Sustained Elevation of Circulating Growth and Differentiation Factor ...

Clinical InvestigationsAcute muscle loss in critically ill patients causes weaknessranging from mild loss of strength and musclewasting to profound muscle weakness. Severe weaknessin this context is known as ICU-acquired paresis (ICUAP),which is associated with significant mortality and morbidity(1). Although it is a common problem, affecting at least 25% ofpatients admitted to critical care (2, 3), our understanding ofthe underlying pathological mechanisms is limited. Sepsis, systemicinflammatory response syndrome, immobility, and hyperglycemiaare all thought to be major etiological risk factorsfor the development of ICUAP (1, 4–7). However, elucidationof the molecular mechanisms involved in human cases is difficult,and evidence has often been contradictory. In part, thisis because data are often obtained on ICU from heterogenousgroups of patients with ICUAP with a range of reasons for ICUadmission and varying lengths of stay.As in other disease processes, muscle mass in critically ill patientsis determined by the balance between dynamic processesmediating muscle breakdown (atrophy, apoptosis, and autophagy)and synthesis (protein synthesis and satellite cell recruitment).Myostatin, a member of the transforming growth factorβ(TGF-β) superfamily, is a well-recognized negative regulatorof muscle mass (8), which has been implicated in the developmentof ICUAP. Elevation of both myostatin messenger RNA(mRNA) and protein was observed in the muscle of critically illpatients (9) and increased circulating levels of myostatin wereobserved in patients with muscle wasting associated with chronicobstructive pulmonary disease (10), cardiac failure (11), andliver failure (12). Other members of the TGF-β family also regulatemuscle phenotype. Growth and differentiation factor-15(GDF-15) is a member of this family that, under normal conditions,is not expressed highly in most tissues (13). However,expression is increased in response to oxidative stress, hypoxia,inflammation, and tissue damage in liver, brain, lung, and vasculartissue (13, 14). These factors are all thought to be importantin the development of ICUAP (7). GDF-15 is increasinglyimplicated in several disease processes, including cardiovasculardisease (14), multiple cancers (15–17), pulmonary artery hypertension(18), and importantly cachexia (19). Furthermore, in amouse model of cardiac hypertrophy, GDF-15 transgenic micewere resistant to cardiac hypertrophy and exposure to circulatingGDF-15 limited cardiac myocyte hypertrophy (20).Insulin-like growth factor-1 (IGF-1) drives many hypertrophypathways. Suppression of IGF-1 was observed in variousanimal models of sepsis and immobility (7) and in patients withsevere ICUAP (21). Although both increased myostatin and decreasedIGF-1 have been implicated in the development of ICU-AP, the relative pattern of change in the circulating levels of thesemediators during the acute development of muscle wasting hasnot been established. Therefore, to identify potential circulatingfactors driving acute muscle wasting, we studied a homogenouscohort of patients undergoing high-risk elective cardiothoracicsurgery who were at risk of prolonged critical care admissionand ICUAP. We hypothesized that the acute insult of surgerywould result in an acute rise in the known circulating drivers ofmuscle breakdown (myostatin) and suppression of those promotingmuscle hypertrophy (IGF-1) and that the dynamic patternof these changes would differ between those patients whodeveloped muscle wasting and those who did not. Furthermore,although a direct effect of GDF-15 on skeletal muscle has notbeen reported previously, the observations discussed above ledus to hypothesize that circulating GDF-15 would rise followingthe acute stress of cardiothoracic surgery and be associated withmuscle wasting. We went on to test our hypothesis further, byexposing cultured myotubes to GDF-15 in vitro.METHODSPatient Recruitment and Study DesignWritten consent was obtained from all those involved in thisprospective observational study (local research ethics committeeapproval 10/H0504/9). Adults undergoing elective cardiac surgeryat the Royal Brompton Hospital were screened for inclusionin the study. The principle inclusion criteria were a high-riskelective procedure requiring postoperative admission to adultcritical care, defined by the surgical team and by the patients’EuroSCORE (22–24). Exclusion criteria included pre-existingmuscle or neuromuscular disease, malignancy, or contraindicationto serial blood sampling. Patients were only included if theywere stable preoperatively and not suffering from acute illness orrequiring emergency surgery. Muscle mass was assessed by measuringthe cross-sectional area of the right rectus femoris muscleby ultrasound (see below) preoperatively and at day 7 of admissionor at discharge from hospital if earlier than day 7. We obtainedblood for analysis preoperatively (“baseline”), and repeatsamples were taken on the first and second postoperative daysand at day 7 or at discharge if earlier; these latter two time pointswere considered identical for the purposes of analysis: only fourpatients (three in the nonwasting patient group and one in thewasting group) were in fact discharged earlier than day 7.Measurement of Cross-Sectional Area of the RectusFemoris by UltrasoundUltrasound cross-sectional area of the rectus femoris (US RF csa)was measured using a previously described technique (25, 26).B-mode ultrasound imaging with a 10-MHz 12L-RS probe wasused (Logiq E, GE Healthcare, Buckinghamshire, UK). The patientwas positioned supine on the bed with their legs flat withleg muscles relaxed. The anterior superior iliac spine (ASIS) andthe point 60% of the distance from the ASIS to the superiorborder of the patella were identified. The ultrasound probe waspositioned perpendicular to the axis of the leg. Where the entireRF csacould not be visualized in a single ultrasound image a point66% or 80% of the distance between the ASIS and patella wasused. The same point was used for follow-up measurements foreach patient. Three separate consecutive images were taken andthe RF csaestimated using Image-J software (National Institutesof Health). The average of these three measurements was used.Inadequate ultrasound images were defined as those where theedges of the RF muscle could not be defined well enough to calculatethe RF csa. This occurred where the patient was very edematousand in these cases the patient data were excluded (n = 3).Critical Care Medicine www.ccmjournal.org 983