Efficacy and Safety of Ursodeoxycholic Acid Versus Cholestyramine ...

GASTROENTEROLOGY 2005;129:894–901Efficacy and Safety of Ursodeoxycholic Acid VersusCholestyramine in Intrahepatic Cholestasis of PregnancyJURATE KONDRACKIENE,* ULRICH BEUERS, ‡ and LIMAS KUPCINSKAS**Department of Gastroenterology, Kaunas University of Medicine, Kaunas, Lithuania; and ‡ Department of Medicine II, Klinikum Grosshadern,University of Munich, Munich, GermanyBackground & Aims: Treatment of intrahepatic cholestasisof pregnancy with ursodeoxycholic acid appearspromising, but data are limited so far. The aim of thisrandomized study was to evaluate the efficacy andsafety of ursodeoxycholic acid in comparison with cholestyramine.Methods: Eighty-four symptomatic patientswith intrahepatic cholestasis of pregnancy were randomizedto receive either ursodeoxycholic acid, 8–10mg/kg body weight daily (n 42), or cholestyramine,8 g daily (n 42), for 14 days. The primary end pointwas a reduction of pruritus by more than 50% after 14days of treatment as evaluated by a pruritus score.Secondary end points were outcome of pregnancy, reductionof serum aminotransferase activities and serumbile acid levels, and drug safety. Intention-to-treat analysiswas applied. Results: Pruritus was more effectivelyreduced by ursodeoxycholic acid than cholestyramine(66.6% vs 19.0%, respectively; P < .005). Babies weredelivered significantly closer to term by patients treatedwith ursodeoxycholic acid than those treated with cholestyramine(38.7 1.7 vs 37.4 1.5 weeks, respectively,P < .05). Serum alanine and aspartate aminotransferaseactivities were markedly reduced by 78.5%and 73.8%, respectively, after ursodeoxycholic acid, butby only 21.4%, each, after cholestyramine therapy (P

September 2005 UDCA VS CHOLESTYRAMINE IN ICP 895dexamethasone, epomediol, S-adenosyl-L-methionine,and cholestyramine have been used without clear evidenceof efficacy. Cholestyramine has been used for reducingpruritus. Observational studies suggest that cholestyraminemay be associated with improved maternalmorbidity without a documented improvement in fetaloutcome. 16–18 Cholestyramine may worsen the malabsorbtionof fat-soluble vitamins, especially vitamin K. Acase report of severe fetal intracranial hemorrhage duringtreatment with cholestyramine for ICP has raised thepossibility that severe maternal vitamin K deficiencymay lead to fetal vitamin K deficiency and coagulopathy.18 Recently, ursodeoxycholic acid (UDCA) has beenproposed for the treatment of ICP. Improvement inmaternal and fetal morbidity was suggested in 8 clinicaltrials and several observational studies, although thesestudies were small and in some aspects inconsistent. 16,17However, data from large, randomized trials of treatmentof ICP are lacking. Therefore, the aim of the presentstudy was to evaluate the efficacy and safety of UDCA inpatients with ICP in comparison with cholestyramine.UDCA and cholestyramine were administered at moderatedoses to alleviate the risk of formation of potentiallytoxic bile acid metabolites and of malabsorption of fatsolublevitamins, respectively.Materials and MethodsStudy DesignAn open, randomized, parallel group study was performedcomparing the efficacy and safety of UDCA and cholestyraminein intrahepatic cholestasis of pregnancy. The studywas approved by the ethics committee of the Kaunas MedicalUniversity. The study was conducted in accordance with theethical guidelines of the Declaration of Helsinki.PatientsPatients with a diagnosis of ICP as defined by thecriteria below were prospectively evaluated for inclusion in thisstudy between October 1999 and September 2002.Inclusion criteria were as follows: skin pruritus starting inthe second or third trimester of pregnancy and elevation of atleast one of the following biochemical parameters above theupper limit of normal: alanine aminotransferase (ALT) 45U/L, aspartate aminotransferase (AST) 40 U/L, and fastingserum bile acids 10 mol/L. Exclusion criteria were asfollows: chronic liver disease, hepatic viral infections (HAV,HBV, HCV, cytomegalovirus, Herpes simplex virus, Epstein-Barr virus), skin disease, allergic disease, and symptomaticcholelithiasis.All patients gave written informed consent before inclusioninto the study and were randomized to receive either UDCA,8–10 mg/kg body weight per day (Ursofalk capsules; Dr. Falk,Pharma GmbH, Freiburg, Germany), or cholestyramine, 8g/day (Ratiopharm GmbH & Co, Ulm, Germany), for 14 days.Randomization was performed by use of sealed envelopes.Study End PointsThe primary study end point was defined as a reductionof severity of pruritus by more than 50% after 14 days oftreatment. Secondary end points were outcome of pregnancy(term of delivery, Apgar score, newborn status), reduction ofserum aminotransferase activities and serum bile acids levels,and drug safety. Self-assessment of pruritus intensity wasperformed by the patients daily by use of a score from 0 to 4(0, no pruritus; 1, occasional; 2, intermittent pruritus everyday with asymptomatic periods prevailing; 3, intermittentpruritus every day with symptomatic periods prevailing; 4,constant pruritus day and night).Serum liver tests, fasting serum bile acids and cholic (CA),chenodeoxycholic (CDCA), deoxycholic (DCA), lithocholic(LCA), and ursodeoxycholic (UDCA) acids were evaluated atentry and on the day after end of therapy. Abdominal ultrasonographyand serologic tests (see inclusion and exclusioncriteria) were performed before treatment. Serum liver testswere determined using routine laboratory techniques. Fastingserum levels of bile acids were determined as described previously.19 In brief, bile acids were extracted with Bond-Elut C18cartridges (Analytichem International, San Diego, CA), solvolysiswas performed to cleave sulfate groups, and enzymatichydrolysis was performed to deconjugate bile acid amidates.Deconjugated bile acids were isolated by extraction on Lipidex1000 (Packard Instruments, Groningen, The Netherlands) andwere then methylated and trimethylsilylated for gas chromatography.Capillary gas chromatography was performed usinga Carlo Erba Fractovap 4160 gas-chromatograph (Carlo ErbaInstruments, Hofheim, Germany). Bile acid derivatives wereseparated on a fused silica capillary CP Sil 19 CB columncoated with chemically bonded OV-1701 (25 m 0.33 mm,Chrompack, Middelburg, The Netherlands). Hydrogen wasthe carrier gas (P .6 kg/cm 2 ). A temperature program from140°C to 270°C with 8°/min was started after on-columninjection. Eluting bile acid derivates were detected by a flameionization detector. Fasting serum samples were stored at20°C until analyzed.The fetal status was monitored in the same hospital everyweek. The outcome of pregnancy and the newborn status wereassessed by routine investigations documented by the obstetriciansand neonatologists. There was no specific protocol forassessing and managing the pregnancy of ICP patients participatingin this trial. The decision to induce labor and carry outcesarean sections was made by the managing obstetriciansindependently of this study. Data recorded during deliveryincluded term and mode of delivery, Apgar score at 1 and 5minutes, and newborn weight.Statistical AnalysisEfficacy and safety analyses were carried out on theintention-to-treat (ITT) population. The results are expressedas mean standard deviation (SD). Comparisons of paramet-

896 KONDRACKIENE ET AL GASTROENTEROLOGY Vol. 129, No. 3ric, normally distributed data were performed by Student ttest. Mann–Whitney U test was used for unpaired and Wilcoxonsigned-rank test for paired nonparametric data. The 2test was used to compare categorical variables. Comparison ofthe change of pruritus score every day of treatment was expressedas area under the curve (AUC). Statistical analysis wasconducted with STATISTICA 5.0. (Statsoft, Tulsa, OK) A P .05 was regarded as significant.ResultsEighty-four patients, aged 18–41 years, between25 and 39 weeks of gestation, who fulfilled the inclusioncriteria were enrolled in the present study. After randomization,42 patients received UDCA, 8–10 mg/kgbody weight/day, and 42 patients received cholestyramine,8 g/day, for 14 days. Median time of onset oftreatment was 35.0 weeks (range, 22.0–39.0 weeks);delivery was at 37–38 weeks. Therefore, most of thepatients were not offered continued medical treatmentuntil delivery when the study was finished and pruritushad improved. Only those patients with early onset ofpruritus (2 in the UDCA group and 2 in the cholestyraminegroup) were offered repeat treatment when pruritusexacerbated in the later course of pregnancy. Baselineclinical characteristics of both groups did not differ(Table 1). During 14 days of treatment, 10 patients inthe UDCA group and 4 in the cholestyramine groupwithdrew or violated the study protocol prior to studycompletion: (1) 4 patients receiving UDCA discontinuedtreatment during the study period; (2) 6 patients in theUDCA group apparently took UDCA before inclusion inthe trial as retrospectively disclosed by serum bile acidanalysis that revealed markedly elevated serum UDCAlevels (15 mol/L) in baseline samples; and (3) 4patients treated with cholestyramine withdrew mainlybecause of adverse events (nausea and vomiting). Intention-to-treat(ITT) analysis included all 84 patients (Figure1).Table 1. Baseline Characteristics of Treatment GroupsCharacteristicUDCA group(n 42)Cholestyraminegroup(n 42) P valueAge (y) 28.9 5.9 27.5 5.3 NSMultiparous 22 18 NSPositive family history 5 6 NSRecurrence 10 11 NSMultiple pregnancy 3 2 NSOnset of pruritus (wk) 31.7 3.1 31.0 3.8 NSOnset of treatment (wk) 34.3 3.1 33.8 2.8 NSWeight (kg) 69.2 5.8 70.8 7.1 NSNOTE. Data are presented as means SD.NS, not significant.The intensity of pruritus was similar in both treatmentgroups at baseline (2.88 0.40 vs 2.95 0.60)and improved after both treatments. However, analysis ofthe change in pruritus score between baseline and theend of the study revealed a significant difference betweentreatments. Relief of pruritus was observed after 3 to 4days of treatment with UDCA, whereas intensity ofpruritus usually diminished only after 7 to 10 daysduring treatment with cholestyramine (Figure 2). After 4days, the pruritus score was significantly lower in patientsreceiving UDCA than in those receiving cholestyramine(pruritus score: 2.08 0.63 vs 2.92 0.62,respectively; P .05), and the difference was even morepronounced after 14 days (pruritus score: 0.44 0.65 vs1.88 0.98, respectively; P .001) (Figure 2). Reductionof the pruritus score by more than 50% was observedin 67% (28 of 42) of patients treated with UDCAvs 19% (8 of 42) of patients treated with cholestyramine(P .0021), indicating higher efficacy of UDCA regardingrelief of pruritus.No stillbirths were observed, and no significant differencesof newborns’ weight were found in both groups.The Apgar score at 1 minute was similar in both groupsbut was significantly higher at 5 minutes in the UDCAgroup than in the cholestyramine group: Apgar score:9.4 0.5 vs 8.7 0.6, respectively; P .05. Inpatients receiving UDCA, delivery occurred significantlycloser to term than in patients who received cholestyramine(38.7 1.7 weeks vs 37.4 1.5 weeks, respectively;P .05). Postnatal development has been normalin all babies. Pregnancy ended prematurely in 3 (7%)patients receiving UDCA and in 5 (12%) patients treatedwith cholestyramine. Seven (16.7%) patients of theUDCA group underwent cesarean section because ofmultiple pregnancies (3 cases), placenta praevia (1 case),cephalo-pelvic disproportion (1 case), fetal distress (1case), and advanced maternal age (1 case), and 3 (7%)patients of the cholestyramine group underwent cesariansection because of fetal distress (1 case), twin pregnancy(1 case), and cephalo-pelvic disproportion (1 case).Baseline serum levels of ALT, AST, bilirubin, -glutamyltransferase(-GT), alkaline phosphatase (AP), andendogenous bile acids were similar in both groups.UDCA significantly reduced serum aminotransferase activitiesand endogenous bile acids serum levels but didnot affect serum bilirubin, -GT, and AP levels (Table2). In contrast, cholestyramine did not significantly affectserum aminotransferases and endogenous bile acids levels,whereas serum levels of bilirubin and AP significantlyincreased (Table 2).Comparison of treatment with UDCA and cholestyraminerevealed that ALT, AST, and endogenous serum

898 KONDRACKIENE ET AL GASTROENTEROLOGY Vol. 129, No. 3Table 2. Biochemical Parameters of Patients With ICP Before and After UDCA or Cholestyramine TreatmentUDCA group (n 42) Cholestyramine group (n 42)Biochemical parametersX SD P value X SD P valueALT (U/L)Before 194.0 155.4 .0001 189.1 115.5 NSAfter 78.2 57.4 222.4 128.0AST (U/L)Before 125.0 101.3 .0001 139.4 87.4 NSAfter 48.8 31.6 151.9 85.2AP (U/L)Before 364.9 124.3 NS 384.0 145.7 .05After 369.4 120.7 425.9 155.9-GT (U/L)Before 27.4 15.1 NS 24.1 14.2 NSAfter 25.1 11.0 24.4 15.1Endogenous bile acids (mol/L)Before 46.5 41.3 .01 38.8 38.5 NSAfter 24.4 29.2 26.2 23.3Bilirubin (mol/L)Before 17.2 15.2 NS 13.3 7.4 .05After 13.2 8.5 15.9 10.0NOTE. Data are presented as means SD.induced apoptosis. 20–24 In ICP, hepatocellular cholestasisand retention of endogenous bile acids and sex hormonemetabolites are predominant features, suggesting thatimprovement of hepatocellular secretion may be a keymechanism of action of UDCA in ICP. Effects of UDCAin ICP resemble the effects observed in other cholestaticdiseases, although, in ICP, the clinical and biochemicaleffects are obtained faster, and they fade quickly whenthe drug is discontinued. 25,26Recent studies have confirmed that patients with ICPhave evident changes in the metabolism of bile acids andsex hormones. The cholestatic potential of some D-ringestrogens, in particular glucuronides such as estradiol-17-D-glucuronide, and mono- or disulfated progesteronemetabolites, mainly 3, 5-isomers, is supported byexperimental and clinical data. 1,8,9 The mechanism bywhich metabolites of sex hormones may induce cholestasisin ICP is still under debate, and mutations in genescoding for biliary transport proteins of physiologicallyoccurring metabolites in pregnancy as well as abnormalmetabolites inhibiting hepatobiliary transport proteinshave been considered to contribute to ICP. In ICP, therelief of cholestasis by UDCA has been suggested to bedue to stimulation of vesicular exocytosis resulting inmobilization of an increased number of transport proteinsto the canalicular membrane and, thereby, stimulationof transport systems involved in the biliary secretionof steroid mono- and disulfates. 1,3,4,21–24,26,27Following oral administration, approximately 30%to 60% of UDCA is absorbed in the gut. 28 The degreeof UDCA enrichment in biliary bile following chronicFigure 3. Serum bile acids before and after (A) UDCA therapy and (B)cholestyramine therapy. Data are means SD; *P .05, **P .0001.

September 2005 UDCA VS CHOLESTYRAMINE IN ICP 899ingestion correlates with the administered dose. Mostclinical trials have used UDCA at a dose of 13–15mg/kg/day. 29–32 Recently, Mazzella et al reported positiveresults and no adverse reactions of high-doseUDCA (1.5–2 g/day). 33,34 Little is known about theefficacy of a moderate dose of UDCA (8–10 mg/kgday) in patients with ICP. We aimed to determine theefficacy of this moderate dose to limit formation of themonohydroxy bile acid LCA, a potentially toxic bacterialproduct of UDCA in the human colon. UDCAgiven at 8–10 mg/kg per day causes an enrichment ofapproximately 40% in biliary bile acids. 35 The resultsof our study demonstrate a beneficial effect of a moderatedose of UDCA on relief of pruritus when comparedwith cholestyramine (50% relief in 67% vs19% of those treated with cholestyramine). Thepresent study confirmed rapid relief of pruritus inpatients receiving UDCA, and most of them experienceda clear relief of pruritus after 3–4 days, whereasduring treatment with cholestyramine, pruritus usuallyattenuated only after 7–10 days. In agreementwith the literature, elevation of aminotransferase activitiesfrom 2- to 15-fold was noticed in 85% ofpatients, bilirubin from 2- to 4-fold in 14% of patients,and fasting serum bile acids from 1.5- to20-fold in 78% of patients. 3,36 Sensitive serum markersfor cholestasis such as -GT and AP were usuallynormal or slightly elevated: -GT was elevated up to3-fold in 11% of patients, and AP was elevated 2- to3-fold in 60% of patients. When -GT levels are high,a mutation of the gene encoding the canalicular phosphatidylcholinetranslocase (multidrug resistance gene3; MDR3) is suspected. 13–15,37 Aminotransferase activitiesand fasting serum bile acids levels were significantlyreduced after treatment with UDCA, whereascholestyraminedid not affect these parameters. Serum bilirubinand -GT levels were normal in 86% and 89% ofpatients, respectively, before treatment and were notsignificantly affected by UDCA or cholestyramine.Serum alkaline phosphatase activity, which is mainlyof placental origin in the third trimester of pregnancy,was not altered by medical treatment.Our results also confirm earlier findings demonstratingthat an increase of serum bile acids and especiallyof CA appears to be a sensitive indicator ofICP. 1,3,8,38–44 The measurement of serum bile acids isparticularly helpful in patients with pruritus but normaltransaminase activities. CA predominated in thespectrum of serum bile acids during ICP when comparedwith healthy pregnant women. In a prospectivecohort study from Sweden, Glantz et al demonstrateda correlation between fetal complications and serumbile acids levels. 44 Elevated CA levels in maternalserum have been found to correlate with strongeruterine muscle contractions and placental chorionicvein vasoconstriction, which may cause fetal distress.41–46 A significant reduction of the concentrationsof CA and CDCA was detected after treatmentwith UDCA, whereas, under treatment with cholestyramine,the serum concentrations of bile acidswere not significantly changed, with the exception ofCDCA. One possible explanation for this decrease ofCDCA could be the binding of bile acids by cholestyramineand the interruption of their enterohepatingcirculation. 18 In the colon, unabsorbed UDCA is partiallyconverted by intestinal bacteria into LCA,known to be embryotoxic in rats. 1,41 Our data haveshown that LCA is not significantly increased duringthe treatment with UDCA at moderate doses. Thedecrease of CA may be clinically relevant because ithas been reported to cause fetal distress. 41–46 No stillbirthsor significant differences of newborn weightwere found in both groups, whereas the Apgar score at5 minutes was significantly higher in the UDCAgroup, and the delivery occurred significantly closer toterm than in patients who received cholestyramine.The present data confirm an excellent safety profile ofUDCA: No adverse effects were observed. By contrast,12 patients (29%) did not tolerate cholestyramine becauseof nausea and vomiting. Both ICP and cholestyraminetreatment may independently lead to vitamin Kdeficiency. Therefore, prolonged high-dose cholestyraminetreatment of pruritus in ICP can increase the risk ofcoagulopathy. 18 However, no cases of coagulopathy werenoticed in our study. This might be due to the shortduration of treatment.In conclusion, the results of our study showed asignificant improvement of pruritus severity, aminotransferaseactivities, and serum bile acid concentrationsand a more favorable outcome of pregnancy andabsence of adverse events after treatment with a moderatedose of UDCA. In contrast, cholestyramine alleviatedpruritus only mildly and caused adverse effects.These promising results obtained in the present andprevious studies confirm the use of UDCA as first-linetherapy for ICP.References1. Reyes H, Sjovall J. Bile acids and progesterone metabolites inintrahepatic cholestasis of pregnancy. Ann Med 2000;32:94–106.2. Reyes H. Review: intrahepatic cholestasis. A puzzling disorder ofpregnancy. Gastoenterol Hepatol 1997;12:211–216.