide, directly into the VMH suppressed hormonal responsesto systemic hypoglycaemia (Evans et al., 2004).The existence of hepatic autoregulation suggests that some peripheral control should exist.Studies producing central euglycaemia and hepatic portal venous hypoglycaemia in dogshave provided evidence for hepatic glucose sensors and suggest that these sensors, as wellas those in the brain, are important in the regulation of glucose (Hamilton-Wessler et al.,1994). However, this topic is somewhat controversial and more recent studies on dogs havefailed to demonstrate an effect of hepatic sensory nerves on the responses to hypoglycaemia(Jackson et al., 2000). Moreover, studies in humans by Heptulla et al. (2001) showedthat providing glucose orally rather than intravenously during a hypoglycaemic hyperinsulinaemicclamp actually enhanced the counterregulatory hormone responses rather thanreduced them.HORMONAL CHANGES DURING HYPOGLYCAEMIAHypoglycaemia induces the secretion of various hormones, some of which are responsible forcounterregulation, many of the physiological changes that occur as a consequence of loweringblood glucose and contribute to symptom generation (see Chapter 2), The stimulation of theautonomic nervous system is central to many of these changes.Activation of the Autonomic Nervous SystemThe autonomic nervous system comprises sympathetic and parasympathetic components(Figure 1.5). Fibres from the sympathetic division leave the spinal cord with the ventral rootsfrom the first thoracic to the third or fourth lumbar nerves to synapse in the sympatheticchain or visceral ganglia, and the long postganglionic fibres are incorporated in somaticnerves. The parasympathetic pathways originate in the nuclei of cranial nerves III, VII, IXand X, and travel with the vagus nerve. A second component, the sacral outflow, suppliesthe pelvic viscera via the pelvic branches of the second to fourth spinal nerves. The gangliain both cases are located near the organs supplied, and the postganglionic neurones aretherefore short.Selective activation of both components of the autonomic system occurs during hypoglycaemia.The sympathetic nervous system in particular is responsible for many of thephysiological changes during hypoglycaemia and the evidence for its activation can beobtained indirectly by observing functional changes such as cardiovascular responses (consideredbelow), measuring plasma catecholamines which gives a general index of sympatheticactivation, or by directly recording sympathetic activity.

HORMONAL CHANGES DURING HYPOGLYCAEMIA 11Figure 1.5 Anatomy of the autonomic nervous system. Pre = preganglionic neurones; post = postganglionicneurones; RC = ramus communicansDirect recordings are possible from sympathetic nerves supplying skeletal muscle and skin.Sympathetic neural activity in skeletal muscle involves vasoconstrictor fibres which innervateblood vessels and are involved in controlling blood pressure. During hypoglycaemia (inducedby insulin), the frequency and amplitude of muscle sympathetic activity are increased asblood glucose falls, with an increase in activity eight minutes after insulin is injectedintravenously, peaking at 25–30 minutes coincident with the glucose nadir, and persisting

12 NORMAL GLUCOSE METABOLISM AND RESPONSESFigure 1.6 (a) Muscle sympathetic activity during euglycaemia and hypoglycaemia. Reproducedfrom Fagius et al. (1986) courtesy of the American Diabetes Association. (b) Skin sympatheticactivity during euglycaemia and hypoglycaemia. Reproduced from Berne and Fagius (1986), with kindpermission from Springer Science and Business Mediafor 90 minutes after euglycaemia is restored (Figure 1.6a) (Fagius et al., 1986). Duringhypoglycaemia, a sudden increase in skin sympathetic activity is seen, which coincideswith the onset of sweating. This sweating leads to vasodilatation of skin blood vessels,which is also contributed to by a reduction in sympathetic stimulation of the vasoconstrictorcomponents of skin arterio-venous anastomoses (Figure 1.6b) (Berne and Fagius, 1986).These effects (at least initially) increase total skin blood flow and promote heat loss fromthe body.Activation of both muscle and skin sympathetic nerve activity are thought to be centrallymediated. Tissue neuroglycopenia can be produced by 2-deoxy-D-glucose, a glucoseanalogue, without increasing insulin. Infusion of this analogue causes stimulation of muscleand skin sympathetic activity demonstrating that it is the hypoglycaemia per se, and not theinsulin used to induce it, which is responsible for the sympathetic activation (Fagius andBerne, 1989).The activation of the parasympathetic nervous system (vagus nerve) during hypoglycaemiacannot be measured directly. The most useful index of parasympathetic function is themeasurement of plasma pancreatic polypeptide, the peptide hormone secreted by the PP cellsof the pancreas, which is released in response to vagal stimulation.

HORMONAL CHANGES DURING HYPOGLYCAEMIA 13Neuroendocrine Activation (Box 1.3)Insulin-induced hypoglycaemia was used to study pituitary function as early as the 1940s.The development of assays for adrenocorticotrophic hormone (ACTH) and growth hormone(GH) allowed the direct measurement of pituitary function during hypoglycaemia in the1960s, and many of the processes governing these changes were unravelled before elucidationof the counterregulatory system. The studies are comparable to those evaluatingcounterregulation, in that potential regulatory factors are blocked to measure the hormonalresponse to hypoglycaemia with and without the regulating factor.Box 1.3Neuroendocrine activationHypothalamus ↑ Corticotrophic releasing hormone↑ Growth hormone releasing hormoneAnterior Pituitary ↑ Adrenocorticotrophic hormone↑ Beta endorphin↑ Growth hormone↑ Prolactin↔ Thyrotrophin↔ GonadotrophinsPosterior pituitary ↑ Vasopressin↑ OxytocinPancreas ↑ Glucagon↑ Pancreatic polypeptide↑ InsulinAdrenal ↑ Cortisol↑ Epinephrine (adrenaline)↑ AldosteroneOthers ↑ Parathyroid hormone↑ Gastrin↑ Somatostatin (28)Hypothalamus and anterior pituitaryACTH, GH and prolactin concentrations increase during hypoglycaemia, but there is nochange in thyrotrophin or gonadotrophin secretion. The secretion of these pituitary hormonesis controlled by releasing factors which are produced in the median eminence of the hypothalamus,secreted into the hypophyseal portal vessels and then pass to the pituitary gland(Figure 1.7). The mechanisms regulating the releasing factors are incompletely understood,but may involve the ventromedial nucleus, one site where brain glucose sensors are situated(Fish et al., 1986).

14 NORMAL GLUCOSE METABOLISM AND RESPONSESFigure 1.7Anatomy of the hypothalamus and pituitary gland• ACTH: Secretion is governed by release of corticotrophin releasing hormone (CRH) fromthe hypothalamus; alpha adrenoceptors stimulate CRH release, and beta adrenoceptorshave an inhibitory action. A variety of neurotransmitters control the release of CRH intothe portal vessels, including serotonin and acetylcholine which are stimulatory and GABAwhich is inhibitory. The increase in ACTH causes cortisol to be secreted from the corticesof the adrenal glands.• Beta endorphins are derived from the same precursors as ACTH and are co-secreted withit. The role of endorphins in counterregulation is uncertain, but they may influence thesecretion of the other pituitary hormones during hypoglycaemia.• GH: Growth hormone secretion is governed by two hypothalamic hormones: growthhormone releasing hormone (GHRH) which stimulates GH secretion, and somatostatinwhich is inhibitory. GHRH secretion is stimulated by dopamine, GABA, opiates andthrough alpha adrenoceptors, whereas it is inhibited by serotonin and beta adrenoceptors.A study in rats showed that bioassayable GH and GHRH are depleted in the pituitary andhypothalamus respectively after insulin-induced hypoglycaemia (Katz et al., 1967).• Prolactin: The mechanisms underlying its secretion are not established. Prolactin secretionis normally under the inhibitory control of dopamine, but evidence also exists forreleasing factors being produced during hypoglycaemia. Prolactin does not contribute tocounterregulation.Posterior pituitaryVasopressin and oxytocin both increase during hypoglycaemia (Fisher et al., 1987). Theirsecretion is under hormonal and neurotransmitter control in a similar way to the hypothalamic

PHYSIOLOGICAL RESPONSES 15hormones. Vasopressin has glycolytic actions and oxytocin increases hepatic glucose outputin dogs, but their contribution to glucose counterregulation is uncertain.Pancreas• Glucagon: The mechanisms of glucagon secretion during hypoglycaemia are still not fullyunderstood. Although activation of the autonomic nervous system stimulates its release,this pathway has been shown to be less important in humans. A reduction in glucoseconcentrations may have a direct effect on the glucagon-secreting pancreatic alpha cells,or the reduced beta cell activity (reduced insulin secretion), which also occurs with lowblood glucose, may release the tonic inhibition of glucagon secretion. However, suchmechanisms would be disturbed in type 1 diabetes, where hypoglycaemia is normallyassociated with high plasma insulin levels and there is no direct effect of beta cell-derivedinsulin on the alpha cells.• Somatostatin: This is thought of as a pancreatic hormone produced from D cells of theislets of Langerhans, but it is also secreted in other parts of the gastrointestinal tract. Thereare a number of structurally different polypeptides derived from prosomatostatin: thesomatostatin-14 peptide is secreted from D cells, and somatostatin-28 from the gastrointestinaltract. The plasma concentration of somatostatin-28 increases during hypoglycaemia(Francis and Ensinck, 1987). The normal action of somatostatin is to inhibit the secretionboth of insulin and glucagon, but somatostatin-28 inhibits insulin ten times more effectivelythan glucagon, and thus may have a role in counterregulation by suppressing insulinrelease.• Pancreatic polypeptide: This peptide has no known role in counterregulation, but its releaseduring hypoglycaemia is stimulated by cholinergic fibres through muscarinic receptorsand is a useful marker of parasympathetic activity.Adrenal and Renin–Angiotensin systemThe processes governing the increase in cortisol during hypoglycaemia are discussed above.The rise in catecholamines, in particular epinephrine from the adrenal medulla, which occurswhen blood glucose is lowered, is controlled by sympathetic fibres in the splanchnic nerve.The increase in renin, and therefore angiotensin and aldosterone, during hypoglycaemia isstimulated primarily by the intra-renal effects of increased catecholamines, mediated throughbeta adrenoceptors, although the increase in ACTH and hypokalaemia due to hypoglycaemiacontributes (Trovati et al., 1988; Jungman et al., 1989). These changes do not have asignificant role in counterregulation, although angiotensin II has glycolytic actions in vitro.PHYSIOLOGICAL RESPONSESHaemodynamic Changes (Box 1.4)The haemodynamic changes during hypoglycaemia (Hilsted, 1993) are mostly caused bythe activation of the sympathetic nervous system and an increase in circulating epinephrine.

16 NORMAL GLUCOSE METABOLISM AND RESPONSESBox 1.4Haemodynamic changes↑↑↑↓↑Heart rateSystolic blood pressureCardiac outputPeripheral resistanceMyocardial contractilityAn increase in heart rate (tachycardia), myocardial contractility and cardiac output occurs,which is mediated through beta 1 adrenoceptors, but increasing vagal tone counteracts thiseffect so the increase is transient. Peripheral resistance, estimated from mean arterial pressuredivided by cardiac output, is reduced. A combination of the increase in cardiac output andreduction in peripheral resistance results in an increase in systolic and a decrease in diastolicpressure, i.e. a widening of pulse pressure without a change in mean arterial pressure.Changes in Regional Blood Flow (Box 1.5 and Figure 1.8)• Cerebral blood flow: Early work produced conflicting results, but these studies werein subjects receiving insulin shock therapy, and the varying effects of convulsions andaltered level of consciousness may have influenced the outcome. Subsequent studies haveconsistently shown an increase in cerebral blood flow during hypoglycaemia despite theuse of different methods of measurement (isotopic, single photon emission computedtomography (SPECT) and Doppler ultrasound). In most of the studies blood glucoseconcentration was less than 2 mmol/l before a change was observed. In animals, hypoglycaemiais associated with loss of cerebral autoregulation (the ability of the brain to maintaincerebral blood flow despite variability in cardiac output) through beta adrenoceptor stimulation,but the exact mechanisms are unknown (Bryan, 1990; Sieber and Traysman,1992).• Gastrointestinal system: Total splanchnic blood flow (supplying the intestines, liver,spleen and stomach) is increased and splanchnic vascular resistance reduced as assessedby the bromosulphthalein extraction technique (Bearn et al., 1952). Superior mesentericartery blood flow measured using Doppler ultrasound increases during hypoglycaemiadue to beta adrenoceptor stimulation (Braatvedt et al., 1993). Radioisotope scanning hasBox 1.5Changes in regional blood flow↑↑↓↑↓Cerebral flowTotal splanchnic flowSplenic flowSkin flow variable (early ↑, late ↓)Muscle flowRenal flow

PHYSIOLOGICAL RESPONSES 17Figure 1.8Changes in regional blood flow during hypoglycaemiademonstrated a reduction in splenic activity during hypoglycaemia (Fisher et al., 1990),which is thought to be a consequence of alpha adrenoceptor-mediated reduction in bloodflow. These changes would all be expected to increase hepatic blood flow.• Skin: The control of blood flow to the skin is complex and different mechanisms predominatein different areas. Studies of the effect of hypoglycaemia on skin blood flow areinconsistent partly because different methods have been used for blood flow measurementand induction of hypoglycaemia, as well as differences in the part of the bodystudied. Definitive conclusions are therefore not possible. Studies using the dorsum ofthe foot and the face (cheek and forehead) have consistently shown an initial vasodilatationand increase in blood flow followed by later vasoconstriction at a blood glucose of2.5 mmol/l (Maggs et al., 1994). These findings are consistent with the clinical pictureof initial flushing and later pallor, with an early rise in skin blood flow followed by alater fall.• Muscle blood flow: A variety of techniques have been used to study muscle blood flow(including venous occlusion plethysmography, isotopic clearance techniques and the useof thermal conductivity meters). All studies have consistently shown an increase in muscle

18 NORMAL GLUCOSE METABOLISM AND RESPONSESblood flow during hypoglycaemia irrespective of skin blood flow. This change is mediatedby beta 2 adrenoceptors (Abramson et al., 1966; Allwood et al., 1959).• Kidney: Inulin and sodium hippurate clearance can be used to estimate glomerular filtrationrate and renal blood flow respectively. Both decrease during hypoglycaemia (Patrick et al.,1989) and catecholamines and renin are implicated in initiating the changes.The changes in blood flow in various organs, like the haemodynamic changes, are mostlymediated by the activation of the sympathetic nervous system or circulating epinephrine. Themajority either protect against hypoglycaemia or increase substrate delivery to vital organs.The increase in cerebral blood flow increases substrate delivery to the brain. Increasingmuscle flow enhances the release and washout of gluconeogenic precursors. The increase insplanchnic blood flow and reduction in splenic blood flow serve to increase hepatic bloodflow to maximise hepatic glucose production. Meanwhile, blood is diverted away fromorgans such as the kidney and spleen, which are not required in the acute response to themetabolic stress.Functional Changes (Box 1.6)• Sweating: Sweating is mediated by sympathetic cholinergic nerves, although other neurotransmitterssuch as vasoactive intestinal peptide and bradykinin may also be involved.The activation of the sympathetic innervation of the skin as described above results inthe sudden onset of sweating. Sweating is one of the first physiological responses tooccur during hypoglycaemia and can be demonstrated within ten minutes of achievinga blood glucose of 2.5 mmol/l (Maggs et al., 1994). It coincides with the onset ofother measures of autonomic activation such as an increase in heart rate and tremor(Figure 1.9).• Tremor: Trembling and shaking are characteristic features of hypoglycaemia and resultfrom an increase in physiological tremor. The rise in cardiac output and vasodilatationoccurring during hypoglycaemia increase the level of physiological tremor and thisis exacerbated by beta adrenoceptor stimulation associated with increased epinephrineconcentrations (Kerr et al., 1990). Since adrenalectomy does not entirely abolish tremor,other components such as the activation of muscle sympathetic activity must beinvolved.Box 1.6Functional changes↑↑↓↓↑↑Sweating (sudden onset)TremorCore temperatureIntraocular pressureJejunal activityGastric emptying

PHYSIOLOGICAL RESPONSES 19We do not have rights to reproduce thisfigure electronicallyFigure 1.9 Sudden onset of sweating, tremor and increase in heart rate during the induction of hypoglycaemia.Reproduced from Hypoglycaemia and Diabetes: Clinical and Physiological Aspects (edsB.M. Frier and M. Fisher), © 1993 Edward Arnold, by permission of Edward Arnold (Publishers) Ltd• Temperature: Despite a beta adrenoceptor-mediated increase in metabolic rate, coretemperature falls during hypoglycaemia. The mechanisms by which this occurs dependon whether the environment is warm or cold. In a warm environment, heat is lost becauseof sweating and increased heat conduction from vasodilatation. Hypoglycaemia reducescore temperature by 03 C and skin temperature up to 2 C (depending on the part of thebody measured) after 60 minutes (Maggs et al., 1994). Shivering is reduced in the cold,and together with vasodilatation and sweating this causes a substantial reduction in core

20 NORMAL GLUCOSE METABOLISM AND RESPONSEStemperature (Gale et al., 1983). In rats, mortality was increased in animals whose coretemperature was prevented from falling during hypoglycaemia (Buchanan et al., 1991).In humans there is anecdotal evidence from subjects undergoing insulin shock therapythat those who had a rise in body temperature showed delayed neurological recovery(Ramos et al., 1968). These findings support the hypothesis that the fall in core temperaturereduces metabolic rate, allowing hypoglycaemia to be better tolerated, and thus thechanges in body temperature are of survival value. The beneficial effects are likely tobe limited, particularly in a cold environment, where the impairment of cerebral functionmeans subjects may not realise they are cold, causing them to be at risk of severehypothermia.• Other functional changes include a reduction in intraocular pressure, greater jejunal butnot gastric motility and inconsistent abnormalities of liver function tests. An increasein gastric emptying occurs during hypoglycaemia (Schvarcz et al., 1995), which maybe protective in that carbohydrate delivery to the intestine is increased, enabling fasterglucose absorption and reversal of hypoglycaemia.CONCLUSIONS• Homeostatic mechanisms exist to maintain glucose concentration within narrow limitsdespite a wide variety of circumstances.• The dependence of the central nervous system on glucose has led to a complex series ofbiochemical, functional and haemodynamic changes aimed at restoring glucose concentrations,producing symptoms and protecting the body in general, and central nervous systemin particular, against the effects of a low blood glucose (Figure 1.10).• Many symptoms of hypoglycaemia result from the activation of the autonomic nervoussystem and help to warn the individual that blood glucose is low. This encourages theingestion of carbohydrate, so helping to restore glucose concentrations in addition tocounterregulation.• Faster gastric emptying and the changes in regional blood flow which also occur as aresult of the activation of the autonomic nervous system increase substrate delivery.• The greater cerebral blood flow increases glucose delivery to the brain (although lossof autoregulation is undesirable), and the increased splanchnic flow results in a greaterdelivery of gluconeogenic precursors to the liver.• Activation of the autonomic nervous system also increases sweating, and together with theinhibition of shivering, this predisposes to hypothermia, which may be neuroprotective.ACKNOWLEDGEMENTSWe would like to thank Professor Robert Tattersall for reading the chapter and for his helpfulsuggestions.

REFERENCES 21NORMAL GLUCOSE HOMEOSTASISINSULIN : GLUCAGON RATIOExcess InsulinHYPOGLYCAEMIADETECTION BY BRAIN GLUCOSE SENSORSPANCREASPITUITARYADRENALACTIVATION OFAUTONOMICNERVOUS SYSTEMGLUCAGON(GH and CORTISOL)CATECHOLAMINESCOUNTERREGULATIONGASTRIC EMPTYINGANDCHANGES IN BLOOD FLOWSWEATING(increased substrateproduction)SYMPTOMS(Autonomic)CEREBRALBLOOD FLOWSPLANCHNICBLOOD FLOW(eating)HYPOTHERMIARESTORE BLOODGLUCOSEINCREASE SUBSTRATEDELIVERYPROTECTION OFVITAL ORGANSFigure 1.10Glucose homeostasis and the correction of hypoglycaemiaREFERENCESAbramson EA, Arky RA, Woeber KA (1966). Effects of propranolol on the hormonal and metabolicresponses to insulin induced hypoglycaemia. Lancet ii: 1386–9.Allwood MJ, Hensel H, Papenberg J (1959). Muscle and skin blood flow in the human forearm duringinsulin hypoglycaemia. Journal of Physiology 147: 269–73.Auer RN, Siesjö BK (1988). Biological differences between ischaemia, hypoglycaemia and epilepsy.Annals of Neurology 24: 699–707.

22 NORMAL GLUCOSE METABOLISM AND RESPONSESBearn AG, Bulling BH, Sherlock S (1952). The response of the liver to insulin in normal subjects andin diabetes mellitus: hepatic vein catheterisation studies. Clinical Science 11: 151–64.Bell GI, Karyano T, Buse JB, Burant CF, Takeda T, Lin D et al. (1990). Molecular biology ofmammalian glucose transporters. Diabetes Care 13: 198–208.Berne C, Fagius J (1986). Skin sympathetic activity during insulin induced hypoglycaemia.Diabetologia 29: 855–60.Borg MA, Sherwin RS, Borg WP, Tamborlane WV, Shulman GI (1997). Local ventromedial hypothalamusglucose perfusion blocks counterregulation during systemic hypoglycemia in awake rats.Journal of Clinical Investigation 99: 361–5.Boyle PJ, Kempers SE, O’Conner AM, Nagy RJ (1995). Brain glucose uptake and unawarenessof hypoglycemia in patients with insulin-dependent diabetes mellitus. New England Journal ofMedicine 333: 1726–31.Braatvedt GD, Flynn MD, Stanners A, Halliwell M, Corrall RJM (1993). Splanchnic blood flow inman: evidence for mediation via a -adrenergic mechanism. Clinical Science 84: 201–7.Bryan RM (1990). Cerebral blood flow and energy metabolism during stress. American Journal ofPhysiology 259: H269–80.Buchanan TA, Cane P, Eng CC, Sipos GF, Lee C (1991). Hypothermia is critical for survival duringprolonged insulin induced hypoglycemia in rats. Metabolism 40: 330–4.Cahill GF (1971). Physiology of insulin in man. Diabetes 20: 785–99.Cryer PE (1981). Glucose counterregulation in man. Diabetes 30: 261–4.Enoksson S, Caprio SK, Rife F, Shulman GI, Tamborlane WV, Sherwin RS (2003). Defective activationof skeletal muscle and adipose tissue lipolysis in type 1 diabetes mellitus during hypoglycemia.Journal of Clinical Endocrinology and Metabolism 88: 1503–11.Evans ML, McCrimmon RJ, Flanagan DE, Keshavarz T, Fan X, McNay EC et al. (2004). HypothalamicATP-sensitive K + channels play a key role in sensing hypoglycemia and triggering counterregulatoryepinephrine and glucagon responses. Diabetes 53: 2542–51.Fagius J, Niklasson F, Berne C (1986). Sympathetic outflow in human muscle nerves increases duringhypoglycemia. Diabetes 35: 1124–9.Fagius J, Berne C (1989). Changes in sympathetic nerve activity induced by 2-deoxy-D-glucoseinfusion in humans. American Journal of Endocrinology 256: E714–21.Fish HR, Chernow B, O’Brian JT (1986). Endocrine and neurophysiologic responses of the pituitaryof insulin-induced hypoglycemia: a review. Metabolism 35: 763–80.Fisher M, Baylis PH, Frier BM (1987). Plasma oxytocin, arginine vasopressin and atrial natriureticpeptide response to insulin-induced hypoglycaemia in man. Clinical Endocrinology 26:179–85.Fisher M, Gillen G, Hepburn DA, Dargie HJ, Barnett E, Frier BM (1990). Splenic responses to acuteinsulin-induced hypoglycaemia in humans. Clinical Science 78: 469–74.Francis BH, Ensinck JW (1987). Differential alterations of the circulating prosomatostatin-derivedpeptides during insulin induced hypoglycemia in man. Journal of Clinical Endocrinology andMetabolism 65: 880–4.Gale EAM, Bennet J, Macdonald IA, Holst JJ, Mathews JA (1983). The physiological effects ofinsulin-induced hypoglycaemia in man: responses at differing levels of blood glucose. ClinicalScience 65: 262–71.Galassetti P, Neill AR, Tate D, Ertl AC, Wasserman DH, Davis SN (2001). Sexual dimorphism in counterregulatoryresponses to hypoglycemia after antecedent exercise. Journal of Clinical Endocrinologyand Metabolism 86: 3516–24.Garber AJ, Cryer PE, Santiago JV, Hammond MW, Pagliara AS, Kipnis DM (1976). The role ofadrenergic mechanisms in the substrate and hormonal response to insulin-induced hypoglycemia inman. Journal of Clinical Investigation 58: 7–15.Gerich JE (1988). Glucose counterregulation and its impact in diabetes mellitus. Diabetes 37:1608–17.

REFERENCES 23Hamilton-Wessler M, Bergman RN, Halter JB, Watanabe RM, Donovan CM (1994). The role ofliver glucosensors in the integrated sympathetic response induced by deep hypoglycemia in dogs.Diabetes 43: 1052–60.Heller SR, Macdonald IA (1996). The measurement of cognitive function during acute hypoglycaemia:experimental limitations and their effect on the study of hypoglycaemia unawareness. DiabeticMedicine 13: 607–15.Heptulla RA, Tamborlane WV, Ma TY-Z, Rife F, Sherwin RS (2001). Oral glucose augments thecounterregulatory hormone response during insulin-induced hypoglycemia in humans. Journal ofClinical Endocrinology and Metabolism 86: 645–8.Hilsted J (1993). Cardiovascular changes during hypoglycaemia. Clinical Physiology 13: 1–10.Jackson PA, Cardin S, Coffey CS, Neal DW, Allen EJ, Penazola AR et al. (2000). Effect of hepaticdenervation on the counterregulatory response to insulin-induced hypoglycemia in the dog. AmericanJournal of Endocrinology 279: E1249–57.Jungman E, Konzog C, Holl E, Fassibinder W, Schoffling K (1989). Effect of a human atrial natureticpeptide on blood glucose concentrations and hormone stimulation during insulin-induced hypoglycaemiain healthy man. European Journal of Clinical Pharmacology 36: 593–7.Katz SH, Dhariwal APS, McCann SM (1967). Effects of hypoglycaemia on the content of pituitarygrowth hormone (GH) and hypothalamic growth hormone releasing factor (GNRH) in the rat.Endocrinology 81: 333–9.Kerr D, Macdonald IA, Tattersall RB (1989). Influence of duration of hypoglycemia on the hormonalcounterregulatory response in normal subjects. Journal of Clinical Endocrinology and Metabolism68: 118–22.Kerr D, Macdonald IA, Heller SR, Tattersall RB (1990). A randomised double-blind placebo controlledtrial of the effects of Metoprolol CR, Atenolol and Propranolol LA on the physiological responsesto hypoglycaemia in the non-diabetic. British Journal of Clinical Pharmacology 29: 685–94.Levin BE, Routh VH, Kang L, Sanders NM, Dunn-Meynell AA (2004). Neuronal glucosensing: Whatdo we know after 50 years? Diabetes 53: 2521–8.Maggs DG, Scott AR, Macdonald IA (1994). Thermoregulatory responses to hyperinsulinemic hypoglycemiaand euglycemia in humans. American Journal of Physiology 267: R1266–72.McCall AL, Fixman LB, Fleming N, Tornheim K, Chick W, Ruderman ND (1986). Chronic hypoglycemiaincreases brain glucose transport. American Journal of Endocrinology 251: E442–5.McCall AL (1993). Effects of glucose deprivation on glucose metabolism in the central nervous system.In: Hypoglycaemia and Diabetes: Clinical and Physiological Aspects. Frier BM and Fisher M,eds. Edward Arnold, London: 56–71.Mitrakou A, Ryan C, Veneman T, Mokan M, Jenssen T, Kiss I et al. (1991). Hierarchy of glycemicthresholds for counterregulatory hormone secretion, symptoms and cerebral dysfunction. AmericanJournal of Endocrinology 266: E67–74.Owen OE, Morgan AP, Kemp HG, Sullivan JM, Herrara MG, Cahill GF Jr (1967). Brain metabolismduring fasting. Journal of Clinical Investigation 46: 1589–95.Patrick AW, Hepburn DA, Craig KJ, Thompson I, Swainson CD, Frier BM (1989). The effects of acuteinsulin-induced hypoglycaemia on renal function in normal human subjects. Diabetic Medicine 6:703–8.Ramos E, Zorilla E, Hadley WB (1968). Fever as a manifestation of hypoglycemia. Journal of theAmerican Medical Association 205: 590–2.Schvarcz E, Palmér M, Åman J, Berne C (1995). Hypoglycemia increases the gastric emptying ratein healthy subjects. Diabetes Care 18: 674–6.Sieber FE, Traysman RJ (1992). Special Issues: Glucose and the brain. Critical Care Medicine 20:104–14.Siegal AM, Kreisberg RA (1975). Metabolic homeostasis: insulin-glucagon interactions. In: DiabetesMellitus (4th edition). Sussman KE and Metz RJS, eds. American Diabetes Association, New York:29–35.

24 NORMAL GLUCOSE METABOLISM AND RESPONSESSokaloff (1989). Circulation and energy metabolism of the brain. In: Basic Neurochemistry. Siegel G,Agranoff B, Albers RW and Molinoff P, eds. Raven Press, New York: 565–90.Trovati M, Massucco P, Mularoni E, Cavalot F, Anfossi G, Matiello L, Emanuelli G (1988). Insulininducedhypoglycaemia increases plasma concentrations of angiotensin II and does not modify atrialnaturetic polypeptide secretion in man. Diabetologia 31: 816–20Wieloch T (1985). Hypoglycemia induced neuronal damage prevented by an N-Methyl D-Aspartateantagonist. Science 230: 681–3.

2 Symptoms of Hypoglycaemiaand Effects on MentalPerformance and EmotionsIan J. DearyINTRODUCTIONThis chapter describes the symptoms that are perceived during acute hypoglycaemia, andthe changes in mental functions and emotions that occur during this metabolic state.The most obvious benefit to a person of knowing about the symptoms of hypoglycaemiais the ability to recognise the onset of a hypoglycaemic episode as early as possible. This isof key importance in informing and educating people with diabetes. Moreover, if a personwith diabetes understands which mental functions are affected by hypoglycaemia he or shecan judge which activities may be most threatened in this state.SYMPTOMS OF HYPOGLYCAEMIAIdentifying the SymptomsThe physiological responses to hypoglycaemia are described in Chapter 1. The responseto hypoglycaemia results in physical symptoms, which raises several questions (McAulayet al., 2001b). Can we compile a comprehensive list of symptoms of hypoglycaemia? Whichare the more common symptoms of hypoglycaemia? Are there early warning symptoms ofhypoglycaemia? Do people differ in how quickly and accurately they detect or recognisehypoglycaemia? Do people differ in the set of symptoms of hypoglycaemia they experience?How can individuals distinguish the symptoms of hypoglycaemia from other bodily changes?The total symptom complexThe most basic question is: what symptoms do people report when they develophypoglycaemia?In humans the symptoms associated with hypoglycaemia were first recorded when insulinbecame available for the treatment of diabetes (Fletcher and Campbell, 1922). A list ofcharacteristic symptoms was described (Table 2.1). It was noted: that some symptomsHypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

Table 2.1 Common symptoms associated with hypoglycaemiaSymptoms associated withhypoglycaemia as derived frompopulation studies (afterHepburn, 1993)Percentage (minimum–maximum) ofpeople reporting the given symptomas associated with hypoglycaemia(after Hepburn, 1993)Symptoms of hypoglycaemiaas noted by Fletcher andCampbell (1922)Symptoms (and percentage [to nearest5%]) of people endorsing symptoms asassociated with hypoglycaemia; after Coxet al., 1993a, Figure 2)Sweating 47–84 Sweating Sweating (80)Trembling 32–78 Tremulousness Trembling (65)Weakness 28–71 Weakness Fatigue/weak (70)Visual disturbance 24–60 Diplopia Blurred vision (20)Hunger 39–49 Excessive hunger Hunger (60)Pounding heart 8–62 Change in pulse rate Pounding heart (55)Difficulty with speaking 7–41 Dysarthria, sensory and Slurred speech (40)motor aphasiaTingling around the mouth 10–39 – Numb lips (50)Dizziness 11–41 Vertigo, faintness,Light-headed/dizzy (60)syncopeHeadache 24–36 Headache (30)Anxiety 10–44 Nervousness, anxiety,excitement,emotional upsetNervous/tense (65)Nausea 5–20 – –Difficulty concentrating 31–75 – Difficulty concentrating (80)Tiredness 38–46 –Drowsiness 16–33 – Drowsy–sleepy (40)Confusion 13–53 Confusion,disorientation,‘goneness’PallorIncoordination Uncoordinated (75)Feeling of heat orcoldEmotional instabilityCold sweats (40)Slowed thinking (70)

SYMPTOMS OF HYPOGLYCAEMIA 27appeared before others during hypoglycaemia; that the blood glucose level at which subjectsbecame aware of hypoglycaemia was characteristic for the individual; that there were largeindividual differences in the levels of blood glucose at which awareness of hypoglycaemiacommenced; and that the preceding blood glucose concentration could affect the onset ofsymptoms.Lists of common symptoms of hypoglycaemia have been compiled from more recentresearch. Hepburn (1993) summarised eight population studies of the symptoms ofhypoglycaemia experienced by adults and children with insulin-treated diabetes, and Coxet al. (1993a) also produced a list of symptoms (Table 2.1). It is evident that the three listsof symptoms do not differ greatly, and that Fletcher and Campbell’s early report (1922)had captured many of the symptoms found in subsequent, more structured investigations.However, their report omitted to mention some symptoms such as tiredness, drowsiness anddifficulty concentrating, though it did include others – such as pallor (a sign rather thana symptom), incoordination and feelings of temperature change – that are emphasised byother researchers as regularly perceived symptoms. Table 2.1 establishes a useful group ofsymptoms that are commonly reported in hypoglycaemia.The way we ask people to describe their symptoms of hypoglycaemia can alter what theytell us. The rank order of symptoms alters considerably if patients are asked to indicate therelevance of each symptom rather than merely to identify that the symptom is associatedwith hypoglycaemia (Cox et al., 1993a). With regard to the criterion of relevance, the mostuseful symptoms in detecting hypoglycaemia are as follows:• sweating;• trembling;• difficulty concentrating;• nervousness, tenseness;• light-headedness, dizziness.The initial symptomsAnother important question is: which hypoglycaemic symptoms appear early during anepisode? The symptoms of hypoglycaemia that appear first and offer early warning of theonset of hypoglycaemia (Hepburn, 1993) are as follows:• trembling;• sweating;• tiredness;• difficulty concentrating;• hunger.This knowledge is obviously useful for the prompt detection and treatment of hypoglycaemia.

28 SYMPTOMS OF HYPOGLYCAEMIAThe Validity of Symptom BeliefsThe individuality of hypoglycaemic symptom clustersA great deal of the interest in symptoms of hypoglycaemia has been stimulated by concernsabout patient education. It is helpful to let patients know the range of symptoms found inhypoglycaemia and to inform them of the early warning symptoms reported by other peoplewith diabetes, much as we all tend to know the range of symptoms that are experiencedwith the common cold. Many surveys and laboratory studies have shown that people differconsiderably in the symptoms of hypoglycaemia they experience (Cox et al., 1993a). Inaddition to learning the generally reported symptoms, individuals with diabetes should beencouraged to learn about their own typical symptoms of hypoglycaemia.Correctly interpreting symptoms as representing hypoglycaemiaSymptoms of hypoglycaemia do not appear on top of the bodily equivalent of a blank sheetof paper. Sometimes we experience symptoms when there is nothing wrong with our bodilyfunctions; on the other hand, sometimes we fail to notice any symptoms when the body ismalfunctioning. The alert person with diabetes who is on the lookout for hypoglycaemiamust make two sorts of decisions. First, symptoms of hypoglycaemia must be detected andcorrectly identified. It would be dangerous for a patient to ignore symptoms of hypoglycaemiabecause he or she thought they were related to something else. Second, symptoms that havenothing to do with hypoglycaemia must be excluded. Unwanted hyperglycaemia could occurif patients treated themselves for hypoglycaemia when the symptoms had another cause.These two main types of error are a failure to treat hypoglycaemia when blood glucose islow, and inappropriate treatment when blood glucose is acceptable or high (Cox et al., 1985;1993a) (Figure 2.1).Blood glucose concentration – symptom report correlationsDo patients’ reports of symptoms of hypoglycaemia bear any relation to their concurrentblood glucose concentrations? After all, the principal aim of educating people to beFigure 2.1Consequences of correct and incorrect perception of hypoglycaemic symptoms

SYMPTOMS OF HYPOGLYCAEMIA 29aware of symptoms of hypoglycaemia is that they become alert to low and potentiallydangerous levels of blood glucose. To answer the above question some researchers haveemployed a field study approach where people are invited to list any symptoms they areexperiencing, and then measure and record their blood glucose concentration several timesa day for weeks. As a result of these studies it is known that each person has somesymptoms that are most reliably associated with their actual blood glucose concentrations(Pennebaker et al., 1981). Some of the symptoms that people report during hypoglycaemiaare more closely related to their actual blood glucose concentrations than are others, and ifwe can identify each individual’s most informative symptoms, we can instruct people to paymore attention to them.The following symptoms are most consistently associated with actual blood glucoseconcentrations (Pennebaker et al., 1981):• hunger (in 53% of people);• trembling (in 33%);• weakness (in 27%);• light-headedness (in 20%);• pounding heart and fast heart rate (both 17%).The same symptoms are not informative for everyone. There were 27% of people for whomweakness was significantly associated with hypoglycaemia, but there were 7% in whom itwas a good symptom of hyperglycaemia! Most people reported more than three symptomsthat were strongly associated with the measured blood glucose concentration.It is evident that an individual’s symptoms are idiosyncratic. If we can help a patientto identify the symptoms of hypoglycaemia peculiar to him or her, which relate toactual blood glucose concentrations, then, by attending to these symptoms, the personshould be especially accurate in recognising hypoglycaemia. People who have one ormore reliable symptom(s) of hypoglycaemia correctly recognise half of their episodes ofhypoglycaemia (defined as a blood glucose less than 3.95 mmol/l [Cox et al., 1993a]).Those who have four or more reliable symptoms recognise a blood glucose below3.95 mmol/l three-quarters of the time. The field study method has suggested that attentionto the following symptoms was particularly useful in detecting actual low blood glucoseconcentrations:• nervousness/tenseness;• slowed thinking;• trembling;• light-headedness/dizziness;• difficulty concentrating;• pounding heart;• lack of co-ordination.

30 SYMPTOMS OF HYPOGLYCAEMIAClassifying Symptoms of HypoglycaemiaUntil now the symptoms of hypoglycaemia have been treated as a homogeneous whole. Canthese symptoms be divided into different groups?Hypoglycaemia has effects on more than one part of the body, and the symptoms ofhypoglycaemia reflect this. First, the direct effects of a low blood glucose concentrationon the brain – especially the cerebral cortex – cause neuroglycopenic symptoms. Second,autonomic symptoms result from activation of parts of the autonomic nervous system.Finally, there may be some non-specific symptoms that are not directly generated by eitherof these two mechanisms. It is only recently that scientific investigations have taken placeto confirm the idea that these separable groups of hypoglycaemic symptoms exist.As suggested above, there are at least two distinct groups of symptoms during the body’sreaction to hypoglycaemia (Hepburn et al., 1991):• Autonomic, with symptoms such as trembling, anxiety, sweating and warmness.• Neuroglycopenic, with symptoms such as drowsiness, confusion, tiredness, inability toconcentrate and difficulty speaking.This information can assist with patient education by supplying evidence for separablegroups of symptoms, and by indicating which symptoms belong to each group. Someneuroglycopenic symptoms, such as the inability to concentrate, weakness and drowsiness,are among the earliest detectable symptoms, but patients tend to rely more on autonomicsymptoms when detecting the onset of hypoglycaemia. Paying more attention to thepotentially useful, early neuroglycopenic symptoms could help with the early detection ofhypoglycaemia.Similar groups of symptoms of hypoglycaemia have been discovered by asking people torecall the symptoms they typically noticed during hypoglycaemia. However, in addition tothe two groups described above, a general feeling of malaise is added (Deary et al., 1993):• Autonomic: e.g. sweating, palpitations, shaking and hunger.• Neuroglycopenic: e.g. confusion, drowsiness, odd behaviour, speech difficulty andincoordination.• General malaise: e.g. headache and nausea.These 11 symptoms are so reliably reported by people and so clearly separable into thesethree groups, that they are used as the ‘Edinburgh Hypoglycaemia Scale’ (Deary et al.,1993). Table 2.2 shows how different researchers have found similar groups of autonomicand hypoglycaemic symptoms.In addition to the above studies that used patients’ self-reported symptoms, physiologicalstudies have also confirmed that the symptoms of hypoglycaemia can be divided into autonomicand neuroglycopenic groups. Symptoms such as sweating, hunger, pounding heart,tingling, nervousness and feeling shaky/tremulous (autonomic symptoms) can be reduced oreven prevented by drugs that block neurotransmission within the autonomic nervous system(Towler et al., 1993), confirming that these symptoms are caused by the autonomic responseto hypoglycaemia. Symptoms such as warmth, weakness, difficulty thinking/confusion,

SYMPTOMS OF HYPOGLYCAEMIA 31Table 2.2Different authors’ lists of autonomic and neuroglycopenic symptoms of hypoglycaemiaAutonomicNeuroglycopenicDeary et al.(1993)Towler et al.(1993)Weinger et al.(1995)Deary et al.(1993)Towler et al. (1993)Sweating Sweaty Sweating Confusion Difficulty thinking/confusedPalpitation Heart pounding Pounding heart, Drowsiness Tired/drowsyfast pulseShaking Shaky/tremulous Trembling OddbehaviourHunger Hungry SpeechDifficulty speakingdifficultyTinglingIncoordinationNervous/anxious Tense WeakBreathing hardWarmfeeling tired/drowsy, feeling faint, difficulty speaking, dizziness and blurred vision (neuroglycopenicsymptoms) are not prevented by drugs that block the autonomic nervous system.Therefore, neuroglycopenic symptoms are not mediated via the autonomic nervous systemand are thought to be caused by the direct effect of glucose deprivation on the brain. Thistype of research has also observed that people tend to rely on autonomic symptoms to detecthypoglycaemia, even when neuroglycopenic symptoms are just as prominent (Towler et al.,1993). Once more, this suggests that more emphasis should be placed on education of thepotential importance of neuroglycopenic symptoms for the early warning of hypoglycaemia.Symptoms might gather into slightly different groupings depending on the situation. Thesymptom groupings in the Edinburgh Hypoglycaemia Scale were developed from diabeticpatients’ retrospective reports. However, when people are asked to rate the same group ofsymptoms during acute, experimentally-induced moderate hypoglycaemia, a slightly differentpattern emerges (McCrimmon et al., 2003). In Table 2.3 there is an autonomic grouping, andthe single neuroglycopenia group has divided into two symptom groups: one with mostlycognitive symptoms and the other with more general symptoms. This division probably arosebecause the subjects in the studies used to form Table 2.3 were engaged in cognitive tasksTable 2.3 Symptom groupings of the Edinburgh Hypoglycaemia Scaleduring experimentally-induced hypoglycaemiaNeuroglycopenic symptomsCognitive dysfunction Neuroglycopenia Autonomic symptomsInability to concentrate Drowsiness SweatingBlurred vision Tiredness TremblingAnxiety Hunger WarmnessConfusionWeaknessDifficulty speakingDouble vision

32 SYMPTOMS OF HYPOGLYCAEMIAduring the period of hypoglycaemia. Therefore, they would be especially aware of cognitiveshortcomings, making this group of symptoms more prominent and coherent.Symptoms in Children and Older PeopleChildren often have difficulty in recognising symptoms of hypoglycaemia, and they showmarked variability in symptoms between episodes of hypoglycaemia (Macfarlane and Smith,1988). Trembling and sweating are often the first symptoms recognised by children. Frominterviews with the parents of children (aged up to 16 years) with type 1 diabetes, andwith some of their children, more is known about the frequency of symptoms of hypoglycaemiain children (McCrimmon et al., 1995; Ross et al., 1998) (Table 2.4). The mostfrequently reported sign that parents observed was pallor (noted by 88%). The parentsTable 2.4 Symptoms of hypoglycaemia in children (derived from Ross et al., 1998)Frequency of rating (%)Symptom Parents’ reports Children’s reportsCorrelation betweenparents’ and children’sintensity ratings aTearful 73 47 040 dHeadache 73 65 033 dIrritable 85 65 016Uncoordinated 62 56 018Naughty 47 31 023 bWeak 79 83 021 bAggressive 75 62 026 cTrembling 79 88 025 bSleepiness 63 69 027 cNightmares 33 19 033 dSweating 76 73 028 cSlurred speech 53 45 028 cBlurred vision 52 55 030 cTummy pain 67 41 036 dFeeling sick 63 53 032 cHungry 74 84 019Yawning 48 45 020 bOdd behaviour 65 50 022 bWarmness 57 68 013Restless 61 57 021 bDaydreaming 70 48 014Argumentative 64 50 021 bPounding heart 21 44 002Confused 75 70 041 dTingling lips 20 24 −001Dizziness 66 87 028 cTired 83 76 026 cFeeling awful 92 79 020a Correlations: p of z (corrected for ties) 1.0 represents perfect agreement; 0 represents no agreement.b p

SYMPTOMS OF HYPOGLYCAEMIA 33frequently reported symptoms of behavioural disturbance such as irritability, argumentativenessand aggression. This latter group of symptoms is not prominent in adults, althoughthe Edinburgh Hypoglycaemia Scale includes ‘odd behaviour’ as an adult neuroglycopenicsymptom. Others had previously noted the prominence of symptoms such as irritability,aggression and disobedience in the parents’ reports of their children’s symptoms of hypoglycaemia(Macfarlane and Smith, 1988; Macfarlane et al., 1989). Parents tend to under-reportthe subjective symptoms of hypoglycaemia, such as weakness and dizziness, but generallythere is good agreement between parents and their children about the most prominentsymptoms of childhood hypoglycaemia (McCrimmon et al., 1995; Ross et al., 1998).Separate groups of autonomic and neuroglycopenic symptoms were not found in childrenwith type 1 diabetes (McCrimmon et al., 1995; Ross et al., 1998). These symptoms arereported together by children and are not distinguished as separate groups, whereas the groupof symptoms related to behavioural disturbance is clearly reported as a distinct group. In arefinement of the earlier study by McCrimmon et al. (1995), Ross et al. (1998) found thatparents could distinguish between autonomic and neuroglycopenic symptoms.People with insulin-treated type 2 diabetes report symptoms during hypoglycaemia thatseparate into autonomic and neuroglycopenic groups (Henderson et al., 2003). Elderlypatients with type 2 diabetes treated with insulin commonly report neurological symptomsof hypoglycaemia which may be misinterpreted as features of cerebrovascular disease, suchas transient ischaemic attacks (Jaap et al., 1998). The age-specific differences in the groupsof hypoglycaemic symptoms, classified using statistical techniques (Principal ComponentAnalysis), are shown in Table 2.5. Health professionals and carers who are involved inthe treatment and education of diabetic patients should be aware of which symptoms arecommon at either end of the age spectrum.From Symptom Perception to ActionPeople with diabetes are better at estimating their blood glucose in natural, everyday situations,as opposed to clinical laboratory settings (Cox et al., 1985). In some ways this issurprising as natural hypoglycaemia often occurs at a time when it is unexpected. In this situation,attention toward symptoms will not be as actively directed toward detection as in thelaboratory setting where it is usually anticipated. Furthermore, hypoglycaemia in everydaylife occurs on the background of other bodily feelings and must be separated from othercauses of the same symptoms. For example, exercise and various acute illnesses can provokesweating in people with diabetes, independently of their association with hypoglycaemia.In a real-life situation a person must detect symptoms of hypoglycaemia and then interpretthem. Failure to detect symptoms can lead to a failure to treat hypoglycaemia, but detectionTable 2.5 Classification of symptoms of hypoglycaemia using Principal Components Analysis inpatients with insulin-treated diabetes depending on age groupChildren (pre pubertal) Adults ElderlyAutonomic/neuroglycopenic Autonomic AutonomicNeuroglycopenicNeuroglycopenicBehavioural Non-specific malaise Neurological

34 SYMPTOMS OF HYPOGLYCAEMIAwithout the correct interpretation of the cause of the symptoms is equally dangerous. Furthermore,it is obvious that someone who does interpret symptoms correctly as being causedby hypoglycaemia, but who does not take action to treat the low blood glucose, will be atthe same risk. These and other steps toward the avoidance of severe hypoglycaemia demonstratethe key role of education about symptoms in people with diabetes (Gonder-Fredericket al., 1997). The psycho-educational programmes of Blood Glucose Awareness Training(BGAT; Schachinger et al., 2005) and Hypoglycemia Anticipation, Awareness and TreatmentTraining (HAATT; Cox et al., 2004) have led to better recognition of hypoglycaemicstates and reduced frequency of hypoglycaemia.Symptom generationFigure 2.2 outlines the stages that intervene between low blood glucose occurring in anindividual and the implementation of effective treatment (Gonder-Frederick et al., 1997). Itis interesting to note the importance of behavioural factors in generating states of low bloodglucose. Episodes of low blood glucose are most likely to come about because of changesin routine aspects of diabetes management, such as taking extra insulin, eating less food ortaking more exercise (Clarke et al., 1997). These factors predict more than 85% of episodesof hypoglycaemia in people with diabetes.In the presence of an intact physiological response to low blood glucose, autonomic andneuroglycopenic symptoms, and symptoms of general malaise, are generated (Figure 2.2).The degree of hypoglycaemia, the person’s quality of glycaemic control and any recentepisodes of hypoglycaemia may all affect the magnitude of the body’s physiological response.Recent, preceding hypoglycaemia can reduce the symptomatic and counterregulatoryhormonal responses to subsequent hypoglycaemia, resulting in a diminished awareness ofsymptoms. This effect of ‘antecedent’ hypoglycaemia is described in Chapter 7. Gender doesnot appear to influence the symptomatic response to hypoglycaemia (Geddes et al., 2006).At the second stage in Figure 2.2 comes the actual generation of physical symptoms.Among the variables that can influence this stage is the prior ingestion of caffeine, whichhas been shown to enhance the intensity of the autonomic and neuroglycopenic symptomsof experimentally induced hypoglycaemia (Debrah et al., 1996) (see Chapter 5). Caffeinemay act by increasing the intensity of symptoms of hypoglycaemia to perceptible levels,much as a magnifying glass enables one to read otherwise too-small print.Symptom detectionThe occurrence of physiological changes in the body does not guarantee that a personwill detect symptoms (Gonder-Frederick et al., 1997). If attention is directed to physicalchanges, people are more likely to detect symptoms than if their attention is held elsewhere.Everyone has had the experience of feeling less discomfort, and being less likely to detecta physical symptom, when being distracted by something diverting. The personal relevanceof the symptom may affect detection; for example, a person with heart disease may be verylikely to detect palpitations (Cox et al., 1993a). The activity of the person at the time ofthe physiological change is obviously important. Hypoglycaemic symptoms will be moreobvious to the person engaged in active mental effort (McCrimmon et al., 2003), such assitting an examination, than to the person relaxing and watching television. A doctor engagedin microsurgery may be very sensitive to the onset of tremor. In one laboratory study the

SYMPTOMS OF HYPOGLYCAEMIA 35Figure 2.2 A model for the occurrence and avoidance of severe hypoglycaemia (after Gonder-Frederick et al., 1997). Note on the right-hand side of each stage the factors that affect its occurrencepeople who had higher anxiety levels were better at detecting symptoms of hypoglycaemia(Ryan et al., 2002).The autonomic symptoms of hypoglycaemia are often emphasised in the detection of hypoglycaemia.However, a strong case can be made for an equal emphasis on neuroglycopenicsymptoms (Gonder-Frederick et al., 1997; McCrimmon et al., 2003) because:• performance on mental tasks deteriorates during hypoglycaemia, and subjective awarenessof this decrement begins at very mild levels of hypoglycaemia• the difference in glycaemic thresholds for the onset of autonomic and neuroglycopenicsymptoms is so small that it is unlikely to be detected when blood glucose declines rapidly

36 SYMPTOMS OF HYPOGLYCAEMIA• neuroglycopenic symptoms are as strongly related to actual blood glucose concentrationsas are autonomic symptoms (Cox et al., 1993a)• people with insulin-treated diabetes cite autonomic and neuroglycopenic symptoms withequal frequency as the primary warning symptoms of hypoglycaemia (Hepburn et al., 1991).Low blood glucose detection – symptom interpretationThe correct detection of a symptom of hypoglycaemia does not always lead to correctinterpretation (Figure 2.2). After correctly detecting relevant symptoms, people fail to detectabout 26% of episodes of low blood glucose (Gonder-Frederick et al., 1997). There areseveral factors that could break a perfect relationship between detection and recognition,and some are discussed below. However, it should be appreciated that symptom detection(internal cues) is not mandatory for the detection of low blood glucose. Self-testing of bloodglucose or the information of family members (external cues) can lead to the successfulrecognition of hypoglycaemia without the patient having detected the episode by symptomaticperception (see Box 2.1).Correct knowledge about symptoms of hypoglycaemia is necessary for the detection oflow blood glucose. The lack of such knowledge among elderly people with diabetes inparticular gives cause for concern (Mutch and Dingwall-Fordyce, 1985). Of 161 diabeticpeople between ages 60 and 87, all of whom were injecting insulin or taking a sulphonylurea,only 22% had ever been told the symptoms of hypoglycaemia and 9% knew no symptomsat all! The percentages of the insulin-treated diabetic patients who knew that the followingsymptoms were associated with hypoglycaemia were as follows:• sweating (82%);• palpitations (62%);• confusion (53%);• hunger (51%);• inability to concentrate (50%);• speech problems (41%);• sleepiness (33%).Box 2.1Identification of hypoglycaemia• Internal cues – autonomic, neuroglycopenic and non-specific symptoms• External cues – relationship of insulin injection to meals, exercise and experienceof self-management• Blood glucose monitoring• Information from observers (e.g. relatives, friends, colleagues)

ACUTE HYPOGLYCAEMIA AND COGNITIVE FUNCTIONING 37In the midst of this ignorance, much hypoglycaemia may not be treated because of a lackof knowledge of the symptoms of hypoglycaemia which would aid their recognition.Most, if not all, of the symptoms of hypoglycaemia can be explained by other physicalconditions. Therefore, correct symptom detection may be usurped by incorrect attribution ofthe cause. For example, having completed some strenuous activity, an athlete may attributethe symptoms of sweating and palpitations to physical exertion. An obvious problem indetecting a low blood glucose is the fact that the organ responsible for the detectionand interpretation of symptoms – the brain, especially the cerebral cortex – is impaired.Thus, impaired concentration and lowered consciousness levels can beget even more severehypoglycaemia.Symptom scoring systemsThe controversy about the effect of human insulin on symptom awareness (Chapter 7)stimulated the development of scoring systems for hypoglycaemia to allow comparativestudies between insulin species. This produced scoring systems such as the EdinburghHypoglycaemia Scale (Deary et al., 1993), and any such system must be validated forresearch application. It is important to note that the nature and intensity of individualsymptoms are as important as, if not more important than, the number of symptoms generatedby hypoglycaemia. The concepts involved are discussed in detail by Hepburn (1993).More information on the symptoms of hypoglycaemia is provided by McAulay et al.(2001b).ACUTE HYPOGLYCAEMIA AND COGNITIVE FUNCTIONINGSymptoms are subjective reports of bodily sensations. With respect to hypoglycaemia someof these reports – especially neuroglycopenic symptoms – pertain to altered cognitive(mental ability) functioning. Do reports of ‘confusion’ and ‘difficulty thinking’ (Table 2.2)concur with objective mental test performance in hypoglycaemia? Before experimentalhypoglycaemia became an accepted investigative tool in diabetes, expert clinical observersnoted impairments of cognitive functions despite clear consciousness during hypoglycaemia(Fletcher and Campbell, 1922; Wilder, 1943). Cognitive functions include the following sortsof mental activity: orientation and attention, perception, memory (verbal and non-verbal),language, construction, reasoning, executive function and motor performance. Early studies(Russell and Rix-Trot, 1975) established that the following abilities become disrupted belowblood glucose levels of about 3.0 mmol/l:• fine motor co-ordination;• mental speed;• concentration;• some memory functions.

38 SYMPTOMS OF HYPOGLYCAEMIAThe hyperinsulinaemic glucose clamp technique allows more controlled experiments ofacute hypo- and hyperglycaemia. However, although this technique is used in most studiesof cognitive function in hypoglycaemia, it does not mimic the physiological or temporalcharacteristics of ‘natural’ or intercurrent episodes of hypoglycaemia experienced by peoplewith type 1 diabetes. From laboratory experiments using the glucose clamp technique it wasfound that blood glucose concentrations between 3.1 and 3.4 mmol/l caused the followingeffects (Holmes, 1987; Deary, 1993):• Slowed reaction times (this experiment involves making a fast response when a lightappears on a computer screen. Hypoglycaemia had more effect on reaction times whenthe reaction involved making a decision).• Slowed mental arithmetic.• Impaired verbal fluency (in this test one has to think of words beginning with a givenletter, probably involving the frontal lobes of the brain).• Impaired performance in parts of the Stroop test (in this test one has to read aloud a seriesof ink colours when words are printed in a different colour from that of the name, e.g. theword RED printed in green ink).Some mental functions were spared during hypoglycaemia, for example:• simple motor (like the speed of tapping) and sensory skills;• the speed of reading words aloud.By 1993 over 16 studies had investigated cognitive functions during acute and mild–moderate hypoglycaemia (Deary, 1993). The levels of blood glucose ranged from 2.0 to3.7 mmol/l. The way that hypoglycaemia was induced varied among studies, as did themethods of blood sampling (e.g. arterialised or venous blood). Moreover, the ability levelsof the people in different studies varied, and there was much heterogeneity in the testbatteries used to assess mental performance. An authoritative statement as to the mentalfunctions disrupted during hypoglycaemia is still not possible. However, in at least one ormore of the studies a number of tests were significantly impaired during hypoglycaemia(Box 2.2).Few areas of mental function are preserved at normal levels during acute hypoglycaemia.There is a general dampening of many abilities that involve conscious mental effort. In theface of so many deleterious effects, what mental functions remain intact during acute hypoglycaemia?At blood glucose concentrations similar to those indicated above, the followingmental tests are not significantly impaired:• finger tapping;• forward digit span (repeating back a list of numbers in the same order);• simple reaction time;• elementary sensory processing.

ACUTE HYPOGLYCAEMIA AND COGNITIVE FUNCTIONING 39Box 2.2Cognitive function tests impaired during acute hypoglycaemia• Trail making (involving visual scanning and mental flexibility)• Digit symbol (speed of replacing a list of numbers with abstract codes)• Reaction time (especially involving a decision)• Mental arithmetic• Verbal fluency• Stroop test• Grooved pegboard (a test of fine manual dexterity)• Pursuit rotor (a test of eye–hand co-ordination)• Letter cancellation (striking out occurrences of a given letter in a page of letters)• Delayed verbal memory• Backward digit span (repeating back a list of numbers backwards)• Story recallThus tests which involve speeded responses and which are more cognitively complex andattention-demanding tend to show impairment during hypoglycaemia (Deary, 1993). Hellerand Macdonald (1996) have concluded that:• even quite severe degrees of hypoglycaemia do not impair simple motor functions;• choice reaction time (where a mental decision of some kind is needed before reacting toa stimulus) is affected at higher blood glucose concentrations more than simple reactiontime;• speed of responding is sometimes slowed in a task in which accuracy is preserved;• many aspects of mental performance become impaired when blood glucose falls belowabout 3.0 mmol/l;• there are important individual differences; some people’s mental performance is alreadyimpaired above a blood glucose of 3.0 mmol/l, whereas others continue to function wellat lower levels;• the speed of response of the brain in making decisions slows down during hypoglycaemia(Tallroth et al., 1990; Jones et al., 1990);• it can take as long as 40 to 90 minutes after blood glucose returns to normal for the brainto recover fully (Blackman et al., 1992; Lindgren et al., 1996).

40 SYMPTOMS OF HYPOGLYCAEMIAFigure 2.3 Cognitive effects of hypoglycaemia. After Deary (1998). This Article was published inDiabetes Annual 11, SM Marshall, PD Home and RA Rizza (eds), 97–118, Copyright Elsevier 1998Determining whether mental performance is impaired at all during mild to moderatehypoglycaemia, while a person is still fully conscious, is only the beginning of this lineof investigation. The next question to ask is whether some particular functions are moresusceptible and some less so? Figure 2.3 encapsulates this problem and illustrates three otherimportant questions about the cognitive effects of acute hypoglycaemia:1. What factors affect the degree of cognitive impairment during hypoglycaemia, other thanthe level of blood glucose?2. Do impairments in laboratory cognitive tasks have a bearing on mental performance inreal life?3. Which basic brain functions are disturbed during acute hypoglycaemia?Influences on the Degree of and Threshold for Cognitive DysfunctionDuring Acute HypoglycaemiaAlthough, on average, impairment of mental performance is worse during hypoglycaemia,some people do not change or may even improve (Pramming et al., 1986; Hoffman et al.,

ACUTE HYPOGLYCAEMIA AND COGNITIVE FUNCTIONING 411989). It is not yet certain whether such individual differences in responses are stable(Gonder-Frederick et al., 1994; Driesen et al., 1995). The following factors might increasea person’s degree of cognitive impairment during acute hypoglycaemia:• male sex (Draelos et al., 1995; but this is disputed for people with type 2 diabetes byBremer et al., 2006);• impaired hypoglycaemia awareness (Gold et al., 1995b);• type 1 diabetes (Wirsen et al., 1992);• high IQ (Gold et al., 1995a).Does glycaemic control affect the cognitive impact of hypoglycaemia? People with type1 diabetes on intensified insulin therapy attain glycaemic control that is nearer to normalthan most people treated with conventional insulin treatment. As a result, the frequency ofsevere hypoglycaemia is increased and is associated with a greater risk of impaired hypoglycaemiaawareness. Neuroendocrine responses to hypoglycaemia are reduced in magnitudeand begin at lower absolute blood glucose concentrations than in people with less strictglycaemic control. However, the hypoglycaemic threshold for cognitive dysfunction maynot change in a similar fashion. Diabetic patients on intensive insulin therapy reported autonomicand neuroglycopenic symptoms at blood glucose concentrations of about 2.4 and2.3 mmol/l respectively, whereas in those with less strict glycaemic control and in nondiabeticindividuals, these symptoms commenced at between 2.8 and 3.0 mmol/l. However,in all three groups, the accuracy and speed in a reaction time test deteriorated significantly atblood glucose concentrations between 2.8 and 3.0 mmol/l (Amiel et al., 1991; Maran et al.,1995). Therefore, people with insulin-treated diabetes who have strict glycaemic controlhave the misfortune that the deterioration in their mental performance begins before theonset of warning symptoms of hypoglycaemia. By contrast, neurophysiological responses(P300 event-related potentials), which have been linked to various measures of cognitivefunction, occur at lower blood glucose concentrations, suggesting that cerebral adaptationhas occurred (Ziegler et al., 1992). The effects of quality of glycaemic control, antecedenthypoglycaemia and impaired hypoglycaemia awareness on the mental performance responsesto hypoglycaemia, and the relation of these responses to the perceptions of the symptoms ofhypoglycaemia, are important topics still under study (see Chapters 7 and 8). The interrelationof these factors makes the field complex, and progress is further hampered by the lackof consensus agreement on a validated battery of cognitive tests for use in hypoglycaemia.Are the Cognitive Changes During Acute Hypoglycaemia Importantand Valid?Do the impairments of mental test performance actually have implications for real-lifefunctions? In addition, are the mental changes during hypoglycaemia a result of impairmentsin basic brain functions?One common, important and potentially dangerous area of real-life functioning is driving(see Chapter 14), which involves many cognitive abilities including psychomotor controland divided attention. Cox and colleagues (1993b; 2000) employed a sophisticated drivingsimulator and had people ‘drive’ on this during controlled hypoglycaemia using a glucose

42 SYMPTOMS OF HYPOGLYCAEMIAclamp technique. With very mild hypoglycaemia (blood glucose below 3.8 mmol/l) thediabetic drivers committed significant driving errors, and during hypoglycaemia the patientsoften drove very slowly, possibly using a compensatory mechanism to avoid errors. Despitethis, more global errors of driving were committed and about half of the participants,despite demonstrating a seriously impaired ability to drive, said they felt competent to driveirrespective of their low blood glucose! It cannot be stated with certainty that the findingsobtained in a driving simulator will apply to real-life driving. However, studies that examinethe practical cognitive effects of hypoglycaemia are invaluable and more are required.Just as more studies that examine the practical cognitive aspects of hypoglycaemia wouldbe useful, so would more studies of the brain’s processing efficiency. Cognitive tests typicallyinvolve a melange of inseparable mental processes, and yet very specific aspects of the humanbrain’s activities can be measured in the clinical laboratory (Massaro, 1993). Studies of thecognitive effects of hypoglycaemia have thus begun to address the impairments to variouscognitive domains in more detail. Basic, specific aspects of visual and auditory processinghave been examined during acute hypoglycaemia in non-diabetic humans. Standard tests ofvisual acuity – those that are measured by an optometrist – are not affected by hypoglycaemia,but other aspects of vision are affected (McCrimmon et al., 1996). These include:• contrast sensitivity (the ability to discriminate faint patterns);• inspection time (the ability to see what is in a pattern when it is shown for a very briefperiod of time);• visual change detection (the ability to spot a small, quick change in a pattern);• visual movement detection (the ability to spot brief movement in a pattern).This means that the ability to see the environment changes in important ways during hypoglycaemia.Visual acuity is preserved, as tested by the ability to read black letters on a whitebackground. However, most visual activity is not like that; many of the things we see happenquickly and in relatively poor light. When the level of contrast falls, or discriminationsmust be made under pressure of time, visual processing is impaired during hypoglycaemia.However, at about the same degree of hypoglycaemia, the ability to distinguish one colourfrom another does not appear to be impaired (Hardy et al., 1995). Speed of auditoryprocessing also appears to be impaired by hypoglycaemia, and the ability to discriminatethe loudness of two tones is disrupted (McCrimmon et al., 1997; Strachan et al., 2003).This suggests that the ability to understand language may be compromised during hypoglycaemia.However, despite there being disruption to central nervous system processing duringhypoglycaemia, no disturbance has been detected in peripheral nerve conduction (Strachanet al., 2001).If basic information processing provides a fundamental limitation to how well the brainis operating, then at a higher level of function, attention is important in carrying out anumber of cognitive functions. A detailed study of a number of different aspects of attentionduring hypoglycaemia found that the abilities to attend selectively and to switch attentionas necessary both deteriorated (McAulay et al., 2001a; 2005).In turn, attention is necessary in order to learn and form new memories. There hasnow been detailed study of the different aspects of memory during acute, insulin-inducedhypoglycaemia (Sommerfield et al., 2003a; 2003b; Deary et al., 2003). Most memory

ACUTE HYPOGLYCAEMIA AND EMOTIONS 43systems are disrupted during hypoglycaemia. However, some are especially badly affected:long-term memory, which is the ability to retain new information after many minutes andmuch distraction; working memory, which is the ability to retain and manipulate informationat the same time; and prospective memory, which is the ability to remember to do things (asin a shopping list) (Warren et al., 2007). Indeed, the ability to perform one tricky workingmemory task was obliterated during hypoglycaemia (Deary et al., 2003). That is, no matterhow good the person was at performing the task during euglycaemia, the same task couldnot be done during moderate hypoglycaemia.Further information on hypoglycaemia and cognitive function is available in a review byWarren and Frier (2005).ACUTE HYPOGLYCAEMIA AND EMOTIONSMood change is part of the experience of hypoglycaemia. Moods are emotion-like experiencesthat are quite general rather than applied to specific situations. Psychologists recognisethree basic moods:• energetic arousal (a tendency to feel lively and active rather than tired and sluggish);• tense arousal (a tendency to feel anxious and nervous versus relaxed and calm);• hedonic tone (a tendency to feel happy versus sad).When people are asked to rate their mood states during hypoglycaemia induced in thelaboratory, changes occur in all of these basic mood states. People feel less energetic, moretense and less happy (Gold et al., 1995c; McCrimmon et al., 1999a; Hermanns et al., 2003).During hypoglycaemia the emotional arousal in response to stimuli becomes more intense(Hermanns et al., 2003). In addition, some people become more irritable and have angryfeelings during hypoglycaemia (Merbis et al., 1996; McCrimmon et al., 1999b). The feelingof low energy takes over half an hour to be restored to normal levels, whereas the feelings oftenseness and unhappiness disappear when blood glucose returns to normal. The prolongedfeeling of low energy after hypoglycaemia may affect work performance, so that when hypoglycaemiahas been treated, an immediate return to the normal state should not be expected.In addition to some people experiencing a low, tense, washed-out, angry mood state,hypoglycaemia alters the way some people look at their life problems. When junior doctorswere asked to assess their career prospects during controlled hypoglycaemia, they weremore pessimistic (McCrimmon et al., 1995) and if a general state of pessimism is commonduring hypoglycaemia, it would be a poor state from which to make personal decisions. It ispossible that the change in mood states during hypoglycaemia is one of the causes of adultsadmitting to more ‘odd behaviour’ (Deary et al., 1993). Altered mood may also accountin part for symptoms of behavioural disturbance that are so prominent in the responses tohypoglycaemia of children with diabetes (McCrimmon et al., 1995).In addition to emotional responses as a result of hypoglycaemia, some people haveemotional responses in anticipation of hypoglycaemia. In Edinburgh one young man withinsulin-treated diabetes developed a phobic anxiety state; his phobia related to becomingcomatose as a result of hypoglycaemia (Gold et al., 1997a). Such a case is exceptional,but many people with diabetes are frightened of hypoglycaemia (see Chapter 14). The

44 SYMPTOMS OF HYPOGLYCAEMIAHypoglycaemia Fear Survey (HFS), which comes in two parts, measures this tendency(Cox et al., 1987). The first part asks people several questions concerning how much theyworry about hypoglycaemia (e.g. ‘Do you worry if you have no one around you during a[hypoglycaemic] reaction?’). The second part asks several questions about what people doto avoid hypoglycaemia (e.g. ‘Do you eat large snacks at bedtime?’). People with greaterfear of hypoglycaemia (Polonsky et al., 1992; Hepburn et al., 1994)• have more anxious personalities in general;• are more likely to confuse symptoms of anxiety for those of hypoglycaemia;• report having had more episodes of hypoglycaemia.It is not yet known whether people who experience more hypoglycaemia become worriersabout it, or whether people who are worriers in general just worry more about hypoglycaemiaas well. Perhaps both are true. However, it does seem likely that the experience of more severehypoglycaemia in the past and the development of impaired awareness of hypoglycaemialead to increased worry about subsequent hypoglycaemia (Gold et al., 1997b).CONCLUSIONS• Because people with diabetes are closely involved in their own treatment it is importantthat they and their educators know about the main side-effects and sequelae of the disorderand its treatments.• Accurate knowledge of the symptoms of hypoglycaemia may be used to avoid the dangersof hypoglycaemia.• The progressively more serious impairment in cognitive function that occurs as bloodglucose declines provides knowledge about the brain’s compromised state during hypoglycaemia:basic functions such as visual processing deteriorate and driving becomesdangerously error-prone. Performance on a host of mental tests becomes worse duringhypoglycaemia.• Some of the neuroglycopenic symptoms of hypoglycaemia are thought to be subjectiveimpressions of impaired cognitive function: these impressions are fully supported by theresults of objective cognitive testing.• Moderate hypoglycaemia may induce a state of anxious tension, unhappiness and lowenergy, and even irritability and anger. Thus hypoglycaemia importantly touches theemotions as well as inducing bodily symptoms and affecting mental performance.REFERENCESAmiel SA, Pottinger RC, Archibald HR, Chusney G, Cunnah DTF, Prior PF, Gale EAM (1991). Effectof antecedent glucose control on cerebral function during hypoglycemia. Diabetes Care 14: 109–18.Blackman JD, Towle VL, Sturis J, Lewis GF, Spire J-P, Polonsky KS (1992). Hypoglycemic thresholdsfor cognitive dysfunction in IDDM. Diabetes 41: 392–9.

REFERENCES 45Bremer JP, Baron M, Peters H, Oltmanns KM, Kern W, Fehm HL et al. (2006). Hormonal, subjectiveand neurocognitive responses to brief hypoglycemia in postmenopausal women and age-matchedmen with type 2 diabetes mellitus. Metabolism 55: 331–8.Clarke WL, Cox DJ, Gonder-Frederick LA, Julian D, Schlundt D, Polonsky W (1997). The relationshipbetween nonroutine use of insulin, food, and exercise and the occurrence of hypoglycemia inadults with IDDM and varying degrees of hypoglycemia awareness and metabolic control. DiabetesEducation 23: 55–8.Cox DJ, Clarke WL, Gonder-Frederick L, Pohl S, Hoover C, Snyder A et al. (1985). Accuracy ofperceiving blood glucose in IDDM. Diabetes Care 8: 529–36.Cox DJ, Irvine A, Gonder-Frederick L, Nowacek G, Butterfield J (1987). Fear of hypoglycemia:quantification, validation and utilization. Diabetes Care 10: 617–21.Cox DJ, Gonder-Frederick L, Antoun B, Cryer PE, Clarke WL (1993a). Perceived symptoms in therecognition of hypoglycemia. Diabetes Care 16: 519–27.Cox D, Gonder-Frederick L, Clarke W (1993b). Driving decrements in type 1 diabetes during moderatehypoglycemia. Diabetes 42: 239–43.Cox DJ, Gonder-Frederick LA, Kovatchev BP, Julian DM, Clarke WL (2000). Progressive hypoglycemia’simpact on driving simulation performance. Occurrence, awareness and correction.Diabetes Care 23: 163–70.Cox DL, Kovatchev B, Koev D, Koeva L, Dachev S, Tcharaktchiev D et al. (2004). Hypoglycemiaanticipation, awareness and treatment training (HAAT) reduces occurrence of severe hypoglycemiaamong adults with type 1 diabetes. International Journal of Behavioral Medicine 11: 212–18.Deary IJ (1993). Effects of hypoglycaemia on cognitive function. In: Hypoglycaemia and Diabetes:Clinical and Physiological Aspects. Frier BM and Fisher M, eds. Edward Arnold, London: 80–92.Deary IJ, Hepburn DA, MacLeod KM, Frier M (1993). Partitioning the symptoms of hypoglycaemiausing multi-sample confirmatory factor analysis. Diabetologia 36: 771–7.Deary IJ (1998). The effects of diabetes on cognitive function. In: Diabetes Annual 11. Marshall SM,Home PD and Rizza RA eds. Elsevier, London: 97–118.Deary IJ, Sommerfield AJ, McAulay V, Frier BM (2003). Moderate hypoglycaemia obliterates workingmemory in humans with and without insulin treated diabetes. Journal of Neurology, Neurosurgeryand Psychiatry 74: 277–82.Debrah T, Sherwin RS, Murphy J, Kerr D (1996). Effect of caffeine on recognition of and physiologicalresponses to hypoglycaemia in insulin-dependent diabetes. Lancet 347: 19–24.Draelos MT, Jacobson AM, Weinger K, Widom B, Ryan CM, Finkelstein DM, Simonson DC (1995).Cognitive function in patients with insulin-dependent diabetes mellitus during hyperglycemia andhypoglycemia. American Journal of Medicine 98: 135–44.Driesen NR, Cox DJ, Gonder-Frederick L, Clarke W (1995). Reaction time impairment in insulindependentdiabetes: task complexity, blood glucose levels, and individual differences. Neuropsychology9: 246–54.Fletcher AA, Campbell WR (1922). The blood sugar following insulin administration and the symptomcomplex-hypoglycemia. Journal of Metabolic Research 2: 637–49.Geddes J, Warren, RE, Sommerfield AJ, McAulay V, Strachan MWJ, Allen KV et al. (2006). Absenceof sexual dimorphism in the symptomatic responses to hypoglycemia in adults with and withouttype 1 diabetes. Diabetes Care 29: 1667–9.Gold AE, Deary IJ, MacLeod KM, Frier BM (1995a). The effect of IQ level on the degree of cognitivedeterioration experienced during acute hypoglycemia in normal humans. Intelligence 20: 267–90.Gold AE, MacLeod KM, Deary IJ, Frier BM (1995b). Hypoglycemia-induced cognitive dysfunctionin diabetes mellitus: effect of hypoglycemia unawareness. Physiology and Behavior 58: 501–11.Gold AE, MacLeod KM, Frier BM, Deary IJ (1995c). Changes in mood during acute hypoglycemiain healthy subjects. Journal of Personality and Social Psychology 68: 498–504.Gold AE, Deary IJ, Frier BM (1997a). Hypoglycaemia and non-cognitive aspects of psychologicalfunction in insulin-dependent (type 1) diabetes mellitus (IDDM). Diabetic Medicine 14: 111–18.

46 SYMPTOMS OF HYPOGLYCAEMIAGold AE, Frier BM, MacLeod KM, Deary IJ (1997b). A structural equation model for predictors ofsevere hypoglycaemia in patients with insulin-dependent diabetes mellitus. Diabetic Medicine 14:309–15.Gonder-Frederick L, Cox D, Driesen NR, Ryan CM, Clarke W (1994). Individual differences inneurobehavioral disruption during mild and moderate hypoglycemia in adults with IDDM. Diabetes43: 1407–12.Gonder-Frederick L, Cox D, Kovatchev B, Schlundt D, Clarke W (1997). A biopsychobehavioralmodel of risk of severe hypoglycemia. Diabetes Care 20: 161–9.Hardy KJ, Scase MO, Foster DH, Scarpello JH (1995). Effect of short term changes in blood glucoseon visual pathway function in insulin dependent diabetes. British Journal of Ophthalmology 79:38–41.Heller SR, Macdonald IA (1996). The measurement of cognitive function during acute hypoglycaemia:experimental limitations and their effects on the study of hypoglycaemia unawareness. DiabeticMedicine 13: 607–15.Henderson JN, Allen KV, Deary IJ, Frier BM (2003). Hypoglycaemia in insulin-treated type 2 diabetes:frequency, symptoms and impaired awareness. Diabetic Medicine 20: 1016–21.Hepburn DA, Deary IJ, Frier BM, Patrick AW, Quinn JD, Fisher M (1991). Symptoms of acute insulininducedhypoglycemia in humans with and without IDDM. Factor-analysis approach. Diabetes Care14: 949–57.Hepburn DA (1993). Symptoms of hypoglycaemia. In: Hypoglycaemia and Diabetes: Clinical andPhysiological Aspects. Frier BM and Fisher M eds. Edward Arnold, London: 93–103.Hepburn DA, Deary IJ, MacLeod KM, Frier BM (1994). Structural equation modeling of symptoms,awareness and fear of hypoglycemia, and personality in patients with insulin-treated diabetes.Diabetes Care 17: 1273–80.Hermanns N, Kubiak T, Kulzer B, Haak T (2003). Emotional changes during experimentally-inducedhypoglycaemia in type 1 diabetes. Biological Psychology 63: 15–44.Hoffman RG, Speelman DJ, Hinnen DA, Conley KL, Guthrie RA, Knapp RK (1989). Changes incortical functioning with acute hypoglycemia and hyperglycemia in type 1 diabetes. Diabetes Care12: 193–7.Holmes CS (1987). Metabolic control and auditory information processing at altered glucose levels ininsulin dependent diabetes. Brain and Cognition 6: 161–74.Jaap AJ, Jones GC, McCrimmon RJ, Deary IJ, Frier BM (1998). Perceived symptoms of hypoglycaemiain elderly type 2 diabetic patients treated with insulin. Diabetic Medicine 15: 398–401.Jones TW, McCarthy G, Tamborlane WV, Caprio S, Roessler E, Kraemer D et al. (1990). Mildhypoglycemia and impairment of brainstem and cortical evoked potentials in healthy subjects.Diabetes 39: 1550–5.Lindgren M, Eckert B, Stenberg G, Agardh C-D (1996). Restitution of neurophysiological functions,performance, and subjective symptoms after moderate insulin-induced hypoglycaemia innon-diabetic men. Diabetic Medicine 13: 218–25.Maran A, Lomas J, Macdonald IA, Amiel SA (1995). Lack of preservation of higher brainfunction during hypoglycaemia in patients with intensively-treated IDDM. Diabetologia 38:1412–18.Massaro DW (1993). Information processing models: microscopes of the mind. Annual Review ofPsychology 44: 383–425.Macfarlane PI, Smith CS (1988). Perceptions of hypoglycaemia in childhood diabetes mellitus: aquestionnaire study. Practical Diabetes 5: 56–8.Macfarlane PI, Walters M, Stutchfield P, Smith CS (1989). A prospective study of symptomatichypoglycaemia in childhood diabetes. Diabetic Medicine 6: 627–30.McAulay V, Deary IJ, Ferguson SC, Frier BM (2001a). Acute hypoglycemia in humanscauses attentional dysfunction while nonverbal intelligence is preserved. Diabetes Care 24:1745–50.

REFERENCES 47McAulay V, Deary IJ, Frier BM (2001b). Symptoms of hypoglycaemia in people with diabetes.Diabetic Medicine 18: 690–705.McAulay V, Deary IJ, Sommerfield AJ, Frier BM (2005). Attentional functioning is impaired duringacute hypoglycaemia in people with type 1 diabetes. Diabetic Medicine 23: 26–31.McCrimmon RJ, Gold AE, Deary IJ, Kelnar CJH, Frier BM (1995). Symptoms of hypoglycemia inchildren with IDDM. Diabetes Care 18: 858–61.McCrimmon RJ, Deary IJ, Huntly BJP, MacLeod KJ, Frier BM (1996). Visual information processingduring controlled hypoglycaemia in humans. Brain 119: 1277–87.McCrimmon RJ, Deary IJ, Frier BM (1997). Auditory information processing during acute insulininducedhypoglycaemia in non-diabetic human subjects. Neuropsychologia 35: 1547–53.McCrimmon RJ, Deary IJ, Frier BM (1999a). Appraisal of mood and personality during hypoglycaemia.Physiology and Behavior 67: 27–33.McCrimmon RJ, Ewing FME, Frier BM, Deary IJ (1999b). Anger-state during acute insulin-inducedhypoglycaemia. Physiology and Behavior 67: 35–9.McCrimmon RJ, Deary IJ, Gold AE, Hepburn DA, MacLeod KM, Ewing FME, Frier BM (2003).Symptoms reported during experimental hypoglycaemia: effect of method of induction of hypoglycaemiaand of diabetes per se. Diabetic Medicine 20: 507–9.Merbis MAE, Snoek FJ, Kanc K, Heine RJ (1996). Hypoglycaemia induces emotional disruption.Patient Education and Counseling 29: 117–22.Mutch WJ, Dingwall-Fordyce I (1985). Is it a hypo? Knowledge of symptoms of hypoglycaemia inelderly diabetic patients. Diabetic Medicine 2: 54–6.Pennebaker JW, Cox DJ, Gonder-Frederick L, Wunsch MG, Evans WS, Pohl S (1981). Physicalsymptoms related to blood glucose in insulin-dependent diabetics. Psychosomatic Medicine 43:489–500.Polonsky WH, Davis CL, Jacobson AM, Anderson BJ (1992). Correlates of hypoglycemic fear in type1 and type 2 diabetes mellitus. Health Psychology 11: 199–202.Pramming S, Thorsteinsson B, Theilgaard A, Pinner EM, Binder C (1986). Cognitive function duringhypoglycaemia in type 1 diabetes mellitus. British Medical Journal 292: 647–50.Ross LA, McCrimmon RJ, Frier BM, Kelnar CJH, Deary IJ (1998). Hypoglycaemic symptoms reportedby children with type 1 diabetes mellitus and by their parents. Diabetic Medicine 15: 836–43.Russell PN, Rix-Trot HM (1975). An exploratory study of some behavioural consequences of insulininducedhypoglycaemia. New Zealand Medical Journal 81: 337–40.Ryan CM, Dulay D, Suprasongsin C, Becker DJ (2002). Detection of symptoms by adolescents andyoung adults with type 1 diabetes during experimental induction of mild hypoglycemia: role ofhormonal and psychological variables. Diabetes Care 25: 852–8.Schachinger H, Hegar K, Hermanns N, Straumann M, Keller U, Fehm-Wolfsdorf G et al. (2005).Randomized controlled clinical trial of Blood Glucose Awareness Training (BGAT III) in Switzerlandand Germany. Journal of Behavioral Medicine 28: 587–94.Sommerfield AJ, Deary IJ, McAulay V, Frier BM (2003a). Moderate hypoglycemia impairs multiplememory functions in healthy adults. Neuropsychology 17: 125–32.Sommerfield AJ, Deary IJ, McAulay V, Frier BM (2003b). Short-term, delayed, and working memoryare impaired during hypoglycemia in individuals with type 1 diabetes. Diabetes Care 26: 390–6.Strachan MWJ, Deary IJ, Ewing FME, Ferguson SC, Young MJ, Frier BM (2001). Acute hypoglycemiaimpairs the functioning of the central but not peripheral nervous system. Physiology & Behavior72: 83–92.Strachan MWJ, Ewing FME, Frier BM, McCrimmon RJ, Deary IJ (2003). Effects of acute hypoglycaemiaon auditory information processing in adults with type 1 diabetes. Diabetologia 46:97–105.Tallroth G, Lindgren M, Stenberg G, Rosen I, Agardh C-D (1990). Neurophysiological changes duringinsulin-induced hypoglycaemia and in the recovery period following glucose infusion in type 1(insulin-dependent) diabetes mellitus and in normal man. Diabetologia 33: 319–23.

48 SYMPTOMS OF HYPOGLYCAEMIATowler DA, Havlin CE, Craft S, Cryer P (1993). Mechanism of awareness of hypoglycemia: perceptionof neurogenic (predominantly cholinergic) rather than neuroglycopenic symptoms. Diabetes 42:1791–8.Warren RE, Frier BM (2005). Hypoglycaemia and cognitive function. Diabetes, Obesity andMetabolism 7: 493–503.Warren RE, Zammitt NN, Deary IJ, Frier BM (2007). The effects of acute hypoglycaemia on memoryacquisition and recall and prospective memory in type 1 diabetes. Diabetologia 50: 178–85.Weinger K, Jacobson AM, Draelos MT, Finkelstein DM, Simonson DC (1995). Blood glucose estimationand symptoms during hyperglycemia and hypoglycemia in patients with insulin-dependentdiabetes mellitus. American Journal of Medicine 98: 22–31.Wilder J (1943). Psychological problems in hypoglycemia. American Journal of Digestive Diseases10: 428–35.Wirsen A, Tallroth G, Lindgren M, Agardh C-D (1992). Neuropsychological performance differsbetween type 1 diabetic and normal men during insulin-induced hypoglycaemia. Diabetic Medicine9: 156–65.Ziegler D, Hubinger A, Muhlen H, Gries FA (1992). Effects of previous glycaemic control on theonset and magnitude of cognitive dysfunction during hypoglycaemia in type 1 (insulin-dependent)diabetic patients. Diabetologia 35: 828–34.

3 Frequency, Causes and RiskFactors for Hypoglycaemia inType 1 DiabetesMark W.J. StrachanINTRODUCTIONHypoglycaemia was first described in humans in the early years of the 20th century, but didnot become firmly established as a pathophysiological entity until the discovery of insulinin 1922. Despite the substantial advances in insulin therapy and blood glucose monitoringthat have occurred in the subsequent 80 years, hypoglycaemia remains the most commoncomplication of type 1 diabetes (The Diabetes Control and Complications Trial ResearchGroup, 1993) and generates as much anxiety in patients as the threat of advanced diabeticcomplications, such as blindness or renal failure (Pramming et al., 1991). Few people withtype 1 diabetes escape intermittent exposure and, as a result, hypoglycaemia is the principallimiting factor in achieving good glycaemic control (Cryer, 1994; Cryer et al., 2003). Themagnitude of the psychological and physical consequences of hypoglycaemia cannot beoverestimated and is considered in detail in other chapters of this book. In this chapterthe frequency of hypoglycaemia is described in people with type 1 diabetes, along with itsunderlying causes and risk factors.DEFINITIONS OF HYPOGLYCAEMIAAny attempt to consider critically the frequency of hypoglycaemia in clinical practice,requires definitive criteria for what constitutes an episode of hypoglycaemia. This poses animmediate difficulty because researchers of the epidemiology of hypoglycaemia have notemployed common definitions with shared specifications.Biochemical Definitions of HypoglycaemiaAt first glance, it would seem sensible to employ a biochemical definition of hypoglycaemia,specifying a given blood glucose concentration, below which hypoglycaemia wouldbe deemed to occur. However, it is not possible to provide such a precise biochemical criterionfor the diagnosis of hypoglycaemia (Service, 1995). As blood glucose concentrationsHypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

50 FREQUENCY, CAUSES AND RISK FACTORSdecline, a hierarchy of events occur at individual glycaemic thresholds, commencing withcounterregulation (arterialised blood glucose ∼38 mmol/l), impairment of different cognitivefunctions (∼32–26 mmol/l) and the onset of symptoms and neurophysiological changes(∼32–24 mmol/l). In clinical practice, however, it is usual for venous or capillary bloodglucose levels to be measured, and these are lower than contemporaneous arterialised bloodglucose concentrations (which are usually measured in research studies of hypoglycaemia)(Heller and Macdonald, 1996).In the non-diabetic individual, venous blood glucose concentrations below 3.0 mmol/lmay occur following an overnight fast or during the course of a prolonged oral glucosetolerance test (Service, 1995). Moreover, as is discussed later, the blood glucose thresholdsfor the onset of symptoms and counterregulation in patients with type 1 diabetesmay vary depending on the preceding or prevailing glycaemic control. Thus, patients withpoor glycaemic control may experience symptoms of hypoglycaemia at venous plasmaglucose concentrations substantially higher than 3.0 mmol/l (Boyle et al., 1988). Patientswith preceding strict glycaemic control may not experience the onset of symptoms of hypoglycaemiauntil venous plasma glucose concentrations have declined to below 2.0 mmol/l(Boyle et al., 1995). On a pragmatic basis, in routine clinical practice, Diabetes UK hasrecommended that individuals with diabetes should try to ensure that their blood glucoseconcentrations do not fall below 4.0 mmol/l (O’Neill, 1997), but this does not define hypoglycaemia.The American Diabetes Association has proposed a blood glucose concentrationof 3.9 mmol/l as representing hypoglycaemia (ADA Workgroup on Hypoglycemia, 2005),but this has been challenged as being too high.Clinical Definitions of HypoglycaemiaThe inability to agree on a biochemical definition for hypoglycaemia requires instead theapplication of clinical criteria (Box 3.1). The difficulty here is that because the symptomsof hypoglycaemia are not specific, and vary between individuals (see Chapter 2), the use ofsymptomatology alone may be unreliable and may result in the inclusion of episodes thatare not true hypoglycaemia. In one prospective study, where capillary blood glucose wasmeasured whenever a patient had symptoms suggestive of hypoglycaemia, only 29% of suchepisodes were accompanied by evidence of biochemical hypoglycaemia (i.e., blood glucose

FREQUENCY OF HYPOGLYCAEMIA 51Box 3.1Clinical definitions of hypoglycaemia• Asymptomatic hypoglycaemia – low blood glucose identified on routine blood test,with no associated symptoms.• Mild symptomatic hypoglycaemia ∗ – symptoms suggestive of hypoglycaemia;episode successfully treated by the patient alone.• Severe hypoglycaemia ∗ – assistance from a third party is required to effect treatment.• Profound hypoglycaemia ∗ – associated with permanent neurological deficits ordeath.∗ Contemporaneous demonstration of a low blood glucose concentration is not mandatory but, particularlyin the case of mild episodes, if it is available it does make the definition more robust.before an episode of hypoglycaemia can be considered to be severe, nor does itspecify any particular method of treating the low blood glucose. There is a substantialdifference in the degree of neuroglycopenia during an episode of hypoglycaemiathat has resolved following the oral administration of carbohydrate, and one that hasresulted in coma and required the administration of intramuscular glucagon or intravenousglucose.Mild hypoglycaemia is considered to occur when an episode is self-treated by the affectedindividual. Although the recording of episodes of severe hypoglycaemia is generally regardedas being a fairly robust measure, mild hypoglycaemia is a much ‘softer’ end-point. Symptomperception by the patient may be difficult and the glycaemic thresholds at which individualsdevelop symptoms of hypoglycaemia are very variable. In many clinical research studies, itis not uncommon to include a composite definition of mild hypoglycaemia, which includessymptomatic episodes that respond to self-treatment (with or without a confirmatory bloodtest) and asymptomatic episodes below an arbitrary biochemical blood glucose value (e.g.3.5 mmol/l) that have been identified through routine blood glucose monitoring (Andersonet al., 1997). The introduction by some investigators of a further category of ‘moderate’hypoglycaemia is not helpful and confuses the distinction between mild and severeepisodes.FREQUENCY OF HYPOGLYCAEMIAThe true frequency of hypoglycaemia in people with type 1 diabetes is difficult to estimateaccurately and not simply because of differences in the definitions of hypoglycaemia. Mostepisodes occur at home, at work or during leisure activities, without involving medical,nursing or paramedical staff, and the subsequent recall of such episodes by patients isgenerally poor, particularly with regard to mild hypoglycaemia. Prospective studies thereforehave a strong advantage over retrospective studies in their respective abilities to documentthe frequency of hypoglycaemia with accuracy. Another crucially important feature is thenature of the population of people with type 1 diabetes under study – factors such as the

52 FREQUENCY, CAUSES AND RISK FACTORSquality of glycaemic control (The Diabetes Control and Complications Trial Research Group,1997) and the presence of impaired awareness of hypoglycaemia (Gold et al., 1997) havea major influence on the frequency of hypoglycaemia. Participants in clinical interventionalstudies are often not representative of the wider body of people with type 1 diabetes. Forexample, in the DCCT (The Diabetes Control and Complications Trial Research Group,1993), the subjects were young, motivated and received substantial professional support(particularly those in the intensive group). They were pre-selected depending on their historyof hypoglycaemia. Thus, individuals were excluded if, in the previous two years, theyhad experienced more than one episode of severe hypoglycaemia causing neurologicalimpairment, without experiencing warning symptoms, or more than two episodes of seizureor coma, regardless of attributed cause. Such exclusions will have inevitably influenced thebaseline frequency of severe hypoglycaemia in the study groups.Small studies estimating the frequency of hypoglycaemia are likely to be biased becausechance variations in the prevalence of risk factors for hypoglycaemia can magnify (ordiminish) the frequency of hypoglycaemic events to a much greater extent than would beobserved in larger investigations. There is also likely to be a major period effect in studiesof the prevalence of hypoglycaemia. In the post-DCCT era, it might be anticipated thatthe incidence of hypoglycaemia would rise as patients and diabetes healthcare professionalssought to tighten glycaemic control (Johnson et al., 2002). This trend may be counterbalancedby the increased use of home blood glucose monitoring, improved educational programmesand the greater use of insulin analogues and continuous subcutaneous insulin infusion therapy(Chase et al., 2001).Thus, estimates of the frequency of hypoglycaemia must be considered in relation to thedefinitions employed and the characteristics of the patients studied. Comparisons betweenpresent-day and historical studies must be made in the knowledge that treatment targets,methods of monitoring blood glucose and insulin therapy have changed greatly in recentyears.Frequency of Mild, Symptomatic HypoglycaemiaThe features of several studies that have examined the frequency of mild hypoglycaemia,either prospectively or retrospectively, in adults with type 1 diabetes are shown in Table 3.1.Rates of mild hypoglycaemia vary substantially, ranging from 8 to 160 episodes per patientper year. However, it is extremely difficult to make direct comparisons between individualstudies because of differences in methodology and patient characteristics.Retrospective studiesTwo retrospective studies by a Danish group (Pedersen-Bjergaard et al., 2001; Pedersen-Bjergaard et al., 2004) were very similar in their patient characteristics, most of whom wereadministering four or more injections of insulin per day and had moderate glycaemic control.On being asked to recall the number of episodes of mild, symptomatic hypoglycaemiaexperienced in the preceding week, the subjects reported a frequency of two episodes perpatient per week. The strength of these studies lies in the large numbers of patients examinedand the short period of recall. This should have minimised inaccuracy, but may have beenan unrepresentative time frame.

Table 3.1 Frequency of mild hypoglycaemia in adults with type 1 diabetesAuthorsNumber ofpatients Follow-up a Treatment b cHbA 1c (%)Definition ofhypoglycaemiaFrequency d(number/pt/yr)Pramming et al., 1991 411 1 week (P) 78% twice daily 87 Symptomatic 94Janssen et al., 2000 31 6 weeks (P) All multiple 72 BG < 35 mmol/l 160Pedersen-Bjergaard 201 1 week (R) 86% multiple 86 Symptomatic 104et al., 2001Pedersen-Bjergaard 170 12 months (P) 87% ≥ 4 inj 84 Symptomatic 88et al., 2003aPedersen-Bjergaard 1076 1 week (R) 72% multiple 86 Symptomatic 104et al., 2004Donnelly et al., 2005 94 4 weeks (P) 49% S&I f 83 Symptomatic 42Leckie et al., 2005 243 (27 T2) e 12 months (P) 80% multiple 91 Symptomatic 8UK HypoglycaemiaStudy Group, 200746 (< 5 yrs)54 (> 15 yrs)9–12 months All ≥ 2 injections 73 Symptomatic and/orBG < 30 mmol/l3578 29This table does not include early studies which tended to focus on rates of mild hypoglycaemia in particular subsets of patients, e.g. with impaired awareness of hypoglycaemia or individuals usingporcine versus human insulin.a Duration of follow-up: P = prospective estimation of hypoglycaemia frequency; R = retrospective.b Multiple refers to preprandial injections (inj) of soluble (or analogue) insulin and bedtime isophane insulin.c Mean HbA1c for the subjects under study.d Frequency of hypoglycaemia in each study has been adjusted to represent an annual rate per person.e 27 Patients in this study had type 2 diabetesf Patients were on soluble and isophane insulins, but frequency of injections not specified.BG = capillary blood glucose concentration.

54 FREQUENCY, CAUSES AND RISK FACTORSProspective studiesProspective studies offer the potential to provide more convincing data on frequency of mildhypoglycaemia, but substantial differences in prevalence were again reported. In an earlierstudy of 441 patients with type 1 diabetes, managed principally with a twice daily insulinregimen containing soluble and isophane insulins, the weekly average was 1.8 episodes ofmild symptomatic hypoglycaemia (Pramming et al., 1991). These patients had moderateglycaemic control and the period of assessment was one week. This study may be of lessrelevance today in view of the current use of intensive insulin therapy and insulin analogues,yet it is interesting that the rate of mild hypoglycaemia is unchanged today.A Danish prospective study (Pedersen-Bjergaard et al., 2003a) included patients withsimilar characteristics to those in two retrospective studies by the same group (Pedersen-Bjergaard et al., 2001; Pedersen-Bjergaard et al., 2004). Hypoglycaemia was recordedmonthly with episodes of mild symptomatic hypoglycaemia being reported for the precedingweek. Mild hypoglycaemia occurred on average 1.7 times per patient per week. Subjectswere also asked to perform a monthly five-point blood glucose profile and to record in additionany blood glucose value below 3.0 mmol/l. Measurements demonstrating biochemicalhypoglycaemia represented 3.7% of all blood glucose readings.In a community-based study in Tayside, Scotland, 94 adults with type 1 diabetes wereselected at random from a regional diabetes database and were asked to record episodes ofhypoglycaemia prospectively over one month (Donnelly et al., 2005). Their median age was40 years, median duration of diabetes was 18 years and median HbA 1c was 8.3%. Biphasicinsulin was used by 35% of participants and 49% used a combination of intermediate andshort-acting insulins (although the frequency of injections was not reported). A total of325 episodes of mild hypoglycaemia occurred, representing a rate of 41.5 episodes perperson per year, i.e., approximately half the rate reported by Pramming et al., (1991) andPedersen-Bjergaard et al., (2003a). The study has weaknesses; it was relatively small anddata on frequency of blood glucose monitoring were limited. The precise criteria for definingmild hypoglycaemia were not clearly described and, in particular, the role of contemporaneousmonitoring data was not specified. Nevertheless, the subjects were probably veryrepresentative of the population of people with type 1 diabetes in that region.Similar data have been reported from a multicentre study from the United Kingdom (UKHypoglycaemia Study Group, 2007). The primary aim of this study was to compare thefrequencies of hypoglycaemia in individuals with different types and durations of diabetes,receiving different treatment modalities. As part of the study, 50 adults with type 1 diabetesof duration less than five years and 57 adults with type 1 diabetes of greater than 15 yearsduration were recruited. All subjects used two or more injections of insulin per day andtheir glycaemic control was good (mean HbA 1c < 80%). The participants were followed forbetween 9–12 months (mean 10 months) and were asked to report all episodes of symptomatic,self-treated hypoglycaemia and episodes where blood glucose was less 3.0 mmol/l,regardless of symptomatology. Subjects were given forms to record such episodes and wereencouraged to record contemporaneous blood glucose levels. To maximise compliance,subjects were asked to send in completed forms every month to the local research centre,including when no episodes of hypoglycaemia had occurred. If no forms were received,telephone contact was made with the subjects. Using this robust methodology, mean ratesof hypoglycaemia of 35 and 29 episodes per person per year were reported for the shortand long duration groups respectively. The distribution of episodes was much skewed, with

FREQUENCY OF HYPOGLYCAEMIA 55some individuals reporting no events and others in excess of 200 per year; overall, approximately85% of individuals experienced at least one episode of mild hypoglycaemia. Theauthors did not report the relative proportion of symptomatic and asymptomatic episodes, orthe proportion of symptomatic episodes with a corroborative blood test, but these data areextremely informative and indicate that duration of diabetes has little impact on the frequencyof mild hypoglycaemia, an observation that has been made before (Pedersen-Bjergaardet al., 2004).In a 12 month prospective study of 243 insulin-treated adults, Leckie and colleaguesreported that mild, symptomatic hypoglycaemia occurred with a frequency of only eightepisodes per patient per year (Leckie et al., 2005). This is one of the lowest rates to bereported in a large group of people, most of whom had type 1 diabetes. However, severalreasons may account for this result. The subjects had suboptimal glycaemic control, with amean HbA 1c of 9.1%. A small number of people with insulin-treated type 2 diabetes wereincluded, who might be expected to have had a lower overall frequency of hypoglycaemia.The prevalence of impaired awareness of hypoglycaemia, a recognised risk factor for hypoglycaemia(see Chapter 7), was exceptionally low in this cohort at 3%. The proportion ofsubjects using insulin analogues was not reported, but it would certainly have been higherthan in studies from the early 1990s. The main strength of the study, namely a long periodof follow up, may have been an inadvertent weakness by causing ‘patient fatigue’, i.e.,participants may have been less assiduous in recording episodes of mild hypoglycaemia asthe study progressed. Finally, this was primarily a study of the impact of hypoglycaemiaoccurring in the work place. Thus, all of the participants were in full-time employment and sorepresented an atypical group of individuals who may have adopted strategies to reduce thefrequency of hypoglycaemia because of the potentially adverse effects that this could haveon their jobs.Frequency of Asymptomatic, Biochemical HypoglycaemiaAside from the problem of specifying a biochemical threshold for hypoglycaemia, therecan be little doubt that asymptomatic hypoglycaemia is even more common than symptomatic,mild hypoglycaemia. However, a major determinant of the frequency of documentedasymptomatic hypoglycaemia is, necessarily, the frequency with which blood glucose ismeasured.Traditionally, two different approaches have been used to investigate this phenomenon:(a) taking multiple sequential blood samples in a controlled, experimental setting, over a prespecifiedtime period; and (b) inviting patients in the community to perform capillary bloodtests at multiple, pre-set time points. Providing the sampling frame is sufficiently frequent,the advantage of the former strategy is that it will capture all episodes of biochemicalhypoglycaemia, however brief, during the time period under study. The disadvantage isthat it is labour intensive for investigators and so only a small number of subjects can bestudied over a relatively short time period, such as 12 or 24 hours. By contrast, the use ofhome blood glucose monitoring, particularly with meters that store the results electronically,allows larger numbers of subjects to be studied over very prolonged time periods. Typically,subjects are asked to perform periodic seven-point profiles, i.e., blood glucose estimationsbefore each meal, two hours after food and during the night. The obvious disadvantages ofthis mode of investigation are that subjects may forget to perform the relevant monitoring,

56 FREQUENCY, CAUSES AND RISK FACTORSor may do so at the wrong times, and that episodes of asymptomatic hypoglycaemia mayoccur outside the sampling time-frames and, thus, be missed.Thorsteinsson et al., (1986) examined seven-point capillary blood glucose profiles in99 adults with type 1 diabetes and demonstrated that the frequency of biochemical hypoglycaemiawas inversely related, in a curvilinear manner, to the median blood glucoseconcentration. Thus, for example, in patients who had a median blood glucose concentrationof 5.0 mmol/l, 10% of blood glucose levels were less than 3.0 mmol/l. By contrast, only 2.5%of blood glucose levels were below 3.0 mmol/l, in patients whose median blood glucoseconcentration was 10 mmol/l (Figure 3.1; Thorsteinsson et al., 1986).In a separate study, nocturnal blood glucose profiles were examined in 31 patients withtype 1 diabetes using multiple injection therapy with soluble and isophane (NPH) insulin(mean HbA 1c 8.6%). Venous blood samples were taken every 30 minutes from 11 p.m. until7.30 a.m.. Nocturnal hypoglycaemia (blood glucose less than 3.0 mmol/l) occurred on 29%of occasions and 67% of these episodes were asymptomatic (Vervoort et al., 1996). Sixindividuals were studied on two separate nights and the blood glucose profiles, perhapspredictably, showed considerable intra-individual variation.Figure 3.1 Correlation between the median blood glucose concentration and the frequencies of bloodglucose concentrations below 4.0, 3.0, 2.5 and 2.0 mmol/l in adults with type 1 diabetes. Solid linesrepresent data from 70 adults on twice daily insulin therapy and dotted lines represent data from20 adults treated with continuous subcutaneous insulin. Approximately 10% of readings were below3.0 mmol/l, when median blood glucose concentration was 5.0 mmol/l. Reproduced with permissionfrom Thorsteinsson et al. (1986) © John Wiley & Sons, Ltd

FREQUENCY OF HYPOGLYCAEMIA 57Janssen et al., (2000) have studied, prospectively, the frequency of biochemical hypoglycaemiaover a six week period in 31 people with type 1 diabetes with a mean HbA 1c of7.2% (all had a HbA 1c ≤ 83%). Subjects were all using a multiple injection regimen withsoluble and isophane insulins and performed one seven-point blood glucose profile and sixfour-point profiles each week. No overnight readings were performed. Patients completeda mean of 82% of the required monitoring schedule and overall experienced a mean of18.7 episodes of hypoglycaemia (i.e., approximately 160 episodes per year). The range waswide, however, at between zero and 41 episodes per individual, over the six weeks ofthe study.The more widespread application of continuous glucose monitoring systems may help toelucidate the frequency of mild hypoglycaemia with greater accuracy. In one study, 65%of people with type 1 diabetes monitored over three days had an episode of asymptomatichypoglycaemia (interstitial glucose < 33 mmol/l) (Bode et al., 2005). However, much workstill needs to be done to clarify the relationship between interstitial glucose and blood glucoseconcentrations, before this can be regarded as a robust tool for detecting hypoglycaemia (seeChapter 5).Frequency of Severe HypoglycaemiaAs with mild hypoglycaemia, direct comparisons between individual studies on the frequencyof severe hypoglycaemia are not straightforward. However, it is a more robust end-pointthan mild hypoglycaemia and, because episodes typically have a more profound effect onindividuals, retrospective recall is much more reliable. At the conclusion of their prospectivestudy, Pedersen-Bjergaard et al. (2003b), showed that 90% of subjects recalled correctlytheir experience of severe hypoglycaemia over the preceding year (Figure 3.2). Subjectswho had experienced a high incidence of severe hypoglycaemia (prospectively recorded),retrospectively underestimated the overall rate by around 15%.Table 3.2 lists some of the large surveys (each in excess of 100 participants) that haveexamined the frequency of severe hypoglycaemia. The table focuses primarily on studiesexamining unselected groups of individuals with type 1 diabetes, and so excludes interventiontrials (The Diabetes Control and Complications Trial Research Group, 1993; Reichard andPihl, 1994; MacLeod et al., 1995) and studies that examined particular sub-groups of patients,such as people with impaired awareness of hypoglycaemia (Gold et al., 1994; MacLeodet al., 1994; Clarke et al., 1995), differing durations of diabetes (UK Hypoglycaemia StudyGroup, 2007) or subjects who have received intensive therapy or education (Muhlhauseret al., 1985; Pampanelli et al., 1996; Bott et al., 1997). A study that examined a mixedgroup of children, adolescents and adults (Allen et al., 2001) has also been excluded. Thefrequency of severe hypoglycaemia (defined as episodes requiring third party assistance) inthe studies listed in Table 3.2 is remarkably consistent at 1.0 to 1.6 episodes per patientper year.However, it is important to note that the frequency of severe hypoglycaemia in unselectedpopulations does not follow a Gaussian distribution (Figure 3.3). The distribution isheavily skewed such that the majority of individuals do not experience any severe hypoglycaemiain a given year, while a small number of individuals have recurrent episodes. Inthe studies highlighted in Table 3.2, between 30–40% of individuals experienced at leastone episode of severe hypoglycaemia over the period in question. The proportion affected

58 FREQUENCY, CAUSES AND RISK FACTORS15Recalled severe hypoglycaemia(episodes per patient-year)10500 5 10 15Prospectively recorded severe hypoglycaemia(episodes per patient-year)Figure 3.2 Correlation between prospectively recorded and retrospectively recalled rate of severehypoglycaemia over the same one year period in 230 people with type 1 diabetes. Marker sizes areweighted by the number of cases. R 2 = 066; p < 0001. Reproduced with permission from Pedersen-Bjergaard et al. (2003b) © John Wiley & Sons, Ltdin the study by Pramming et al., (1991) was much lower, but the period of follow up wasonly one week. This skewed distribution serves to emphasise further the importance ofpatient selection in ascertaining the frequency of severe hypoglycaemia with accuracy, as theexclusion of a relatively small number of people at high risk would substantially reduce theoverall risk.In three of the studies of unselected adults with type 1 diabetes, severe hypoglycaemia wasfurther subdivided to examine episodes associated with more significant neuroglycopenia,i.e., those resulting in coma and/or seizures (Table 3.2) (ter Braak et al., 2000; Pedersen-Bjergaard et al., 2003a; Pedersen-Bjergaard et al., 2004). Furthermore, in the study byMuhlhauser et al., (1998) only episodes treated with intra-muscular glucagon or intravenousglucose were addressed. Predictably, such events were rarer and represented about onequarter of all episodes of severe hypoglycaemia.Emergency and hospital services will occasionally be involved in the management ofsevere hypoglycaemia and there are data on the use of such agencies. This clearly hasto be interpreted with caution, as the majority of all hypoglycaemia is managed in thecommunity, without involvement of (para)clinical staff. Individuals admitted to hospitalwith hypoglycaemia are probably atypical, and have an increased prevalence of alcoholdependence and mental illness (Hart and Frier, 1998). An early study from Australia reportedthat, over one year, 3.5% of people attending an urban diabetes clinic had an episode ofhypoglycaemia severe enough to warrant referral to hospital (Moses et al., 1985). Morerecent data from Tayside, Scotland, demonstrated that 7.1% of people with type 1 diabetes

Table 3.2 Frequency of severe hypoglycaemia in adults with type 1 diabetesAuthors Number of patients Follow-up a Treatment b HbA 1cc(%)Definition ofhypoglycaemiaFrequency d(episodes/pt/yr)ProportionAffected (%)Pramming et al., 1991 411 1 week (P) 78% twice daily 87 Third party 14 3MacLeod et al., 1993 600 (56 T2) e 12 months (R) 76% twice daily 107 A1 Third party 16 29Muhlhauser et al.,1998684 12 months (R) 70% > 2 inj 80 Glucagon/glucose 021 13ter Braak et al., 2000 195 12 months (R) 82% intensive 78 Third partyComa/seizurePedersen-Bjergaardet al., 2001Pedersen-Bjergaardet al., 2003aPedersen-Bjergaardet al., 200415 4104207 24 months (R) 86% ≥ 4 inj 86 Third party 11 NR170 12 months (P) 87% ≥ 4 inj 84 Third partyComa/seizure1076 12 months (R) 72% ≥ 4 inj 86 Third partyComa/seizure11 390313 37035Leckie et al., 2005 243 (27 T2) e 12 months (P) 80% multiple 91 Third party 098 34UK HypoglycaemiaStudy Group46 (15 years)9–12 months All ≥ 2 injections 7.37.8Third party 1.13.22246This table only considers studies examining in excess of 100 subjects and does not include early studies which tended to focus on rates of severe hypoglycaemia in particular subsets of patients, e.g.with impaired awareness of hypoglycaemia or individuals using porcine versus human insulin.a Duration of follow-up: P = prospective estimation of hypoglycaemia frequency; R = retrospective.b Multiple refers to preprandial injections (inj) of soluble (or analogue) insulin and bedtime isophane insulin.c Mean HbA1c for the subjects under study; A1 = figure is HbA 1 .d Frequency of hypoglycaemia in each study has been adjusted to represent an annual rate per person.NR = data not reported.e 56 include patients with type 2 diabetes.

60 FREQUENCY, CAUSES AND RISK FACTORS90807060% of patients504030201000 2 4 6 8 10Severe hypoglycaemia, episodes per yearFigure 3.3 Distribution of self-reported number of episodes of severe hypoglycaemia during thepreceding year in 1049 unselected patients with type 1 diabetes (light bars) and 209 patients selectedby criteria to mimic the characteristics of the DCCT cohort (dark bars). Reproduced with permissionfrom Pedersen-Bjergaard et al. (2004) © John Wiley & Sons, Ltdhad an episode of hypoglycaemia that required contact with emergency medical teams overone year, with an incidence rate of 0.12 per patient per year (Leese et al., 2003). Thus, onlyabout one in 10 of all episodes of severe hypoglycaemia result in contact with emergencyservices.Frequency of Hypoglycaemia: Summary and ConclusionsThe literature on frequency of hypoglycaemia is heterogeneous and inconsistent and, thusthere are considerable methodological limitations in our ability to ascertain accurately thefrequency of hypoglycaemia. The definitive study, a prolonged, prospective evaluation of alarge number of unselected people with type 1 diabetes, outside ‘trial’ conditions, has yet to beperformed. Existing data on severe hypoglycaemia are probably fairly accurate, but estimatesof mild hypoglycaemia should be regarded with caution. However, the ‘average’ patientwith type 1 diabetes will probably experience about 1–2 episodes of mild hypoglycaemiaper week and be exposed to several thousand episodes over a lifetime with diabetes. Overall,severe hypoglycaemia may be expected to occur once or twice each year, but the distributionis heavily skewed, such that most individuals will be unaffected while a small number willhave multiple episodes. In the early 1980s, Robert Tattersall’s group in Nottingham reportedthat admission to hospital as a consequence of hypoglycaemia represented the ‘tip of aniceberg’ of all episodes of hypoglycaemia (Potter et al., 1982), and this observation doesnot appear to have changed.

CAUSES OF HYPOGLYCAEMIACAUSES OF HYPOGLYCAEMIA 61Although advances in insulin therapy have been made over the last 80 years, the administrationof exogenous insulin remains a very crude means of managing type 1 diabetes. Thetime-action profiles of the modern insulin analogues do not mimic the physiological changesin plasma insulin concentrations that occur in non-diabetic individuals (Figure 3.4) and,crucially, concentrations of exogenous insulin cannot respond to changes in blood glucoseconcentration. Therefore, at its most fundamental level, hypoglycaemia in people with type1 diabetes is the result of an imbalance between insulin-mediated glucose efflux from theblood stream and the amount of glucose entering the circulation from ingested carbohydrateand from the liver. Cryer et al. (2003), have grouped the causes of hypoglycaemia intype 1 diabetes into six categories (Box 3.2) depending on their relative effects on insulinFigure 3.4 Mean 24 hour plasma glucose and insulin profiles in 12 healthy non-diabetic individuals.Shaded areas represent 95% confidence intervals. Glucose levels remain within tight limits, while thereis considerable variation in insulin concentrations, particularly around meal times. Reprinted from TheLancet, 358, Owens et al., Insulins today and beyond 739–746 (2001), with permission from ElsevierBox 3.2Causes of hypoglycaemia in type 1 diabetes1. Inappropriate insulin injection – e.g. excessive dose, inappropriate time, inappropriateinsulin formulation.2. Inadequate exogenous carbohydrate – e.g. missed meal or snack, overnight fast.3. Increased carbohydrate utilisation – e.g. exercise.4. Decreased endogenous glucose production – e.g. excessive alcohol consumption.5. Increased insulin sensitivity – e.g. night time, exercise, weight loss.6. Decreased insulin clearance – e.g. renal failure.Modified from Cryer et al., 2003.

62 FREQUENCY, CAUSES AND RISK FACTORSconcentrations or sensitivity and on glucose entry into the circulation. Although this classificationis not perfect, it is a useful starting point for considering the causes of an episodeof hypoglycaemia.Patient ErrorIt is common after an episode of hypoglycaemia for the person with diabetes or, indeeda member of the diabetes team, to try to ascertain why the episode occurred. Althoughthere is always a risk of ‘spurious attribution’, in many instances an obvious cause canbe identified and this is often the result of an error of judgement. Carbohydrate intakemay have been inadequate because a meal was missed or delayed, or simply contained aninsufficient content of carbohydrate. Alternatively, the patient may have injected too muchinsulin relative to the amount of carbohydrate in the meal. It is also not uncommon forinsulin to be administered at an inappropriate time or for the ‘wrong’ insulin to be injectedaccidentally, e.g. a rapid-acting insulin analogue is administered at a time when basal insulinshould have been given. An often unrecognised problem is the inadequate re-suspensionof isophane (or lente) insulins or of fixed mixtures of insulin. If these insulins are notfully resuspended prior to subcutaneous injection, the insulin may be absorbed at variablerates resulting in unpredictable insulin levels and, thus, an increased risk of hypoglycaemia(Owens et al., 2001). Deliberate overdose of insulin is rare, but may result in protractedhypoglycaemia.AlcoholThe relationship of alcohol to hypoglycaemia is considered in detail in Chapter 5. Surveysbased on patient interviews have implicated alcohol in up to one fifth of episodes ofsevere hypoglycaemia requiring hospital admission (Potter et al., 1982; Moses et al., 1985;Feher et al., 1989; Hart and Frier, 1998). In a recent study of 141 people treated forsevere hypoglycaemia in three emergency centres in Copenhagen, alcohol was detectedin the blood of 17% of the subjects (Pedersen-Bjergaard et al., 2005). The median bloodalcohol concentration was 11 mmol/l. Alcohol inhibits gluconeogenesis and so may directlycontribute to the development of hypoglycaemia. In addition, there are some data to suggestthat alcohol attenuates the counterregulatory response to hypoglycaemia (Avogaro et al.,1993; Turner et al., 2001). However, the main impact of alcohol probably centres onits ability to impair awareness of hypoglycaemia and so hinder the ability of individualsto take appropriate corrective therapy (Kerr et al., 1990). Thus, an individual underthe influence of alcohol may not recognise that he or she is hypoglycaemic, and evenif the symptoms are recognised, the person may not have the capacity to self-treat. Anepisode of mild hypoglycaemia may therefore be converted rapidly into a severe episode.Moreover, friends, colleagues or bystanders may presume that the neuroglycopenic symptomsand signs exhibited by the individual are a consequence of alcohol intoxicationand so again appropriate treatment and medical help may not be provided. It is forthese reasons that alcohol is implicated in many instances of profound and protractedinsulin-induced hypoglycaemia associated with permanent neurological damage (Arky et al.,1968).

RISK FACTORS FOR SEVERE HYPOGLYCAEMIA 63ExerciseExercise is recommended for people with type 1 diabetes because of its positive physiologicaland psychological effects. However, exercise can increase the risk of hypoglycaemia bothduring the physical activity itself and in the recovery period (see Chapter 14). The reasonsfor the high incidence of exercise-induced hypoglycaemia in adults with type 1 diabeteshave not been fully elucidated. Moderate exercise in non-diabetic individuals causes a fallin plasma insulin to 40–50% of pre-exercise levels (Galassetti et al., 2003). This normalphysiological decline cannot occur when insulin is being administered exogenously, unlessthe dose is reduced before exercise commences (Sonnenberg et al., 1990). An acute increasein insulin sensitivity following exercise also increases the risk of hypoglycaemia (Sonnenberget al., 1990).The metabolic and counterregulatory hormonal responses to acute hypoglycaemia andexercise are qualitatively very similar. Antecedent exercise blunts the counterregulatoryresponse to subsequent acute hypoglycaemia (Galassetti et al., 2001a; Galassetti et al.,2001b). The inverse situation also applies, i.e., antecedent hypoglycaemia diminishes thecounterregulatory response to subsequent exercise (Galassetti et al., 2003). This means thatat the very time that metabolic substrate requirements are increasing, there is the potentialfor an acute failure of endogenous glucose production. Thus, impaired counterregulatoryresponses may be an important mechanism in promoting exercise-related hypoglycaemia intype 1 diabetes.RISK FACTORS FOR SEVERE HYPOGLYCAEMIAIn a relatively high proportion of cases – in some series as high as 40% (Potter et al., 1982;Feher et al., 1989) – it is not possible to identify the precipitating cause of an episode ofhypoglycaemia. Indeed, this figure may be even higher, because the phenomenon of ‘spuriousattribution’ means that in some instances the perceived precipitant of a given episode mayhave been an innocent bystander. It has, therefore, been increasingly recognised that diabeteshealthcare professionals must look beyond conventional precipitating factors and considerother phenomena which may be associated with an increased risk of hypoglycaemia, namely,the risk factors for hypoglycaemia (Table 3.3). Many of these are discussed in more detailin other chapters.Intensive Insulin TherapyRandomised trials, most notably the DCCT, have provided substantial data on the epidemiologyof hypoglycaemia in adults with type 1 diabetes and, in particular, on the impact ofintensive insulin therapy.The Diabetes Control and Complications Trial (DCCT)The DDCT was a landmark study and provided diabetes specialists with the long-awaitedproof that strict glycaemic control limited the incidence and severity of microvascularcomplications in people with type 1 diabetes. A total of 1441 patients with type 1 diabetes

Table 3.3 Risk factors for severe hypoglycaemia in adults with type 1 diabetesStudyIntensivePrevioustherapy Low HbA 1c episode DurationImpairedawarenessC-peptidenegative Sex AgeInsulindose ACE activityMuhlhauser et al., 1985 − − −Muhlhauser et al., 1987 − −MacLeod et al., 1993 − + − − −Gold et al., 1994 +EURODIAB, 1994 + +MacLeod et al., 1994 +SDIS, 1994 +Clarke et al., 1995 +Pampanelli, et al., 1996 −Bott et al., 1997 − + + + +DCCT, 1997 + + + + + M + +Gold et al., 1997 − + + + +Muhlhauser et al., 1998 − − + − + + − − −Pedersen-Bjergaard et al., 2001 + ∗ + + + +ter Braak et al., 2000 − − + − − +Leese et al., 2003 − + − +Pedersen-Bjergaard et al., 2003 − + ∗ + − − − +Pedersen-Bjergaard et al., 2004 − − + − − − −Leckie et al., 2005 − + − F − −UK Hypoglycaemia + +Study Group, 2007This table examines the risk factors for severe hypoglycaemia that have been most commonly examined in adults with type 1 diabetes. Studies examining predominantly mild hypoglycaemia werenot included. +=positive association between risk factor and severe hypoglycaemia; −=no association between risk factor and severe hypoglycaemia. ∗ = only a risk factor if awareness ofhypoglycaemia not included. M = severe hypoglycaemia more common in men; F = severe hypoglycaemia more common in women.

RISK FACTORS FOR SEVERE HYPOGLYCAEMIA 65were randomly assigned either to intensive insulin therapy (based on multiple injection insulinregimens or continuous subcutaneous insulin infusion therapy) or to conventional insulintherapy (one or two insulin injections daily) (The Diabetes Control and Complications TrialResearch Group, 1993). In the conventional group, patients did not generally perform homeblood glucose monitoring, clinical reviews were undertaken every three months and patientswere not informed about their HbA 1c result, unless it was in excess of 13%. By contrast, inthe intensive therapy group, subjects performed frequent home blood glucose monitoring,had monthly visits with the study team and also had frequent telephone contact to achieve asstrict glycaemic control as possible. Over a mean follow-up of 6.5 years, the average HbA 1cconcentration in the intensive group was approximately 7.0% and in the conventional groupapproximately 8.8% (The Diabetes Control and Complications Trial Research Group, 1993).The subjects recruited to this study were not typical of the wider population of people withtype 1 diabetes. The participants had greater motivation and individuals were not permittedto take part if, in the preceding two years, they had experienced more than one episodeof severe neurological impairment without warning symptoms of hypoglycaemia, or morethan two episodes of seizure or coma, regardless of attributed cause. Moreover, during thestudy itself, the occurrence of an episode of severe hypoglycaemia in an individual patientprompted a review of conventional risk factors and, in instances where probable causes wereidentified, corrective actions such as re-educating the patient were undertaken (The DiabetesControl and Complications Trial Research Group, 1997).Despite all these factors, 3788 episodes of severe hypoglycaemia occurred in the 1441patients over the course of the study and, of these, 1027 were associated with coma and/orseizure (The Diabetes Control and Complications Trial Research Group, 1997). The rateof severe hypoglycaemia in the intensively-treated patients was 0.61 per patient per year,while that in the conventionally-treated group was 0.19 per patient per year, i.e., a three-folddifference. Over the mean of 6.5 years of follow-up, 65% of the intensive group experiencedat least one episode of severe hypoglycaemia, compared with 35% of the individuals in theconventional group.The Stockholm Diabetes Intervention Study (SDIS)The SDIS was a much smaller study than the DCCT, but had similar aims and objectives.A group of 102 patients with type 1 diabetes were recruited and 89 remained after 7.5 yearsof follow-up (Reichard and Pihl, 1994). The mean HbA 1c was 7.1% in the intensive groupand 8.5% in the conventional group. Severe hypoglycaemia, defined as episodes requiringthird party assistance, occurred in 80% of subjects in the intensive group and 58% of those inthe conventional group over the follow-up period. Overall, the rate of severe hypoglycaemiawas 1.1 per patient per year in the intensive group and 0.4 per patient per year in theintensive group. Other risk factors for severe hypoglycaemia were not reported.The Bucharest-Dusseldorf StudyThe DCCT was a multicentre study; 27 out of the 29 centres reported that intensive therapywas associated with an increased risk of severe hypoglycaemia, but in two centres noincreased risk was observed (The Diabetes Control and Complications Trial Research Group,1997). This led some investigators to claim that with appropriate education and training

66 FREQUENCY, CAUSES AND RISK FACTORSintensive insulin therapy need not necessarily be associated with an increased risk of hypoglycaemia(Plank et al., 2004). In the Bucharest-Dusseldorf Study, 300 individuals withtype 1 diabetes were randomised to: (a) conventional therapy for one year and then to oneyear of intensive therapy; (b) two years of intensive therapy; or (c) a four-day in-patientgroup teaching programme with conventional insulin therapy for one year (Muhlhauseret al., 1987). Glycated haemoglobin (HbA 1 ), remained unchanged at around 12–13% duringconventional therapy, but fell to ∼95% during intensive therapy. In the first year of thestudy, severe hypoglycaemia occurred in 6% of the intensively-treated patients and 12% ofthe conventionally-treated patients, and in year two the proportion of patients experiencingsevere hypoglycaemia in both intensive groups fell to 3–4%.Observational dataThe Dusseldorf team reported observational data on the impact of intensive therapy on thefrequency of severe hypoglycaemia (Bott et al., 1997). A total of 636 people with type1 diabetes who had participated in a structured five-day in-patient treatment and teachingprogramme for intensification of insulin therapy in one of ten different hospitals in Germany,were re-examined at intervals over six years. The mean HbA 1c fell from 8.3% to 7.6%.Severe hypoglycaemia, defined as episodes treated with intramuscular glucagon or intravenousglucose, decreased from 0.28 episodes per patient per year in the year preceding theprogramme to 0.17 episodes per patient per year afterwards (Bott et al., 1997). The variationin incidence of severe hypoglycaemia between different centres ranged from 0.05–0.27episodes per patient per year.Pampanelli and colleagues (1996) from Perugia reported retrospective data on 112 individualswho had been commenced on intensive insulin therapy (preprandial soluble insulinand bedtime isophane insulin) at diagnosis of diabetes and who were attending clinic atleast four times per year. Mean HbA 1c was 7.2% and mean duration of diabetes was 7.8years. Frequency of mild hypoglycaemia was estimated by a review of the patients’ bloodglucose monitoring diaries in the nine months prior to study, and the overall rate was 35.6episodes per patient per year. Severe hypoglycaemia, defined as episodes requiring thirdparty assistance, was assessed retrospectively over the duration of diabetes and was recalledby only six patients (representing an overall rate of 0.001 episodes per patient per year).SummaryThus, in the DCCT and the SDIS, intensive insulin therapy was associated with a nearlythree-fold increase in the risk of severe hypoglycaemia. By contrast, in other studies, theincidence of severe hypoglycaemia fell when intensive therapy was coupled with a detailededucation programme. The teams from Dusseldorf and Perugia would argue that the highquality of their education programmes resulted in the low frequencies of observed severehypoglycaemia. Although this may be the case, the importance of patient selection in allof these studies cannot be over-emphasised. In their study of 1076 adults with type 1diabetes from the United Kingdom and Denmark, Pedersen-Bjergaard et al. (2004) examineda subset of 230 individuals whose clinical characteristics were similar to patients enrolledin the DCCT. This sub-group accounted for only 5.4% of all reported episodes of severehypoglycaemia, with an overall rate of 0.35 episodes per patient per year, i.e., approximately

RISK FACTORS FOR SEVERE HYPOGLYCAEMIA 67one quarter that of the entire group. Risk factors for severe hypoglycaemia in this group(impaired awareness and retinopathy) were different from that of the study population as awhole (see Table 3.3). Thus, the DCCT patients were not representative of people with type1 diabetes, whereas the education programme undertaken in Dusseldorf is not somethingthat is widely replicated in mainstream diabetes practice. Risk factors for hypoglycaemiamay differ according to the specific characteristics of individuals being studied.Thus, in conclusion, although most specialists would accept that severe hypoglycaemia ismore common in patients receiving intensive insulin therapy, it is not inevitable and betterpatient education may actually reduce the incidence.Strict Glycaemic ControlStrict glycaemic control, as evidenced by glycated haemoglobin concentrations that approachthe upper end of the non-diabetic range, is closely linked to intensive insulin therapy. Sincethe DCCT results were published, glycated haemoglobin concentrations at or around thislevel have become the usual target for many people with type 1 diabetes. In the DCCT, therewas a quadratic relationship between HbA 1c and risk of severe hypoglycaemia (Figure 3.5),with the risk of hypoglycaemia increasing as HbA 1c decreased (The Diabetes Control andComplications Trial Research Group, 1997). However, the attained HbA 1c did not accountfor all the difference in risk of severe hypoglycaemia between the two arms of the study,as subjects in the intensive group still had an excess risk of severe hypoglycaemia after10080Rate per 100 Patient Years604020*** **0* * ** * * **5 6 7 8 9 10 11 12 13 14HbA 1c (%) During StudyFigure 3.5 Risk of severe hypoglycaemia as a function of monthly updated HbA 1c in the DiabetesControl and Complications Trial. The circles represent data from the intensive group and asterisks datafrom the conventional group. The bold solid and bold dashed lines represent the regression plots foreach group, and the non-bold dashed lines on either side show the upper and lower 95% confidencebands; DCCT (1997). Copyright © 1997 American Diabetes Association. Reprinted with permissionfrom The American Diabetes Association

68 FREQUENCY, CAUSES AND RISK FACTORSstatistical adjustment for HbA 1c concentration. Indeed, in the intensive group, only about 5%of the variation in frequency of severe hypoglycaemia could be explained by the glycatedhaemoglobin concentration (The Diabetes Control and Complications Trial Research Group,1997).Other studies have reported variable relationships between glycaemic control andfrequency of severe hypoglycaemia. In the study by Bott et al., (1997) a lower mean HbA 1cwas associated with severe hypoglycaemia, but there was no linear or quadratic relationshipbetween HbA 1c and severe hypoglycaemia. In the EURODIAB IDDM ComplicationsStudy, severe hypoglycaemia (defined as episodes requiring third party assistance) occurredin 32% of individuals over 12 months. A clear relationship to glycated haemoglobin wasevident, in that 40% of individuals with an HbA 1c < 54% were affected compared with24% of individuals with a HbA 1c > 78% (The EURODIAB IDDM Complications StudyGroup, 1994). In the study of workplace hypoglycaemia, recurrent severe hypoglycaemiawas associated with strict glycaemic control, but no link was found in individuals who hadexperienced only one episode (Leckie et al., 2005). In several other studies, no relationshipwas observed between severe hypoglycaemia and glycated haemoglobin (Muhlhauseret al., 1985; Muhlhauser et al., 1987; MacLeod et al., 1993; Gold et al., 1997; Muhlhauseret al., 1998; ter Braak et al., 2000; Pedersen-Bjergaard et al., 2001; Leese et al., 2003;Pedersen-Bjergaard et al., 2003a; Pedersen-Bjergaard et al., 2004) after adjustment for otherrisk factors.Glycated haemoglobin does not, of course, provide the entire picture about an individual’sglycaemic control and although HbA 1c may not predict risk of hypoglycaemia, low meanhome blood glucose concentrations and excessive variability in blood glucose can identifyindividuals more prone to hypoglycaemia (Cox et al., 1994; Janssen et al., 2000).Thus, a straightforward relationship does not exist between severe hypoglycaemia andglycaemic control. For an episode of mild hypoglycaemia to progress to one that causessignificant neuroglycopenia and impairs consciousness, other factors must operate, whichnegate the normal symptomatic and hormonal responses to hypoglycaemia.Acquired Hypoglycaemia SyndromesWithin each treatment group of the DCCT, the number of previous episodes of severehypoglycaemia was the strongest predictor of risk of future episodes (The Diabetes Controland Complications Trial Research Group, 1997). Moreover, approximately 30% of patientsin each group experienced a second episode of severe hypoglycaemia within four monthsfollowing an initial episode. The importance of a previous history of severe hypoglycaemiain predicting future risk has been replicated in several other studies (MacLeod et al., 1993;Bott et al., 1997; Gold et al., 1997; Muhlhauser et al., 1998). Moreover, as demonstrated inTable 3.3, many studies have also linked increased duration of diabetes with an increasedrisk of severe hypoglycaemia.Cryer has suggested that the integrity of the glucose counterregulatory system may be apivotal factor in determining whether the relative or absolute hyperinsulinism that frequentlyoccurs in insulin-treated diabetes ultimately results in the development of hypoglycaemia(Cryer, 1994; Cryer et al., 2003). Three acquired hypoglycaemia syndromes are associatedwith an increased risk of severe hypoglycaemia in people with type 1 diabetes. These areconsidered in greater detail in Chapters 6 and 7.

RISK FACTORS FOR SEVERE HYPOGLYCAEMIA 69Counterregulatory hormonal deficienciesHypoglycaemia-induced secretion of glucagon declines in most patients within five years ofdeveloping type 1 diabetes (Gerich et al., 1973; Bolli et al., 1983). A defective epinephrineresponse to hypoglycaemia may develop some years later (Bolli et al., 1983; Hirsch andShamoon, 1987; Dagogo-Jack et al., 1993). As with the glucagon response, the impairedepinephrine response is hypoglycaemia-specific but, in contrast to glucagon, it exhibits athreshold effect – i.e., an epinephrine response can still be elicited by hypoglycaemia, but onlyat a lower blood glucose concentration (Dagogo-Jack et al., 1993). If hypoglycaemia developsin patients who have this combined counterregulatory hormonal deficiency, glucose recoverymay be severely compromised (see Chapter 6). Subjecting such patients to intensified insulintherapy increased the risk of severe hypoglycaemia by 25 times, compared with subjectswho had an intact epinephrine response (White et al., 1983).Impaired awareness of hypoglycaemiaIn many patients with insulin-treated diabetes, the hypoglycaemia symptom profile alterswith time, resulting in impaired perception of the onset of hypoglycaemia (see Chapter 7).Commonly, autonomic warning symptoms are diminished and neuroglycopenic symptomspredominate. Around 25% of people with type 1 diabetes develop impaired awareness ofhypoglycaemia and the prevalence of this problem increases with the duration of insulintreatment (Hepburn et al., 1990; Gerich et al., 1991; Pramming et al., 1991). Prospectivestudies have demonstrated that the frequency of severe hypoglycaemia is increased up tosix-fold in patients with impaired awareness compared to those with normal hypoglycaemiaawareness (Gold et al., 1994; Clarke et al., 1995).Hypoglycaemia-associated central autonomic failureThe above acquired hypoglycaemia syndromes tend to segregate together clinically. Patientswith glycated haemoglobin concentrations close to the non-diabetic range are at greater riskof developing impaired awareness (Boyle et al., 1995; Kinsley et al., 1995; Pampanelli et al.,1996), while the glycaemic thresholds for the onset of symptoms and responses are alteredin patients with impaired awareness (Grimaldi et al., 1990; Hepburn et al., 1991; Mokanet al., 1994; Bacatselos et al., 1995). Cryer has suggested that these acquired abnormalitiesrepresent a form of central ‘Hypoglycaemia-Associated Autonomic Failure’ (HAAF) in type1 diabetes, speculating that recurrent severe hypoglycaemia is the primary cause (Cryer,1992; Cryer et al., 2003). If hypoglycaemia is the precipitant, then it is possible to see how avicious cycle may become established with the development of the acquired hypoglycaemiasyndromes promoting further episodes of severe hypoglycaemia.SummaryThere is a well-known adage that ‘hypoglycaemia begets hypoglycaemia’. People who haveexperienced one episode of severe hypoglycaemia are much more likely to develop furtherepisodes, and the greatest risk occurs in the weeks and months after the index event. Severehypoglycaemia becomes a more common problem in people with long-standing type 1

70 FREQUENCY, CAUSES AND RISK FACTORSdiabetes (UK Hypoglycaemia Group, 2007). This is the legacy of the burden ofhypoglycaemia that such individuals have experienced over many years with diabetes.The impaired symptomatic and counterregulatory responses to hypoglycaemia dramaticallyincrease the likelihood that an episode of mild hypoglycaemia will progress to a more severeevent.Genetic Predisposition to HypoglycaemiaMost of the precipitants and risk factors for hypoglycaemia have been known about formany years. In 2001, Pedersen-Bjergaard et al. (2001) reported a novel risk factor for severehypoglycaemia in adults with type 1 diabetes and raised the notion that some individualsmay have an inherent genetic susceptibility to hypoglycaemia. The Danish investigatorsnoted the similarity between endurance exercise and hypoglycaemia in that both are statesof limited metabolic fuel availability. Previous studies had linked exercise performance to aparticular polymorphism of the angiotensin-converting enzyme (ACE) gene. Specifically, theinsertion (I) allele, which resulted in low serum ACE activity, was associated with superiorperformance capacity compared with the deletion (D) allele. In an initial retrospective surveyof 207 adults with type 1 diabetes, patients with the DD genotype had a 3.2-fold increasedrisk of severe hypoglycaemia in the preceding two years, compared with individuals withthe II genotype. There was also a significant relationship between serum ACE activity,with a 1.4 increment in risk of severe hypoglycaemia for every 10 U/l rise in serum ACEconcentration (Figure 3.6). The serum ACE activity was directly linked to ACE genotype,and it remained a significant risk factor even after adjustment for conventional risk factors.Moreover, serum ACE activity was a stronger risk factor for severe hypoglycaemia inC-peptide negative individuals with impaired awareness, than in other groups (relative risk1.7 per 10 U/l; Figure 3.7). No significant relationship was observed between serum ACEactivity or genotype and frequency of mild hypoglycaemia.Figure 3.6 Risk of severe hypoglycaemia according to serum ACE activity in 207 patients withtype 1 diabetes, untreated with ACE inhibitors or angiotensin-2 receptor antagonists. Broken linesrepresent 95% confidence intervals. Reprinted from The Lancet, Pedersen-Bjergaard et al. (2001) withpermission from Elsevier

RISK FACTORS FOR SEVERE HYPOGLYCAEMIA 71Figure 3.7 Association between severe hypoglycaemia and serum ACE activity according toC-peptide status and self-estimated awareness of hypoglycaemia. Reprinted from The Lancet, Pedersen-Bjergaard et al. (2001) with permission from ElsevierThese findings were replicated in a prospective study for one year in 107 adults (Pedersen-Bjergaard et al., 2003a). Serum ACE activity in the fourth quartile was associated with a2.7-fold increased risk of severe hypoglycaemia compared to activity in the lowest quartile.Compared to subjects with the II genotype, individuals with the DD genotype had a 1.8-foldincreased risk of severe hypoglycaemia, although this did not reach statistical significance.Higher serum ACE concentrations were also associated with an increased risk of severehypoglycaemia in Swedish children and adolescents with type 1 diabetes (Nordfeldt andSamuelsson, 2003).The Danish group have put forward a number of possible mechanisms to explain theirobservations (Pedersen-Bjergaard et al., 2001; Pedersen-Bjergaard et al., 2003a). They speculatethat low levels of serum ACE may be associated with less cognitive deterioration duringacute hypoglycaemia, thereby increasing the likelihood that remedial action to correct hypoglycaemiais taken. Alternatively, low serum ACE activity might enhance counterregulationor reduce production of toxic substances, e.g. reactive oxygen species, during hypoglycaemia.All these putative mechanisms remain highly speculative, but the authors also raiseone other intriguing possibility: namely that ACE inhibition might reduce the frequency ofhypoglycaemia. This may seem counterintuitive, because previous population-based studiessuggested an association between severe hypoglycaemia and use of ACE inhibitors (Heringset al., 1996; Morris et al., 1997). However, these studies are flawed and a re-examination ofthe role of ACE inhibitors in ameliorating the impact of hypoglycaemia seems warranted.Recent data from one centre in Scotland have not demonstrated such a strong link betweenserum ACE and severe hypoglycaemia in adults with type 1 diabetes (Zammitt et al., 2007),nor has a study of children and adolescents with type 1 diabetes in Australia (Bulsaraet al., 2007). This should, therefore, serve as a reminder that genetic susceptibility to anybiological variable may not be the same in different populations and reinforces the need forthe Danish observations to be examined in other countries and ethnic groups. However, thestudies by Pedersen-Bjergaard and colleagues raise the intriguing prospect that there maybe yet other genetic factors that influence susceptibility to hypoglycaemia. Identification of

72 FREQUENCY, CAUSES AND RISK FACTORSthese may help stratify risk in individuals with type 1 diabetes and may ultimately lead tothe development of novel therapeutic interventions to prevent or ameliorate the impact ofhypoglycaemia.Absent Endogenous Insulin SecretionSeveral studies have demonstrated the importance of endogenous insulin secretion in definingrisk of hypoglycaemia. Individuals who are C-peptide negative, i.e., who have no endogenousinsulin production, have an approximately two- to fourfold increased risk of severe hypoglycaemiacompared to people with detectable C-peptide (Bott et al., 1997; The Diabetes Controland Complications Trial Research Group, 1997; Muhlhauser et al., 1998; Pedersen-Bjergaardet al., 2001). These data mirror the clinical experience that severe hypoglycaemia is rare inthe 12 months after the diagnosis of type 1 diabetes (Davis et al., 1997), when significantconcentrations of endogenous insulin can be measured. The presumption is that the abilityof endogenous insulin levels to fall in the face of a declining blood glucose concentrationprovides an additional layer of protection in the defences against hypoglycaemia and thusreduces overall risk.Dose and Type of InsulinFew diabetes specialists who were practising in the mid-1980s will forget the controversythat surrounded the introduction of human insulin. A substantial and vocal minority ofpatients with type 1 diabetes claimed that the change from animal-derived to human insulinwas associated with a loss of warning symptoms of hypoglycaemia and thus an increasedrisk. These claims were subject to considerable scientific scrutiny and ultimately a systematicreview of the evidence found no evidence to support the premise that treatment with human,as opposed to animal, insulin was associated with an increased risk of hypoglycaemia (Aireyet al., 2000).Since that time, several insulin analogues have been developed and their introductioninto clinical practice has been accompanied by the publication of studies that purport todemonstrate that use of the insulin analogues is associated with a lower frequency ofhypoglycaemia, in the face of stable or improved glycaemic control (Anderson et al., 1997;Garg et al., 2004; Hermansen et al., 2004). However, recent systematic reviews suggestthat neither short-acting (Siebenhofer et al., 2004) nor long-acting (Warren et al., 2004)insulin analogues are associated with clinically significant lower rates of hypoglycaemia. Theusual caveats of such clinical trials apply in terms of patient characteristics and recordingof hypoglycaemia. In one observational study from Colorado, the frequency of severehypoglycaemia rose in clinic patients in the immediate aftermath of the DCCT as effortsto intensify insulin therapy were instituted, but from 1996 onwards, the rates of severehypoglycaemia declined and the authors linked this to the introduction of insulin lispro(Chase et al., 2001). This study was subject to the effects of numerous confounders, butfurther data from routine clinical practice are required to help clarify the impact of insulinanalogues on risk of hypoglycaemia. The introduction of inhaled insulin (Exubera) has notbeen associated with a lower risk of hypoglycaemia in either type 1 or type 2 diabetes; whencompared with insulin administered by the subcutaneous route, rates of severe hypoglycaemiawere either equivalent (Hollander et al., 2004) or higher (Skyler et al., 2005).

RISK FACTORS FOR SEVERE HYPOGLYCAEMIA 73Higher doses of insulin have also been associated with an increased risk of hypoglycaemiain some studies (The Diabetes Control and Complications Trial Research Group, 1997;ter Braak et al., 2000), but not in others (Table 3.3). Several possible explanations mayaccount for this relationship – high insulin doses may be a sign of less endogenous insulinproduction, or of efforts to achieve strict glycaemic control. It may also reflect sub-optimalcompliance with insulin therapy and so indicate a pattern of behaviour that predisposes tomarked fluctuations in glucose control.SleepAs has already been discussed in the section on asymptomatic hypoglycaemia, nocturnalhypoglycaemia is very common (Vervoort et al., 1996). In the DCCT, 43% of episodesof severe hypoglycaemia occurred between midnight and 8.00 a.m., and 55% of episodesoccurred when individuals were asleep (The DCCT Research Group, 1991). The potentialfor nocturnal hypoglycaemia engenders significant anxiety among patients, particularly inindividuals who live alone. Patients worry that they will not awaken when hypoglycaemiaoccurs and that they may be left incapacitated or die as a consequence. Many individuals,as a result, maintain higher blood glucose concentrations at bedtime to reduce the risk ofnocturnal hypoglycaemia. Hypoglycaemia appears to be more common at night becausecounterregulatory hormone responses are blunted during sleep in people with type 1 diabetes(Jones et al., 1998; Banarer and Cryer, 2003). Moreover, hypoglycaemia awareness is alsoreduced during sleep (Banarer and Cryer, 2003), and so ultimately sleep impairs both thephysiological and behavioural responses to hypoglycaemia. Unsurprisingly, considerableeffort has been directed at developing strategies to reduce the risk of nocturnal hypoglycaemiaand these are discussed in Chapter 4.Microvascular ComplicationsIn a retrospective study of 44 patients with type 1 diabetes with impaired renal function(serum creatinine ≥ 133 umol/l and proteinuria), the incidence of severe hypoglycaemiawas five times higher than in matched subjects with normal renal function (Muhlhauseret al., 1991). In other studies, severe hypoglycaemia was linked with nephropathy (terBraak et al., 2000), peripheral neuropathy and retinopathy (Pedersen-Bjergaard et al., 2004).Although it is generally recognised that insulin requirement declines in advanced renaldisease, with reduced clearance of insulin, this association between hypoglycaemia andnephropathy (and other microvascular complications) could be confounded by many otherfactors, e.g. concomitant drug therapy (ter Braak et al., 2000) and the co-existence ofacquired hypoglycaemia syndromes which, like microvascular complications, are linked withincreasing duration of diabetes.The role of peripheral autonomic neuropathy in increasing the risk of severe hypoglycaemiahas been considered in several studies and deserves mention. In the EURODIABIDDM Complications Study, the presence of abnormal cardiovascular reflexes was associatedwith a 1.7-fold increased risk of severe hypoglycaemia (Stephenson et al., 1996).Gold et al. (1997) also demonstrated that autonomic neuropathy was associated with asmall increased risk of severe hypoglycaemia in 60 patients with type 1 diabetes, but nosuch relationship was demonstrated in the DCCT (The DCCT Research Group, 1991). The

74 FREQUENCY, CAUSES AND RISK FACTORSmechanism underlying this association remains to be fully elucidated. Peripheral autonomicneuropathy often coexists with impaired hypoglycaemia awareness in patients with type 1diabetes, presumably because both conditions are associated with diabetes of long duration(Hepburn et al., 1990), but impaired awareness can readily occur in the absence of peripheralautonomic neuropathy (Hepburn et al., 1990; Ryder et al., 1990; Bacatselos et al., 1995).It is well established that severe autonomic neuropathy is associated with ‘gastroparesisdiabeticorum’, which can cause marked swings in blood glucose concentration. Althoughsuch severe gastroparesis is now rare, delayed gastric emptying is relatively common (Konget al., 1996) and may explain at least part of the association between hypoglycaemia andautonomic neuropathy.Social and Psychological FactorsPsychological factors clearly play a crucial role in determining an individual’s likelihood ofdeveloping severe hypoglycaemia. Low mood (Gonder-Frederick and Cox, 1997), emotionalcoping (Bott et al., 1997) and socio-economic status (Muhlhauser et al., 1998; Leese et al.,2003) have been linked to risk of severe hypoglycaemia and so too have other morestraightforward behavioural factors such an individual’s propensity to carry a supply ofcarbohydrate for emergency use (Bott et al., 1997) and their determination to achievenormoglycaemia (Muhlhauser et al., 1998).Since the 1980s, Cox and colleagues at the University of Virginia, USA, have carried outseminal investigations to explore the psychological impact of hypoglycaemia on people withtype 1 diabetes. They developed a Fear of Hypoglycaemia scale, which sought to quantifythe anxieties that people with type 1 diabetes have with respect to hypoglycaemia and theextent to which individuals take steps to avoid experiencing such episodes (Cox et al., 1987).Unsurprisingly, there is a close association between fear of hypoglycaemia and perceivedrisk of future severe hypoglycaemia (Gonder-Frederick et al., 1997). In many instances thisis appropriate, in that ‘fear’ ratings are often high in people who have impaired awarenessof hypoglycaemia and/or have experienced multiple episodes in the past (Gold et al., 1996).However, in other people, fear of hypoglycaemia may be high while absolute risk is low –such individuals often display high levels of trait anxiety or have had a traumatic previousexperience of hypoglycaemia. It is often extremely difficult to persuade such individuals tomaintain strict glycaemic control. Conversely, there are people who have a very low fearof hypoglycaemia, despite a propensity to recurrent episodes. Such individuals may fail totake appropriate precautionary measures, thereby putting themselves and others at risk ifhypoglycaemia occurs, for example, when that individual is driving a car.EndocrinopathiesEndocrine disorders, such as Addison’s disease and hypopituitarism, which are associatedwith a deficiency of counterregulatory hormones, can be associated with an increased riskof hypoglycaemia in adults with type 1 diabetes. These are uncommon in everyday diabetespractice, but clinicians should maintain a high index of suspicion particularly in the patientwho simultaneously demonstrates a marked and otherwise unexplained decline in insulinrequirements.

CONCLUSIONS 75Other Risk FactorsThe list of other possible risk factors for hypoglycaemia is a long one. Smoking is a relativelynovel marker of increased risk (Pedersen-Bjergaard et al., 2004), but may be confoundedby other diabetes-related and lifestyle factors, e.g. regular use of alcohol (ter Braak et al.,2000). In the DCCT, men and adolescents were at increased risk (The Diabetes Control andComplications Trial Research Group, 1997), but these associations have not been replicatedwith any consistency in other studies (Table 3.3). Risks associated with pregnancy areaddressed in Chapter 10.Causes and Risk Factors for Hypoglycaemia: Summary andConclusionsMany of the risk factors described above are inter-related and any given individual mayhave more than one, which clearly will increase their overall risk. However, at its mostfundamental level, hypoglycaemia in adults with type 1 diabetes is a consequence of theinability of exogenous insulin levels to fall in response to a declining blood glucose concentration.In people who have had type 1 diabetes for several years, the situation is exacerbatedbecause of a failure of the normal physiological counterregulatory defence mechanisms,which in non-diabetic individuals serve to increase exogenous glucose production andgenerate warning symptoms. This counterregulatory failure worsens with increasing duration ofdiabetes, particularly if glycaemic control is strict and the individual has experienced recurrentepisodes of severe hypoglycaemia. Other factors may come in to play, for example particularbehavioural patterns and the time-action profiles of the exogenous insulin. Recent data alsoraise the possibility that there may be inherent genetic susceptibility to hypoglycaemia, butthis needs to be affirmed in wider populations and the underlying mechanisms more clearlydissected. Novel strategies to reduce overall risk of hypoglycaemia are urgently required.CONCLUSIONS• There are no definitive criteria for what constitutes hypoglycaemia, but most specialistsdistinguish mild from severe episodes depending on whether or not the individual is ableto self-treat.• The ‘average’ adult with type 1 diabetes will experience many thousands of episodes of mildhypoglycaemia over a lifetime, with a typical frequency of one to two episodes per week.• Severe hypoglycaemia is less common, and on average occurs once or twice every year,with an annual prevalence of around 30%. However, the distribution is heavily skewed,such that many individuals are unaffected over a calendar year, while a small numberexperience recurrent episodes.• Severe hypoglycaemia requiring treatment with intramuscular glucagon and/or intravenousglucose is even less common, and the majority of all episodes of hypoglycaemia aremanaged in the community by the patient and/or relatives and friends, without recourseto emergency services.

76 FREQUENCY, CAUSES AND RISK FACTORS• Hypoglycaemia occurs when there is an imbalance between insulin-mediated glucosedisposal and glucose influx into the circulation from the liver and exogenous carbohydrate.Typical precipitants include patient error in insulin dosage, alcohol andexercise.• Plasma concentrations of exogenous insulin cannot decline in response to falling bloodglucose and the time action profiles of current insulins do not accurately mimic the normalphysiological variation in insulin. In a high proportion of cases, no underlying precipitantof a given episode of hypoglycaemia can be identified.• The DCCT and other intervention studies have provided substantial information of theepidemiology of severe hypoglycaemia, but individuals participating in such studies maynot be representative of the wider population who have type 1 diabetes.• A major determinant of increased risk of severe hypoglycaemia is Hypoglycaemia-Associated Autonomic Failure, which is a consequence of exposure to recurrent episodesof hypoglycaemia. This is a feature of individuals with long-standing diabetes, intensiveinsulin therapy and strict glycaemic control.• Recent data suggest that there may be specific genetic factors that predispose to anincreased risk of hypoglycaemia.REFERENCESAirey CM, Williams DRR, Martin PG, Bennett CMT, Spoor PA (2000). Hypoglycaemia induced byexogenous insulin – ‘human’ and animal insulin compared. Diabetic Medicine 17: 416–32.Allen C, LeClaire T, Palta M, Daniels K, Meredith M, D’Alessio DJ, for the Wisconsin DiabetesRegistry Project (2001). Risk factors for frequent and severe hypoglycemia in type 1 diabetes.Diabetes Care 24: 1878–81.American Diabetes Association Workgroup on Hypoglycemia (2005). Defining and Reporting Hypoglycemiain Diabetes. Diabetes Care 28: 1245–9.Anderson JH, Brunelle RL, Koivisto VA, Pfutzner A, Trautmann ME, Vignati L, DiMarchi R, TheMulticenter Insulin Lispro Study Group (1997). Reduction of postprandial hyperglycemia andfrequency of hypoglycemia in IDDM patients on insulin-analog treatment. Diabetes 46: 265–70.Arky RA, Veverbrants E, Abramson EA (1968). Irreversible hypoglycemia. A complication of alcoholand insulin. Journal of the American Medical Association 206: 575–8.Avogaro A, Beltramello P, Gnudi L, Maran A, Valerio A, Miola M et al. (1993). Alcohol intakeimpairs glucose counterregulation during acute insulin-induced hypoglycemia in IDDM patients.Evidence for a critical role of free fatty acids. Diabetes 42: 1626–34.Bacatselos SO, Karamitsos DT, Kourtoglou GI, Zambulis CX, Yovos JG, Vyzantiadis AT (1995).Hypoglycaemia unawareness in type 1 diabetic patients under conventional insulin treatment.Diabetes Nutrition and Metabolism 8: 267–75.Banarer S, Cryer PE (2003). Sleep-related hypoglycemia-associated autonomic failure in type 1diabetes. Diabetes 52: 1195–1203.Bode BW, Schwartz S, Stubbs HA, Block JE (2005). Glycemic characteristics in continuously monitoredpatients with type 1 and type 2 diabetes. Normative values. Diabetes Care 28: 2361–6.Bolli G, De Feo P, Compagnucci P, Cartechini MG, Angeletti G, Santeusanio F et al. (1983).Abnormal glucose counterregulation in insulin-dependent diabetes mellitus. Interaction of antiinsulinantibodies and impaired glucagon and epinephrine secretion. Diabetes 32: 134–41.

REFERENCES 77Bott S, Bott U, Berger M, Muhlhauser I (1997). Intensified insulin therapy and the risk of severehypoglycaemia. Diabetologia 40: 926–32.Boyle PJ, Schwartz NS, Shah SD, Clutter WE, Cryer PE (1988). Plasma glucose concentrations at theonset of hypoglycemic symptoms in patients with poorly controlled diabetes and in non-diabetics.New England Journal of Medicine 318: 1487–92.Boyle PJ, Kempers SF, O’Connor AM, Nagy RJ (1995). Brain glucose uptake and unawarenessof hypoglycemia in patients with insulin-dependent diabetes mellitus. New England Journal ofMedicine 333: 1726–31.Bulsara MK, Holman CDJ, van Bockxmeer FM, Davis EA, Gallego PH, Beilby JP, Palmer LJet al. (2007). The relationship between ACE genotype and risk of severe hypoglycemia in a largepopulation-based cohort of children and adolescents with type 1 diabetes. Diabetologia 50: 965–71.Chase HP, Lockspeiser T, Peery B, Shepherd M, MacKenzie T, Anderson J, Garg SK (2001). Theimpact of the Diabetes Control and Complications Trial and Humalog insulin on glycohemoglobinlevels and severe hypoglycemia in type 1 diabetes. Diabetes Care 24: 430–4.Clarke WL, Cox DJ, Gonder-Frederick LA, Julian D, Schlundt D, Polonsky W (1995). Reducedawareness of hypoglycemia in adults with IDDM. A prospective study of hypoglycemic frequencyand associated symptoms. Diabetes Care 18: 517–22.Cox DJ, Irvine A, Gonder-Frederick L, Nowacek G, Butterfield J (1987). Fear of hypoglycemia:quantification, validation and utilization. Diabetes Care 10: 617–21.Cox DJ, Kovatchev BP, Julian DM, Gonder-Frederick LA, Polonsky WH, Schlundt DG, Clarke WL(1994). Frequency of severe hypoglycemia in insulin-dependent diabetes mellitus can be predictedfrom self-monitoring blood glucose data. Journal of Clinical Endocrinology and Metabolism 79:1659–62.Cryer PE (1992). Iatrogenic hypoglycemia as a cause of hypoglycemia-associated autonomic failurein IDDM. A vicious cycle. Diabetes 41: 255–260.Cryer PE (1994). Hypoglycemia: the limiting factor in the management of IDDM. Diabetes 43:1378–89.Cryer PE, Davis SN, Shamoon H (2003). Hypoglycemia in diabetes. Diabetes Care 26: 1902–12.Dagogo-Jack SE, Craft S, Cryer PE (1993). Hypoglycemia-associated autonomic failure in insulindependentdiabetes mellitus. Recent antecedent hypoglycemia reduces autonomic responses to,symptoms of, and defense against subsequent hypoglycemia. Journal of Clinical Investigation 91:819–28.Davis EA, Keating B, Byrne GC, Russell M, Jones TW (1997). Hypoglycemia: Incidence and clinicalpredictors in a large population-based sample of children and adolescents with IDDM. DiabetesCare 20: 22–5.Donnelly LA, Morris AD, Frier BM, Ellis JD, Donnan PT, Durrant R et al., for the DARTS/MEMOCollaboration (2005). Frequency and predictors of hypoglycaemia in type 1 and insulin-treatedtype 2 diabetes: a population-based study. Diabetic Medicine 22: 749–55.Feher MD, Grout P, Kennedy A, Elkeles RS, Touquet R (1989). Hypoglycaemia in an inner-cityaccident and emergency department: a 12-month survey. Archives of Emergency Medicine 6: 183–8.Galassetti P, Mann S, Tate D, Neill RA, Costa F, Wasserman DH, Davis SN (2001a). Effects ofantecedent prolonged exercise on subsequent counterregulatory responses to hypoglycemia. AmericanJournal of Endocrinology 280: E908–17.Galassetti P, Neill AR, Tate D, Ertl AC, Wasserman DH, Davis SN (2001b). Sexual dimorphismin counterregulatory responses to hypoglycemia after antecedent exercise. Journal of ClinicalEndocrinology and Metabolism 86: 3516–24.Galassetti P, Tate D, Neill RA, Morrey S, Wasserman DH, Davis SN (2003). Effect of antecedenthypoglycemia on counterregulatory responses to subsequent euglycemic exercise in type 1 diabetes.Diabetes 52: 1761–9.Garg SK, Gottlieb PA, Hisatomi ME, D’Souza A, Walker AJ, Izuora KE, Chase HP (2004). Improvedglycemic control without an increase in severe hypoglycemic episodes in intensively treated patients

78 FREQUENCY, CAUSES AND RISK FACTORSwith type 1 diabetes receiving morning, evening, or split dose insulin glargine. Diabetes Researchand Clinical Practice 66: 49–56.Gerich J, Langlois M, Noacco C, Karam J, Forsham P (1973). Lack of glucagon response to hypoglycemiain diabetes: evidence for an intrinsic pancreatic alpha-cell defect. Science 182: 171–3.Gerich JE, Mokan M, Veneman T, Korytkowski M, Mitrakou A (1991). Hypoglycemia unawareness.Endocrine Reviews 12: 356–71.Gold AE, MacLeod KM, Frier BM (1994). Frequency of severe hypoglycemia in patients with type 1diabetes with impaired awareness of hypoglycemia. Diabetes Care 17: 697–703.Gold AE, Frier BM, MacLeod KM, Deary IJ (1997). A structural equation model for predictors ofsevere hypoglycaemia in patients with insulin-dependent diabetes mellitus. Diabetic Medicine 14:309–15.Gonder-Frederick L, Clarke WL, Cox DJ (1997). The emotional, social, and behavioural implicationsof insulin-induced hypoglycemia. Seminars in Clinical Neuropsychiatry 2: 57–65.Gonder-Frederick LA, Cox D (1997). The relationship between psychological depression and frequencyof severe hypoglycemia (SH) in IDDM. Diabetes 46: 268A(Abstract)Grimaldi A, Bosquet F, Davidoff P, Digy JP, Sachon C, Landault C et al. (1990). Unawareness ofhypoglycemia by insulin-dependent diabetics. Hormone and Metabolic Research 22: 90–5.Hart SP, Frier BM (1998). Causes, management and morbidity of acute hypoglycaemia in adultsrequiring hospital admission. Quarterly Journal of Medicine 91: 505–10.Heller SR, Macdonald IA (1996). The measurement of cognitive function during acute hypoglycaemia:experimental limitations and their effect on the study of hypoglycaemia unawareness. DiabeticMedicine 13: 607–15.Hepburn DA, Patrick AW, Eadington DW, Ewing DJ, Frier BM (1990). Unawareness of hypoglycaemiain insulin-treated diabetic patients: prevalence and relationship to autonomic neuropathy. DiabeticMedicine 7: 711–17.Hepburn DA, Patrick AW, Brash HM, Thomson I, Frier BM (1991). Hypoglycaemia unawarenessin type 1 diabetes: a lower plasma glucose is required to stimulate sympatho-adrenal activation.Diabetic Medicine 8: 934–45.Herings RMC, de Boer A, Stricker BHC, Leufkens HGM, Porsius A (1996). Hypoglycemia associatedwith the use of inhibitors of angiotensin-converting enzyme. Lancet 345: 1195–8.Hermansen K, Fontaine P, Kukolja KK, Peterkova V, Leth G, Gall M-A (2004). Insulin analogues(insulin detemir and insulin aspart) versus traditional human insulins (NPH insulin and regularhuman insulin) in basal-bolus therapy for patients with type 1 diabetes. Diabetologia 47: 622–9.Hirsch BR, Shamoon H (1987). Defective epinephrine and growth hormone responses in type 1diabetes are stimulus specific. Diabetes 36: 20–6.Hollander PA, Blonde L, Rowe R, Mehta AE, Milburn JL, Hershon KS et al., the Exubera Phase IIIStudy Group (2004). Efficacy and safety of inhaled insulin (Exubera) compared with subcutaneousinsulin therapy in patients with type 2 diabetes. Diabetes Care 27: 2356–62.Janssen MMJ, Snoek FJ, de Jongh RT, Casteleijn S, Deville W, Heine RJ (2000). Biological andbehavioural determinants of the frequency of mild, biochemical hypoglycaemia in patients with Type 1diabetes on multiple insulin injection therapy. Diabetes Metabolism Research and Reviews 16: 157–63.Johnson ES, Koepsell TD, Reiber G, Stergachis A, Platt R (2002). Increasing incidence of serioushypoglycemia in insulin users. Journal of Clinical Epidemiology 55: 253–9.Jones TW, Porter P, Sherwin RS, Davis EA, O’Leary P, Frazer F et al. (1998). Decreased epinephrineresponses to hypoglycemia during sleep. New England Journal of Medicine 338: 1657–62.Kerr D, Macdonald IA, Heller SR, Tattersall RB (1990). Alcohol causes hypoglycaemic unawarenessin healthy volunteers and patients with type 1 (insulin-dependent) diabetes. Diabetologia 33: 216–21.Kinsley BT, Widom B, Simonson DC (1995). Differential regulation of counterregulatory hormonesecretion and symptoms during hypoglycemia in IDDM. Diabetes Care 18: 17–26.Kong M-F, Macdonald IA, Tattersall RB (1996). Gastric emptying in diabetes. Diabetic Medicine 13:112–9.

REFERENCES 79Leckie AM, Graham MK, Grant JB, Ritchie PJ, Frier BM (2005). Frequency, severity, and morbidityof hypoglycemia occurring in the workplace in people with insulin-treated diabetes. Diabetes Care28: 1333–8.Leese GP, Wang J, Broomhall J, Kelly P, Marsden A, Morrison W et al., for the DARTS/MEMOCollaboration (2003). Frequency of severe hypoglycemia requiring emergency department treatmentin type 1 and type 2 diabetes. A population-based study of health service resource use. DiabetesCare 26: 1176–80.MacLeod KM, Hepburn DA, Frier BM (1993). Frequency and morbidity of severe hypoglycaemia ininsulin-treated diabetic patients. Diabetic Medicine 10: 238–45.MacLeod KM, Deary IJ, Graham KS, Hepburn DA, Frier BM (1994). Hypoglycaemia unawareness inadult patients with type 1 diabetes: relationship to severe hypoglycaemia and cognitive impairment.Diabetes, Nutrition and Metabolism 7: 205–11.MacLeod KM, Gold AE, Frier BM (1995). Frequency, severity and symptomatology of hypoglycaemia:a comparative trial of human and porcine insulins in type 1 diabetic patients. Diabetic Medicine 12:134–41.Mokan M, Mitrakou A, Veneman T, Ryan C, Korytkowski M, Cryer PE, Gerich J (1994). Hypoglycemiaunawareness in IDDM. Diabetes Care 17: 1397–403.Morris AD, Boyle DIR, McMahon AD, Pearce H, Evans JMM, Newton RW et al., for theDARTS/MEMO Collaboration (1997). ACE inhibitor use is associated with hospitalization forsevere hypoglycemia in patients with diabetes. Diabetes Care 20: 1363–7.Moses RG, Hubert PA, Lewis-Driver DJ (1985). Severe hypoglycaemia. A one-year prospective surveyin Wollongong. Medical Journal of Australia 142: 294–6.Muhlhauser I, Berger M, Sonnenberg G, Koch J, Jorgens V, Schernthaner G, Scholz V (1985).Incidence and management of severe hypoglycemia in 434 adults with insulin-dependent diabetesmellitus. Diabetes Care 8: 268–73.Muhlhauser I, Bruckner I, Berger M, Cheta D, Jorgens V, Ionescu-Tirgoviste C et al. (1987). Evaluationof an intensified insulin treatment and teaching programme as routine management of Type 1(insulin-dependent) diabetes. The Bucharest-Düsseldorf Study. Diabetologia 30: 681–90.Muhlhauser I, Toth G, Sawicki PT, Berger M (1991). Severe hypoglycemia in type 1 diabetic patientswith impaired kidney function. Diabetes Care 14: 344–6.Muhlhauser I, Overmann H, Bender R, Bott U, Berger M (1998). Risk factors for severe hypoglycaemiain adult patients with type 1 diabetes. Diabetologia 41: 1274–82.Nordfeldt S, Samuelsson U (2003). Serum ACE predicts severe hypoglycemia in children and adolescentswith type 1 diabetes. Diabetes Care 26: 274–8.O’Neill, S (1997). Make four the floor. Balance Jan/Feb: 20–3.Owens DR, Zinman B, Bolli GB (2001). Insulins today and beyond. Lancet 358: 739–46.Pampanelli S, Fanelli C, Lalli C, Ciofetta M, Del Sindaco P, Lepore M et al. (1996). Long-termintensive insulin therapy in IDDM: effects on HbA 1c , risk for severe and mild hypoglycaemia, statusof counterregulation and awareness of hypoglycaemia. Diabetologia 39: 677–86.Pedersen-Bjergaard U, Agerholm-Larsen B, Pramming S, Hougaard P, Thorsteinsson B (2001).Activity of angiotensin-converting enzyme and risk of severe hypoglycaemia in type 1 diabetesmellitus. Lancet 357: 1248–53.Pedersen-Bjergaard U, Agerholm-Larsen B, Pramming S, Hougaard P, Thorsteinsson B (2003a).Prediction of severe hypoglycaemia by angiotensin-converting enzyme activity and genotype in type1 diabetes. Diabetologia 46: 89–96.Pedersen-Bjergaard U, Pramming S, Thorsteinsson B (2003b). Recall of severe hypoglycaemia andself-estimated state of awareness in type 1 diabetes. Diabetes Metabolism Research and Reviews19: 232–40.Pedersen-Bjergaard U, Pramming S, Heller SR, Wallace TM, Jorgensen HV, Matthews DR et al.(2004). Severe hypoglycaemia in 1076 adult patients with type 1 diabetes: influence of risk markersand selection. Diabetes Metabolism Research and Reviews 20: 479–86.

80 FREQUENCY, CAUSES AND RISK FACTORSPedersen-Bjergaard U, Reubsaet JLE, Nielsen SL, Pedersen-Bjergaard S, Perrild H, Pramming S,Thorsteinsson B (2005). Psychoactive drugs, alcohol, and severe hypoglycemia in insulin-treateddiabetes: Analysis of 141 cases. American Journal of Medicine 118: 307–10.Plank J, Kohler G, Rakovac I, Semlitsch BM, Horvath K, Bock G, et al. (2004). Long-term evaluationof a structured out-patient education programme for intensified insulin therapy in patients with type1 diabetes: a 12 year follow-up. Diabetologia 47: 1370–5.Potter J, Clarke P, Gale EAM, Dave SH, Tattersall RB (1982). Insulin-induced hypoglycaemiain an accident and emergency department: the tip of an iceberg? British Medical Journal 285:1180–2.Pramming S, Thorsteinsson B, Bendtson I, Binder C (1990). The relationship between symptomaticand biochemical hypoglycaemia in insulin-dependent diabetic patients. Journal of Internal Medicine228: 641–6.Pramming S, Thorsteinsson B, Bendtson I, Binder C (1991). Symptomatic hypoglycaemia in 411 type1 diabetic patients. Diabetic Medicine 8: 217–22.Reichard P, Pihl M (1994). Mortality and treatment side-effects during long-term intensified conventionalinsulin treatment in the Stockholm Diabetes Intervention Study. Diabetes 43: 313–7.Ryder REJ, Owens DR, Hayes TM, Ghatei MA, Bloom SR (1990). Unawareness of hypoglycaemiaand inadequate hypoglycaemic counterregulation: no causal relationship with diabetic autonomicneuropathy. British Medical Journal 301: 783–7.Service FJ (1995). Hypoglycemic disorders. New England Journal of Medicine 332: 1144–52.Siebenhofer A, Plank J, Berghold A, Horvath K, Sawicki PT, Beck P, Pieber TR (2004). Meta-analysisof short-acting insulin analogues in adult patients with type 1 diabetes: continuous insulin infusionversus injection therapy. Diabetologia 47: 1895–1905.Skyler JS, Weinstock RS, Raskin P, Yale J-F, Barrett E, Gerich JE, Gerstein HC, the Inhaled InsulinPhase III type 1 Diabetes Study Group (2005). Diabetes Care 28: 1630–35.Sonnenberg GE, Kemmer FW, Berger M (1990). Exercise in type 1 (insulin-dependent) diabeticpatients treated with continuous subcutaneous insulin infusion. Prevention of exercise inducedhypoglycaemia. Diabetologia 33: 696–703.Stephenson JM, Kempler P, Cavallo Perin P, Fuller JH, the EURODIAB IDDM Complications StudyGroup (1996). Is autonomic neuropathy a risk factor for severe hypoglycaemia? The EURODIABIDDM Complications Study. Diabetologia 39: 1372–6.ter Braak EWMT, Appelman AMMF, van de Laak MF, Stolk RP, Van Haeften TW, Erkelens DW(2000). Clinical characteristics of type 1 diabetic patients with and without severe hypoglycemia.Diabetes Care 23: 1467–71.The DCCT Research Group (1991). Epidemiology of severe hypoglycemia in the Diabetes Controland Complications Trial. American Journal of Medicine 90: 450–9.The Diabetes Control and Complications Trial Research Group (1993). The effect of intensive treatmentof diabetes on the development and progression of long-term complications in insulin-dependentdiabetes mellitus. New England Journal of Medicine 329: 977–86.The Diabetes Control and Complications Trial Research Group (1997). Hypoglycemia in the DiabetesControl and Complications Trial. Diabetes 46: 271–86.The EURODIAB IDDM Complications Study Group (1994). Microvascular and acute complicationsin IDDM patients: the EURODIAB IDDM Complications Study. Diabetologia 37:278–85.Thorsteinsson B, Pramming S, Lauritzen T, Binder C (1986). Frequency of daytime biochemical hypoglycaemiain insulin-treated diabetic patients: relation to daily median blood glucose concentrations.Diabetic Medicine 3: 147–51.Turner BC, Jenkins E, Kerr D, Sherwin RS, Cavan DA (2001). The effect of evening alcohol consumptionon next-morning glucose control in type 1 diabetes. Diabetes Care 24: 1888–93.UK Hypoglycaemia Study Group (2007). Risk of hypoglycaemia in types 1 and 2 diabetes: effects oftreatment modalities and their duration. Diabetologia 50: 1140–47.

REFERENCES 81Vervoort G, Goldschmidt HMG, van Doorn LG (1996). Nocturnal blood glucose profiles in patientswith type 1 diabetes mellitus on multiple (> 4) daily insulin injection regimens. Diabetic Medicine13: 794–9.Warren E, Weatherley-Jones E, Chilcott J, Beverley C (2004). Systematic review and economicevaluation of a long-acting insulin analogue, insulin glargine. Health Technology Assessment 8 (45):iii, 1–57.White NH, Skor DA, Cryer PE, Levandoski LA, Bier DM, Santiago JV (1983). Identification of type 1diabetic patients at increased risk for hypoglycemia during intensive therapy. New England Journalof Medicine 308: 485–91.Zammitt NN, Geddes J, Warren RE, Marioni R, Ashby JP, Frier BM (2007). Serum Angiotensin-Converting Enzyme and frequency of severe hypoglycaemia in type 1 diabetes: does a relationshipexist? Diabetic Medicine, in press.

4 Nocturnal HypoglycaemiaSimon R. HellerINTRODUCTIONPeople with type 1 diabetes, who owe their lives to insulin, worry as much abouthypoglycaemia as they do about the prospect of developing serious diabetic complications,and episodes occurring during sleep are feared more than any other. Several studies haveconfirmed how commonly hypoglycaemia occurs during the night, and the possibility ofthe development of severe nocturnal episodes can drive parents to sleep in their child’sbedroom for many years. However, nocturnal hypoglycaemia may have consequences beyonda domestic emergency. Asymptomatic episodes may contribute to impaired hypoglycaemiaawareness and deficient counterregulation, while nocturnal hypoglycaemia may also be associatedwith cognitive impairment and is implicated in the ‘Dead in Bed’ syndrome (seeChapter 12) with the increased risk of sudden death during sleep in young people with type1 diabetes. In this chapter, the reasons why nocturnal hypoglycaemia occurs so frequentlyin type 1 diabetes are discussed. Also, some clinical approaches are described that mighthelp to reduce its frequency and severity and some controversial topics are considered suchas the Somogyi effect.EPIDEMIOLOGY – HOW COMMON IS NOCTURNALHYPOGLYCAEMIA?Nocturnal hypoglycaemia has always been of major concern to individuals with diabetes,with many patients experiencing severe episodes, but it was not until the 1970s that thisproblem was studied systematically. Patients with type 1 diabetes on routine insulin therapywere admitted for overnight monitoring using intermittent venous blood sampling (Gale andTattersall, 1979). Nocturnal hypoglycaemia was reported in 22 of 39 adults, who had whatwould now be considered poorly controlled diabetes and were being treated with one ortwo injections of insulin each day. Since then a number of investigators have examined thefrequency of nocturnal hypoglycaemia, either using patients’ self-reports or by measuringblood glucose at intervals overnight. Although some studies have reported relatively lowrates, others have demonstrated how common nocturnal hypoglycaemia can be. Reportedrates of hypoglycaemia vary between 7 and 60% both in children and adults (Table 4.1).Hypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

Table 4.1 Frequency of nocturnal hypoglycaemia in insulin-treated diabetesStudy NumberAdults (A) orchildren (C) TreatmentSampling (bloodglucose (BG) orcontinuous(CGM))Percentage withglucose below:(mmol/l)3.0 2.5 2.0Proportionasymptomatic Mean HbA 1cGale andTattersall, 1979Dornan et al.,1981Pramming et al.,1985Whincup andMilner, 1987Bendtson et al.,1988Shalwitz et al.,1990Vervoort et al.,199639 A BD insulin BG 56 6482 A BD insulin BG 27 27 9.5 (A1)58 A Conventional BG 29 22 9 9 9.1 (A1)71 C BD insulin ODinsulinBG 34 111023 A MDI BG 30 9 9 57 8.3 (A1)25 CSII 44 36 12 66 7.0 (A1c)135 C Conventional BG 14 7 2 100 10.3 (A1)31 A Basal bolus BG 29 67 8.6 (A1c)Porter et al., 1997 61 C BD BG 17 13 8Beregszaszi et al., 150 C BD/MIT BG 47 491997Matyka et al., 29 C BD BG 45 100 8.8 (A1c)1999bBoland et al., 2001 56 C MIT/CSII CGM 70 100Kaufman et al.,200247 C MIT/CSII CGM 35 27 100 8.6BD = twice daily; CSII = continuous subcutaneous insulin infusion; OD = once daily; MDI = multiple doses of insulin; CGM = continuous glucose monitoring; MIT = multiple injection therapy;BG = blood glucose

EPIDEMIOLOGY OF NOCTURNAL HYPOGLYCAEMIA? 8576Glucose (mmol/l)54321021:0023:00 01:00 03:00Time (hours)05:00 07:00Figure 4.1 Overnight glucose profiles of one child with type 1 diabetes who was hypoglycaemic onboth study nights. The shaded area represents the range of glucose values from the overnight profilesof the children without diabetesAlthough this is clearly dependent upon the blood glucose level that is designated asrepresenting hypoglycaemia, it is interesting that the studies that reported the lowest ratesof hypoglycaemia had sampled blood glucose less frequently, and it is conceivable thathypoglycaemia at other times of the night was simply not identified. Furthermore, most earlystudies involved patients with poor glycaemic control. It seems likely that with intensiveinsulin therapy now being offered to most patients, rates of nocturnal hypoglycaemia areeven higher. This certainly appears to be true in children. In one alarming study of childrentreated with multiple injection therapy, the rates of nocturnal hypoglycaemia exceeded 50%(Beregszaszi et al., 1997). It is noteworthy that children have been directly observed tosleep through most nocturnal episodes despite having a blood glucose value below 2 mmol/l(Matyka et al., 1999b) (Figure 4.1).The introduction of continuous blood glucose monitoring with glucose being measuredevery few minutes has allowed more frequent recording during sleep. Initial reports usingthese methods have suggested that hypoglycaemic events may be even more commonthan was proposed previously (Table 4.1). However, although continuous blood glucoserecording appears to have revealed a high rate of nocturnal episodes in some studies,it may over-estimate the overall frequency (McGowan et al., 2002). The technologymeasures glucose in the extra cellular space and not in the blood vessels, and the relationshipbetween the glucose concentrations in these sites is not clear (Kulcu et al., 2003).It is also possible that the technology itself produces an inbuilt bias (Wentholt et al.,2005). Nevertheless, not only is nocturnal hypoglycaemia common when measured byconventional blood glucose sampling but it is also often of long duration with someepisodes lasting for more than three hours (Gale and Tattersall, 1979; Matyka et al.,1999b).

86 NOCTURNAL HYPOGLYCAEMIACAUSES OF NOCTURNAL HYPOGLYCAEMIAThe Limitations of Therapeutic Insulin DeliveryThe ability of people with insulin-treated diabetes to maintain strict glucose targets andprevent long-term tissue damage is compromised by the deficiencies of currently availablemethods of insulin delivery. In non-diabetic individuals, endogenous secretion of insulinprecisely meets demand. When food is eaten, insulin secretion increases rapidly to matchnutrient intake, particularly when the ingested food contains carbohydrate. In between meals,basal insulin secretion falls to a low but consistent level to maintain basal metabolism,without the risk of hypoglycaemia, even during prolonged fasting that lasts for hours ordays. In contrast, the limitations of delivering insulin by subcutaneous injection to patientswho can no longer produce endogenous insulin, leads not only to inadequate plasma insulinconcentrations during eating and the immediate postprandial period, but also to inappropriatelyraised plasma insulin levels in the post-absorptive phase (Rizza et al., 1980). Thisresults in high postprandial blood glucose concentrations in the hour or so after eating anda tendency to cause hypoglycaemia in the period before the next meal. Individuals areparticularly likely to be affected during the night, when the interval between ingestion offood may be several hours (Box 4.1). As discussed below, developments in insulin deliveryby using insulin analogues or continuous subcutaneous insulin infusion offer some benefitover conventional insulin, although these approaches mitigate rather than cure the problemof nocturnal hypoglycaemia.Impaired Counterregulatory Responses to Hypoglycaemia duringNocturnal HypoglycaemiaVarious mechanisms contribute to the additional risk of developing hypoglycaemia during thenight. Although an earlier study suggested that hormonal responses to hypoglycaemia mightbe increased during nocturnal episodes (Bendtson et al., 1993), recent work has indicatedthat counterregulatory defences are generally impaired. Matyka et al. (1999a) studied 29 prepubertalchildren overnight in their own homes on two separate occasions. They confirmedthat not only was nocturnal hypoglycaemia common, but also that prolonged episodes ofhypoglycaemia were not accompanied by increases in epinephrine (adrenaline) and otherBox 4.1Factors contributing to the development of nocturnal hypoglycaemia1. Long period between meals (especially in children).2. Inconsistency of conventional subcutaneous basal insulin delivery – nocturnalhyperinsulinaemia.3. Unawareness of early symptoms of hypoglycaemia when asleep.4. Diminished counterregulatory hormone release and symptomatic response in supineposture and effect of sleep per se.

CAUSES OF NOCTURNAL HYPOGLYCAEMIA 87protective endocrine responses. Since protective counterregulatory responses mitigate theseverity of hypoglycaemia by increasing hepatic glucose output and reducing peripheralglucose uptake, as well as enhancing some of the warning autonomic symptoms, defectiveresponses during the night may have a major effect in increasing the risk of nocturnalhypoglycaemia.One important question that emerges from this work is why nocturnal hypoglycaemiashould provoke different physiological responses to those that occur during the day. It nowappears that impaired responses are the result of additional factors at night, in particular asupine posture and also sleep per se.Supine PostureTwo studies have investigated the effect of posture on counterregulatory hormonal defencesto hypoglycaemia. Hirsch et al. (1991) compared the physiological responses to hypoglycaemiainduced using a hyperinsulinaemic clamp in young people with type 1 diabetes, whowere either in an upright or a horizontal position. Increments in hypoglycaemic symptomscores were more than 50% lower when patients remained supine compared to the usualincrease that was observed when they were standing erect. These findings have beenconfirmed by other researchers, who also demonstrated that plasma epinephrine levels werethree times lower during hypoglycaemia in the supine position compared to concentrationswhen subjects were standing (Robinson et al., 1994). The precise physiological explanationis not entirely clear but may be related to the recruitment of adrenoreceptors which areprimed by an upright posture. Whatever the cause, these observations suggest that patientsare more vulnerable to progressing to a severe episode of hypoglycaemia when lying horizontalin bed, because of a reduction in symptom intensity and in magnitude of hormonalcounterregulation.SleepMost studies have reported suppression of physiological protection by counterregulatorymechanisms during hypoglycaemia. Jones et al. (1998) lowered blood glucose to 2.8 mmol/lboth during the daytime and at night, and demonstrated diminished epinephrine responsesin diabetic and non-diabetic adolescents while they were asleep when compared to thebrisk responses while they were awake, whether during daytime hours or during the night(Figure 4.2). Banerer and Cryer (2003) confirmed these observations in patients with type 1diabetes but, interestingly, in non-diabetic subjects no difference in epinephrine responsewas observed between these states. Diminished counterregulatory responses, includingepinephrine, have also been demonstrated during spontaneous nocturnal hypoglycaemicepisodes in children with diabetes (Matyka et al., 1999a). Therefore, sleep appears to beassociated with diminished catecholamine and symptomatic responses to hypoglycaemiawith a reduction in wakening during a hypoglycaemic episode. The diminished responsemay occur because the glycaemic thresholds for activation of these responses have beenreset to a lower glucose level (Gais et al., 2003), so that more profound hypoglycaemia isnecessary to provoke a similar response to that observed in subjects when they are awake.As the investigators involved in these studies have pointed out, this increases the risk of a

88 NOCTURNAL HYPOGLYCAEMIAPlasma Epinephrine (pg/ml)350300250200150100500Patients with DiabetesNormal Subjects450Daytime, awakeDaytime, awake400Nighttime, asleepNighttime, asleepNighttime, awake350300250200150100500–60 –40 –20 0 20 40 60 –60 –40 –20 0 20 40 60Time (minutes)Figure 4.2 Mean (± SE) plasma epinephrine concentrations in eight patients with type 1 diabetesand six normal subjects during periods of hypoglycaemia when they were awake during the day, awakeat night and asleep at night. (To convert plasma epinephrine values to picomoles per litre, multiply by5.458. The zero on the x-axis indicates the beginning of the hypoglycaemic period). Reproduced withpermission from Jones et al. (1998). Copyright © 1998 Massachusetts Medical Societysevere hypoglycaemic episode occurring at night. However, the mechanisms that disturb thephysiological responses to hypoglycaemia during sleep remain unknown.Sleep is not a unitary process (Oswald, 1987). Sleep is dominated by Slow Wave Sleep(SWS) during the first third of the night, and by the cyclical appearance of Rapid EyeMovement (REM) sleep during the latter two thirds. Autonomic activity during SWS isrelatively steady but in REM sleep (desynchronisation of the EEG, with absence of activity inthe anti-gravity and periodic eye movements) modulations in respiratory and cardiovascularevents occur with other changes in the autonomic nervous system. These differences inautonomic activity between SWS and REM sleep suggest that the effect of hypoglycaemiaon autonomic responses may vary depending on which stage of sleep is being experienced.However, to date no studies have been published that have explored the effect of differentphases of sleep on counterregulatory responses to hypoglycaemia.CONSEQUENCES OF NOCTURNAL HYPOGLYCAEMIAImpaired Awareness of HypoglycaemiaThe demonstration in the early 1990s that repeated exposure to hypoglycaemia leads toimpaired physiological defences to subsequent episodes and a reduction in the intensity ofsymptoms (Heller and Cryer, 1991; Dagogo-Jack et al., 1993) identified a mechanism thatexplains why some individuals lose their hypoglycaemia warning symptoms. Further studiesdemonstrated that such episodes did not need to be symptomatic to produce alterations inphysiological responses. Veneman et al. (1993) induced hypoglycaemia in 10 non-diabeticsubjects overnight and tested their physiological responses to hypoglycaemia the followingmorning. They reported lower symptomatic and hormonal responses when compared to anon-hypoglycaemia control night. These relatively mild levels of hypoglycaemia are rarely

CONSEQUENCES OF NOCTURNAL HYPOGLYCAEMIA 89reported by patients because they remain asleep, but may contribute to the development ofimpaired awareness of hypoglycaemia.The ‘Dead in Bed’ SyndromeThe possible contribution of hypoglycaemia to cardiac arrhythmias and sudden death isdiscussed in detail in Chapter 12. However, since it has been proposed that nocturnalhypoglycaemia may be a specific precipitant, it merits a brief mention here. A detailed surveyof unexpected deaths in young people with type 1 diabetes in the early 1990s highlighted arare but distinct mode of death in which young people were found lying in an undisturbed bed.Autopsy failed to reveal a structural cause but circumstantial evidence implicated nocturnalhypoglycaemia as a precipitant. Not all affected individuals had strict glycaemic control, butmany were known to have been susceptible to developing nocturnal hypoglycaemia. Thistype of sudden and unexpected death has since been confirmed in epidemiological surveysin other countries, possibly accounting for between 5–10% of all deaths under the age of40 in people with diabetes. Hypoglycaemia causes an increase in the QT interval and onepossible explanation is that an episode of nocturnal hypoglycaemia triggers a ventriculararrhythmia in a susceptible individual. However, further work is required to establish thisplausible hypothesis as the cause of these deaths.Neurological Consequences of Nocturnal Hypoglycaemia on CerebralFunctionThe possibility that recurrent exposure to hypoglycaemia, particularly occurring during sleep,might insidiously damage cerebral function and cause permanent cognitive impairment wasraised by the finding that children who had developed type 1 diabetes before the age offive years, exhibited cognitive impairment when compared with non-diabetic controls (Ryanet al., 1985). This observation was replicated in different studies using a range of cognitivetests (Bjorgaas et al., 1997; Rovet and Ehrlich, 1999).However, the relationship to hypoglycaemia, particularly during the night, was unclear(Golden et al., 1989); hypoglycaemia-induced convulsions have also been implicated, and itis possible that other factors associated with early-onset diabetes may contribute to cognitiveimpairment.Several studies have explored the extent to which nocturnal episodes may affect performanceon the following day. Bendtson et al. (1992) found no difference in cognitive performanceamong adults with type 1 diabetes when tested on the morning after an episode ofnocturnal hypoglycaemia in comparison with a night with no hypoglycaemia. Similar findingshave been reported after the induction of experimental hypoglycaemia during the night (Kinget al., 1998) although the subjects were more fatigued on the following morning. It seemsthat even children are relatively unaffected by a single episode of nocturnal hypoglycaemia.Matyka et al. (1999b) tested pre-pubertal children after episodes of prolonged, spontaneouslyoccurring, hypoglycaemia that occurred during sleep, many of which lasted several hours,and found no deleterious effect on cognitive function on the following morning, althoughmood was adversely affected.In summary, although it seems plausible that recurrent nocturnal hypoglycaemia mightcontribute to cognitive decline, on the available evidence the verdict remains unproven.

90 NOCTURNAL HYPOGLYCAEMIACAN NOCTURNAL HYPOGLYCAEMIA BE PREDICTED?The ability to predict whether a nocturnal hypoglycaemic episode is likely, by measuringglucose concentrations at bedtime, has generally been tested in studies using intermittentblood glucose sampling to detect hypoglycaemia. It is likely that such a design will failto identify all episodes. The results of two early studies that observed patients who werereceiving one or two daily injections of insulin, suggested that a bedtime glucose concentrationof 6.0 mmol/l or less indicated an 80% chance of a period of biochemical hypoglycaemiaduring the night (Pramming et al., 1985; Whincup and Milner, 1987). Furthermore, theseearly studies were generally undertaken among patients who were taking two injectionsof conventional insulins each day. Studies of children have suggested that blood glucosemeasurements at bedtime are less predictive of hypoglycaemia in the first half of thenight, although a value of < 75 mmol/l does indicate an increased risk (Matyka et al.,1999b). That study and others have also shown that a low fasting blood glucose is astrong indicator that hypoglycaemia has occurred in the latter half of the night (Matyka,2002).Since severe nocturnal hypoglycaemia is relatively rare, it is much more difficult toestablish to what extent bedtime blood glucose values are predictive. A previous historyof severe hypoglycaemia, impaired awareness of hypoglycaemia or strict control with lowHbA 1c values, will confer a greater chance of a severe episode but are a poor guide to therisk of developing a severe nocturnal episode on any particular night.THE SOMOGYI PHENOMENON: THE CONCEPT OF REBOUNDHYPERGLYCAEMIAIn the late 1930s, a Hungarian biochemist, Michael Somogyi, working in St Louis, USA,suggested that nocturnal hypoglycaemia might provoke rebound hyperglycaemia on thefollowing morning, and he supported his hypothesis with a demonstration that reducingevening doses of insulin led to a reduction in fasting urinary glycosuria (Somogyi, 1959). Heproposed that nocturnal hypoglycaemia provokes a counterregulatory response with rises inplasma epinephrine, cortisol and growth hormone resulting in the release of glucose from theliver and inhibition of the effects of insulin over the next few hours. The logical conclusionfrom his hypothesis was that this ‘rebound’ elevated fasting blood glucose in the morningshould be treated, not by an increase in the evening dose of insulin, but paradoxically by areduction. The idea of ‘rebound hyperglycaemia’ following nocturnal hypoglycaemia, (alsoknown as the Somogyi phenomenon) as an explanation for a high fasting blood glucosein insulin-treated patients, has proved to be very attractive to many diabetes healthcareprofessionals who firmly believe in its existence. The consequences are important as patientsare often advised to reduce their evening insulin dose, particularly if they complain ofnocturnal hypoglycaemia. However, its clinical relevance was challenged over 20 years agoand repeated studies have established that fasting hyperglycaemia is largely a result of fallingplasma insulin concentrations during the night, as the subcutaneous depot of insulin that wasinjected the day before is dissipated.Gale et al. (1980) demonstrated that periods of hypoglycaemia during the night wereoften prolonged and were not accompanied by a large rise in counterregulatory hormones.

THE SOMOGYI PHENOMENON 91Probability of hypoglycaemic episodes in the night1. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21Morning blood glucose (mmol/I)Figure 4.3 Risk of nocturnal hypoglycaemia according to fasting morning blood glucose (95% Cl) in594 nights. Black bars = hypoglycaemic nights; shaded bars = possibly hypoglycaemic nights. Reproducedfrom Hoi-Hansen et al. (2005). With kind permission from Springer Science and Business MediaAlthough fasting glucose concentrations were frequently high in the patients they studied,this was related directly to a waning of circulating plasma insulin concentrations. Someinvestigators have demonstrated that when hypoglycaemia is experimentally-induced duringthe night, this can raise blood glucose on the following morning, even if circulating plasmainsulin concentrations are maintained (Perriello et al., 1988). However, the additional increasein the fasting blood glucose concentration is modest (around 2.0 mmol/l) and its clinicalrelevance is questionable. Other researchers have found no effect on daytime concentrationsof blood glucose after lowering blood glucose to hypoglycaemic levels during the night(Hirsch et al., 1990). Careful analysis of data collected both by self-monitoring of bloodglucose (Havlin and Cryer, 1987) and by continuous glucose monitoring (Hoi-Hansen et al.,2005) (Figure 4.3) during everyday activities, has also shown that nocturnal hypoglycaemiadoes not provoke rebound fasting hyperglycaemia.Many insulin-treated diabetic patients experience high fasting blood glucose levels butthis common clinical problem is essentially a consequence of inadequate basal insulinreplacement overnight. The important mechanisms that contribute to fasting hyperglycaemiaappear to be a combination of waning plasma insulin levels and glucose release fromthe liver secondary to nocturnal spikes of growth hormone secretion (Campbell et al.,1985), a physiological process termed the ‘dawn phenomenon’. High blood glucose levelsfollowing symptomatic nocturnal hypoglycaemia may also result from excessive intakeof oral carbohydrate, ingested as treatment of the hypoglycaemia, rather than a powerfulcounterregulatory response, which is usually suppressed at night. The clinical message istherefore clear: fasting hyperglycaemia indicates a need for adjusting basal insulin in termsof type or timing rather than reducing the dose. Some useful clinical steps to be undertakenin patients presenting with this problem are listed in Box 4.2.

92 NOCTURNAL HYPOGLYCAEMIABox 4.2 Clinical approach to high fasting blood glucose complicated by nocturnalhypoglycaemia1. Measure blood glucose at 2–3 a.m. over a few days.2. If nocturnal hypoglycaemia is present, ensure that basal insulin is taken at bedtime(i.e., split pre-mixed evening insulin).3. Progressively increase bedtime long-acting insulin in doses of 2–4 units whilechecking with 3 a.m. blood glucose measurements that this is not precipitatingnocturnal hypoglycaemia4. Use a long-acting insulin analogue, either glargine or detemir.5. Teach patients to take an appropriate (but not excessive) quantity of a high energyglucose drink, orange juice or glucose as sweets or tablets to treat nocturnal hypoglycaemicepisodes.6. When available, consider obtaining a continuous glucose monitoring profile.CLINICAL SOLUTIONS (BOX 4.3)Dietary MeasuresA time honoured approach to reducing the frequency of nocturnal hypoglycaemia has beento counteract the effect of insulin by ensuring that patients eat a bedtime snack. This seemedparticularly important for children who go to bed early and sleep for many hours betweentheir evening insulin dose and breakfast on the following day. It clearly makes sense for allindividuals taking insulin to measure their blood glucose before bed and to take additionalfood if their blood glucose is low. However, the extent to which protective eating can preventnocturnal hypoglycaemia in patients taking intermittent injections of insulin is limited.Box 4.3Potential remedies for problematical nocturnal hypoglycaemia1. Long and rapid-acting insulin analogues.2. -agonists (terbutaline or salbutamol).3. Appropriate snacks:– uncooked cornstarch– high protein foods.4. CSII.

CLINICAL SOLUTIONS 93The effectiveness of different approaches has generally been assessed by measuring theeffectiveness of different snacks in preventing biochemical or symptomatic episodes ofhypoglycaemia. Before the advent of continuous blood glucose monitoring, such studiesdemanded regular blood sampling for measurement of glucose and since this generallyrequired admission to hospital for study, these investigations were not exactly examininga typical clinical situation. Furthermore, the number of subjects studied has generally beenrelatively small, and although sufficient to measure changes in blood glucose concentrationor frequency of biochemical hypoglycaemia, the studies have been inadequately poweredto determine effects on the frequency of severe episodes. Some studies have examinedthe potential of carbohydrate foods that are absorbed slowly and thus able to counter thehypoglycaemic effect of insulin over longer periods. Uncooked cornstarch, which has a verylow glycaemic index, has been studied intensively, particularly because it is used successfullyto prevent severe hypoglycaemia in glycogen storage diseases (Goldberg and Slonim, 1993).In one study, when young people were given uncooked cornstarch incorporated into a normalbedtime snack, the incidence of symptomatic and biochemical nocturnal hypoglycaemiawas three-fold lower (Kaufman and Devgan, 1996). A blinded randomised trial reported asimilarly lower frequency of nocturnal episodes (Kaufman et al., 1997). Other work hasexamined the effect of snacks rich in protein or fat.Different types of snack were compared against placebo in a trial using a crossoverdesign in 15 adults with type 1 diabetes (Kalergis et al., 2003). When patients retired tobed with blood glucose greater than 10.0 mmol/l, nocturnal episodes were not observed.At a pre-bedtime glucose below 7.0 mmol/l, a high protein snack prevented any episodeof nocturnal hypoglycaemia by contrast with either a standard snack or one containingcornstarch (Figure 4.4). Thus the clinical data measuring the effectiveness of cornstarch are100%80%% Nights60%40%20%0% Pl S CS Prot Pl S CS Prot Pl S CS Prot10 mmol/LBedtime blood glucose categoryFigure 4.4 Effect of snacks and bedtime blood glucose concentration on frequency of nocturnalhypoglycaemia. Open box = neither nocturnal hypoglycaemia nor morning hypoglycaemia; solidblack box = hypoglycaemia; striped box = morning hypoglycaemia only. Snack type: Pl = placebo;S = standard diet; CS = cornstarch; prot = protein

94 NOCTURNAL HYPOGLYCAEMIAconflicting. Since uncooked cornstarch is not easy to prepare in a digestible form, it is notsurprising that it is not used widely to prevent nocturnal hypoglycaemia.An alternative approach to dietary supplements is to reduce the rate of absorption ofcarbohydrates using a disaccharidase inhibitor such as acarbose. Three studies of theseagents have examined their effects in patients with type 1 diabetes. McCulloch et al. (1983)studied the effect of acarbose on the risk of overnight hypoglycaemia, and found that therisk of symptomatic nocturnal hypoglycaemia was lowered by 39%. Taira et al. (2000)reported similar benefits using voglibose. However, a recent study reported no benefit ofacarbose over placebo when both pharmaceutical and snacking interventions were comparedwith respect to their effect on preventing hypoglycaemia (Raju et al., 2006). In the light ofthese limited and conflicting data it seems unlikely that acarbose will never be widely used,particularly as sucrose-containing products cannot be used as a treatment for hypoglycaemiawhen acarbose is being taken.Pharmaceutical InterventionsThere are indications that -agonists may have some use in reducing the risk of nocturnalhypoglycaemia. For some years inhaled terbutaline has been proposed as a method ofelevating blood glucose. In the early 1990s, Wiethop and Cryer (1993) demonstrated thatits oral or subcutaneous delivery following induced hypoglycaemia, led to a rise in bloodglucose compared to placebo, an effect that lasted for some hours. More recently, Wright andWales (2003) reported that children with type 1 diabetes who were receiving treatment forasthma had fewer episodes of nocturnal hypoglycaemia when compared to a non-asthmaticgroup of children in a survey lasting three months, and they implicated a beneficial effectof -agonist therapy. Raju et al. (2006) also compared the effect of inhaled terbutaline withother therapeutic remedies such as cornstarch, standard snacks and acarbose on nocturnalblood glucose levels in patients with type 1 diabetes. They found that terbutaline preventednocturnal hypoglycaemia in all 15 subjects. However, although this treatment offered thegreatest protection against hypoglycaemia at night, it also led to the highest fasting bloodglucose among the different remedies, indicating that additional work needs to be done toestablish this treatment as a realistic therapeutic option.Timing and Type of Insulin, Including Insulin AnaloguesThe introduction of insulin analogues with pharmacokinetic properties that bear more resemblanceto physiological insulin profiles in non-diabetic individuals highlighted the potentialof such preparations to lower the risk of hypoglycaemia. Over a full 24 hours, the overallfrequency of symptomatic hypoglycaemia in trials of rapid-acting insulin analogues hasbeen modestly lower than with conventional insulins, but a consistent finding has been alower rate of nocturnal hypoglycaemia. It appears that the tendency of conventional solubleinsulin to self-associate into hexamers at therapeutic concentrations leads to increasingplasma insulin levels with repeated injection. Since rapid-acting insulin analogues separateinto single molecules much more readily, accumulation of insulin is less likely and therisk of nocturnal hypoglycaemia is subsequently lower. The frequency of nocturnal hypoglycaemiaobserved in clinical trials has been variable. Rates of nocturnal hypoglycaemiahave been over 50% lower in some trials involving patients with type 1 diabetes with strict

CONCLUSIONS 95glycaemic control (Heller et al., 1999; Heller et al., 2004), and although this has generallybeen demonstrated for symptomatic episodes, these data might reflect a lower frequency ofsevere hypoglycaemia (Holleman et al., 1997).The long-acting analogues – insulins glargine and detemir – which provide basal insulinreplacement, also appear to lower the risk of nocturnal hypoglycaemia. They have a longerduration of action when compared to isophane (NPH) insulin, together with less of a peakin their time-action profile and a more consistent duration of action (Barnett, 2003). Bothof these properties probably contribute to the lower rates of nocturnal hypoglycaemia thathave been observed during clinical trials. Relative risk reductions of around 30% have beenreported for long-acting preparations in trials involving patients with type 1 (De Leeuw et al.,2005; Pieber et al., 2000) and with type 2 diabetes (Hermansen et al., 2006; Yki-Jarvinenet al., 2000).The combination of both rapid and long-acting insulin analogues might be expected to havea particularly powerful effect in lowering the risk of nocturnal hypoglycaemia. In the fewstudies comparing combinations of insulin analogues to conventional insulins this appears tobe the case. Hermansen et al. (2004) reported a 55% lower rate of symptomatic nocturnalhypoglycaemia when using an insulin detemir/aspart combination as basal-bolus therapy inpatients with type 1 diabetes, which was accompanied by a modest but significant fall in HbA 1cof 0.2%. Ashwell et al. (2006) observed a fall in HbA 1c of 0.5% using insulin glargine andlispro in a basal-bolus regimen in patients with type 1 diabetes, while nocturnal hypoglycaemiawas 44% lower in frequency. However, nocturnal hypoglycaemia was not eradicated,leading to the conclusion that although insulin analogues may help to reduce the side-effectsof insulin therapy, they do not approach the requirements of therapeutic insulin delivery.Continuous Subcutaneous Insulin Infusion (CSII)Since nocturnal hypoglycaemia is largely the result of inadequate basal insulin replacement,one would expect that the most effective method of basal insulin delivery currently available,CSII, could limit the frequency of nocturnal hypoglycaemia (Pickup and Keen, 2002).However, early studies reported surprisingly little effect on hypoglycaemia, perhaps becauseof a failure to train patients in the essential related skills of carbohydrate and insulindose adjustment. More recent work has indicated that CSII can reduce overall rates ofhypoglycaemia (Bode et al., 1996; Boland et al., 1999; Kanc et al., 1998), but specificreporting on rates of nocturnal episodes is unusual. Furthermore, few trials have used arandomised design, suggesting that a lower risk might relate to other characteristics ofthose who use CSII rather than to the technology itself. Thus the amount to which modernpump therapy lowers the risk of nocturnal hypoglycaemia has still be to be established.Nevertheless, in those who have experienced recurrent nocturnal episodes and who have notimproved after a trial of insulin analogues, it seems worthwhile undertaking a trial of CSII.CONCLUSIONS• Nocturnal hypoglycaemia remains an unresolved clinical side-effect of insulin therapypreventing the attainment of strict glycaemic control for many people. It contributes tomorbidity, and perhaps mortality, in patients with type 1 diabetes.

96 NOCTURNAL HYPOGLYCAEMIA• Nocturnal hypoglycaemia is caused chiefly by the limitations of current methods ofinsulin delivery. The inability of subcutaneous insulin therapy, particularly conventionalbasal insulins, to maintain low-level stable insulin concentrations overnight, leads both tonocturnal hypoglycaemia and fasting hyperglycaemia.• In addition to the limitations of insulin delivery, hypoglycaemia is also more commonas a result of diminished counterregulatory responses overnight, associated in part withsleep, which has a specific inhibitory effect on physiological defences to hypoglycaemia,and a supine posture which suppresses autonomic responses.• Nocturnal hypoglycaemia is a particular problem in children, partly as a consequence ofthe long period of fasting between their evening meal and their breakfast.• The risk of nocturnal hypoglycaemia is greatest in those who have a bedtime glucosebelow 7.0 mmol/l. Bedtime snacks can reduce the risk during the early part of the night.Specific foods (uncooked cornstarch or protein rich snacks) reduce the risk of nocturnalhypoglycaemia in some studies.• Rebound hyperglycaemia (the ‘Somogyi phenomenon’) is rarely caused by counterregulatoryhormone release provoked by an overnight hypoglycaemic episode, since hormonalsecretion is generally suppressed. It is mainly a consequence of waning of circulatingplasma insulin levels and should be treated by an adjustment in the timing and type ofinsulin used rather than a reduction in insulin dose.• The problem of nocturnal hypoglycaemia may respond to the use of insulin analogues(both rapid and long-acting) or to CSII with an insulin pump. However, its eradicationwill depend upon new methods of insulin delivery.REFERENCESAshwell SG, Amiel SA, Bilous RW, Dashora U, Heller SR, Hepburn DA et al. (2006). Improvedglycaemic control with insulin glargine plus insulin lispro: a multicentre, randomized, cross-overtrial in people with type 1 diabetes. Diabetic Medicine 23: 285–92.Banarer S, Cryer PE (2003). Sleep-related hypoglycemia-associated autonomic failure in type 1diabetes: reduced awakening from sleep during hypoglycemia. Diabetes 52: 1195–203.Barnett AH (2003). A review of basal insulins. Diabetic Medicine 20: 873–85.Bendtson I, Kverneland A, Pramming S, Binder C (1988). Incidence of nocturnal hypoglycaemia ininsulin-dependent diabetes. Acta Endocrinologica 223: 543–8.Bendtson I, Gade J, Theilgard A, Binder C (1992). Cognitive function in type 1 IDD patients afternocturnal hypoglycaemia. Diabetologia 35: 898–903.Bendtson I, Rosenfalck AM, Binder C (1993). Nocturnal versus diurnal hormonal counterregulation tohypoglycemia in type 1 (insulin-dependent) diabetic patients. Acta Endocrinologica 128: 109–15.Beregszaszi M, Tubiana-Rufi N, Benali K, Noel M, Bloch J, Czernichow P (1997). Nocturnal hypoglycemiain children and adolescents with insulin-dependent diabetes mellitus: prevalence and riskfactors. Journal of Pediatrics 131: 27–33.Bjorgaas M, Gimse R, Vik T, Sand T (1997). Cognitive function in type 1 diabetic children with andwithout episodes of severe hypoglycaemia. Acta Paediatrica 86: 148–53.Bode BW, Steed RD, Davidson PC (1996). Reduction in severe hypoglycemia with long-term continuoussubcutaneous insulin infusion in type I diabetes. Diabetes Care 19: 324–7.

REFERENCES 97Boland EA, Grey M, Oesterle A, Fredrickson L, Tamborlane WV (1999). Continuous subcutaneousinsulin infusion. A new way to lower risk of severe hypoglycemia, improve metabolic control, andenhance coping in adolescents with type 1 diabetes. Diabetes Care 22: 1779–84.Boland E, Monsod T, Delucia M, Brandt CA, Fernando S, Tamborlane WV (2001). Limitations ofconventional methods of self-monitoring of blood glucose: lessons learned from 3 days of continuousglucose sensing in pediatric patients with type 1 diabetes. Diabetes Care 24: 1858–62.Campbell PJ, Bolli GB, Cryer PE, Gerich JE (1985). Pathogenesis of the dawn phenomenon in patientswith insulin-dependent diabetes mellitus. Accelerated glucose production and impaired glucoseutilization due to nocturnal surges in growth hormone secretion. New England Journal of Medicine312: 1473–9.Dagogo-Jack SE, Craft S, Cryer PE (1993). Hypoglycemia-associated autonomic failure in insulindependentdiabetes mellitus. Journal of Clinical Investigation 91: 819–28.De Leeuw I, Vague P, Selam JL, Skeie S, Lang H, Draeger E, Elte JW (2005). Insulin detemir usedin basal-bolus therapy in people with type 1 diabetes is associated with a lower risk of nocturnalhypoglycaemia and less weight gain over 12 months in comparison to NPH insulin. Diabetes Obesityand Metabolism 7: 73–82.Dornan TL, Orde Peckar C, Mayon-White VA, Knight AH, Moore RA, Hockaday TDR et al. (1981).Unsuspected hypoglycaemia, haemoglobin A1 and diabetic control. Quarterly Journal of Medicine197: 31–8.Gais S, Born J, Peters A, Schultes B, Heindl B, Fehm HL, Werner K (2003). Hypoglycemia counterregulationduring sleep. Sleep 26: 55–9.Gale EA, Tattersall RB (1979). Unrecognised nocturnal hypoglycaemia in insulin-treated diabetics.Lancet 1: 1049–52.Gale EA, Kurtz AB, Tattersall RB (1980). In search of the Somogyi effect. Lancet 2: 279–82.Goldberg T, Slonim AE (1993). Nutrition therapy for hepatic glycogen storage diseases. Journal ofthe American Dietetic Association 93: 1423–30.Golden MP, Ingersoll GM, Brack CJ, Russell BA, Wright JC, Huberty TJ (1989). Longitudinalrelationship of asymptomatic hypoglycemia to cognitive function in IDDM. Diabetes Care 12:89–93.Havlin CE, Cryer PE (1987). Nocturnal hypoglycemia does not commonly result in major morninghyperglycemia in patients with diabetes mellitus. Diabetes Care 10: 141–7.Heller SR, Cryer PE (1991). Reduced neuroendocrine and symptomatic responses to subsequenthypoglycemia after 1 episode of hypoglycemia in non-diabetic humans. Diabetes 40: 223–6.Heller SR, Amiel SA, Mansell P (1999). Effect of the fast-acting insulin analog lispro on the risk ofnocturnal hypoglycemia during intensified insulin therapy. Diabetes Care 22: 1607–11.Heller SR, Colagiuri S, Vaaler S, Wolffenbuttel BH, Koelendorf K, Friberg HH et al. (2004). Hypoglycaemiawith insulin aspart: a double-blind, randomised, crossover trial in subjects with type 1diabetes. Diabetic Medicine 21: 769–75.Hermansen K, Fontaine P, Kukolja KK, Peterkova V, Leth G, Gall MA (2004). Insulin analogues(insulin detemir and insulin aspart) versus traditional human insulins (NPH insulin and regularhuman insulin) in basal-bolus therapy for patients with type 1 diabetes. Diabetologia 47: 622–9.Hermansen K, Davies M, Derezinski T, Martinez Ravn G, Clauson P, Home P (2006). A 26-week,randomized, parallel, treat-to-target trial comparing insulin detemir with NPH insulin as add-ontherapy to oral glucose-lowering drugs in insulin-naive people with type 2 diabetes. Diabetes Care29: 1269–74.Hirsch IB, Smith LJ, Havlin CE, Shah SD, Clutter WE, Cryer PE (1990). Failure of nocturnalhypoglycemia to cause daytime hyperglycemia in patients with IDDM. Diabetes Care 13: 133–42.Hirsch IB, Heller SR, Cryer PE (1991). Increased symptoms of hypoglycaemia in the standing positionin insulin-dependent diabetes mellitus. Clinical Science 80: 583–6.Hoi-Hansen T, Pedersen-Bjergaard U, Thorsteinsson B (2005). The Somogyi phenomenon revisitedusing continuous glucose monitoring in daily life. Diabetologia 48: 2437–8.

98 NOCTURNAL HYPOGLYCAEMIAHolleman F, Schmitt H, Rottiers R, Rees A, Symanowski S, Anderson JH, The Benelux-UK InsulinLispro Study Group (1997). Reduced frequency of severe hypoglycemia and coma in well-controlledIDDM patients treated with insulin lispro. Diabetes Care 20: 1827–32.Jones TW, Porter P, Sherwin RS, Davis EA, O’Leary P, Frazer F et al. (1998). Decreased epinephrineresponses to hypoglycemia during sleep. New England Journal of Medicine 338: 1657–62.Kalergis M, Schiffrin A, Gougeon R, Jones PJ, Yale JF (2003). Impact of bedtime snack compositionon prevention of nocturnal hypoglycemia in adults with type 1 diabetes undergoing intensive insulinmanagement using lispro insulin before meals: a randomized, placebo-controlled, crossover trial.Diabetes Care 26: 9–15.Kanc K, Janssen MM, Keulen ET, Jacobs MA, Popp-Snijders C, Snoek FJ, Heine RJ (1998). Substitutionof night-time continuous subcutaneous insulin infusion therapy for bedtime NPH insulin in amultiple injection regimen improves counterregulatory hormonal responses and warning symptomsof hypoglycaemia in IDDM. Diabetologia 41: 322–9.Kaufman FR, Devgan S (1996). Use of uncooked cornstarch to avert nocturnal hypoglycemia inchildren and adolescents with type I diabetes. Journal of Diabetes and Its Complications 10: 84–7.Kaufman FR, Halvorson M, Kaufman ND (1997). Evaluation of a snack bar containing uncookedcornstarch in subjects with diabetes. Diabetes Research and Clinical Practice 35: 27–33.Kaufman FR, Austin J, Neinstein A, Jeng L, Halvorson M, Devoe DJ, Pitukcheewanont P (2002).Nocturnal hypoglycemia detected with the Continuous Glucose Monitoring System in pediatricpatients with type 1 diabetes. Journal of Pediatrics 141: 625–30.King P, Kong MF, Parkin H, Macdonald IA, Tattersall RB (1998). Well-being, cerebral function, andphysical fatigue after nocturnal hypoglycemia in IDDM. Diabetes Care 21: 341–5.Kulcu E, Tamada JA, Reach G, Potts RO, Lesho MJ (2003). Physiological differences betweeninterstitial glucose and blood glucose measured in human subjects. Diabetes Care 26: 2405–9.Matyka KA, Crowne EC, Havel PJ, Macdonald IA, Matthews D, Dunger DB (1999a). Counterregulationduring spontaneous nocturnal hypoglycemia in prepubertal children with type 1 diabetes.Diabetes Care 22: 1144–50.Matyka KA, Wigg L, Pramming S, Stores G, Dunger DB (1999b). Cognitive function and mood afterprofound nocturnal hypoglycaemia in prepubertal children with conventional insulin treatment fordiabetes. Archives of Disease in Childhood, 81: 138–42.Matyka KA (2002). Sweet dreams? Nocturnal hypoglycemia in children with type 1 diabetes. PediatricDiabetes 3: 74–81.McCulloch DK, Kurtz AB, Tattersall RB (1983). A new approach to the treatment of nocturnalhypoglycemia using alpha-glucosidase inhibition. Diabetes Care 6: 483–7.McGowan K, Thomas W, Moran A (2002). Spurious reporting of nocturnal hypoglycemia by CGMSin patients with tightly controlled type 1 diabetes. Diabetes Care 25: 1499–503.Oswald I (1987). The normal record of sleep. In: A Textbook of Clinical Neurophysiology. Halliday AM,Butler SR and Paul R, eds. John Wiley & Sons Inc., New York: 173–85.Perriello G, De Feo P, Torlone E, Calcinaro F, Ventura MM, Basta G et al. (1988). The effectof asymptomatic nocturnal hypoglycemia on glycemic control in diabetes mellitus. New EnglandJournal of Medicine 319: 1233–9.Pickup J, Keen H (2002). Continuous subcutaneous insulin infusion at 25 years: evidence base for theexpanding use of insulin pump therapy in type 1 diabetes. Diabetes Care 25: 593–8.Pieber TR, Eugene-Jolchine I, Derobert E (2000). Efficacy and safety of HOE 901 versus NPH insulinin patients with type 1 diabetes. The European Study Group of HOE 901 in type 1 diabetes. DiabetesCare 23: 157–62.Porter PA, Keating B, Byrne G, Jones TW (1997). Incidence and predictive criteria of nocturnalhypoglycemia in young children with insulin-dependent diabetes mellitus. Journal of Pediatrics130: 366–72.Pramming S, Thorsteinsson B, Bendtson I, Ronn B, Binder C (1985). Nocturnal hypoglycaemia inpatients receiving conventional treatment with insulin. British Medical Journal 291: 376–9.

REFERENCES 99Raju B, Arbelaez AM, Breckenridge SM, Cryer PE (2006). Nocturnal hypoglycemia in type 1 diabetes:an assessment of preventive bedtime treatments. Journal of Clinical Endocrinology and Metabolism91: 2087–92.Rizza RA, Gerich JE, Haymond MW, Westland RE, Hall LD, Clemens AH, Service FJ (1980).Control of blood sugar in insulin-dependent diabetes: comparison of an artificial endocrine pancreas,continuous subcutaneous insulin infusion and intensified conventional insulin therapy. New EnglandJournal of Medicine 303: 1313–18.Robinson AM, Parkin HM, Macdonald IA, Tattersall RB (1994). Physiological response to posturalchange during mild hypoglycaemia in patients with IDDM. Diabetologia 37: 1241–50.Rovet JF, Ehrlich RM (1999). The effect of hypoglycemic seizures on cognitive function in childrenwith diabetes: a 7-year prospective study. Journal of Pediatrics 134: 503–6.Ryan C, Vega A, Drash A (1985). Cognitive deficits in adolescents who developed diabetes early inlife. Pediatrics 75: 921–7.Shalwitz RA, Farkas-Hirsch R, White NH, Santiago JV (1990). Prevalence and consequences ofnocturnal hypoglycemia among conventionally treated children with diabetes mellitus. Journal ofPediatrics 116: 685–9.Somogyi M (1959). Exacerbation of diabetes by excess insulin action. American Journal of Medicine26: 169–91.Taira M, Takasu N, Komiya I, Taira T, Tanaka H (2000). Voglibose administration before the eveningmeal improves nocturnal hypoglycemia in insulin-dependent diabetic patients with intensive insulintherapy. Metabolism 49: 440–3.Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J (1993.) Induction of hypoglycemia unawarenessby asymptomatic nocturnal hypoglycemia. Diabetes 42: 1233–7.Vervoort G, Goldschmidt HM, van Doorn LG (1996). Nocturnal blood glucose profiles in patients withtype 1 diabetes mellitus on multiple (> or = 4) daily insulin injection regimens. Diabetic Medicine13: 794–9.Wentholt IM, Vollebregt MA, Hart AA, Hoekstra JB, DeVries JH (2005). Comparison of a needletypeand a microdialysis continuous glucose monitor in type 1 diabetic patients. Diabetes Care 28:2871–6.Whincup G, Milner RD (1987). Prediction and management of nocturnal hypoglycaemia in diabetes.Archives of Disease in Childhood 62: 333–7.Wiethop BV, Cryer PE (1993). Alanine and terbutaline in treatment of hypoglycemia in IDDM.Diabetes Care 16: 1131–6.Wright NP, Wales JK (2003). The incidence of hypoglycaemia in children with type 1 diabetes andtreated asthma. Archives of Disease in Childhood 88: 155–6.Yki-Jarvinen H, Dressler A, Ziemen M (2000). Less nocturnal hypoglycemia and better post-dinnerglucose control with bedtime insulin glargine compared with bedtime NPH insulin during insulincombination therapy in type 2 diabetes. HOE 901/3002 Study Group. Diabetes Care 23: 1130–6.

5 Moderators, Monitoring andManagement of HypoglycaemiaTristan Richardson and David KerrINTRODUCTIONDespite advances in insulin pharmacology and delivery and in patient education, the lifetimefrequency of symptomatic hypoglycaemia remains substantial, with the average patient likelyto experience thousands of episodes over the course of their life with insulin-treated diabetes.Furthermore, there are a number of everyday factors that continue to influence (moderate) therisk, presentation and rate of recovery from low blood glucose concentrations (Table 5.1, andsee Chapter 3). An understanding of the predisposing factors that influence hypoglycaemia isimportant to allow appropriate advice and education to be given to patients, and to appreciatethe relative importance of different moderators and reduce their risks (real and perceived),particularly of causing recurrent severe hypoglycaemia.RISK FACTORS FOR THE DEVELOPMENTOF HYPOGLYCAEMIAIn physiological terms, current insulin regimens are less than ideal. Imperfect insulin is oftengiven at the ‘wrong’ time, in the ‘wrong’ place and at the ‘wrong’ dose. Unsurprisingly, amismatch between insulin and carbohydrate absorption frequently occurs, leading to overinsulinisationand the potential risk of hypoglycaemia. Several other factors can influenceinsulin absorption after subcutaneous injection including:• depth of the injection;• angle of the needle for giving the injection;• site of the injection;• presence of lipohypertrophy at injection site;• time of day that injection is given;• phase of the menstrual cycle;• relationship with exercise;Hypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

102 MODERATORS, MONITORING AND MANAGEMENTTable 5.1Causes of hypoglycaemia in patients with diabetesChange in insulinsensitivityChange in insulinpharmaco-dynamicsAltered insulin:carbohydrate ratioOther related conditions‘Honeymoon period’in newly diagnosedtype 1 diabetesPost-partumChange in injection site Unplanned exercise Endocrine dysfunction(Addison’s disease,hypopituitarism)Change in insulin Change in social PsychologicalformulationcircumstancesMenstruation Change in temperature BreastfeedingAlcoholGastroparesisRenal failureExerciseMalabsorption e.g.Coeliac Disease• ambient temperature;• psychological factors such as mood;• use of other medicines affecting skin blood flow;• for pre-mixed insulin preparations – whether the insulin has been shaken adequately beforethe injection.Although there is now greater awareness of the importance of educating patients aboutthe nuances of dietary carbohydrate counting, it is important to remember that the absorptionof food is also variable within individuals, being affected by the constituents and size of ameal, and the speed with which it is eaten. In addition, gastric emptying is variable withinan individual with diabetes and is influenced by the status of the autonomic nervous system(Feldman and Schiller, 1983; Vinik et al., 2003).In addition, improving glycaemic control and lowering HbA 1c per se are also associatedwith a trebling of risk for hypoglycaemia (The Diabetes Control and Complications TrialResearch Group, 1995). However, this alone does not wholly predict the occurrence ofhypoglycaemia. In the intensively treated group in the DCCT, HbA 1c only accounted for60% of the risk of severe hypoglycaemia (The Diabetes Control and Complications TrialResearch Group, 1997). Additional risk factors have also been identified, many of which arediscussed in detail in Chapter 3:• previous episodes of severe hypoglycaemia (Bott et al., 1997; The Diabetes Control andComplications Trial Research Group, 1997);• long duration of type 1 diabetes (Cox et al., 1994; The Diabetes Control and ComplicationsTrial Research Group, 1997);• intensive insulin therapy (The DCCT Research Group, 1991);• strict glycaemic control (The Diabetes Control and Complications Trial Research Group,1997);

LIFESTYLE MODERATORS 103• absolute insulin deficiency (Bott et al., 1997; The Diabetes Control and ComplicationsTrial Research Group, 1997);• sleep (Pramming et al., 1990; The Diabetes Control and Complications Trial ResearchGroup, 1997);• impaired hypoglycaemia awareness (Gold et al., 1994);• alcohol (Richardson et al., 2005b);• exercise (MacDonald, 1987);• pregnancy (Rosen et al., 1995);• impaired renal function (Muhlhauser et al., 1991).Despite these risk factors being recognised, it is often difficult to untangle the majorprecipitating factors for a given episode of severe hypoglycaemia. Patient recall is oftenvague, preconceptions may have been applied and amnesia for an event is common. Anecdotally,around a third to half of episodes remain unexplained in routine clinical practice,although patient ‘error’ is still perceived (by healthcare professionals) as the most likely‘cause’ of a hypoglycaemic episode (The DCCT Research Group, 1991).LIFESTYLE MODERATORSNumerous lifestyle influences predispose towards hypoglycaemia, and some of these, suchas pregnancy (Chapter 10) and exercise (Chapter 14), are discussed elsewhere in this book.Alcohol and HypoglycaemiaAlcohol is an important risk factor for hypoglycaemia for individuals treated with insulin,with estimates suggesting that up to 20% of severe hypoglycaemic events are attributableto its use (Potter et al., 1982; Nilsson et al., 1988). However, there is nothing to suggestthat (in general terms) people with type 1 diabetes adopt a different approach to their use ofalcohol than the rest of the population.Alcohol has been associated with hypoglycaemia in several ways:• Ingestion of even small amounts may impair the ability of the individual to detect theonset of hypoglycaemia at a stage when they are still able to take appropriate action, i.e.,eat some carbohydrate.• Hypoglycaemia per se may be mistaken for intoxication by observers, with legal andhealth consequences.• Alcohol has been shown in some studies to impact directly on gluconeogenesis and/or thecounterregulatory responses to hypoglycaemia (Turner et al., 2001; Kerr et al., 2007).• Recent data indicate that small amounts of alcohol can augment the cognitive deficitassociated with hypoglycaemia in individuals with type 1 diabetes (Cheyne et al., 2004).

104 MODERATORS, MONITORING AND MANAGEMENTIn the past, biochemical hypoglycaemia (in non-diabetic individuals) associated withalcohol intoxication was attributed to the toxic effects of impurities associated with theproduction of illicit drinks (‘hooch’ or ‘moonshine’). These included methanol, gasoline andethyl acetate (Brown and Harvey, 1941). More recently, it was thought that the biochemicaleffects of alcohol were associated with hypoglycaemia in three ways:• inhibition of gluconeogenesis;• potentiation of the effects of exercise and glucose-lowering agents;• causing reactive hypoglycaemia in susceptible individuals.More than 90% of an ethanol load is metabolised by the liver, being catalysed by alcoholdehydrogenase into acetate. The rate-limiting step is ultimately dependent upon the availabilityof nicotinamide di-nucleotide (NAD+) in a glycogen-replete state. In conditionswhere glycogen stores are depleted and blood glucose is being maintained by hepatic glucoseproduction (e.g. in malnourished individuals and after prolonged exercise), alcohol ingestionmay lead directly to a fall in blood glucose. In contrast, studies have failed to showany short-term effect of alcohol consumed with a meal (Kerr et al., 1990; Avogaro et al.,1993), or when given intravenously after an overnight fast (Kolaczynski et al., 1988). This isprobably because the suppression of gluconeogenesis by ethanol has little effect on hepaticglucose output in well-fed subjects (Gin et al., 1992) and may in fact be counterbalancedby a reduction in peripheral glucose uptake (Koivisto et al., 1993). In well-nourished, nondiabeticsubjects, very little evidence is available to suggest that alcohol has any significanteffect on glucose homeostasis (Trojan et al., 1999).However, alcohol has been shown to suppress lipolysis acutely (Avogaro et al., 1993).Following ingestion of alcohol, a reduction in plasma levels of free fatty acids is associatedwith a reduction in gluconeogenesis and an increased risk of hypoglycaemia in type1 diabetes (Avogaro et al., 1993). In other words, any potential effect of alcohol uponprevailing glucose is likely to be maximal at a time when glucose homeostasis is dependentupon free fatty acid production, e.g. overnight, when lipolysis increases to promote gluconeogenesis(Hagstrom-Toft et al., 1997). The alcohol-induced suppression of lipolysis maythen predispose to hypoglycaemia the next morning (Figures 5.1 and 5.2). The predispositionto delayed hypoglycaemia in type 1 diabetes may be augmented by relative hyperinsulinaemia,which in turn further suppresses lipolysis. However, it is likely that other factorsare involved, as in well-fed subjects the suppression of gluconeogenesis by alcohol per semay have little effect on hepatic glucose output (Gin et al., 1992), and is considered to becounterbalanced by a reduction in peripheral glucose uptake (Koivisto et al., 1993).In a clinical context the main concern for insulin-treated individuals is that moderatealcohol intake (6–9 units) can acutely diminish hypoglycaemia awareness (Moriarty et al.,1993) and impair counterregulatory responses to insulin-induced hypoglycaemia (Yki-Jarvinen and Nikkila, 1985; Pukakainen et al., 1991). Glucagon release has been shown tobe suppressed by alcohol in some studies (Rasmussen et al., 2001), but not in others (Kerret al., 1990). However, the prolonged hypoglycaemic effect of alcohol following its ingestionis more likely to implicate either cortisol or growth hormone responses to hypoglycaemia(Figure 5.3). In animal studies, alcohol has been shown to stimulate corticotrophin-releasinghormone and thus increase plasma cortisol (Rivier et al., 1984). A rise in circulating glucocorticoidcan inhibit growth hormone release (Wehrenberg et al., 1990), which in turn may

LIFESTYLE MODERATORS 10520 22 24 2 4 6 8 10 12 14TimeAdrenaline FFA Growth Hormone GlucoseFigure 5.1 Pictoral showing steady nocturnal and next-day glucose concentrations, a normal rise ingrowth hormone and free fatty acids overnight, and a normal counterregulatory response should therebe a predisposition to hypoglycaemiaAlcoholPotential forhypoglycaemia20 22 24 2 4 6 8 10 12 14TimeAdrenaline FFA GH GlucoseFigure 5.2 Pictoral indicating a slow decline in plasma glucose following evening alcohol ingestion,suppression of free fatty acids and overnight growth hormone release, and the inability to counterregulatean increased predisposition to hypoglycaemia and impaired awareness of hypoglycaemiabe further reduced through the direct inhibition of growth hormone release through the acuteingestion of alcohol (Conway and Mauceri, 1991).In type 1 diabetes, the influence of alcohol ingestion on the immediate counterregulatoryresponses has been investigated by a hyperinsulinaemic clamp study (Kerr et al.,2007). Ingestion of modest amounts of alcohol (to plasma levels of less than 50 mg/dl)

106 MODERATORS, MONITORING AND MANAGEMENT5004003002001000Free insulin (pmol/l)WineWater20 22 24 02 04 06 08 10 12Time (24 h)8007006005004003002001000WineWaterCortisol (nmol/l)20 22 24 02 04 06 08 10 12Time (24 h) hormone (µg/l)Time (24 h)WineWater20 22 24 02 04 06 08 10 129080706050403020100Glucagon (ng/l)Time (24 h)WineWater20 22 24 02 04 06 08 10 12Figure 5.3 Counterregulatory hormone responses following earlier ingestion of alcohol indicatingsuppression of overnight growth hormone secretion. Reproduced from Turner et al., 2001, withpermission from The American Diabetes Associationappears to attenuate the usual growth hormone response to mild hypoglycaemia (Figure 5.4).Other investigators have reported that acute and sustained alcohol ingestion in non-diabeticsubjects can suppress growth hormone release in response to insulin-induced hypoglycaemia(Kolaczynski et al., 1988; Kerr et al., 1990), and also in individuals suffering from reactivehypoglycaemia (Avogaro et al., 1993). In association with blunting of the hormonal counterregulatoryresponses to hypoglycaemia following alcohol ingestion, insulin sensitivity alsoappears to be increased (Ting and Lautt, 2006), thus perhaps directly suppressing hepaticglucose production.Recent studies have also reported blunting of the epinephrine response to hypoglycaemiaafter alcohol (Figure 5.5). The reduction is catecholamine response was mirrored by areduction in hypoglycaemia awareness, which has been described previously with alcohol(Kerr et al., 1990).This blunting of the adreno-medullary response could result from a number of differentmechanisms:• Antecedent nocturnal hypoglycaemia could predispose to delayed hypoglycaemia(Veneman et al., 1993) (i.e., hypoglycaemia leads to more hypoglycaemia).• Sleep patterns are altered by alcohol, which may increase the time spent in ‘deep’ non-REMsleep, thereby increasing the predisposition towards hypoglycaemia through a reductionin the ability to counterregulate (Jones et al., 1998) (see Chapter 4).

LIFESTYLE MODERATORS 107Growthhormone(µg / l)35Growth hormone (mean and SE)Glc = 4.5 Glc = 4.5 / 2.8 Glc = 4.530252015alcoholE + PE + AH + PH + A10500 30 60 90 120 150 180 210 240Time (mins)Figure 5.4 Growth hormone concentrations during the four study conditions (E = euglycaemia,P = placebo, H = hypoglycaemia, A = alcohol and Glc = glucose)1200Plasma adrenaline during hyperinsulinaemic hypoglycaemic clamp1000Adrenaline (ng/ml)8006004002000Baseline euglycemia4.5mmol / lInitial hypoglycaemia2.5mmol / lEnd hypoglycaemia2.5mmol / lEuglycaemia4.5mmol / lAlcoholPlaceboFigure 5.5 Epinephrine (adrenaline) response during a hypoglycaemic clamp following alcoholingested 12 hours previously (unpublished data of authors)

108 MODERATORS, MONITORING AND MANAGEMENT• Other mechanisms, which could account for a failure to counterregulate through an inadequatecatecholamine response, could be explained by a reduction in GH-associated primingof the catecholamine response.The blunted catecholamine response to hypoglycaemia may explain why symptom scoresare lower; this may account for the previously described, alcohol-induced impairment ofhypoglycaemia awareness that is associated with delayed hypoglycaemia following consumptionof alcohol at an earlier time (Kerr et al., 1990).The predisposition of alcohol to cause prolonged, recurrent hypoglycaemia with bluntingof the counterregulatory response, may lead to delayed hypoglycaemia and impaired awarenessof hypoglycaemia. Knowledge of the increased risks and dangers involved with anincreased risk of severe hypoglycaemia is important so that patients can make appropriateadjustments to their lifestyles.Although patients often ask for guidance about alcohol and diabetes, the advice offeredcan be variable and conflicting. For people treated with insulin, it is often recommendedthat dietary carbohydrate should not be omitted and alcohol should be taken with, or shortlybefore, food. Patients should be advised that the risk of hypoglycaemia may extend for‘several hours’ after drinking. It is an important clinical observation that, at blood alcohollevels that remain within the statutory limits for driving in the UK, autonomic and neuroglycopenicwarning symptoms of early hypoglycaemia can be impaired and the cognitivedeficits that are usually associated with mild hypoglycaemia are augmented (Cheyne et al.,2004). As is recommended for all non-diabetic drivers, the advice should be to consume noalcohol if an individual is planning to drive.A further consideration is the ‘morning after the night before’ phenomenon in terms ofhypoglycaemia risk. In a laboratory-based study, Turner et al. (2001) reported that ingestionof alcohol with an evening meal increased the risk of hypoglycaemia the next morning(Figure 5.6). More recently, a study of people with type 1 diabetes during their normal dailylives and using a continuous glucose monitoring system (CGMS), confirmed a predisposition2016WineWaterGlucose(mmol / l)1284018 20 22 24 02 04 06 08 10 12Time (24 h)Figure 5.6 Change in overnight plasma glucose following alcohol or placebo (Turner et al., 2001).The period of drinking is indicated by the shaded bar and the times of symptomatic hypoglycaemia areindicated by the vertical arrows. Reprinted with permission from The American Diabetes Association

LIFESTYLE MODERATORS 109mean change in interstitial glucose (mmol / l) 20:00 22:00 00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00–1.0–2.0–3.0–4.0–5.0period of increasedrisk ofhypoglycaemiaFigure 5.7 Mean difference in interstitial glucose between alcohol and placebo beverages followingstandardised meal at 19:00–20:00to delayed hypoglycaemia (Richardson et al., 2005b). In this study alcohol was shown toreduce average interstitial tissue glucose (Figure 5.7) and increase the risk of hypoglycaemiaover the following 24 hours, with patients reporting more than twice as many hypoglycaemicepisodes per day.In the study of Richardson et al. (2005b), the average interstitial glucose level was 1–2 mmol/l lower with alcohol, but the rate of hypoglycaemia the next day depended on theprevailing levels of glucose overnight. Patients who had a lower baseline glucose at thebeginning of the night or in the morning, were at a greater risk of developing alcohol-inducedhypoglycaemia than those with preceding hyperglycaemia. This increased risk persisted fornearly a full day after alcohol ingestion (Richardson et al., 2005b).CaffeineThe consumption of caffeine has occurred for over 8000 years. Its impact on health –both good and bad – has been widely reported and disputed. Coffee (and caffeine) is themost widely used stimulant in the world and is consumed by more than 50% of Britonsregularly. This amounts to about 75 million cups of coffee consumed every day. Caffeineis also consumed in a variety of different formats such as tea and soft drinks and in ‘overthe counter’ remedies for coughs and colds. During the period from 1960 to 1982, thelevel of consumption of caffeine-containing products rose by 231% (Gilbert, 1984), andin the UK, the average caffeine consumption by adults was estimated to be 400 mg daily(Gilbert, 1984). Children, who do not drink coffee, may consume equivalent amounts insoft drinks.Caffeine exerts a variety of pharmacological actions at diverse sites, both centrallyand peripherally, principally through adenosine receptor antagonism. Although it has beensuggested that the amount of caffeine consumed is related to the risk of developing orprotecting against type 2 diabetes (Pereira et al., 2006), the discussion here is focused onthe influence that caffeine has on the physiological responses to a fall in blood glucose.

110 MODERATORS, MONITORING AND MANAGEMENTMinutes of interstitial hypoglycaemia20016012080400DayplaceboDaycaffeineNightplaceboNightcaffeineFigure 5.8 Diurnal variation in time spent in hypoglycaemic range (interstitial glucose < 35 mmol/l)comparing caffeine and placebo. Error bars indicate the confidence intervals of the meansPrevious studies have shown that ingestion of moderate amounts of caffeine may be usefulby augmenting the symptomatic and hormonal responses to mild hypoglycaemia. Caffeine(3–4 cups of drip-brewed coffee each day) enhances the symptomatic and sympathoadrenalresponses to hypoglycaemia in healthy non-diabetic volunteers and in patients with type 1diabetes (Kerr et al., 1993; Debrah et al., 1996). This may enhance the ability of individualsto perceive the onset of symptoms and take appropriate action by ingesting carbohydratebefore neuroglycopenia develops. The beneficial effect of caffeine on hypoglycaemia risk isindependent of a change in glycaemic control (Watson et al., 2000). Putative mechanismshave included an increase in counterregulatory hormones, including epinephrine, growthhormone and cortisol, but not norepinephrine (Debrah et al., 1996).More recently the influence of caffeine on frequency of hypoglycaemia in patients withlong-standing type 1 diabetes has been studied using CGMS. By using continuous monitoring,caffeine appears to reduce the duration of nocturnal hypoglycaemia in subjects with type 1diabetes by almost 50% (Richardson et al., 2005a) (Figure 5.8). It has been suggested thata beneficial effect of the caffeine-associated reduction in nocturnal hypoglycaemia may beto reduce the risk of developing impaired awareness of hypoglycaemia on the next day. Thecaffeine-associated reduction in ‘antecedent’ nocturnal hypoglycaemia seen in these studiesmay explain the augmentation in the symptomatic and hormonal responses to mild daytimehypoglycaemia described previously (Debrah et al., 1996).The relationship between autonomic dysfunction and the development of impaired awarenessof hypoglycaemia was unclear for many years (see Chapter 7). Most studies havesuggested that peripheral autonomic neuropathy is not associated with an increased riskof severe hypoglycaemia (Polinsky et al., 1980; Bjork et al., 1990; Ryder et al., 1990;The DCCT Research Group, 1991) although central autonomic dysfunction may be important(Evans et al., 2003). Although caffeine improves parasympathetic autonomic function(Richardson et al., 2004), no correlation was found with the observed reduction in nocturnalhypoglycaemia associated with caffeine. As mentioned earlier, it is possible that caffeineuncouples cerebral blood flow and glucose utilisation via antagonism of adenosine receptors(Laurienti et al., 2003), attenuating the glucose supply to the brain (reduced cerebral blood

MONITORING 111flow) while simultaneously increasing glucose demand, thus resulting in relative neuroglycopeniaand earlier release of counterregulatory hormones.Caffeine may also act through an alteration in sleep pattern as adenosine has been implicatedin the physiological regulation of sleep. Caffeine reduces non-rapid eye movement(REM) sleep (Landolt et al., 1995), a stage of sleep that is associated with attenuatedcounterregulatory responses to hypoglycaemia (Jones et al., 1998). Therefore, it couldbe hypothesised that caffeine reduces the time spent in non-REM sleep, lessening theperiod during which counterregulatory responses are suppressed. This may protect againstprolonged hypoglycaemia and may explain the findings of fewer and shorter moderateepisodes of nocturnal hypoglycaemia in people using caffeine. The suggested beneficialeffects of caffeine on nocturnal hypoglycaemia would support the notion that caffeineuse should be encouraged in a population that is prone to the risks of severe nocturnalhypoglycaemia.MONITORINGIt is a sine qua non that the best defence against hypoglycaemia is the ability to recogniseit at an early stage and take appropriate action. Nevertheless, some people lose their abilityto detect hypoglycaemia:• Individuals may fail to develop appropriate warning symptoms.• Individuals may fail to recognise the warning symptoms as being related to hypoglycaemia.• Individuals may recognise the warning symptoms but may be unable to take appropriateaction because of neuroglycopenia.Memory impairment is commonly associated with hypoglycaemia (see Chapter 2) andthus patient recall may underestimate the true frequency of the problem. Furthermore thedefinition of hypoglycaemia needs to be agreed by the patient, their relatives and their healthprofessionals especially for more modest events, which are often ‘accepted’ by patients asan inevitable part of insulin treatment.As a consequence, it is important that patients have additional methods of detecting lowblood glucose levels. Invariably this involves finger stick measurements of blood glucoselevels using glucose meters. However, this method has problems per se related to:• poor technical performance, e.g. inadequate samples;• contamination (actually rare in routine practice);• technical limitations of the devices (Melki et al., 2006).Recently novel methods for measuring glucose levels have been introduced although atpresent none fulfils the ‘holy grail’ of non-invasive glucose monitoring. It is noteworthythat the newer methods of measuring interstitial glucose have highlighted the fact thathypoglycaemia remains a common problem in type 1 diabetes management and that manyepisodes remain unrecognised (Cheyne and Kerr, 2002). These methods can, however, be auseful aid, allowing patients (and their healthcare professionals) to determine the modulating

112 MODERATORS, MONITORING AND MANAGEMENT25.0Sensor Profile – before pump therapy25.0Sensor Profile – after pump therapySensor value (mmol/l) 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00Timesensor value20. 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00TimeFigure 5.9 CGMS profile before and after conversion to CSII in a patient with > 15 years of poorlycontrolled type 1 diabetesGlucose Concentration (mmol/L) 4:00 8:00 12:00 16:00 20:00 24:00TimeFigure 5.10 CGMS profile indicating the extent of postprandial hyperglycaemia. Capillary bloodglucose measurements are shown in square boxesinfluence of a number of factors on glycaemic control and insulin use (Figures 5.9 and 5.10).Unfortunately they have a number of limitations, as detailed in the next section.Continuous Glucose Monitoring Systems (CGMS)There are limitations to the use of traditional blood glucose measurements for detectinglow blood glucose levels. Consequently, a great deal of time, effort and money has beenspent on developing new methods for the detection of low blood glucose levels. The mostpopular system at present is the use of a continuous glucose monitoring system (CGMS)(Hoi-Hansen et al., 2005). However, technical considerations influence the use of interstitialglucose monitoring for the detection of hypoglycaemia, specifically:• There is a physiological lag between changes in blood and interstitial glucose levels whichsuggest that CGMS may overestimate the duration of a hypoglycaemic event.• It is unclear as to the significance of changes at the level of interstitial fluid compared tofluctuations in blood glucose level.

MONITORING 113• It is unclear whether changes in interstitial glucose levels, measured in the anteriorabdominal wall, mirror those occurring at the level of glucose sensing neurones in thehypothalamus and elsewhere.• The recordings from the CGMS are markedly influenced by the accuracy and frequencyof the results of fingerstick blood testing for calibration purposes.The first commercially available device, the CGMS (Medtronic Minimed, Minneapolis,USA), was designed to monitor glucose levels continuously in tissue fluid. The systemcomprised a disposable subcutaneous glucose sensor connected by a cable to a pager-sizedglucose monitor. The sensor measured interstitial glucose levels every 10 seconds andaveraged over 5 minutes, i.e., 288 measurements each day, and provided measurementsof glucose in the range of 2.2–22 mmol/l. The first version provided retrospective data.The latest generation sensor transfers data to an on-screen display in real-time for thepatient to adjust his or her own medication (Bode et al., 2004). One problem has been indefining hypoglycaemia using these devices, although some groups have defined interstitialhypoglycaemia according to the level at which there is activation of the counterregulatoryhormone cascade and onset of neuroglycopenic symptoms respectively (Fanelli et al., 1994)(Table 5.2).There are also other limitations (including expense) in the use of continuous interstitialmonitoring devices. Continuous glucose monitoring is a technique in its infancy and as suchthere is debate whether it has the ability to detect ‘true’ hypoglycaemia (McGowan et al.,2002). This uncertainty relates to a number of factors including:Table 5.2 Guidelines for interpreting interstitial hypoglycaemia using CGMS as developedby the UK Hypoglycaemia Study Group (2007), and published in Department forTransport Road Safety Research Report No. 61, “Stratifying hypoglycemic event risk ininsulin-treated diabetes” (2006), pp. 68–9Defining an episode of hypoglycaemiaThere must be four consecutive readings of 3.5 mmol/l or lower for an episode to beclassified as hypoglycaemia.The first of the readings (of 3.5 mmol/l or less) signifies the start of hypoglycaemia.The first reading of 3.5 mmol/l or above signifies the end of hypoglycaemia.Hypoglycaemia ends fully when there are four or more consecutive readings above3.5 mmol/l.If during hypoglycaemia the sensor value is 3.5 mmol/l or higher for 1–3 readings andthen goes back down below 3.5 mmol/l, then the whole episode is counted as onehypoglycaemic event and the short period above 3.5 mmol/l is included in theduration of hypoglycaemia.To be labelled as moderate hypoglycaemia, the sensor has to read 3.0 mmol/l orbelow for at least four consecutive readings.If during mild hypoglycaemia the reading falls to 3.0 mmol/l or below for four ormore consecutive readings, the whole hypoglycaemic episode will be considered asmoderate hypoglycaemia.Prolonged hypoglycaemia is defined as lasting > 2 hours.

114 MODERATORS, MONITORING AND MANAGEMENT6.00Sensor value compared to blood glucoseGlucose (mmol/l) averageaverage BG0.000:00 1:00 2:00 3:00 4:00Time (hours)Figure 5.11 Comparison between interstitial (CGMS) and plasma glucose under hyperinsulinaemichypoglycaemic clamp conditions• the potential error in recording glucose values at the limit of the detection range;• the potential for artefact (a flat line might indicate a low glucose value or a technicalproblem with the recording);• differences between blood and extracellular interstitial glucose.Continuous glucose monitoring methodology does not directly sample blood glucosebut records glucose values in the extracellular interstitial space. Glucose diffuses acrossthe capillary wall into the interstitial space before being transported into cells where it ismetabolised or stored. The relationship between blood and interstitial glucose is not wellunderstood and there are physiological and pharmacological reasons why the two may differ.Laboratory studies have demonstrated a variable relationship between blood and interstitialglucose concentrations according to whether blood glucose is falling or rising, and circulatinginsulin concentrations may also be important (Caplin et al., 2003).Finally, it is important to note that the use of CGMS is best used to corroborate themore traditional method of recording hypoglycaemia events, which is relying on patients’self-reports. Even taking into account the uncertainty of which CGMS value representstrue hypoglycaemia, rates of low glucose are comparable whether measured by prospectivecollection of self-reported episodes or by continuous glucose monitoring.Although there is increasing experience with the CGMS system for detection of hypoglycaemia,the relationship between interstitial glucose levels measured from the anterior abdominalwall and cerebral interstitial levels is unknown. Furthermore, in non-diabetic individuals,the CGMS may overestimate the duration of hypoglycaemia as there appears to be a timelag between sensor-measured interstitial tissue glucose and peripheral blood glucose levelsduring recovery from hypoglycaemia (Cheyne et al., 2002) (Figure 5.11). Overestimation ofnocturnal hypoglycaemia in patients with strictly controlled type 1 diabetes has also beenreported (McGowan et al., 2002). However, in more customary populations where capillaryglucose sensor calibration was more widely dispersed, accurate prediction of hypoglycaemiawith the CGMS has been demonstrated (McGowan et al., 2002; Caplin et al., 2003).

MANAGEMENT OF HYPOGLYCAEMIA 115MANAGEMENT OF HYPOGLYCAEMIAIt goes without saying that prevention is better than cure and patients with diabetes needto be thoroughly and continuously educated about the potential risks of hypoglycaemia.Diabetes UK recommends a policy of 4 is the floor as a capillary blood glucose level forintervention with regard to treating hypoglycaemia. The treatment of hypoglycaemia is bestregarded as a spectrum of increasing intervention with at one end the simple ingestion of oralcarbohydrate and at the other, acute medical therapy in an intensive care unit (Table 5.3).The simplest treatment, when the patient recognises the early warning symptoms (seeChapter 2), is to eat carbohydrate, which must be palatable, concentrated and portable.Glucose tablets (Dextrosol) are usually recommended in the UK, barley sugar in the USAand, in France, lumps of sugar (sucrose). Beverages such as soft drinks or orange juicewith a high glucose content are also suitable. The important factor is that short-actingcarbohydrate should be followed by some form of longer-acting carbohydrate such as bread orbiscuits.The second level of treatment is when the patient is clearly hypoglycaemic but cannotor will not take oral fast-acting carbohydrate. People on insulin will often not admit tobeing hypoglycaemic and may react adversely to attempts to give them carbohydrate. Liquidglucose solutions are often unsatisfactory because the patient can spit them out. It is betterto use a commercially-available glucose gel such as GlucoGel (Diabetic Bio-diagnostics),which can be squeezed like toothpaste into the mouth and is absorbed through the buccalmucosa. Although some doubts have been expressed about the effectiveness of oral glucosegels, relatives often seem to prefer to try this before resorting to injecting glucagon. It shouldnot be used in semi-comatose patients or those at risk of aspirating. Jam or honey may bejust as effective.Glucagon promotes hepatic glycogenolysis and the glycaemic response to a dose of1 mg is essentially the same whether it is injected subcutaneously, intramuscularly or intravenously(Muhlhauser et al., 1985a). The advantage of glucagon is that it can be givenTable 5.3 The therapeutic spectrum of hypoglycaemia; complexity of treatment depends primarilyon duration (adapted from MacCuish, 1993)Duration of hypoglycaemiaInitial Management(minutes)By patient By family By paramedicsIn hospital A&EdepartmentOngoingmanagement (hours)In intensive careOral carbohydrate(> 20 g)Oral carbohydrate(liquid/solid)Glucagon 1mgim or iv25 g glucose iv Mannitol (20%,20 ml)Glucagon 1 mg iv DexamethasoneGlucose gel 25 g glucose iv Dextrose/insulininfusionGlucagon 1 mg imOxygenAnticonvulsantsSedation

116 MODERATORS, MONITORING AND MANAGEMENTby relatives or friends after minimal training. Paramedics can also use it at the patient’shome or in an ambulance. The disadvantages are that it takes longer (approximately 10minutes) than intravenous glucose to restore consciousness and does not work in patientswho have deficient or absent hepatic glycogen stores (alcoholics or people with cachexia).Unfortunately, even where glucagon is available, relatives or friends may not use it. Inone study (Muhlhauser et al., 1985b), 53 of 123 episodes of severe hypoglycaemia weretreated by relatives or friends with glucagon, 30 by assisting physicians and 44 requiredhospital admission. When glucagon was available but not used, it was because those whoknew how to use it were not present (20 cases) or were too anxious to do so (24 cases).In children with diabetes, Daneman et al. (1989) found that glucagon was used in only athird of households in which it was available – presumably because relatives were eithertoo frightened or poorly educated. Another problem is the limited shelf life; Ward et al.(1990) found that nearly three-quarters of patients knew about glucagon but only 20% hada supply that was in date. Its effect is less certain if coma has been prolonged. In a studyof 100 patients brought to a hospital outpatient department, glucagon was immediatelyeffective in only 41% of patients whose mean estimated duration of coma was 50 minutes(MacCuish et al., 1970).Within five minutes the traditional dose of 50 ml of a 50% glucose (dextrose) solutionraises blood glucose from below 1 to > 12 mmol/l (Collier et al., 1987). This dose isnow considered to be too large by using too concentrated a solution, and 20% dextrosewill suffice. The main problem is the difficulty of giving an intravenous injection toan uncooperative patient who may be refusing or resisting treatment or who has to bephysically restrained to receive an intravenous injection. Extravasation of the concentratedglucose solution outside the vein causes painful phlebitis and, because of its hypertonicity,even an intravascular injection can cause phlebitis or thrombosis. It is not thereforeused much outside hospitals, and few GPs carry glass vials of dextrose for thispurpose.When a patient known to have type 1 diabetes is admitted to hospital in hypoglycaemiccoma, but fails to recover after being given intravenous glucose, other causes ofcoma such as excessive ingestion of alcohol, self-poisoning with opiates or other drugs,acute vascular events such as subarachnoid haemorrhage or stroke, head injury or otherintracranial catastrophes must be excluded. Cerebral oedema is a recognised sequel ofsevere hypoglycaemia, and urgent neuroimaging is required to establish its presence; treatment(Table 5.3) is usually in an intensive care unit as this complication has a highmortality.The most difficult decision is to know for how long to continue treatment. Patients,who have made a full recovery after being unconscious for several days, may (anecdotally)appear subsequently to have significant cognitive impairment or permanent brain damage.It is beyond the scope of this chapter to discuss the investigation and management ofhypoglycaemia in non-diabetic individuals.In conclusion, hypoglycaemia continues to be a common problem in the management ofindividuals with type 1 diabetes. The use of newer technologies of continuous glucose monitoringhas highlighted that it is almost impossible to eliminate hypoglycaemia completelywith present insulin therapy, although understanding moderating factors such as alcohol andincluding them as a component of education programmes for people with insulin-treateddiabetes may help to alleviate some of the anxiety associated with the risk of living constantlywith the threat of hypoglycaemia.

REFERENCES 117CONCLUSIONS• Hypoglycaemia is associated with multiple risk factors.• Alcohol is a frequent cause of severe hypoglycaemia.• Alcohol is associated with delayed hypoglycaemia through impaired hypoglycaemiaawareness and an abnormal counterregulatory response.• Understanding of the nature of alcohol-associated hypoglycaemia is the first step toreducing the risks associated with alcohol.• Caffeine may augment the symptomatic and hormonal responses to hypoglycaemia andreduce nocturnal hypoglycaemia in individuals with type 1 diabetes.• Continuous Glucose Monitoring is a useful adjunct to the management of the patient withtype 1 diabetes.• Treatment of hypoglycaemia can be regarded as a spectrum of increasing therapeuticcomplexity depending on the severity of the hypoglycaemia and the clinical status of thepatient.REFERENCESAvogaro A, Beltramello P, Gnudi L, Maran A, Valerio A, Miola M et al. (1993). Alcohol intakeimpairs glucose counterregulation during acute insulin-induced hypoglycemia in IDDM patients.Evidence for a critical role of free fatty acids. Diabetes 42: 1626–34.Bjork E, Palmer M, Schvarcz E, Berne C (1990). Incidence of severe hypoglycaemia in an unselectedpopulation of patients with insulin-treated diabetes mellitus, with special reference to autonomicneuropathy. Diabetes Nutrition and Metabolism 4: 303–9.Bode B, Gross K, Rikalo N, Schwartz S, Wahl T, Page C et al. (2004). Alarms based on real-timesensor glucose values alert patients to hypo- and hyperglycemia: the guardian continuous monitoringsystem. Diabetes Technology and Therapeutics 6: 105–13.Bott S, Bott U, Berger M, Muhlhauser I (1997). Intensified insulin therapy and the risk of severehypoglycaemia. Diabetologia 40: 926–32.Brown TM, Harvey AM (1941). Spontaneous hypoglycemia in ‘smoke’ drinkers. Journal of theAmerican Medical Association 17: 12–5.Caplin NJ, O’Leary P, Bulsara M, Davis EA, Jones TW (2003). Subcutaneous glucose sensor valuesclosely parallel blood glucose during insulin-induced hypoglycaemia. Diabetic Medicine 20: 238–41.Cheyne EH, Cavan DA, Kerr D (2002). Performance of a continuous glucose monitoring system duringcontrolled hypoglycemia in healthy volunteers. Diabetes Technology and Therapeutics 4: 607–13.Cheyne E, Kerr D (2002). Making ‘sense’ of diabetes: using a continuous glucose sensor in clinicalpractice. Diabetes Metabolism Research and Reviews 18 Suppl 1: S43–8.Cheyne EH, Sherwin RS, Lunt MJ, Cavan DA, Thomas PW, Kerr D (2004). Influence of alcoholon cognitive performance during mild hypoglycaemia; implications for type 1 diabetes. DiabeticMedicine 21: 230–7.Collier A, Steedman DJ, Patrick AW, Nimmo GR, Matthews DM, MacIntyre CA et al. (1987).Comparison of intravenous glucagon and dextrose in treatment of severe hypoglycemia in anAccident and Emergency Department. Diabetes Care 10: 712–5.Conway S, Mauceri H (1991). The influence of acute ethanol exposure on growth hormone release infemale rats. Alcohol 8: 159–64.

118 MODERATORS, MONITORING AND MANAGEMENTCox DJ, Gonder-Frederick L, Julian DM, Clarke W (1994). Long-term follow-up of blood glucoseawareness training. Diabetes Care 17: 1–5.Daneman D, Frank M, Perlman K, Tamm J, Ehrlich R (1989). Severe hypoglycemia in children withinsulin-dependent diabetes mellitus: frequency and predisposing factors. Journal of Pediatrics 115:681–5.Debrah K, Sherwin RS, Murphy J, Kerr D (1996). Effect of caffeine on recognition of and physiologicalresponses to hypoglycaemia in insulin-dependent diabetes. Lancet 347: 19–24.Evans SB, Wilkinson CW, Gronbeck P, Bennett JL, Taborsky GJ Jr, Figlewicz DP (2003). Inactivationof the PVN during hypoglycemia partially simulates hypoglycemia-associated autonomic failure.American Journal of Physiology. 284: R57–65.Fanelli C, Pampanelli S, Epifano L, Rambotti AM, Ciofetta M, Moda F et al. (1994). Relative rolesof insulin and hypoglycaemia on induction of neuroendocrine responses to, symptoms of, anddeterioration of cognitive function in hypoglycaemia in male and female humans. Diabetologia 37:797–807.Feldman M, Schiller LR (1983). Disorders of gastrointestinal motility associated with diabetes mellitus.Annals of Internal Medicine 98: 378–84.Gilbert, RM (1984). Caffeine consumption. In: The Methylxanthines Beverages and Foods: Chemistry,Consumption and Health Effects. Spiller GA, ed. Alan R Liss, New York: 235–301.Gin H, Morlat P, Ragnaud JM, Aubertin J (1992). Short-term effect of red wine (consumed duringmeals) on insulin requirement and glucose tolerance in diabetic patients. Diabetes Care 15: 546–8.Gold AE, MacLeod KM, Frier BM (1994). Frequency of severe hypoglycemia in patients with type Idiabetes with impaired awareness of hypoglycemia. Diabetes Care 17: 697–703.Hagstrom-Toft E, Bolinder J, Ungerstedt U, Arner P (1997). A circadian rhythm in lipid mobilizationwhich is altered in IDDM. Diabetologia 40: 1070–8.Hoi-Hansen T, Pedersen-Bjergaard U, Thorsteinsson B (2005). Reproducibility and reliability ofhypoglycaemic episodes recorded with continuous glucose monitoring system (CGMS) in daily life.Diabetic Medicine 7: 858–62.Jones TW, Porter P, Sherwin RS, Davis EA, O’Leary P, Frazer F et al. (1998). Decreased epinephrineresponses to hypoglycemia during sleep. New England Journal of Medicine 338: 1657–62.Kerr D, Macdonald IA, Heller SR, Tattersall RB (1990). Alcohol causes hypoglycaemic unawareness inhealthy volunteers and patients with Type 1 (insulin-dependent) diabetes. Diabetologia 33: 216–21.Kerr D, Sherwin RS, Pavalkis F, Fayad PB, Sikorski L, Rife F et al. (1993). Effect of caffeine onthe recognition of and responses to hypoglycemia in humans. Annals of Internal Medicine 119:799–804.Kerr D, Cheyne EH, Thomas P, Sherwin RS (2007). Influence of acute alcohol ingestion on thehormonal responses to modest hypoglycaemia in patients with Type 1 diabetes. Diabetic Medicine24: 312–16.Kolaczynski JW, Ylikahri R, Harkonen M, Koivisto VA (1988). The acute effect of ethanol oncounterregulatory response and recovery from insulin-induced hypoglycemia. Journal of ClinicalEndocrinology and Metabolism 67: 384–8.Koivisto VA, Tulokas S, Toivonen M, Haapa E, Pelkonen R (1993). Alcohol with a meal has noadverse effects on postprandial glucose homeostasis in diabetic patients. Diabetes Care 16: 1612–4.Landolt HP, Werth E, Borbely AA, Dijk DJ (1995). Caffeine intake (200 mg) in the morning affectshuman sleep and EEG power spectra at night. Brain Research 675: 67–74.Laurienti PJ, Field AS, Burdette JH, Maldjian JA, Yen YF, Moody DM (2003). Relationship betweencaffeine-induced changes in resting cerebral perfusion and blood oxygenation level-dependent signal.American Journal of Neuroradiology 24: 1607–11.MacCuish AC, Munro JF, Duncan LJP (1970). Treatment of hypoglycaemic coma with glucagon,intravenous dextrose, and mannitol infusion in a hundred diabetics. Lancet ii: 946–9.MacCuish AC (1993). Treatment of hypoglycaemia. In: Hypoglycaemia and Diabetes: Clinical andPhysiological Aspects. Frier BM and Fisher M, eds. Edward Arnold, London: 212–21.

REFERENCES 119MacDonald MJ (1987). Postexercise late-onset hypoglycemia in insulin-dependent diabetic patients.Diabetes Care 10: 584–8.McGowan K, Thomas W, Moran A (2002). Spurious reporting of nocturnal hypoglycemia by CGMSin patients with tightly controlled type 1 diabetes. Diabetes Care 25: 1499–503.Melki V, Ayon F, Fernandez M, Hanaire-Broutin H (2006). Value and limitations of the ContinuousGlucose Monitoring System in the management of type 1 diabetes. Diabetes and Metabolism 32:123–9.Moriarty KT, Maggs DG, Macdonald IA, Tattersall RB (1993). Does ethanol cause hypoglycaemia inovernight fasted patients with type 1 diabetes? Diabetic Medicine 10: 61–5.Muhlhauser I, Koch J, Berger M (1985a). Pharmacokinetics and bioavailability of injected glucagon:differences between intramuscular, subcutaneous, and intravenous administration. Diabetes Care 8:39–42.Muhlhauser I, Berger M, Sonnenberg G, Koch J, Jorgens V, Schernthaner G et al. (1985b). Incidenceand management of severe hypoglycemia in 434 adults with insulin-dependent diabetes mellitus.Diabetes Care 8: 268–73.Muhlhauser I, Toth G, Sawicki PT, Berger M (1991). Severe hypoglycemia in type 1 diabetic patientswith impaired kidney function. Diabetes Care 14: 344–46.Nilsson A, Tideholm B, Kalen J, Katzman P (1988). Incidence of severe hypoglycaemia and its causesin insulin-treated diabetics. Acta Medica Scandinavica 224: 257–2.Pereira MA, Parker ED, Folsom AR (2006). Coffee consumption and risk of type 2 diabetes mellitus:an 11-year prospective study of 28 812 postmenopausal women. Archives of Internal Medicine 26:1311–6.Polinsky RJ, Kopin IJ, Ebert MH, Weise V (1980). The adrenal medullary response to hypoglycemiain patients with orthostatic hypotension. Journal of Clinical Endocrinology and Metabolism 51:1401–6.Potter J, Clarke P, Gale EA, Dave SH, Tattersall RB (1982). Insulin-induced hypoglycaemia inan Accident and Emergency Department: the tip of an iceberg? British Medical Journal 285:1180–2.Pramming S, Thorsteinsson B, Bendtson I, Binder C (1990). The relationship between symptomaticand biochemical hypoglycaemia in insulin-dependent diabetic patients. Journal of Internal Medicine228: 641–6.Pukakainen I, Koivisto VA, Yki-Jarvinen H (1991). No reduction in total hepatic glucose output byinhibition of gluconeogenesis with ethanol in NIDDM patients. Diabetes 40: 1319–27.Rasmussen BM, Orskov L, Schmitz O, Hermansen K (2001). Alcohol and glucose counterregulationduring acute insulin-induced hypoglycemia in type 2 diabetic subjects. Metabolism 50:451–7.Richardson T, Rozkovec A, Thomas P, Ryder J, Meckes C, Kerr D (2004). Influence of caffeine onheart rate variability in patients with long-standing type 1 diabetes. Diabetes Care 27: 1127–31.Richardson T, Thomas P, Ryder J, Kerr D (2005a). Influence of caffeine on frequency of hypoglycemiadetected by continuous interstitial glucose monitoring system in patients with long-standing type 1diabetes. Diabetes Care 28: 1316–20.Richardson T, Weiss M, Thomas P, Kerr D (2005b). Day after the night before: influence ofevening alcohol on risk of hypoglycemia in patients with type 1 diabetes. Diabetes Care 28:1801–2.Rivier C, Bruhn T, Vale W (1984). Effect of ethanol on the hypothalamic-pituitary-adrenal axis inthe rat: role of corticotropin-releasing factor (CRF). Journal of Pharmacology and ExperimentalTherapy 229: 127–31.Rosen B, Miodovnik M, Holcberg G, Khoury J, Siddiqi T (1995). Hypoglycaemia: the price ofintensive insulin therapy for pregnant women with insulin dependent diabetes mellitus. Obstetricsand Gynaecology 85: 417–22.

120 MODERATORS, MONITORING AND MANAGEMENTRyder RE, Owens DR, Hayes TM, Ghatei MA, Bloom SR (1990). Unawareness of hypoglycaemia andinadequate hypoglycaemic counterregulation: no causal relation with diabetic autonomic neuropathy.British Medical Journal 301: 783–7.The DCCT Research Group (1991). Epidemiology of severe hypoglycemia in the Diabetes Controland Complications Trial. American Journal of Medicine 90: 450–9.The Diabetes Control and Complications Trial Research Group (1995). Adverse events and theirassociation with treatment regimens in the Diabetes Control and Complications Trial. Diabetes Care18: 1415–27.The Diabetes Control and Complications Trial Research Group (1997). Hypoglycemia in the DiabetesControl and Complications Trial. The Diabetes Control and Complications Trial Research Group.Diabetes 46: 271–86.Ting JW, Lautt WW (2006). The effect of acute, chronic, and prenatal ethanol exposure on insulinsensitivity. Pharmacology and Therapeutics 111: 346–73.Trojan N, Pavan P, Iori E, Vettore M, Marescotti MC, Macdonald IA et al. (1999). Effect of differenttimes of administration of a single ethanol dose on insulin action, insulin secretion and redox state.Diabetic Medicine 16: 400–7.Turner BC, Jenkins E, Kerr D, Sherwin RS, Cavan DA (2001). The effect of evening alcohol consumptionon next-morning glucose control in type 1 diabetes. Diabetes Care 24: 1888–93.UK Hypoglycaemia Group (2007). Risk of hypoglycaemia in types 1 and 2 diabetes: effects oftreatment modalities and their duration. Diabetologia 50: 1140–7.Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J (1993). Induction of hypoglycemia unawarenessby asymptomatic nocturnal hypoglycemia. Diabetes 42: 1233–7.Vinik AI, Maser RE, Mitchell BD, Freeman R (2003). Diabetic autonomic neuropathy. Diabetes Care26: 1553–79.Ward CM, Stewart AW, Cutfield RG (1990). Hypoglycaemia in insulin dependent diabetic patientsattending an outpatients’ clinic. New Zealand Medical Journal 25: 339–41.Watson JM, Jenkins EJ, Hamilton P, Lunt MJ, Kerr D (2000). Influence of caffeine on the frequencyand perception of hypoglycemia in free-living patients with type 1 diabetes. Diabetes Care 23:455–9.Wehrenberg WB, Janowski BA, Piering AW, Culler F, Jones KL (1990). Glucocorticoids: potentinhibitors and stimulators of growth hormone secretion. Endocrinology 126: 3200–3.Yki-Jarvinen H, Nikkila EA (1985). Ethanol decreases glucose utilisation in healthy man. Journal ofClinical Endocrinology and Metabolism 61: 941–5.

6 Counterregulatory Deficienciesin DiabetesDavid Kerr and Tristan RichardsonINTRODUCTIONModern intensive education programmes for people with insulin-treated diabetes have failedto eliminate hypoglycaemia, and this important side-effect of insulin treatment may add tothe psychological, as well as the physical, burden associated with this condition. Attempts toimprove blood glucose control can increase the risk of severe hypoglycaemia. For example,within the Diabetes Control and Complications Trial (DCCT), improvements in HbA 1c levelswith intensive insulin therapy were associated with a three-fold increase in the risk of severehypoglycaemia compared to individuals treated conventionally (The Diabetes Control andComplications Trial Research Group,1993). Despite the development of strategies to improveglycaemic control, while simultaneously trying to reduce the risk of recurrent severe hypoglycaemia,less than half of people with type 1 diabetes consistently achieve HbA 1c levels < 75%(Jacqueminet et al., 2005), and for some individuals, fear of hypoglycaemia is the main barrierto achieving optimal glycaemic control (Cox et al., 1987) (see Chapter 14).The use of continuous glucose monitoring has indicated that the frequency and duration ofhypoglycaemic events in patients with type 1 diabetes have probably been under-recognised(Figure 6.1) (Cheyne and Kerr, 2002), especially in children (Weintrob et al., 2004). In theUK in recent years, a number of structured education programmes have been developed, mostof which are based on a German model (Muhlhauser et al., 1987), and have been focusedon empowering patients to alter insulin doses more accurately according to the carbohydratecontent of meals, the level of planned exercise, work demands and so on (DAFNE StudyGroup, 2002). Disappointingly, despite increasing numbers of patients having access to theseeducational programmes, along with the increased availability of analogue insulins, and withthe evidence that these measures are reducing the number of patients who have microvascularcomplications (Figure 6.2), the rates of severe hypoglycaemia have not altered significantly(Figure 6.3) (Bulsara et al., 2004).Why is hypoglycaemia still such an important problem for people with type 1 diabetesdespite these significant improvements in technology and the delivery of care? The answerfrequently relates to the problem of defective glucose counterregulation and the often associateddifficulty of recognising a fall in blood glucose (impaired awareness of hypoglycaemia –see Chapter 7) at a time when appropriate self-corrective action can be taken.Hypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

122 COUNTERREGULATORY DEFICIENCIES IN DIABETES20.0Glucose (mmol/L)–5.012:00AM4:00AM8:00AM12:00PM4:00PM8:00PM12:00AMFigure 6.1 Interstitial glucose values over three days measured by continuous glucose monitoringand showing persistent nocturnal hypoglycaemia50%40%RetinopathyKidney disease30%20%10%0%25 yearsFigure 6.2 Reduction in prevalence of microvascular complications over the past 25 years fromspecialist diabetes centres. Adapted from Rossing (2005), with kind permission from Springer Scienceand Business MediaNORMAL GLUCOSE COUNTERREGULATIONUnder normal circumstances, the brain uses glucose as its predominant fuel and is almostcompletely dependent upon a continuous supply of glucose from the peripheral circulationto maintain normal function. Consequently, to protect the delivery of glucose to the brain,a hierarchy of (counterregulatory) responses (Figure 6.4) are activated as peripheral bloodglucose falls below normal (Mitrakou et al., 1991). The main components of this systemof normal (i.e. non-diabetic) glucose counterregulation, which prevents or quickly correctshypoglycaemia, are as follows:• A reduction in pancreatic -cell insulin secretion.• An increase in pancreatic -cell glucagon secretion. If hypoglycaemia is prolonged anumber of other hormones are also released. These include epinephrine (adrenaline),growth hormone and cortisol.

NORMAL GLUCOSE COUNTERREGULATION 123A22Rate / 100 Patient years201816141210864Incidence rate21992 1994 1996Calender Year1998 20002002Figure 6.3 Rates of severe hypoglycaemia in 1335 children between 1992 and 2002. Hypoglycaemiahas remained problematical despite a fall in HbA 1c of 0.2% per year and increased use of multiple dailyinjections, insulin analogues and pump therapy. Adapted from Bulsara et al. (2004) and reproducedcourtesy of The American Diabetes Association• Activation of the autonomic nervous system with the development of characteristicwarning symptoms.• During more profound hypoglycaemia (blood glucose < 20 mmol/l), glucose delivery tothe brain is enhanced as a consequence of increasing cerebral blood flow (Thomas et al.,1997).• Hepatic autoregulation, whereby the liver is able to respond to hypoglycaemia by directlyincreasing glucose production in the absence of detectable hormonal stimulation. Hepaticautoregulation only appears to have an influential role in glucose counterregulation duringprolonged and severe hypoglycaemia (Tappy et al., 1999).The principal counterregulatory hormones, glucagon and epinephrine, directly increase theproduction of glucose by the liver as a consequence of the breakdown of hepatic glycogenstores (glycogenolysis) and the manufacture of glucose by gluconeogenesis. Epinephrinealso promotes muscle glycogenolysis, proteolysis and lipolysis to provide substrates (lactate,alanine and glycerol) for further gluconeogenesis (Figure 6.5).As blood glucose falls below normal, these hormonal responses are not secreted on an ‘allor nothing’ basis. Individual hormones have specific blood glucose thresholds at which levelsbegin to rise above their baseline levels (Figure 6.6). For example, the glycaemic thresholdsfor the release of glucagon and epinephrine are well above the thresholds for the generationof warning symptoms and impairment of higher brain (cognitive) function (Mitrakou et al.,1991). These thresholds are not fixed but can be altered upwards or downwards according to

124 COUNTERREGULATORY DEFICIENCIES IN DIABETESHypothalamusPituitary glandGrowth hormoneAutonomic Nervous SystemACTHPancreasAdrenalglandCortisolGlucagonNorepinephrineEpinephrineFigure 6.4The brain is the key regulatory organ involved in glucose counterregulationthe prevailing quality of glycaemic control. They do not appear, however, to be influencedby the rate of fall of blood glucose within the hyper- or euglycaemic range.The key organ in coordinating the hormonal and other responses to hypoglycaemia isthe brain. Although numerous neural areas have been proposed as the control centres forcounterregulation, it is likely that neurones located within the ventro-medial nuclei of thehypothalamus are essential for integrating the hormonal responses to a fall in peripheralblood glucose (Borg et al., 1994), possibly via ATP-sensitive K+ channels (McCrimmonet al., 2005), although other glucoreceptors outside the brain are involved in initiating thecounterregulatory responses, most notably within the liver (Smith et al., 2002). Althoughthe brain was once thought to be insensitive to insulin, there is clinical evidence to suggestthat insulin can act on the central nervous system (CNS) to influence the physiologicalresponses to hypoglycaemia (Kerr et al., 1991). Recently, in studies using brain/neuronalinsulin receptor knockout mice (i.e., mice with absent insulin receptor proteins in the brain),the induction of hypoglycaemia was associated with an attenuated epinephrine and an almostcompletely absent norepinephrine response, although glucagon release was unaffected whencompared to control animals. Therefore it appears that insulin has a role in protecting theCNS against hypoglycaemia (Fisher et al., 2005).

NORMAL GLUCOSE COUNTERREGULATION 125Figure 6.5Principal metabolic effects of counterregulation in response to hypoglycaemiaGlucose (mmol / l)4.0Release of hormonesWarning symptoms3.0Cognitive impairmentIncreased brain blood flow2.0Figure 6.6 Blood glucose thresholds for release of counterregulatory hormones, onset of warningsymptoms of hypoglycaemia and cognitive impairment as blood glucose falls below normal

126 COUNTERREGULATORY DEFICIENCIES IN DIABETESIn type 1 diabetes there may be multiple defects in normal glucose counterregulationwhich significantly increase an individual’s risk of hypoglycaemia. Defects in hormonalcounterregulation can result as a consequence of the following:• secretion of inadequate amounts of counterregulatory hormones;• alteration of the blood glucose threshold at which the hormones are released; i.e., a moreprofound hypoglycaemic stimulus is required before hormonal secretion increases abovebaseline;• diminished tissue sensitivity to a given plasma concentration of hormone.In type 1 diabetes, the most common scenario for increasing the risk of recurrent hypoglycaemiaincludes a reduced (but not absent) epinephrine response to falling blood glucoselevels, sustained peripheral hyperinsulinaemia and an absent glucagon response – this constellationof defects is associated with a 25-fold increased risk of severe hypoglycaemia inpeople who are on intensive insulin therapy. If impaired hypoglycaemia awareness is present,the risk is increased six fold (see Chapter 7).DEFECTIVE HORMONAL GLUCOSE COUNTERREGULATIONThe most important defects in glucose counterregulation in type 1 diabetes are:• failure of circulating plasma insulin levels to decline (i.e., as a consequence of exogenousinsulin administration);• failure of glucagon secretion from pancreatic -cells;• attenuated epinephrine response to hypoglycaemia.Several factors are known to increase the risk of counterregulatory failure, including:• long duration of diabetes;• extremes of age;• improving glycaemic control;• sleep;• exercise;• recurrent hypoglycaemia – related to the degree rather than duration of antecedent hypoglycaemia(Davis et al., 2000a).The defects in glucose counterregulation are not ‘all or nothing’ and are influenced by anumber of factors. Some of the defects may be reversible.

DEFECTIVE HORMONAL GLUCOSE COUNTERREGULATION 127GlucagonIn non-diabetic individuals, glucagon is secreted by pancreatic -cells within 30 minutes ofthe blood glucose falling below normal. It is released directly as a consequence of local tissueglucopenia and indirectly by sympathetic neural inputs to the pancreas. It is unclear whetherlocal (pancreatic) tissue glucopenia or autonomic activation is the key process involved instimulating the release of glucagon during hypoglycaemia.In people with type 1 diabetes, the glucagon secretory response to hypoglycaemia isinitially diminished and is subsequently lost within a few years of the onset of diabetes(Gerich et al., 1973), although it can be released in response to other stimuli, such as exerciseor an intravenous infusion of arginine (Gerich and Bolli, 1993). This indicates that the failureof the glucagon response is most probably a signalling rather than a structural defect. Loss ofthe glucagon response to hypoglycaemia often coexists with clinical evidence of autonomicneuropathy but the latter is not invariably present (Bolli et al. 1983). Although recent studieshave suggested that an early sympathetic neuropathy limited to the pancreatic islets may bea potential mechanism, patients with pancreatic transplants (i.e., denervated islets) can stillproduce glucagon in response to hypoglycaemia (Diem et al., 1990). The cause of the defectin glucose counterregulation is not known but may include:• reduction in pancreatic -cell mass;• autonomic neuropathy;• local effect of insulin on normal -cell function;• generalised hyperinsulinaemia;• chronic hyperglycaemia;• increased pancreatic production of somatostatin;• accumulation of amylin within the islets;• local effect of insulin-like growth factor-1.An alternative hypothesis to explain the glucagon-counterregulatory defect suggests thatin healthy individuals a decrease in intra-islet insulin in association with a decrease in-cell glucose is the usual signal for release of glucagon as peripheral blood glucose levelsfall (Samols et al., 1972). Support for this comes from the observation that infusion ofthe -cell secretagogue, tolbutamide, prevents the glucagon response to hypoglycaemia innon-diabetic individuals (Banarer and Cryer, 2003). Similarly, supraphysiological levels ofinsulin have been shown to impair glucagon release in response to moderate hypoglycaemia(Kerr et al., 1991). Blunting of the normal glucagon response to hypoglycaemia can alsobe achieved in healthy volunteers by infusion of insulin-like growth factor-1 (IGF-1), theputative mediator of the somatotrophic action of growth hormone, at least in non-diabetichumans (Kerr et al., 1993) (Figure 6.7). Whether IGF-1 is involved in the pathogenesis ofglucagon counterregulatory failure in patients with type 1 diabetes is unclear.Alternatively, it is possible that chronic hyperglycaemia may directly impair pancreatic-cell function through the mechanism of glucose toxicity similar to the effect that this has on-cell function. Nevertheless, improvements in glycaemic control, following introduction of

128 COUNTERREGULATORY DEFICIENCIES IN DIABETESFigure 6.7 Infusion of insulin-like growth factor 1 abolishes the expected rise in glucagon whenblood glucose was lowered to, and maintained at, 2.8 mmol/l in healthy volunteers. Reproduced fromKerr et al. (1993) by permission of The Journal of Clinical Investigationintensive insulin therapy, invariably fail to restore the glucagon responses to hypoglycaemia(Amiel, 1991). The early appearance of an impaired response of glucagon, together with atemporal dissociation from deficiencies in other counterregulatory hormones, argues againstthis abnormality being mediated from within the central nervous system.Impairment of the glucagon response to hypoglycaemia results in greater suppressionof hepatic glucose production by insulin. Since the ability of insulin to increase glucoseutilisation is unaltered, glucose uptake by muscle and other tissues will exceed hepaticglucose production for much longer, resulting in more profound and prolonged hypoglycaemia.Whether or not the lack of a glucagon response is by itself sufficient to increase thefrequency of severe hypoglycaemia is unclear; most patients with recurrent severe hypoglycaemiahave impairment both of glucagon and epinephrine release.CatecholaminesAn attenuated epinephrine response (from the adrenal medulla) to falling blood glucoselevels together with an absent glucagon response markedly increases the risk of severehypoglycaemia in individuals with type 1 diabetes. It is unknown how common epinephrinecounterregulatory failure is in type 1 diabetes, but it is related to the duration of diabetesand the prevailing quality of glycaemic control. Estimates suggest this defect may exist inup to 45% of people with type 1 diabetes of long duration (Gerich and Bolli, 1993).The epinephrine response to hypoglycaemia can be subnormal in patients with type 1diabetes who have no clinical evidence of autonomic neuropathy and also in individualswith secondary (pancreatic) diabetes. Like glucagon, the epinephrine secretory deficiency isstimulus-specific to hypoglycaemia, remaining intact in response to exercise. The isolated

DEFECTIVE HORMONAL GLUCOSE COUNTERREGULATION 129failure of plasma epinephrine concentrations to rise in response to hypoglycaemia doesnot appreciably impair glucose counterregulation – unless it is combined with glucagondeficiency (as occurs in patients with long-standing type 1 diabetes), when the risk of severeand prolonged hypoglycaemia is significantly increased.As mentioned above, deficient sympathoadrenal responses to hypoglycaemia are majorcomponents of the clinical problem of defective glucose counterregulation and impairedawareness of hypoglycaemia in type 1 diabetes (Cryer, 2004; 2005). This may occur as aresult of the following:• reduced release of catecholamines and acetyl choline;• possibly a reduction in tissue sensitivity to the actions of these substances, although thisis controversial.Of clinical importance is the observation that a hypoglycaemic episode per se, adverselyalters the subsequent sympathoadrenal responses to hypoglycaemia (Heller and Cryer, 1991).Cryer has termed this Hypoglycaemia Associated Autonomic Failure (HAAF) (Figure 6.8).In other words, a recent episode of hypoglycaemia, known as ‘antecedent hypoglycaemia’,(including an asymptomatic event) diminishes the sympathoadrenal, symptomatic and cognitiveresponses to subsequent hypoglycaemia. As a corollary, avoidance of hypoglycaemia(for as little as two weeks) markedly improves the responses (Cranston et al., 1994). Sympathoadrenalresponses to hypoglycaemia are also affected by recent exercise, whether the fallin blood glucose level occurs when an individual is awake or asleep (Jones et al., 1998;Banarer and Cryer, 2003) (Figure 6.9) and by the time of day or night (Merl et al., 2004).Insulin Deficient Diabetes(Imperfect Insulin Replacement)(No Insulin, No Glucagon)SleepAntecedent HypoglycaemiaReduced SympathoadrenalResponses to HypoglycaemiaAntecedentExerciseReduced SympatheticNeural ResponsesReduced EpinephrineResponsesHypoglycaemiaUnawarenessDefective GlucoseCounterregulationRecurrent HypoglycaemiaFigure 6.8 Hypoglycaemia Associated Autonomic Failure (HAAF) in type 1 diabetes. Adapted fromCryer (2005) and reproduced courtesy of The American Diabetes Association

130 COUNTERREGULATORY DEFICIENCIES IN DIABETES(a)Nominal glucose (mg / dl)(b)Nominal glucose (mg / dl)8575 65 55 458575 65 55 45Epinephrine (pg/ml)60050040030020010000 60 120 180 240 300Time (min)300025002000150010005000pmol / lEpinephrine (pg / ml)6005004003002001000Night awakeMorning awake0 60 120 180 240 300Time (min)300025002000150010005000pmol / lAsleepFigure 6.9 Plasma epinephrine (adrenaline) in non-diabetic subjects (a) and in patients with type 1diabetes (b) studied in the morning while awake and during the night while awake and asleep. Adaptedfrom Banarer and Cryer (2003) and reproduced courtesy of The American Diabetes Association(□ morning awake; • night awake; asleep)Cortisol and Growth HormoneGrowth hormone (GH) and cortisol are thought to become important glucose-raisinghormones only after hypoglycaemia has been prolonged for more than one hour. However,defects in cortisol and GH release can cause profound and prolonged hypoglycaemia becauseof a reduction in hepatic glucose production and, to a lesser extent, by exaggeration ofinsulin-stimulated glucose uptake by muscle.Abnormalities in growth hormone and cortisol secretion in response to hypoglycaemiaare characteristic of long-standing type 1 diabetes, affecting up to a quarter of patientswho have had diabetes for more than ten years. In rare cases, coexistent endocrine failuresuch as Addison’s disease or hypopituitarism also predisposes patients to severe hypoglycaemia.Pituitary failure, although uncommonly associated with type 1 diabetes, occasionallydevelops in young women as a consequence of ante-partum pituitary infarction. As anintact hypothalamic–pituitary–adrenal axis is important for adequate counterregulation, thisaxis should be formally assessed in any individual with brittle diabetes presenting withunexplained, recurrent hypoglycaemia (Hardy et al., 1994; Flanagan and Kerr, 1996).More commonly, ingestion of even modest amounts of alcohol can significantly attenuatenormal growth hormone secretion and increase the risk of hypoglycaemia especially thefollowing morning (see Chapter 5).MECHANISMS OF COUNTERREGULATORY FAILUREAt the onset of type 1 diabetes, hormonal counterregulation is usually normal but within fiveyears of diagnosis, glucagon responses to hypoglycaemia become markedly impaired or evenabsent, although a glucagon response can occur if the hypoglycaemic stimulus is sufficientlyprofound (Frier et al. 1988; Hvidberg et al. 1998). After ten years of diabetes, patients usuallyhave a sub-optimal epinephrine response to compound the absent glucagon response to a fall

MECHANISMS OF COUNTERREGULATORY FAILURE 131in blood glucose (White et al., 1985) (Figure 6.10). Thus, patients with type 1 diabetes oflong duration are at risk of severe and prolonged neuroglycopenia during hypoglycaemia asa direct consequence of inadequate glucose counterregulation. Although attenuated growthhormone and cortisol responses are less common, they are late manifestations in terms ofdiabetes duration.As mentioned previously, these defects in glucose counterregulation are not ‘all or nothing’changes but can be influenced by the prevailing standard of glycaemic control and by thefrequency of hypoglycaemic episodes. Various theories relate to the clinical observation thatblood glucose thresholds for the release of counterregulatory hormone levels can change afterperiods of recurrent hypoglycaemia (Cryer, 2005). These may relate to changes at the levelof the CNS, which co-ordinates the usual responses to low blood glucose levels. At presentthere is little evidence to suggest that the alterations associated with recurrent hypoglycaemiaoccur at glucose sensors outside the CNS, for example, within the portal vein.(a)type 1 diabetes1–5 years(b)type 1 diabetes14–31 yearsInsulinInsulinGlucagon (pg / ml) Epinephrine (pg / ml) Glucose (mmol / I)54323002001000200150100Glucagon (pg / ml) Epinephrine (pg / ml) Glucose (mmol / I)543230020010002001501000 30 60 90 120 150 0 30 60 90 120 150Time (min)Time (min)Figure 6.10 Influence of duration of diabetes on glucagon and epinephrine responses to hypoglycaemiain patients with type 1 diabetes (•) after (a) 1–5 years (glucagon response is blunted whereasepinephrine release is preserved); and (b) with long-standing diabetes, both responses become severelyimpaired. = non-diabetic controls. Reproduced from Textbook of Diabetes, 2nd edition (1997)Pickup J. and William G. (eds) by permission of Blackwell Science Ltd. Data sourced from Bolliet al. (1983). Copyright © 1983 American Diabetes Association. Reprinted with permission from TheAmerican Diabetes Association

132 COUNTERREGULATORY DEFICIENCIES IN DIABETESSystemic MediatorThe systemic mediator theory suggests that a substance is released in response to hypoglycaemiawhich attenuates subsequent sympathoadrenal responses to further episodes of hypoglycaemia.The initial candidate for this was cortisol, based on two observations: first, theattenuating effect of antecedent hypoglycaemia on later sympathoadrenal responses is absentin patients with primary adrenocortical failure; and second, in healthy volunteers, followinginfusions of cortisol (to supraphysiological levels) during euglycaemia, adrenomedullaryepinephrine secretion and muscle sympathetic neural activity were reduced during subsequenthypoglycaemia (Davis et al., 1996; 1997). However, this effect of cortisol is lost ifthe prevailing cortisol levels are lowered towards those seen during hypoglycaemia or ifrecurrent hypoglycaemia is induced in animals that are genetically modified to have absentadrenocortical responses (Raju et al., 2003; McGuiness et al., 2005).Brain Fuel TransportThis mechanism is based on the hypothesis that following antecedent hypoglycaemia, glucosetransport from blood into brain tissue is increased – in animals by increasing GLUT-1 transportacross the brain microvasculature (McCall et al., 1986; Kumagai et al., 1995). In patients withtype 1 diabetes whose treatment resulted in near normal glucose levels, impaired awareness ofhypoglycaemia can develop – such patients are at increased risk of seizures and coma. Boyleet al. (1995) tested the hypothesis that during hypoglycaemia, these patients would have normalglucose uptake in the brain and consequently that sympathoadrenal activation would not occur,resulting in impaired awareness of hypoglycaemia. They found that there was no significantchange in the uptake of glucose in the brain among the patients with type 1 diabetes who hadthe lowest HbA 1c levels. Conversely, glucose uptake in the brain fell in patients with less wellcontrolledtype 1 diabetes. The responses of plasma epinephrine and pancreatic polypeptide andthe frequency of symptoms of hypoglycaemia were also lowest in the group with the lowestHbA 1c values. They concluded that during hypoglycaemia, patients with nearly normal HbA 1cvalues have normal glucose uptake in the brain, preserving cerebral metabolism, reducing theresponses of counterregulatory hormones, and causing impaired awareness of hypoglycaemia(Boyle et al., 1995). However, these findings occurred after days of prolonged hypoglycaemiawhich is in contrast to the clinical observation that attenuated sympathoadrenal responses occurwithin hours of a hypoglycaemic event.More recently, studies using positron emission tomography have found no change in bloodto-brainglucose transport 24 hours after an episode of hypoglycaemia and no differencesbetween individuals with and without hypoglycaemia awareness (Segel et al., 2001; Binghamet al., 2005). It remains possible, however, that there are changes in the transport of alternativecerebral fuels following antecedent hypoglycaemia.Brain MetabolismIt has been hypothesised that brain metabolism per se is altered following an episode ofhypoglycaemia. Most research in this area has focused on the ventromedial nucleus of thehypothalamus. Glucose deprivation in the VMH (by administration of 2-deoxyglucose) activatesthe sympathoadrenal system and increases glucagon secretion whereas local perfusion

AGE, OBESITY AND GLUCOSE COUNTERREGULATION 133GlucoseDeliveryUtilisationAdenosineCaffeineFigure 6.11 Caffeine may act by uncoupling brain glucose demand (increased) and substrate delivery(decreased) through its actions on adenosine receptors. Reproduced from Brain Research Reviews, 17,Nehlig et al., 139–169, Copyright (1992), with permission from Elsevierof the area with glucose suppresses these responses during systemic hypoglycaemia (Borget al., 1995; Borg et al., 1997). The mechanisms involved are unknown but may be a consequenceof increased glucokinase activity to enhance glucose metabolism in neurones in theregion (Gabriely and Shamoon, 2005). However, it is likely that brain metabolism in areasand other signalling mechanisms within the CNS are influenced by recurrent antecedenthypoglycaemia (Cryer, 2005).The brain glycogen supercompensation hypothesis suggests that after a single episodeof hypoglycaemia, there is a rebound increase in glycogen formation in brain astrocytes toprovide additional substrates (e.g. lactate) for brain metabolism (Choi et al., 2003). Alterationsof substrate delivery to the brain do appear to influence the magnitude of the hormonalcounterregulatory response to hypoglycaemia in healthy volunteers and in patients with type1 diabetes. Infusions of acetazolamide, a potent cerebral vasodilator, markedly attenuatesthese responses (Thomas et al., 1997) whereas ingestion of modest amounts of caffeine (toreduce substrate delivery) augments the responses (Debrah et al., 1996). The mechanismsof the latter is unknown but may involve antagonism of central adenosine receptors withuncoupling of brain blood flow (i.e., substrate delivery) and brain glucose metabolism (i.e.,brain glucose demand) resulting in relative cerebral neuroglycopenia (Figure 6.11). This isdiscussed in more detail in Chapter 5.AGE, OBESITY AND GLUCOSE COUNTERREGULATIONIn children with type 1 diabetes, the glucagon response to hypoglycaemia is markedly attenuatedcompared to non-diabetic individuals but compensated for by vigorous secretion of othercounterregulatory hormones, particularly epinephrine, with the peak epinephrine responsesbeing almost two-fold higher than in adults (Amiel et al., 1987). The total sympathoadrenalresponses to hypoglycaemia are also influenced by pubertal stage (Ross et al., 2005).

134 COUNTERREGULATORY DEFICIENCIES IN DIABETESFurthermore, it appears that the glycaemic thresholds for the secretion of epinephrine andgrowth hormone are set at a higher blood glucose level in non-diabetic children comparedto adults (Jones et al., 1991). In children with type 1 diabetes, the secretion of epinephrinein response to hypoglycaemia commences at an even higher level. In children who havemarkedly elevated HbA 1c values, there is a further shift of the blood glucose threshold to ahigher level for the release of counterregulatory hormones.Advanced age, in otherwise healthy people, does not appear to diminish or delay counterregulatoryresponses to hypoglycaemia (Brierley et al., 1995), although the magnitudeof responses of epinephrine and glucagon is lower at milder hypoglycaemic levels (around3.4 mmol/l ) compared to younger non-diabetic subjects, but is much more comparablewith a more profound hypoglycaemic stimulus (2.8 mmol/l) (Ortiz-Alonso et al., 1994) (SeeChapter 11). The magnitude of counterregulatory responses to low blood glucose levelsfollowing preceding hypoglycaemia also appears to depend on the gender of experimentalsubjects, with men having blunted responses compared to women (Davis et al., 2000b).Both the autonomic nervous system and the hypothalamic–pituitary–adrenal axis areactivated in excess in the morbidly obese. Before and after bariatric surgery (averageweight loss 40 kg over 12 months), severely obese non-diabetic subjects, underwenta hyperinsulinaemic hypoglycaemic clamp (blood glucose 3.4 mmol/l). Before weightreduction, patients demonstrated brisk peak responses in glucagon, epinephrine, pancreaticpolypeptide, and norepinephrine. After surgery and during hypoglycaemia, all theseresponses were attenuated and most markedly so for glucagon, which was totally abolishedin association with a marked improvement in insulin sensitivity. In contrast,the growth hormone response was increased after weight reduction (Guldstrand et al.,2003).HUMAN INSULIN AND COUNTERREGULATIONAt present there is no consistent evidence that the species of insulin is an important determinantof the counterregulatory response to hypoglycaemia. Over 25 clinical laboratory studieshave examined the effect of insulin species on the counterregulatory response to hypoglycaemiainduced by an intravenous bolus, intravenous infusion, or subcutaneous injection ofinsulin (Fisher and Frier, 1993; Jorgensson et al., 1994). Most of the studies showed nosignificant differences between the hormonal responses. Two studies showed a reduction inthe epinephrine response to hypoglycaemia, and both of these studies also reported diminishedautonomic symptoms to hypoglycaemia after human insulin (Schluter et al., 1982;Heine et al., 1989).A meta-analysis comparison of the effects of human and animal insulin as well as ofthe adverse reaction profiles did not show clinically relevant differences between speciesespecially in terms of risk and responses to hypoglycaemia (Richter and Neises, 2005).TREATMENT OF COUNTERREGULATORY FAILUREAt present no treatment is available that will reverse the glucagon deficit that developswithin a few years of the onset of type 1 diabetes (Figure 6.12). However, there are strategies

TREATMENT OF COUNTERREGULATORY FAILURE 135GlucagonGlucose Glucose GlucoseTimeTimeGlucagonEpinephrineEpinephrineGlucagon andEpinephrineGlucagon andEpinephrineTimeFigure 6.12 Schematic representation of the consequences of defective glucagon, epinephrine or acombined defect of glucagon and epinephrine release during recovery from hypoglycaemiato reduce the risk of hypoglycaemia from causing further hypoglycaemia and promotingHypoglycaemia Associated Autonomic Failure:• relaxation of glycaemic targets;• use of multiple daily injections of insulin, utilising rapid-acting and basal analogues (Bolli,2006);• consideration of switching to continuous subcutaneous insulin infusion therapy (Chaseet al., 2006);

136 COUNTERREGULATORY DEFICIENCIES IN DIABETES• intensive education including carbohydrate counting and appropriate blood glucose monitoring;• use of novel technologies to aid diagnosis, e.g. continuous glucose monitoring (Cheyneand Kerr, 2002);• discussion of patient factors, e.g. lipohypertrophy and other injection site problems,alcohol, caffeine consumption.Common psychological problems known to affect diabetes management adversely, such asanxiety, depression and eating disorders, have been extensively reported (Jacqueminet et al.,2005). Of particular relevance is the recognition of the role played by high levels of anxiety.Evidence-based treatment interventions are available for treating anxiety in the non-diabeticpopulation; however a systematic review and meta-analysis of randomised controlled trialsof psychological interventions for adults with diabetes has yet to be conducted. In clinicalpractice it continues to be recognised that there is a group of patients whose lives arecompletely disrupted by recurrent episodes of hyper and hypoglycaemia – the so-called‘brittle diabetic’. The outlook for such patients is usually poor (Tattersall et al., 1991).CONCLUSIONS• In non-diabetic individuals, clinically significant hypoglycaemia is an extremely rare eventbecause of effective glucose counterregulation. This includes suppression of endogenouspancreatic insulin secretion, and release of glucagon, catecholamines, cortisol and growthhormone.• The brain is the critical organ for co-ordination of the physiological responses to lowblood glucose levels.• People with diabetes almost inevitably lose their ability to release glucagon in responseto a fall in blood glucose, within five years of diagnosis. After ten years, a significantproportion of patients also has deficient epinephrine responses and is at increased risk ofmore protracted hypoglycaemia and neuroglycopenia.• Modern treatment of type 1 diabetes, including intensive education and treatment regimensutilising new technologies, has failed to eradicate completely the problem of recurrenthypoglycaemia.REFERENCESAmiel SA, Simonson DC, Sherwin RS, Lauriano AA, Tamborlane WV (1987). Exaggeratedepinephrine responses to hypoglycemia in normal and insulin-dependent diabetic children. Journalof Pediatrics 110: 832–7.Amiel S (1991). Glucose counter-regulation in health and disease: current concepts in hypoglycaemiarecognition and response. Quarterly Journal of Medicine 293: 707–27.Banarer S, Cryer PE (2003). Sleep-related hypoglycemia-associated autonomic failure in type 1diabetes: reduced awakening from sleep during hypoglycemia. Diabetes 52: 1195–203.

REFERENCES 137Bingham EM, Dunn JT, Smith D, Sutcliffe-Goulden J, Reed LJ, Mansden PK, Amiel SA (2005).Differential changes in brain glucose metabolism during hypoglycaemia accompany loss of hypoglycaemiaawareness in men with type 1 diabetes mellitus. An [11C]-3-O-methyl-D-glucose PETstudy. Diabetologia 48: 2080–9.Bolli G, De Feo P, Compagnucci P, Cartechini MG, Angeletti G, Santeusanio F et al. (1983).Abnormal glucose counterregulation in insulin-dependent diabetes mellitus. Interaction of antiinsulinantibodies and impaired glucagon and epinephrine secretion. Diabetes 32: 134–41.Bolli GM (2006). Insulin treatment in type 1 diabetes. Endocrine Practice 12 Suppl 1: 105–9.Borg WP, During MJ, Sherwin RS, Borg MA, Brines ML, Shulman GI (1994). Ventromedial hypothalamiclesions in rats suppress counter regulatory responses to hypoglycemia. Journal of ClinicalInvestigation 93: 1677–82.Borg WP, Sherwin RS, During MJ, Borg MA, Shulman GI (1995). Local ventromedial hypothalamusglucopenia triggers counterregulatory hormone release. Diabetes 44: 180–4.Borg MA, Sherwin RS, Borg WP, Tamborlane WV, Shulman GI (1997). Local ventromedial hypothalamusglucose perfusion blocks counterregulation during systemic hypoglycemia in awake rats.Journal of Clinical Investigation 99: 361–5.Boyle PJ, Kempers SF, O’Connor AM, Nagy RJ (1995). Brain glucose uptake and unawarenessof hypoglycemia in patients with insulin dependent diabetes mellitus. New England Journal ofMedicine 333: 1726–31.Brierley EJ, Broughton DL, James OFW, Alberti KGMM (1995). Reduced awareness of hypoglycaemiain the elderly despite an intact counter-regulatory response. Quarterly Journal of Medicine 88: 439–45.Bulsara M, Holman CD, Davies EA, Jones TW (2004). The impact of a decade of changing treatmenton rates of severe hypoglycemia in a population-based cohort of children with type 1 diabetes.Diabetes Care 27: 2293–8.Chase HP, Horner BP, McFann K, Yetzer H, Gaston J, Banion C et al. (2006). The use of insulin pumpswith meal bolus alarms in children with type 1 diabetes to improve glycemic control. Diabetes Care29: 1012–5.Cheyne E, Kerr D (2002) Making ‘sense’ of diabetes: using a continuous glucose sensor in clinicalpractice. Diabetes Metabolism Research Reviews 18 Suppl 1: S43–8.Choi I Y, Seaquist ER, Gruetter R (2003). Effect of hypoglycemia on brain glycogen metabolism invivo. Journal of Neuroscience Research 72: 25–32.Cox DJ, Irving GA, Gonder-Frederick L, Nowacek G, Butterfield J (1987). Fear of hypoglycemia:quantification, validation and utilization. Diabetes Care 10: 617–21.Cranston I, Lomas J, Maran A, Macdonald IA, Amiel SA (1994). Restoration of hypoglycaemiaawareness in patients with long-duration insulin-dependent diabetes. Lancet 344: 283–7.Cryer PE (2004). Diverse causes of hypoglycemia-associated autonomic failure in diabetes. NewEngland Journal of Medicine 350: 2272–9.Cryer PE (2005). Mechanisms of hypoglycemia-associated autonomic failure and its componentsyndromes in diabetes. Diabetes 54: 3592–601.DAFNE Study Group (2002). Training in flexible, intensive insulin management to enable dietaryfreedom in people with type 1 diabetes: dose adjustment for normal eating (DAFNE) randomisedcontrolled trial. British Medical Journal 325: 746.Davis SN, Shavers C, Costa F, Mosqueda-Garcia R (1996). Role of cortisol in the pathogenesis ofdeficient counterregulation after antecedent hypoglycemia in normal humans. Journal of ClinicalInvestigation 98: 680–91.Davis SN, Shavers C, Davis B, Costa F (1997). Prevention of an increase in plasma cortisol duringhypoglycemia preserves subsequent counterregulatory responses. Journal of Clinical Investigation100: 429–38.Davis SN, Mann S, Galassetti P, Neil R, Tate D, Ertl AC, Costa F (2000a). Effects of differingdurations of antecedent hypoglycemia on counterregulatory responses to subsequent hypoglycemiain normal humans. Diabetes 49: 1897–903.

138 COUNTERREGULATORY DEFICIENCIES IN DIABETESDavis SN, Fowler F, Costa S (2000b). Hypoglycemic counterregulatory responses differ between menand women with type 1 diabetes. Diabetes 49: 65–72.Debrah K, Sherwin RS, Murphy J, Kerr D (1996). Effect of caffeine on recognition of and physiologicalresponses to hypoglycaemia in insulin-dependent diabetes. Lancet 347: 19–24Diem P, Redmon JB, Abid M, Moran A, Sutherland DE, Halter JB, Robertson RP (1990). Glucagon,catecholamine and pancreatic polypeptide secretion in type I diabetic recipients of pancreas allografts.Journal of Clinical Investigation 86: 2008–13.Fisher M, Frier BM (1993). Hypoglycaemia and human insulin. In: Hypoglycaemia and Diabetes:Clinical and Physiological Aspects. Frier BM and Fisher M, eds. Edward Arnold, London: 314–27.Fisher S, Brunning J, Lannon S, Kahn CR (2005). Insulin signaling in the central nervous system iscritical for the normal sympathoadrenal response to hypoglycemia. Diabetes 54: 1447–51.Flanagan D, Kerr D (1996). Recurrent hypoglycemia and long-standing brittle diabetes. Endocrinologyand Metabolism 3: 67–69.Frier BM, Fisher M, Gray CE, Beastall GH (1988). Counterregulatory hormonal responses to hypoglycaemiain type 1 (insulin-dependent) diabetes: evidence for diminished hypothalamic-pituitaryhormonal secretion. Diabetologia 31: 421–9.Gabriely I, Shamoon H (2005). Fructose normalizes specific counterregulatory responses to hypoglycemiain patients with type 1 diabetes. Diabetes 54: 609–16.Gerich JE, Langlois M, Noacco C, Karam JH, Forsham PH (1973). Lack of glucagon response tohypoglycemia in diabetes: evidence for an intrinsic pancreatic alpha cell defect. Science 182: 171–3.Gerich JE, Bolli GB (1993). Counterregulatory failure. In: Hypoglycaemia and Diabetes: Clinical andPhysiological Aspects. Frier BM and Fisher M, eds. Edward Arnold, London: 253–67.Guldstrand M, Ahren B, Wredling R, Backman L, Lins P, Adamson U (2003). Alteration of thecounterregulatory responses to insulin-induced hypoglycemia and of cognitive function after massiveweight reduction in severely obese subjects. Metabolism 52: 900–7.Hardy KJ, Burge MR, Boyle P, Scarpello JHB (1994). A treatable cause of recurrent severe hypoglycemia.Diabetes Care 17: 722–4.Heine RJ, Van der Heyden EAP, Van der Veen EA (1989). Responses to human and porcine insulinin healthy subjects. Lancet 334: 946–9.Heller SR, Cryer PE (1991). Reduced neuroendocrine and symptomatic responses to subsequenthypoglycemia after 1 episode of hypoglycemia in non-diabetic humans. Diabetes 40: 223–6.Hvidberg A, Juel Christensen N, Hilsted J (1998). Counterregulatory hormones in insulin-treateddiabetic patients admitted to an Accident and Emergency Department with hypoglycaemia. DiabeticMedicine 15: 199–204.Jacqueminet S, Masseboeuff N, Rolland M, Grimaldi A, Sachon C (2005). Limitations of the so-called‘intensified’ insulin therapy in type 1 diabetes mellitus. Diabetes Metabolism 31(4 Pt 2): 4S45–50.Jones TW, Boulware SD, Kraemer DT, Caprio S, Sherwin RS, Tamborlane WV (1991). Independenteffects of youth and poor diabetes control on responses to hypoglycemia in children. Diabetes 40:358–63.Jones TW, Porter P, Sherwin RS, Davis EA, O’Leary P, Frazer F et al. (1998). Decreased epinephrineresponses to hypoglycemia during sleep. New England Journal of Medicine 338: 1657–62.Jorgensen L, Dejgaard A, Pramming SK (1994). Human insulin and hypoglycaemia: a literature survey.Diabetic Medicine 11: 925–34.Kerr D, Reza M, Smith N, Leatherdale BA (1991). Importance of insulin in subjective, cognitive andhormonal responses to hypoglycemia in patients with IDDM. Diabetes 40: 1057–62.Kerr D, Tamborlane WV, Rife F, Sherwin RS (1993). Effect of insulin-like growth factor-1 on theresponses to and recognition of hypoglycemia in humans. Journal of Clinical Investigation 91:141–7.Kumagai AK, Kang YS, Boado RJ, Pardridge WM (1995). Upregulation of blood-brain barrier GLUT-1 glucose transporter protein and mRNA in experimental chronic hypoglycemia. Diabetes 44:1399–404.

REFERENCES 139McCall AL, Fixman LB, Fleming N, Tornheim K, Chick W, Ruderman NB (1986). Chronic hypoglycemiaincreases brain glucose transport. American Journal of Endocrinology 251: E442–7.McCrimmon RJ, Evans ML, Fan X, McNay EC, Chan O, Ding Y et al. (2005). Activation ofATP-sensitive K+ channels in the ventromedial hypothalamus amplifies counterregulatory hormoneresponses to hypoglycemia in normal and recurrently hypoglycemic rats. Diabetes 54: 3169–74.McGuiness O, Ansari T, Potts J, Jacobson L (2005). Glucocorticoid insufficiency in corticotrophinreleasing hormone knockout mice does not preserve autonomic responses during repeated hypoglycemiaclamps. Diabetes 54 (Suppl): A26 (abstract).Merl V, Kern W, Peters A, Oltmanns KM, Gais S, Born J et al. (2004). Differences between nighttimeand daytime hypoglycemia counterregulation in healthy humans. Metabolism 53: 894–8.Mitrakou A, Ryan C, Veneman T, Mokan M, Jenssen T, Kiss I et al. (1991). Hierarchy of glycemicthresholds for counterregulatory hormonal secretion, symptoms, and cerebral dysfunction. AmericanJournal of Endocrinology 260: E67–74.Muhlhauser I, Bruckner I, Berger M, Cheta D, Jorgens V, Lonescu-Tirgoviste C et al. (1987). Evaluationof an intensified insulin treatment and teaching programme as routine management of type 1(insulin-dependent) diabetes. The Bucharest–Dusseldorf Study. Diabetologia 30: 681–90.Nehlig A, Daval JL, Debry G (1992). Caffeine and the central nervous system: mechanisms of action,biochemical, metabolic and psychostimulant effects. Brain Research Reviews 17: 139–69.Ortiz-Alonso FJ, Galecki A, Herman WH, Smith MJ, Jacquez JA, Halter JB (1994). Hypoglycemiacounterregulation in elderly humans: relationship to glucose levels. American Journal ofEndocrinology 267: E497–506.Raju B, McGregor VP, Cryer PE (2003). Cortisol elevations comparable to those that occurduring hypoglycemia do not cause hypoglycemia-associated autonomic failure. Diabetes 52:2083–89.Richter B, Neises G (2005). ‘Human’ insulin versus animal insulin in people with diabetes mellitus.Cochrane Database Systematic Reviews Jan 25(1): CD003816.Ross LA, Warren RE, Kelnar CJ, Frier BM. (2005). Pubertal stage and hypoglycaemia counterregulationin type 1 diabetes. Archives of Disease in Childhood 90: 190–4.Rossing P. (2005) The changing epidemiology of diabetic microangiopathy in type 1 diabetes.Diabetologia 2005; 48: 1439–44.Samols E, Tyler J, Marks V (1972): Glucagon-insulin interrelationships. In: Glucagon: MolecularPhysiology, Clinical and Therapeutic Implications. Lefebvre P, Unger RH, eds. Pergamon, ElmsfordNY: 151–174.Schluter KJ, Petersen KG, Sontheimer K, Enzmann F, Kerp L (1982). Different counterregulatoryresponses to human insulin (recombinant DNA) and purified pork insulin. Diabetes Care 5 (suppl 2):78–81.Segel SA, Fanelli CG, Dence CS, Markham J, Videen TO, Paramore DS et al. (2001). Blood-to-brainglucose transport, cerebral glucose metabolism and cerebral blood flow are not increased followinghypoglycemia. Diabetes 50: 1911–17.Smith D, Pernet D, Reid H, Bingham E, Rosenthal JM, Macdonald IA et al. (2002). The role ofhepatic portal glucose sensing in modulating responses to hypoglycaemia in man. Diabetologia 45:1416–24.Tappy L, Chiorelo R, Berger M (1999). Autoregulation of glucose production in health and disease.Current Opinions in Clinical Nutrition and Metabolic Care 2: 161–4.Tattersall RB, Gregory R, Selby C, Kerr D, Heller S (1991). Course of brittle diabetes: 12 year followup. British Medical Journal 302: 1240–3.The Diabetes Control and Complications Trial Research Group (1993). The effect of intensive treatmentof diabetes on the development and progression of long-term complications in insulin-dependentdiabetes mellitus. New England Journal of Medicine 329: 977–86.Thomas M, Sherwin R, Murphy J, Kerr D (1997). Importance of cerebral blood flow on recognitionof and physiological responses to hypoglycemia. Diabetes 46: 829–33.

140 COUNTERREGULATORY DEFICIENCIES IN DIABETESWeintrob N, Schechter A, Benzaquen H, Shalitin S, Lilos P, Galatzer A, Phillip M (2004). Glycemicpatterns detected by continuous subcutaneous glucose sensing in children and adolescents withtype 1 diabetes mellitus treated by multiple daily injections versus continuous subcutaneous insulininfusion. Archives of Pediatric and Adolescent Medicine 158: 677–84.White NH, Gingerich RL, Levandoski LA, Cryer PE, Santiago JV (1985). Plasma pancreatic polypeptideresponse to insulin-induced hypoglycemia as a marker for defective glucose counterregulationin insulin-dependent diabetes. Diabetes 34: 870–5.

7 Impaired Awareness ofHypoglycaemiaBrian M. FrierDangerous hypoglycaemia may occur without warning symptoms– E.P. Joslin et al. (1922)INTRODUCTIONThe generation of symptoms in response to hypoglycaemia provides a fundamental defencefor the brain, by alerting the affected individual to the imminent development of neuroglycopenia(Chapter 2). This should provoke an appropriate response – obtaining and ingestingsome form of carbohydrate to reverse the low blood glucose. If these warning symptomsfail to occur, or they are delayed until the blood glucose has fallen to a level that causesdisabling neuroglycopenia, serious consequences may ensue. When the normal warningmechanisms are deficient or are ignored and no avoiding action is taken, severe hypoglycaemiamay occur, with progression to confusion, altered consciousness and eventual coma.An inadequate symptomatic warning often occurs in people with insulin-treated diabetes, invarious circumstances and with differing causes, and is described as impaired awarenessof hypoglycaemia or hypoglycaemia unawareness. This is an acquired abnormality that iseffectively a complication of insulin therapy, and should be ranked alongside the microvascularcomplications of diabetes such as retinopathy, neuropathy or nephropathy, because itsmorbidity can be just as serious and disabling.NORMAL RESPONSES TO HYPOGLYCAEMIAAcute hypoglycaemia induces a series of changes – hormonal, neurophysiological, symptomaticand cognitive – which occur at different and defined blood glucose concentrations(Figure 7.1). The thresholds at which these changes are triggered have been described innon-diabetic humans, most occurring within a relatively narrow range of blood glucoseconcentrations. In diabetic individuals these glycaemic thresholds are not static and permanent,but are dynamic and display plasticity, altering in response to external influences suchas changes in glycaemic control and exposure to extremes of blood glucose. Thus the bloodglucose level at which symptoms are activated can be modified through the ability of thebrain to adapt to environmental change, that is, its exposure to prevailing blood glucoseconcentrations.Hypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

142 IMPAIRED AWARENESS OF HYPOGLYCAEMIAArterialised venous blood glucose concentration (mmol/L) mmol/LInhibition of endogenousinsulin secretion3.8 mmol/LCounterregulatoryhormone release• Glucagon• Epinephrine3.2–2.8 mmol/LOnset of symptoms• Autonomic• Neuroglycopenic3.0–2.4 mmol/LNeurophysiologicaldysfunction• Evoked responses3.0 mmol/LWidespreadEEG changes2.8 mmol/LCognitivedysfunction• Inability toperform complextasks< 1.5 mmol/LSevereneuroglycopenia• Reducedconscious level• Convulsions• Coma0Figure 7.1 Hierarchy of endocrine, symptomatic and neurological responses to acute hypoglycaemiain non-diabetic subjects. Glycaemic thresholds are based on glucose concentrations in arterialisedvenous blood. Modified from Textbook of Diabetes, 2nd edition (1997) (eds J. Pickup and G. Williams),by permission of Blackwell Science LtdDepriving the brain of glucose causes it to malfunction, and cognitive impairment quicklybecomes evident as an overt manifestation of neuroglycopenia. Some of these features arerelatively subtle, and may not be detected immediately by the patient. A fall in blood glucosetriggers activation of the peripheral autonomic nervous system via central hypothalamic autonomiccentres within the brain, and stimulates the sympathoadrenal system. This promotestypical physiological responses including sweating, an increase in rate and contractility of theheart (sensed as a pounding heart), and tremor, these being some of the classical features ofthe autonomic reaction (Figure 7.2). Epinephrine (adrenaline) is secreted in large quantitiesfrom the adrenal medullae and contributes to some of the symptoms mainly by heighteningthe magnitude of the response. The early literature on hypoglycaemia and diabetes providesaccurate descriptions of the autonomic features of acute hypoglycaemia, and patients andphysicians alike commonly discussed hypoglycaemic ‘reactions’, a term that regrettably isnow seldom used. It emphatically describes the sudden, and often florid, onset of the autonomicfeatures of hypoglycaemia, which drive the individual to seek assistance or obtain asupply of glucose to relieve these unpleasant symptoms.‘Awareness’ of HypoglycaemiaThe generation of typical physiological responses to hypoglycaemia is perceived throughsensory feedback to the brain, and after central processing, an appropriate motor responseis made. Much has been made by some commentators of the predominant importanceof autonomic symptoms in the detection of the onset of hypoglycaemia. This premise is

NORMAL RESPONSES TO HYPOGLYCAEMIA 143We do not have rights to reproduce thisfigure electronicallyFigure 7.2 Generation of neuroglycopenic and autonomic symptoms in response to hypoglycaemia.Autonomic activation and the involvement of the sympatho-adrenal system in the stimulation of representativeend-organs associated with common autonomic symptoms of hypoglycaemia. Reproducedfrom Hypoglycaemia and Diabetes (eds B.M. Frier and M. Fisher), © 1993 Edward Arnold, bypermission of Edward Arnold (Publishers) Ltdbased partly on the laboratory-based observation of non-diabetic subjects that autonomicsymptoms commence at a higher blood glucose concentration (around 0.5 mmol/l) thanneuroglycopenic symptoms (Mitrakou et al., 1991). In everyday experience reported bypeople with insulin-treated diabetes, a rapid decline in blood glucose does not permit asubjective distinction to be made between these different thresholds for the developmentof autonomic and neuroglycopenic symptoms, and people treated with insulin identify bothtypes with equal frequency as their initial warning symptoms (Hepburn et al., 1992).It has been assumed that because neuroglycopenia may interfere with cognitive function,this will affect the individual’s ability to perceive and interpret neuroglycopenic cues such asthe inability to concentrate, drowsiness or difficulty with mentation. This may be true whena falling blood glucose is not treated and is allowed to drop to a level associated with severe

144 IMPAIRED AWARENESS OF HYPOGLYCAEMIAneuroglycopenia, but most patients detect (and often rely upon) neuroglycopenic symptomsduring early hypoglycaemia, and rate these as important as autonomic symptoms in providinga warning. It is the initial perception of any symptom of hypoglycaemia, irrespective ofwhether this is autonomic, neuroglycopenic or simply a vague sensation of apprehension orloss of well-being (a common early feature described by many) which constitutes ‘awareness’of hypoglycaemia. Only the initial warning symptoms are important in this respect, and notthe total spectrum or absolute number of symptoms, some of which occur too late to haveany value in alerting the patient to the impending risk of a falling blood glucose. A majordifference between the autonomic and the neuroglycopenic symptomatic response is that,once triggered, the autonomic response quickly reaches a maximum intensity which thengradually declines with time, whereas the neuroglycopenic response becomes more profoundthe further the blood glucose falls. This qualitative difference in response becomes importantif early cues are ignored or are not detected, as progressive neuroglycopenia will eventuallyinterfere with the individual’s ability to identify and self-treat the low blood glucose.When a person is fully awake, alert and on guard against possible hypoglycaemia, this symptomaticwarning system generally works very effectively (Chapter 2). However, there are manytimes in everyday life when the symptoms may be either diminished or disregarded. This isparticularly so during sleep when symptoms are seldom detected, or they may be ignoredif a person is distracted by other activities, such as watching an interesting programme ontelevision, participating in sport or concentrating on a task. Circumstances can modify thevalue of specific warning symptoms, making them difficult to interpret as features of hypoglycaemia.Examples include sweating on a hot day, shivering when the weather is cold orfeeling drowsy during a boring meeting! All of these may represent early hypoglycaemiabut are attributed to other causes by the affected person. A list of the factors that influencenormal awareness of hypoglycaemia is shown in Box 7.1. The intensity of symptoms canvary and the value of individual symptoms as warning features may not be constant in anysingle individual. This is often not appreciated in the assessment of research findings, and it isdifficult to extrapolate the careful measurement of symptomatic responses to hypoglycaemiain studies performed in a laboratory setting, to the hurly-burly of everyday life.Box 7.1Factors influencing normal awareness of hypoglycaemiaInternalPhysiologicalRecent glycaemic controlDegree of neuroglycopeniaSymptom intensity/sensitivityPsychologicalFocused attentionCongruence; denialCompeting explanationsEducationKnowledgeSymptom beliefExternalDrugsBeta-adrenoceptor blockers (non-selective)Hypnotics, tranquillisersAlcoholEnvironmentalPostureDistraction

IMPAIRED AWARENESS OF HYPOGLYCAEMIA 145Warning symptoms provide internal cues, but most people with insulin-treated diabetesalso rely on external cues based on their experience of the timing of insulin administrationin relation to food, the effect of delaying meals or the amount of food ingested, the effect ofexercise on blood glucose and many other factors that can influence short-term glycaemiccontrol. These cues are supplemented by blood glucose monitoring, which gives an exactand objective measure of prevailing glycaemia. Further useful feedback may be obtainedfrom observers such as relatives or friends, many of whom become adept at noticing earlyneuroglycopenia before the onset of the patient’s subjective warning symptoms. ‘Awareness’of hypoglycaemia is therefore distilled from a combination of resources, and has to belearned by people with newly diagnosed diabetes commencing insulin therapy. They have noprevious experience of symptoms of hypoglycaemia and must receive appropriate educationon the potential range of symptoms. Through experience they will recognise the clusterof symptoms peculiar to themselves, because symptoms are idiosyncratic. Awareness ofhypoglycaemia therefore assists in protecting the individual from the risk of an unexpectedfall in blood glucose. When awareness of hypoglycaemia becomes impaired or is absentwhile a person is awake, the individual becomes progressively vulnerable to the developmentof severe hypoglycaemia.IMPAIRED AWARENESS OF HYPOGLYCAEMIADefinitionNo satisfactory or comprehensive definition of impaired hypoglycaemia awareness has beensuggested to date. Many laboratory-based studies of experimental hypoglycaemia have usedarbitrary definitions based on witnessed observations of subjects who fail to develop classicalfeatures of hypoglycaemia, or the failure of physiological or hormonal responses toexceed twice the standard deviation from mean basal levels. These are statistical devices,which take no account of subjective reality, require the application of sophisticated andunphysiological glucose clamp procedures, and have little direct application to clinicalmanagement.Asymptomatic biochemical hypoglycaemia occurs more frequently during routine bloodglucose monitoring in diabetic patients who report impaired awareness of hypoglycaemia(Gold et al., 1994; Clarke et al., 1995) and such a record may alert the clinician to thepossibility that an individual is developing this problem. A much higher rate of undetectedhypoglycaemia in people with impaired awareness has been demonstrated during wakinghours using continuous blood glucose monitoring (Kubiak et al., 2004). However, in clinicalpractice a careful history is essential in determining whether reduced warning symptoms ofhypoglycaemia are a significant problem, and if this is occurring consistently. Patients whoassert that they have a problem with perceiving the onset of symptoms of hypoglycaemiaare generally correct in this belief (Clarke et al., 1995), so that the identification of impairedawareness of hypoglycaemia should be based principally on clinical history. Validatedscoring systems to assess awareness of hypoglycaemia have been described by Gold et al.(1994) and Clarke et al. (1995), and supportive information can be derived from simultaneousinspection of the individual’s blood glucose results. Detailed questioning of a patient about hisor her ability to detect the onset of hypoglycaemic symptoms may need to be supplementedby questioning close relatives, who often report a much higher rate of severe hypoglycaemia

146 IMPAIRED AWARENESS OF HYPOGLYCAEMIA(Heller et al., 1995; Jorgensen et al., 2003). This will provide a witnessed descriptionof how hypoglycaemia develops in a patient, with information on its true frequency andseverity. Patients often underestimate the frequency of severe hypoglycaemia, partly becauseof post-hypoglycaemia amnesia.ClassificationIn one study, Hepburn et al. (1990) subdivided hypoglycaemia awareness into three categories:normal, partial and absent awareness. These were defined as follows:• Normal awareness: the individual is always aware of the onset of hypoglycaemia.• Partial awareness: the symptom profile has changed with a reduction either in the intensityor in the number of symptoms and, in addition, the individual may be aware of someepisodes of some episodes of hypoglycaemia but not of others.• Absent awareness: the individual is no longer aware of any episode of hypoglycaemia.Although the subdivision into partial and absent awareness is artificial, it reflects the naturalhistory of this clinical problem, illustrating the gradual progression of this disability, andemphasising that in some patients the abnormality is severe (absent awareness) althoughtotal absence of clinical manifestations of hypoglycaemia (particularly the neuroglycopenicfeatures) is exceptionally rare (Gold et al., 1994, Clarke et al., 1995). The problem maynot be simply an absence of symptoms, but rather that the time during which warningsymptoms can be detected is extremely short, allowing the affected individual a very limitedopportunity to take avoiding action. Some patients describe how the onset of hypoglycaemiaappears to have become much more rapid compared with their previous experience andprogresses quickly to severe neuroglycopenia. However, impaired awareness may not necessarilyevolve into total unawareness of hypoglycaemia, and may vary over time, presumablybecause of major influences of environmental factors on the generation and perception ofsymptoms.The above classification of awareness of hypoglycaemia is far from comprehensive. Inaddition, the state of hypoglycaemia awareness can be ascertained only when the individualis in a physical state in which recognition of the onset of hypoglycaemia is possible.Therefore, if the person is asleep, intoxicated, inebriated, anaesthetised or sedated, so thattheir conscious level is reduced, they are not able to perceive (as subjective symptoms)the normal physiological manifestations of hypoglycaemia. An individual’s awareness ofhypoglycaemia can be evaluated only if hypoglycaemia occurs while the individual is awake.A further prerequisite is that the person must have had previous experience of hypoglycaemiaat some time during treatment with insulin. In assessing the present state ofhypoglycaemia awareness, it is desirable that the patient should have experienced one ormore episodes of hypoglycaemia (confirmed biochemically) within a recent time intervalsuch as the preceding year, so that a comparison of the symptoms can be made with earlierepisodes of hypoglycaemia. A diagnosis of impaired hypoglycaemia awareness cannot beentertained or surmised if a patient has either never been exposed previously to acute hypoglycaemiaor has only started to experience hypoglycaemic events very recently. Becausehypoglycaemia awareness and its impairment is a continuum ranging from normality to

PREVALENCE OF IMPAIRED AWARENESS OF HYPOGLYCAEMIA 147complete inability to detect the onset of hypoglycaemia, a classification of this conditionwill need to consider alterations in symptom intensity as well as detection of hypoglycaemiaby any means and the ability of the patient to self-treat low blood glucose.PREVALENCE OF IMPAIRED AWARENESS OFHYPOGLYCAEMIAImpaired awareness of hypoglycaemia is common in people treated with insulin. Althoughthe chronic form of this acquired condition mainly affects those with type 1 diabetes, itappears that a similar problem does eventually emerge in patients with type 2 diabetes whohave been treated with insulin for several years (Hepburn et al., 1993a). A prevalence of 8%was observed in a cohort of 215 patients in Edinburgh (Henderson et al., 2003). Becausefew patients with type 2 diabetes who require insulin therapy survive for a sufficientlylong period to permit this complication to develop, impaired awareness is principally aproblem associated with type 1 diabetes. It is not known whether impaired awareness ofhypoglycaemia occurs in diabetic patients treated with oral antidiabetic agents.Impaired awareness of hypoglycaemia has been shown to be associated with strictglycaemic control (see Chapter 8), but significant modification of the symptomatic responseto hypoglycaemia does not occur unless the glycated haemoglobin concentration is withinthe non-diabetic range (Boyle et al., 1995; Kinsley et al., 1995; Pampanelli et al., 1996).Only a small proportion of people with insulin-treated diabetes can sustain this degree ofsuper-optimal glycaemic control indefinitely. In the Diabetes Control and ComplicationsTrial (DCCT), with its extensive resources devoted to maintaining intensive insulin therapy,more than 40% of the patients in the group with strict glycaemic control achieved a HbA 1c of6.05% or less (the upper limit of the non-diabetic range) at some time during the study, butonly 5% were able to maintain this level of glycaemic control continuously (The DiabetesControl and Complications Trial Research Group, 1993). The proportion of any insulintreateddiabetic population that can achieve this therapeutic goal will depend on local policiesregarding insulin therapy, the expertise of local diabetes specialist teams, available resourcesand the enthusiasm of individual patients. With the exception of a few highly motivatedpatients, most people treated with insulin are unable to maintain strict glycaemic controlfor protracted periods. In clinical practice this ‘acute’ form of hypoglycaemia unawarenessis probably relatively uncommon. Nonetheless, the influence of strict glycaemic control onsymptomatic and counterregulatory responses to hypoglycaemia has been studied extensively,and has provided insights into the potential pathogenetic mechanisms underlyingimpaired awareness of hypoglycaemia.Reduced warning symptoms of hypoglycaemia (of varying severity) occur in approximatelyone quarter of all insulin-treated patients. Cross-sectional population surveys indifferent European and North American populations of insulin-treated diabetic patients, usingsimilar methods of assessment, have given remarkably consistent estimates (Table 7.1).Impaired awareness of hypoglycaemia becomes more common with increasing duration ofinsulin-treated diabetes (Hepburn et al., 1990), and almost 50% of patients experience hypoglycaemiawithout warning symptoms after 25 years or more of treatment (Pramming et al.,1991) (Figure 7.3). It appears therefore to be an acquired abnormality associated with insulintherapy.

148 IMPAIRED AWARENESS OF HYPOGLYCAEMIATable 7.1 Prevalence of Hypoglycaemia Unawareness in population studies ofinsulin-treated diabetesCountryNumber ofpatientsImpaired awareness ofhypoglycaemia (%)ReferenceScotland 302 23 Hepburn et al. (1990)Germany 523 25 Muhlhauser et al. (1991)Denmark 411 27 Pramming et al. (1991)USA 628 20 Orchard et al. (1991)Figure 7.3 Comparisons between the duration of diabetes and the percentage of 411 type 1 diabeticpatients reporting (a) changes in symptoms of hypoglycaemia, (b) sweating and/or tremor as one of thetwo cardinal autonomic symptoms of hypoglycaemia, and (c) severe hypoglycaemic episodes withoutwarning symptoms. Values are medians; shaded areas show 95% confidence limits. Reproduced fromPramming et al. (1991) by permission of John Wiley & Sons, Ltd

PATHOGENESIS OF IMPAIRED AWARENESS OF HYPOGLYCAEMIA 149Frequency of Associated Severe HypoglycaemiaIt is apparent that impaired awareness of hypoglycaemia is a major risk factor for severehypoglycaemia. In the DCCT, 36% of all episodes of severe hypoglycaemia occurred withno warning symptoms in patients while they were awake (The DCCT Research Group,1991). In a population study in Edinburgh, retrospective assessment of the frequency ofsevere hypoglycaemia revealed that 90% of patients with impaired awareness of symptomsexperienced severe hypoglycaemia in the preceding year, compared to 18% in a comparablegroup who had retained normal awareness (Hepburn et al., 1990). Prospective studies haveconfirmed the increase in frequency of mild and severe hypoglycaemia associated withimpaired awareness of hypoglycaemia (Gold et al., 1994; Clarke et al., 1995), with a sixfoldhigher frequency of severe hypoglycaemia being documented in people with impairedawareness (Gold et al., 1994) (Figure 7.4).Annual Prevalence(% Subjects)Incidence(Events per subject per year)Figure 7.4 Proportion of patients affected and event rates for severe hypoglycaemia in patientswith type 1 diabetes with normal (□, n= 31) or impaired (, n= 29) awareness of hypoglycaemia.Reproduced from Cryer and Frier (2004) by permission of John Wiley & Sons, Ltd. Data derived fromGold et al. (1994)PATHOGENESIS OF IMPAIRED AWARENESS OFHYPOGLYCAEMIAThe mechanisms underlying impaired awareness of hypoglycaemia are not known and maybe multifactorial. Possible mechanisms are listed in Box 7.2.

150 IMPAIRED AWARENESS OF HYPOGLYCAEMIABox 7.2Impaired awareness of hypoglycaemia: possible mechanismsCNS adaptationChronic exposure to low blood glucose• glucose clamp (2.9 mmol/l) for 56 hours in non-diabetic subjects• insulinoma in non-diabetic patients• strict glycaemic control in diabetic patientsRecurrent transient exposure to low blood glucose• antecedent hypoglycaemiaCNS glucoregulatory failure• counterregulatory deficiency (hypothalamic defect?)• hypoglycaemia associated (central) autonomic failure (HAAF)Peripheral nervous system dysfunction• peripheral autonomic neuropathy• reduced peripheral adrenoceptor sensitivityAltered Glycaemic Threshold for Initiation of SymptomsSymptoms of hypoglycaemia commence when the blood glucose reaches a specific level, andalthough this threshold may differ between individuals, it is usually constant and reproduciblein the non-diabetic state (Vea et al., 1992). This blood glucose threshold for symptoms canbe modified by protracted hypoglycaemia (Boyle et al., 1994) and is not fixed in people withdiabetes who are treated with insulin, with its dynamic nature being demonstrated in varioussituations. In clinical practice, it has long been recognised that insulin-treated diabetic patientswho have poor glycaemic control experience symptoms of hypoglycaemia when their bloodglucose declines within a hyperglycaemic range (Maddock and Krall, 1953) and this hasbeen shown to be associated with the onset of hypoglycaemic symptoms at a significantlyhigher blood glucose (4.3 mmol/l) compared to non-diabetic subjects (2.9 mmol/l) (Boyleet al., 1988). Conversely, strict glycaemic control modifies the glycaemic threshold for theonset of symptoms, which do not commence until blood glucose has declined to a lowerlevel than that required in less well controlled patients to initiate a symptomatic response(see Chapter 8).The terminology that is used in relation to a change in the glycaemic threshold is potentiallyconfusing. When a lower blood glucose is required to initiate a response, whethersymptomatic, physiological or counterregulatory, the glycaemic threshold is said to be raised

PATHOGENESIS OF IMPAIRED AWARENESS OF HYPOGLYCAEMIA 151or elevated; that is, a more profound hypoglycaemic stimulus is necessary to trigger therelevant response. Thus, strict glycaemic control raises the glycaemic threshold for theonset of symptoms, which do not occur until blood glucose has declined to a much lowerconcentration than would be observed in non-diabetic subjects.For many years, clinicians have recognised that the glycaemic threshold for the onset ofhypoglycaemic symptoms is higher in patients with a long duration of type 1 diabetes whorequire a much lower blood glucose to provoke a symptomatic response. Lawrence (1941)wrote that ‘as years of insulin life go on, sometimes only after 5–10 years, I find it almostthe rule that the type of insulin reactions change, the premonitory autonomic symptoms aremissed out and the patient proceeds directly to the more serious manifestations affecting thecentral nervous system’. He astutely suggested that ‘the tissues may become attuned to a lowersugar concentration’. Recent studies in animals and humans have shown that the brain doesadapt to chronic exposure to low blood glucose (see below) but this may not be beneficialto the individual with diabetes who is treated with insulin, i.e., it is a maladaptive response.An early study by Sussman et al. (1963) – revisited and extended by Hepburn et al.(1991) – showed that diabetic patients who had self-reported unawareness of hypoglycaemiadid mount a sympatho-adrenal response to acute hypoglycaemia, but that this occurred ata lower blood glucose concentration than comparable diabetic subjects who had normalsymptomatic awareness (Figure 7.5). However, the autonomic response was preceded by theWe do not have rights to reproduce thisfigure electronicallyFigure 7.5 Venous blood glucose concentrations for the onset of the autonomic reaction in responseto insulin-induced hypoglycaemia in individual non-diabetic control subjects, and in type 1 diabeticpatients with normal and impaired awareness of hypoglycaemia and with autonomic neuropathy.Mean + SEM is shown for each group. Data derived from Hepburn et al. (1991) and reproduced fromHypoglycaemia and Diabetes (eds B.M. Frier and M. Fisher), © 1993 Edward Arnold, by permissionof Edward Arnold (Publishers) Ltd

152 IMPAIRED AWARENESS OF HYPOGLYCAEMIAdevelopment of overt neuroglycopenia, which interfered with perception of the autonomicwarning symptoms when they did eventually occur. This sequence of responses disrupts theability of the individual subject to take appropriate action to self-treat low blood glucose.Similar findings have been reported by others (Grimaldi et al., 1990; Mokan et al., 1994;Bacatselos et al., 1995).In these studies, the counterregulatory hormonal responses to hypoglycaemia were delayed(Grimaldi et al., 1990; Hepburn et al. 1991) and their glycaemic thresholds had also shiftedto occur at lower blood glucose concentrations (Mokan et al., 1994). This is consistentwith the reported observation that impaired awareness of hypoglycaemia co-segregates withcounterregulatory hormonal deficiency in people with longstanding type 1 diabetes (Ryderet al., 1990). In addition, Mokan et al. (1994) reported that cognitive dysfunction andneuroglycopenic symptoms in people with impaired awareness occurred at lower bloodglucose levels than in people with type 1 diabetes who had normal awareness. This suggeststhat people with impaired awareness can function effectively with very low blood glucoseconcentrations, at which symptoms and cognitive impairment would normally occur in nondiabeticand aware diabetic subjects. The potential risk of this situation is apparent: it is akinto walking along the edge of a cliff on a dark night. With such a narrow glycaemic warningzone the propensity to rapidly develop severe neuroglycopenia is high and the margin forerror is dangerously narrow.The results of these laboratory-based experimental studies of diabetic patients who haveestablished hypoglycaemia unawareness are consistent with clinical observations of peoplewith this acquired problem. At one moment they appear to be cerebrating normally (despitetheir blood glucose being low) then they rapidly become confused or drowsy, often with avacant or dazed appearance and an inertia to seek some form of carbohydrate to reverse theneuroglycopenia. They may have to rely on relatives, friends or colleagues to identify thehypoglycaemia and provide treatment. This becomes a serious emergency if the patient isalone or if the insidious, but often rapid, development of neuroglycopenia goes unobserved.This explains the increased risk of progression to severe hypoglycaemia, and the higher ratesreported in people with hypoglycaemia unawareness.Studies examining the effects of strict glycaemic control on symptomatic and counterregulatoryresponses to hypoglycaemia have also demonstrated a similar shift in glycaemicthreshold for autonomic symptoms and an acute sympatho-adrenal response. However, theeffect on glycaemic thresholds for neuroglycopenic symptoms and cognitive dysfunctionremains controversial (see Chapter 8).Peripheral Autonomic NeuropathyFor many years, peripheral autonomic neuropathy was considered to be the principalcause of impaired awareness of hypoglycaemia (Hoeldtke et al., 1982). This was basedon the assumption that the diminished secretion of epinephrine in response to hypoglycaemia(Hilsted et al., 1981; Bottini et al., 1997) would either prevent the generationof autonomic symptoms (such as sweating or a pounding heart) or reduce their intensity,resulting in an inability to perceive the onset of hypoglycaemia. Thus, autonomicneuropathy would interfere with the normal physiological responses stimulated by autonomicactivation.

PATHOGENESIS OF IMPAIRED AWARENESS OF HYPOGLYCAEMIA 153There are various reasons why this hypothesis is unlikely:• Although epinephrine can augment the intensity of a few autonomic symptoms of hypoglycaemia,it has a very limited role in their generation, which is modulated by sympatheticneural activation; the reduced secretory response of epinephrine in autonomic neuropathyis compensated by an increase in sensitivity of peripheral beta-adrenoceptors (Hilstedet al., 1987).• Diabetic subjects with autonomic neuropathy have normal physiological responses andexperience typical autonomic symptoms during hypoglycaemia (Hilsted et al., 1981;Hepburn et al., 1993b), and no relationship has been found between autonomic dysfunctionand hypoglycaemic symptoms (Berlin et al., 1987).• Impaired awareness of hypoglycaemia co-segregates with deficient counterregulatoryhormonal responses and not with autonomic neuropathy (Ryder et al., 1990).• The prevalence of autonomic neuropathy is similar in patients with type 1 diabetes oflong duration (more than 15 years), whether or not they have impaired awareness ofhypoglycaemia (Hepburn et al., 1990).• Although impaired awareness of hypoglycaemia is a major risk factor for severe hypoglycaemia,the latter is either no more common in type 1 diabetic patients with autonomicneuropathy (Bjork et al., 1990; The DCCT Research Group, 1991), or is only modestlyincreased (Stephenson et al., 1996).• Autonomic neuropathy is not a determinant of whether glycaemic thresholds for autonomic(including symptomatic) responses to hypoglycaemia are affected by antecedenthypoglycaemia (Dagogo-Jack et al., 1993).Both impaired awareness of hypoglycaemia and peripheral autonomic neuropathy arecommon in people with type 1 diabetes of long duration, and frequently coexist. This doesnot prove a causal relationship, and it would appear that peripheral autonomic dysfunctiondoes not have a prominent role in the pathogenesis of this syndrome.However, reduced sensitivity of cardiac beta-adrenoceptors to catecholamines has beenobserved in patients with type 1 diabetes who have impaired awareness of hypoglycaemia(Berlin et al., 1987). Hypoglycaemia per se reduces beta-adrenergic sensitivity in type 1diabetes (Fritsche et al., 1998), and this sensitivity is increased after avoidance of hypoglycaemiafor four months in people who have impaired awareness (Fritsche et al., 2001).The improved beta-adrenergic sensitivity correlated with a rise in autonomic symptomscores. Maladaptation of tissue sensitivity to catecholamines may therefore contribute tothe development of hypoglycaemia unawareness even though autonomic neuropathy is notpresent.Hypoglycaemia Associated Autonomic FailureThe co-segregation of impaired hypoglycaemia awareness with counterregulatory deficiencysuggests that they share a common underlying pathogenetic mechanism. These acquiredabnormalities associated with hypoglycaemia in type 1 diabetes (Box 7.3) are characterisedby a high frequency of severe hypoglycaemia and a common pathophysiological

154 IMPAIRED AWARENESS OF HYPOGLYCAEMIABox 7.3Acquired syndromes associated with hypoglycaemia in type 1 diabetes• Counterregulatory deficiency• Impaired hypoglycaemia awareness• Altered glycaemic thresholds for counterregulatory and symptomatic responsesFigure 7.6 Schematic diagram of the concept of hypoglycaemia associated autonomic failure(HAAF), based on Cryer (1992)feature, namely the elevated glycaemic thresholds (or lower blood glucose concentrations)that are required to trigger symptomatic and hormonal secretory responses. Inother words, more profound hypoglycaemia is necessary to produce the usual symptomaticand counterregulatory responses to acute hypoglycaemia. Cryer (1992) has designatedthis group of abnormalities as a form of ‘hypoglycaemia associated autonomicfailure’ (HAAF), and has speculated that recurrent severe hypoglycaemia may be theprimary problem which establishes a vicious circle (Figure 7.6). It seems likely that thisdefect resides within the central nervous system. The possible mechanisms underlyingHAAF and related hypoglycaemia syndromes in diabetes have been reviewed in detail(Cryer, 2005).

PATHOGENESIS OF IMPAIRED AWARENESS OF HYPOGLYCAEMIA 155Central Nervous System Adaptation to HypoglycaemiaSome people with insulin-treated diabetes remain lucid, with no evidence of impaired cognitivefunction, when their blood glucose is low (often well below 3.5 mmol/l). Biochemicalhypoglycaemia that is asymptomatic is commonly recorded by patients who have impairedawareness of hypoglycaemia, and they appear to have developed a neurological adaptationto chronic neuroglycopenia. The altered glycaemic threshold prevents the onset of warningsymptoms and cognitive dysfunction until the blood glucose falls to a dangerously low level,which is extremely undesirable when striving for safe clinical management of insulin-treateddiabetes.Although the human brain is dependent on a continuous supply of glucose for normalfunction, it can adapt to prolonged exposure to hypoglycaemia. This adaptation processtakes at least several hours and possibly a few days to occur. Short-term exposure toacute hypoglycaemia (blood glucose 2.5 mmol/l) for 60 minutes in non-diabetic subjectsshowed no improvement in cognitive function and no reduction in symptom scoresduring this brief time interval (Gold et al., 1995a). However, when non-diabetic subjectswere subjected to chronic hypoglycaemia (blood glucose 2.9 mmol/l) for 56 hours, usinga glucose clamp, significant cerebral adaptation did occur (Boyle et al., 1994). Theresponses to acute hypoglycaemia (blood glucose 2.5 mmol/l) were compared before andafter the period of chronic hypoglycaemia. Brain glucose uptake was initially reducedwhen blood glucose was below 3.6 mmol/l, but after a period of chronic hypoglycaemiauptake was preserved and cerebral function was maintained (Figure 7.7), demonstratingan effect of cerebral adaptation to chronic neuroglycopenia. The glycaemic thresholds forthe onset of symptoms, counterregulatory hormonal secretion and cognitive dysfunction,were all modified and occurred at much lower blood glucose concentrations. A similarphenomenon has been observed in non-diabetic patients who had an insulinoma causingchronic hypoglycaemia; symptomatic responses to acute hypoglycaemia were blunted andcounterregulatory hormonal responses were impaired but cognitive function was unaffected(Mitrakou et al., 1993). Surgical removal of the insulin-secreting tumour reversedthese abnormalities indicating that they had resulted from cerebral adaptation to chronichypoglycaemia.Strict glycaemic control in people with insulin-treated diabetes also alters the glycaemicthresholds for the development of counterregulatory hormones and symptoms (Chapter 8),so that a lower blood glucose concentration is required to trigger these responses. Theobservation that this requires a reduction in HbA 1c to within the non-diabetic range(Kinsley et al., 1995) suggests that the median daily blood glucose in these individualsis relatively low, and the frequency of biochemical (and symptomatic) hypoglycaemiawill be greater than in insulin-treated diabetic patients who are not as wellcontrolled (Thorsteinsson et al., 1986). Boyle et al. (1995) have shown that those patientswho had near normal HbA 1c values, maintained normal uptake of glucose by the brainduring hypoglycaemia, so preserving cerebral metabolism, reducing the counterregulatoryresponses to hypoglycaemia and diminishing symptomatic awareness. Although thiscapacity to maintain, and even increase, cerebral blood glucose uptake during hypoglycaemiais a protective response for the brain in these patients with strict glycaemiccontrol, it is considered to be maladaptive, because it suppresses the normal symptomaticwarning of responses and so risks the development of much more profoundneuroglycopenia.

156 IMPAIRED AWARENESS OF HYPOGLYCAEMIAFigure 7.7 Rates of brain glucose uptake, epinephrine concentration and total symptoms of hypoglycaemiain non-diabetic subjects before and after prolonged hypoglycaemia. Initial day of investigation(hatched); after 56 hours of hypoglycaemia (solid). ∗ Significant difference from baseline for each ofthe two days. Reproduced from Boyle (1997), Diabetologia, 40, S69–S74. With kind permission ofSpringer Science and Business MediaDuring hypoglycaemia, glucose transport into the brain becomes rate-limiting, and brainenergy metabolism deteriorates. The adaptive response results from an increased utilisationof glucose by the brain. In rodents, the transport of glucose across the blood–brain barrier isincreased after several days of chronic hypoglycaemia (McCall et al., 1986). Further studiesin rats of glucose transport activity across the blood–brain barrier have shown that whenblood glucose was kept below 2.0 mmol/l for several days, changes in expression of theglucose transporter, GLUT-1, in brain microvasculature occurred in response to the chronichypoglycaemia (Kumagai et al., 1995). This increase in GLUT-1 activity was responsiblefor the compensatory increase in glucose transport across the blood–brain barrier.

PATHOGENESIS OF IMPAIRED AWARENESS OF HYPOGLYCAEMIA 157Antecedent (Episodic) HypoglycaemiaIt has been recognised for many years that severe hypoglycaemia is associated with furtherepisodes of severe hypoglycaemia, and one episode may influence the clinical manifestationsof another occurring soon afterwards (Severinghaus, 1926). In recent years, several studieshave shown that the symptomatic and counterregulatory responses to an episode of acutehypoglycaemia are diminished if a preceding (or antecedent) episode of hypoglycaemiahas occurred within the previous 24 hours. Several studies have been performed in nondiabeticsubjects (Table 7.2) and in people with insulin-treated diabetes (Table 7.3). Althoughthese studies differ considerably in design and methods of inducing hypoglycaemia, ingeneral it appears that antecedent hypoglycaemia of between one and two hours durationhas a significant influence on the magnitude of the symptomatic and counterregulatoryresponses to subsequent hypoglycaemia occurring within the following 24 to 48 hours(Figure 7.8).The glycaemic thresholds for symptomatic and counterregulatory hormonal responses arealtered by antecedent hypoglycaemia, and the degree to which subsequent responses areblunted are determined by the duration and depth of antecedent hypoglycaemia (Davis et al.,1997). Some of the physiological responses (e.g. sweating) may be blunted for longer thanother responses following antecedent hypoglycaemia (George et al., 1995). Recurrent, shortlived(15 minutes) episodes of hypoglycaemia on four consecutive days, had no effect oncounterregulatory and symptomatic responses in non-diabetic subjects (Peters et al., 1995),and so transient reductions in blood glucose may not produce this effect. Davis et al. (2000)have observed that a short duration of antecedent hypoglycaemia (20 minutes to lowerand raise blood glucose from 3.9 to 2.9 mmol/l with the blood glucose being maintainedfor five minutes at 2.9 mmol/l) did not affect symptomatic awareness of hypoglycaemia,but did blunt the counterregulatory hormonal responses. Antecedent hypoglycaemia alsoreduces the counterregulatory responses to exercise on the following day, both in nondiabetic(Davis et al., 2000b) and type 1 diabetic subjects (Galassetti et al., 2003) andinfluences the metabolic responses, particularly diminishing endogenous glucose productionin response to exercise. This may promote exercise-induced hypoglycaemia in type 1diabetes.Some studies have examined the effect of antecedent hypoglycaemia on cognitive function,but in many the methods of assessment were inadequate and insufficient to provide definitiveevidence of a change in cognitive response. Although some maintain that the glycaemicthreshold for cognitive dysfunction is not altered by hypoglycaemia (see Chapter 8), anincreasing number of studies have suggested that this does shift to a lower blood glucoseconcentration in the same manner as the thresholds for autonomic and counterregulatoryresponses (Veneman et al., 1993; Ovalle et al., 1998; Fanelli et al., 1998). Consistentwith these observations, a study from Germany in non-diabetic men has shown thatafter a single episode of antecedent hypoglycaemia, subsequent hypoglycaemia had lesseffect on auditory-evoked brain potentials and short-term memory (Fruehwald-Schulteset al., 2000), demonstrating cerebral adaptation that preserves cognitive function. Nocturnal(episodic) hypoglycaemia, which is frequently not identified by patients, has been proposedas a mechanism for the induction of hypoglycaemia unawareness in people who giveno history of recurrent hypoglycaemia (Veneman et al., 1993). The possible mechanismsof cerebral adaptation causing impaired awareness of hypoglycaemia are summarised inBox 7.4.

Table 7.2 Studies of antecedent hypoglycaemia (AH) in non-diabetic humansReferences No. of subjectsMethod of induction,nadir BG (mmol/l) andduration of AHInterval before testhypoglycaemiaTest hypo: methodand BG nadir(mmol/l)Effect of AH on subsequent responsesto hypoglycaemiaHeller and Cryer(1991)Widom and Simonson(1992)9 Clamp (3.0) 2 h 18 h Clamp (2.8) Reduced symptoms and CR responses10 Clamp (2.2–2.8) 1 h 4 consecutive dailyhyposStepped clamp(2.2)Veneman et al. (1993) 10 iv. infusion (2.2–2.5) 2 h 7 h Stepped clamp(2.3)Mellman et al. (1994) 9 Clamp (3.2) 2 h 1.5 h Stepped clamp(2.8)Elevated BG thresholds for symptomsand CR responsesNocturnal hypo: BG thresholds forsymptoms and CR responseselevatedElevated BG thresholds for symptomsand CR responsesRobinson et al. (1995) 10 Clamp (3.0) 2 h 24 h and 6 days Clamp (2.5) Reduced adrenaline and sweating at 6daysPeters et al. (1995) 10 iv. bolus injection ofinsulin (< 28)< 15 min4 consecutive dailyhyposiv. bolus (< 28)X4No effect on symptoms or CRresponsesGeorge et al. (1995) 8 Clamp (2.9) 2 h 2 and 5 days Clamp (2.5) Physiological responses still reducedafter 1 weekDavis et al. (1997) 8 Stepped clamp (2.9) 2 h < 24 h Clamp (2.9) Magnitude of reduced CR responserelated to depth of AHDavis et al. (2000a) 31 Clamp (2.9) 2 h (n = 15);30 min (n = 16); 5 min(n = 10)24 h Clamp (2.9) Reduced CR responses in all studies.Symptom scores unaffected byshort duration hypoglycaemiaDavis et al. (2000b) 16 Clamp (2.9) 2 h 24 h Exercise (90 min) Blunting of CR responses to exercise.Attenuated metabolic responses(reduced endogenous glucoseproduction)Fruehwald-Schulteset al. (2000)30 Stepped clamp (2.6) 2.5 h 18–24 h Clamp (3.1) Auditory-evoked brain potentials andshort-term memory less affectedBG = blood glucose; AH = antecedent hypoglycaemia; CR = counterregulatory (hormonal); iv = intravenous

Table 7.3 Studies of antecedent hypoglycaemia in people with insulin-treated diabetesReference No. of subjectsMethod of induction,nadir BG (mmol/l)and duration of AHInterval before testhypoglycaemiaTest hypo: methodand BG nadir(mmol/l)Effect of AH on responses tohypoglycaemiaDavis et al.(1992)Dagogo-Jacket al. (1993)Lingenfelseret al. (1993)George et al.(1997)Ovalle et al.(1998)Fanelli et al.(1998)Galassettiet al. (2003)13 Clamp (3.0) 2 h 60 min Clamp (3.0) Reduced CR responses and hepaticglucose output26 (± AN) 12(non-diabetic)18 iv. bolus injectionof insulin (< 2.2)mins X 3Clamp (2.7) 2 h 15 h Stepped clamp(2.6)3 days Stepped clamp(1.7)8 Clamp (2.8) 2 h 2 days Stepped clamp(2.2)6 Clamp (2.8) 2 h,twice weekly for1 month15 Clamp (2.6) 3.5 h,nocturnal48 h Stepped clamp(2.5)5–10 h Stepped clamp(2.5)16 Clamp (2.9) 2 h 24 h Exercise for 90minElevated BG thresholds forsymptoms and autonomicresponsesReduced symptoms, CR andneurophysiological responsesPhysiological responses diminishedfor 2 daysReduced symptoms, CR andcognitive dysfunction. Fewersymptomatic episodes of clinicalhypoglycaemia.Less deterioration in cognitivefunction after nocturnalhypoglycaemia (versuseuglycaemia). Elevated BGthreshold for cognitivedysfunctionBlunting of CR responses toexercise. Three fold increase inglucose needAH = antecedent hypoglycaemia; AN = autonomic neuropathy; iv = intravenous; CR = counterregulatory (hormonal)

160 IMPAIRED AWARENESS OF HYPOGLYCAEMIAFigure 7.8 Schematic representation of the effect of antecedent hypoglycaemia on the neuroendocrineand symptomatic responses to subsequent hypoglycaemiaBox 7.4 Mechanisms of cerebral adaptation causing impaired awareness ofhypoglycaemiaSymptomatic and Neuroendocrine responses to hypoglycaemia in insulin-treateddiabetes are diminished in association with:• strict glycaemic control (HbA 1c in non-diabetic range)• antecedent (episodic) hypoglycaemia• chronic (protracted) hypoglycaemiaThey may be restored by:• relaxation of glycaemic control• scrupulous avoidance of hypoglycaemiaAlthough antecedent hypoglycaemia may induce transient impairment of awareness ofhypoglycaemia it is unclear how this mechanism would induce chronic or prolonged loss ofsymptomatic perception. Although frequent, recurrent hypoglycaemia may have a contributoryeffect to inducing hypoglycaemia unawareness, presumably the hypoglycaemia has tobe relatively protracted to induce prolonged cerebral adaptation, and the phenomenon is notlimited to patients who have strict glycaemic control. The problem remains of explainingthe induction of protracted or chronic hypoglycaemia unawareness, which often appears tobe a permanent defect. Presumably repetitive hypoglycaemic insults to the brain (which arenot necessarily severe) eventually ‘downregulate’ the central mechanisms that sense a lowblood glucose and activate the glucoregulatory responses within the hypothalamus. There isevidence of a permanent redistribution of regional cerebral blood flow in diabetic patientswith a history of recurrent severe hypoglycaemia (MacLeod et al., 1994a) with, in particular,a relative increase in blood flow to the frontal lobes. This may represent a chronic adaptive

IMPAIRED AWARENESS OF HYPOGLYCAEMIA 161response to protect vulnerable areas of the brain from recurrent, severe neuroglycopenia.However, a further study showed that the changes in regional cerebral blood flow in responseto controlled hypoglycaemia in patients with type 1 diabetes occurred independently of thestate of awareness of hypoglycaemia (MacLeod et al., 1996). The EEG changes associatedwith modest hypoglycaemia are more pronounced in patients with type 1 diabetes who haveimpaired awareness of hypoglycaemia (Tribl et al. 1996). Most studies suggest that a diffusefunctional abnormality is present in the anterior part of the brain in diabetes, and this may beimplicated in the impaired perception of hypoglycaemia. The pre-frontal areas of the cortexare closely connected to sub-cortical areas, and localised dysfunction could theoreticallyreduce the ability of the brain to perceive symptomatic hypoglycaemia.Further clues may be provided by neuroimaging studies. A study using positron emissiontomography (PET) compared changes in global and regional brain glucose metabolism duringeuglycaemia and hypoglycaemia in 12 men with type 1 diabetes, six of whom had impairedawareness of hypoglycaemia (Bingham et al., 2005). Brain glucose content was reducedby hypoglycaemia in both groups with a relative increase in tracer uptake on the prefrontalcortical regions. Differences between the groups were observed in glucose handling in regionsof the brain, and whereas the cerebral metabolic rate for glucose showed a relative rise in theaware subjects, it fell in the unaware subjects. Global neuronal activation was observed withhypoglycaemia in the aware patients, but was absent in the unaware, suggesting that corticalactivation is a necessary correlate of hypoglycaemia awareness. Further studies using PETscans or other forms of neuroimaging may identify regional differences in the response tohypoglycaemia within the brains of people with impaired awareness, which will help toelucidate the functional abnormalities associated with this syndrome.IMPAIRED AWARENESS OF HYPOGLYCAEMIA ANDLONG-TERM EFFECT ON COGNITIVE FUNCTIONImpaired awareness of hypoglycaemia is a major risk factor for severe hypoglycaemia, andpatients with the chronic form of this condition have a six-fold higher frequency (Gold et al.,1994). It is possible therefore that impaired hypoglycaemia awareness may be associatedwith evidence of a decline in cognitive function. Hepburn et al. (1991) noted that diabeticpatients with a history of impaired awareness of hypoglycaemia performed less well thanthose with normal awareness of hypoglycaemia on limited cognitive function testing, bothat a normal blood glucose and during hypoglycaemia. This suggested that an acquiredcognitive impairment may have been superimposed upon an increased susceptibility toneuroglycopenia. A modest, but insignificant, decline in intellectual function was notedwith progressive loss of hypoglycaemia awareness in a population study (MacLeod et al.,1994b).Formal measurement of cognitive function during controlled hypoglycaemia (bloodglucose 2.5 mmol/l) showed that patients with type 1 diabetes who had impaired hypoglycaemiaawareness exhibited more profound cognitive dysfunction during acute hypoglycaemiathan patients with normal awareness, and that this persisted for longer followingrecovery of blood glucose (Gold et al., 1995b). By contrast, a more recent study of peoplewith type 1 diabetes, which examined the rate of recovery of differing domains of cognitivefunction after hypoglycaemia, showed that during hypoglycaemia (blood glucose 2.5 mmol/l)cognitive function did not deteriorate in those with impaired awareness, suggesting that

162 IMPAIRED AWARENESS OF HYPOGLYCAEMIAcerebral adaptation to hypoglycaemia had occurred in these patients (Zammitt et al., 2005).These apparently discrepant results leave this issue unresolved.HUMAN INSULINFor 60 years after its discovery, insulin for therapeutic use was obtained from the pancreataof cattle and pigs. With the development of recombinant DNA technology it was possibleto ‘genetically engineer’ molecules and insulin was the first protein to be made in thisway, becoming available for the treatment of humans in the 1980s. Several of the existinganimal insulin formulations were withdrawn, principally for commercial reasons, and humaninsulin rapidly became the most commonly prescribed form of insulin. The structure ofhuman insulin differs from porcine insulin by a single amino acid and from bovine insulinby three amino acids. In initial trials it was not expected that human insulin would differsubstantially in potency from animal insulin, but because human insulin was slightly purerthan some of the animal insulins, patients were advised to reduce the dose by around 10%when converting from animal to human insulin. Detailed pharmacokinetic studies comparinghuman and animal insulins did not demonstrate any major differences, but human insulin hasa slightly faster onset of action, a slightly shorter duration of action, and is less immunogenicthan equivalent animal insulins. Most clinical studies, conducted on a worldwide scale,showed no significant differences between human and animal insulins in their clinicalapplication.In Switzerland, however, one group of clinicians reported encountering serious clinicalproblems with the use of human insulin in patients with type 1 diabetes (Teuscher and Berger,1987). In particular, they claimed that patients experienced more frequent hypoglycaemiawith human insulin, and that warning symptoms were modified by human insulin, as a resultof which many patients were unable to detect the onset of hypoglycaemia. A pathologist in theUnited Kingdom then claimed that the number of patients dying from severe hypoglycaemiahad increased since the introduction of human insulin (see Chapter 12). The evidence forthis irresponsible statement did not withstand scrutiny, but in the UK anecdotal reportsemerged of problems experienced by patients with human insulin, and solicitors acting onbehalf of over 400 patients tried to bring a legal action against the insulin manufacturers,alleging that human insulin gave less warning of hypoglycaemia. Additional claims includedallegations that human insulin may have caused personality changes in individuals and evenother disease states such as multiple sclerosis. This group action was abandoned in 1993because of the lack of robust scientific evidence for these claims.However, this issue generated much controversy and heated debate and stimulated severalstudies comparing human with animal insulins, which are not reviewed here. A systematicreview of the extensive literature on this topic examined whether published evidencesupported a difference in the frequency and awareness of hypoglycaemia induced by humanand animal insulins (Airey et al., 2000). A total of 52 randomized controlled trials wereidentified, 37 of which were of double-blind design, whereas others reported hypoglycemicoutcomes as a secondary or incidental outcome during comparative investigations of efficacyor immunogenicity. Seven of the double-blind studies reported differences in frequency ofhypoglycaemia or of symptomatic awareness, and four of the unblinded trials reported differencesin hypoglycaemia. None of the four population time trend studies found any relationshipbetween the increasing use of human insulin and hospital admission for hypoglycaemia

TREATMENT STRATEGIES 163or unexplained death among people with insulin-treated diabetes. The authors observed thatthe least rigorous scientific studies gave the greatest support to the premise that treatmentwith human insulin influences the frequency, severity or symptoms of hypoglycaemia. Thereport concluded that the published evidence did not support the contention that humaninsulin is responsible either for the alleged problems with impaired awareness of hypoglycaemiaor for a higher risk of severe hypoglycaemia, but were unable to decide whether someof the phenomena associated with the use of human insulin resulted from stricter glycaemiccontrol (Airey et al., 2000).In clinical practice there are undoubtedly a small number of people with insulin-treateddiabetes in whom the use of human insulin has not been satisfactory, being associated withfrequent and unpredictable hypoglycaemia and a diminished sense of well-being. Whetherthis is related to the different pharmacokinetics of human insulin or represents an idiosyncraticresponse in individuals is unclear, but such patients clearly wish to retain the option touse animal species of insulin, and it is to be hoped that the availability of the animal speciesof insulins will be maintained. No problem with symptomatic awareness of hypoglycaemiahas been reported with the use of the newer insulin analogues.TREATMENT STRATEGIESWhen impaired awareness of hypoglycaemia is therapy-related, that is, resulting from strictglycaemic control, the approach to management is relatively simple. The total insulin doseshould be reduced, attention paid to the appropriateness of the insulin regimen, and overallglycaemic control should be relaxed. Liu et al. (1996) reported an improvement in symptomaticand counterregulatory hormonal responses to hypoglycaemia after three months ofless strict glycaemic control in a small group of insulin-treated patients, in whom the meanHbA 1c rose from 6.9% to 8.0%.It has been claimed that impaired awareness of hypoglycaemia (and to some extent counterregulatoryhormonal deficiency) can be reversed by scrupulous avoidance of hypoglycaemiathrough meticulous attention to diabetic management (Cranston et al., 1994; Dagogo-Jacket al., 1994; Fanelli et al., 1994). The effect that this had on glycaemic thresholds for cognitivedysfunction and the recovery of counterregulatory hormonal secretion to hypoglycaemiadiffered between these studies, but all demonstrated an improved symptomatic responsefollowing avoidance of hypoglycaemia for periods varying from three weeks to one year.However, the studies can be criticised for the following reasons:• Only a small number of patients were studied.• The definition of hypoglycaemia unawareness was based on an increased frequency ofasymptomatic biochemical hypoglycaemia, and with the exception of the study by Dagogo-Jack et al. (1994), was not based on having a history of hypoglycaemia unawareness.• In all studies there was a small but definite rise in glycated haemoglobin, suggesting thatthe improved symptomatic awareness was related primarily to relaxation of glycaemiccontrol.Although the scrupulous avoidance of hypoglycaemia is clearly desirable, and may bebeneficial to reducing the severity of hypoglycaemia unawareness, it is very difficult to

164 IMPAIRED AWARENESS OF HYPOGLYCAEMIAFigure 7.9 Augmentation of the normal secretory response of epinephrine (adrenaline) to, andawareness of, acute hypoglycaemia (blood glucose 2.8 mmol/l) by the prior ingestion of caffeine ininsulin-treated diabetic patients. Derived from data in Debrah et al. (1996)achieve as it is extremely time-consuming and labour-intensive both for patients and healthprofessionals. The use of continuous subcutaneous insulin infusion overnight instead ofisophane (NPH) insulin at bedtime has been shown to be beneficial in diabetic patientswith impaired awareness of hypoglycaemia, improving warning symptoms and counterregulatoryresponses to hypoglycaemia, presumably by reducing the frequency of nocturnalhypoglycaemia (Kanc et al., 1998).The ingestion of caffeine uncouples the relationship between cerebral blood flow andglucose utilisation via antagonism of adenosine receptors, causing relative neuroglycopeniaand earlier release of counterregulatory hormones during moderate hypoglycaemia. Theprior consumption of caffeine augments the symptomatic and counterregulatory hormonalresponses to a modest reduction of blood glucose in non-diabetic subjects (Kerr et al., 1993),and a similar phenomenon occurs in people with type 1 diabetes following the ingestion of adose of caffeine equivalent to two or three cups of coffee (Debrah et al., 1996). The reductionin cerebral blood flow is sustained, the counterregulatory response is augmented (Figure 7.9)and greater awareness of hypoglycaemia occurs (see Chapter 5). This raises the prospectof identifying some form of therapeutic intervention, which utilises a similar mechanism toheighten the residual symptomatic response in people with type 1 diabetes who have impairedawareness of hypoglycaemia. The adenosine-receptor antagonist, theophylline, stimulatesthe secretion of catecholamines and reduces cerebral blood flow, and a single intravenousdose has been shown to enhance counterregulatory hormone responses to hypoglycaemiaand partially restore perception of hypoglycaemic symptoms in patients with type 1 diabeteswith impaired awareness of hypoglycaemia (de Galan et al., 2002). Glycaemic thresholdsfor haemodynamic and symptomatic responses were restored to normal. It is not knownwhether oral theophylline would be as effective, and whether the effects can be sustained,as the development of tolerance to these drugs is common.It is clearly desirable to avoid severe hypoglycaemia at all costs, and treatment strategiesshould be adopted to achieve this aim (Box 7.5). Frequent blood glucose monitoring isessential in affected patients, and may require occasional nocturnal measurements to detect

CONCLUSIONS 165Box 7.5 Treatment strategies for patients with impaired awareness ofhypoglycaemia• Frequent blood glucose monitoring (including nocturnal measurements).• Avoid blood glucose values < 40 mmol/l.• Set target range of blood glucose higher than for ‘aware’ patients (e.g. preprandialbetween 6.0–12.0 mmol/l; bedtime > 80 mmol/l)• Avoid HbA 1c in non-diabetic range.• Use predominantly short-acting insulins (e.g. basal-bolus regimen; insulinanalogues).• Regular snacks between meals and at bedtime, containing unrefined carbohydrate.• Appropriate additional carbohydrate consumption and/or insulin dose adjustmentfor premeditated exercise.• Learn to identify subtle neuroglycopenic cues to low blood glucose.low blood glucose during the night. Blood glucose awareness training has been developedin the USA, with re-education of affected patients to recognise neuroglycopenic cues (Coxet al., 1995), but this also requires facilities and resources that are not available in mostcentres. Intensive insulin therapy is contraindicated in patients who have impaired awarenessof hypoglycaemia and treatment goals have to be considered individually. The avoidanceof severe hypoglycaemia is paramount as this may exacerbate the problem, and the use ofmostly short-acting insulin (and possibly insulin analogues) in basal-bolus regimens maybe particularly useful in avoiding biochemical and symptomatic hypoglycaemia withoutcompromising overall glycaemic control.CONCLUSIONS• An inadequate symptomatic warning to hypoglycaemia is common in people withinsulin-treated diabetes and is described as impaired awareness of hypoglycaemia orhypoglycaemia unawareness. It increases in prevalence with duration of insulin-treateddiabetes.• In people who report impaired awareness of hypoglycaemia, asymptomatic hypoglycaemiaoccurs more frequently during routine blood glucose monitoring. This may alert theclinician to the possibility that an individual is developing this problem.• Impaired awareness of hypoglycaemia may be associated with strict glycaemic control;significant modification of the symptomatic response occurs when the HbA 1c concentrationis within the non-diabetic range.

166 IMPAIRED AWARENESS OF HYPOGLYCAEMIA• The mechanisms underlying impaired awareness of hypoglycaemia may be multifactorial.Possible mechanisms include chronic exposure to a low blood glucose, antecedenthypoglycaemia, and central autonomic and glucoregulatory failure.• Antecedent hypoglycaemia has a significant influence on the magnitude of the symptomaticand counterregulatory responses to subsequent hypoglycaemia occurring withinthe following 48 hours.• When impaired awareness of hypoglycaemia results from strict glycaemic control, thetotal insulin dose should be reduced, attention paid to the suitability of the insulin regimen,and overall glycaemic control should be relaxed.• Impaired awareness of hypoglycaemia, and to some extent counterregulatory hormonaldeficiency, can probably be reversed by scrupulous avoidance of hypoglycaemia throughmeticulous attention to diabetic management.• Intensive insulin therapy is contraindicated in patients who have impaired awareness ofhypoglycaemia. The use of mostly short-acting insulins (including insulin analogues) inbasal-bolus regimens may be particularly useful in avoiding biochemical and symptomatichypoglycaemia without compromising overall glycaemic control.REFERENCESAirey CM, Williams DRR, Martin PG, Bennett, CMT, Spoor PA (2000). Hypoglycaemia induced byexogenous insulin – ‘human’ and animal insulin compared. Diabetic Medicine 17: 416–32.Bacatselos SO, Karamitsos DT, Kourtoglou GI, Zambulis CX, Yovos JG, Vyzantiadis AT (1995).Hypoglycaemia unawareness in type 1 diabetic patients under conventional insulin treatment.Diabetes, Nutrition and Metabolism 8: 267–75.Berlin I, Grimaldi A, Payan C, Sachon C, Bosquet F, Thervet F, Puech AJ (1987). Hypoglycemicsymptoms and decreased B-adrenergic sensitivity in insulin-dependent diabetic patients. DiabetesCare 10: 742–7.Bingham EM, Dunn JT, Smith D, Sutcliffe-Goulden J, Reed LJ, Marsden PK, Amiel SA (2005).Differential changes in brain glucose metabolism during hypoglycaemia accompany loss of hypoglycaemiaawareness in men with type 1 diabetes mellitus. An ( 11 C)-3-O-methyl-D-glucose PETstudy. Diabetologia 48: 2080–9.Bjork E, Palmer M, Schvarcz E, Berne C (1990). Incidence of severe hypoglycaemia in an unselectedpopulation of patients with insulin-treated diabetes mellitus, with special reference to autonomicneuropathy. Diabetes, Nutrition and Metabolism 4: 303–9.Bottini P, Boschetti E, Pampanelli S, Ciofetta M, Del Sindaco P, Scionti L et al. (1997). Contribution ofautonomic neuropathy to reduced plasma adrenaline responses to hypoglycemia in IDDM. Evidencefor a nonselective defect. Diabetes 46: 814–23.Boyle PJ, Schwartz NS, Shah SD, Clutter WE, Cryer PE (1988). Plasma glucose concentrations at theonset of hypoglycemic symptoms in patients with poorly controlled diabetes and in non-diabetics.New England Journal of Medicine 318: 1487–92.Boyle PJ, Nagy RJ, O’Connor AM, Kempers SF, Yeo RA, Qualls C (1994). Adaptation in brainglucose uptake following recurrent hypoglycemia. Proceedings of National Academy of Science,USA 91: 9352–6.Boyle PJ, Kempers SF, O’Connor AM, Nagy RJ (1995). Brain glucose uptake and unawarenessof hypoglycemia in patients with insulin-dependent diabetes mellitus. New England Journal ofMedicine 333: 1726–31.

REFERENCES 167Boyle PJ (1997). Alteration in brain glucose metabolism induced by hypoglycaemia in man,Diabetologia, 40, Suppl. S69–S74.Clarke WL, Cox DJ, Gonder-Frederick LA, Julian D, Schlundt D, Polonsky W (1995). Reducedawareness of hypoglycemia in adults with IDDM. A prospective study of hypoglycemic frequencyand associated symptoms. Diabetes Care 18: 517–22.Cox D, Gonder-Frederick L, Polonsky W, Schlundt D, Julian D, Clarke W (1995). A multicenterevaluation of blood glucose awareness training – II. Diabetes Care 18: 523–8.Cranston I, Lomas J, Maran A, Macdonald I, Amiel SA (1994). Restoration of hypoglycaemia awarenessin patients with long-duration insulin-dependent diabetes. Lancet 344: 283–7.Cryer PE (1992). Iatrogenic hypoglycemia as a cause of hypoglycemia-associated autonomic failurein IDDM. A vicious circle. Diabetes 41: 255–60.Cryer P and Frier BM (2004). Hypoglycaemia, International Textbook of Diabetes, 3rd edn,Ferrannini E., Zimmet P., De Fronzo R. and Keen H. (eds), 1079–100.Cryer PE (2005). Mechanisms of hypoglycemia-associated autonomic failure and its componentsyndromes in diabetes. Diabetes 54: 3592–601.Dagogo-Jack SE, Craft S, Cryer PE (1993). Hypoglycemia-associated autonomic failure in insulindependentdiabetes mellitus. Journal of Clinical Investigation 91: 819–28.Dagogo-Jack S, Rattarasarn C, Cryer PE (1994). Reversal of hypoglycemia unawareness, but notdefective glucose counterregulation, in IDDM. Diabetes 43: 1426–34.Davis MR, Mellman M, Shamoon H (1992). Further defects in counterregulatory responses inducedby recurrent hypoglycemia in IDDM. Diabetes 41: 1335–40.Davis SN, Shevers C, Mosqueda-Garcia R, Costa F (1997). Effects of differing antecedent hypoglycemiaon subsequent counterregulation in normal humans. Diabetes 46: 1328–35.Davis SN, Mann S, Galassetti P, Neill RA, Tate D, Ertl AC, Costa F (2000a). Effects of differingdurations of antecedent hypoglycemia on counterregulatory responses to subsequent hypoglycemiain normal humans. Diabetes 49: 1897–903.Davis SN, Galassetti P, Wasserman DH, Tate D (2000b). Effects of antecedent hypoglycemia onsubsequent counterregulatory responses to exercise. Diabetes 49: 73–81.de Galan BE, Tack CJ, Lenders JW, Pasman JW, Elving LD, Russel FG et al. (2002). Theophyllineimproves hypoglycemia unawareness in type 1 diabetes. Diabetes 51: 790–6.Debrah K, Sherwin RS, Murphy J, Kerr D (1996). Effect of caffeine on recognition of and physiologicalresponses to hypoglycaemia in insulin dependent diabetes. Lancet 347: 19–24.Fanelli C, Pampanelli S, Epifano L, Rambotti AM, Di Vincenzo A, Modarelli F et al. (1994). Long-termrecovery from unawareness, deficient counterregulation and lack of cognitive dysfunction duringhypoglycaemia, following institution of rational, intensive insulin therapy in IDDM. Diabetologia37: 1265–76.Fanelli CG, Paramore DS, Hershey T, Terkamp C, Ovalle F, Craft S, Cryer PE (1998). Impact of nocturnalhypoglycemia on hypoglycemic cognitive dysfunction in type 1 diabetes. Diabetes 47: 1920–7.Fritsche A, Stumvoll M, Grub M, Sieslack S, Renn W, Schmulling RM et al. (1998). Effect ofhypoglycemia on beta-adrenergic sensitivity in normal and type 1 diabetic subjects. Diabetes Care21: 1505–10.Fritsche A, Stefan N, Haring H, Gerich J, Stumvoll M (2001). Avoidance of hypoglycemia awarenessby increasing beta-adrenergic sensitivity in type 1 diabetes. Annals of Internal Medicine 134: 729–36.Fruehwald-Schultes B, Born J, Kern W, Peters A, Fehm HL (2000). Adaptation of cognitive functionto hypoglycemia in healthy men. Diabetes Care 23: 1059–66.Galassetti P, Tate D, Neill RA, Morrey S, Wasserman DH, Davis SN (2003). Effect of antecedenthypoglycemia on counterregulatory responses to subsequent euglycemic exercise in type 1 diabetes.Diabetes 52: 1761–9.George E, Harris N, Bedford C, Macdonald IA, Hardisty CA, Heller SR (1995). Prolonged but partialimpairment of the hypoglycaemic physiological response following short-term hypoglycaemia innormal subjects. Diabetologia 38: 1183–90.

168 IMPAIRED AWARENESS OF HYPOGLYCAEMIAGeorge E, Marques JL, Harris ND, Macdonald IA, Hardisty CA, Heller SR (1997). Preservation ofphysiological responses to hypoglycemia 2 days after antecedent hypoglycemia in patients withIDDM. Diabetes Care 20: 1293–8.Gold AE, MacLeod KM, Frier BM (1994). Frequency of severe hypoglycemia in patients with type 1diabetes with impaired awareness of hypoglycemia. Diabetes Care 17: 697–703.Gold AE, Deary IJ, MacLeod KM, Thomson KJ, Frier BM (1995a). Cognitive function during insulininducedhypoglycemia in humans: short-term cerebral adaptation does not occur. Psychopharmacology119: 325–33.Gold AE, MacLeod KM, Deary IJ, Frier BM (1995b). Hypoglycemia-induced cognitive dysfunctionin diabetes mellitus: effect of hypoglycemia unawareness. Physiology and Behavior 58: 501–11.Grimaldi A, Bosquet F, Davidoff P, Digy JP, Sachon C, Landault C et al. (1990). Unawareness ofhypoglycemia by insulin-dependent diabetics. Hormone and Metabolic Research 22: 90–5.Heller SR, Cryer PE (1991). Reduced neuroendocrine and symptomatic responses to subsequenthypoglycemia after 1 episode of hypoglycemia in non-diabetic humans. Diabetes 40: 223–6.Heller, SR, Chapman J, McCloud J, Ward J (1995). Unreliability of reports of hypoglycaemia bydiabetic patients. British Diabetic Journal 310: 440.Henderson JN, Allen KV, Deary IJ, Frier BM (2003). Hypoglycaemia in insulin-treated type 2 diabetes:frequency, symptoms and impaired awareness. Diabetic Medicine 20: 1016–21.Hepburn DA, Patrick AW, Eadington DW, Ewing DJ, Frier BM (1990). Unawareness of hypoglycaemiain insulin-treated diabetic patients: prevalence and relationship to autonomic neuropathy. DiabeticMedicine 7: 711–7.Hepburn DA, Patrick AW, Brash HM. Thomson I, Frier BM (1991). Hypoglycaemia unawarenessin type 1 diabetes: a lower plasma glucose is required to stimulate sympatho-adrenal activation.Diabetic Medicine 8: 934–45.Hepburn DA, Deary IJ, Frier BM (1992). Classification of symptoms of hypoglycaemia in insulintreateddiabetic patients using factor analysis: relationship to hypoglycaemia unawareness. DiabeticMedicine 9: 70–5.Hepburn DA, MacLeod KM, Pell ACH, Scougal IJ, Frier BM (1993a). Frequency and symptoms ofhypoglycaemia experienced by patients with type 2 diabetes treated with insulin. Diabetic Medicine10: 231–7.Hepburn DA, MacLeod KM, Frier BM (1993b). Physiological, symptomatic and hormonal responsesto acute hypoglycaemia in type 1 diabetic patients with autonomic neuropathy. Diabetic Medicine10: 940–9.Hilsted J, Madsbad S, Krarup T, Sestoft L, Christensen NJ, Tronier B, Galbo H (1981). Hormonal,metabolic, and cardiovascular responses to hypoglycemia in diabetic autonomic neuropathy.Diabetes 30: 626–33.Hilsted J, Richter E, Madsbad S, Tronier B, Christensen NJ, Hildebrandt P et al. (1987). Metabolic andcardiovascular responses to epinephrine in diabetic autonomic neuropathy. New England Journal ofMedicine 317: 421–6.Hoeldtke RD, Boden G, Shuman CR, Owen OE (1982). Reduced epinephrine secretion and hypoglycemiaunawareness in diabetic autonomic neuropathy. Annals of Internal Medicine 96: 459–62.Jorgensen HV, Pedersen-Bjergaard U, Rasmussen AK, Borch-Johnsen K (2003). The impact of severehypoglycemia and impaired awareness of hypoglycemia on relatives of patients with type 1 diabetes.Diabetes Care 26: 1106–9.Joslin EP, Gray H, Root HF. (1922). Insulin in hospital and home. Journal of Metabolic Research2: 651–99.Kanc K, Janssen MMJ, Keulen ETP, Jacobs MAJM, Popp-Snijders C, Snoek FJ, Heine RJ (1998).Substitution of night-time continuous subcutaneous insulin infusion therapy for bedtime NPH insulinin a multiple injection regimen improves counterregulatory hormonal responses and warning symptomsof hypoglycaemia in IDDM. Diabetologia 41: 322–9.

REFERENCES 169Kerr D, Sherwin RS, Pavalkis F, Fayad P, Sikorski L, Rife F et al. (1993). Effect of caffeine on therecognition of and responses to hypoglycemia in humans. Annals of Internal Medicine 119: 799–804.Kinsley BT, Widom B, Simonson DC (1995). Differential regulation of counterregulatory hormonesecretion and symptoms during hypoglycemia in IDDM. Effect of glycemic control. Diabetes Care18: 17–26.Kubiak T, Hermanns HJ, Schreckling HJ, Kulzer B, Haak T (2004). Assessment of hypoglycaemiaawareness using continuous glucose monitoring. Diabetic Medicine 21: 487–90.Kumagai AK, Kang Y-S, Boado RJ, Pardridge WM (1995). Upregulation of blood–brain barrierGLUT 1 glucose transporter protein and mRNA in experimental chronic hypoglycemia. Diabetes44: 1399–404.Lawrence RD (1941). Insulin hypoglycaemia. Changes in nervous manifestations. Lancet ii: 602.Lingenfelser T, Renn W, Sommerwerck U, Jung MF, Buettner UW, Zaiser-Kaschel H et al. (1993).Compromised hormonal counterregulation, symptom awareness, and neurophysiological functionafter recurrent short-term episodes of insulin-induced hypoglycemia in IDDM patients. Diabetes42: 610–8.Liu D, McManus RM, Ryan EA (1996). Improved counter-regulatory hormonal and symptomaticresponses to hypoglycemia in patients with insulin-dependent diabetes mellitus after 3 months ofless strict glycemic control. Clinical and Investigative Medicine 19: 71–82.MacLeod KM, Hepburn DA, Deary IJ, Goodwin GM, Dougall N, Ebmeier KP, Frier BM (1994a).Regional cerebral blood flow in IDDM patients: effects of diabetes and of recurrent severe hypoglycaemia.Diabetologia 37: 257–63.MacLeod KM, Deary IJ, Graham KS, Hepburn DA, Frier BM (1994b). Hypoglycaemia unawareness inadult patients with type 1 diabetes: relationship to severe hypoglycaemia and cognitive impairment.Diabetes, Nutrition and Metabolism 7: 205–12.MacLeod KM, Gold AE, Ebmeier KP, Hepburn DA, Deary, IJ, Goodwin GM, Frier BM (1996). Theeffects of acute hypoglycemia on relative cerebral blood flow distribution in patients with type 1(insulin-dependent) diabetes and impaired hypoglycemia awareness. Metabolism 45:974–80.McCall AL, Fixman LB, Fleming N, Tornheim K, Chick W, Ruderman NB (1986). Chronic hypoglycaemiaincreases brain glucose transport. American Journal of Endocrinology 251: E442–7.Maddock RK, Krall LP (1953). Insulin reactions. Manifestations and need for recognition of longactinginsulin reactions. Archives of Internal Medicine 91: 695–703.Mellman MJ, Davis MR, Brisman M, Shamoon H (1994). Effect of antecedent hypoglycemia oncognitive function and on glycemic thresholds for counterregulatory hormone secretion in healthyhumans. Diabetes Care 17: 183–8.Mitrakou A, Ryan C, Veneman T, Mokan M, Jenssen T, Kiss I et al. (1991). Hierarchy of glycemicthresholds for counterregulatory hormone secretion, symptoms and cerebral dysfunction. AmericanJournal of Endocrinology 260: E67–74.Mitrakou A, Fanelli C, Veneman T, Perriello G, Calderone S, Platanisiotis D et al. (1993). Reversibilityof unawareness of hypoglycemia in patients with insulinomas. New England Journal of Medicine329: 834–9.Mokan M, Mitrakou A, Veneman T, Ryan C, Korytkowski M, Cryer P, Gerich J (1994). Hypoglycemiaunawareness in IDDM. Diabetes Care 17: 1397–403.Muhlhauser I, Heirnemann L, Fritsche E, von Lennep K, Berger M (1991). Hypoglycemic symptomsand frequency of severe hypoglycemia in patients treated with human and animal insulinpreparations. Diabetes Care 14: 745–9.Orchard TJ, Maser RE, Becker DJ, Dorman JS, Drash AL (1991). Human insulin use and hypoglycaemia:insights from the Pittsburgh Epidemiology of Diabetes Complications Study. DiabeticMedicine 8: 469–74.Ovalle F, Fanelli CG, Poramore DS, Hersley T, Craft S, Cryer PE (1998). Brief twice-weekly episodesof hypoglycemia reduce detection of clinical hypoglycemia in type 1 diabetes mellitus. Diabetes47: 1472–9.

170 IMPAIRED AWARENESS OF HYPOGLYCAEMIAPampanelli S, Fanelli C, Lalli C, Ciofetta M, Del Sindorco P, Lepore M et al. (1996). Long-termintensive insulin therapy in IDDM: effects on HbA 1c , risk for severe and mild hypoglycaemia, statusof counterregulation and awareness of hypoglycaemia. Diabetologia 39: 677–86.Peters A, Rohloff F, Kerner W (1995). Preserved counterregulatory hormone release and symptomsafter short term hypoglycemic episodes in normal men. Journal of Clinical Endocrinology andMetabolism 80: 2894–8.Pramming S, Thorsteinsson B, Bendtson I, Binder C (1991). Symptomatic hypoglycaemia in 411 type1 diabetic patients. Diabetic Medicine 8: 217–22.Robinson AM, Parkin HM, Macdonald IA, Tattersall RB (1995). Antecedent hypoglycaemia in nondiabeticsubjects reduces the adrenaline response for 6 days but does not affect the catecholamineresponse to other stimuli. Clinical Science 89: 359–66.Ryder REJ, Owens DR, Hayes TM, Ghatei MA, Bloom SR (1990). Unawareness of hypoglycaemia andinadequate hypoglycaemic counterregulation: no causal relation with diabetic autonomic neuropathy.British Medical Journal 301: 783–7.Severinghaus EL (1926). Hypoglycemic coma due to repeated insulin overdosage. American Journalof Medical Sciences 172: 573–80.Stephenson JM, Kempler P, Cavallo Peria P, Fuller JH, EURODIAB IDDM Complications StudyGroup (1996). Is autonomic neuropathy a risk factor for severe hypoglycaemia? The EURODIABIDDM Complications Study. Diabetologia 39: 1372–6.Sussman KE, Crout JR, Marble A (1963). Failure of warning in insulin-induced hypoglycemic reactions.Diabetes 12: 38–45.Teuscher A, Berger WG (1987). Hypoglycaemia unawareness in diabetics transferred from beef/porcineinsulin to human insulin. Lancet ii: 382–5.The DCCT Research Group (1991). Epidemiology of severe hypoglycemia in the Diabetes Controland Complications Trial. American Journal of Medicine 90: 450–9.The Diabetes Control and Complications Trial Research Group (1993). The effect of intensive treatmentof diabetes on the development and progression of long-term complications in insulin-dependentdiabetes mellitus. The New England Journal of Medicine 329: 977–86.Thorsteinsson B, Pramming S, Lauritzen T, Binder C (1986). Frequency of daytime biochemical hypoglycaemiain insulin-treated diabetic patients: relation to daily median blood glucose concentrations.Diabetic Medicine 3: 147–51.Tribl G, Howorka K, Heger G, Anderer P, Thoma H, Zeitlhofer J (1996). EEG topography duringinsulin-induced hypoglycemia in patients with insulin-dependent diabetes mellitus. EuropeanNeurology 36: 303–9.Vea H, Jorde R, Sager G, Vaaler S, Sundsfjord J (1992). Reproducibility of glycaemic thresholdsfor activation of counterregulatory hormones and hypoglycaemic symptoms in healthy subjects.Diabetologia 35: 958–61.Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J (1993). Induction of hypoglycemia unawarenessby asymptomatic nocturnal hypoglycemia. Diabetes 42: 1233–7.Widom B, Simonson DC (1992). Intermittent hypoglycemia impairs glucose counterregulation.Diabetes 41: 1335–40.Zammitt NN, Warren RE, Deary IJ, Frier BM (2005). Recovery of cognitive function after insulininducedhypoglycaemia in people with type 1 diabetes with either normal or impaired awareness ofhypoglycaemia. Diabetologia 48 (Suppl 1): A293 (abstract).

8 Risks of Strict GlycaemicControlStephanie A. AmielINTRODUCTIONThe benefit of strict glycaemic control in diminishing the risks of the development of longtermcomplications of diabetes is beyond doubt, but the negative aspects of such therapiesneed to be considered, and their risks identified, understood and minimised. Modern intensifiedinsulin management need not necessarily increase the risk of iatrogenic problems andcan deliver better glycaemic control more safely than in the past, although substantial scoperemains for improvement. New drugs for type 2 diabetes may offer greater opportunities toachieve near-normoglycaemia but may also bring new risks. These risks need to be explainedcarefully to every patient, who can then make an individual, informed choice about themanagement of their diabetes.The risks of intensified insulin therapy, the focus of this chapter, are those of insulinitself – intensified. Thus the major side-effects are weight gain (The Diabetes Control andComplications Trial Research Group, 1988) and hypoglycaemia (The Diabetes Control andComplications Trial Research Group, 1993; 1995a; 1997). Both of these problems mayappear to be minimised with modern strategies for patient self-management, at least inpublished studies (Jorgens et al., 1993; DAFNE Study Group, 2002; Plank 2004 et al.;Samann et al. 2005), yet they remain serious issues for large numbers of people. Weight gain,attributed primarily to the resolution of caloric loss in glycosuria (Carlson and Campbell,2003), is theoretically responsive to dietary strategies, but insulin and peripheral insulinsensitizers do cause lipogenesis and fluid retention, both of which contribute to a rise inweight that may be unacceptable to patients. Evidence is accumulating about the potentialeffects of insulin and other anti-diabetic agents on appetite control and satiety that maymake the control of weight more difficult. Although the long-term diminution of risk ofvascular complications is now established beyond doubt, the sudden institution of strictglycaemic control after a prolonged period of hyperglycaemia, can transiently, but sometimesseriously, destabilise microvascular disease (Agardh et al., 1992; The Diabetes Control andComplications Trial Research Group, 1998). In type 1 diabetes, the long-term follow up ofthe DCCT cohorts unequivocally has extended the evidence to include slowed progressionof macrovascular, as well as microvascular, disease (Nathan et al., 2003; Writing team forthe Diabetes Control and Complications Trial/Epidemiology of Diabetes Intervention andComplications Research Group, 2002), and so the risks of intensified therapy need to bebalanced against the potentially large gains. A new risk, of unknown magnitude, is theincreasing use of novel insulins, which have different properties from endogenous humaninsulin and thus, at least in theory, may have different side-effects.Hypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

172 RISKS OF STRICT GLYCAEMIC CONTROLWhen cohorts of patients are studied rather than individuals, other potential risks ofthe demands of intensified insulin therapies, and in particular the inherent psychosocialstrains, do not emerge as a problem. Concerns have been expressed that greater use ofhome blood glucose monitoring may increase anxiety, particularly in type 2 patients whoare not taking insulin, and who have limited means at their disposal of responding tohigh blood glucose values (Franciosi et al., 2001). In contrast, in type 1 diabetes, researchsuggests that patients may prefer intensified treatment regimens. In the Diabetes Controland Complications Trial, patients in the intensive treatment arm of the study had an overallimprovement in their subjective feelings of control and well-being, although it was offset bya greater fear of hypoglycaemia (The Diabetes Control and Complications Trial ResearchGroup, 1996a). However, this balance may be particularly positive in people who activelychoose to use intensified therapies. In the DAFNE Trial, all participants had selected theintensive programme of treatment, and significant and apparently lasting benefits in qualityof life measures were demonstrated using the intensified management strategy (DAFNEResearch Group, 2002). However, it must be acknowledged that when individual patientsare exhorted to achieve perfection in glycaemic control, they may experience difficultiesand frustration with the impossible task of trying to eliminate blood glucose readings thatlie outside the normal range, especially if they are not equipped to act upon such readings.In general, the main risk of intensified diabetes therapy remains hypoglycaemia. Thischapter examines the problem of hypoglycaemia that is specifically associated with strictglycaemic control, an area that has aroused much concern and controversy. Most commentsrelate to patients with type 1 diabetes. The risks of severe hypoglycaemia associated withstrict control in insulin-treated type 2 diabetes are likely to be similar, but occur much laterin the natural history of the disease, when insulin deficiency is profound (see Chapter 11).DEFINITION OF HYPOGLYCAEMIAIt is difficult to determine a frequency of hypoglycaemia without first defining what is meant by‘hypoglycaemia’. In many studies, hypoglycaemia is documented by self-reporting, which maybe very unreliable (Heller et al., 1995). Retrospective analyses suffer from problems of recall,and accurate documentation of hypoglycaemia is obtained only in prospective research studiesthat require biochemical verification of low blood glucose concentrations (see Chapter 3).Hypoglycaemia can be categorised by its symptomatology and its severity, but no realconsensus exists. ‘Mild’ hypoglycaemia is usually defined as an episode that a person recognisesand treats themselves and does not significantly disrupt daily living; ‘severe’ hypoglycaemiais an episode in which blood glucose has fallen to a level where the patient hasbecome so disabled that assistance is required from another person (The Diabetes Controland Complications Trial Research Group, 1991). Alternatively, ‘severe’ hypoglycaemia maybe defined by the requirement for parenteral treatment (intramuscular glucagon or intravenousdextrose), with or without hospital admission, or by the development of coma (TheDiabetes Control and Complications Trial Research Group, 1987). A category of ‘moderate’hypoglycaemia, in which an individual requires external assistance but which falls shortof requiring parenteral therapy or developing a coma, or the division of hypoglycaemiainto grades of severity has also been used (Limbert et al., 1993). Obviously the definitionused will affect the estimate of incidence, and if severe hypoglycaemia is definedsolely as coma, rates will be lower than if all episodes requiring assistance are included.

CONTRIBUTORS TO INCREASED RISK OF SEVERE HYPOGLYCAEMIA 173Various levels of blood glucose concentration have been used to define mild and moderatehypoglycaemia biochemically, but there is now recognition that the blood glucose concentrationswhich have been used arbitrarily to define pathological ‘spontaneous’ hypoglycaemia(such as < 22 mmol/l) are unsuitable for defining hypoglycaemia in people with diabetes.An arterialised plasma glucose concentration of around 3.6 mmol/l may be sufficient tocause physiological autonomic responses in healthy volunteers (Chapter 1). Subtle changesin cognitive function can initially be detected by formal testing below 4.0 mmol/l, althoughclinically relevant cognitive impairment does not occur until the arterialised plasma glucoseconcentration has fallen to approximately 3.0 mmol/l (Chapter 2). There is evidence thatif plasma glucose falls below 3.0 mmol/l for a period of time, this can reduce the symptomaticresponses to a further episode of hypoglycaemia occurring within the following 24hours (Heller and Cryer, 1991), so-called antecedent hypoglycaemia (Chapter 7). The continuousavoidance of low blood glucose concentrations (below 3.0 mmol/l) can restore normalhormonal and symptomatic responses to hypoglycaemia (Fanelli et al., 1993; Cranston et al.,1994), and it follows that values below 3.0 mmol/l can be considered to be hypoglycaemic.Diabetes UK coined the phrase ‘make four the floor’ to protect against potentially dangeroushypoglycaemia and, most recently, a panel convened by the American Diabetes Associationconcluded in favour of a concentration of 3.9 mmol/l (Working Group on Hypoglycemia,American Diabetes Association, 2005). This was based on research data, showing evidenceof counterregulatory responses at blood glucose concentrations (often arterial or arterialised)of 3.9 mmol/l and the demonstration of a small reduction in the glucagon response tohypoglycaemia induced immediately afterwards. However, this defines ‘hypoglycaemia’ atblood glucose concentrations that commonly occur in healthy non-diabetic people and maytherefore encourage over-treatment. It is probably more appropriate to differentiate betweentargets for adjusting therapy, which may properly be above 4.0 mmol/l, and ‘hypoglycaemia’,which implies a pathological cause and requires intervention.CONTRIBUTORS TO INCREASED RISK OF SEVEREHYPOGLYCAEMIA IN PATIENTS UNDERTAKING INTENSIFIEDINSULIN THERAPYFactors Predisposing Patients to Severe Hypoglycaemia in IntensifiedInsulin TherapyThe relationship between impaired symptomatic awareness of hypoglycaemia and anincreased rate of severe hypoglycaemia is well established (Hepburn et al., 1990; Gold et al.,1994; Clarke et al., 1995), although affected patients in these studies were not subject to strictglycaemic control. The association between counterregulatory failure and increased risk ofsevere hypoglycaemia is also well recognised (Ryder et al., 1990). Indeed, counterregulatoryfailure was proposed as a predictor of risk of severe hypoglycaemia in the subsequentapplication of intensified therapy (White et al., 1983), and it was not until later that theability of intensified therapy to cause counterregulatory failure was suggested (Simonsonet al., 1985a). It is indeed very important to appreciate that neither asymptomatic nor severehypoglycaemia are restricted to people using intensified insulin therapy.

174 RISKS OF STRICT GLYCAEMIC CONTROLApart from a previous history of severe hypoglycaemia, the greatest risk may be the degreeof insulin deficiency, as reflected by the absence of C-peptide (Muhlhauser et al., 1998), aswell as the glycaemic control prior to embarking upon intensified therapy and the determinationto reach the glycaemic targets (Bott et al., 1994; Muhlhauser et al., 1998). Preservationof endogenous insulin is not affected by intensification of insulin therapy, althoughthere is evidence to suggest that if strict glycaemic control is imposed when diabetes is diagnosed,this may result in more prolonged preservation of endogenous insulin secretion (Shahet al., 1989, The Diabetes Control and Complications Trial Research Group, 1998b). Otherfactors, related to the patient rather than to treatment, may increase the risk of severe hypoglycaemia,including social class (Muhlhauser et al., 1998) and possibly genetics. In a study fromDenmark, much of the risk of severe hypoglycaemia was attributed to ACE genotype (Pedersen-Bjergaard et al., 2003), although this has not been confirmed and has aroused controversy;also, the absence of traditional risk factors in the Danish study is a cause for concern.The Effects of Intensified Insulin Therapy Upon Risk of SevereHypoglycaemiaIn the DCCT, a clear link was demonstrated between intensified insulin therapy and thefrequency of severe hypoglycaemia. In that trial, a three-fold higher rate of severe hypoglycaemiawas recorded by the patients in the intensive treatment arm when compared with thoseon conventional therapy (The Diabetes Control and Complications Trial Research Group,1991; 1993; 1997). This persisted throughout the entire study, although absolute rates declinedgradually in both groups. Furthermore, the risk of severe hypoglycaemia was higher for anygiven HbA 1c , for the people receiving intensive treatment. This phenomenon has not beenadequately explained. It is now known that exposure to hypoglycaemia per se can inducedefects in counterregulation and loss of subjective awareness of hypoglycaemia (Heller andCryer, 1991; George et al., 1995; 1997; Davis et al., 1997). It has been assumed that intensivetherapy exposes the patient to a greater frequency of mild hypoglycaemia that is sufficientto induce such defects and thereby increase the risk of severe hypoglycaemia by thatmechanism. However, methods for delivering intensified diabetes therapy have subsequentlyimproved. Modern methods that focus on transferring skills of insulin adjustment to the patientsthemselves are reported to achieve improvements in HbA 1c with multiple daily injectiontherapy regimens, without causing more episodes of severe hypoglycaemia, and in their mostsuccessful forms achieve a parallel reduction of hypoglycaemia (Jorgens et al., 1993; DAFNEStudy Group, 2002; Plank et al., 2004; Samann et al., 2005). The judicious use of insulinanalogues in intensified regimens may be associated with slightly less risk of hypoglycaemia(Ashwell et al., 2006), whereas the use of continuous subcutaneous insulin infusion (CSII)with pumps is associated with a much lower frequency of severe hypoglycaemia, and has beenused successfully as treatment for patients with problematical hypoglycaemia (Bode et al.,1996, Rodrigues et al., 2005) and in the context of clinical trials (Hoogma et al., 2006).The Link Between Intensified Insulin Therapy and Risk of SevereHypoglycaemiaPatients describe symptoms of hypoglycaemia at a wide range of blood glucose concentrations.In an individual patient, the main determinant of the blood glucose concentration

CONTRIBUTORS TO INCREASED RISK OF SEVERE HYPOGLYCAEMIA 175at which protective responses commence is probably the recent prevailing range of bloodglucose concentration to which the patient has been exposed. For example, when patientswith poorly-controlled type 2 diabetes were studied with a controlled hypoglycaemic challengeafter blood glucose had been normalised overnight, their epinephrine responses tohypoglycaemia were triggered at higher blood glucose values than in well-controlled patients(Korzon-Burakowska et al., 1998).As mentioned earlier, the first indication that strict glycaemic control might cause abnormalresponses to hypoglycaemia was observed when controlled hypoglycaemia was induced ina small group of patients with type 1 diabetes before, and after, they had been treatedwith intensified insulin therapy (Simonson et al., 1985a). Following the improvement inglycaemic control, the magnitude of the counterregulatory hormonal response to an abruptlowering of blood glucose to 2.8 mmol/l was significantly less than observed previously.This study had been planned to investigate the potential of better glycaemic control to restoresome of the defects of normal counterregulation that develop in people with type 1 diabetes(see Chapter 6), so these results were unexpected. The importance of these preliminaryobservations was underlined by a subsequent study in which patients with type 1 diabetesreceiving intensified insulin treatment were found to have impaired glucose counterregulation(Amiel et al., 1987). During an intravenous infusion of insulin, most patients were unableto maintain arterialised plasma glucose above 3.0 mmol/l, in contrast with conventionallytreateddiabetic patients whose glycaemic control was not as good (as demonstrated by higherglycated haemoglobin concentrations) or non-diabetic volunteers. The intensively-treateddiabetic patients were less symptomatic, and although the rise in their plasma epinephrinewas of similar magnitude to the other groups, this occurred only when the hypoglycaemia wasmore profound. Further studies of hypoglycaemia, using a stepped glucose clamp to producea controlled reduction of blood glucose, confirmed that the symptomatic and hormonalresponses started at lower blood glucose concentrations in patients with strict glycaemiccontrol, and were delayed in onset and diminished in magnitude for any given blood glucoseconcentration (Amiel et al., 1988) (Figure 8.1).The delayed onset and diminished vigour of symptomatic and hormonal responses to hypoglycaemiain strictly-controlled diabetic subjects offers a partial explanation for the increasedFigure 8.1 The effect of intensified diabetes therapy (IRx) on epinephrine responses to a slowreduction in plasma glucose over four hours. Copyright © 1988 American Diabetes Association. FromAmiel et al., 1988. Reprinted with permission from The American Diabetes Association

176 RISKS OF STRICT GLYCAEMIC CONTROLoccurrence of asymptomatic biochemical hypoglycaemia. The risk may be particularly manifestedwhen the glycated haemoglobin concentration is reduced to within, or just above, thenon-diabetic range (Box 8.1). This was shown in a study of 34 subjects with type 1 diabeteswho had a wide range of total HbA 1 values (Kinsley et al., 1995). They were subjected to astepped glucose clamp to lower arterialised blood glucose to 2.3 mmol/l and the responses werecompared with a non-diabetic control group. Symptomatic responses (particularly autonomic)and some counterregulatory hormonal responses were diminished in the seven diabetic subjectswho had a total HbA 1 of 7.85% or less, i.e., glycaemic control that was within their local nondiabeticrange of total HbA 1 . A very similar study by Pampanelli et al. (1996) produced identicalobservations in 10 of 33 subjects, whose HbA 1c was within the local non-diabetic range, and inwhom it was also noted that the onset of some aspects of cognitive dysfunction was delayed.Current evidence would suggest that it is the increased exposure to episodic hypoglycaemia,associated with the treatment strategy that is promoting the problem. Most importantly, a seriesof studies has shown that hypoglycaemia awareness and counterregulatory hormone responsescan be restored in well-controlled diabetic subjects by avoidance of blood glucose concentrationsbelow 3.0 mmol/l in daily life, confirming the circular link between hypoglycaemiaexposure and impaired awareness of hypoglycaemia (Fanelli et al., 1993; Cranston et al., 1994).Thus, although impaired awareness of hypoglycaemia is a major problem in clinicalpractice, it is by no means exclusively confined to intensified therapy. Although the riskremains greater with lower mean glucose and glycated haemoglobin concentrations, impairedawareness is reversible, at least in the setting of carefully controlled research studies, byscrupulous avoidance of even modest hypoglycaemia in daily life (Fanelli et al., 1993;Cranston et al., 1994). Although this may result in a deterioration of glycaemic control asthe problem was reversed, with a rise in mean HbA 1c from 6.9% to 8.0% in one smallstudy of seven patients with impaired awareness of hypoglycaemia (Liu et al., 1996), thisis not inevitable (Cranston et al., 1994). It is possible for avoidance of hypoglycaemia toresult in an improvement of glycated haemoglobin, as post-hypoglycaemia hyperglycaemiais eradicated.Much less work has been done in type 2 diabetes, although there is increasing evidencethat in patients with insulin-treated type 2 diabetes of long duration, the prevalence of severehypoglycaemia is not greatly different from people with type 1 diabetes (see Chapter 11).Modern trends of starting insulin earlier in type 2 diabetes, when insulin deficiency is notBox 8.1Effects of strict glycaemic control in type 1 diabetes• Reduction in microvascular and macrovascular complications.• Potential increase in risk of severe hypoglycaemia.• Diminished counterregulatory and symptomatic responses to hypoglycaemia.• Altered glycaemic thresholds for activation of responses (i.e., lower blood glucoserequired).• Promotion of increased frequency of exposure to hypoglycaemia which exacerbatesimpaired awareness of hypoglycaemia.• Tendency to weight gain.

CEREBRAL ADAPTATION 177severe, are likely to reduce the overall risk of hypoglycaemia in patients with insulin-treatedtype 2 diabetes. Recent studies using bedtime basal insulin as the first line of intensifyingdiabetes treatment for type 2 patients who are not achieving glycaemic targets, have reporteda low risk of severe hypoglycaemia, even when using conventional insulins (Yki-Jarvinenet al., 2006). However, caution is indicated when patients require conversion to full insulintherapy. In a small study of poorly-controlled patients treated with oral medication, in whichresponses to hypoglycaemia were measured before and after improving glycaemic controlwith insulin, counterregulatory responses and the blood glucose thresholds at which thesewere initiated were modified, as occurs in type 1 diabetes (Korson-Burakowska et al., 1998).CEREBRAL ADAPTATIONWhen hypoglycaemia occurs, the stimulus for counterregulation appears to be a fall in thecerebral metabolic rate of glucose. Boyle et al. (1994) measured arteriovenous differencesin glucose concentration in the human brain during hypoglycaemia to show that the rate ofuptake of glucose (and by implication of metabolism) falls before most of the counterregulatoryresponses and cognitive changes occur. They also demonstrated that this fall in metabolicrate of the brain was reduced in healthy volunteers who were made acutely hypoglycaemicfollowing a period of 56 hours of protracted moderate hypoglycaemia, suggesting that themetabolism of the human brain can adapt to prolonged exposure to low blood glucose.This enables the brain to maintain its metabolism and continue to function in response tosubsequent hypoglycaemia. A further study in diabetic patients showed that diabetic patientswith strict glycaemic control and impaired awareness of hypoglycaemia were able to maintainthe rate of cerebral uptake of glucose during experimental hypoglycaemia, while otherswith normal symptomatic awareness exhibited a marked fall in cerebral uptake of glucose,associated with symptomatic and counterregulatory hormonal responses (Figure 8.2) (Boyleet al., 1995). These data led to the hypothesis that impaired awareness of hypoglycaemiaand defective glucose counterregulation may result from an adaptation in the sensitivity ofthe cerebral glucose sensor, which allows it to sustain its metabolic rate (and so not triggercounterregulation) during subsequent hypoglycaemia (see Chapter 7).However, the expectation that patients with impaired awareness of hypoglycaemia willshow an increase in brain glucose metabolic rate at any given blood glucose concentrationhas not been supported by neuroimaging studies. In studies utilising positron emissiontomography that used various tracers for glucose to measure either the metabolic rate ofbrain glucose or glucose tracer uptake in humans, several investigators have failed to finddifferences during euglycaemia or hypoglycaemia that could be in accord with prevailingglycaemic control (Cranston et al., 2001; Segal et al., 2001). One study found evidence duringhypoglycaemia of a difference in the change in uptake of the glucose tracer, de-oxyglucose,in the brain region around the hypothalamus in intensively-treated diabetic subjects who hadimpaired awareness of hypoglycaemia (Cranston et al., 2001), which is of interest becauseanimal studies have implicated this region (among others) in sensing hypoglycaemia. Inmore recent studies, the difference in brain glucose metabolism in subjects with impairedawareness of hypoglycaemia was a failure of increase of cerebral metabolic rate duringhypoglycaemia, associated with a failure to generate or perceive symptoms (Bingham et al.,2005). These data are compatible with the concept that cortical activation is important forperception of symptoms and that this fails in people who develop a loss of awareness of

178 RISKS OF STRICT GLYCAEMIC CONTROLFigure 8.2 Changes from baseline (mean ± SD) in (a) glucose uptake in the brain and (b) hypoglycaemiasymptom scores, and plasma concentrations of (c) epinephrine and (d) pancreatic polypeptideduring hypoglycaemia in patients with type 1 diabetes with differing degrees of glycaemic control(black bars), and in non-diabetic subjects (grey bars). Reproduced from Boyle et al. (1995) withpermission. Copyright © 1995 Massachusetts Medical Society

CEREBRAL ADAPTATION 179hypoglycaemia. It is becoming evident that changes in symptomatic responses and corticalfunction in hypoglycaemia are driven by complex mechanisms associated with, but notexclusively controlled directly by, the changes in the glucose metabolic rate of neurones.Some cognitive functions are better preserved than others during hypoglycaemia in subjectswho have previous experience of hypoglycaemia than in hypoglycaemia-naive subjects whohave normal counterregulation (Fanelli et al., 1993; Boyle et al., 1995). This does not entirelyfit the clinical picture of patients becoming significantly confused during hypoglycaemiawhile remaining asymptomatic.One measure of cognitive function, the choice reaction time, does not appear to adapt, andwhen hypoglycaemia is induced slowly, it deteriorates at similar levels of blood glucose inmost subjects, irrespective of their previous glycaemic experience and their state of hypoglycaemiaawareness (Maran et al., 1995). Other measures of cognitive function also deteriorateat similar levels of blood glucose in diabetic subjects who have had very disparate experiencesof preceding glycaemia (Widom and Simonson, 1990; Amiel et al., 1991; Hvidberget al., 1996). The ability of the brain to adapt its metabolic and functional capacity accordingto previous glycaemic experience varies across different regions of the brain. Regions of thebrain that detect hypoglycaemia, and some parts of the cerebral cortex, may be able to adaptmore effectively to antecedent hypoglycaemia than other areas, to sustain glucose metabolismduring subsequent exposure. As blood glucose falls this would effectively destroy the normalprotective hierarchy of corrective and symptomatic responses that precede cognitive impairment,replacing it with the dangerous situation whereby cognitive impairment is the initialresponse to hypoglycaemia, with autonomic responses not occurring until the blood glucosedeclines to a much lower level. In this situation the patient becomes too confused and unableto recognise the warning symptoms and so take appropriate corrective action (Figure 8.3).Figure 8.3 The change in hierarchy of responses to hypoglycaemia (a) before and (b) after intensifiedinsulin therapy in type 1 diabetes mellitus

180 RISKS OF STRICT GLYCAEMIC CONTROLThe magnitude of the change in glycaemic thresholds for various functions of the brainin response to strict control of diabetes is variable. Where glucose thresholds for cognitivedysfunction do alter in people with impaired awareness of hypoglycaemia, the differencesbetween the blood glucose thresholds for the symptomatic and autonomic responses and thosefor the onset of cognitive impairment are much smaller. As a result, the window of opportunityfor the patient to recognise that hypoglycaemia is developing is much narrower, givingless time for corrective action to be taken. As described above, the molecular mechanismscontrolling the thresholds for activation of the various components of the counterregulatoryresponses remain the subject of intense research.OTHER RISKS OF INTENSIFIED INSULIN THERAPYDiabetic Ketoacidosis and HyperinsulinaemiaAlthough severe hypoglycaemia was indisputably the major metabolic side-effect of intensiveinsulin therapy in the DCCT, concerns have been expressed that some intensive treatmentregimens may also increase the risk of developing ketosis. This was primarily related tothe use of CSII (with insulin pump therapy) and was thought to relate to the absence ofany intermediate-acting or background insulin in the event of pump failure. In insulin pumptherapy, soluble or fast-acting analogue insulin is delivered steadily by a slow infusion ofvery low doses throughout the day. The insulin delivery is accelerated before meals to deliverboluses, akin to giving intermittent subcutaneous injections of short-acting insulin. Becausethe basal insulin is delivered in a very low volume and there is no depot of intermediateactinginsulin in the subcutaneous tissues to act as a reservoir, an interruption in the deliveryof insulin can rapidly lead to hyperglycaemia and even ketosis, especially if the patient’sblood glucose is already elevated (Castilloa et al., 1996). This may occur as a result ofdisconnection of the pump, air in the delivery system, blockage in the tubing or morerarely, mechanical failure of the pump. The apparently high risk of diabetic ketoacidosis(DKA) with insulin pump therapy was first described when pumps left the experimentalcentres where they had been invented and were taken up for more general use (Knightet al., 1985), although the same centre that reported a problem with DKA also reportedsatisfactory experience overall with pump therapy (Knight et al., 1986). In 1997, a metaanalysisof trials of CSII has indicated that the rate of DKA was significantly higher (Eggeret al., 1997) and the rate of development of DKA was also slightly greater in intensivelytreatedpatients in the DCCT, although many of those patients used multiple injectionsof insulin to improve their glycaemic control (The Diabetes Control and ComplicationsTrial Research Group, 1995b). However, a more recent meta-analysis, including studiesusing more modern equipment and up-dated algorithms for using pump therapy, is morereassuring (Pickup et al., 2002). Current intensive treatment regimens focus more on transferringskills of flexible insulin dose adjustment more effectively to the patients and in thissetting, DKA rates are not higher. Indeed, experienced centres have utilised insulin pumptherapy to help people avoid recurrent DKA (Rodrigues et al., 2005). However, the riskis worth reiterating, as it can return when new technologies are applied in inexperiencedsettings.Intensive insulin therapy often leads to a redistribution of the times of administrationof insulin rather than a straightforward increase in dosage, and concerns have been raised

THERAPEUTIC MANIPULATION 181Figure 8.4 Severe hypoglycaemia (events per 100 patient-years) at baseline with multiple dailyinjections (MDI) and by year on continuous subcutaneous insulin infusion (CSII). Copyright © 1996American Diabetes Association. From Bode et al., 1996. Reprinted with permission from The AmericanDiabetes Associationthat continuous peripheral hyperinsulinaemia may be deleterious. This anxiety may be moretheoretical than real, as improved glycaemic control in type 1 diabetes improves insulinsensitivity (Simonson et al., 1985b), which ultimately should prevent hyperinsulinaemia.However, achieving adequate plasma concentrations of insulin in the hepatic circulation isalways likely to be at the cost of promoting hyperinsulinaemia in the systemic circulation,as insulin has to be delivered by subcutaneous injection. This potential over-insulinisationmay contribute to the risk of hypoglycaemia, and when insulin is delivered into the portalsystem, as with intraperitoneal infusion systems, hypoglycaemia is less frequent at any givenblood glucose level (Lassmann-Vague et al., 1996; Dunn et al., 1997). However, one study(Figure 8.4) demonstrated a pronounced and sustained reduction in the frequency of severehypoglycaemia following the transfer of patients from multiple injections of insulin to CSII(Bode et al., 1996), and so appropriate temporal distribution of the action of insulin may,be the critical factor in preventing hypoglycaemia.THERAPEUTIC MANIPULATIONAvoidance of HypoglycaemiaIt is important to stress that management of the potentially devastating syndrome of impairedawareness of hypoglycaemia and deficient counterregulation, with its high risk of severehypoglycaemia, should not be an excuse for encouraging poor glycaemic control. However,some patients with these acquired syndromes are not suitable for intensive insulin therapy and

182 RISKS OF STRICT GLYCAEMIC CONTROLvery strict glycaemic control, and blood glucose targets may have to be higher when facedwith these problems. Frequent blood glucose monitoring is essential to identify biochemicalhypoglycaemia, and the use of basal-bolus insulin regimens (which predominantly use shortactinginsulins) may be beneficial in avoiding recurrent hypoglycaemia. However, frequentblood glucose monitoring alone can sometimes exacerbate the problem, unless the patientis instructed in how to use the data to adjust insulin regimens prospectively to avoidproblems, rather than to react by immediate treatment of the inevitable occasional highreading. There have been few detailed behavioural studies in people with impaired awarenessof hypoglycaemia but it may be a particular problem where the links between high bloodglucose and risk of vascular complications have been very well accepted by the patient, whomay need convincing that transient hyperglycaemia is not a problem. Clinical experiencesuggests that many patients with such problems ‘glucose chase’, taking corrective actionwhenever they identify a high blood glucose concentration. Correcting this behaviour, inparticular avoiding postprandial glucose correction, and re-training patients to use glucosemonitoring to seek patterns for future dose adjustment, can be very beneficial. Programmesthat teach patients to test blood glucose before eating and to use the information to adjust thedose of insulin required for the immediate meal, allow people to live with greater flexibility.By recording the blood glucose results, recurrent patterns in changes can be sought againstwhich prospective adjustments to insulin regimens can be made. These measures allow HbA 1cto be improved while lowering the risk of severe hypoglycaemia. It is important to recognisethat the preservation of physiological defences to hypoglycaemia is dependent upon theavoidance of hypoglycaemia in daily life, and not on tolerance of chronic hyperglycaemiaand an elevated glycated haemoglobin. As discussed above, the studies that have attempted torestore awareness of hypoglycaemia by avoidance of hypoglycaemia did not cause any majordeterioration of glycaemic control, although a modest increase in HbA 1c of around 0.5–1.0%occurred in two studies (Fanelli et al., 1993; Dagogo-Jack et al., 1994). Anecdotally, averageblood glucose concentrations and HbA 1c may even improve with strategies for avoidinghypoglycaemia, by preventing wide fluctuations in blood glucose.Strategies to avoid hypoglycaemia can be very time-consuming and labour-intensive, forthe patient as well as for the physician, and require several supportive measures, such ashaving to maintain daily telephone contact between the patient and the medical and nursingstaff (Fanelli et al., 1993). It took Cranston and colleagues (1994) up to 12 months for thesubjects taking part in their study to achieve three consecutive weeks when the home bloodglucose readings did not fall below 3.0 mmol/l. Two of three studies demonstrated partialrecovery of the counterregulatory responses to hypoglycaemia (Fanelli et al., 1993; Cranstonet al., 1994), but one did not (Dagogo-Jack et al., 1994), although the symptoms of hypoglycaemiawere restored. The number of subjects was small in all of these studies. However,the studies were often done in patients with an established problem of hypoglycaemia andnewer tools are at hand to help prevent or reverse the problems.Educational strategies cannot be emphasised too strongly. They remain to be testedspecifically in people who have major problems with recurrent severe hypoglycaemia buttheir ability to achieve better glycaemic control with less frequent hypoglycaemia is amajor improvement over the outcomes of the DCCT. For patients with type 1 diabetes,the introduction of fast-acting insulin analogues for insulin replacement at meal-times hasreduced hypoglycaemia, particularly at night, because of the much shorter duration of actionof fast-acting analogues (Brunelle et al., 1998; Home et al., 2000). Likewise, the longeractinginsulin analogues have been associated with a lower risk of nocturnal hypoglycaemia

PATIENTS UNSUITABLE FOR STRICT CONTROL 183(Pieber et al., 2000; Vague et al., 2003) and regimens that combine short and longer-actinginsulin analogues claim to have a lower risk of hypoglycaemia in patients with type 1 diabetes(Hermansen et al., 2004; Ashwell et al., 2005), although the methods of measurement ofhypoglycaemia in many of these trials was much less robust than was discussed earlier.The potential of CSII to reduce hypoglycaemia has also been highlighted. The use offaster and less painful glucose monitoring devices will facilitate home monitoring, althoughit remains critical that the data obtained are utilised through appropriate patient education.The advent of real-time glucose monitors may allow patients to avoid hypoglycaemia, as theycan take action to interrupt a steady decline in blood glucose concentration that they wouldnot previously have had the opportunity to observe. The value of these new technologiesremains to be proven in routine clinical use, but they do hold promise.The main defence against recurrent hypoglycaemia with its consequent blunting of subjectivesymptomatic awareness remains the establishment of therapeutic goals that are realisticfor individual patients. The physician’s tendency to concentrate on eliminating hyperglycaemiahas led to subnormal blood glucose values being ignored, a practice worsened by thebelief of some physicians and patients alike that, because an episode of biochemical hypoglycaemiais asymptomatic, it is not important. There is no doubt that a clinically detectabledeterioration in performance of some aspects of cognitive function occurs in human subjectsat arterialised blood glucose concentrations of 3.0 mmol/l (see Chapter 2), and an absenceof symptoms at that level should ring alarm bells with the patient’s physician. Given thathealthy non-diabetic subjects do not commonly exhibit fasting blood glucose concentrationsbelow 4.0 mmol/l, it seems wholly unnecessary to encourage or even permit such subnormalityin patients with diabetes (one exception to this maxim being pregnancy where healthynon-diabetic women do exhibit lower blood glucose levels as discussed in Chapter 10). Withintensive insulin therapy the therapeutic targets should be near-normal blood glucose levels(before meals 4.0–7.0 mmol/l, after meals 4.0–9.0 mmol/l, depending on time of testing),with a slightly higher than normal glucose at bedtime (7.0–9.0 mmol/l) to reduce the risk ofhypoglycaemia occurring during the night. Blood glucose measured during the night may bea little lower (≥ 36 mmol/l), but in view of the evidence presented above, patients shouldavoid allowing it to fall any lower than this.PATIENTS UNSUITABLE FOR STRICT CONTROLThe DCCT demonstrated that any reduction in glycated haemoglobin is associated with areduced risk of microvascular complications over time, and the benefits are greater withhigher glycated haemoglobin concentrations (The Diabetes Control and Complications TrialResearch Group, 1996b). A cross-sectional study which suggested that the risk reductionfor nephropathy is near-maximal at a glycated haemoglobin of 8% (Krolewski et al., 1995)cannot be extrapolated to other microvascular complications, because in the DCCT noglycaemic threshold (estimated by glycated haemoglobin) for the development of retinopathywas demonstrated in a patient group whose average HbA 1c was 7%. The long-term followupof the DCCT cohort has confirmed that intensive therapy benefits macrovascular aswell as microvascular risk and that its effects are sustained (Writing team for the DiabetesControl and Complications Trial/ Epidemiology of Diabetes Interventions and ComplicationsResearch Group, 2002). Thus, unless an individual already has normal glycated haemoglobinwith no problematical hypoglycaemia, no patient who has diabetes is unsuitable for attempts

184 RISKS OF STRICT GLYCAEMIC CONTROLto improve their glycaemic control, especially with modern techniques that are able to deliverimproved control without increasing hypoglycaemia risk.In practice, however, there are patients in whom attempts to achieve a near normal glycatedhaemoglobin are not appropriate (Box 8.2). Patients with advanced complications, especiallyretinopathy, have not been shown to benefit and a sudden improvement in glycaemic controlmay cause acceleration in severity of pre-proliferative or early proliferative retinopathy(Hanssen et al., 1986). Although some authorities claim that this should not be a contraindicationto improving glycaemic control (Chantelau and Kohner, 1997), as yet there is no realevidence for benefit in advanced cases and the retinopathy should be treated appropriatelybefore glycaemic control is intensified. Similarly, in patients with established renal impairmentand severe macrovascular disease, attempts to treat elevated blood pressure and plasmalipids and to encourage patients to stop smoking may be more beneficial than targetingglycaemic control alone. As intensive insulin therapy is aimed at achieving benefit over aperiod of five to ten years or more, patients with a reduced life expectancy should not beexposed to the risks and rigours associated with this treatment regimen. This applies also toelderly patients, who may be frail and physically inactive.Very young patients may not be good candidates for very strict glycaemic control. Poorcontrol should not be encouraged in children, as growth may be jeopardised, and there issome evidence that pre-pubertal glycaemic control may influence the later risk of complications(Donaghue et al., 1997; Holl et al., 1998). However, very small children, who arevery insulin-sensitive, may be at risk of intellectual damage if exposed to recurrent severehypoglycaemia (see Chapters 9 and 13).Box 8.2Application of strict glycaemic control in type 1 diabetesCaution required:• Long duration of insulin-treated diabetes (counterregulatory deficiences).• Previous history of severe hypoglycaemia.• Established impaired awareness of hypoglycaemia.• History of epilepsy.• Patient unwilling to do home blood glucose monitoring.Contraindicated:• Extremes of age.• Ischaemic heart disease.• Unstable diabetic retinopathy (can be instituted after treatment).• Advanced diabetic complications.• Limited life expectancy (e.g. serious coexisting disease).

REFERENCES 185It is the patient who determines the degree of glycaemic control that they feel is worththe effort. Patients with very erratic life styles, and those who are not prepared to committhemselves to regular self-monitoring of blood glucose, with frequent attention to the timingof injection and adjustment of dosage of insulin, cannot safely undertake measures toachieve near-normoglycaemia. A compromise must be reached after a full discussion of therisks. Patients currently experiencing problematical hypoglycaemia may not wish to aim forglycaemic targets near the normal range, although the regimens of intensive insulin therapymay still be appropriate for them if they can eliminate hypoglycaemia from their daily lives.This may also be true of people undertaking dangerous or physically demanding jobs, whomay deliberately set higher blood glucose targets to protect against hypoglycaemia, but whoshould be encouraged to practice regular self-monitoring and adjustment of insulin doses. Itis the informed patient who must determine their own therapeutic aims at any given time.The doctor’s role is to try to ensure that the patient has the knowledge to make appropriatedecisions and to provide the tools to achieve these aims.CONCLUSIONS• The principal risks of intensive insulin therapy are hypoglycaemia and weight gain.• In earlier studies such as the DCCT, patients using intensive insulin regimens were threetimes more likely to have an episode of severe hypoglycaemia than those on conventionalinsulin regimens, but newer techniques for training patients to use insulin flexibly candeliver improved glycaemic control without any increase in severe hypoglycaemia, andsometimes with reduced hypoglycaemia occurrence.• It is likely that exposure to hypoglycaemia during intensive therapy impairs the counterregulatoryresponses to hypoglycaemia and symptomatic awareness and this may be seenparticularly if glycated haemoglobin is within the non-diabetic range.• Total avoidance of hypoglycaemia can restore the symptomatic response to hypoglycaemia,but achieving this once problematic hypoglycaemia has been established may bedemanding and time-consuming, both for patients and healthcare professionals. Nevertheless,the long-term benefits of good diabetic control in type 1 diabetes are unequivocaland current technologies should help more patients achieve it.• It is essential that all definitions of good glycaemic control include an absence of severehypoglycaemia as well as near-normal glycated haemoglobin. Intensified treatment regimensshould be adjusted to incorporate both.REFERENCESAgardh CD, Eckert B, Agardh E (1992). Irreversible progression of severe retinopathy in young type-1insulin-dependent diabetes mellitus patients after improved metabolic control. Journal of DiabeticComplications 6: 96–100.Amiel SA, Tamborlane WV, Simonson DC, Sherwin RS (1987). Defective glucose counterregulationafter strict glycemic control of insulin-dependent diabetes mellitus. New England Journal ofMedicine 316: 1376–83.

186 RISKS OF STRICT GLYCAEMIC CONTROLAmiel SA, Sherwin RS, Simonson DC, Tamborlane WV (1988). Effect of intensive insulin therapyon glycemic thresholds for counterregulatory hormone release. Diabetes 37: 901–7.Amiel SA, Pottinger RC, Archibald HR, Chusney G, Cunnah DTF, Prior PF, Gale EAM (1991). Effectof antecedent glucose control on cerebral function during hypoglycemia. Diabetes Care 14: 109–18.Ashwell SG, Amiel SA, Bilous RW, Dashora U, Heller SR, Hepburn DA et al. (2006). Improvedglycaemic control with insulin glargine plus insulin lispro: a multicentre, randomized, cross-overtrial in people with type 1 diabetes. Diabetic Medicine 23: 285–92.Bingham EM, Dunn JT, Smith D, Sutcliffe-Goulden J, Reed LJ, Marsden PK, Amiel SA (2005).Differential changes in brain glucose metabolism during hypoglycaemia accompany loss of hypoglycaemiaawareness in men with type 1 diabetes mellitus. An [11C]-3-O-methyl-D-glucose PETstudy. Diabetologia 48: 2080–9.Bode BW, Steed RD, Davidson PC (1996). Reduction in severe hypoglycemia with long-term continuoussubcutaneous insulin infusion in type 1 diabetes. Diabetes Care 19: 324–7.Bott U, Jorgens V, Grusser M, Bender R, Muhlhauser I, Berger M (1994). Predictors of glycaemiccontrol in type 1 diabetic patients after participation in an intensified treatment and teachingprogramme. Diabetic Medicine 11: 362–71.Boyle PJ, Nagy RJ, O’Connor AM, Kempers SF, Yeo RA, Qualls C (1994). Adaptation in brainglucose uptake following recurrent hypoglycemia. Proceedings of the National Academy of Sciencesof the United States of America 91: 9352–6.Boyle PJ, Kempers SF, O’Connor AM, Nagy RJ (1995). Brain glucose uptake and unawarenessof hypoglycemia in patients with insulin-dependent diabetes mellitus. New England Journal ofMedicine 333: 1726–31.Brunelle BL, Llewelyn J, Anderson JH Jr, Gale EA, Koivisto VA (1998). Meta-analysis of the effect ofinsulin lispro on severe hypoglycemia in patients with type 1 diabetes. Diabetes Care 21: 1726–31.Carlson MG, Campbell PJ (1993). Intensive insulin therapy and weight gain in IDDM. Diabetes 42:1700–7.Castilloa MJ, Sheen AJ, Lefebvre PJ (1996). The degree/rapidity of the metabolic deteriorationfollowing interruption of a continuous subcutaneous insulin infusion is influenced by the prevailingblood glucose level. Journal of Clinical Endocrinology and Metabolism 81: 1975–8.Chantelau E, Kohner EM (1997). Why some cases of retinopathy worsen when diabetic controlimproves. British Medical Journal 315: 1105–6.Clarke WL, Cox DJ, Gonder-Frederick LA, Julian D, Schlundt D, Polonsky W (1995). Reducedawareness of hypoglycemia in adults with IDDM. A prospective study of hypoglycemic frequencyand associated symptoms. Diabetes Care 18: 517–22.Cranston I, Lomas J, Maran A, Macdonald IA, Amiel SA (1994). Restoration of hypoglycaemiaawareness in patients with long-duration insulin-dependent diabetes. Lancet 344: 283–7.Cranston I, Reed LJ, Marsden PK, Amiel SA (2001). Changes in regional brain (18)Ffluorodeoxyglucoseuptake at hypoglycemia in type 1 diabetic men associated with hypoglycemiaunawareness and counter-regulatory failure. Diabetes 50: 2329–36.DAFNE Study Group (2002). Training in flexible, intensive insulin management to enable dietaryfreedom in people with type 1 diabetes: Dose Adjustment For Normal Eating (DAFNE) randomisedcontrolled trial. British Medical Journal 325: 746.Dagogo-Jack S, Rattarasarn C, Cryer PE (1994). Reversal of hypoglycemia unawareness, but notdefective glucose counterregulation, in IDDM. Diabetes 43: 1426–34.Davis SN, Shavers C, Mosqueda-Garcia R, Costa F (1997). Effects of differing antecedent hypoglycemiaon subsequent counterregulation in normal humans. Diabetes 46: 1328–35.Donaghue KC, Fung AT, Hing S, Fairchild J, King J, Chan A et al. (1997). The effect of prepubertaldiabetes duration on diabetes microvascular complications in early and late adolescence. DiabetesCare 20: 77–80.Dunn FL, Nathan DM, Scavini M, Selam JL, Wingrove TG (1997). Long term therapy of IDDM with animplantable insulin pump. The Plantable Insulin Pump Trial Study Group. Diabetes Care 20: 59–63.

REFERENCES 187Egger M, Davey Smith G, Stettler C, Diem P (1997). Risk of adverse effects of intensified treatmentin insulin-dependent diabetes mellitus: a meta-analysis. Diabetic Medicine 14: 919–28.Fanelli CG, Epifano L, Rambotti AM, Pampanelli S, Di Vincenzo A, Modanelli F et al. (1993).Meticulous prevention of hypoglycemia normalizes the glycemic thresholds and magnitude of mostof neuroendocrine responses to, symptoms of, and cognitive function during hypoglycemia inintensively treated patients with short-term IDDM. Diabetes 42: 1683–9.Franciosi M, Pellegrini F, De Berardis G, Belfiglio M, Cavaliere D, Di Nardo B et al., QuED StudyGroup (2001). Diabetic patients: an urgent need for better educational strategies. Diabetes Care 24:1870–7.George E, Harris N, Bedford C, Macdonald IA, Hardisty CA, Heller SR (1995). Prolonged but partialimpairment of the hypoglycaemic physiological response following short-term hypoglycaemia innormal subjects. Diabetologia 38: 1183–90.George E, Marques JL, Harris ND, Macdonald IA, Hardisty CA, Heller SR (1997). Preservation ofphysiological responses to hypoglycemia 2 days after antecedent hypoglycemia in patients withIDDM. Diabetes Care 20: 1293–8.Gold AE, MacLeod KM, Frier BM (1994). Frequency of severe hypoglycemia in patients with type 1diabetes with impaired awareness of hypoglycemia. Diabetes Care 17: 697–703.Hanssen KF, Dahl-Jorgensen K, Lauritzen T, Feldt-Rasmussen B, Brinchmann-Hansen O, DeckertT (1986). Diabetic control and microvascular complications: the near-normoglycaemic experience.Diabetologia 29: 677–84.Heller SR, Cryer PE (1991). Reduced neuroendocrine and symptomatic responses to subsequenthypoglycemia after 1 episode of hypoglycemia in non-diabetic humans. Diabetes 40: 223–6.Heller S, Chapman J, McLoud J, Ward J (1995). Unreliability of reports of hypoglycaemia by diabeticpatients. British Medical Journal 310: 440.Hepburn DA, Patrick AW, Eadington DW, Ewing DJ, Frier BM (1990). Unawareness of hypoglycaemiain insulin-treated diabetic patients: prevalence and relationship to autonomic neuropathy. DiabeticMedicine 7: 711–7.Hermansen K, Fontaine P, Kukolja KK, Peterkova Y, Leth G, Gall MA (2004). Insulin analogues(insulin detemir and insulin aspart) versus traditional human insulins (NPH insulin and regularhuman insulin) in basal-bolus therapy for patients with type 1 diabetes. Diabetologia 47: 622–9.Holl RW, Lang GE, Grabert M, Heinze E, Lang GK, Debatin KM (1998). Diabetic retinopathy inpediatric patients with type 1 diabetes: effect of diabetes duration, prepubertal and pubertal onsetof diabetes, and metabolic control. Journal of Pediatrics 132: 790–4.Home PD, Lindholm A, Riis A, European Insulin Aspart Study Group (2000). Insulin aspart versushuman insulin in the management of long-term blood glucose control in type 1 diabetes mellitus: arandomized controlled trial. Diabetic Medicine 17: 762–70.Hoogma RP, Hammond PJ, Gomis R, Kerr D, Bruttomesso D, Bouter KP et al. 5 Nations StudyGroup (2006). Comparison of the effects of continuous subcutaneous insulin infusion (CSII) andNPH-based multiple daily insulin injections (MDI) on glycaemic control and quality of life: resultsof the 5-nations trial. Diabetic Medicine 23: 141–7.Hvidberg A, Fanelli CG, Hershey T, Terkamp C, Craft S, Cryer PE (1996). Impact of recent antecedenthypoglycemia on hypoglycemic cognitive dysfunction in non-diabetic humans. Diabetes 45: 1030–6.Jorgens V, Grusser M, Bott U, Mulhauser I, Berger M (1993). Effective and safe translation ofintensified insulin therapy to general internal medicine departments. Diabetologia 36: 99–105.Kinsley BT, Widom B, Simonson DC (1995). Differential regulation of counterregulatory hormonesecretion and symptoms during hypoglycemia in IDDM. Effect of glycemic control. Diabetes Care18: 17–26.Knight G, Jennings AM, Boulton AJ, Tomlinson S, Ward JD (1985). Severe hyperkalaemia andketoacidosis during routine treatment with an insulin pump. British Medical Journal 291: 371–2.Knight G, Boulton AJ, Drury J, Ward JD (1986). Long term glycaemic control by alternative regimensin a feasibility study of continuous subcutaneous insulin infusion. Diabetes Research 3: 335–8.

188 RISKS OF STRICT GLYCAEMIC CONTROLKorzon-Burakowska A, Hopkins D, Matyka K, Lomas J, Pernet A, Macdonald I, Amiel S (1998).Effects of glycemic control on protective responses against hypoglycemia in type 2 diabetes.Diabetes Care 21: 283–90.Krolewski AS, Laffel LM, Krolewski M, Quinn M, Warram JH (1995). Glycosylated hemoglobinand the risk of microalbuminuria in patients with insulin-dependent diabetes mellitus. New EnglandJournal of Medicine 332: 1251–5.Lassmann-Vague V, Belicar P, Alessis C, Raccah D, Vialettes B, Vague P (1996). Insulin kinetics intype I diabetic patients treated by continuous intraperitoneal insulin infusion: influence of anti-insulinantibodies. Diabetic Medicine 13: 1051–5.Limbert C, Schwingshandl J, Haas J, Roth R, Borkenstein M (1993). Severe hypoglycemia in childrenand adolescents with IDDM: frequency and associated factors. Journal of Diabetes Complications7: 216–20.Liu D, McManus RM, Ryan EA (1996). Improved counter-regulatory hormonal and symptomaticresponses to hypoglycemia in patients with insulin-dependent diabetes mellitus after 3 months ofless strict glycemic control. Clinical and Investigative Medicine 19: 71–82.Maran A, Lomas J, Macdonald IA, Amiel SA (1995). Lack of preservation of higher brain functionduring hypoglycaemia in patients with intensively-treated IDDM. Diabetologia 38: 1412–18.Muhlhauser I, Overmann H, Bender R, Bott U, Berger M (1998). Risk factors of severe hypoglycaemia inadult patients with type I diabetes – a prospective population based study. Diabetologia 41: 1274–82.Nathan DM, Lachin J, Cleary P, Orchard T, Brillon DJ, Backlund JY et al. (2003). Intensive diabetestherapy and carotid intima-media thickness in type 1 diabetes mellitus. New England Journal ofMedicine 348: 2294–303.Pampanelli S, Fanelli C, Lalli C, Gofetta M, Del Sindaco P, Lepore M et al. (1996). Long-termintensive therapy in IDDM: effects on HbA 1c , risk for severe and mild hypoglycaemia, status ofcounterregulation and awareness of hypoglycaemia. Diabetologia 39: 677–86.Pedersen-Bjergaard U, Agerholm-Larsen B, Pramming S, Hougaard P, Thorsteinsson B (2003). Predictionof severe hypoglycaemia by angiotensin-converting enzyme activity and genotype in type 1diabetes. Diabetologia 46: 89–96.Pickup J, Mattock M, Kerry S (2002). Glycaemic control with continuous subcutaneous insulininfusion compared with intensive insulin injections in patients with type 1 diabetes: meta-analysisof randomised controlled trials. British Medical Journal 324: 705–8.Pieber TR, Eugene-Jolchine I, Derobert E (2000). Efficacy and safety of HOE 901 versus NPH insulinin patients with type 1 diabetes. The European Study Group of HOE 901 in type 1 diabetes. DiabetesCare 23: 157–62.Plank J, Kohler G, Rakovac I, Semlitsch BM, Horvath K, Bock G et al. (2004). Long-term evaluationof a structured outpatient education programme for intensified insulin therapy in patients with type1 diabetes: a 12-year follow-up. Diabetologia 47: 1370–75.Rodrigues I, Reid HA, Ismail K, Amiel SA (2005). Indications and efficacy of continuous subcutaneousinsulin infusion (CSII) therapy in type 1 diabetes mellitus: a clinical audit in a specialist service.Diabetic Medicine: 22: 842–9.Ryder RE, Owens DR, Hayes TM, Ghatei MA, Bloom SR (1990). Unawareness of hypoglycaemia andinadequate hypoglycaemic counterregulation: no causal relation with diabetic autonomic neuropathy.British Medical Journal 301: 783–7.Samann A, Muhlhauser I, Bender R, Kloos C, Muller UA (2005). Glycaemic control and severehypoglycaemia following training in flexible, intensive insulin therapy to enable dietary freedom inpeople with type 1 diabetes: a prospective implementation study. Diabetologia 48: 1965–70.Segal SA, Fanelli CG, Dence CS, Markham J, Videen TO, Paramore DS et al. (2001). Blood-to-brainglucose transport, cerebral glucose metabolism, and cerebral blood flow are not increased afterhypoglycemia. Diabetes 50: 1911–7.Shah SC, Malone JL, Simpson NE (1989). A randomized trial of intensive insulin therapy in newlydiagnosed insulin-dependent diabetes mellitus. New England Journal of Medicine 320: 550–4.

REFERENCES 189Simonson DC, Tamborlane WV, DeFronzo RA, Sherwin RS (1985a). Intensive insulin therapy reducescounterregulatory hormone responses to hypoglycemia in patients with type 1 diabetes. Annals ofInternal Medicine 103: 184–90.Simonson DC, Tamborlane WV, Sherwin RS, Smith JD, DeFronzo RA (1985b). Improved insulinsensitivity in patients with type 1 diabetes mellitus after CSII. Diabetes 34 (suppl 3): 80–6.The Diabetes Control and Complications Trial Research Group (1987). The Diabetes Control andComplications Trial (DCCT). Results of a feasibility study. Diabetes Care 10: 1–19.The Diabetes Control and Complications Trial Research Group (1988). Weight gain associated withintensive therapy in the Diabetes Control and Complications Trial. Diabetes Care 11: 567–73.The Diabetes Control and Complications Trial Research Group (1991). Epidemiology of severe hypoglycemiain the Diabetes Control and Complications Trial. American Journal of Medicine 90:450–9.The Diabetes Control and Complications Trial Research Group (1993). The effect of intensive treatmentof diabetes on the development and progression of long-term complications in insulin-dependentdiabetes mellitus. New England Journal of Medicine 329: 977–86.The Diabetes Control and Complications Trial Research Group (1995a). Adverse events and theirassociation with treatment regimens in the Diabetes Control and Complications Trial. Diabetes Care18: 1415–27.The Diabetes Control and Complications Trial Research Group (1995b). Implementation of treatmentprotocols in the Diabetes Control and Complications Trial. Diabetes Care 18: 361–76.The Diabetes Control and Complications Trial Research Group (1996a). Influence of intensive diabetestreatment on quality-of-life outcomes in the Diabetes Control and Complications Trial. DiabetesCare 19: 195–203.The Diabetes Control and Complications Trial Research Group (1996b). The absence of a glycemicthreshold for the development of long-term complications: the perspective of the Diabetes Controland Complications Trial. Diabetes 45: 1289–98.The Diabetes Control and Complications Trial Research Group (1997). Hypoglycemia in the DiabetesControl and Complications Trial. Diabetes 46: 271–86.The Diabetes Control and Complications Trial Research Group (1998). Effect of intensive therapy onresidual beta-cell function in patients with type 1 diabetes in the Diabetes Control and ComplicationsTrial. A randomized, controlled trial. Annals of Internal Medicine 128: 517–23.Vague P, Selam JL, Skeie S, De Leeuw I, Elte JW, Haahr H et al. (2003). Insulin detemir is associatedwith more predictable glycemic control and reduced risk of hypoglycemia than NPH insulin inpatients with type 1 diabetes on a basal-bolus regimen with premeal insulin aspart. Diabetes Care26: 590–6.White NH, Skor DA, Cryer PE, Levandoski LA, Bier DM, Santiago JV (1983). Identification of type 1diabetic patients at increased risk for hypoglycemia during intensive therapy. New England Journalof Medicine 308: 485–91.Widom B, Simonson DC (1990). Glycemic control and neuropsychologic function during hypoglycemiain patients with insulin-dependent diabetes mellitus. Annals of Internal Medicine 112:904–12.Working Group on Hypoglycemia, American Diabetes Association (2005). Defining and reportinghypoglycemia in diabetes: a report from the American Diabetes Association Workgroup on Hypoglycemia.Diabetes Care 28: 1245–9.Writing team for the Diabetes Control and Complications Trial/ Epidemiology of Diabetes Interventionsand Complications Research Group (2002). Effect of intensive therapy on the microvascularcomplications of type 1 diabetes mellitus. Journal of the American Medical Association 287: 2563–9.Yki-Jarvinen H, Kauppinen-Makelin R, Tiikkainen M, Vahatalo M, Virtamo H, Nikkila K et al.(2006). Insulin glargine or NPH combined with metformin in type 2 diabetes: the LANMET study.Diabetologia 49: 442–51.

9 Hypoglycaemia in Childrenwith DiabetesKrystyna A. MatykaINTRODUCTIONSub-optimal care of children with type 1 diabetes mellitus carries devastating consequences.Young children, previously thought to be protected from the early development of microvascularcomplications, have been found to be at significant risk of these complications that canpresent in adolescence (Danne et al., 1994; Solders et al., 1997; Schultz et al., 1999). Yetthey are also at risk of detrimental neuropsychological sequelae, which are thought to berelated to recurrent episodes of hypoglycaemia that may accompany intensification of insulintherapy aimed at decreasing the risk of these microvascular complications. The exaggeratedmetabolic demands of the growing child, combined with a lifestyle that is unpredictableeven on a day to day basis, make children very vulnerable to both repeated and severeepisodes of hypoglycaemia (Allen et al., 2001). This chapter examines the aetiology, physiology,consequences and management of episodes of hypoglycaemia during this dynamictime of life.DEFINITION OF HYPOGLYCAEMIA IN CHILDHOODThe definition of hypoglycaemia in childhood has been extremely controversial. It hasbeen suggested that children can tolerate lower levels of blood glucose, especially as thedeveloping brain can use alternative substrates for cerebral metabolism. This has beensupported by the clinical finding that some children with diabetes appear to be ‘normal’ whentheir blood glucose concentrations are low as demonstrated by home blood glucose monitoring.However, difficulties arise as young children are not expected to perform complexpsychomotor tasks and clinical detection of mild changes in performance can be difficult.This subject is not easy to study because of the ethical difficulties of performing studies ofnormal glucose homeostasis during fasting in young children, which could be used to definethe limits of normality of blood glucose concentrations.Fasting glucose requirements will, in part, be determined by brain glucose uptake andthe central nervous system has a pivotal role in carbohydrate metabolism throughout life.The brain of infants and children can use glucose at a rate of 3–5 mg/kg/min, equivalent toalmost all endogenous production as defined by stable isotope studies of cerebral glucosemetabolism, and a linear correlation exists between glucose production and estimated brainHypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

192 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESsize, from premature infants to adult life (Bier et al., 1977). There is a marked changein the correlation between body weight and hepatic glucose production corresponding tomid-puberty (approximately 40 kg) indicating that brain growth is virtually complete (Bieret al., 1977). However, the developing brain also has the ability to use alternative substratesfor cerebral metabolism, and during fasting young children have higher concentrations ofketones and lactate when compared to adults (Haymond et al., 1982).A number of studies have examined metabolic parameters, including blood glucoseconcentrations, following fasts of varying duration in children. On the whole these studieshave been used to provide normative data for use in the clinical evaluation of children withpotential disorders of metabolism; as a result the fasts have been prolonged and samplinginfrequent. There is little detailed information on metabolic variables during a short fast asis experienced by children with diabetes, for example when they go to sleep. One study ofnocturnal glucose homeostasis in 39 children, subsequently found to have a constitutionalshort stature, demonstrated a cyclical variation in blood glucose concentrations during thenight with a periodicity of 80–120 minutes (Stirling et al., 1991). At some stage of thenight blood glucose levels fell transiently to below 3 mmol/l in 5% of the children, but it isdifficult to know if these children are representative of those who grow normally.Other studies have examined glucose and metabolite profiles intermittently duringprolonged fasts of 24–36 hour duration (Chaussain, 1973; Chaussain et al., 1974; Chaussainet al., 1977; Saudubray et al., 1981; Haymond et al., 1982; Kerr et al., 1983; Lamerset al., 1985a; Lamers et al., 1985b). Fasting glucose concentrations varied depending on theduration of the fast and age of child studied. After 24 hours, blood glucose was found torange from 3.0–3.5 mmol/l. Some of these studies suggested a positive correlation betweenfasting blood glucose concentrations and age (Chaussain et al., 1977; Saudubray et al., 1981;Lamers et al., 1985a).Definition of Hypoglycaemia in Childhood DiabetesAlthough there has been great controversy regarding the biochemical definition of hypoglycaemiafor the diagnosis of pathological states in the paediatric population (Koh et al., 1988),these arguments should not apply to type 1 diabetes. The management of type 1 diabetesinvolves striving towards the maintenance of glucose concentrations well within the physiologicalrange rather than merely outside of the pathological range. The only study from thosediscussed above to examine glucose concentrations after a duration of fasting that wouldrepresent that occurring as part of normal daily living (14 hours) found a fasting glucoseconcentration of 43 ± 01 mmol/l in children aged 3–15 years old (Lamers et al., 1985b).Guidelines now suggest that diabetes management should aim to keep plasma glucose above4 mmol/l (‘four should be the floor’ was recommended in a Diabetes UK report in 1996),and this is probably a reasonable recommendation for children.Both in research studies and in clinical care, hypoglycaemia is often subdivided intodegrees of severity based on the intervention required. An example of such a classificationthat was suggested by the International Society for Paediatric and Adolescent Diabetologists(ISPAD) is shown in Table 9.1. It should be noted that the usual adult definition of ‘mild’(i.e., self-treated) hypoglycaemia cannot be applied in young children who rely on theiradult carers for their diabetes management, because they are unlikely to be able to treat theepisodes themselves.

SIGNS AND SYMPTOMS OF HYPOGLYCAEMIA 193Table 9.1 Example of classification of hypoglycaemia (data sourced from ISPAD Consensus Guidelines,2000)GradeMild (Grade 1)Moderate (Grade 2)Severe (Grade 3)DescriptionAware of, responds to and self-treats the hypoglycaemiaCannot respond to hypoglycaemia and requires help fromsomeone else, but oral treatment is successfulSemi-conscious or unconscious or in coma ± convulsions andmay require parenteral therapy (glucagon or IV glucose)PREVALENCE OF HYPOGLYCAEMIAAs the definition of hypoglycaemia has been controversial, studies of prevalence have oftenused variable definitions, both for daytime hypoglycaemia and for that occurring duringsleep. Hypoglycaemia is notoriously under-reported, as episodes, particularly mild ones, arequickly forgotten. Even severe episodes may be overlooked. The individual themselves mayhave amnesia for the event and, if it occurred away from their normal environment, nobodymay document what took place.A number of studies have examined the prevalence of severe hypoglycaemia in thepaediatric population (Table 9.2). Some studies included only episodes of coma/convulsionwhereas others also included those events in which neurological impairment was severeenough to require intervention. Hypoglycaemia has been shown to be a significant problembut, given the methodological complexities, prevalence rates of severe hypoglycaemia havevaried greatly from 3.1 to 85.7 episodes per 100 patient-years. It is likely that less severeepisodes are much more common and under-reported.Nocturnal HypoglycaemiaStudies have also examined the prevalence of nocturnal hypoglycaemia. In these studiesa biochemical rather than a symptomatic definition of hypoglycaemia was used and hasbeen very variable, ranging from 3.0 to 3.8 mmol/l. The prevalence of hypoglycaemia hasvaried from 10 to 55% but it is noticeable that the more recent studies have detecteda higher prevalence (Beregszaszi et al., 1997; Lopez et al., 1997; Porter et al., 1997;Matyka et al., 1999a). The majority of these episodes do not wake the patient and, ineveryday life, a great number of episodes of nocturnal hypoglycaemia will be completelyundetected.SIGNS AND SYMPTOMS OF HYPOGLYCAEMIADaytime HypoglycaemiaThe classical symptoms of hypoglycaemia, as described in adults (Chapter 2), are classifiedinto three distinct groups: autonomic, neuroglycopenic and non-specific (Deary et al.,1993). In adults the sympathoadrenal response during hypoglycaemia is primarily responsiblefor the classical autonomic symptoms which alert the individual to the falling glucose

Table 9.2 Summary of studies examining prevalence of severe hypoglycaemia since the Diabetes Control and Complications trial in 1993StudyAge(years)Number ofsubjects Definition Study typeNo. episodes/ 100patient years CorrelationsDCCT (1994) 13–17 195 coma, seizure or requiringassistanceprospective 85.7 lower HbA 1cBognetti et al. (1997) 3.2–25.5 187 retrospective 14.9 younger ageMortensen et al.(1997)Nordfelt andLudvigsson (1997)0–18 2897 coma ± seizures cross-sectional 22 younger age;lower HbA 1c1–18 146 coma ± seizures;prospective 15–19lower age at onset;requiring assistanceprospective 10.1–12.6 longer disease duration;lower HbA 1cprospective 4.8younger age;13.1lower HbA 1cprospective 7.8younger age15.4lower HbA 1cDavis et al. (1997) 0–18 657 coma ± seizures;requiring assistanceDavis et al. (1998) 0–18 709 coma ± seizures;requiring assistanceRosilio et al. (1998) 1–19 2579 coma ± seizures orglucagon injectionTupola et al. (1998b) 1–24 376 coma ± seizures orglucagon injectioncross-sectional 45 lower HbA 1c ;more exercise;more daily bloodglucose measurementsprospective 3.1 lower HbA 1c ;higher insulin doseThomsett et al. (1999) 1–19 268 coma ± seizures retrospective 5 younger age;lower HbA 1c ;number of clinic visitsAllen et al. (2001) 0–34 415 coma prospective 7% of patients4 yrs afterdiagnosis;4% after6.5 yearslower HbA 1c ;older age

Levine et al. (2001) 7–16 300 episodes requiring outsideassistanceprospective 62 lower HbA 1c ;younger ageRewers et al. (2002) 0–19 1243 coma ± seizures prospective 19 young age;increased diabetesduration;lower HbA 1c ;underinsuranceCraig et al. (2002) 1.2–15.8 1190 coma ± seizures prospective 36 younger age;males;> 3 insulininjections/day;longer diabetes durationHoll et al. (2003) 11–18 872 coma ± seizures cross-sectional 17.9 not evaluatedVanelli et al. (2005) 1.6–17.1 3560 coma ± seizures ;episodes requiringparenteral therapy;episodes requiring outsideassistancecross-sectional 17.6 none apparentWagner et al. (2005) 0–9 6309 episodes requiring outsideassistanceprospective 22.6 younger age;longer diabetes duration;higher insulin dose;insulin regimen;centre experience

196 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESconcentration so that corrective action can be taken. The classical autonomic symptomsoccur in adults between 3.0 and 3.6 mmol/l and include: sweating, palpitations, hunger andshaking. If glucose concentrations fall further, neuroglycopenia will develop, usually at aglucose concentration around 2.8 mmol/l, and if it falls further the individual may not be ableto correct the hypoglycaemia themselves. The most common neuroglycopenic symptomsare: confusion, drowsiness, odd behaviour, speech difficulty and incoordination. If bloodglucose continues to fall, coma or convulsion could ensue although the glycaemic thresholdat which this occurs is not certain.The symptoms of hypoglycaemia in childhood differ from adults. In one study, childrenand parents were asked which symptoms and signs alerted them to an episode of hypoglycaemia(McCrimmon et al., 1995). The authors reported that symptoms did not separate intodistinct autonomic and neuroglycopenic categories and behavioural symptoms were prominent.Another study examined the frequency of hypoglycaemia in a group of children andadolescents using a three-month diary (Tupola et al., 1998a). Episodes of hypoglycaemia(defined as a blood glucose ≤ 3 mmol/l) were documented along with symptoms. Of thepatients (aged 2.5–21 years), 52% had a total of 287 episodes of hypoglycaemia, the majorityof which (77%) were mild. The most common presenting symptoms were weakness, tremor,hunger and drowsiness; 39% of symptoms were classified as ‘adrenergic’ (autonomic) and61% as neuroglycopenic or non-specific behavioural (Table 9.3). The dominant symptomswere different in different age groups of children. Those children less than six years ofage had fewer autonomic symptoms than adolescents, with the commonest symptom beingdrowsiness in young children and tremor in the older children (Tupola et al., 1998a).Nocturnal HypoglycaemiaThe majority of episodes of nocturnal hypoglycaemia do not awaken the child from sleepand thus go undetected, even in those who have normal awareness of hypoglycaemiaduring waking hours (Gale and Tattersall, 1979; Bendtson et al., 1993; Porter et al., 1996;Table 9.3 Reported symptoms during 221 episodes of mild hypoglycaemia(data sourced from Tupola et al., 1998b)Symptom Prevalence (%)autonomic tremor 22hunger 14sweating 4neuroglycopenic drowsiness 34irritability/aggressiveness 4dizziness 1poor concentration 1blurred vision 1non-specific weakness 7nausea 7abdominal pain 2headache 2tearfulness 1

RISK FACTORS FOR HYPOGLYCAEMIA 197Beregszaszi et al., 1997; Lopez et al., 1997; Matyka et al., 1999a). The reasons for thislack of awareness of nocturnal episodes of hypoglycaemia are unclear, but are discussedin Chapter 4. As symptom generation depends on intact autonomic and counterregulatorydefence mechanisms, it has been postulated that diminished counterregulation overnight mayresult in lack of arousal when hypoglycaemia occurs during sleep (Bendtson et al., 1993;Jones et al., 1998).RISK FACTORS FOR HYPOGLYCAEMIAThe discussion of risk factors naturally overlaps with that of prevention of hypoglycaemia,and a fuller review of specific interventions is provided in the section entitled ‘Prevention’ onpage 207. Table 9.3 presents the results of studies that have been performed to examine theprevalence of severe hypoglycaemia in childhood since the Diabetes Control and ComplicationsTrial (DCCT) was published (The Diabetes Control and Complications Trial ResearchGroup, 1993). Although a number of studies have suggested correlations between youngerage and strict glycaemic control, it is important to note that many individual episodes ofhypoglycaemia may be explained by missed meals or unplanned exercise, but this wouldnot have been addressed in epidemiological studies.Glycaemic ControlInsulin requirements vary with age and are approximately 0.5–1 U/kg/day before puberty and1.5–2 U/kg/day during adolescence, reflecting the insulin resistance that is present during thisperiod of rapid growth and development (Dunger, 1992). Despite numerous developmentsin terms of novel insulin preparations and modes of delivery, people with type 1 diabetesstill experience varying states of insulin deficiency or excess that are difficult to controland predict. This is probably most evident in adolescents with type 1 diabetes in whomperipheral hyperinsulinaemia is achieved in an attempt to replace adequate levels of insulinin the portal circulation during the pubertal growth spurt (Dunger, 1992).The DCCT highlighted the dilemma faced by all patients with type 1 diabetes (TheDiabetes Control and Complications Trial Research Group, 1993). Attempts at improvingglycaemic control, by intensifying diabetes management, in an effort to decrease the likelihoodof the long-term microvascular complications of diabetes led to a significant increasein the risk of severe hypoglycaemia (SH). In the DCCT, subjects in the intensified treatmentgroup had a three-fold higher risk of SH (The Diabetes Control and Complications TrialResearch Group, 1993). A group of 195 adolescents, aged between 13 and 17 years, took partin this trial (The Diabetes Control and Complications Trial Research Group, 1994). Althoughthe benefits of improved glycaemic control in terms of microvascular complications werestill significant, the adolescents found it more difficult to achieve the low HbA 1c concentrationthan adults (806 ± 013 versus 712 ± 003%; p

198 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESWe do not have rights to reproduce thisfigure electronicallyFigure 9.1 Rates of (a) Severe hypoglycaemia, and (b) average HbA 1c by calender year. Reproducedfrom Bulsara et al. (2004) with permission from The American Diabetes Associationhypoglycaemia was as common in those centres where metabolic control was poor, as inthose centres that achieved better control as judged by HbA 1c , suggesting that research doesnot always reflect clinical experience (Mortensen et al., 1997).Since 1993, ample opportunity has been present to refine the approaches to intensiveinsulin therapy and to improve education both for patients and physicians. Longitudinalstudies of the incidence of hypoglycaemia are unusual but one audit study from a largepaediatric clinic in Western Australia demonstrated an interesting trend in incidence of SHover a period of ten years (Bulsara et al., 2004). Over the first five years of the study,the incidence increased by 29% in conjunction with a decline in the average HbA 1c ofabout 0.2% per year. Despite a continued improvement in glycaemic control, the incidenceof SH appeared to plateau at this clinical centre suggesting that improved diabetesmanagement, from more effective insulin regimens or better education, can improve bloodglucose concentrations without a concomitant increase in incidence of hypoglycaemia(Figure 9.1).Nocturnal Insulin RequirementsThe mismatch of insulin delivery and insulin requirements on standard insulin regimensis particularly evident during the night and most episodes of severe hypoglycaemia occurduring sleep (Edge et al., 1990a; The Diabetes Control and Complications Trial ResearchGroup, 1997). Current insulin replacement regimens tend to result in hyperinsulinaemia inthe early part of the night, although physiological insulin requirement is at its nadir between24:00–03:00 hours, and so exacerbates the risk of hypoglycaemia at this time (Matykaet al., 1999a; Mohn et al., 1999; Ford-Adams et al., 2003). Insulin requirements then peakbetween 04:00–08:00 hours and a ‘dawn phenomenon’ occurs which can lead to fastinghyperglycaemia (Bolli and Gerich, 1984; Edge et al., 1990b) and is thought to result fromGH secretion during the later part of the night (De Feo et al., 1990; Edge et al., 1990b)

RISK FACTORS FOR HYPOGLYCAEMIA 199exacerbated further by the delayed effects of daytime physical activity on muscle glucosemetabolism and the prolonged period of fasting that occurs overnight, especially in youngchildren. This suggests that the overnight period is the time of greatest hypoglycaemia risk(The Diabetes Control and Complications Trial Research Group, 1997).Intensive Insulin RegimensFew studies have examined the impact of insulin regimen on the risk of hypoglycaemiain children. The DCCT did find a significantly higher risk of hypoglycaemia among the195 adolescents who participated in the study although this was a comparison of overallglycaemic control and not of specific regimens (The Diabetes Control and ComplicationsTrial Research Group, 1994). A number of studies examining prevalence of hypoglycaemiahave found an inverse correlation between hypoglycaemia risk and glycated haemoglobin(Table 9.2). The Hvidore Study Group has formed a collaboration between 21 internationalpaediatric centres from 18 countries (Holl et al., 2003). The group surveyed paediatricdiabetes management of 2873 children aged up to 18 years in 1995 and restudied 872 ofthese children in 1998. Although the use of multiple injection regimens increased from 42%to 71% this did not lead to an improvement in glycaemic control as judged by glycatedhaemoglobin concentrations. Although there was a tendency towards an increase in thefrequency of severe hypoglycaemia in the group of children/adolescents who had had anintensification of insulin regimen, this did not reach statistical significance, perhaps becauseof the low number of events recorded (Holl et al., 2003).Another study of more than 6000 children has suggested that injection regimen and centreexperience, as judged by the size of the clinic, may be significant risk factors for severehypoglycaemia (Wagner et al., 2005). In this study of children aged up to nine years, anincreased risk of hypoglycaemia was observed in those children taking four insulin injectionsdaily or on insulin pump therapy, compared to those children taking one to three injectionsdaily.It is important to note that even the more recent studies do not include data acquired sincethe introduction of the insulin analogues or the use of more physiological and intensiveinsulin regimens. Small studies of a few children in which insulin analogues have beencompared with human insulins suggest that the risk of hypoglycaemia may be lower withanalogues without compromising glycaemic control (see later section on hypoglycaemiaprevention on page 209).DietChildren with type 1 diabetes have the same nutritional requirements as children withoutdiabetes. However, meals and snacks containing a high proportion of carbohydrate, have tobe regularly distributed throughout the day to avoid the extremes of hypo- and hyperglycaemia(Magrath et al., 1993). This can be an issue for children who do not want to bedifferent from their peers and do not want to eat at times when there friends are playing.Toddlers can present a special problem as many do not eat regular meals but graze duringthe day.Surprisingly, there has been little systematic study of the role of both quantity andquality of dietary components on the risk of hypoglycaemia. However, some studies of the

200 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESprevalence of hypoglycaemia have found that a number of episodes can be attributed tomissed meals (Daneman et al., 1989; Davis et al., 1997; Tupola et al., 1998b). The impactof dietary interventions in the avoidance of hypoglycaemia, mainly during sleep, have beenexamined, and this is discussed later in the section on hypoglycaemia prevention.Physical ActivityAs early as 1926, it was found that exercise could potentiate the hypoglycaemic effect ofinsulin in patients with type 1 diabetes (Lawrence, 1926). During the first 5–10 minutes ofexercising, muscle glycogen is used as the primary source of energy (Price et al., 1994).Subsequently, fuel is provided increasingly by circulating glucose, through gluconeogenesisand free fatty acids (FFAs), the release of which are under hormonal control and dependentpredominantly on the portal glucagon : insulin ratio (Ahlborg et al., 1974). The acute effectsof exercise are followed by restoration of the metabolic milieu. Muscle glucose uptakeremains elevated as glycogen stores are replenished and although insulin sensitivity isenhanced in the period after exercise, increased glucose uptake by skeletal muscle can occureven in the absence of insulin (Cartee and Holloszy, 1990). The time taken to restore muscleglycogen to pre-exercise levels will depend on the intensity and duration of the exerciseperformed, and the timing and amount of dietary carbohydrate intake, but it can take severalhours – typically 6–20 hours (Ivy and Holloszy, 1981; Richter, 1996).Until recently there has been little systematic study of the impact of exercise in childhoodon glucose homeostasis. One study used continuous glucose monitoring to study a standardisedexercise protocol in a group of children who were using continuous subcutaneousinsulin infusion (CSII). Glucose profiles were examined both during and after exercise ona cycle ergometer with the infusion pump either switched on or off (Admon et al., 2005).Hypoglycaemia was more common after exercise than during it, and this was true whetherCSII was on or off. All subjects had one to three episodes of symptomatic hypoglycaemiawithin 2.5 to 12 hours after exercise and four subjects had asymptomatic hypoglycaemiaduring exercise, only one of whom had consumed extra carbohydrate because their preexerciseblood glucose had been below 5.5 mmol/l. Another study examined the impact ofdaytime exercise on overnight blood glucose profiles in 50 subjects aged 10–18 years onintensive insulin regimens (Tsalikian et al., 2005). On one occasion they were studied duringa day of afternoon exercise, involving four periods of 15 minutes each on a treadmill at aheart rate estimated to be 55% of maximum effort for this age group, and on a separateoccasion during a rest day. In this study, 22% of subjects developed hypoglycaemia duringexercise; overnight hypoglycaemia was more common during the night after the afternoonexercise than during a night after the rest day (Tsalikian et al., 2005).AgeStudies of prevalence of hypoglycaemia have consistently found that younger children,especially those under the age of five years, are at increased risk of hypoglycaemia. Thismay be a consequence of increased insulin sensitivity, irregular eating patterns or impairedsymptomatic awareness.

COUNTERREGULATION IN CHILDHOOD 201GeneticsRecent studies suggest that molecular markers may influence hypoglycaemia risk. A polymorphismin the gene encoding for angiotensin converting enzyme (ACE) has been describedindicating the presence (insertion, I) or absence (deletion, D) of a 287 base pair sequencewithin intron 16 resulting in three genotypes: II, ID and DD. These genotypes are stronglyrelated to serum ACE concentration with the highest values in DD and the lowest valuesin II genotypes (Rigat et al., 1990). Danish studies of adults with type 1 diabetes observedthat patients with an II (insertion) genotype, who had a low serum ACE activity, had asignificantly lower frequency of severe hypoglycaemia (Pedersen-Bjergaard et al., 2001;Pedersen-Bjergaard et al., 2003). A Swedish study of children with type 1 diabetes hasreported a six-fold lower frequency of severe hypoglycaemia in those patients who had lowserum ACE activity (Nordfelt and Samuelsson, 2003). However, a study of 585 children andadolescents in an Australian centre has not confirmed this association (Bulsara et al., 2007),so this remains controversial.Clinic ExperienceRecent therapeutic developments, with the availability of novel insulin analogues and thegreater use of intensive insulin regimens, have required re-education not only of patientsbut also the multidisciplinary team. Few studies have examined the impact of the clinicstructure on risk of hypoglycaemia. One large multicentre study in Germany did show alink between small clinic size (< 50 children) and an increased risk of severe hypoglycaemia(Wagner et al., 2005). Another longitudinal study of prevalence of severe hypoglycaemiain Australia showed an increase in hypoglycaemia risk as the mean glycatedhaemoglobin in the clinic declined. However, over the final five years of the study the riskof hypoglycaemia plateaued while the HbA 1c continued to decrease, suggesting the possiblebenefit of familiarity among patients, healthcare professionals or both (Bulsara et al., 2004)(Figure 9.1).COUNTERREGULATION IN CHILDHOODThe physiology of counterregulation is the subject of Chapters 1 and 6, but a brief overviewof studies of counterregulation in childhood will be presented here. Despite the ethical andpractical problems of inducing hypoglycaemia in children for research purposes, a numberof studies have been performed (Amiel et al., 1987; Brambilla et al., 1987; Singer-Granicket al., 1988; Hoffman et al., 1991; Jones et al., 1991; Bjorgaas et al., 1997a; Ross et al.,2005). These studies have all examined experimentally-induced hypoglycaemia, either usingan insulin infusion method or a hyperinsulinaemic, hypoglycaemic glucose clamp technique.The results of individual hormonal responses are discussed briefly.GlucagonAs in adults the glucagon response during hypoglycaemia is lost in children with diabetes(Amiel et al., 1987; Jones et al., 1991; Ross et al., 2005). This is also the case in toddlers,

202 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESaged 18–57 months old, who have a very short duration of diabetes (Brambilla et al., 1987).This means that individuals with diabetes are more reliant on adequate epinephrine responsesto correct hypoglycaemia.EpinephrineThe majority of studies suggest that children have exaggerated epinephrine responses tohypoglycaemia, with peak values that are two-fold higher than those found in adults (Amielet al., 1987). Data from prepubertal children have been analysed separately from pubertalchildren and no significant differences were found in epinephrine responses (Amiel et al.,1987; Ross et al., 2005). One study, using an insulin infusion as opposed to a hyperinsulinaemicclamp, did suggest that epinephrine responses were blunted in a group ofpoorly-controlled adolescents (Bjorgaas et al., 1997a). The reason for the discrepancy inresults is not clear.Glucose thresholds for counterregulatory responses have received very little attention.In a study of poorly-controlled adolescents (average total HbA 1 : 15%) glucose thresholdsfor epinephrine secretion were significantly higher, with the poorly-controlled adolescentsreleasing epinephrine at a glucose concentration of 4.7 mmol/l compared to 3.9 mmol/l inhealthy adolescents (Jones et al., 1991).Growth HormoneNone of the reported studies have identified defects in GH release during hypoglycaemia.Generally GH has been found to increase in response to hypoglycaemia both in childrenwith diabetes and in non-diabetic controls (Amiel et al., 1987; Brambilla et al., 1987; Joneset al., 1991).CortisolCortisol, like growth hormone, becomes more important as hypoglycaemia becomesprolonged. Studies of hypoglycaemia in children have shown variable results. Brambillafound no increase in cortisol in either the diabetic or control group of toddlers studied(Brambilla et al., 1987), whereas others have documented an increase in cortisol both inchildren with and without diabetes (Amiel et al., 1987; Jones et al., 1991).Effect of Sleep Stage on CounterregulationOne study of nocturnal hypoglycaemia in prepubertal children on conventional insulinregimens, found that the median glucose nadir during episodes of nocturnal hypoglycaemiawas 1.9 mmol/l (range: 1.1–3.3 mmol/l) and the median duration was 270 minutes (range:30–630 minutes) (Matyka et al., 1999a). This is similar to adults in whom the averageduration of hypoglycaemia (glucose below 2 mmol/l) during the night was found to be threehours in a group of adults with poorly-controlled diabetes (Gale and Tattersall, 1979).Conventional wisdom would argue that prolonged episodes of hypoglycaemia are unusualas hypoglycaemia is promptly corrected by counterregulatory defence mechanisms. However,

CONSEQUENCES OF HYPOGLYCAEMIA 203studies of nocturnal hypoglycaemia suggest that counterregulatory responses may be bluntedduring sleep (see Chapter 4). A study of spontaneous nocturnal hypoglycaemia in prepubertalchildren with diabetes demonstrated blunted and delayed counterregulatory hormoneresponses during sleep (Matyka et al., 1999b). Peak epinephrine response was only 0.9 nmol/land there was a marked delay between the glucose falling below 3.5 mmol/l and a significantrise in epinephrine; the mean delay was 170 minutes. Another study examined counterregulatoryhormone responses during hypoglycaemia that was experimentally-induced duringthe time of night when slow wave sleep predominates, and these responses were comparedwith those to hypoglycaemia induced during the day with the subjects awake and then againwhen they were awake during the night (Jones et al., 1998). Adolescents with diabetesand healthy controls participated in the study. Epinephrine responses during hypoglycaemiawere blunted when hypoglycaemia was induced during slow wave sleep compared to whenhypoglycaemia was induced when subjects were awake during the day or during the night(Jones et al., 1998). Studies of the physiology of sleep have demonstrated both variationsin autonomic tone and cerebral glucose metabolism going from slow wave sleep through torapid eye movement sleep which may influence the counterregulatory response during sleep(Maquet et al., 1990; Parmeggiani and Morrison, 1990).CONSEQUENCES OF HYPOGLYCAEMIACognitive ImpairmentSevere hypoglycaemia can cause catastrophic cerebral damage when it is profound andprolonged, and in very young children this may be a risk associated with a variety of causes(Lucas et al., 1988). Glucose is critical not only as the major fuel for cerebral metabolism butalso as a precursor of essential substrates which are essential for normal brain development(Glazer and Weber, 1971). Concerns have been raised that recurrent hypoglycaemia couldaffect long-term academic achievement in children with type 1 diabetes, from the effectsof hypoglycaemia disrupting school performance to the possibility of damage accumulatingover time.Acute effectsFew studies have studied the impact of acute hypoglycaemia on cognitive performance in childrenor adolescents. One study of the effects of experimentally-induced mild hypoglycaemia(3.1–3.6 mmol/l) found decrements in tests of mental flexibility and on measures that requiredplanning and decision making and a rapid response, although the results were variable withinsubjects (Ryan et al., 1990). The learning ability of children, who spend much of their day atschool assimilating information, could be seriously compromised if they experience frequentepisodes of even mild hypoglycaemia during their time in class. Children can also be affectedif they miss lessons because of severe hypoglycaemic events and could be compromised byasymptomatic episodes of hypoglycaemia that go undetected and therefore untreated. Thereis some evidence to suggest that children may be particularly susceptible to mild episodes ofhypoglycaemia – studies examining P300 evoked potentials and EEG changes in response to

204 HYPOGLYCAEMIA IN CHILDREN WITH DIABETEShypoglycaemia have found that abnormalities commence at a higher blood glucose level inchildren than in adults (Jones et al., 1995; Bjorgaas et al., 1998).The effect of nocturnal hypoglycaemia on cognitive function has received little attention.Cognitive performance could be directly affected by hypoglycaemia, but also by sleepdisturbance. The few studies of nocturnal hypoglycaemia in children and adolescents withtype 1 diabetes have shown no deleterious effect on overall sleep physiology (Bendtsonet al., 1992; Porter et al., 1996; Matyka et al. 2000, Pillar et al., 2003) or on cognitivefunction the following morning.Long-term effectsEarly studies of children with diabetes suggested that they were ‘mentally superior’ (Westet al., 1934). However, in recent years concern has been expressed about the potential impactof recurrent episodes of hypoglycaemia on intellectual performance. It is beyond the scopeof this chapter to provide a comprehensive review on this topic and only more recent studiesare reviewed.The developing brain is extremely vulnerable to all types of cerebral trauma. Studies ofchildren who have experienced closed head injuries suggest that the consequences may bedelayed, with subtle cerebral damage becoming evident with time as normal developmentalmilestones are delayed. A large number of studies have examined the impact of recurrenthypoglycaemia in childhood (Ryan et al., 1985; Golden et al., 1989; Bjorgaas et al., 1997b;Hershey et al., 1999; Rovet and Ehrlich, 1999; Northam et al., 2001; Wysocki et al.,2003). Almost without exception, the results have shown a possible link between severehypoglycaemia and decrements in cognitive performance and that those children most atrisk of cognitive impairment have been those diagnosed early in life – usually less than fiveyears of age (Ryan et al., 1985; Golden et al., 1989; Bjorgaas et al., 1997b; Northam et al.,2001). Deficiencies have been found in several cognitive domains but are more likely inthose originating in the frontal lobe. One of the most impressive longitudinal studies hasbeen that of Northam and colleagues. Cognitive performance was assessed in a large groupof children (123 at baseline) with newly-diagnosed diabetes, aged 3–14 years, who werecompared to healthy controls at three months, two years and six years following diagnosis(Northam et al., 2001). At six years, children with diabetes performed more poorly inmeasures of intelligence, attention, processing speed, long-term memory and executive skills.Attention, processing speed and executive skills were especially affected in those childrenwho had developed diabetes when less than four years of age (Northam et al., 2001). Severehypoglycaemia was associated with lower verbal and full-scale intelligence quotient (IQ)scores. The authors concluded that recurrent hypoglycaemia was a potential explanation forthese cognitive deficits but could not exclude an effect of chronic hyperglycaemia.Another study has performed a cognitive test battery and structural neuroimaging usingMagnetic Resonance Imaging in a group of young adults with type 1 diabetes and comparedthe findings between those with either early onset disease, defined as less than sevenyears, and late onset disease (Ferguson et al., 2005). Physiological risk variables such asdiabetes duration, evidence of microvascular disease and retrospective reporting of precedingexposure to severe hypoglycaemia were also assessed. The patients with early onset diabeteswere found to have deficits in non-verbal intelligence, information processing ability andpsychomotor speed. The authors also found higher ventricular volumes and higher frequency

CONSEQUENCES OF HYPOGLYCAEMIA 205of mild ventricular atrophy in those with early onset diabetes. None of the findings wererelated to the presence of microvascular disease or diabetes duration, suggesting that thecumulative effects of hyperglycaemia were unlikely to be causative. However, no definite linkwas confirmed between exposure to severe hypoglycaemia and any of these defects, althoughthe study was limited by retrospective reporting of the history of preceding hypoglycaemiaover a long period of time (Ferguson et al., 2005). Another study did not find a differencein tests of cognitive function over a period of 18 months in 142 patients taking part in astudy examining the impact of intensive therapy versus usual care (Wysocki et al., 2003).The study group had a wide age range from 6–15 years and data from younger childrenwere not analysed separately.Although there are methodological problems in designing studies to assess long-termcognitive function in children who have an ongoing chronic disorder, these studies do raiseanxieties. It is felt that until hypoglycaemia can be reliably avoided, glycaemic control shouldbe less intensively managed in younger children to avoid the risk of severe hypoglycaemia.This would put younger children at greater risk of developing the long-term microvascularcomplications of diabetes in an attempt to avoid the possible cognitive defects that may beassociated with recurrent hypoglycaemia.Hypoglycaemic HemiplegiaThis is a rare complication of acute hypoglycaemia in which the patient recovers fromthe hypoglycaemia with a transient hemiparesis lasting no more than 24 hours. Whenneuroimaging is performed it is rare to find an abnormality. There is no evidence of anysevere sequelae to this neurological manifestation of severe neuroglycopenia (Pocecco andRonfani, 1998).Fear of HypoglycaemiaBoth the children with diabetes and their parents worry about the prospect of a severe episodeof hypoglycaemia (Gold et al., 1997; Clarke et al., 1998; Gonder-Frederick et al., 1997;Marrero et al., 1997; Nordfelt and Ludvigson, 2005) (see Chapter 14). In one study severehypoglycaemia caused more fear than the prospect of an episode of diabetic ketoacidosis(Nordfelt and Ludvigson, 2005). Although there is little evidence that this modifies behaviourto attempt hypoglycaemia avoidance, such as relaxing glycaemic control or eating moresnacks, the adverse effects on quality of life should not be underestimated (Gold et al.,1997).Prediction of Nocturnal HypoglycaemiaOvernight glucose profiles have been shown to be extremely variable. As a result studies ofovernight hypoglycaemia have been unable to provide a ‘safe’ glucose value with which togo to bed. What is more useful is the fasting blood glucose concentration on the followingmorning. One study showed that the median fasting blood glucose at 07:00 hours wassignificantly lower following hypoglycaemia than a night with no hypoglycaemia (3.7[1.4–10.6] versus 8.5 [3.8–19.2] mmol/l, p = 000001) (Matyka et al., 1999a) (Figure 9.2).

206 HYPOGLYCAEMIA IN CHILDREN WITH DIABETES2015Glucose (mmol / l)1050hypoglycaemiano hypoglycaemiaFigure 9.2 Glucose values at 07.00 hours following a night of hypoglycaemia () and a night of nohypoglycaemia (△). Reproduced by permission of K. Matyka, PhD thesis ‘Nocturnal hypoglycaemiain prepubertal children with type 1 diabetes mellitus’, University of LondonThe occurrence of a ‘Somogyi phenomenon’, whereby overnight hypoglycaemia promotesglucose counterregulation and causes fasting hyperglycaemia, has not been demonstrated instudies of nocturnal hypoglycaemia in children (Porter et al., 1996; Beregszaszi et al., 1997;Matyka et al., 1999a).MANAGEMENT OF HYPOGLYCAEMIAThe management of acute hypoglycaemia will depend on the severity of the episode. TheInternational Society for Paediatric and Adolescent Diabetologists has provided guidelinesfor the management of acute episodes of hypoglycaemia based on severity (ISPAD, 2000).Blood glucose measurement is the only way to confirm hypoglycaemia if the diagnosisis uncertain and also confirm the return of the blood glucose to normal after treatment.Figure 9.3 shows the flow diagram for the management of hypoglycaemia.PreventionWhen faced in clinic with a child who is having recurrent episodes of hypoglycaemia, adetailed history should be obtained regarding the timing of hypoglycaemia, insulin regimen,dietary intake and the relation to periods of physical activity. This will enable an assessmentto be made of possible risk factors and inform how these may be avoided. If no obviouscause is found then other pathology should be sought, such as coincidental coeliac diseaseor the possibility of Addison’s disease, although these are relatively rare causes of recurrenthypoglycaemia (see Chapter 3).When contemplating preventive management the following aspects should be considered.

MANAGEMENT OF HYPOGLYCAEMIA 207Check blood glucose – Glucose meter test and formal laboratory glucose if possibleHypoglycaemiaMildSevereCONSCIOUSi.e. gag reflex intactUNCONSCIOUSOral glucosee.g. 5–15 grams of glucoseor 100 mls sweet drinkGlucogelOne ampoule orallyI.M. Glucagon5 yrs 1.0 mgSuccessNo successIf no response in15 mins give 1–2mls/kg of 10%dextrose I.V.Repeat until thereis a clinicalresponse.As symptoms improve or normoglycaemia is restored add oral complex carbohydratee.g. biscuit, bread and so on.If unable to tolerate oral carbohydrate may need a glucose infusione.g. 5–10% glucose at maintenance rateFigure 9.3Management of hypoglycaemiaEducationGreat importance is placed on education in the management of type 1 diabetes and structurededucation programmes are now an essential part of any diabetes service provision. Despitethis, little validation has been made of the use of structured educational programmes inchildren. One study from Scandinavia has shown the benefits of a focused education package(Nordfelt et al., 2003). In this study families were given both written and video informationon diabetes. One group were given general information on diabetes management and theother was given material to educate about the importance and means of hypoglycaemiaavoidance. Although no differences in glycated haemoglobin were found between the two

208 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESgroups, those who had the targeted intervention had a significantly reduced rate of severehypoglycaemia (Nordfelt et al., 2003).InsulinAs noted earlier, the DCCT suggested that attempts to intensify insulin regimens may increasethe risk of severe hypoglycaemia. Recent introductions of analogue insulins, however,indicate that improved glycaemic control does not always lead to hypoglycaemia, althoughfew studies have been designed specifically to examine the benefits of different insulinregimens on the risk of hypoglycaemia. One study that compared insulin lispro with soluble(short-acting) insulin as part of a basal bolus regimen, showed small but statistically significantreductions in the prevalence of hypoglycaemia over a 30 day period when insulinlispro was being used (Holcombe et al., 2002). Other studies have found benefits of insulinanalogue regimens on nocturnal hypoglycaemia. One randomised cross-over study in adolescentscompared insulin lispro and glargine, as part of a daily multiple injection regimen,to human soluble and isophane insulins. Nocturnal hypoglycaemia was 43% lower withthe analogue regimen, although no difference was observed in self-reported symptomatichypoglycaemia (Murphy et al., 2003). Another study in prepubertal children examined thebenefits of a thrice daily insulin regimen, where the evening dose of mixed insulin wasreplaced by a rapid-acting insulin analogue with the evening meal and isophane insulin beforebed (Ford-Adams et al., 2003). Although there was no difference in glycated haemoglobinbetween the two treatment arms, the prevalence of hypoglycaemia was lower in the earlypart of the night (22.00–04.00 hours) when the analogue was used.Although not every patient is suited to using insulin pump therapy, clinic-based studiesof CSII therapy have shown that more stable blood glucose control can be achieved withoutan increased risk of hypoglycaemia. In an American study describing the experience ofusing insulin pumps in a paediatric clinic, it was found that 50 adolescents on multipledaily injections experienced 134 episodes of severe hypoglycaemia per 100 patient-yearscompared to 76 episodes per 100 patient-years in the 25 adolescents who opted for pumptherapy (Boland et al., 1999).DietFew studies have examined the impact of dietary interventions on hypoglycaemia risk exceptfor nocturnal hypoglycaemia. The major dietary modification has been that of the introductionof a larger proportion of starch, as a form of long-acting carbohydrate, as part of the eveningsnack (Ververs et al., 1993; Kaufman et al., 1995; Detlofson et al., 1999; Matyka et al.,1999a). In these studies the benefits of starch have been inconsistent. One study found alower frequency of nocturnal hypoglycaemia, although capillary sampling was performedonly intermittently during the night and some episodes of hypoglycaemia may have beenundetected (Kaufman et al., 1995). Others found no beneficial effect of cornstarch on theprevention of nocturnal hypoglycaemia although blood glucose concentrations fell moreslowly, but in one study this occurred at the expense of promoting hyperglycaemia (Ververset al., 1993; Matyka et al., 1999a). Although not designed to examine the impact of diet onhypoglycaemia, a study of a low glycaemic index diet has been shown to improve glycaemic

CONCLUSIONS 209control without an increase in rate of hypoglycaemia, and appeared to have enhanced qualityof life (Gilbertson et al., 2001)ExerciseWhen adequate plasma insulin concentrations are available, exercise can lead to acutehypoglycaemia. However if exercise is performed at a time of relative insulin deficiency,hyperglycaemia with ketosis can occur. In addition, delayed hypoglycaemia may occur asmuscle glycogen stores recover mainly overnight (Admon et al., 2005; Tsalikian et al., 2005).Although few data are currently available regarding the most appropriate management ofplanned periods of physical activity, a number of guidelines have been proposed. The InternationalSociety for Paediatric and Adolescent Diabetologists has published guidelines onthe Internet (www.ispad.org). These recommend that careful monitoring of blood glucose isessential to match food and insulin to the intensity of exercise and that a reduction of insulinshould be considered. Additional slowly absorbed carbohydrate will be necessary, especiallyat bedtime, if exercise has been performed in the afternoon or early evening. From the dataavailable so far (Admon et al., 2005; Tsalikian et al., 2005), these guidelines do seem areasonable approach to the avoidance of both exercise related hypo- and hyperglycaemia.It is important, however, to work with the child and family to provide individually-tailoredrecommendations that are tried and tested for the child. The management of unpredictableepisodes of physical activity are likely to remain a problem until a cure for diabetesis found.CONCLUSIONS• Hypoglycaemia is a common problem for children with type 1 diabetes, especially youngchildren, and for their families.• Behavioural symptoms are more common in childhood than more typical autonomicsymptoms. The majority of episodes of nocturnal hypoglycaemia are totally asymptomatic.• Episodes of hypoglycaemia remain a significant barrier when striving for a degree ofglycaemic control that will delay or prevent the development of the microvascular complicationsof diabetes.• Glucagon responses during hypoglycaemia are lost early in the course of type 1 diabetes.Epinephrine responses during overnight hypoglycaemia are blunted in both healthy childrenand those with type 1 diabetes and may contribute to the lack of symptoms of manyepisodes of hypoglycaemia overnight.• Concerns remain about the possible long-term implications of hypoglycaemia in terms ofcognitive dysfunction although hyperglycaemia may also be important.• More recent data suggest that novel insulin analogues and regimens may enable improvementsin glycaemic control to be achieved without a concomitant increase in the risk ofhypoglycaemia.

210 HYPOGLYCAEMIA IN CHILDREN WITH DIABETES• It is important to consider both the child and their family when assessing the causes ofrecurrent episodes of hypoglycaemia and when suggesting possible avoidance strategies.• All children with diabetes should have a personalised action plan for hypoglycaemiaavoidance.REFERENCESAdmon G, Weinstein Y, Falk B, Weintrob N, Benzaquen H, Ofan R et al. (2005). Exercise with andwithout an insulin pump among children and adolescents with type 1 diabetes mellitus. Pediatrics116: 348–55.Ahlborg G, Felig P, Hagenfeldt L, Hemdler R, Wahren J (1974). Substrate turnover during prolongedexercise in man. Journal of Clinical Investigation 53: 1080–90.Allen C, LeCaire T, Palta M, Daniles K, Meredith M, D’Alessio DJ, for the Wisconsin DiabetesRegistry Project (2001) Risk factors for frequent and severe hypoglycemia in type 1 diabetes.Diabetes Care 24: 1878–81.Amiel SA, Simonson DC, Sherwin RS, Lauritano AA, Tamborlane WV (1987). Exaggeratedepinephrine responses to hypoglycemia in normal and insulin-dependent diabetic children. Journalof Pediatrics 110: 832–7.Bendtson I, Gade J, Thomsen CE, Rosenfalck A, Wildschiodtz G (1992). Sleep disturbances in IDDMpatients with nocturnal hypoglycemia. Sleep 15: 74–81Bendtson I, Rosenfalck AM, Binder C (1993). Nocturnal versus diurnal counterregulation to hypoglycemiain type 1 (insulin-dependent) diabetic patients. Acta Endocrinologica 128: 109–15.Beregszaszi M, Tubiana-Rufi N, Benali K, Noel M, Bloch J, Czernichow P (1997). Nocturnal hypoglycemiain children and adolescents with insulin-dependent diabetes mellitus: Prevalence and riskfactors. Journal of Pediatrics 131: 27–33.Bier DM, Leake RD, Haymond MW, Arnold KJ, Gruenke LD, Sperling MA, Kipnis DM (1977).Measurement of ‘true’ glucose production rates in infancy and childhood with 6,6-dideuteroglucose.Diabetes 26: 1016–23.Bjorgaas M, Vik T, Sand T, Birkeland K, Sager G, Vea H, Jorde R (1997a). Counterregulatory andsymptom responses to hypoglycaemia in diabetic children. Diabetic Medicine 14: 433–41.Bjorgaas M, Gimse R, Vik T, Sand T (1997b). Cognitive function in type 1 diabetic children with andwithout episodes of severe hypoglycaemia. Acta Paediatrica 86: 148–53.Bjorgaas M, Sand T, Vik T, Jorde R (1998). Quantitative EEG during controlled hypoglycaemia indiabetic and non-diabetic children. Diabetic Medicine 15: 30–7.Bognetti F, Brunelli A, Meschi F, Viscardi M, Bonfanti R, Chiumello G (1997). Frequency andcorrelates of severe hypoglycaemia in children and adolescents with diabetes mellitus. EuropeanJournal of Paediatrics 156: 589–91.Boland EA, Grey M, Oesterle A, Fredrickson L, Tamborlane WV (1999). Continuous subcutaneousinsulin infusion. A new way to lower risk of severe hypoglycemia, improve metabolic control, andenhance coping in adolescents with type 1 diabetes. Diabetes Care 22: 1779–84.Bolli G, Gerich J (1984). The dawn phenomenon – a common occurrence in both non-insulin dependentand insulin dependent diabetes mellitus. New England Journal of Medicine 310: 746–50.Brambilla P, Bougneres PF, Santiago JV, Chaussin JL, Pouplard A, Castano L (1987). Glucose counterregulationin pre-school-age children with recurrent hypoglycemia during conventional treatment.Diabetes 36: 300–4.Bulsara MK, Holman CDJ, Davis EA, Jones TW (2004). The impact of a decade of changing treatmenton rates of severe hypoglycemia in a population-based cohort of children with type 1 diabetes.Diabetes Care 27: 2293–8.

REFERENCES 211Bulsara MK, Holman CDJ, van Bockxmeer FM, Davis EA, Gallego PH, Beilby JP et al. (2007). Therelationship between ACE genotype and risk of severe hypoglycaemia in a large population-basedcohort of children and adolescents with type 1 diabetes. Diabetologia 50: 965–71.Cartee GD, Holloszy JO (1990). Exercise increases susceptibility of muscle glucose transport toactivation by various stimuli. American Journal of Endocrinology 258: E390–3.Chaussain J-L (1973). Glycemic response to 24-hour fast in normal children and children with ketotichypoglycemia. Journal of Pediatrics 82: 438–43.Chaussain J-L, Georges P, Olive G, Job J-C (1974). Glycemic response to 24-hour fast in normalchildren and children with ketotic hypoglycemia: II Hormonal and metabolic changes. Journal ofPediatrics 85: 776–81.Chaussain J-L, Georges P, Calzada L, Job J-C (1977). Glycemic response to 24-hour fast in normalchildren: III Influence of age. Journal of Pediatrics 91: 711–4.Clarke WL, Gonder-Frederick A, Snyder AL, Cox DJ (1998). Maternal fear of hypoglycemia intheir children with insulin dependent diabetes mellitus. Journal of Pediatric Endocrinology andMetabolism 11: 189–94.Craig ME, Handelsman P, Donaghue KC, Chan A, Blades B, Laina R et al., NSW/ACTHbA(1c) Study Group (2002). Predictors of glycaemic control and hypoglycaemia in childrenand adolescents with type 1 diabetes from NSW and ACT. Medical Journal of Australia 177:235–8.Daneman D, Frank M, Perlman K, Tamm J, Ehrlich R (1989). Severe hypoglycemia in children withinsulin-dependent diabetes mellitus: frequency and predisposing factors. Journal of Pediatrics 5:681–5.Danne T, Weber B, Hartmann R, Enders I, Burger W, Hovener G (1994). Long-term glycemic controlhas a nonlinear association to the frequency of background retinopathy in adolescents with diabetes.Follow-up of the Berlin Retinopathy Study. Diabetes Care 17: 1390–6.Davis EA, Keating B, Byrne GC, Russell M, Jones TW (1997). Hypoglycemia: incidence and clinicalpredictors in a large population based sample of children and adolescents with IDDM. DiabetesCare 20: 20–5.Davis EA, Keating B, Byrne GC, Russell, Jones TW (1998). Impact of improved glycaemic controlon rates of hypoglycaemia in IDDM. Archives of Disease in Childhood 78: 111–5.Deary IJ, Hepburn DA, MacLeod KM, Frier BM (1993). Partitioning the symptoms of hypoglycaemiausing multi-sample confirmatory factor analysis. Diabetologia 36: 771–7.De Feo P, Torlone E, Fanelli C, Santeusanio F, Brunetti P, Bolli GB (1990). Nocturnal spikes of growthhormone secretion cause the dawn phenomenon in type 1 (insulin dependent) diabetes mellitusby decreasing hepatic (and extrahepatic) sensitivity to insulin in the absence of insulin waning.Diabetologia 33: 52–9.Detlofson I, Kroon M, Aman J (1999). Oral bedtime cornstarch supplementation reduces the risk fornocturnal hypoglycaemia in young children with type 1 diabetes. Acta Paediatrica 88: 595–7.Dunger DB (1992). Diabetes in puberty. Archives of Disease in Childhood 67: 569–70.Edge JA, Matthews DR, Dunger DB (1990a). Failure of current insulin regimes to meet the overnightrequirements of adolescents with insulin-dependent diabetes. Diabetes Research 15: 109–12.Edge JA, Matthews DR, Dunger DB (1990b). The dawn phenomenon is related to overnight growthhormone release in adolescent diabetics. Clinical Endocrinology 33: 729–37.Ferguson SC, Blane A, Wardlaw J, Frier BM, Perros P, McCrimmon RJ, Deary IJ (2005). Influenceof an early-onset age of type 1 diabetes on cerebral structure and cognitive function. Diabetes Care28: 1431–7.Ford-Adams ME, Murphy NP, Moore EJ, Edge JA, Ong KL, Watts AP et al. (2003). Insulin lispro:a potential role in preventing nocturnal hypoglycaemia in young children with diabetes mellitus.Diabetic Medicine 20: 656–60.Gale EAM, Tattersall RB (1979). Unrecognised nocturnal hypoglycaemia in insulin-treated diabetics.Lancet i: 1049–52.

212 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESGilbertson HR, Brand-Miller JC, Thorburn AW, Evans S, Chondros P, Werther GA (2001). The effectof flexible low glycemic index dietary advice versus measured carbohydrate exchange diets onglycemic control in children with type 1 diabetes. Diabetes Care 24: 1137–43.Glazer RI, Weber G (1971). Incorporation of [6- 3 H] glucose into lipid, protein, RNA and DNA slicesof differentiating rat cerebral cortex. Journal of Neurochemistry 18: 1569–76.Gold AE, Deary IJ, Frier BM (1997). Hypoglycaemia and non-cognitive aspects of psychologicalfunction in insulin-dependent (type 1) diabetes mellitus (IDDM). Diabetic Medicine 14: 111–8.GoldenMP,IngersollGM,BrackCJ,RussellBA,WrightJC,HubertyTJ(1989).Longitudinalrelationshipof asymptomatic hypoglycemia to cognitive function in IDDM. Diabetes Care 12: 89–93.Gonder-Frederick L, Cox D, Kovatchev B, Julian D, Clarke W (1997). The psychosocial impact ofsevere hypoglycemic episodes on spouses of patients with IDDM. Diabetes Care. 20: 1543–6.Haymond MW, Karl IE, Clarke WL, Pagliara AS, Santiago JV (1982). Differences in circulatinggluconeogenic substrates during short-term fasting in men, women and children. Metabolism 31:33–42.Hershey T, Bhargava N, Sadler M, White NH, Craft S (1999). Conventional versus intensive diabetestherapy in children with type 1 diabetes – effects on memory and motor speed. Diabetes Care 22:1318–24.Hoffman RP, Singer-Granick C, Drash AL, Becker DJ (1991). Plasma catecholamine responses tohypoglycemia in children and adolescents with IDDM. Diabetes Care 14: 81–8.Holcombe JH, Zalani S, Arora VK, Mast CJ (2002). Lispro in Adolescents Study Group. Comparisonof insulin lispro with regular human insulin for the treatment of type 1 diabetes in adolescents.Clinical Therapeutics 24: 629–38.Holl RW, Swift PG, Mortensen HB, Lynggaard H, Hougaard P, Aanstoot HJ et al., on behalf of theHvidore Study Group (2003). Insulin injection regimens and metabolic control in an internationalsurvey of adolescents with type 1 diabetes over 3 years: results from the Hvidore Study Group.European Journal of Paediatrics 162: 22–9.ISPAD Consensus Guidelines for the Management of type 1 Diabetes Mellitus in Children andAdolescents. Swift P, ed. Medical Forum International, The Netherlands. 2000.Ivy JL, Holloszy JO (1981). Persistent increase in glucose uptake by rat skeletal muscle followingexercise. American Journal of Physiology 241: C200–3.Jones TW, Boulware SD, Kraemer DT, Caprio S, Sherwin RS, Tamborlane WV (1991). Independenteffects of youth and poor diabetes control on responses to hypoglycemia in children. Diabetes 40:358–63.Jones TW, Borg WP, Boulware SD, McCarthy G, Sherwin RS, Tamborlane WV (1995). Enhancedadrenomedullary response and increased susceptibility to neuroglycopenia: Mechanisms underlyingthe adverse effects of sugar ingestion in healthy children. Journal of Pediatrics 126: 171–7.Jones TW, Porter P, Sherwin RS, Davis EA, O’Leary P, Frazer F et al. (1998). Decreased epinephrineresponses to hypoglycemia during sleep. New England Journal of Medicine 338: 1657–62.Kaufman FR, Halvorson M, Kaufman ND (1995). A randomized, blinded trial of uncooked cornstarchto diminish nocturnal hypoglycemia at diabetes camp. Diabetes Research and Clinical Practice. 30:205–9.Kerr DS, Hansen IL, Levy MM (1983). Metabolic and hormonal responses of children and adolescentsto fasting and 2-deoxyglucose. Metabolism 22: 951–9.Koh THHG, Eyre JA, Aynsley-Green A (1988). Neonatal hypoglycaemia – the controversy regardingdefinition. Archives of Disease in Childhood 63: 1386–8.Lamers KJB, Doesburg WH, Gabreels FJM, Lemmens WAJG, Romson AC, Wevers RA, Renier WO(1985a). The concentration of blood components related to fuel metabolism during prolonged fastingin children. Clinica Chimica Acta 152: 155–63.Lamers KJB, Doesburg WH, Gabreels FJM, Romson AC, Renier WO, Wevers RA, Lemmens WAJG(1985b). Reference values of blood components related to fuel metabolism in children after anovernight fast. Clinica Chimica Acta 145: 17–26.

REFERENCES 213Lawrence RD (1926). The effects of exercise on insulin action in diabetes. British Medical Journal 1:648–52.Levine B, Anderson BJ, Butler DA, Antisdel JE, Brackett J, Laffel LMB (2001). Predictors of glycemiccontrol and short-term adverse outcomes in youths with type 1 diabetes. Journal of Pediatrics 139:197–203.Lopez MJ, Oyarzabal M, Barrio R, Hermoso F, Lopez JP, Rodriguez M et al. (1997). Nocturnalhypoglycaemia in IDDM patients younger than 18 years. Diabetic Medicine 14: 772–7.Lucas A, Morley R, Cole TJ (1988). Adverse neurodevelopmental outcome of moderate neonatalhypoglycaemia. British Medical Journal 297: 1304–8.McCrimmon RJ, Gold AE, Deary IJ, Kelnar CJ, Frier BM (1995). Symptoms of hypoglycemia inchildren with IDDM. Diabetes Care 18: 858–61.Magrath G, Hartland BV, the Nutrition Subcommittee of the British Diabetic Association’s ProfessionalAdvisory Committee (1993). Dietary recommendations for children with diabetes: an implementationpaper. Journal of Human Nutrition and Dietetics 6: 491–507.Maquet P, Dive D, Salmon E, Sadzot B, Franco G, Poirrier R et al. (1990). Cerebral glucose utilizationduring the sleep-wake cycle in man determined by positron emission tomography and 18F-2-fluoro-2-deoxy-D-glucose method. Brain Research 513: 136–43.Marrero DG, Guare JC, Vandagriff JL, Fineberg NS (1997). Fear of hypoglycemia in the parents ofchildren and adolescents with diabetes: maladaptive or healthy response? Diabetes Educator 23:281–6.Matyka KA, Wiggs L, Pramming S, Stores G, Dunger DB (1999a). Cognitive function and moodfollowing profound nocturnal hypoglycaemia during conventional therapy of diabetes. Archives ofDisease in Childhood 81: 138–42.Matyka KA, Crowne EC, Havel PJ, Macdonald IA, Matthews D, Dunger DB (1999b). Counterregulationduring spontaneous nocturnal hypoglycemia in prepubertal children with insulin dependentdiabetes mellitus. Diabetes Care 22: 1144–50.Matyka KA, Crawford C, Wiggs L, Dunger DB, Stores G (2000). Alterations in sleep physiology inyoung children with insulin dependent diabetes mellitus: relationship to nocturnal hypoglycemia.Journal of Pediatrics 137: 233–8.Mohn A, Matyka KA, Harris DA, Ross KM, Edge JA, Dunger DB (1999). Lispro or regular insulin formultiple injection therapy in adolescence. Differences in free insulin and glucose levels overnight.Diabetes Care 22: 27–32.Mortensen HB, Hougaard P, for the Hvidore Study Group on Childhood Diabetes (1997). Comparisonof metabolic control in a cross-sectional study of 2,873 children and adolescents with IDDM from18 countries. Diabetes Care 20: 714–20.Murphy NP, Keane SM, Ong KK, Ford-Adams M, Edge JA, Acerini C, Dunger DB (2003).Randomized cross-over trial of insulin glargine plus lispro or NPH insulin plus regular humaninsulin in adolescents with type 1 diabetes on intensive insulin regimens. Diabetes Care 26:799–804.Nordfelt S, Ludvigsson J (1997). Severe hypoglycemia in children with IDDM: a prospective populationstudy 1992–1994. Diabetes Care 20: 497–503.Nordfelt S, Samuelsson U (2003). Serum ACE predicts severe hypoglycemia in children and adolescentswith type 1 diabetes. Diabetes Care 26: 274–8.Nordfelt S, Johansson C, Carlsson E, Hammersjo J-A (2003). Prevention of severe hypoglycaemia intype 1 diabetes: a randomised controlled population study. Archives of Disease in Childhood 88:240–5.Nordfelt S, Ludvigsson J (2005). Fear and other disturbances of severe hypoglycaemia in children andadolescents with type 1 diabetes. Journal of Pediatric Endocrinology 18: 83–91.Northam EA, Anderson PJ, Jacobs R, Hughes M, Warne GL, Werther GA (2001). Neuropsychologicalprofiles of children with type 1 diabetes 6 years after disease onset. Diabetes Care24: 1541–6.

214 HYPOGLYCAEMIA IN CHILDREN WITH DIABETESParmeggiani PL, Morrison AR (1990). Alterations in autonomic functions during sleep. In: CentralRegulation of Autonomic Functions. Loewy AD and Spyer KM, eds. Oxford University Press,Oxford: 367–86.Pedersen-Bjergaard U, Agerholm-Larsen, Pramming S, Hougaard P, Thorsteinsson B (2001). Activityof angiotensin converting enzyme and risk of severe hypoglycaemia in type 1 diabetes mellitus.Lancet 357: 1248–53.Pedersen-Bjergaard U, Agerholm-Larsen, Pramming S, Hougaard P, Thorsteinsson B (2003). Predictionof severe hypoglycaemia by angiotensin converting enzyme activity and genotype in type 1 diabetes.Diabetologia 46: 89–96.Pillar G, Schuscheim G, Weiss R, Malhotra A, McCowen KC, Shlitner A et al. (2003). Interactionsbetween hypoglycemia and sleep architecture in children with type 1 diabetes mellitus. Journal ofPediatrics 142: 163–8.Pocecco M, Ronfani L (1998). Transient focal neurologic deficits associated with hypoglycaemiain children with insulin-dependent diabetes mellitus. Italian Collaborative Paediatric DiabetologicGroup. Acta Paediatrica 87: 542–4.Porter P, Byrne G, Stick S, Jones TW (1996). Nocturnal hypoglycaemia and sleep disturbances inyoung teenagers with insulin dependent diabetes mellitus. Archives of Disease in Childhood 75:120–3.Porter PA, Keating B, Byrne G, Jones TW (1997). Incidence and predictive criteria of nocturnalhypoglycemia in young children with insulin-dependent diabetes mellitus. Journal of Pediatrics130: 366–72.Price TB, Rothman DL, Taylor R, Avison MJ, Shulman GI, Shulman RG (1994). Human muscleglycogen resynthesis after exercise: insulin-dependent and independent phases. Journal of AppliedPhysiology 76: 104–11.Rewers A, Chase HP, Mackenzie T, Walravens P, Roback M, Rewers M et al. (2002). Predictors ofacute complications in children with type 1 diabetes. Journal of the American Medical Association287: 2511–8.Richter EA (1996). Glucose utilization. In: Exercise: Regulation and Integration of Multiple Systems.Rowell LB and JT Shepherd JT, eds. Oxford University Press, New York: 912–51.Rigat B, Hubert C, Alhenc-Gelas F, Cambien F, Corvol P, Soubrier F (1990). An insertion/deletionpolymorphism in the angiotensin-1 converting enzyme gene accounting for half the variance ofserum enzyme levels. Journal of Clinical Investigation 86: 1343–6.Rosilio MR, Cotton JB, Wieliczko MC, Genrault B, Carel JC, Couvaras O et al., on behalf of theFrench Pediatric Diabetes Group (1998). Factors associated with glycemic control. Diabetes Care21: 1146–53.Ross LA, Warren RE, Kelnar CJH, Frier BM (2005). Pubertal stage and hypoglycaemia counterregulationin type 1 diabetes. Archives of Disease in Childhood 90: 190–4.Rovet JF, Ehrlich RM (1999). The effect of hypoglycemic seizures on cognitive function in childrenwith diabetes: a seven year prospective study. Journal of Pediatrics 134: 503–6.Ryan C, Vega A, Drash A (1985). Cognitive deficits in adolescents who developed diabetes early inlife. Pediatrics 75: 921–7.Ryan CM, Atchison J, Puczynski S, Puczynski M, Arslanian S, Becker D (1990). Mild hypoglycemiaassociated with deterioration of mental efficiency in children with insulin-dependent diabetesmellitus. Journal of Pediatrics 117: 32–8.Saudubray JM, Marsac C, Limal JM, Dumurgier E, Charpentier C, Ogier H, Coude FX (1981).Variation in plasma ketone bodies during a 24-hour fast in normal and hypoglycemic children:Relationship to age. Journal of Pediatrics 98: 904–8.Schultz CJ, Konopelska-Bahu T, Dalton RN, Carroll TA, Stratton I, Gale EA et al. (1999). Microalbuminuriaprevalence varies with age, sex, and puberty in children with type 1 diabetes followedfrom diagnosis in a longitudinal study. Oxford Regional Prospective Study Group. Diabetes Care22: 495–502.

REFERENCES 215Singer-Granick C, Hoffman RP, Kerensky K, Drash AL, Becker DJ (1988). Glucagon responses tohypoglycemia in children and adolescents with IDDM. Diabetes Care 11: 643–9.Solders G, Thalme B, Aguirre-Aquino M, Brandt L, Berg U, Persson A (1997). Nerve conductionand autonomic nerve function in diabetic children. A 10 year follow-up study. Acta Paediatrica 86:361–6.Stirling HF, Darling JAB, Kelnar CJH (1991). Nocturnal glucose homeostasis in normal children.Hormone Research 35: 54 (abstract).The Diabetes Control and Complications Trial Research Group (1993). The effect of intensive treatmentof diabetes on the development and progression of long-term complications in insulin-dependentdiabetes mellitus. New England Journal of Medicine 329: 977–86.The Diabetes Control and Complications Trial Research Group (1994). Effect of intensive diabetestreatment on the development and progression of long-term complications in adolescents withinsulin-dependent diabetes mellitus: Diabetes Control and Complications Trial. Journal of Pediatrics125: 177–88.The Diabetes Control and Complications Trial Research Group (1997). Hypoglycemia in the DiabetesControl and Complications Trial. Diabetes 46: 271–86.Thomsett M, Shield G, Batch J, Cotterill A (1999). How well are we doing? Metabolic control inpatients with diabetes. Journal of Paediatrics and Child Health 35: 479–82.Trumper BG, Reschke K, Molling J (1995). Circadian variation of insulin requirement in IDDM: therelationship between circadian change in insulin demand and diurnal patterns of growth hormone,cortisol and glucagon during euglycaemia. Hormone and Metabolic Research 27: 141–7.Tsalikian E, Mauras N, Beck RW, Tamborlane WV, Janz KF, Chase HP et al., Diabetes Research InChildren Network Direcnet Study Group (2005). Impact of exercise on overnight glycemic controlin children with type 1 diabetes. Journal of Pediatrics 147: 528–34.Tupola S, Rajantie J (1998a). Documented symptomatic hypoglycaemia in children and adolescentsusing multiple daily injection therapy. Diabetic Medicine 15: 492–6.Tupola S, Rajantie J, Maenpaa J (1998b). Severe hypoglycaemia in children and adolescents duringmultiple-dose insulin therapy. Diabetic Medicine 15: 695–9.Vanelli M, Cerutti F, Chiarelli F, Lorini R, Meschi F, MCDC – Italy Group (2005). Nationwide crosssectionalsurvey of 3560 children and adolescents with diabetes in Italy. Journal of EndocrinologicalInvestigation 28: 692–9.Ververs MT, Rouwe C, Smit GP (1993). Complex carbohydrates in the prevention of nocturnalhypoglycaemia in diabetic children. European Journal of Clinical Nutrition 47: 268–73.Wagner VM, Grabert M, Holl RW (2005). Severe hypoglycaemia, metabolic control and diabetesmanagement in children with type 1 diabetes in the decade after the Diabetes Control and ComplicationsTrial – a large-scale multicentre study. European Journal of Paediatrics 164: 73–9.West H, Richey A, Eyre MB (1934). Study of intelligence levels of juvenile diabetics. PsychologicalBulletin 31: 598.Wysocki T, Harris MA, Mauras N, Fox L, Taylor A, Jackson SC, White NH (2003). Absence ofadverse effects of severe hypoglycemia on cognitive function in school-aged children with diabetesover 18 months. Diabetes Care 26: 1100–5.

10 Hypoglycaemia in PregnancyAnn E. Gold and Donald W.M. PearsonINTRODUCTIONDiabetes mellitus is one of the most common medical conditions affecting women duringtheir reproductive years. A successful outcome of diabetic pregnancy can usually be anticipatedwith current management strategies, although an adverse outcome is still more commonthan in the non-diabetic population (Casson et al., 1997; Penney et al., 2003a; Evers et al.,2004; Jensen et al., 2004; Confidential Enquiry into Maternal and Child Health, 2005).Meticulous control of blood glucose before conception and throughout gestation is the cornerstoneof management to reduce congenital anomalies, neonatal morbidity and mortality.However, striving for continuous normoglycaemia comes at a cost. Many women experiencean increased frequency of hypoglycaemia, accompanied by impaired awareness ofhypoglycaemia or modification of their hypoglycaemic symptoms. This chapter describeswhy hypoglycaemia is a recognised problem during pregnancy and how this influences themanagement of diabetic pregnancies.Population studies have shown that in many countries the average age of mothers withdiabetes (type 1 and type 2 diabetes) during pregnancy is around 30 years (Penney et al.,2003a). At the time of their first pregnancy, women with type 1 diabetes have on average haddiabetes for over ten years, whereas some will have been exposed to the long-term effectsof chronic hyperglycaemia for much longer when they conceive. Since the microvascularcomplications of diabetes are associated with the duration of the condition, many women haveestablished microangiopathy at the time of conception. Careful preparation for pregnancyand regular obstetrical and medical surveillance throughout pregnancy and delivery aremandatory (SIGN, 2001; American Diabetes Association, 2003), along with rapid access tospecialist paediatric facilities for the neonate who may be heavy for dates and prematuredelivery, often by caesarean section.METABOLIC CHANGES DURING PREGNANCYFundamental changes occur in maternal metabolism and physiology during pregnancy. Over280 days the mother’s weight increases on average by 12.5 kg. The main increase in weightoccurs in the second half of pregnancy and is caused by the growth of the conceptus, theenlargement of maternal organs, maternal storage of fat and protein and an increase inmaternal blood volume and interstitial fluid. An increase in the basal metabolic rate results inthe need for increased energy intake. In addition throughout pregnancy maternal metabolismHypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

218 HYPOGLYCAEMIA IN PREGNANCYadapts to ensure an adequate supply of nutrients to the growing fetus and developing placenta.A normal pregnancy is characterised by major alterations of glucose homeostasis. Fastingglucose declines and, although the plasma glucose is elevated after an oral glucose tolerancetest, the mean plasma glucose level is around 4 mmol/l during the third trimester of a normalnon-diabetic pregnancy on a normal diet (Paretti et al., 2001).Development of the placenta in the uterus during the first trimester of pregnancy occursin a low oxygen environment when maternal blood supply is restricted. During this timefetal metabolism is heavily anaerobic, which may serve to protect the developing embryofrom oxygen free radical-mediated teratogenesis. At the start of the second trimester, whenorganogenesis is complete, the maternal circulation develops to support fetal growth.In the second and third trimester the development of insulin resistance leads to increasedinsulin secretion to avoid abnormal increases in glucose, free fatty acids and amino acids.In normal pregnancy insulin sensitivity is decreased by between 30 and 60%. Changinghormonal levels make a major contribution to insulin resistance. Human placental lactogen(HPL) has actions similar to growth hormone. It increases lipolysis with a rise in freefatty acids which are a steady source of energy for the mother and fetus during periodsof starvation. Progesterone is also associated with insulin resistance. Maternal lipid storesincrease during pregnancy and adipokines and cytokines may play a role in the developmentof increasing insulin resistance. The cytokine tumour necrosis factor-alpha (TNF-) risesas the fat mass increases and can be related to insulin resistance. In pregnant women adecrease of adiponectin has been shown to relate to increasing insulin resistance in the thirdtrimester. In women with type 1 diabetes the physiological development of insulin resistanceduring pregnancy poses challenges to the expectant mother who is attempting to maintainnormoglycaemia.Many other changes in physiology occur in pregnancy. The complex process of placentaldevelopment is mostly complete by the end of the second trimester though the placentacontinues to expand with the growing fetus. In the third trimester maternal metabolismswitches from anabolism to catabolism, permitting an enhanced transfer of nutrients acrossthe placenta to sustain rapid fetal growth. The placenta is an active organ in this process. Inaddition to synthesising various hormones the placenta regulates the transfer of maternal fuelsto the fetus and facilitates maternal metabolic adaptation at different stages of pregnancy.Cells in contact with the maternal circulation and fetal circulation have a range of receptors,transporters and channels on both placental surfaces.At term the placenta of the mother with diabetes shows a number of differences fromthose in women who do not have diabetes. These include changes in morphology, bloodflow, transport and metabolism. This is important since transplacental transport of glucoseis a facilitated process and net transfer is strongly dependent on the concentration gradientof glucose between the maternal and fetal blood. However, the correlation between variousindices of glucose control – e.g. HbA 1c and fetal growth – is poor, suggesting that factorsother than maternal hyperglycaemia contribute to accelerated fetal growth (Penney et al.,2003b). Up-regulation of placental glucose transporters in type 1 diabetes may contributeto increased placental glucose transfer and stimulate fetal growth even if the mother hasexcellent glycaemic control. Transport of amino acids across the human placenta is an activeprocess resulting in amino acid concentrations in the fetal circulation that are substantiallyhigher than those in the maternal circulation.Management strategies in women with type 1 diabetes need to take the metabolic adaptationsof pregnancy into account. Although insulin resistance is the characteristic feature

FREQUENCY OF HYPOGLYCAEMIA IN DIABETIC PREGNANCY 219of the later stage of pregnancy, in the first trimester a modest increase in insulin sensitivityoccurs. The Diabetes In Early Pregnancy study (DIEP) reported declining insulinrequirements in the middle of the first trimester of pregnancy in women with type1 diabetes (Jovanovic et al., 2001). Over-insulinisation at this stage may be an issuesince women will be striving for optimal glycaemic control during the crucial periodof organogenesis. Hyperemesis gravidarum may also contribute to an increased risk ofhypoglycaemia.FREQUENCY OF HYPOGLYCAEMIA IN DIABETIC PREGNANCYHypoglycaemia is a frequent problem experienced by women with diabetes during pregnancy.A number of studies have reported the frequency of hypoglycaemia during pregnancy inwomen with diabetes but comparison is made difficult by the variations used in the definitionof hypoglycaemia and the methods used to collect the data. Table 10.1 summarises thefrequency of hypoglycaemia in some studies of women with pre-gestational diabetes. In all ofthe studies, with the exception of that by Persson and Hanson (1993), severe hypoglycaemiawas common during pregnancy in women with pre-gestational diabetes. In the study byPersson and Hanson (1993), a lower incidence of hypoglycaemia was reported using anintensive insulin regimen combined with very frequent self-monitoring, which may partlyaccount for the difference in frequency as compared with the other studies.Most studies have demonstrated that the peak incidence of hypoglycaemia occurs duringthe first and second trimesters. Kimmerle et al. (1992) observed that 84% of hypoglycaemicepisodes that resulted in impaired consciousness occurred before week 20. A peak incidenceof hypoglycaemia was observed during weeks 10–15 in a study of women receivingintensive insulin therapy (Rosenn et al., 1995). In the Diabetes Control and ComplicationsTrial (DCCT), a similar number of episodes of hypoglycaemia was recorded in the firstand second trimesters and fewer episodes were reported during the third trimester. In alarger study of 323 women, severe hypoglycaemia was almost 2.5 times more frequent inthe first trimester compared with the third trimester (Evers et al., 2002a; 2004). However,the reported incidence varies considerably between studies, which may represent differencesin patient groups and management strategies as well as varying definitions of severehypoglycaemia.Nocturnal hypoglycaemia is particularly common during pregnancy (Kimmerle et al.,1992). The advent of continuous blood glucose monitoring (CBGM) has confirmed this highincidence of nocturnal hypoglycaemia during pregnancy in mothers with type 1 diabetes(Yogev et al., 2003). In this study, 34 women were monitored for a 72-hour period betweenweeks 16 and 32 of pregnancy. During this short time period, nocturnal hypoglycaemic eventswere recorded in 26 (76%) women but only 17 of the patients experienced symptoms. In allof the affected patients an interval of 1–4 hours elapsed before any clinical manifestationsof hypoglycaemia were apparent.Comparison of two national audits of diabetic pregnancy in Scotland has shown thatsignificantly more women experienced severe hypoglycaemia in 1998–1999 than in 2003–2004 (41 versus 30%). This difference may be explained by a number of factors, such asmore women attending for pre-pregnancy care in 2003–2004 and the application of newerinsulin regimens that employed short-acting analogues (Scottish Diabetes in Pregnancy StudyGroup, 2004).

220 HYPOGLYCAEMIA IN PREGNANCYTable 10.1pregnancySummary of studies of the incidence of ‘severe’ hypoglycaemia in pre-gestational diabeticReferenceType ofdiabetesDefinition of severehypoglycaemiaWomen experiencingseverehypoglycaemiaduring pregnancynKimmerleet al., 1992Persson et al.,1993Rosenn et al.,1995Type 1Type 1Type 1Impaired consciouslevel responding toglucose/glucagonRequiring externalhelp for recoverySubdivided into:a) requiringexternal helpb) coma/seizureDCCT, 1996 Type 1 Seizure or loss ofconsciousnessMasson et al.,2003Garg et al.,2003Evers et al.2002a; 2004ScottishDiabetes inPregnancyStudy Group,2004Type 1 (alltaking lisproinsulin)Type 1 (alltaking lisproinsulin)Type 1Type 1Requiring externalhelp for recoveryRequiring externalhelp for recoveryRequiring externalhelp for recoveryRequiring externalhelp for recovery41% (77% episodesoccurred duringsleep)4.4% 11371% 34% 8417% in intensivegroup19.8% inconventionalgroup27% 7623% overall 6241% 1st trimester17% 3rd trimester30% overall20% 1st trimester17% 2nd trimester10% 3rd trimester77(85pregnancies)180(270pregnancies)278 (2002)323 (2004)155Hypoglycaemia also occurs in women with type 2 diabetes, who often require insulin inthe pre-pregnancy period in order to optimise glycaemic control, and also in women whodevelop gestational diabetes (typically in the late second and third trimesters). However,the frequency of hypoglycaemia has not been documented with accuracy in these groups; itis the authors’ impression that hypoglycaemia occurs much less frequently than in type 1diabetes.Why are Women with Pre-gestational Diabetes at Greater Risk ofHypoglycaemia during Pregnancy?Great importance is placed upon maintaining good glycaemic control throughout diabeticpregnancy when women are highly motivated and often achieve levels of glycated

FREQUENCY OF HYPOGLYCAEMIA IN DIABETIC PREGNANCY 221haemoglobin that are close to the non-diabetic range. In addition, because of the changesin insulin sensitivity in the first trimester, their risk of severe hypoglycaemia is increased.It has been shown that women with a previous history of hypoglycaemia are at increasedrisk of experiencing severe hypoglycaemia during pregnancy (Kimmerle et al., 1992; Everset al., 2002a). Women who demonstrate a greater fluctuation in blood glucose during pregnancyare also at greater risk of experiencing severe hypoglycaemia (Rosenn et al., 1995)and women who are not using an ‘intensive’ insulin regimen have an enhanced risk. Everset al. (2002a) observed that the risks of experiencing severe hypoglycaemia during the firsttrimester were significantly greater in women who had duration of diabetes greater than tenyears, HbA 1c less than 6.5% or an insulin dose of greater than 0.1 units of insulin per kgbody weight.It is possible that impaired counterregulation could contribute to the increased risk ofhypoglycaemia. A rat model demonstrated that the glucagon and epinephrine (but notnorepinephrine) responses to hypoglycaemia (plasma glucose 3.4 mmol/l) were suppressedduring pregnancy (Rossi et al., 1993); these data suggest that the counterregulatoryresponses to hypoglycaemia in rats may be impaired. A few studies have examinedhormonal counterregulatory responses in pregnant women with type 1 diabetes duringexperimental hypoglycaemia during pregnancy. All studies utilised a hyperinsulinaemichypoglycaemic clamp technique, except that by Nisell et al. (1994), which used aninsulin bolus injection to induce hypoglycaemia. The results of these studies are shown inTable 10.2.The studies vary in terms of the time of gestation at the time of study and theglucose nadir reached and it is difficult to determine whether counterregulation is impairedduring pregnancy as a result of pregnancy per se or because of differences in glycaemiccontrol and the presence of diabetes. The study by Rosenn et al. (1996) probably hadthe most appropriate study design to answer this question. One criticism, however, isthat the glucose nadir achieved was only 3.3 mmol/l, which may not have been sufficientto stimulate a counterregulatory response in the non-diabetic control group. The studiesprovide conflicting evidence as to whether the counterregulatory response to hypoglycaemiais deficient during pregnancy. No studies have examined counterregulatory effectsduring the first trimester at the time when the frequency of, and exposure to, hypoglycaemiaare at a peak. However, to conduct such studies would raise major ethicalconcerns.Few studies have examined the development of impaired hypoglycaemia awarenessduring pregnancy although most clinicians would agree that this is particularly problematicalduring the first trimester. Evers et al. (2002a) observed that severe hypoglycaemiain the first trimester was more likely to occur in women with reduced symptomaticawareness of hypoglycaemia. In laboratory-induced hypoglycaemia Björklundet al. (1998a) did measure symptomatic responses to hypoglycaemia during the thirdtrimester, and also postnatally, and found that symptoms such as ‘inability to concentrate’,‘headache’ and ‘pounding heart’ were less prominent during pregnancy comparedwith during the postnatal period. However, it is difficult to ascertain whether thisis a consequence of differences in glycaemic control or the incidence of hypoglycaemiaduring these two time periods. Fear of hypoglycaemia is also greater in womenwho have experienced severe hypoglycaemia (Evers et al., 2002a) and this is animportant problem for the diabetic mother which should be addressed during antenatalcare.

222 HYPOGLYCAEMIA IN PREGNANCYTable 10.2 Summary of studies examining counterregulatory responses to hypoglycaemia in womenwith pre-gestational diabetesReferenceGestationat time ofstudyControl groupBloodglucose nadirduring studyResultsnDiamondet al.,1992Nisellet al.,1994Rosennet al.,1996Björklundet al.,1998aBjörklundet al.,1998b21–37 weeks Non-pregnant,non-diabeticage matchedwomenLasttrimester,and 8–12weekspost-partumNon-pregnant(1);24–28 (2);32–34 (3)30–34 weeks(1);5–13 monthspostnatal(2)Acted as owncontrolsNon-diabetic,age-matched.Cases werealso studiedon 3occasionsi.e. acted asown controlsActed as owncontrols2.5 mmol/l No glucagon response incases.Epinephrine releasesuppressed in cases; cf.controls.Lower blood glucose forepinephrine and growthhormone release incases; cf. controls3.2 mmol/l No glucagon or cortisolresponse at either timepoint.Similar increases inepinephrine andnorepinephrine onboth occasions3.3 mmol/l Reduced epinephrineresponse in cases; cf.controls at all studytimes.Reduced epinephrineresponse in casesduring pregnancy; cf.pre-pregnancy.Growth hormone responsesreduced in pregnancy incases and controls; cf.pre-pregnancy2.3 mmol/l Epinephrine responsesimilar in pregnancy andpostnatal.Dehydroepiandrosteroneincreased more rapidlyand less sustainedduring pregnancy.Growth hormone responsesreduced in pregnancy30–34 weeks No controls 2.3 mmol/l Increase in placentalgrowth hormone, but nochanges in otherplacental hormonesduring hypoglycaemia9 type 1cases;7 controls8 pregestationaltype 1patients;1 gestational17 type 1cases;10 controls1010

CLINICAL MANAGEMENT BEFORE AND DURING PREGNANCY 223CLINICAL MANAGEMENT BEFORE AND DURING PREGNANCYPre-conception CareThe advantages of planned pregnancy should be regularly emphasised to women withdiabetes during their reproductive years and effective contraceptive advice should be part ofroutine clinical care. After menarche, all girls with type 1 diabetes should be aware of theimportance of pregnancy planning, because the infants of mothers who have attended forpre-pregnancy care have fewer major congenital malformations and require shorter periodsin special care facilities than infants of mothers who do not attend for pre-conception care(Fuhrmann et al., 1983; Steel et al., 1989; 1990; Kitzmiller et al., 1991; Ray et al., 2001).Attendance for structured pre-pregnancy care is also associated with a reduction in the rate ofspontaneous abortion. Pre-pregnancy counselling should address ways of minimising the riskof severe hypoglycaemia both before and during pregnancy. Ideally this should be discussedduring the pre-pregnancy period. However, if the pregnancy has not been planned, womenshould be made aware of their increased risk of hypoglycaemia during pregnancy and howto avoid and manage potential episodes, particularly as the greatest risk is during the firsttrimester.Organisation of Clinical CareIn many centres clinical care is delivered by a multidisciplinary combined obstetric/diabeticteam with very regular out-patient reviews to assess metabolic control and obstetric progress(Figure 10.1). Home blood glucose monitoring results are assessed and insulin regimenand dietary intake modified to optimise glycaemic control and HbA 1c (Figure 10.2). Mostwomen present for booking at around eight weeks gestation when an early scan will providean accurate estimate of gestational age. This is important to allow the optimal time ofdelivery to be determined. Screening is performed routinely for Down’s syndrome andneural-tube defects. Although the prevalence of congenital anomaly has declined followingthe introduction of pre-conceptional counselling, the incidence of congenital malformationis still higher than in the non-diabetic population. A detailed ultrasound scan at around 20weeks is performed to detect severe congenital anomalies, particularly to identify majormalformation of the heart and the central nervous system. Frequent scanning is performedlater in pregnancy to monitor fetal growth. In the third trimester regular cardiotocography,Doppler ultrasound and fetal movement charts are used to monitor fetal progress. Figure 10.3demonstrates the measurement of abdominal circumference (AC). Sequential measurementsof AC are used to monitor fetal growth.Good glycaemic control reduces stillbirth rate, neonatal hypoglycaemia and respiratorydistress syndrome. Women should strive to maintain blood glucose levels as near to the nondiabeticrange as possible without an excessive risk of hypoglycaemia. This usually meansblood glucose target levels between 4 and 7 mmol/l. The diabetes team, but in particularthe diabetes specialist nurses and specialist midwives, have an important role in educatingwomen on the need for home blood glucose monitoring (usually four to six times a day) andin introducing intensive insulin regimens if the women are not already on these programmes.Maternal issues in pregnancy are shown in Table 10.3.

224 HYPOGLYCAEMIA IN PREGNANCYType 1 and type 2 diabetes in pregnancy;Gestational diabetes on insulinBlood tests Other Tests Scanning Miscellaneous CounsellingPre-pregnancy HbA 1c Hb U&E TFT Rubella Eyes MSU/MA Folic acidDieticianMidwife

Figure 10.2 Example of home blood glucose monitoring chart used in the combined antenatal diabetic clinic at Aberdeen Maternity Hospital, Scotland

226 HYPOGLYCAEMIA IN PREGNANCYFigure 10.3 Ultrasound scan of at 32 weeks gestation demonstrating measurement of abdominalcircumference (AC)Optimising Insulin RegimensData is limited with regard to studies of different insulin regimens during pregnancy andmany studies are small and observational in nature. The effects of conventional versusintensive insulin therapy were studied in the women participating in the DCCT (The DiabetesControl and Complications Trial Research Group, 1996). In women who had initially been

CLINICAL MANAGEMENT BEFORE AND DURING PREGNANCY 227Table 10.3Maternal issues in pregnancyDiabetes-relatedHypoglycaemiaIncreased risk of ketoacidosisRetinopathy may deteriorateNephropathy and fluid retentionPre-eclampsia/hypertensionIncreased caesarean section rateObstetricPreterm deliveryIncreased caesarean section rateMacrosomia and shoulder dystociaassigned to conventional treatment, the occurrence of hypoglycaemia resulting in impairedconscious level was 4.0 per 100 pregnant-months compared with 2.8 in those receivingintensive treatment. Those women in the conventional group who had changed to intensivetherapy before conception experienced significantly fewer episodes of hypoglycaemia duringpregnancy than those who had not intensified their regimen until after conception (2.0 versus5.0 episodes per 100 patient years). This suggests that it may be preferable, where possible,to intensify the insulin regimen during the pre-pregnancy period rather than making thechange once conception has occurred.Another study of 118 women with pre-gestational diabetes compared twice daily freemixing of soluble and isophane insulins with a basal-bolus regimen of soluble insulinbefore meals and bedtime isophane insulin (Nachum et al., 1999). The authors reported nodifference between the regimens in the incidence of hypoglycaemia, although those takingfour injections of insulin per day achieved better glycaemic control. One interpretation ofthe data is that the greater the insulin dosage, the better the glycaemic control. However, theoverall incidence of recorded hypoglycaemia was remarkably low (10%) and almost half thewomen had type 2 diabetes.Short-acting insulin analogues have received some attention and one randomised trial ofinsulin lispro in pregnancies complicated by type 1 diabetes has been reported (Perssonet al., 2002). Thirty-three women were studied with 16 being randomised to lispro and 17 tosoluble insulin at gestational week 15. Most of the women had very good glycaemic controlat entry to the study (mean HbA 1c 6.5%, range 4.5–8.6%). Only two patients experiencedsevere hypoglycaemia during the remainder of their pregnancies and both were on humansoluble insulin, although it is important to remember that the study was undertaken inthe latter half of pregnancy when the risk of hypoglycaemia is usually lower. However, asignificantly higher rate of biochemical hypoglycaemia (< 30 mmol/l) was documented inthe women using lispro (who were monitoring blood glucose six times daily). No differenceswere observed between the groups with regard to glycated haemoglobin, although womentaking lispro had lower postprandial glucose excursions. In an observational study in whichwomen received lispro before and during pregnancy, the prevalence of severe hypoglycaemiawas 23%, with no greater incidence of adverse outcomes compared to other regimens(Garg et al., 2003). A similar prospective study that was designed to assess progressionof retinopathy during pregnancy in women on lispro, demonstrated that the number ofsubjective hypoglycaemic events did not differ between those taking lispro and those takingsoluble insulin, although the overall rates of hypoglycaemia were relatively low (Loukovaaraet al., 2003).

228 HYPOGLYCAEMIA IN PREGNANCYIn gestational diabetes, when lispro was compared with human soluble (regular) insulin,a significant difference was demonstrated in the incidence of hypoglycaemia (Jovanovicet al., 1999). However, the recorded incidence of hypoglycaemia was again very low and itis not appropriate to extrapolate this data to pregnancy in type 1 diabetes. In a recent largestudy of diabetic pregnancy from the Netherlands (Evers et al., 2004), the rate of severehypoglycaemia in a subgroup of 35 patients who were using lispro, did not differ fromthat observed in patients using other insulin regimens. The congenital malformation ratesin those women using lispro have not differed from those on soluble insulin (Wyatt et al.,2004). A recently reported multi-centre study by Mathies on et al. (2007) has demonstratedthe safe use of insulin aspart in Type 1 diabetic pregnancy, with a non-significantly lowerincidence of severe hypoglycaemia compared to human soluble insulin at comparable levelsof glycaemic control (Hb A 1c mostly

CLINICAL MANAGEMENT BEFORE AND DURING PREGNANCY 229Table 10.4 Insulin treatment in Gestational Diabetes Mellitus (GDM) (SIGN 2001)• Intensive insulin regimens are occasionally recommended to achieve normal bloodglucose levels and fetal growth• Women requiring insulin require a comprehensive education package regarding allaspects of insulin treatmentConsider insulin treatment for GDM if:• The fasting blood glucose is over 6 mmol/l or the 2 hr postprandial glucose is over7 mmol/l, despite appropriate lifestyle adjustments• The ultrasound scans show excessive growth with an abdominal circumference greaterthan the 95 percentile for gestation (some authorities recommend > 75 percentile)pregnancy and labour. In some European countries, a significant percentage of womenuse this type of treatment during pregnancy but in many countries only a relatively smallpercentage use CSII.Although an appropriate insulin regimen is very important many other factors can influenceinsulin absorption and effectiveness. Pregnancy is a good time to review all factors that couldcontribute to suboptimal glycaemic control. All practical aspects of insulin administrationshould be reviewed and injection sites examined.Some women with gestational diabetes may require insulin during the last few weeks ofpregnancy. The management of insulin treatment in gestational diabetes is summarised inTable 10.4.Dietary and Lifestyle ManagementDietary advice is also essential before, during and after pregnancy. Such advice willencourage the intake of foods with a high level of complex carbohydrates, soluble fibreand vitamins. Balancing of carbohydrate intake and insulin dose is important and manywomen may find it helpful to count carbohydrate portions and learn appropriate correctionfactors for high blood glucose levels. Dietetic review is therefore appropriate and a reviewof ‘sick-day rules’ should be undertaken. In the early weeks of pregnancy, where nauseaand vomiting are common, this is particularly important. During the later weeks of pregnancy,women may prefer to take smaller but more frequent meals and their insulin regimenshould be adjusted accordingly. In general, women are also advised to reduce their intake ofsaturated fat.Neural tube defects in high risk pregnancies are associated with low levels of folate. A largestudy of non-diabetic women has shown that the prescription of 4 mg of folate supplement,pre- and peri-conception, provides protection against neural tube defects in women at highrisk. In Scotland, all women are advised to take 5 mg of folic acid pre-conception, andcontinue this preparation during pregnancy until around 12 weeks of gestation. Smokingshould be discouraged. If women are used to taking regular physical activity this should beencouraged, along with advice about reducing insulin dosages before exercise and ensuringappropriate carbohydrate intake.

230 HYPOGLYCAEMIA IN PREGNANCYCorrection of HypoglycaemiaAdvice concerning the correction of hypoglycaemia by appropriate dietary intake shouldbe given and is similar to that for the non-pregnant patient with diabetes (see Chapter 5).Patients, their partners and other family members should be advised on the use of glucosegels, such as GlucoGel, for emergency use, and should keep a supply of this available.Instruction about the use of glucagon should also be given to partners or close familymembers. This should be done at the first review, since hypoglycaemia is more common inthe first trimester.Management of DeliveryThe infant of a diabetic mother is at increased risk during labour and delivery. Mothers withdiabetes must be recognised to be a high risk obstetric patient and delivered in a unit thatcan provide experienced obstetric care and immediate access to neonatal intensive support.Timing of DeliveryThere is no evidence to support the continuation of a type 1 diabetic pregnancy beyond40 weeks and, provided glycaemic control has been satisfactory and fetal growth has notbeen excessive, induction of labour and delivery at around 39 weeks appears to provide acompromise between delivering the baby too early and avoiding late unexplained intrauterinedeath.Management of Diabetes during LabourManagement of diabetes in labour is summarised in Table 10.5. The aim is to achievematernal normoglycaemia during labour using an intravenous dextrose/insulin infusion.The insulin infusion rate must be halved immediately after delivery to prevent postnatalhypoglycaemia, since there is a rapid increase in insulin sensitivity after separation of theplacenta. After delivery, the insulin dose should be reduced to the dosages used beforepregnancy. This will usually necessitate halving the doses required during pregnancy. If themother decides on breast feeding, a further reduction of the insulin dosage may be requiredto avoid hypoglycaemia.MATERNAL COMPLICATIONS OF DIABETES DURINGPREGNANCYThe complications experienced during pregnancy include those resulting from hypoglycaemiaand from the microvascular complications of diabetes.

MATERNAL COMPLICATIONS OF DIABETES DURING PREGNANCY 231Table 10.5 Management of diabetes during labour and delivery (clinicalcircumstances may lead to modifications in this regimen)• Fast the patient on the day of delivery for an elective section; otherwisefast from the onset of established labour.• Set up an intravenous insulin infusion using a syringe pump, e.g. 50 unitsshort-acting analogue in 50 ml saline. Start at 1–2 units per hour.• Simultaneously commence a 10% dextrose +01% KCI intravenous infusionat a rate of 100 ml/hr.• Monitor bedside glucose hourly and set a target blood glucose of4–6 mmol/l.• Use the algorithm (figure) to adjust the insulin infusion.• Halve the rate of the insulin infusion on separation and delivery of theplacenta. Continue the dextrose infusion as before.• Revert to the pre-pregnancy dosages of subcutaneous insulin when anormal eating pattern is established.Insulin infusion scaleStart insulin at 2 units per hour.Monitor capillary bedside glucose hourly.Capillary blood glucose (mmol/l)< 4 4–6 > 6Reduce insulin by Continue current rate Increase by 0.5 unit/hr0.5 unit/hrThe Risks of Maternal Hypoglycaemia to the MotherThe morbidity and mortality of hypoglycaemia in diabetes is discussed in Chapter 12.During pregnancy the risks are similar but compounded by the potential risk to the fetuscaused by injury. A maternal death was reported in one series (Rayburn et al., 1986).One particular concern in the pre-pregnancy and pregnancy clinics is the risk of hypoglycaemiawhile driving. One woman was involved in a motor vehicle accident after experiencinga hypoglycaemia-induced convulsion and fracturing her tibia and fibula. Whenplanning pregnancy, advice on the avoidance of hypoglycaemia while driving should bereinforced for patients treated with insulin (see Chapter 14), particularly with respectto testing blood glucose before every journey and ensuring appropriate management ofhypoglycaemia if it occurs while driving. They should also be warned that they willbe advised to cease driving if their warning symptoms of hypoglycaemia are significantlyreduced. Women who develop gestational diabetes and require insulin treatmentshould be given similar advice although the risk of developing severe hypoglycaemiais lower.

232 HYPOGLYCAEMIA IN PREGNANCYMicrovascular Complications of PregnancyRetinopathy is common and may progress during pregnancy. Therefore retinal screeningshould be undertaken during each trimester (Figure 10.1). In the longer term, parous womenwith type 1 diabetes have significantly lower levels of all types of retinopathy comparedwith non-parous women. The associated significant differences in HbA 1c suggest that theimproved glycaemic control that is associated with pregnancy may be sustained over timewith beneficial effects on long-term complications. Thus women should be reassured thatstrict glycaemic control during, and immediately after, pregnancy can effectively reducethe long-term risk of retinopathy. In pregnancy, diabetic nephropathy of any degree is lesscommon than retinopathy, and requires specialist management.COMPLICATIONS IN THE INFANT OF THE DIABETIC MOTHERThe potential complications for the infant of a mother with diabetes are shown in Table 10.6.Newly born infants with diabetic mothers are at increased risk of developing hypoglycaemia.To avoid this risk, maternal blood glucose should be kept as close to normal as possibleduring pregnancy and particularly during labour and delivery. Early feeding of the newlyborn infant is essential and careful monitoring is mandatory.The Risks of Maternal Hypoglycaemia to the Fetus/infantThe potential risks of hypoglycaemia to the offspring can be considered in three main ways:the teratogenic effects of hypoglycaemia, other immediate effects on the fetus and delayedeffects.Table 10.6 Potential problems affecting the infant ofa mother with diabetesNeonatal and perinatal• Hypoglycaemia• Stillbirth• Increased perinatal mortality• Congenital anomalies• Preterm delivery• Macrosomia• Birth trauma at delivery• Polycythaemia• Respiratory distress syndrome

COMPLICATIONS IN THE INFANT OF THE DIABETIC MOTHER 233Data in animals suggest that hypoglycaemia in the first trimester could be teratogenic(Buchanan et al., 1986; Akazawa et al., 1987, 1989; Ellington, 1987; Smoak and Sadler,1990). However, it is very difficult to extrapolate these results to human pregnancy as theexposure to hypoglycaemia in these animal studies would equate to sustained hypoglycaemiain humans of at least five hours. There also appears to be a differential effect betweenspecies making it inappropriate to extrapolate results from one species to another (Erikssonet al., 1986).In non-diabetic women in whom hypoglycaemic coma was induced as a treatment forpsychiatric illness, the fetus was affected more frequently when hypoglycaemia was inducedduring the first trimester (Impastato et al., 1964). However many potential confoundingvariables were present, and the blood glucose profiles in women with type 1 diabetes arecompletely different from the non-diabetic woman.When viewing the outcomes of diabetic pregnancy in humans, the overwhelming evidencesupports the achievement of strict glycaemic control in early pregnancy, which is usuallyassociated with a higher risk of hypoglycaemia but greatly reduces the risks of congenitalmalformation (Kitzmiller et al., 1991; Kimmerle et al., 1992; Evers et al., 2004; Rosennet al., 1995; The Diabetes Control and Complications Trial Research Group, 1996). Thissuggests that the risk of any teratogenic effect of hypoglycaemia is greatly outweighed bythe benefits of good glycaemic control. A large follow-up study of 329 infants born tomothers with diabetes demonstrated a high perinatal fetal mortality rate of 21%, but noconclusive evidence could implicate hypoglycaemia or any other iatrogenic factor (Farquhar,1969). In another study of women who had relatively good glycaemic control (Mills et al.,1988), all episodes of documented hypoglycaemia between gestational weeks 5 and 12 wererecorded and no association with fetal malformations was observed. Similarly, in a study byKitzmiller et al. (1991), all episodes of symptomatic hypoglycaemia were documented andno association with fetal malformation was found.In addition to the reduced risk of congenital malformations, other complications of pregnancyappear to be fewer in women who experience hypoglycaemia during pregnancy. Theincidence of macrosomia has been shown to be significantly lower in women affected bysevere hypoglycaemia during the third trimester (relative risk 0.66) (Evers et al., 2002b,Evers et al., 2004).With respect to the immediate effects of maternal hypoglycaemia on the fetus in utero,some data are available from studies that have induced experimental hypoglycaemia duringpregnancy. Early studies suggested that fetal heart rate variability may be decreased duringmaternal hypoglycaemia (Stangenberg et al., 1983). However, later studies did not demonstratechanges in fetal movements, breathing or heart rate during maternal hypoglycaemia(Nisell et al., 1994, Reece et al., 1995). Björklund et al. (1996) also studied the effectsof maternal blood glucose of 2.2 mmol/l on fetal heart rate. They observed an increase infrequency and amplitude of fetal heart rate accelerations, but no potentially harmful effectson fetal heart rate or on umbilical Doppler waveform analysis were seen. In a more recentstudy of 116 women who experienced ‘hypoglycaemia’ rather than hyperglycaemia whileundergoing a glucose tolerance test to screen for gestational diabetes, no adverse effects onfetal growth or other perinatal outcomes were observed, although the level of blood glucosethat was designated as ‘hypoglycaemia’ (4.9 mmol/l) may well be considered too high torepresent this state (Calfee et al., 1999).The later developmental effects in the infant that has been subjected to maternal hypoglycaemiamay not become apparent until later in childhood. However, there are many

234 HYPOGLYCAEMIA IN PREGNANCYpotential confounding factors when studying this aspect. For example, a woman who hasbeen exposed to hypoglycaemia may also have experienced hyperglycaemia, which couldalso have a potentially deleterious effect on cognition and it is not possible to separatethese effects. Early studies did demonstrate that the children of diabetic mothers developedcerebral dysfunction (Yssing, 1974) and one study reported an association between the IQof offspring and the presence of acetonuria, but not hypoglycaemia, during pregnancy in themother (Churchill et al., 1969).A Japanese study of 33 children born to mothers with diabetes demonstrated significantlylower intelligence scores than a control group when tested at three years of age (Yamashitaet al., 1996). However, no correlation was observed between IQ and maternal HbA 1c scoresduring pregnancy. An electrophysiological study of 60 infants (34 controls and 26 childrenfrom mothers with diabetes) performed at six months of age, also demonstrated deficits inrecognition memory (suggesting hippocampal damage) in the offspring from mothers withdiabetes that was unrelated to maternal glycaemic control, although the incidence of maternalhypoglycaemia was not measured (Nelson et al., 2000).In summary, there are no data that suggest that hypoglycaemia is teratogenic in humansor hypoglycaemia has an immediate adverse effect on the fetus. Some studies suggest thatthe offspring of diabetic mothers may have some differences in cerebral function, althoughthese differences have not been attributed to maternal hypoglycaemia.CONCLUSIONS• The outlook for the offspring of women with type 1 diabetes has improved dramaticallyover the last 30 years, and the benefit to the infant from optimal glycaemic control duringpregnancy is evident.• The risk of maternal hypoglycaemia is considerable, particularly in the first and secondtrimesters, and severe hypoglycaemia is common.• Factors that may contribute to increased risk of hypoglycaemia include the requirementfor strict glycaemic control, possible effects of pregnancy to cause counterregulatory deficiencyand altered symptomatic awareness, differing insulin regimens and lifestyle/dietaryfactors.• Detailed education is essential to minimise the risk of hypoglycaemia.• There is no evidence to suggest that hypoglycaemia has an adverse effect on the humanfetus or the infant of a diabetic mother, although significant maternal morbidity may occur.• Good organisation of pre-pregnancy and pregnancy services is required to reduce the risksof maternal and infant morbidity during diabetic pregnancy.REFERENCESAkazawa M, Akazawa S, Hashimoto M, Yamaguchi Y, Kuriya N, Toyama K et al. (1987).Effects of hypoglycaemia on early embryogenesis in rat embryo organ culture. Diabetologia30: 791–6.

REFERENCES 235Akazawa M, Akazawa S, Hashimoto M, Akashi M, Yamazaki H, Tahara D et al. (1989). Effects ofbrief exposure to insulin-induced hypoglycemic serum during organogenesis in rat embryo culture.Diabetes 38: 1573–8.American Diabetes Association (2003). Preconception care of women with diabetes. Diabetes Care26 (Suppl 1): S91–3.Björklund AO, Adamson UKC, Almström NHH, Enocksson EA, Gennser GM, Lins PE, Westgren LM(1996). Effects of hypoglycaemia on fetal heart activity and umbilical artery Doppler velocity waveformsin pregnant women with insulin-dependent diabetes mellitus. British Journal of Obstetricsand Gynaecology 103: 413–20.Björklund A, Adamson U, Andreasson K, Carlstrom K, Hennen G, Ogout A, Lins PE, Westgren M(1998a). Hormonal counterregulation and subjective symptoms during induced hypoglycemia ininsulin-dependent diabetes mellitus patients during and after pregnancy. Acta Obstetrica GynecologicaScandinavica 77: 625–34.Björklund AO, Adamson UKC, Carlström KAM, Hennen G, Igout A, Lins PE, Westgren LM (1998b).Placental hormones during induced hypoglycaemia in pregnant women with insulin-dependentdiabetes mellitus: evidence of an active role for placenta in hormonal counter-regulation. BritishJournal of Obstetrics and Gynaecology 105: 649–55.Buchanan TA, Schemmer JK, Freinkel N (1986). Embryotoxic effects of brief maternal insulinhypoglycemiaduring organogenesis in the rat. Journal of Clinical Investigation 78: 643–9.Calfee EF, Rust OA, Bofill JA, Ross EL, Morrison JC (1999). Maternal hypoglycemia: is it associatedwith adverse perinatal outcome? Journal of Perinatology 19: 379–82.Casson IF, Clarke CA, Howard CV, McKendrick O, Pennycook S, Pharoah PO et al. (1997). Outcomesof pregnancy in insulin dependent diabetic women: results of a five year population cohort study.British Medical Journal 315: 275–8.Churchill JA, Berendes HW, Nemore J (1969). Neuropsychological deficits in children of diabeticmothers. American Journal of Obstetrics and Gynecology 105: 257–68.Confidential Enquiry into Maternal and Child Health (2005). Pregnancy in women with type 1 andtype 2 diabetes in 2002–2003. England, Wales and Northern Ireland. CEMACH, London.Diamond MP, Reece EA, Caprio S, Jones TW, Amiel SA, DeGennaro N, Laudano A et al. (1992).Impairment of counterregulatory hormone responses to hypoglycemia in pregnant women withinsulin-dependent diabetes mellitus. American Journal of Obstetrics and Gynecology 166: 70–7.Ellington SK (1987). Development of rat embryos cultured in glucose-deficient media. Diabetes 36:1372–8.Eriksson UT, Dahlström VE, Lithell HO (1986). Diabetes in pregnancy: influence of background andmaternal diabetic state on the incidence of skeletal malformations in the fetal rat. Acta Endocrinologica(Copenhagen) 112 (Suppl 277): 66–73.Evers IM, ter Braak EWMT, de Valk HW, der Schout BV, Janssen N, Visser GA (2002a). Riskindicators predictive of severe hypoglycemia during the first trimester of type 1 diabetic pregnancy.Diabetes Care 25: 554–9.Evers IM, de Walk HW, Mol BWJ, ter Braak EWMT, Visser GHA (2002b). Macrosomia despite goodglycaemic control in type 1 diabetic pregnancy; results of a nationwide study in The Netherlands.Diabetologia 45: 1484–9.Evers IM, de Valk HW, Visser GHA (2004). Risk of complications of pregnancy in women with type1 diabetes: nationwide prospective study in the Netherlands. British Medical Journal 328: 915–8.Farquhar JW (1969). Prognosis for babies born to diabetic mothers in Edinburgh. Archives of Diseasein Childhood 44: 36–47.Fuhrmann K, Reiker H, Semmler K, Glockner E (1983). Prevention of congenital malformation ininfants of insulin dependent diabetic mothers. Diabetes Care 6: 219–23.Garg SK, Frias JP, Anil S, Gottlieb PA, MacKenzie T, Jackson WE. (2003). Insulin lispro therapyin pregnancies complicated by type 1 diabetes: glycemic control and maternal and fetal outcomes.Endocrine Practice 9: 187-93.

236 HYPOGLYCAEMIA IN PREGNANCYImpastato DJ, Gabriel AR, Lardaro EH (1964). Electric and insulin shock therapy during pregnancy.Diseases of the Nervous System 25: 542–6.Jensen DM, Damm P, Moelsted-Pedersen L, Ovesen P, Westergaard JG, Moeller M, Beck-Nielsen H(2004). Outcomes in type 1 diabetic pregnancies: a nationwide, population-based study. DiabetesCare, 27: 2819–23.Jovanovic L, Ilic S, Pettitt DJ, Hugo K, Gutierrez M, Bowsher RR, Bastyr EJ (1999).Metabolic and immunologic effects of insulin lispro in gestational diabetes. Diabetes Care 22:1422–7.Jovanovic L, Knopp RH, Brown, Conleg MR, Park E, Mills JL et al. (2001). Declining insulinrequirement in the late first trimester of diabetic pregnancy. Diabetes Care 24: 1130–6.Kimmerle R, Heinemann L, Delecki A, Berger M (1992). Severe hypoglycemia incidence and predisposingfactors in 85 pregnancies of type 1 diabetic women. Diabetes Care 15: 1034–7.Kitzmiller JL, Gavin LA, Gin GD, Jovanovic-Petersen L, Main EK, Zingran WD (1991). Preconceptioncare of diabetes. Glycemic control prevents congenital abnormalities. Journal of the AmericanMedical Association 265: 731–6.Lapolla A, Dalfrà MG, Masin M, Bruttomesso D, Piva I, Crepaldi C et al. (2003). Analysis of outcomeof pregnancy in type 1 diabetics treated with insulin pump or conventional insulin therapy. ActaDiabetologica 40: 143–9.Loukovaara S, Immonen I, Teramo KA, Kaaja R (2003). Progression of retinopathy during pregnancyin type 1 diabetic women treated with insulin lispro. Diabetes Care 26: 1193–8.Masson EA, Patmore JE, Brash PD, Baxter M, Caldwell G, Gallen IW et al. (2003). Pregnancyoutcome in type 1 diabetes mellitus treated with insulin lispro (Humalog). Diabetic Medicine 20:46–50.Mathieson ER, Kinsley B, Amiel SA, Heller S, McCance D, Duran S, Bellaire S, Raben A on behalfof the Insulin Aspart Pregnancy Study Group (2007). Maternal glycemic control and hypoglycaemiain Type 1 diabetic pregnancy. Diabetes Care 30: 771–6.Mills JL, Knopp RH, Simpson JL, Jovanovic-Peterson L, Metzger BE, Holmes, LB et al. (1988). Lackof relation of increased malformation rates in infant of diabetic mothers to glycemic control duringorganogenesis. New England Journal of Medicine 318: 671–6.Nachum Z, Ben-Shlomo I, Weiner E, Shalev E (1999). Twice versus four times daily insulin doseregimens for diabetes in pregnancy: randomised controlled trial. British Medical Journal 319:1223–7.Nelson CA, Wewerka S, Thomas KM, Tribby-Walbridge S, deRegnier R-A, Georgieff M (2000).Neurocognitive sequelae of infants of diabetic mothers. Behavioral Neuroscience 114: 950–6.Nisell H, Persson B, Hanson U, Lunell NO, Nylund L, Sarby B, Thronstrom S (1994). Hormonal,metabolic, and circulatory responses to insulin-induced hypoglycemia in pregnant and nonpregnantwomen with insulin-dependent diabetes. American Journal of Perinatology 11: 231–6.Paretti E, Mecacci I, Papini M, Cloni R, Carignani L, Mignosa M et al. (2001). Third trimester maternalglucose levels from diurnal profiles in non-diabetic pregnancies. Correlation with sonographicperameters of fetal growth. Diabetes Care 24: 1319–23.Penney GC, Mair G, Pearson DW (2003a). Outcomes of pregnancies in women with type 1 diabetesin Scotland: a national population-based study. British Journal of Obstetrics and Gynaecology 110:315–18.Penney GC, Mair G, Pearson DW (2003b). The relationship between birth weight and maternalglycated haemoglobin (HbA 1c ) concentration in pregnancies complicated by type 1 diabetes. DiabeticMedicine 20: 162–6.Persson B, Hansson U (1993). Hypoglycaemia in pregnancy. Bailliere’s Clinical Endocrinology andMetabolism 7: 731–9.Persson B, Swahn M-L, Hjertberg R, Hanson U, Nord E, Nordlander E, Hansson L-O (2002). Insulinlispro therapy in pregnancies complicated by type 1 diabetes mellitus. Diabetes Research andClinical Practice 58: 115–21.

REFERENCES 237Ray JG, O’Brien TE, Chan WS (2001). Preconception care and the risk of congenital anomalies inthe offspring of women with diabetes mellitus: a meta-analysis. Quarterly Journal of Medicine 94:435–44.Rayburn W, Piehl E, Jacober S, Schork A, Ploughman L (1986). Severe hypoglycemia during pregnancy:its frequency and predisposing factors in diabetic women. International Journal of Obstetricsand Gynecology 24: 263–8.Reece EA, Hagay Z, Roberts AB, DeGennaro N, Homko CJ, Connolly-Diamond M, Sherwin R et al.(1995). Fetal Doppler and behavioral responses during hypoglycemia induced with the insulin clamptechnique in pregnant diabetic women. American Journal of Obstetrics and Gynecology 172: 151–5.Rosenn BM, Miodovnik M, Holcberg G, Khoury JC, Siddiqi TA (1995). Hypoglycemia: the price ofintensive insulin therapy for pregnant women with insulin-dependent diabetes mellitus. Obstetricsand Gynecology 85: 417–22.Rosenn BM, Miodovnik M, Khoury JC, Siddiqi TA (1996). Counterregulatory hormonal responses tohypoglycemia during pregnancy. Obstetrics and Gynecology 87: 568–74.Rossi G, Lapaczewski P, Diamond MP, Jacob RJ, Shulman GI, Sherwin RS (1993). Inhibitory effectof pregnancy on counterregulatory hormone responses to hypoglycemia in awake rat. Diabetes 42:1440–5.Scottish Diabetes in Pregnancy Study Group (2004). Audit of pregnancies in women with diabetes.Findings from a one year audit. Report to Scottish Programme for Clinical Effectiveness In ReproductiveHealth (SPCERH).SIGN Executive (2001). SIGN 55 Management of Diabetes. Scottish Intercollegiate GuidelinesNetwork, Edinburgh: 34–8.Smoak IW, Sadler TW (1990). Embryopathic effects of short-term exposure to hypoglycemia in mouseembryos in vitro. American Journal of Obstetrics and Gynecology 63: 619–24.Stangenberg M, Persson B, Stange L, Carlstrom K (1983). Insulin-induced hypoglycemia in pregnantdiabetics. Maternal and fetal cardiovascular reactions. Acta Obstetrica Gynecologica Scandinavica62: 249–52.Steel JM, Johnstone FD, Smith AF (1989). Pre-pregnancy preparation. In: Carbohydrate Metabolismin Pregnancy and the Newborn: IV. Sutherland HW, Stowers JM and Pearson DWM, eds. Springer-Verlag, Berlin: 129–39.Steel JM, Johnstone FD, Hepburn D, Smith AF (1990). Can pre-pregnancy care of diabetic womenreduce the risk of abnormal babies? British Medical Journal 301: 1070–3.The Diabetes Control and Complications Trial Research Group (1996). Pregnancy outcomes in theDiabetes Control and Complications Trial. American Journal of Obstetrics and Gynecology 174:1343–53.Wyatt JW, Frias JL, Hoymet HE, Jovanovic L, Kaaja R, Brown F et al., the IONS study group (2004).Congenital anomaly rate in offspring of mothers with diabetes treated with insulin lispro duringpregnancy. Diabetic Medicine 22: 803–7.Yamashita Y, Kawano Y, Kuriya N, Murakami Y, Matsuishi T, Yoshimatsu K, Kato H (1996).Intellectual development of offspring of diabetic mothers. Acta Paediatrica 85: 1192–6.Yssing M (1974). Oestriol excretion in pregnant diabetes related to long term prognosis of survivingchildren. Acta Endocrinologica 182: 95–104.Yogev Y, Chen R, Ben-Haroush A, Phillip M, Jovanovic L, Hod M (2003). Continuous glucosemonitoring for the evaluation of gravid women with type 1 diabetes mellitus. Obstetrics andGynecology 101: 633–8.

11 Hypoglycaemia in Type 2Diabetes and in Elderly PeopleNicola N. Zammitt and Brian M. FrierINTRODUCTIONHypoglycaemia has been thought to be related almost exclusively to the management of type1 diabetes and has received limited attention as a problem associated with the treatment oftype 2 diabetes. However, as the prevalence of type 2 diabetes escalates and life expectancyincreases, it is inevitable that the number of older people with insulin-treated diabetes willincrease steadily. Furthermore, the current emphasis on adherence to stricter glycaemictargets is encouraging the earlier use of insulin to treat type 2 diabetes, so influencing theprevalence and severity of hypoglycaemia in this group.Because hypoglycaemia is increasing as a clinical problem in people with type 2 diabetes,many of whom are elderly, it is pertinent to consider how age and type 2 diabetes per seaffect the pathophysiology and frequency of hypoglycaemia.PATHOPHYSIOLOGY OF HYPOGLYCAEMIAMost studies that have examined counterregulatory responses to hypoglycaemia have beenconducted in young adults with, and without, type 1 diabetes. The pathophysiological differencesthat exist between type 1 and type 2 diabetes have the potential to modify counterregulatorymechanisms, as does the increasing prevalence of type 2 diabetes in middle-aged andelderly individuals. The increasing age of the population with diabetes will also influencecounterregulation.The responses to hypoglycaemia can be considered in terms of generation of symptomresponses and counterregulatory hormonal changes. The two are independent, although somehormones such as epinephrine (adrenaline) can contribute to symptom generation; bothresponses are mediated centrally through glucose sensors in the brain. It is convenient toexamine symptoms and counterregulation separately.The Effects of Ageing on the Responses to HypoglycaemiaSymptomsThe symptoms of hypoglycaemia (Chapter 2) have been classified as autonomic, neuroglycopenicand non-specific (Deary et al., 1993). In elderly people who have diabetes,Hypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

240 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLETable 11.1 Common symptoms of hypoglycaemia in the elderly derivedby Principal Components Analysis (adapted from Jaap et al. (1998) withpermission from John Wiley & Sons, Ltd)Neuroglycopenic Autonomic NeurologicalPoor concentration Sweating UnsteadinessLightheadedness Shaking Poor co-ordinationWeakness Anxiety Blurred visionConfusion Pounding heart Double visionDrowsinessSlurred speechDizzinessan additional group of neurological symptoms affecting vision and co-ordination has beenidentified (Jaap et al., 1998; McAulay et al., 2001) (Table 11.1) and non-specific symptomsof general malaise are less prominent than in young adults. These neurological symptoms(and signs) that are generated as manifestations of hypoglycaemia in elderly people may beconfused with other conditions such as a transient ischaemic attack or a vaso-vagal episode.Many elderly people with type 2 diabetes have limited knowledge of the symptoms andtreatment of hypoglycaemia (Pegg et al., 1991; Thomson et al., 1991) and their relatives andcarers are equally ill-informed (Mutch and Dingwall-Fordyce, 1985). This magnifies the riskthat early hypoglycaemia will neither be identified nor treated, and will therefore progress tothe point where external help is required. Even in young adults, knowledge of diabetes andits treatment declines with time (Lawrence and Cheely, 1980), and in elderly people who arevulnerable to the effects of age-related cognitive decline, the need for regular educationalreinforcement is greater, although seldom undertaken.Elderly people report a profile of hypoglycaemic symptoms that is different from thatof young adults, and the intensity of their symptoms is less (Brierley et al., 1995; Matykaet al., 1997). A small British study that compared the responses to hypoglycaemia ofyoung and elderly non-diabetic adults, showed that symptom scores were significantlylower in the older group, with autonomic and neuroglycopenic symptoms being affectedequally (Brierley et al., 1995). Two Canadian studies have implicated diminished autonomicactivation as the reason for the lower magnitude of symptomatic response in theelderly and have suggested that the attenuation in symptom intensity is a feature ofincreasing age, independent of any effects of diabetes (Meneilly et al., 1994a; Meneillyet al., 1994b).The glycaemic thresholds at which symptomatic responses to hypoglycaemia are generated(Chapter 7) alter with age. In young adults, symptoms are generated at a blood glucose levelthat is approximately 1.0 mmol/l higher than the level at which cognitive function becomesimpaired (Schwartz et al., 1987). This allows sufficient time for remedial action to be takenbefore the development of severe neuroglycopenia (Mitrakou et al., 1991). However, inolder subjects these thresholds are much closer together and occur almost simultaneously(around 30 ± 02 mmol/l) (Matyka et al., 1997) (Figure 11.1), which truncates the timeavailable for self-treatment before the development of incapacitating neuroglycopenia. Thus,the altered symptom profile, lower symptom intensity and altered glycaemic thresholds inthe elderly combine to increase the risk of progression to severe hypoglycaemia in this agegroup.

PATHOPHYSIOLOGY OF HYPOGLYCAEMIA 241Arterialised bloodglucose (mmol/l)4.03.5SYMPTOMSYounger men(3.6 mmol/l)3.0SYMPTOMSREACTION TIMEOlder men (3.0 mmol/l)2.5FOUR CHOICEREACTION TIMEYounger men(2.6 mmol/l)2.0Figure 11.1 Glycaemic thresholds for subjective symptomatic awareness of hypoglycaemia and forthe onset of cognitive dysfunction in young and elderly non-diabetic males. Figure based on dataderived from Matyka et al. (1997) and reproduced from McAulay and Frier (2001), with permissionfrom John Wiley and Sons, LtdCounterregulationEarly studies examining counterregulatory responses in elderly non-diabetic adults yieldedconflicting results, partly because interpretation of the data was confounded by the presenceof co-morbidities in many of the participants. However, subsequent studies suggest that theremay be age-related alterations in counterregulation. One study used an intravenous infusionof insulin to compare the counterregulatory responses to hypoglycaemia in non-diabeticelderly and young adults (Marker et al., 1992). This study suggested that with advancingage the secretion of growth hormone and cortisol is diminished, with modest attenuation ofblood glucose recovery in older non-diabetic adults, in whom the rise in plasma epinephrineis slower than in younger subjects (Marker et al., 1992). The secretion of glucagon and therate of insulin clearance were also lower. A reduced rate of clearance of insulin has beennoted in several studies (Minaker et al., 1982; Reaven et al., 1982; Fink et al., 1985). Thechanges observed by Marker and colleagues were unaffected by preceding physical training,suggesting that they are not simply a consequence of a more sedentary lifestyle associatedwith ageing (Marker et al., 1992). The glycaemic thresholds for the secretion of glucagonand epinephrine in response to hypoglycaemia also occur at a lower blood glucose level thanin younger subjects. In young non-diabetic adults, these hormones are released at a bloodglucose level of 3.3 mmol/l, compared to approximately 2.8 mmol/l in older adults (Meneillyet al., 1994a).With increasing age, the intensity of the hypoglycaemic stimulus (as demonstrated byblood glucose nadir) appears to influence the magnitude of the counterregulatory response.

242 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLEIn clamp studies comparing elderly and young non-diabetic subjects, the magnitude ofthe glucagon and epinephrine responses was lower in the elderly group during mildhypoglycaemia (blood glucose 3.3 mmol/l). However, both age groups achieved a similarmagnitude of response at a lower blood glucose of 2.8 mmol/l, indicating that in theelderly, counterregulatory responses are preserved during more profound hypoglycaemia(Ortiz-Alonso et al., 1994).Two other similar studies in non-diabetic elderly subjects have not demonstrated anysignificant age-related impairment of the counterregulatory hormonal responses to hypoglycaemia(Brierley et al., 1995; Matyka et al., 1997). Furthermore, although the symptomaticand counterregulatory responses to hypoglycaemia may be modified by advancing age, it isnot known at what age these changes become apparent or whether these effects are influencedby gender or the menopause in women. When designing studies and analysing results,few investigations of hypoglycaemia in type 2 diabetes have taken age or sex into accountand few studies have included people aged over 70 years.The Effects of Type 2 diabetes on the Responses to HypoglycaemiaInterpretation of the results of early studies of the counterregulatory responses to hypoglycaemiais constrained by their technical limitations and only a few of the more recentstudies were designed to control for potentially confounding variables and embraced modernmethodology.SymptomsType 2 diabetes is a heterogeneous disorder and individuals with this condition may betreated with one or more treatment modalities. This raises the possibility that the agent usedto treat the condition could influence the symptoms of hypoglycaemia. However, when thesymptomatic responses to hypoglycaemia were examined in a group of non-diabetic subjectsby inducing hypoglycaemia with insulin or a sulphonylurea (tolbutamide), neither the naturenor the intensity of symptoms differed with either agent (Peacey et al., 1996).Some studies have compared the symptoms associated with insulin-induced hypoglycaemiain people with type 1 and type 2 diabetes. In a retrospective population survey,people with insulin-treated type 2 diabetes reported a similar symptom profile to peoplewith type 1 diabetes, who were matched for duration of insulin therapy, but not for ageor duration of diabetes (Hepburn et al., 1993). Furthermore, glucose clamp studies haveshown that after controlling for age, no significant differences exist in the symptoms ofhypoglycaemia between people with type 1 and type 2 diabetes (Levy et al., 1998). Thesymptoms of hypoglycaemia in type 2 diabetes appear to be unaffected by the nature of thedisorder or by the choice of treatment modality.CounterregulationEarly studies of the counterregulatory responses to hypoglycaemia in type 2 diabetes werelimited by factors such as differences in blood glucose nadir between the diabetic and controlgroups, poorly matched or absent control groups and the inconsistent methods used to induce

PATHOPHYSIOLOGY OF HYPOGLYCAEMIA 243hypoglycaemia with techniques such as intravenous or subcutaneous bolus injections ofinsulin (De Galan and Hoekstra, 2001).Three studies that have examined counterregulatory responses to hypoglycaemia in peoplewith type 2 diabetes, who were being treated with diet alone or with oral medication,have provided important observations. These studies demonstrated that the secretion ofcounterregulatory hormones occurs at higher blood glucose levels in individuals with type2 diabetes than in non-diabetic subjects (Korzon-Burakowska et al., 1998; Spyer et al.,2000) and in people with type 1 diabetes (Levy et al., 1998). In one study, a steppedglucose clamp was used to examine how glycaemic control influenced the counterregulatoryresponse to hypoglycaemia in 11 subjects with type 2 diabetes who were diet-treated orwere taking sulphonylureas, and they were compared with 10 subjects with type 1 diabetes(Levy et al., 1998). The group with type 2 diabetes was older than the group with type1 diabetes, so two non-diabetic control groups were matched for age and body weightto the diabetic subjects. In the subjects with type 2 diabetes, counterregulatory hormoneswere released at higher blood glucose levels than in those with type 1 diabetes. However,males were over-represented in the group with type 2 diabetes, and six of the subjects withtype 2 diabetes required high insulin infusion rates, so introducing potentially confoundingvariables and limiting interpretation of the results of this study. There is evidence thathyperinsulinaemia per se can suppress the release of glucagon in response to hypoglycaemia(Diamond et al., 1991; Liu et al., 1991a; Liu et al., 1991b; Fanelli et al., 1994) whileincreasing catecholamine and cortisol release (Davis et al., 1993) in non-diabetic and type1 diabetic subjects. Similarly, in non-diabetic and in type 1 diabetic subjects, women haveattenuated counterregulatory responses to hypoglycaemia compared to men (Amiel et al.,1993; Davis et al., 2000; Sandoval et al., 2003). The effects of gender and hyperinsulinaemiaon counterregulation have not yet been determined in type 2 diabetes.People with type 2 diabetes may have greater protection against hypoglycaemia becausethe counterregulatory responses commence at higher blood glucose levels than in nondiabeticpeople or in those with type 1 diabetes (Korzon-Burakowska et al., 1998; Levyet al., 1998; Spyer et al., 2000). However, when HbA 1c is lowered with intensive therapyin type 1 diabetes, the thresholds for the counterregulatory responses are shifted to a lowerglycaemic level (Cryer, 2002; Cryer et al., 2003) and the same phenomenon appears to occurin type 2 diabetes (Korzon-Burakowska et al., 1998; Levy et al., 1998).Counterregulatory deficiencies are associated with increasing duration of type 1 diabetes;it has been suggested that similar counterregulatory deficiencies may develop in type 2diabetes as the disorder progresses in severity. The secretory response of glucagon tohypoglycaemia is lost early in the natural history of type 1 diabetes (Chapter 6) and, althoughthe catecholamine response compensates for several years, it too declines with time (Cryer,2002). There is disagreement over whether the glucagon response in type 2 diabetes ismodestly diminished (Bolli et al., 1984; Meneilly et al., 1994b; Shamoon et al., 1994) orpreserved (Boden et al., 1983; Heller et al., 1987; Korzon-Burakowska et al., 1998; Levyet al., 1998). It is probable that the ability of an individual to secrete glucagon in responseto hypoglycaemia is related to their residual insulin secretory capacity and the ability tosuppress endogenous insulin secretion when blood glucose falls (Israelian et al., 2005).People with type 2 diabetes constitute a heterogeneous group, with varying degrees of insulindeficiency. Most investigators who have reported an intact glucagon response have studiedpeople with type 2 diabetes who were treated satisfactorily with oral anti-diabetic agentsand were therefore unlikely to have been insulin-deficient (Boden et al., 1983; Heller et al.,

244 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLE1987; Levy et al., 1998; Spyer et al., 2000). Furthermore, with one exception (Meneillyet al., 1994b) all of these studies examined middle-aged subjects in their fifth or sixth decade,despite the fact that most people with type 2 diabetes are older than this. The effects ofageing on counterregulation have not been addressed in these studies.To assess the role of insulin deficiency in promoting counterregulatory hormonal deficiencies,the counterregulatory responses to hypoglycaemia were examined in 15 non-diabeticcontrols and in 13 people with type 2 diabetes, six of whom were treated with insulin, and whowere demonstrated to be insulin-deficient with low plasma C-peptide, while the remainingseven were being treated with oral anti-diabetic agents (Segel et al., 2002). The glucagonresponse to hypoglycaemia was preserved in the patients on oral agents and in the nondiabeticcontrols, but was virtually absent in the insulin-deficient patients, demonstrating anassociation between acquired counterregulatory abnormalities and insulin deficiency in type2 diabetes. Deficient counterregulatory responses to hypoglycaemia were also observed in agroup of patients with type 2 diabetes who had moderate beta cell failure (Israelian et al.,2006). In this study, the reduction in endogenous insulin secretion that normally occursduring hypoglycaemia was delayed and reduced, and responses of glucagon and growthhormone were impaired (Israelian et al., 2006) (Figure 11.2). Thus, multiple defects in thecounterregulatory responses to hypoglycaemia emerge as beta cell function fails in type 2diabetes.In type 1 diabetes, individuals who experience frequent hypoglycaemia can developa condition that has been termed Hypoglycaemia Associated Autonomic Failure (HAAF)(Cryer, 1992; Cryer, 2004). The underlying premise is that recurrent hypoglycaemia leadsGlucagon (pg/ml)15012510075Growth hormone (µg/l) (pM × 10 –3 ) Non-diabetic subjects= T2DM SubjectsMEAN + SEM–30 0 30 60 90 120 –30 0 30 60Time (min)Cortisol (nM)7505002500Time (min)90 120Figure 11.2 Hypoglycaemia-induced responses of glucagon, epinephrine, cortisol and growthhormone in non-diabetic subjects and in subjects with type 2 diabetes. In type 2 diabetes significantlylower responses to hypoglycaemia were observed for glucagon (45% reduction, p

PATHOPHYSIOLOGY OF HYPOGLYCAEMIA 245to failure of the centrally mediated counterregulatory response to hypoglycaemia, resultingin impaired awareness of hypoglycaemia (Chapter 7). Impaired hypoglycaemia awarenesshas been associated primarily with type 1 diabetes but people with insulin-deficient type 2diabetes are also at risk of developing HAAF. In the study by Segel and colleagues (Segelet al., 2002), subjects with type 2 diabetes underwent a hypoglycaemic clamp on the first dayof the study followed by a subsequent period of hypoglycaemia later that same day. Theywere then exposed to a second hypoglycaemic clamp during the following day, at which timethe plasma glucose levels required to activate the hormonal and symptomatic responses weresignificantly lower than during the first hypoglycaemic clamp. These findings implicate arole for antecedent hypoglycaemia in modifying the glycaemic thresholds at which responsesto hypoglycaemia are triggered in type 2 diabetes; this suggests that individuals with insulintreatedtype 2 diabetes can develop HAAF.The preservation of the epinephrine response to hypoglycaemia, in conjunction withpossible insulin resistance in some individuals, may provide further defensive mechanismsto protect people with type 2 diabetes against hypoglycaemia. The lipolytic effects ofepinephrine can outweigh the anabolic effects of insulin on insulin-resistant adipose tissue,resulting in a rise in plasma free fatty acids in response to hypoglycaemia in type 2 diabetes(Bolli et al., 1984; Heller et al., 1987; Shamoon et al., 1994), but not in type 1 diabetes(Maggs et al., 1997). Epinephrine secretion during hypoglycaemia may therefore have agreater protective effect in insulin-resistant patients. Epinephrine also stimulates the mobilisationof glucose stored in the kidney and this may compensate in part for the impairedproduction of hepatic glucose observed in individuals who have a deficient glucagon responseto hypoglycaemia (Woerle et al., 2003).Moderators of Hypoglycaemia in type 2 diabetesAlthough factors such as sleep, consumption of alcohol and the timing of exercise are knownto influence the risk of hypoglycaemia in type 1 diabetes (see Chapters 3 and 5), their effectsare not known in type 2 diabetes. Prolonged treatment with insulin (greater than ten years) isa reliable predictor of increased risk of severe hypoglycaemia in type 2 diabetes (Donnellyet al., 2005) and when people with type 2 diabetes become insulin-deficient, their frequencyof severe hypoglycaemia rises towards that of individuals with type 1 diabetes who arematched for duration of treatment with insulin (Hepburn et al., 1993).Impaired awareness of hypoglycaemia is a major significant risk factor for severe hypoglycaemiain people with type 1 diabetes (Gold et al., 1994), but is less common in type2 diabetes (Hepburn et al., 1993; Henderson et al., 2003) with only 8% having impairedawareness in a retrospective survey of 215 individuals with insulin-treated type 2 diabetes.However, the incidence of hypoglycaemia was nine times greater in this small sub-groupthan in those with normal awareness (Henderson et al., 2003). Continuous glucose monitoringsystems (CGMS) can detect asymptomatic hypoglycaemia. In a prospective studyof 30 individuals with type 2 diabetes and 40 patients with type 1 diabetes, asymptomatichypoglycaemia was detected in 14 (47%) of the group with type 2 diabetes and in 25 (63%)of the group with type 1 diabetes (Chico et al., 2003). An Australian study used CGMS toexamine the frequency of hypoglycaemia over six days in 25 patients treated with sulphonylureas.In 14 (56%) of these subjects, readings of < 22 mmol/l lasting for at least 15minutes were recorded. The participants had no symptomatic awareness of these episodes

246 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLE(Hay et al., 2003). Although impaired awareness of hypoglycaemia is more common in type1 diabetes, it may be more frequent in type 2 diabetes than is appreciated.Variability of blood glucose levels also appears to be an important moderator of the risk ofhypoglycaemia. In a prospective observational trial, a predominantly male cohort of peoplewith insulin-treated type 2 diabetes recorded blood glucose profiles for eight weeks andreported all episodes of hypoglycaemia over one year. The probability of all hypoglycaemiawas highest in those who had a low mean blood glucose with a high standard deviation,implicating the variability of blood glucose values as a predictor of risk of hypoglycaemiain insulin-treated type 2 diabetes (Murata et al., 2004).Exercise and alcohol can also influence the risk of hypoglycaemia. The rate of skeletalmuscle glucose uptake during physical exertion is normal in people with type 2 diabetes,but hepatic glucose output is impaired compared to non-diabetic individuals (Chipkin et al.,2001); this can result in exercise-related hypoglycaemia during physical exertion. Exercisereduces postprandial blood glucose concentrations (Minuk et al., 1981; Larsen et al., 1997)and improves insulin sensitivity for up to 16 hours after physical activity (Trovati et al., 1984;Devlin et al., 1987), exposing the individual to an increased risk of hypoglycaemia for aprolonged period. However, a combination of moderate exercise and ingestion of alcohol didnot provoke acute hypoglycaemia in C-peptide-positive, middle-aged individuals with type2 diabetes, regardless of whether they had eaten or fasted before exercise (Rasmussen et al.,1999). Although alcohol impairs the counterregulatory responses to hypoglycaemia in type1 diabetes (Avogaro et al., 1993), it does not appear to affect recovery from hypoglycaemiain type 2 diabetes (Rasmussen et al., 2001).FREQUENCY OF HYPOGLYCAEMIA IN TYPE 2 DIABETESMild hypoglycaemia is defined by the ability to self-treat, while the need for externalassistance denotes a severe episode. The frequency of hypoglycaemia in people withtype 1 diabetes is described in Chapter 3. Mild hypoglycaemia occurs on average aroundtwice weekly (Pramming et al., 1991; Pedersen-Bjergaard et al., 2004) and the estimatedincidence of severe hypoglycaemia ranges from 1.0 to 1.7 episodes per patient per year(MacLeod et al., 1993; ter Braak et al., 2000; Pedersen-Bjergaard et al., 2004) with an annualprevalence of between 30% (MacLeod et al., 1993; The Diabetes Control and ComplicationsTrial Research Group, 1993; Stephenson et al., 1994) and 40% (ter Braak et al., 2000).The frequency of hypoglycaemia in type 2 diabetes cannot be summarised in an equallysuccinct manner because of the heterogeneity of this disorder and the range of treatmentmodalities available. Furthermore, many people with type 2 diabetes are elderly and thefrequency of hypoglycaemia is often underestimated in this group (McAulay et al., 2001).Most studies have been conducted retrospectively and both the definitions of hypoglycaemiaand the nature of the treatment modalities that have been examined differ between studies,thus hindering comparison. Studies have demonstrated that individuals with type 1 diabetescan reliably remember the occurrence of severe hypoglycaemia after an interval of oneyear, but that recall of mild hypoglycaemia becomes unreliable after a one week period(Pramming et al., 1991; Pedersen-Bjergaard et al., 2003). In people with insulin-treatedtype 2 diabetes, recall of severe hypoglycaemia is similarly robust over a period of oneyear (Akram et al., 2003) but the reliability of their recall of mild hypoglycaemia has notbeen examined. Finally, the treatment regimens and patient inclusion criteria used in clinical

FREQUENCY OF HYPOGLYCAEMIA IN TYPE 2 DIABETES 247trials may not be representative of the patient characteristics observed and the therapeuticinterventions employed within an unselected diabetic out-patient population and this maylimit the extent to which such data can be extrapolated. Table 11.2b summarises availableepidemiological data on hypoglycaemia in type 2 diabetes.Hypoglycaemia and Oral Antidiabetic AgentsHypoglycaemia caused by oral antidiabetic agents is primarily associated with the insulinsecretagogues. It is seldom a side-effect of treatment with thiazolidinediones or alphaglucosidaseinhibitors, and has been reported only infrequently in association with metformin,usually when food intake is restricted (UK Prospective Diabetes Study (UKPDS) Group1998a, United Kingdom Prospective Diabetes Study Group 1998c). The frequency of hypoglycaemiasecondary to sulphonylurea therapy is considerably lower than that attributedto insulin (UK Prospective Diabetes Study (UKPDS) Group 1998b; Miller et al., 2001;Leese et al., 2003) but is probably underestimated (Hartling et al., 1987). In a retrospectivesix-month study in England of 219 people with type 2 diabetes treated with sulphonylureasand metformin, 20% of those taking sulphonylureas (either alone or in combination withmetformin) had experienced hypoglycaemic symptoms (Jennings et al., 1989). Sulphonylureascan also cause severe hypoglycaemia. Over a seven-year period, the Swedish AdverseDrug Reactions Advisory Committee reported 19 cases of glipizide-associated severe hypoglycaemiapresenting with reduced consciousness, with two deaths (Asplund et al., 1991).Incidences of severe hypoglycaemia of 0.224 episodes per 100 person-years with longactingsulphonylureas versus 0.075 episodes per 100 patient-years were reported among 28patients (median age 73 years) who were admitted to hospital with severe hypoglycaemia inSwitzerland (Stahl and Berger, 1999). The frequency of severe hypoglycaemia may be underestimatedin this study as patients receiving treatment in the community were excluded. Arecent study, A Diabetes Outcome Progression Trial (ADOPT), in which newly diagnosedpatients with type 2 diabetes were randomised to treatment with rosiglitazone, metformin andglibenclamide for a median of four years, prospectively recorded the hypoglycaemic eventsassociated with each treatment (Kahn et al., 2006). One (0.1%) serious and 142 (9.8%) totalevents were recorded for rosiglitazone, one (0.1%) serious and 168 (11.6%) total events formetformin and eight (0.6%) serious and 557 (38.7%) total events for glibenclamide.The frequency of hypoglycaemia relates to the individual pharmacokinetic propertiesof each sulphonylurea (Table 11.3), with the long-acting agents such as chlorpropamide,glibenclamide and long-acting glipizide being associated with the greatest risk (Stahl andBerger, 1999; Del Prato et al., 2002; Rendell 2004). Glibenclamide is associated with agreater risk of severe hypoglycaemia than gliclazide (Tessier et al., 1994) because activemetabolites prolong its hypoglycaemic effects for 24 hours (Jonsson et al., 2001; Rendell2004). Glibenclamide also attentuates the glucagon response to hypoglycaemia in nondiabeticvolunteers (ter Braak et al., 2002) and in people with type 2 diabetes (Landstedt-Hallin et al., 1999; Banarer et al., 2002). Several drugs may potentiate the hypoglycaemiceffects of sulphonylureas (Table 11.4). Risk factors for severe hypoglycaemia associated withsulphonylurea therapy include age, a past history of cardiovascular disease or stroke, renalfailure, reduced food intake, alcohol ingestion and interactions with other drugs (Hartlinget al., 1987; Seltzer et al., 1989; Asplund et al., 1991; Campbell et al., 1994; Shorr et al.,1997; Ben-Ami et al., 1999; Burge et al., 1999; Harrigan, et al., 2001).

Table 11.2a Prevalence of severe hypoglycaemia (hypos) in type 2 diabetes (adapted from Zammitt and Frier, 2005).Study Jennings, 1989UKPDS 33,1998Hepburn, Abraira,1993 ∗ 1995Miller,2001Henderson,2003Leese,2003Donnelly,2005Akram,2006Design Retrospective,structured interviewProspective,multicentre RCTSubjects T2DM, OHA only T2DM, OHA andinsulinNumber 219 (203: SU, 16:metformin)RetrospectivequestionnaireInsulin-treatedT1DM andT2DM3935 104 T1DM, 104T2DMProspective,multicentre RCTRetrospectiveinterviewRetrospectivequestionnaireAll T2DM Type 2 Insulin-treatedT2DMPopulationbaseddatasetanalysis69 T1DM;T2DM SU(22), insulin(66)Prospective RetrospectivequestionnaireInsulin-treatedT1DM &T2DM153 1055 215 160 94 T1DM173 T2DMDuration 6 months 10 years 1 year 18–35 months 7 months 1 year I year One month 1 yearInsulin-treatedT2DM401PREVALENCEOHA: allhyposMetformin 0% ⎫⎬⎭NA NA 16% NA NA NA NASUs 202%overallGlibenclamide 313%Chlorpropamide 136%Gliclazide 131%Glibenclamide 17.0%Chlorpropramide11.0%OHA: SH Glibenclamide0.6%Insulin: allhyposChlorpropramide0.4%NA NA 0% NA 0.8% NA NANA 36.5% 82.7% 56% conventional 30% 64% NA 45% NA93% intensiveInsulin: SH NA 2.3% 10% NA 0% 15% 7.3% 3% 16.5%Abbreviations: T1DM = type 1 diabetes; T2DM = type 2 diabetes; NA = not applicable; SH = severe hypoglycaemia; SU = sulphonylurea; LOC = loss of consciousness; RCT = randomised clinicaltrial; OHA = oral hypoglycaemic agent; Hypos = hypoglycaemia; ∗ = only figures for type 2 DM given

FREQUENCY OF HYPOGLYCAEMIA IN TYPE 2 DIABETES 249Table 11.2b Incidence of severe hypoglycaemia (hypos) in type 2 diabetes. Figures are expressed asevents per patient per year (adapted from Zammitt and Frier, 2005)Study VA-CSDM 1995 Henderson 2003 Leese 2003 Donnelly 2005 Akram, 2006INCIDENCEOHA: all hypos NA NA NA NA NAOHA: SH NA NA SU: 0.009NANAMetformin: 0.0005Insulin: all hypos 1.5 (standard) NA NA 16.37 NA16.5 (intensive)Insulin: SH 0.02 0.28 0.12 (T1 and T2DM) 0.35 0.44Abbreviations: T1DM = type 1 diabetes; T2DM = type 2 diabetes; NA = not applicable; SH = severe hypoglycaemia; SU =sulphonylurea; LOC = loss of consciousness; RCT = randomised clinical trial; OHA = oral hypoglycaemic agent; Hypos =hypoglycaemiaTable 11.3 Pharmacokinetics of the sulphonylureas (data sourced from Kobayashi et al. (1984);Campbell et al. (1994); DeFronzo (1999); Harrower (2000); Harrigan et al. (2001); Schernthaner(2003))Generation Name t max (hours)t 1/2(hours)Duration ofaction (hours)Renal excretion ofactive metaboliteFirst Chlorpropamide 2–7 36 60 YesTolbutamide 3–4 3–28 6–12 InsignificantSecond Glipizide 1–3 7 12–24 NoGlipizide GITS 6–12 7 24 No(Glucotrol XL)Glibenclamide 2–6 10 12–24 YesGliclazide 2–3 12–14 Minimal (5%)Gliclazide MR 4–6 24 MinimalThird Glimepiride 2–3 5–9 16–24 Yes (?)t max = time to peak; t 1/2 = half-lifeTable 11.4 Drugs that potentiate the effects of sulphonylurea drugs used to treat type 2 diabetes (datasourced from Campbell et al. (1994); Harrigan et al. (2001))MechanismDecreased hepaticmetabolismDecreasedrenalexcretionDisplacement ofSU from albuminbinding sitesIncreasesplasmaconcentrationInhibition ofgluconeogenesisDrugs Chloramphenicol Allopurinol Fibrates Fluconazole, AlcoholH2 blockers Probenecid Trimethoprim MiconazoleCiprofloxacin Aspirin AspirinWarfarinWarfarinSulphonamidesSulphonamidesMonoamineoxidaseinhibitors

250 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLESome of the newer sulphonylureas such as glimepiride are associated with a lower risk ofhypoglycaemia. In a population-based study in Germany, the incidence of hypoglycaemiawas examined in people with type 2 diabetes who had attended a hospital emergency departmentover a four-year period (Holstein et al., 2001). A total of 45 episodes were recorded inindividuals who had been taking a sulphonylurea, of which 38 were associated with glibenclamide.Glimepiride was implicated in only six episodes, despite more frequent prescribingof glimepiride. Glimepiride was compared with a modified release form of gliclazide ina multicentre, European trial (Schernthaner et al., 2004a). Glycated haemoglobin valuesimproved by around 1.0% in both groups but modified release gliclazide was implicatedin fewer cases of hypoglycaemia (3.7%) compared to glimepiride (8.9%). Severe hypoglycaemiadid not occur in either group.The oral glucose prandial regulators, repaglinide and nateglinide, induce less frequenthypoglycaemia than the sulphonylureas because of their rapid onset of action and theirselective stimulation of insulin secretion in the presence of carbohydrate but not in thefasting state (Strange et al., 1999; Nattrass and Lauritzen, 2000; Culy and Jarvis, 2001). In arandomised multi-centre trial comparing repaglinide with nateglinide, slightly lower HbA 1cvalues were achieved after 16 weeks on repaglinide at the expense of mild hypoglycaemicepisodes in 7% of patients compared to no episodes in the nateglinide group (Rosenstocket al., 2004).Studies Comparing Hypoglycaemia Secondary to Insulin or OralAntidiabetic AgentsIt is often difficult to compare the frequency of hypoglycaemia ascribed to insulin andoral antidiabetic agents because in many studies the participants were receiving treatmentwith both modalities. The United Kingdom Prospective Diabetes Study Group (1998a)reported the prevalence of hypoglycaemia in people with type 2 diabetes using differenttherapies. A higher frequency of hypoglycaemia was observed in association with intensive,compared to conventional, treatment whether using sulphonylureas or insulin. The prevalenceof hypoglycaemia across the different treatment groups is summarised in Table 11.5. Of theoral agents used in this study, metformin had the lowest recorded rate of hypoglycaemiawhile glibenclamide had the highest. With intensive treatment, hypoglycaemia occurred mostfrequently in the insulin-treated patients, and the prevalence of hypoglycaemia was lowerin the first decade of the study than in later years. The prevalence of hypoglycaemia waslower when the groups were analysed on an ‘intention to treat’ basis because several patientsin the conventional-treatment groups required an increasing number of different therapieswith escalating doses as their glycaemic control deteriorated. Although the patients werequestioned about the occurrence of hypoglycaemia at every four-monthly review, only themost severe episode was recorded each time. This study cannot therefore provide an accurateestimate of the incidence of hypoglycaemia in type 2 diabetes and the overall prevalence ofsevere hypoglycaemia of 2% is deceptive as it does not indicate the rise in prevalence withincreasing duration of type 2 diabetes (United Kingdom Prospective Diabetes Study Group,1998a).A systematic review of randomised controlled trials comparing insulin monotherapy withcombination therapy with insulin and oral antidiabetic agents confirmed the relative safety ofthe latter treatment regimen and 13 out of 14 studies did not show any significant difference

FREQUENCY OF HYPOGLYCAEMIA IN TYPE 2 DIABETES 251Table 11.5 Hypoglycaemia episodes per year by intention to treat analysis and actual therapy forintensive and conventional treatment in the UKPDS: non-overweight patients (data sourced from UKProspective Diabetes Study (UKPDS) Group (1998a; 1998b))Mean proportion of patients per year with hypoglycaemia (%)Conventional Chlorpropamide Glibenclamide Insulin Metformin≥ 1 episodeMH in first10 years≥ 1 episodeSH in first10 years≥ 1 episodeMH in first10 years(intentionto treat)≥ 1 episodeSH in first10 years(intentionto treat)Normal BMI 12 110 177 365 –Overweight 09 121 175 340 42participantsNormal BMI 01 04 06 23 –Overweight 07 06 25 03 0participantsNormal BMIOverweightparticipantsNormal BMIOverweightparticipants10 16 21 28 –79 152 205 255 8307 10 14 18 –0.7 1.2 1.0 2.0 0.6MH = mild hypoglycaemia; SH = severe hypoglycaemia; BMI = body mass indexin hypoglycaemia rates between the two therapeutic strategies (Goudswaard et al., 2004).A randomised multicentre study compared the use of twice daily human Mixtard 30 (NovoNordisk) with once daily insulin glargine in combination with metformin and glimepiride andobserved that the combination regimen of insulin and oral agents achieved better glycaemiccontrol with fewer confirmed episodes of hypoglycaemia (4.07 versus 9.97 episodes per patientyear,p

252 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLEAfrican-Americans (Miller et al., 2001). A quarter of the group had experienced at leastone episode of hypoglycaemia during the study period. The prevalence of hypoglycaemiaincreased with more intensive treatment, the highest rate being associated with insulin.Severe hypoglycaemia occurred in 0.5% of patients, all of whom had been treated withinsulin. However, this study is limited by its reliance on patient recall of hypoglycaemia andthe ethnicity and gender of the study group, which makes it impossible to extrapolate theoutcomes to European populations.A 12-month prospective multicentre British study tested the hypothesis that the riskof hypoglycaemia in individuals with insulin-treated type 2 diabetes of short duration iscomparable to those taking sulphonylureas and is lower than in patients with newly diagnosedtype 1 diabetes (UK Hypoglycaemia Study Group, 2007). Monthly questionnaires were usedto record self-reported hypoglycaemia and a 72-hour period of continuous blood glucosemonitoring was used at the beginning and end of the study period to detect asymptomatichypoglycaemia. No difference was observed in the prevalences of severe hypoglycaemiain those with type 2 diabetes treated with sulphonylureas, and in individuals with type 2diabetes treated with insulin for less than two years (7% in both groups). Symptomatichypoglycaemia in patients with type 2 diabetes recently commenced on insulin therapy(< 2 years) was considerably lower than in those with type 1 diabetes of less than five yearsduration (median rate one versus 22 episodes per subject per year, p 5 years, showing a significantly higherprevalence in those with a longer duration of insulin therapy.1.0Proportion reporting at least one servere hypo0. 2 treatedwithsulphonylureastype2 < 2 yrsinsulintype2 > 5 yrsinsulintype1 < 5 yrsinsulintype1 > 15 yrsinsulinFigure 11.3 Proportion of patients with type 1 diabetes for < 5 years and > 15 years, and type2 diabetes in different treatment groups (sulphonylureas, insulin < 2 years, insulin > 5 years), whoexperienced one or more episodes of self-reported severe hypoglycaemia during 9–12 months offollow-up in the UK Hypoglycaemia Group Study. Reproduced from UK Hypoglycaemia Study Group(2007), Diabetologia, with kind permission of Springer Science and Business Media

FREQUENCY OF HYPOGLYCAEMIA IN TYPE 2 DIABETES 253Hypoglycaemia and InsulinIn the USA, the Veterans Affairs Cooperative Study in type 2 Diabetes (VA CSDM)examined glycaemic control and complications and compared a simple (‘standard’) insulinregimen (administered once daily) with an intensive (‘stepped’) regimen (Abraira et al.,1995). The participants in this trial had diabetes of relatively short duration (mean ± SD78 ± 4 years), were all insulin-treated males, and were followed up for only 18–35 months.The overall incidence of severe hypoglycaemia was 0.02 episodes per patient per year withno significant difference between the standard and stepped treatment groups. The frequencyof mild hypoglycaemia was significantly higher in the intensively treated group (steppedversus standard: 16.5 versus 1.5 episodes per patient per year). However, blood glucosewas monitored less frequently in the standard treatment group, which may have causedunder-reporting of asymptomatic hypoglycaemia.The participants in the VA CSDM study had a relatively short duration of diabetes. Aretrospective survey in Edinburgh of 215 people with insulin-treated type 2 diabetes observedthat the frequency of hypoglycaemia increased with the duration of insulin therapy andthe duration of type 2 diabetes (Henderson et al., 2003) (Figure 11.4) and was inverselyproportional to HbA 1c concentration. The annual prevalence of severe hypoglycaemia was15% with an overall incidence of 0.28 episodes per patient per year.The relationship between duration of insulin treatment and prevalence of severe hypoglycaemiahas been replicated in a more recent 12-month prospective multicentre Britishsurvey (UK Hypoglycaemia Study Group 2007) (Figure 11.3). A similar incidence wasreported in a retrospective study in Denmark of 401 patients with insulin-treated type 2diabetes, 66 (16.5%) of whom had experienced at least one episode of severe hypoglycaemiain the preceding year giving an overall incidence of 0.44 episodes per patient4035n = 2930Prevalence (%)25201510n = 147n = 39501–5 6–10 >10Duration of insulin therapy (years)Figure 11.4 Prevalence of severe hypoglycaemia in relation to duration of insulin therapy inpatients with type 2 diabetes. Reproduced from Henderson et al. (2003) by permission of BlackwellPublishing

254 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLEper year, but no relationship to HbA 1c was observed (Akram et al., 2006). A retrospectivestudy of 600 unselected insulin-treated diabetic patients performed a decade earlierin Edinburgh had observed an incidence of severe hypoglycaemia of 0.73 episodes perpatient per year in the 56 people with type 2 diabetes compared with 1.7 episodes perpatient per year in the 544 with type 1 diabetes (MacLeod et al., 1993). An earlier surveyin the same centre compared the frequency of severe hypoglycaemia in 86 people withinsulin-treated type 2 diabetes with 86 people with type 1 diabetes, matched for duration ofinsulin treatment and insulin dose (Hepburn et al., 1993). The frequency of severe hypoglycaemiawas similar in the two groups and a positive correlation was found betweenthe frequency of severe hypoglycaemia and duration of treatment with insulin (r = 039,p

MORBIDITY OF HYPOGLYCAEMIA AND NEED FOR EMERGENCY TREATMENT 255Incretin mimeticsGlucagon-like peptide 1 (GLP-1) is an incretin hormone that promotes glucose-dependentinsulin secretion and inhibition of glucagon production. GLP-1 is rapidly degraded in vivoby the ubiquitous enzyme Dipeptidyl Peptidase IV (DPP-IV) and no stable oral preparationis available. GLP-1 mimetics are associated with improvements in glycaemic control(Zander et al., 2002; Fineman et al., 2003; Koltermann et al., 2003; Degn et al., 2004a)without appearing to cause hypoglycaemia in people with type 2 diabetes (Knop et al., 2003;Madsbad et al., 2004). Exenatide, a synthetic GLP-1 receptor agonist, stimulates insulinrelease only in the presence of glucose (Degn et al., 2004b) and, although it suppressesglucagon production, the effect on glucagon suppression is abolished during hypoglycaemiaand the counterregulatory response to insulin-induced hypoglycaemia is preserved (Naucket al., 2002, Degn et al., 2004b). Although incretin mimetics may cause reactive hypoglycaemiain non-diabetic individuals (Meier and Nauck, 2005), they do not appear to causehypoglycaemia in people with type 2 diabetes (Knop et al., 2003, Vilsbøll et al., 2001).Few studies to date have quantified the risk of hypoglycaemia associated with DPP-IVinhibitors. In a one-year placebo-controlled trial in which the DPP-IV inhibitor, vildagliptinwas added to metformin, only four confirmed episodes of mild (self-treated) hypoglycaemiawere recorded in the 51 patients in the treatment arm (with no episodes in the placebo arm)and severe hypoglycaemia did not occur (Ahren et al., 2004).MORBIDITY OF HYPOGLYCAEMIA AND NEED FOREMERGENCY TREATMENTMany people with type 1 diabetes regard severe hypoglycaemia with the same degree oftrepidation as that reserved for the advanced complications of diabetes such as loss ofsight or renal failure (Pramming et al., 1991). Hypoglycaemia is not simply extremelyunpleasant for the individual concerned; it has the potential risk of severe morbidity andmay precipitate major vascular events such as stroke, myocardial infarction, acute cardiacfailure and ventricular arrhythmias (Landstedt-Hallin et al., 1999; McAulay and Frier, 2001;Desouza et al., 2003) (see Chapter 12). Healthcare professionals may not always recognise thecausative role of hypoglycaemia when treating these secondary events, especially if they areunfamiliar with some of the age-related neurological manifestations of hypoglycaemia. Theelderly are particularly at risk of hypoglycaemia-related physical injury and bone fracturesas a result of their general frailty and the presence of co-morbidities, such as osteoporosis(McAulay and Frier 2001). In a seven-year review of 102 cases of hypoglycaemic comasecondary to either insulin or glibenclamide, 92 patients had type 2 diabetes; seven sustainedphysical injury, five died, two suffered myocardial ischaemia and one patient had a stroke(Ben-Ami et al., 1999).In type 1 diabetes, relatives often treat severe hypoglycaemia at home, but while peoplewith insulin-treated type 2 diabetes experience severe hypoglycaemia less frequently, theyappear to require the assistance of the emergency services with equal frequency. This mightsuggest that people with insulin-treated diabetes are at greater risk of morbidity and disabilityduring hypoglycaemia, and they and their relatives may be less able to cope than youngerpeople with type 1 diabetes. In addition, many people with insulin-treated type 2 diabeteslive alone. A population survey in the region of Tayside in Scotland indicated that the annual

256 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLErate of severe hypoglycaemia requiring emergency medical intervention was similar in thesegroups (Leese et al., 2003). All episodes of severe hypoglycaemia requiring input from theemergency medical services in one year were identified. A total of 160 people with diabetesrequired treatment for 244 episodes of severe hypoglycaemia. Emergency treatment wasrequired for 7.1% of those with type 1 diabetes, 7.3% of those with insulin-treated type 2diabetes and 0.8% of people taking oral antidiabetic agents.In a different prospective survey in the same region of Tayside, the occurrence of hypoglycaemiawas monitored over a period of one month in a cohort of 267 people withinsulin-treated diabetes (both type 1 and type 2) (Donnelly et al., 2005). The prevalence ofall forms of hypoglycaemia in the group with insulin-treated type 2 diabetes was 45% withan incidence of 16.4 episodes per patient per year, compared to an incidence of 42.9 episodesper patient per year in type 1 diabetes. The incidence of severe hypoglycaemia was 0.35episodes per patient per year in the group with type 2 diabetes and 1.15 episodes per patientper year in those with type 1 diabetes. The figures for the incidences were extrapolated fromprospective data collected over one month but these calculated rates for people with type 1diabetes are consistent with those recorded in other European studies (Pramming et al., 1991;MacLeod et al., 1993; ter Braak et al., 2000; Pedersen-Bjergaard et al., 2004), suggestingthat the data collected were representative of the annual event rate. In this study, only 10%of the group with type 1 diabetes experiencing severe hypoglycaemia required emergencyservice treatment compared to one in three of the group with type 2 diabetes. Thus thefrequency of severe hypoglycaemia recorded in people with type 2 diabetes was higher thananticipated and their need to enlist the help of the emergency services was greater than inthose with type 1 diabetes.CONCLUSIONS• Ageing modifies the counterregulatory and symptomatic responses to hypoglycaemia.• In older people, effective self-treatment of hypoglycaemia may be compromised by theclose proximity of the glycaemic thresholds for the onset of symptoms and cognitivedysfunction, which occur almost simultaneously.• In type 2 diabetes, counterregulatory responses to hypoglycaemia commence at higherblood glucose levels than those observed in non-diabetic adults or in people with type 1diabetes, and this may have a protective effect. When insulin therapy is introduced andHbA 1c is reduced, these thresholds are shifted to lower blood glucose levels.• With progressive insulin deficiency, people with type 2 diabetes develop counterregulatoryhormonal deficiencies and impaired symptomatic awareness, similar to the acquiredhypoglycaemia syndromes of type 1 diabetes.• Hypoglycaemia in type 2 diabetes occurs most frequently with insulin therapy, butsulphonylurea-induced hypoglycaemia is also a significant but under-estimated problem.• Although less common than in type 1 diabetes, the frequency of hypoglycaemia in insulintreatedtype 2 diabetes rises progressively with increasing duration of insulin treatment.• People with insulin-treated type 2 diabetes are more likely to require the assistance ofemergency medical services to treat severe hypoglycaemia than those with type 1 diabetes.

REFERENCES 257• The paucity of data from elderly people with diabetes is of concern, as the morbidity ofhypoglycaemia is greater in this age group, their presenting features of hypoglycaemiamay be misinterpreted and prompt treatment may not be provided.REFERENCESAbraira C, Colwell JA, Nuttall FQ, Sawin CT, Nagel NJ, Comstock JP et al., the VA CSDM Group(1995). Veterans Affairs Cooperative Study on glycemic control and complications in Type IIDiabetes (VA CSDM). Diabetes Care 18: 1113–23.Ahren B, Landin-Olsson M, Jansson PA, Svensson M, Holmes D, Schweizer A (2004). Inhibition ofdipetidyl-peptidase-4 reduces glycemia, sustains insulin levels, and reduces glucagon levels in type2 diabetes. Journal of Clinical Endocrinology and Metabolism 89: 2078–84.Akram K, Pedersen-Bjergaard U, Borch-Johnsen K, Thorsteinsson B (2003). Recall of severe hypoglycaemicepisodes and course of hypoglycaemia awareness in insulin treated type 2 diabetes inone year follow-up. Diabetologia 46 (Suppl 2) A304 (Abstract).Akram K, Pedersen-Bjergaard U, Carstensen B, Borch-Johnsen K, Thorsteinsson B (2006). Frequencyand risk factors of severe hypoglycaemia in insulin-treated type 2 diabetes: a cross-sectional survey.Diabetic Medicine 23: 750–6.Allen KV, McAulay V, Sommerfield AJ, Frier BM (2004). Hypoglycaemia is uncommon with acombination of antidiabetic drugs and bedtime NPH insulin for type 2 diabetes. Practical DiabetesInternational 21: 179–82.Amiel SA, Maran A, Powrie JK, Umpleby AM, Macdonald IA (1993). Gender differences in counterregulationto hypoglycaemia. Diabetologia 36: 460–4.Anderson JH Jr, Brunelle RL, Keohane P, Koivisto VA, Trautmann ME, Vignati L, DiMarchi R (1997).Mealtime treatment with insulin analog improves postprandial hyperglycemia and hypoglycemiain patients with non-insulin-dependent diabetes mellitus. Multicenter Insulin Lispro Study Group.Archives of Internal Medicine 157: 1249–55.Asplund K, Wilholm B-E, Lundman B (1991). Severe hypoglycaemia during treatment with glipizide.Diabetic Medicine 8: 726–31.Avogaro A, Beltramello P, Gnudi L, Maran A, Valerio A, Miola M et al. (1993). Alcohol intakeimpairs glucose counterregulation during acute insulin-induced hypoglycemia in IDDM patients.Evidence for a critical role of free fatty acids. Diabetes 42: 1626–34.Banarer S, McGregor V, Cryer PE (2002). Intraislet hyperinsulinemia prevents the glucagon responseto hypoglycemia despite an intact autonomic response. Diabetes 51: 958–65.Bastyr EJ, Huang Y, Brunelle RL, Vignati L, Cox DJ, Kotsamos JG (2000). Factors associated withnocturnal hypoglycemia among patients with type 2 diabetes new to insulin therapy. DiabetesObesity and Metabolism 2: 39–46.Ben-Ami H, Nagachandran P, Mendelson A, Edoute Y (1999). Drug-induced hypoglycemic coma in102 diabetic patients. Arch Intern Med 159: 281–4.Boden G, Soriano M, Hoeldtke RD, Owen OE (1983). Counterregulatory hormone release and glucoserecovery after hypoglycemia in non-insulin-dependent patients. Diabetes 32: 231–7.Bolli G, Tsalikian E, Haymond M, Cryer P, Gerich JE (1984). Defective glucose counterregulation aftersubcutaneous insulin in noninsulin dependent diabetes mellitus: paradoxical lack of compensatoryincrease in glucose production, roles of insulin resistance, abnormal neuroendocrine responses andislet paracrine interactions. Journal of Clinical Investigation 73: 1532–41.Brierley EJ, Broughton DL, James OFW, Alberti KGMM (1995). Reduced awareness of hypoglycaemiain the elderly despite an intact counterregulatory response. Quarterly Journal of Medicine 88:439–45.

258 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLEBurge MR, Zeise T-M, Sobhy TA, Rassam AG, Schade DS (1999). Low-dose ethanol predisposeselderly fasted patients with type 2 diabetes to sulfonylurea-induced low blood glucose. DiabetesCare 22: 2037–43.Campbell IW, Chalmers J, Herlihy OM (1994). Sulphonylurea-induced hypoglycaemia in elderlypeople with diabetes. Practical Diabetes 11: 102–3.Chico A, Vidal-Rios P, Subira M, Novials A (2003). The continuous glucose monitoring system isuseful for detecting unrecognised hypoglycemias in patients with type 1 and type 2 diabetes but isnot better than frequent capillary glucose measurements for improving metabolic control. DiabetesCare 26: 1153–7.Chipkin SR, Klugh SA, Chasan-Taber L (2001). Exercise and diabetes. Cardiology Clinics 19: 489–505.Cryer PE (1992). Iatrogenic hypoglycemia as a cause of hypoglycemia-associated autonomic failurein IDDM. A vicious cycle. Diabetes. 41: 255–60Cryer PE (2002). Hypoglycaemia: The limiting factor in the glycaemic management of type I andtype II diabetes. Diabetologia 45: 937–48.Cryer PE, Davis SN, Shamoon H (2003). Hypoglycemia in diabetes. Diabetes Care 26: 1902–12.Cryer PE (2004). Diverse causes of hypoglycemia-associated autonomic failure in diabetes. NewEngland Journal of Medicine 350: 2272–9.Culy CR, Jarvis B (2001). Repaglinide: a review of its therapeutic use in type 2 diabetes mellitus.Drugs 61: 1625–60.Dailey G, Rosenstock J, Moses RG, Ways K (2004). Insulin glulisine provides improved glycemiccontrol in patients with type 2 diabetes. Diabetes Care 27: 2363–8.Davis SN, Goldstein RE, Jacobs J, Price L, Wolfe R, Cherrington AD (1993). The effects of differinginsulin levels on the hormonal and metabolic response to equivalent hypoglycemia in normalhumans. Diabetes 42: 263–72.Davis SN, Fowler S, Costa F (2000). Hypoglycemic counterregulatory responses differ between menand women with type 1 diabetes. Diabetes 49: 56–72.De Galan BE, Hoekstra JBL (2001). Glucose counterregulation in type 2 diabetes mellitus. DiabeticMedicine 18: 519–27.Deary IJ, Hepburn DA, MacLeod KM, Frier BM (1993). Partitioning the symptoms of hypoglycaemiausing multi-sample confirmatory factor analysis. Diabetologia 36: 771–7.DeFronzo RA (1999). Pharmacologic therapy for type 2 diabetes mellitus. Annals of Internal Medicine131: 281–303.Degn KB, Juhl CB, Sturis J, Jakobsen G, Brock B, Chandramouli V et al. (2004a). One week’streatment with the long-acting glucagon-like peptide 1 derivative liraglutide (NN2211) markedlyimproves 24h glycemia and - and -cell function and reduces endogenous glucose release inpatients with type 2 diabetes. Diabetes 53: 1187–94.Degn KB, Brock B, Juhl CB, Djurhuus CB, Grubert J, Kim D et al. (2004b). Effect of intravenousinfusion of exenatide (synthetic Exendin-4) on glucose-dependent insulin secretion and counterregulationduring hypoglycemia. Diabetes 53: 2397–403.Del Prato, Aragona M, Coppelli A (2002). Sulfonylureas and hypoglycaemia. Diabetes Nutrition andMetabolism 15: 444–51.Desouza C, Salazar H, Cheong B, Murgo J, Fonseca V (2003). Association of hypoglycemia andcardiac ischemia. Diabetes Care 26: 1485–89.Devlin JT, Hirshman M, Horton ED, Horton ES (1987). Enhanced peripheral and splanchnic insulinsensitivity in NIDDM men after single bout of exercise. Diabetes 36: 434–39.Diamond MP, Hallerman L, Starick-Zych K, Jones TW, Connolly-Howard M, Tamborlane WV,Sherwin RS (1991). Suppression of counterregulatory hormone response to hypoglycemia by insulinper se. Journal of Clinical Endocrinology and Metabolism 72: 1388–90.Donnelly LA, Morris AD, Frier BM, Ellis JD, Donnan PT, Durrant R et al., for the DARTS/MEMOcollaboration (2005). Frequency and predictors of hypoglycaemia in type 1 and insulin-treated type2 diabetes: a population based study. Diabetic Medicine 22: 449–55.

REFERENCES 259Fanelli C, Pampanelli S, Epifano L, Rambotti AM, Ciofetta M, Modarelli F et al. (1994). Relativeroles of insulin and hypoglycaemia on induction of neuroendocrine responses to, symptoms of anddeterioration of cognitive function in hypoglycaemia in male and female humans. Diabetologia 37:797–807.Fineman MS, Bicsak TA, Shen LZ, Taylor K, Gaines E, Varns A et al. (2003). Effect on glycemiccontrol of exenatide (synthetic exendin-4) additive to existing metformin and/or sulfonylurea treatmentin patients with type 2 diabetes. Diabetes Care 26: 2370–7.Fink RI, Revers RR, Kolterman OG, Olefsky JM (1985). The metabolic clearance of insulin and thefeedback inhibition of insulin secretion are altered with ageing. Diabetes 34: 275–80.Gold AE, MacLeod KM, Frier BM (1994). Frequency of severe hypoglycemia in patients with type 1diabetes with impaired awareness of hypoglycemia. Diabetes Care 17: 697–703.Goudswaard A, Furlong NJ, Rutten G, Stolk R, Valk G (2004). Insulin monotherapy versus combinationsof insulin with oral hypoglycaemic agents in patients with type 2 diabetes mellitus (Cochranereview). Cochrane Database of Systematic Reviews 4: CD003418.Haak T, Tiengo A, Draeger E, Suntum M, Waldhausl W (2005). Lower within-subject variability offasting blood glucose and reduced weight gain with insulin detemir compared to NPH insulin inpatients with type 2 diabetes. Diabetes Obesity and Metabolism 7: 56–64.Harrigan RA, Nathan MS, Beattie P (2001). Oral agents for the treatment of type 2 diabetes mellitus:pharmacology, toxicity and treatment. Annals of Emergency Medicine 38: 68–78.Harrower A (2000). Gliclazide modified release: from once-daily administration to 24-hour bloodglucose control. Metabolism 49 (Suppl 2): 7–11.Hartling SG, Faber OK, Wegmann M-L, Wahlin-Boll E, Melander A (1987). Interaction of ethanoland glipizide in humans. Diabetes Care 10: 683–6.Hay LC, Wilmhurst EG, Fulcher G (2003). Unrecognized hypo- and hyperglycemia in well-controlledpatients with type 2 diabetes mellitus: the results of continuous glucose monitoring. DiabetesTechnology and Therapeutics 5: 19–26.Heller SR, Macdonald IA, Tattersall RB (1987). Counterregulation in type II (non-insulin-dependent)diabetes mellitus. Normal endocrine and glycaemic responses up to 10 years after diagnosis.Diabetologia 30: 924–9.Henderson JN, Allen KV, Deary IJ, Frier BM (2003). Hypoglycaemia in insulin-treated type 2 diabetes:frequency, symptoms and impaired awareness. Diabetic Medicine 20: 1016–21.Hepburn DA, MacLeod KM, Pell ACH, Scougal IJ, Frier BM (1993). Frequency and symptoms ofhypoglycaemia experienced by patients with type 2 diabetes treated with insulin. Diabetic Medicine10: 231–7.Herman WH, Ilag LL, Johnson SL, Martin CL, Sinding J, AL Harthi A et al. (2005). A clinical trial ofcontinuous subcutaneous insulin infusion versus multiple daily injections in older adults with type2 diabetes. Diabetes Care 28: 1568–73.Hermansen K, Ronnemaa T, Petersen AH, Bellaire S, Adamson U (2004). Intensive therapy withinhaled insulin via the AERx insulin diabetes management system: a 12-week proof-of-concept trialin patients with type 2 diabetes. Diabetes Care 27: 162–7.Hermansen K, Davies M, Derezinski T, Martinez Ravn G, Clauson P, Home P, on behalf of the LevemirTreat-To-Target Group (2006). A 26-week, randomised, parallel, treat-to-target trial comparinginsulin detemir with NPH insulin as add-on therapy to oral glucose-lowering drugs in insulin-naïvepeople with type 2 diabetes. Diabetes Care 29: 1269–74.Hollander PA, Blonde L, Rowe R, Mehta AE, Milburn JL, Hershon KS et al., for theExubera Phase III Study Group (2004). Efficacy and safety of inhaled insulin (Exubera)compared with subcutaneous insulin therapy in patients with type 2 diabetes. Diabetes Care 27:2356–62.Holstein A, Plaschke A, Egberts E-H (2001). Lower incidence of severe hypoglycemia in patients withtype 2 diabetes treated with glimepiride versus glibenclamide. Diabetes Metabolism Research andReviews 17: 467–73.

260 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLEIsraelian Z, Gosmanov NR, Szoke E, Schorr M, Bokhari S, Cryer PE et al. (2005). Increasing thedecrement in insulin secretion improves glucagon responses to hypoglycemia in advanced type 2diabetes. Diabetes Care 28: 2691–6.Israelian Z, Szoke E, Woerle J, Bokhari S, Schorr M, Schwenke DC et al. (2006). Multiple defects incounterregulation of hypoglycemia in modestly advanced type 2 diabetes mellitus. Metabolism 55:593–8.Jaap AJ, Jones GC, McCrimmon RJ, Deary IJ, Frier BM (1998). Perceived symptoms of hypoglycaemiain elderly type 2 diabetic patients treated with insulin. Diabetic Medicine 15: 398–401.Janka HU, Plewe G, Riddle MC, Klieb-Frisch C, Schweitzer MA, Yki-Järvinen H (2005). Comparisonof basal insulin added to oral agents versus twice-daily premixed insulin as initial therapy for type2 diabetes. Diabetes Care 28: 254–9.Jennings AM, Wilson RM, Ward JD (1989). Symptomatic hypoglycemia in NIDDM patients treatedwith oral hypoglycemic agents. Diabetes Care 12: 203–7.Jonsson A, Chan JC, Rydberg T, Vaaler S, Hallengren B, Cockram CS et al. (2001). Pharmacodynamicsand pharmacokinetics of intravenous glibenclamide in Caucasian and Chinese patients with type-2diabetes. European Journal of Clinical Pharmacology 55: 721–7.Kahn S, Haffner SM, Heise MA, Herman WH, Holman RR, Jones NP et al., on behalf of the ADOPTstudy group (2006). Glycemic durability of rosiglitazone, metformin, or glyburide monotherapy.New England Journal of Medicine 355: 2427–43.Knop FK, Vilsbøll T, Larsen S, Madsbad S, Holst J, Krarup T (2003). No reactive hypoglycemiaafter subcutaneous administration of glucagon-like peptide-1 in lean type 2 diabetic patients and inpatients with diabetes secondary to chronic pancreatitis. Diabetes Care 26: 2581–7.Kobayashi K, Kimura M, Sakuguchi T, Hase A, Matsuoka A (1984). Pharmacokinetics of gliclazidein healthy and diabetic subjects. Journal of Pharmacological Science 73: 1684–7.Kolterman OG, Buse JB, Fineman MS, Gaines E, Heintz S, Bisack TA et al. (2003).Synthetic Exendin-4 (exenatide) significantly reduces postprandial and fasting plasma glucosein subjects with type 2 diabetes. Journal of Clinical Endocrinology and Metabolism 88:3082–8.Korzon-Burakowska A, Hopkins D, Matyka K, Lomas J, Pernet A, Macdonald I, Amiel S (1998).Effects of glycemic control on protective responses against hypoglycemia in type 2 diabetes.Diabetes Care 21: 283–90.Landstedt-Hallin L, Adamson U, Lins P-E (1999). Oral glibenclamide suppresses glucagon secretionduring insulin-induced hypoglycemia in patients with type 2 diabetes. Journal of ClinicalEndocrinology and Metabolism 84: 3140–5.Larsen JJ, Dela F, Kjaer M, Galbo H (1997). The effect of moderate exercise on postprandial glucosehomeostasis in NIDDM patients. Diabetologia 40: 447–53.Lawrence PA, Cheely J (1980). Deterioration of diabetic patients’ knowledge and management skillsas determined during out-patient visits. Diabetes Care 3: 214–8.Leese GP, Wang J, Broomhall JKelly P, Marsden A, Morrison W et al., for the DARTS/MEMO Collaboration(2003). Frequency of severe hypoglycemia requiring emergency treatment in type 1 and type 2diabetes: a population-based study of health service resource use. Diabetes Care 26: 1176–80.Levy CJ, Kinsley BT, Bajaj M, Simonson DC (1998). Effect of glycemic control on glucose counterregulationduring hypoglycemia in NIDDM. Diabetes Care 21: 1330–8.Liu D, Moberg E, Kollind M, Lins PE, Adamson U (1991a). A high concentration of circulatinginsulin suppresses the glucagon response to hypoglycemia in normal man. Journal of ClinicalEndocrinology and Metabolism 73: 1123–8.Liu DT, Adamson UC, Lins PE, Kollind ME, Moberg EA, Andreasson K (1991b). Inhibitory effect ofcirculating insulin on glucagon secretion during hypoglycemia in type I diabetic patients. DiabetesCare 15: 59–65.MacLeod KM, Hepburn DA, Frier BM (1993). Frequency and morbidity of severe hypoglycaemia ininsulin-treated diabetic patients. Diabetic Medicine 10: 238–45.

REFERENCES 261Madsbad S, Schmitz O, Ranstam J, Jakobsen G, Matthews DR, on behalf of the NN2211-1310 internationalstudy group (2004). Improved glycemic control with no weight increase in patients with type 2diabetes after once-daily treatment with the long-acting Glucagon-Like Peptide 1 analog Liraglutide(NN2211): a 12-week, double-blind, randomized, controlled trial. Diabetes Care 27: 1335–42.Maggs DG, Jacob R, Rife F, Caprio S, Tamborlane WV, Sherwin RS (1997). Counterregulation inperipheral tissues: effect of systemic hypoglycemia on levels of substrates and catecholamines inhuman skeletal muscle and adipose tissue. Diabetes 46: 70–6.Marker JC, Cryer PE, Clutter WE (1992). Attenuated glucose recovery from hypoglycemia in theelderly. Diabetes 41: 671–8.Matyka K, Evans M, Lomas J, Cranston I, Macdonald I, Amiel SA (1997). Altered hierarchy ofprotective responses against severe hypoglycemia in normal aging in healthy men. Diabetes Care20: 135–41.McAulay V, Deary IJ, Frier BM (2001). Symptoms of hypoglycaemia in people with diabetes. DiabeticMedicine 18: 690–705.McAulay V, Frier BM (2001) Hypoglycaemia. In: Diabetes in Old Age. Second edition. Sinclair AJ,Finucane P, eds. John Wiley and Sons, Ltd, Chichester: 133–52.McAulay V, Frier BM (2003). Insulin analogues and other developments in insulin therapy for diabetes.Expert Opinion in Pharmacotherapy 4: 1141–56.Meier JJ, Nauck MA (2005). Glucagon-like peptide 1 (GLP-1) in biology and pathology. DiabetesMetabolism Research and Reviews 21: 91–117.Meneilly GS, Cheung E, Tuokko H (1994a). Altered responses to hypoglycemia of healthy elderlypeople. Journal of Clinical Endocrinology and Metabolism 78: 1341–8.Meneilly GS, Cheung E, Tuokko H (1994b). Counterregulatory hormone responses to hypoglycemiain the elderly patient with diabetes. Diabetes 43: 403–10.Miller CD, Philips LS, Ziemer DC, Gallina DL, Cook CB, El-Kebbi IM (2001). Hypoglycemia inpatients with type 2 diabetes mellitus. Archives of Internal Medicine 161: 1653–9.Minaker KL, Rowe JW, Torino R, Pallotta JA (1982). Influence of age on clearance of insulin in man.Diabetes 31: 851–5.Minuk HL, Vranic M, Marliss EB, Hanna AK, Albisser AM, Zinman B (1981). Glucoregulatoryand metabolic response to exercise in obese noninsulin-dependent diabetes. American Journal ofEndocrinology 240: E458–64.Mitrakou A, Ryan C, Veneman T, Mokan M, Jenssen T, Kiss I et al. (1991). Hierarchy of glycemicthresholds for counterregulatory hormone secretion, symptoms and cerebral dysfunction. AmericanJournal of Endocrinology 266: E67–74.Murata GH, Hoffman RM, Shah JH, Wendel CS, Duckworth WC (2004). A probabilistic model forpredicting hypoglycemia in type 2 diabetes mellitus. Archives of Internal Medicine 164: 1445–50.Mutch WJ, Dingwall-Fordyce I (1985). Is it a hypo? Knowledge of the symptoms of hypoglycaemiain elderly diabetic patients. Diabetic Medicine 2: 54–6.Nattrass M, Lauritzen T (2000). Review of prandial glucose regulation with repaglinide: a solution tothe problem of hypoglycemia in the treatment of type 2 diabetes? International Journal of Obesityand Metabolic Diseases 24 (Suppl 3): S21–31.Nauck MA, Heimesaat MM, Behle K, Holst JJ, Nauck MS, Ritzel R et al. (2002). Effects of glucagonlikepeptide 1 on counterregulatory hormone responses, cognitive functions and insulin secretionduring hyperinsulinemic stepped hypoglycemic clamp experiments on healthy volunteers. Journalof Clinical Endocrinology and Metabolism 87: 1239–46.Ortiz-Alonso FJ, Galecki A, Herman WH, Smith MJ, Jacquez JA, Halter JB (1994). Hypoglycemiacounterregulation in elderly humans: relationship to glucose levels. American Journal ofEndocrinology 267: E497–506.Peacey SR, George E, Rostami-Hodjegan A, Bedford C, Harris N, Hardisty CA et al. (1996). Similarphysiological and symptomatic responses to sulphonylurea and insulin induced hypoglycaemia innormal subjects. Diabetic Medicine 13: 634–41.

262 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLEPedersen-Bjergaard U, Pramming S, Thorsteinsson B (2003). Recall of severe hypoglycaemia andself-estimated state of awareness in type 1 diabetes. Diabetes Metabolism Research and Reviews19: 232–40.Pedersen-Bjergaard U, Pramming S, Heller SR, Wallace TM, Rasmussen AK, Jorgensen HV et al.(2004). Severe hypoglycaemia in 1076 adult patients with type 1 diabetes: influence of risk markersand selection. Diabetes Metabolism Research and Reviews 20: 479–86.Pegg A, Fitzgerald D, Wise D, Singh BM, Wise PH (1991). A community-based study of diabetesrelatedskills and knowledge in elderly people with insulin-requiring diabetes. Diabetic Medicine 8:778–81.Pramming S, Thorsteinsson B, Bendtson I, Binder C (1991). Symptomatic hypoglycaemia in 411type 1 diabetic patients. Diabetic Medicine 8: 217–22.Raskin P, Allen E, Hollander P, Gabbay RA, Hu P, Bode B, Garber A, for the INITIATE Study Group(2005). Initiating insulin therapy in type 2 diabetes: a comparison of biphasic and basal insulinanalogs. Diabetes Care 26: 260–5.Raslova K, Bogoev M, Raz I, Leth G, Gall MA, Hancu N (2004). Insulin detemir and insulin aspart:a promising basal-bolus regimen for type 2 diabetes. Diabetes Research and Clinical Practice 66:193–201.Rasmussen BM, Christiansen C, Rasmussen OW, Hansen C, Hermansen K (1999). Alcohol andpostexercise metabolic responses in type 2 diabetes. Metabolism 48: 597–602.Rasmussen BM, Orskov L, Schmitz O, Hermansen K (2001). Alcohol and glucose counterregulationduring acute insulin-induced hypoglycemia in type 2 diabetic subjects. Metabolism 50: 451–7.Reaven GM, Greenfield MS, Mondon CE, Rosenthal M, Wright D, Reaven EP (1982). Does insulinremoval rate from plasma decline with age? Diabetes 31: 670–73.Rendell M (2004). The role of sulphonylureas in the management of type 2 diabetes mellitus. Drugs64: 1339–58.Riddle MC, Rosenstock J, Gerich J, on behalf of the insulin glargine 4002 study investigators (2003).Randomized addition of glargine or human NPH insulin to oral therapy of type 2 diabetic patients.Diabetes Care 26: 3080–6.Rosenstock J, Schwartz SL, Clark CM, Park GD, Donley DW, Edwards MB (2001). Basal insulintherapy in type 2 diabetes. Diabetes Care 24: 631–6.Rosenstock J, on behalf of the HOE 901/204 Study Investigators Group (2003). Safety and efficacyof insulin glargine (HOE 901) versus NPH insulin in combination with oral treatment in type 2diabetic patients. Diabetic Medicine 20: 545–51.Rosenstock J, Hassman DR, Madder RD Brazinsky, Farrell J, Khutoryansky N, Hale PM: (2004).Repaglinide versus nateglinide monotherapy. Diabetes Care 27: 1265–70.Ross SA, Zinman B, Campos RV, Strack T, Canadian Lispro Study Group (2001). A comparativestudy of insulin lispro and human regular insulin in patients with type 2 diabetes mellitus andsecondary failure of oral hypoglycemic agents. Clinical Investigative Medicine 24: 292–8.Sandoval DA, Ertl AC, Richardson MA, Tate DB, Davis SN (2003). Estrogen blunts neuroendocrineand metabolic responses to hypoglycemia. Diabetes 52: 1749–55.Schernthaner G (2003). Gliclazide modified release: A critical review of pharmacodynamic, metabolicand vasoprotective effects. Metabolism 52 (Supplement 1): 29–34.Schernthaner G, Grimaldi A, Di Mario U, Drzewoski J, Kempler P, Kvapil M et al. (2004a). GUIDEstudy: double blind comparison of once-daily gliclazide MR and glimepiride in type 2 diabeticpatients. European Journal of Clinical Investigation 34: 535–42.Schernthaner G, Kopp HP, Ristic S, Muzyka B, Peter L, Mitteregger G (2004b). Metabolic control inpatients with type 2 diabetes using Humalog Mix50 injected three times daily: crossover comparisonwith human insulin 30/70. Hormone and Metabolic Research 36: 188–93.Schwartz NS, Clutter WE, Shah SD, Cryer PE (1987). Glycemic thresholds for activation of glucosecounterregulatory systems are higher than the thresholds for symptoms. Journal of Clinical Investigation79: 777–81.

REFERENCES 263Segel SA, Paramore DA, Cryer PE (2002). Hypoglycemia-Associated Autonomic Failure in advancedtype 2 diabetes. Diabetes 51: 724–33.Seltzer HS (1989). Drug-induced hypoglycemia. Endocrinology and Metabolism Clinics of NorthAmerica 18: 163–82.Shamoon H, Friedman A, Canton C, Zacharowicz L, Hu M, Rossetti L (1994). Increased epinephrineand skeletal muscle responses to hypoglycemia in non-insulin-dependent diabetes mellitus. Journalof Clinical Investigation 93: 2562–71.Shorr RI, Ray WA, Daugherty JR, Griffin MR (1997). Incidence and risk factors for serious hypoglycemiain older persons using insulin or sulfonylureas. Archives of Internal Medicine 157: 1681–6.Spyer G, Hattersley A, Macdonald IA, Amiel S, MacLeod KM (2000). Hypoglycemic counterregulationat normal blood glucose concentrations in patients with well controlled type 2 diabetes. Lancet 356:1970–4.Stahl M, Berger W (1999). Higher incidence of severe hypoglycaemia leading to hospital admissionin type 2 diabetic patients treated with long-acting versus short-acting sulphonylureas. DiabeticMedicine 16: 586–90.Stephenson J, Fuller JH, on behalf of the EURODIAB IDDM Complications Study Group (1994).Microvascular and acute complications in IDDM patients: the EURODIAB IDDM complicationsstudy. Diabetologia 37: 278–85.Strange P, Schwartz SL, Graf RJ, Polvino W, Weston I, Marbury TC et al. (1999). Pharmacokinetics,pharmacodynamics, and dose-response relationship of repaglinide in type 2 diabetes. DiabetesTechnology and Therapeutics 1: 247–56.ter Braak EWMT, Appelman AMMF, van de Laak M, Stolk RP, van Haeften TW, Erkelens DW(2000). Clinical characteristics of type 1 diabetic patients with and without severe hypoglycemia.Diabetes Care 23: 1467–71.ter Braak EWMT, Appelman AMMF, van Der Tweel I, Erkelens DW, van Haeften TW (2002). Thesulfonylurea glibenclamide induces impairment of glucagon and growth hormone responses duringmild insulin-induced hypoglycemia. Diabetes Care 25: 107–12.Tessier D, Dawson K, Tetrault JP, Bravo G, Meneilly GS (1994). Glibenclamide versus gliclazide intype 2 diabetes of the elderly. Diabetic Medicine 11: 974–80.The Diabetes Control and Complications Trial Research Group (1993). The effect of intensive treatmentof diabetes on the development and progression of long-term complications in insulin-dependentdiabetes mellitus. New England Journal of Medicine 329: 977–86.Thomson FJ, Masson EA, Leeming JT, Boulton AJM (1991). Lack of knowledge of symptoms ofhypoglycaemia by elderly diabetic patients. Age and Ageing 20: 404–6.Trovati M, Carta Q, Cavalot F, Vitali S, Banaudi C, Lucchina PG et al. (1984). Influence of physicaltraining on blood glucose control, glucose tolerance, insulin secretion and insulin action in noninsulin-dependentdiabetic patients. Diabetes Care 7: 416–20.UK Hypoglycaemia Study Group (2007). Risk of hypoglycaemia in types 1 and 2 diabetes: effects oftreatment modalities and their duration. Diabetologia 50: 1140–7.UK Prospective Diabetes Study (UKPDS) Group (1998a). Effect of intensive blood-glucose controlwith metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet352: 854–65.UK Prospective Diabetes Study (UKPDS) Group (1998b). Intensive blood-glucose control with sulphonylureasor insulin compared with conventional treatment and risk of complications in patients withtype 2 diabetes (UKPDS 33). Lancet 352: 837–52.United Kingdom Prospective Diabetes Study Group (1998c). A 6 year randomised controlled trialcomparing sulfonylurea, insulin and metformin therapy in patients with newly diagnosed type 2 diabetesthat could not be controlled with diet therapy (UKPDS 24). Annals of Internal Medicine 128: 165–75.Vilsbøll T, Krarup T, Madsbad S, Holst JJ (2001). No reactive hypoglycaemia in type 2 diabeticpatients after subcutaneous administration of GLP-1 and intravenous glucose. Diabetic Medicine18: 144–9.

264 HYPOGLYCAEMIA IN TYPE 2 DIABETES AND IN ELDERLY PEOPLEWeissberg-Benchell J, Antisdel-Lomaglio J, Seshadri R (2003). Insulin pump therapy. A meta-analysis.Diabetes Care 26: 1079–87.Woerle HJ, Meyer C, Popa EM, Cryer PE, Gerich JE (2003). Renal compensation for impaired hepaticglucose release during hypoglycemia in type 2 diabetes. Further evidence for hepatorenal reciprocity.Diabetes 52: 1386–92.Yki-Järvinen H, Dressler A, Ziemen M, the HOE 901/3002 study group (2000). Less nocturnal hypoglycemiaand better post-dinner glucose control with bedtime insulin glargine compared with bedtimeNPH insulin during insulin combination therapy in type 2 diabetes. Diabetes Care 23: 1130–6.Zammitt NN, Frier BM (2005). Hypoglycemia in type 2 diabetes: pathophysiology, frequency andeffects of different treatment modalities. Diabetes Care 28: 2948–61.Zander M, Madsbad S, Madsen JL, Holst JJ (2002). Effect of 6-week course of glucagon-like peptide1 on glycaemic control, insulin sensitivity and beta-cell function in type 2 diabetes: a parallel groupstudy. Lancet 359: 824–30.

12 Mortality, CardiovascularMorbidity and Possible Effectsof Hypoglycaemia on DiabeticComplicationsMiles Fisher and Simon R. HellerINTRODUCTIONFor patients with type 1 diabetes, insulin-induced hypoglycaemia is one of the most fearedconsequences of the disorder (Pramming et al., 1991). It is the major factor that restrainsmany patients from pursuing intensive insulin therapy and trying to achieve the levels of strictglycaemic control that are necessary to prevent the development of diabetic complications.Some fear the immediate lack of self control which can accompany the impairment ofcognitive function during acute hypoglycaemia. Others are embarrassed by the dependenceon other people for assistance during an episode of severe hypoglycaemia. Many patientsshare the worries expressed by some diabetes healthcare professionals about the possiblelong-term effects of recurrent hypoglycaemia on the brain (see Chapter 13).An additional factor which may dissuade many patients from improving their glycaemiccontrol is the fear of dying during an episode of hypoglycaemia, especially when lowblood glucose occurs during sleep. These anxieties may be shared by the patient’s relativeswho may have witnessed previous episodes of nocturnal hypoglycaemia or convulsions,about which the patient has no recollection. Fear of hypoglycaemia was heightened in the1980s by the publicity that surrounded the possible adverse effects of human insulin, andin particular the knowledge that some young people with type 1 diabetes had died suddenlyand unexpectedly, the so-called ‘dead in bed syndrome’ (Campbell, 1991).It would be wrong for healthcare professionals to dismiss such fears as irrational. Manyprofessionals will have first- or second-hand experience of the sudden death of a patientwith type 1 diabetes in circumstances that have implicated acute hypoglycaemia. Thischapter examines the epidemiology and causes of death from hypoglycaemia in patients withdiabetes, including those risk factors that appear to be associated with sudden death. The‘dead in bed syndrome’ is explored in detail, and comparisons drawn with other syndromesof sudden death in people who do not have diabetes. Putative mechanisms and risk factorsfor sudden death are described. Hypoglycaemia may also cause significant cardiovascularmorbidity in people with diabetes, and the effects on heart disease and cardiovascularHypoglycaemia in Clinical Diabetes, 2nd Edition.© 2007 John Wiley & Sons, LtdEdited by B.M. Frier and M. Fisher

266 MORTALITY, CARDIOVASCULAR MORBIDITY AND DIABETIC COMPLICATIONSdisease are examined. Finally, the hypothesis that hypoglycaemia may worsen the chronicmicrovascular complications of diabetes is examined.DEATHS ATTRIBUTABLE TO HYPOGLYCAEMIAProblems with Death CertificationIn people with diabetes the cause of death is often recorded inaccurately on the deathcertificate, even to the extent of omitting ‘diabetes’ altogether (Mulhauser et al., 2002;Waernbaum et al., 2006). Subsequent problems with analysis of death certificates can becompounded in the United Kingdom by the fact that ‘hypoglycaemia’ is not coded underone single heading. Furthermore, the exact coding of the cause of death is often left to thediscretion of individual coding clerks who usually have no knowledge of the clinical details(Tattersall and Gale, 1993). These factors make it impossible to obtain any precise estimatesof the frequency of sudden death, in contrast to records of the frequency of episodes of severehypoglycaemia. Many episodes of hypoglycaemia, including nocturnal hypoglycaemia, arenot recognised by patients or carers. If we add the difficulty of confirming hypoglycaemiaat post-mortem, it is not surprising that considerable uncertainty and variation surround theestimated number of deaths that are attributed to hypoglycaemia in people with diabetes.However, since hypoglycaemia is so common, we can conclude that the risk of death duringan individual episode is extremely low.Problems of Establishing the Cause of Death at Post-mortemIn attempting to establish a post-mortem diagnosis of hypoglycaemia, the pathologist needs toperform biochemical tests, examine the brain for evidence of hypoglycaemic brain damage,and exclude any other possible cause of death (Tattersall and Gale, 1993). Carbohydratemetabolism continues after death, and post-mortem changes in blood glucose can causedifficulties in confirming a hypoglycaemic death forensically. The continuing breakdownof glycogen (glycogenolysis) increases the blood glucose concentration in the inferior venacava, so that the presence of a normal or high blood glucose concentration on the rightside of the heart does not exclude ante-mortem hypoglycaemia (a false negative result for adiagnosis of hypoglycaemia). In the peripheral circulation, glucose continues to be utilisedby red blood cells, so that the presence of a low glucose concentration does not necessarilyindicate ante-mortem hypoglycaemia. Indeed low blood glucose is often found after deathin those without diabetes (a false positive result for a diagnosis of hypoglycaemia).The measurement of the glucose concentration in the vitreous humour presents similarproblems because of continued post-mortem glucose utilisation, and so this cannot be usedto confirm ante-mortem hypoglycaemia (false positive). A normal or raised glucose concentrationin the vitreous humour after death, however, excludes hypoglycaemia at the time ofdeath (true negative). Thus, the sensitivities and specificities of blood and vitreous humourmeasurements of glucose in diagnosing ante-mortem hypoglycaemia are unknown.In addition to the biochemical problems in diagnosing hypoglycaemia after death, errorsmay be introduced by attribution bias of the pathologist performing the post-mortem. Deathmay be attributed to minor degrees of coronary heart disease, since it is so common in the

RISKS OF DEATH FROM HYPOGLYCAEMIA 267diabetic population (false negative result for a diagnosis of hypoglycaemia). Alternatively, thepathologist, even when unsure may attribute death to hypoglycaemia rather than indicatingno cause on a certificate (false positive).CRUDE ESTIMATES OF MORTALITY FROM HYPOGLYCAEMIABecause of the problems detailed above, any estimate of mortality from hypoglycaemiawill be crude. Published reports range from no deaths attributable to hypoglycaemia at oneextreme to between 20% and 25% in reports from some Scandinavian centres. Most studiessuggest that the proportion of deaths caused by hypoglycaemia is between 2% and 6%, alower frequency than those associated with ketoacidosis. For a more detailed analysis, thereader is referred to the review by Tattersall and Gale (1993).If deaths caused by renal failure or coronary heart disease in people with diabetes continueto decline as diabetes care improves, then the relative proportion of deaths caused byhypoglycaemia may increase. This is particularly likely if intensive insulin therapy continuesto be adopted more widely in an attempt to prevent or reduce microvascular disease (TheDCCT Research Group 1991; The Diabetes Control and Complications Trial ResearchGroup 1993).RISKS OF DEATH FROM HYPOGLYCAEMIAThe risk factors that are commonly cited as increasing the risk of death from hypoglycaemiaare often anecdotal, and may owe more to the prejudices of individual clinicians than to scientificevidence. Those suggested are detailed in Box 12.1 and include alcohol abuse and/orinebriation (Arky et al., 1968; Kalimo and Olsson, 1980; Critchley et al., 1984; MacCuish,1993), psychiatric illness or personality disorder (Shenfield et al., 1980; Tunbridge, 1981),self-neglect (Tunbridge, 1981), resistance to education (Shenfield et al., 1980), hypopituitarismfollowing pituitary ablation therapy for proliferative retinopathy (Nabarro et al.,1979; Shenfield et al., 1980), and patients who have diabetes secondary to pancreatic disease(MacCuish, 1993).Box 12.1Possible risk factors for death from hypoglycaemia• Alcoholism and/or inebriation• Psychiatric illness or personality disorder• Self-neglect; inanition• Fecklessness/resistance to education• Diabetes secondary to pancreatic disease• Hypopituitarism following pituitary ablation

268 MORTALITY, CARDIOVASCULAR MORBIDITY AND DIABETIC COMPLICATIONSSUDDEN DEATHThe sudden and unexpected death of a young person is an infrequent event, and a series ofdeaths either from accidents or from natural causes creates considerable media interest. At aBritish inquest in the early 1990s, it was suggested that the frequency of sudden death inpeople with type 1 diabetes was increasing, that these deaths were caused by hypoglycaemiaand that there might be a link with the clinical use of human insulin preparations. Thesecomments received widespread coverage in the British media, and fuelled the controversyabout hypoglycaemia associated with human insulin.Sudden death does occur, albeit very infrequently, in young people who do not havediabetes (Box 12.2), and examination of the recognised causes may give some insight intopossible mechanisms of sudden death in those with type 1 diabetes (Box 12.3). Indeed, eventhe phenomenon of unexplained death in diabetes is not a new one. In the 1960s, Malins(1968) described 14 patients who had been attending his diabetes outpatient clinic who diedin circumstances implicating hypoglycaemia and in whom no alternative cause of death wasidentified at autopsy. Eight were over 60 years of age, and the clinical records revealeda history of poor nutrition, treatment with a large dose of long-acting insulin, nocturnalhypoglycaemia, and the absence of an alert family member. This description is strikinglysimilar to the circumstances described by Tattersall and Gill (1991) more than 20 yearslater, although, since the patients described by Malins were much older, their deaths aremore likely to have been related to established cardiovascular disease. Sudden, unexpecteddeath has also been described in type 1 diabetic patients with advanced diabetic autonomicneuropathy (Ewing et al., 1991).Unexplained Deaths of Type 1 Diabetic PatientsFollowing the publicity generated by the assertion that there had been an increase in suddendeaths from acute hypoglycaemia, the British Diabetic Association (now Diabetes UK)Box 12.2Syndromes of sudden death in non-diabetic young people• Hypertrophic obstructive cardiomyopathy (HOCM)• Coronary heart disease (severe coronary artery occlusion or myocardial infarction)• Other cardiac anatomical abnormalities (congenital anomalies of the coronaryarteries, right ventricular dysplasia)• Syndromes of QT prolongation• Epilepsy• Phaeochromocytoma• Sudden death in water• Toxic substance abuse

SUDDEN DEATH 269Box 12.3Possible mechanisms contributing to sudden death• Ventricular arrhythmias/fibrillation (HOCM, coronary artery occlusion)• Increased epinephrine (sport, phaeochromocytoma, sudden death in water)• Decreased potassium (sport, phaeochromocytoma)• Autonomic stimulation/bradycardia (cold water immersion)• Severe hyperkalaemia and red cell lysis (fresh water drowning)• Respiratory arrest (autonomic neuropathy)• Asystole (epilepsy)requested notification of sudden, unexpected deaths of young people with type 1 diabetes.Cases were referred by the forensic chemist who had publicised the issue, relatives or friendswho had heard of the television coverage, and physicians with an interest in diabetes whoreported recent sudden deaths of patients under their care. Detailed analysis was published byTattersall and Gill (1991) on behalf of the British Diabetic Association. A total of 53 caseswere referred, but the analysis was confined to the 50 cases who were under 50 years of age.Five cases were excluded because a definite cause of death was identified at post-mortem,and in 11 cases, death was the result of suicide or self-poisoning. Six patients were thoughtto have died from ketoacidosis, two from hypoglycaemic brain damage, and in four cases thedeath was totally unexplained. The largest group comprised 22 patients who were classifiedas ‘dead in bed’, and in an accompanying editorial the term ‘dead in bed syndrome’ wasused (Campbell, 1991) (Figure 12.1).Analysis of the 22 ‘dead in bed’ patients showed that they were aged between 12 and43 years, with duration of diabetes from 3 to 27 years. All had been treated with humaninsulin, and three were taking four injections a day, 18 twice daily insulin and one wasinjecting insulin once a day. Information on diabetic complications was not available forall cases, but 13 had no complications and only four had severe complications, suggestingthat undiagnosed autonomic neuropathy was not a factor. All died outside hospital, 19 weresleeping alone at the time of death, and 15 died during the night. Twenty patients werefound lying in an undisturbed bed.Because 14 patients had a history of severe nocturnal hypoglycaemia, and most wereapparently well on retiring to bed but were found dead in the morning, the scenario wasconsistent with an episode of severe or protracted nocturnal hypoglycaemia having precipitatedsudden death. Although all had been taking human insulin at the time of death, mosthad been transferred from animal insulin between six months and two years earlier, andthe authors concluded that no temporal relationship between the change in insulin speciesand the fatal event could be demonstrated. They proposed that ‘circumstantial evidenceimplicates nocturnal hypoglycaemia in many cases’. Neuropathological evidence of hypoglycaemiawas rare however, suggesting that protracted neuroglycopenia had not occurredin most patients, and implicating sudden cardiac or respiratory arrest as a direct consequenceof hypoglycaemia.

270 MORTALITY, CARDIOVASCULAR MORBIDITY AND DIABETIC COMPLICATIONS211522“Dead in bed”UnexplainedDiabetic ketoacidosisSuicideDefinite cause of death64Hypoglycaemic brain damageFigure 12.1Gill (1991)Sudden deaths of 50 young type 1 diabetic patients. Data derived from Tattersall andA study in Sweden of over 4000 patients with type 1 diabetes diagnosed before the age of14 and followed up for 13 years showed a three-fold increase in the standardised mortalityrate in comparison to a non-diabetic population (Sartor and Dahlquist, 1995). In nine of the33 deaths the subjects were found ’dead in bed’, eight of whom had gone to bed apparentlyin good health but subsequently were found dead. One subject had a cardiac arrest after hermorning injection of insulin, and at autopsy one had lacerations inside the mouth, suggestingpreceding convulsions. The authors suggested that hypoglycaemia was the most likely causeof death in all patients. Very little additional information was provided, and because glycatedhaemoglobin values could not be standardised it was not possible to ascertain whether anassociation existed between strict glycaemic control and sudden death. A further publicationfrom the same Swedish Childhood Diabetes Register has extended the dataset for another tenyears, and identified another eight subjects who were found deceased in bed at home by closerelatives without any cause of death found at forensic autopsy (Dahlquist and Kallen, 2005).Another study from Norway of patients under the age of 40, identified 240 deaths fromall causes and 16 cases that fulfilled the criteria of ‘dead in bed syndrome’ (Thordarsonand Sovik, 1995). This represented 6.7% of all deaths in this age group. All were found inan undisturbed bed, and nine had been on regimens requiring multiple injections of insulin(eight were taking five injections a day, with one taking a total of seven injections a day).Frequent episodes of hypoglycaemia were documented in 12 cases, in ten of whom nocturnalhypoglycaemia had occurred. Although autopsy had been performed in 13 patients, a causeof death was not evident in any case. Again, the authors concluded that hypoglycaemia wasthe most likely precipitant of death. A similar study from Denmark showed an associationbetween chronic alcohol abuse or acute alcohol intoxication and subjects who were founddead in bed (Borch-Johnsen and Helweg-Larsen, 1993). The number of subjects who werefound dead in bed remained remarkably constant over a seven year period, and no associationwas observed with an increasing usage of human insulin in Denmark during that time(Figures 12.2 and 12.3).

SUDDEN DEATH 271Figure 12.2 Number of sudden deaths and human insulin as % of total sales of insulin in Denmarkfrom 1982 to 1988. Reproduced from Borch-Jensen and Helweg-Larsen (1993) with permission fromJohn Wiley & Sons, LtdFigure 12.3 Number of deaths due to definite (closed bars) and possible (open bars) hypoglycaemiafrom 1982 to 1988. Reproduced from Borch-Jensen and Helweg-Larsen (1993) with permission fromJohn Wiley & Sons, LtdA recent study from Norway reported on long-term mortality in all people with type 1diabetes who had been diagnosed between 1973 and 1982, and who were less than 15 yearsof age at diagnosis (Skrivarhaug et al., 2006). Mortality was recorded from diabetes onsetuntil the end of 2002. They reported 103 deaths with 17 sudden or unexpected deaths, andfour of these met the criteria for ‘dead in bed’ (patients found dead in an undisturbed bed;observed to be in good health the day before; autopsy not informative). Common causes ofdeath were acute metabolic complications of diabetes, violence and cardiovascular disease,including a 16 year old male with myocardial infarction confirmed by autopsy.

272 MORTALITY, CARDIOVASCULAR MORBIDITY AND DIABETIC COMPLICATIONSImportance of Young People Without Typical Cardiac DiseaseWhen assessing the risk of sudden death in people with diabetes, the risks have to becompared with people without diabetes. Sudden unexpected death can occur in any youngperson irrespective of whether they have diabetes or not. The critical issue is whether suddendeath occurs more frequently in individuals with type 1 diabetes. Studies that have measuredthe frequency of sudden death in young people have reported rates around 1.3 to 8.5 per100 000 patient-years. Since they are based on large numbers of subjects these estimatesare probably more accurate than data reporting the risks in patients with type 1 diabetes.However, the figures suggest that the risk of sudden death is considerably greater in thosewith diabetes. Although it is difficult to be precise, the risk in patients with diabetes seemsto be around three times higher.RISK FACTORS FOR SUDDEN DEATHIn the general population, the most frequent cause of sudden death is a cardiac arrhythmia,mostly related to coronary heart disease; it is likely that the same problem occurs in peoplewith diabetes. Nevertheless, if sudden death is occurring more frequently in patients, thenadditional factors are probably responsible. Some of this increase may be a consequence ofthe more advanced or premature ischaemic heart disease which is associated with diabetes.In some people the development of hypoglycaemic convulsions may impose an additionalinsult, although the fact that most subjects were found with their bedclothes undisturbedis against the pre-terminal development of tonic-clonic convulsions. This scenario does notexclude other forms of seizure activity and a striking similarity exists between the syndromesof sudden death in epilepsy and diabetes (Brown et al., 1990; Nashef and Brown, 1996).A study using an implantable ECG recorder in 20 patients with epilepsy identified threepatients with potentially fatal asystole, and permanent pacemakers were inserted in fourpatients (Rugg-Gunn et al., 2004).When considering other factors, the strongest candidates are probably coexisting autonomicneuropathy and hypoglycaemia.Autonomic NeuropathyAutonomic neuropathy increases the risk of sudden death in patients with diabetes in some(Ewing et al., 1980) but not all (Sampson et al., 1990) studies. The exact cause of deathremains uncertain although most groups that reported increased risk of death have suggestedthat a cardiac arrhythmia is responsible. Some groups have reported lengthened QT intervalsin patients with autonomic neuropathy (Ewing and Neilson, 1990) highlighting the associationbetween prolonged QT intervals and the risk of sudden death in other conditionssuch as the congenital long QT syndrome (Ewing et al., 1991). However, it seems unlikelythat autonomic neuropathy alone could account for the greater risk of sudden death inyoung patients with diabetes. In those who died suddenly, autonomic function had seldombeen formally tested. Some subjects had advanced diabetic complications and would probablyhave had some degree of autonomic neuropathy, but a significant proportion of those

RISK FACTORS FOR SUDDEN DEATH 273who died had a relatively short duration of diabetes with no evidence of microvasculardisease. However, other indicators of autonomic dysfunction, such as reduced heart ratevariability and diminished baroceptor sensitivity, are often present in patients with type 1diabetes even when formal cardiovascular tests of autonomic function are normal (Westonet al., 1996).HypoglycaemiaAll of the investigators who have reported sudden death in patients with type 1 diabeteshave implicated hypoglycaemia as a contributing factor. Death had occurred during thenight when hypoglycaemia is a common problem, and some had strict glycaemic controland had experienced nocturnal hypoglycaemia in the past (Box 12.4). The crucial questionis whether any mechanism exists through which hypoglycaemia could cause sudden death.Hypoglycaemia can cause irreversible brain damage but those who suffer this complicationrequire prolonged exposure to a low blood glucose and are unlikely to die suddenly. Previousauthors have emphasised the likelihood of an arrhythmic death and have pointed out that thedemonstration of a plausible mechanism by which hypoglycaemia caused cardiac arrhythmiaswould strongly implicate hypoglycaemia (Tattersall and Gale, 1993). How then doeshypoglycaemia affect electrical activity in the heart?The evidence concerning the effects of hypoglycaemia on the electrocardiogram (ECG)has been obtained from different sources. First, there is anecdotal clinical evidence wherearrhythmias have been detected during episodes of hypoglycaemia and resolved when bloodglucose recovered. Second, there are experimental data where the electrocardiogram hasbeen measured both in diabetic and non-diabetic subjects during controlled hypoglycaemiainduced in the laboratory.Hypoglycaemia causes an increase in autonomic neural activity affecting both parasympatheticand sympathetic nerves, an increase in plasma epinephrine, and a fall in potassium.Simple electrocardiographic techniques have demonstrated flattening or inversion of the Twave and some studies have also reported prolongation of the QT interval (Fisher and Frier,1993). A less consistent effect has been ST segment depression. In the presence of ischaemicheart disease it is not difficult to see how some of these changes might lead to malignantcardiac tachydysrhythmias and sudden death (see below). It is less easy to explain how evenBox 12.4Possible risk factors for ‘dead in bed syndrome’• Previous nocturnal hypoglycaemia• Living/sleeping alone• Intensive therapy• Multiple injections of insulin• Alcohol ingestion

274 MORTALITY, CARDIOVASCULAR MORBIDITY AND DIABETIC COMPLICATIONSthese profound, but brief, physiological changes could precipitate a fatal cardiac event inthose whose heart is otherwise healthy.Lengthening of the QT interval is associated with sudden death in other conditions, bothcongenital and acquired. The congenital long QT syndrome is an inherited disorder in whichmutations within the genes coding for the membrane ion channels which contribute to thecardiac action potential cause profound lengthening of the QT interval (Jervell and Lange-Neilsen, 1957; Jackman et al., 1988; Curran et al., 1995). Among the mutations that havebeen reported, one (within the SCN5A gene causing LQT3) leads to impaired sodium channelinactivation whereas others (LQT1-2, LQT5-6) produce abnormalities within potassiumchannels leading to decreased outward potassium current (Roden et al., 2002). Many ofthose people affected can be identified by marked QT lengthening on their surface ECG andare at high risk of developing fatal polymorphic ventricular tachycardia (Torsade de PointeVT). However, members of some families are at risk for VT despite an ECG that appearsnormal. Triggers of VT (drugs that block K + channels, exercise, sudden noise, sleep) mayvary according to the type of mutation. However, although the lifetime risk of death in thecongenital long QT syndrome is around 70% without treatment, patients can clearly survivewith a long QT interval for many years. At first sight it therefore seems unlikely that a shortperiod of hypoglycaemia lasting for an hour or so could cause a malignant arrhythmia. Yet,under other conditions, shorter periods of altered depolarisation can precipitate ventriculartachycardia.In addition to congenital QT lengthening, certain therapeutic agents (including antiarrhythmicagents, antibiotics, antihistamines) can cause an acquired long QT syndrome andsudden death by blocking movement of K + ions through one of the cardiac ion channels (therapid component of the delayed rectifier current, I Kr coded by the HERG gene) responsiblefor cardiac repolarization (Roden et al., 2002). It appears that interactions between geneticfactors (mutations and polymorphisms in the genes coding for proteins contributing to thecardiac action potential and its physiological regulation) and environmental effects (includingage, sex, state of sympathoadrenal activation and K + concentration) can determine whethera cardiac arrhythmia is triggered. One can therefore hypothesise that during clinical episodesof hypoglycaemia, alterations in cardiac repolarisation may be sufficient to precipitate a fatalperiod of ventricular tachycardia.It has been demonstrated that the QT interval can lengthen during experimental hypoglycaemiaboth in diabetic and non-diabetic subjects (Marques et al., 1997) (Figure 12.4).Some subjects had quite pronounced increments in QT interval and a strong relationshipwas observed between the increase in QT interval and the rise in plasma epinephrine. Betablockadeand potassium infusion both prevented QT lengthening during hypoglycaemia(Robinson et al., 2003). Epinephrine infusion in normal subjects also caused QT lengtheningthat was partially prevented by simultaneous potassium infusion (Lee et al., 2003). Similarchanges in QTc have subsequently been described during clinical episodes of nocturnalhypoglycaemia in adults and children with type 1 diabetes (Robinson et al., 2004; Murphyet al., 2004).The major problem with the hypothesis is that because sudden death occurs so rarely, itis very difficult to test directly. Isolated case reports of transient cardiac dysrhythmias havenot provided much additional useful information. There are reports of atrial fibrillation andsupraventricular tachycardia occurring during clinical episodes of hypoglycaemia but thereis no direct evidence that profound hypoglycaemia can cause a life threatening disturbancein cardiac rhythm.

RISK FACTORS FOR SUDDEN DEATH 275EuglycaemiaHypoglycaemiaFigure 12.4 Typical QT measurement with a screen cursor placement from one subject (a)during euglycaemia showing a clearly defined T wave, and (b) during hypoglycaemia showingprolonged repolarisation and prominent U wave. Horizontal: 700 ms epoch, vertical 1.33 mVfull scale. Reproduced from Marques et al. (1997) with permission from John Wiley & Sons,LtdAnother predisposing factor for an arrhythmia during hypoglycaemia might be animbalance of autonomic activation during hypoglycaemia in people with diabetes. Asmentioned above, hypoglycaemia leads to activation of the sympathetic and parasympatheticnervous system. Recent studies examining heart rate variability have describedstrong sympathetic cardiac activation with a concomitant increase in parasympathetictone in normal subjects during insulin-induced hypoglycaemia (Schachinger et al., 2004),although no change in heart rate variability was demonstrated in another similar study(Laitinen et al., 2003). A further study in patients with type 1 diabetes and control subjectsdemonstrated that more prolonged hypoglycaemia resulted in a reduction of cardiac vagaloutflow (Koivikko et al., 2005). Lee et al. (2004) appeared to refute the hypothesis thatautonomic neuropathy contributes to hypoglycaemia-induced QT lengthening by demonstratingthat those with autonomic neuropathy had the smallest increments in QT intervalinduced by experimental hypoglycaemia when compared to other diabetic groups. However,since these individuals also had the longest duration of diabetes, they unsurprisingly alsoexperienced the smallest rise in sympathoadrenal activation. In the absence of a studycomparing comparable levels of sympathoadrenal stimulation on cardiac electrophysiologicalresponses, the degree to which autonomic neuropathy contributes to this phenomenon remainsuncertain.Fatal events may be infrequent because the combination of factors that together lead toa fatal cardiac arrhythmia occur only rarely. Indeed severe hypoglycaemia, which causes aprofound sympathoadrenal discharge without alerting the patient or partner, is itself relativelyunusual. Thus, the ‘substrate’ for a lethal cardiac arrhythmia might be the combination ofa severe hypoglycaemic attack at night (which fails to wake people unless symptoms areintense) and a cardiac conduction system affected by sub-clinical autonomic neuropathy in

276 MORTALITY, CARDIOVASCULAR MORBIDITY AND DIABETIC COMPLICATIONSan individual who has inherited polymorphisms in the LQT genes causing exaggerated QTlengthening during sympathoadrenal activation.In summary, the majority of sudden deaths in young patients with diabetes remain unexplainedand we have hypothesised that these were deaths due to ventricular arrhythmia.These might have resulted from an increase in plasma epinephrine and a fall in potassiumwhich accompany hypoglycaemia, so producing prolongation of the QT interval. It ispossible that this occurs on a background of autonomic instability caused by early autonomicneuropathy.What Do We Say to Patients?It is clear that the frequency of deaths resulting from hypoglycaemia is underestimated,and some deaths related to hypoglycaemia may be attributed to other causes at the time ofcertification. Nevertheless, when we consider that nocturnal hypoglycaemia is very common,affecting up to 60% of patients every night, we can reassure patients that sudden deathas a consequence of acute hypoglycaemia is very rare. However, the available evidenceprevents us stating that there is no risk of death from hypoglycaemia during sleep, or atother vulnerable times when treatment is not rendered promptly.EFFECT OF HYPOGLYCAEMIA ON CARDIOVASCULARDISEASEAcute hypoglycaemia provokes an intense haemodynamic response secondary to activationof the autonomic nervous system with the secretion of epinephrine (adrenaline) (DeRosaand Cryer, 2004). The heart rate increases over a period of 15 to 20 minutes, but rarelyrises above 100 beats/minute. A modest but significant increase in systolic blood pressureis accompanied by a slight but significant fall in diastolic blood pressure (Fisheret al., 1987; Russell et al., 2001). The pulse pressure widens, with a substantial increasein cardiac output and a fall in total peripheral vascular resistance (Figure 12.5). Thesehaemodynamic changes are relatively short-lived, and exert no significant after-effects onthe 24-hour heart rate or blood pressure (Avogaro et al., 1994). In a person with a normalheart these haemodynamic changes are probably of no great significance, but in a patientwho has underlying coronary heart disease the profound increase in cardiac workloadmay provoke a cardiac arrhythmia, myocardial ischaemia and even myocardial infarction(Box 12.5).Arrhythmias and Coronary Heart DiseaseOccasional cardiac arrhythmias have been demonstrated in normal subjects during experimentalhypoglycaemia studies. It would now be considered unethical to perform hypoglycaemiastudies in patients with known heart disease, but many studies were performed in anearlier era both in diabetic and non-diabetic patients with coronary heart disease to examinethe effects of acute hypoglycaemia (Fisher and Frier, 1993). Sinus bradycardia has beenreported in a very small number of cases (Pollock et al., 1996; Navarro-Gutierrez et al.,

EFFECT OF HYPOGLYCAEMIA ON CARDIOVASCULAR DISEASE 277Figure 12.5 Mean response of (a) heart rate, (b) systolic blood pressure and mean arterial bloodpressure, and (c) left ventricular ejection fraction following intravenous injection of insulin at time 0.(R = autonomic reaction). Reproduced from Frier et al. (1987), with kind permission from SpringerScience and Business Media2003). Atrial fibrillation has been described in some patients and in addition there are severalcase reports of atrial fibrillation following hypoglycaemia in insulin-treated patients whohad no overt evidence of heart disease (Collier et al., 1987; Baxter et al., 1990; Odeh et al.,1990; Navarro-Gutierrez et al., 2003).There is a single report of a transient ventricular tachycardia occurring duringexperimental hypoglycaemia in a non-diabetic patient with coronary heart disease, andventricular tachycardia was recently documented in an elderly non-diabetic man whodeveloped hypoglycaemia during emergency surgery (Chelliah, 2000). There have been case

278 MORTALITY, CARDIOVASCULAR MORBIDITY AND DIABETIC COMPLICATIONSBox 12.5Cardiac effects of acute hypoglycaemia• Increased heart rate• Widening of pulse pressure• Arrhythmias• Silent myocardial ischaemia• Angina• Myocardial infarctionreports, with ECG evidence, of ventricular ectopics, sustained ventricular tachycardia,ventricular fibrillation and asystole during hypoglycaemia in diabetic patients (Shimadaet al., 1984; Burke and Kearney, 1999). Obviously this does not exclude the possibility ofthese arrhythmias occurring more frequently in clinical practice, as these arrhythmias willbe fatal if uncorrected. In most instances it is unlikely that any precipitating cause of thearrhythmia would be sought; hypoglycaemia may not have been recognised and we havealready alluded to the difficulties in establishing a putative diagnosis of hypoglycaemia atpost-mortem.Angina and Myocardial IschaemiaThe provocation of angina and myocardial ischaemia by exercise is well documentedin clinical practice. By contrast, acute hypoglycaemia, which provokes a more intensehaemodynamic response and in particular a greater increase in plasma epinephrine, hasrarely been documented as provoking anginal chest pain, either in the experimentalsituation or in anecdotal case reports. A literature search of over 6000 insulin tolerancetests recorded only two episodes of angina. This may reflect the fact that coronaryheart disease would be considered a contraindication to insulin tolerance testing,and in clinical practice clinicians may accept higher ambient blood glucose concentrationsin diabetic patients with known coronary heart disease to avoid hypoglycaemia.It is also possible that the haemodynamic changes of hypoglycaemia are soprofound that they are frequently fatal in patients with coronary heart disease, and hypoglycaemiais probably overlooked as a provoking cause when determining cause ofdeath.It is now well established that many episodes of ST segment depression on the ECG arenot associated with angina, and constitute ‘silent ischaemia’. One case has been describedduring 24-hour ECG monitoring of hypoglycaemia that provoked silent ischaemia in adiabetic patient with suspected coronary heart disease (Pladziewicz and Nesto, 1989). Morerecently, 72-hour continuous glucose monitoring with simultaneous cardiac Holter monitoringwas performed in 21 patients with coronary heart disease and insulin-treated type2 diabetes (Desouza et al., 2003). A total of 26 episodes of symptomatic hypoglycaemia