The potent bile acid sequestrant colesevelam is not effective in cholestatic pruritus: Results of a double-blind, randomized, placebo-controlled trial


  • Potential conflict of interest: Nothing to report.

  • The identifier is NCT00756171.


Colesevelam is an anion-exchange resin with a 7-fold higher bile acid–binding capacity and fewer side effects than cholestyramine, the current first-line treatment option for cholestatic pruritus. The aim of this trial was to compare the effects of colesevelam and a placebo in patients with cholestatic pruritus. In a randomized, double-blind, investigator-initiated, multicenter trial, patients with cholestatic pruritus, both treatment-naive and previously treated, received 1875 mg of colesevelam or an identical placebo twice daily for 3 weeks. The effect on pruritus was assessed with daily visual analogue scales, quality-of-life scores, and evaluations of cutaneous scratch lesions. The predefined primary endpoint was the proportion of patients with at least a 40% reduction in pruritus visual analogue scale scores. Thirty-eight patients were included, and 35 were evaluable: 17 took colesevelam, 18 took the placebo, 22 were female, 8 were treatment-naive, 14 had primary biliary cirrhosis, and 14 had primary sclerosing cholangitis. The mean serum bile acid levels were comparable between the groups before treatment (P = 0.74), but they were significantly different after treatment (P = 0.01) in favor of patients treated with colesevelam. Thirty-six percent of patients in the colesevelam group reached the primary endpoint versus 35% in the placebo group (P = 1.0). There were no significant differences between the groups with respect to pruritus scores, quality-of-life scores, and severity of cutaneous scratch lesions. Mild side effects occurred in one colesevelam-treated patient and four placebo-treated patients. Conclusion: Although colesevelam significantly decreased serum bile acid levels, this trial was unable to demonstrate that it was more effective than a placebo in alleviating the severity of pruritus of cholestasis. (HEPATOLOGY 2010)

Pruritis is a frequent and debilitating symptom of cholestatic liver disease.1 Although the pathophysiology of pruritus secondary to cholestasis remains largely unknown, it is widely assumed that bile acids are etiologically involved.2, 3 The principal pharmacological treatment options currently available and recommended in recent guidelines4 are cholestyramine5, 6 (a nonabsorbable, bile acid–binding resin), rifampicin,7, 8 naltrexone,9, 10 and sertraline.11 However, the efficacy of these drugs is variable, and side effects are common. Cholestyramine frequently causes constipation and nausea, rifampicin is known for its potential hepatotoxicity, and patients on naltrexone may experience symptoms of endogenous opioid-withdrawal syndrome. Therefore, the treatment of cholestatic pruritus is currently often problematic and unsatisfactory, and alternative treatment options are warranted.

Colesevelam (Cholestagel) is a bile acid sequestrant taken in tablet form that hydrates to a gel and is being used for the treatment of hypercholesterolemia. This agent differs from other sequestrants in that the hydrophilic polymer backbone has abundant hydrophobic side chains facilitating the binding of bile acids. Also, it is better tolerated than other bile acid sequestrants such as cholestyramine and colestipol12, 13: specifically, no clinically significant difference has been shown between a placebo and colesevelam with respect to gastrointestinal complaints such as flatulence, constipation, diarrhea, nausea, and dyspepsia.14 Until now, the efficacy of colesevelam in ameliorating the pruritus of cholestasis was explored only in a small open study,15 which showed that colesevelam was effective in ameliorating pruritus in five of eight patients. The aim of the current trial was to assess the effect of colesevelam on cholestatic pruritus in a double-blind, randomized, placebo-controlled trial.


ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; LLN, lower limit of normal; MELD, Model for End-Stage Liver Disease; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; SD, standard deviation; UDCA, ursodeoxycholic acid; ULN, upper limit of normal; VAS, visual analogue scale.

Patients and Methods

This study was a double-blind, randomized, placebo-controlled, multicenter trial. The aim was to assess the efficacy of colesevelam versus a placebo in cholestatic pruritus. Inclusion started in September 2008, and the follow-up was completed in October 2009. With approval of their medical ethical committees, three Dutch university hospitals participated in this trial: Erasmus University Medical Centre Rotterdam, University Medical Centre Utrecht, and University Medical Centre Amsterdam. The trial was registered at (NCT00756171) and conducted, recorded, and reported in compliance with International Conference on Harmonisation Good Clinical Practice and national regulations. All consecutive patients with cholestatic pruritus, both treatment-naive and previously treated, who visited our outpatient clinic were asked to participate. Inclusion criteria were the presence of pruritus of any severity as a result of any cholestatic disorder, age >18 years, and informed consent. Exclusion criteria were the use of cholestyramine within 3 weeks before the evaluation, pregnancy, no ability to understand or speak the Dutch language, malignancy, or a life expectancy less than 6 months.

Patients received either Cholestagel (colesevelam hydrochloride), a hydrophilic, water-insoluble, nonabsorbable agent (taken orally as three 625-mg tablets twice daily), or placebo tablets of identical shape and taste. The total individual treatment period was 21 days. Randomization was centralized with opaque, serial-numbered envelopes prepared by the trial statistician. Participants were assigned to one of the two arms according to a standard randomization schedule (1:1) in blocks of four and were stratified per trial center (Erasmus University Medical Centre Rotterdam, University Medical Centre Amsterdam, and University Medical Centre Utrecht). Both participants and investigators were blinded.

Participants visited the outpatient clinic three times: the first time for screening, the second time before the start of treatment, and the third time at the end of treatment. After 1 week of treatment, they were contacted by telephone to address potential problems and questions and to assess potential side effects. Patients were further instructed to call the principal investigator when necessary.

The intensity of symptoms was quantified by means of visual analogue scales (VASs) in the morning and in the evening. The pruritus VAS score varied from score 0 (no pruritus) to 10 (severe pruritus). Fatigue and quality of sleep were evaluated with comparable VAS scores. Patients were asked to begin completing the pruritus VAS scores 7 days before the actual start of treatment (day 0). The severity of pruritus was further quantified by open categorized questions and by descriptions of the nature and extent of cutaneous scratch lesions, which were classified as follows: (1) excoriations (mild), (2) plaques (moderate), (3) nodules (severe), or (4) indifferent scars. Additionally, with identical cameras with the same settings, photographs were taken from the front and back of the extremities and the trunk before treatment and at the end of treatment. The severity of skin lesions was scored on the basis of these photographs by an experienced dermatologist (H.B.T.) who was blinded with respect to the treatment allocation and sequence of photographs. The participants completed two validated quality-of-life scores (Short Form 36 and Liver Disease Symptom Index 2.0) before and after treatment.16, 17 Finally, laboratory investigations, including measurements of the total serum bile acid, total bilirubin, albumin, alkaline phosphatase, aspartate aminotransferase, and alanine aminotransferase levels, were performed before and at the end of treatment. Possible adverse events were assessed after 7 and 21 days of treatment.

When patients were using ursodeoxycholic acid, this treatment was continued. To prevent interference with the absorption of ursodeoxycholic acid and other drugs including levothyroxine, glyburide, and oral contraceptives, a minimal interval of 4 hours between the intake of the study medication and these drugs was advised. Antipruritic drugs other than the study medication were stopped with a washout period of at least 3 weeks. However, patients were allowed to continue rifampicin and naltrexone at a stable dose if they felt these agents were of (some) benefit. No other antipruritic drug was allowed during the trial when patients experienced worsening of pruritus.

Definitions and Statistical Methods.

The predefined primary endpoint of this study was the proportion of patients per group with at least a 40% reduction of pruritus based on VAS scores; this was based on a comparison of the mean scores on days 18, 19, and 20 with the mean scores on days −2, −1, and 0. Secondary endpoints were an improvement in quality-of-life scores and a reduction in the severity of scratch lesions after 21 days in comparison with pretreatment scores.

The power calculation (Fisher's exact test) was based on the primary outcome of a 40% reduction in the severity of pruritus. On the basis of a study of Mayo et al.,11 we expected this outcome in at least 50% of patients treated with colesevelam and in 5% of those treated with placebo. To detect this difference with a significance level of 0.05 and a power of 80% in a two-tailed test, 17 participants had to be included in each treatment arm. Considering a 10% dropout rate, we determined that the total number to be included should be 38 patients. An interim analysis was not performed. Both a modified intention-to-treat analysis and a per protocol analysis were performed. Statistical differences were evaluated for the two groups by both parametric and nonparametric tests. A P value <0.05 (two-tailed) was considered statistically significant. SPSS 15.0 was used to perform analyses.



We included and randomized 38 patients, 35 of whom were analyzed because one patient withdrew from participation after randomization and before the start of treatment, one patient stopped the intake of naltrexone during the trial period, and one patient was unable to fill out the questionnaires. Both a modified intention-to-treat analysis (n = 36) and a per protocol analysis (n = 35) were performed, and the results were concordant. For matters of clarity, we decided to describe the 35 patients randomized and treated according to protocol.

Of the remaining 35 participants, 17 were treated with colesevelam, and 18 were treated with a placebo. Eight patients were treatment-naive, whereas 27 patients had already been treated with one or more antipruritic drugs. Symptoms had been present for a median period of 24 months (range = 1-360 months). All 35 participants completed the trial. The collection of study data, which included the questionnaires, VAS scores, and laboratory studies, was complete.

Both groups were comparable with respect to age, baseline biochemistries, and use of ursodeoxycholic acid (Table 1). With respect to etiology, however, the majority of patients with primary biliary cirrhosis were assigned to the placebo group (10/14). Conversely, the majority of patients with primary sclerosing cholangitis were assigned to the colesevelam group (10/14). Because primary biliary cirrhosis is a disease mostly affecting females, whereas primary sclerosing cholangitis predominantly affects males, this distribution explains the observed difference in the male/female ratio between the two groups. Other etiologies of cholestatic pruritus were alcoholic cirrhosis, cirrhotic hepatitis C, biliary atresia, sarcoidosis hepatis, and adenosine triphosphate–binding cassette B4 (multidrug resistance protein 3) deficiency in one case each. In two cases, the etiology of the liver disease was cryptogenic. No patient reported an atopic constitution.

Table 1. Baseline Characteristics
CharacteristicColesevelam (n = 17)Placebo (n = 18)
  • Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; LLN, lower limit of normal; MELD, Model for End-Stage Liver Disease; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; SD, standard deviation; UDCA, ursodeoxycholic acid; ULN, upper limit of normal.

  • *

    Normal level <10 μmol/L.

Sex, male/female8/95/13
Age, years [mean (SD)]50 (13)54 (13)
Etiology, PBC/PSC/other (n)4/10/310/4/4
Bilirubin [ULN (SD)]2.1 (2.8)2.0 (2.4)
ALP [ULN (SD)]2.5 (2.3)1.9 (1.2)
Albumin [LLN (SD)]1.1 (0.2)1.1 (0.2)
AST [ULN (SD)]2.2 (1.7)2.2 (1.5)
ALT [ULN (SD)]2.0 (1.5)1.6 (1.0)
MELD score [mean (SD)]7.2 (5.5)6.8 (4.8)
Concurrent antipruritic treatment, rifampicin/naltrexone/sertraline (n)0/2/00/2/0
Patients treated with UDCA (n)1615
UDCA, mg/kg/day [mean (SD)]14.9 (6.0)16.2 (4.0)
Total serum bile acids, μmol/L [mean (range)]*140 (32-440)155 (9-550)

At entry, all participants graded the severity of pruritus as severe. In most patients (89%), pruritus was most severe in the evening and/or at night. Scratch lesions of any type or severity were present in 55% of cases. These lesions were found primarily on the extremities and the back. On average, 10% to 30% of the total body area showed abnormalities secondary to scratching. Excoriations were seen in 34% of cases, plaques were seen in 37%, noduli were seen in 26%, and scars were seen in 23%.

The use of other antipruritic medications was equal in the two treatment arms, with two patients in each group using naltrexone with incomplete effects. Side effects (no more than mild stool changes) were reported by four patients in the placebo group and by one patient in the colesevelam group. Dose reduction was not necessary; all patients continued the treatment during the 3-week period. The reported trial medication intake was 100%.

Primary Outcome.

For the primary outcome, the proportion of patients with at least a 40% reduction of the pruritus VAS score after treatment, there was no significant difference between the colesevelam and placebo groups. In the colesevelam group, 36% of patients reached the defined 40% reduction of the pruritus VAS score in the morning versus 35% in the placebo group (P = 1.0). With respect to the pruritus VAS score in the evening, a 40% reduction was noted in 40% and 50% of colesevelam-treated and placebo-treated patients, respectively (P = 0.74).

According to an open categorized question, 100% of participants experienced severe pruritus before treatment. At the end of the treatment period, 76% of the colesevelam-treated patients and 72% of the placebo-treated patients reported severe pruritus.

With respect to the quality of sleep and fatigue VAS scores, no statistically significant differences were found.

Serum Bile Acid and Bilirubin Levels Before and After Treatment.

The median total serum bile acid levels at the baseline were 140 and 155 μmol/L for the colesevelam and placebo groups, respectively (P = 0.74). During treatment, levels decreased significantly in the colesevelam group to 73 μmol/L (P = 0.003), whereas levels tended to increase to 212 μmol/L in the placebo group (P = 0.67; Fig. 1). After treatment, the serum bile acid level was significantly lower in the colesevelam group versus the placebo group (P = 0.01).

Figure 1.

Individual serum bile acid levels before and after treatment with colesevelam (left) and placebo (right). The solid, black line is the interpolation line.

Figure 2 shows the relation between changes in morning pruritus scores and changes in serum bile acid levels. In the majority of patients, pruritus scores decreased, and this was associated with slightly increased mean serum bile acid levels in the placebo group and reduced levels in the colesevelam group. There was no significant correlation between these changes in either group (Spearman test). Bilirubin levels were comparable for placebo-treated patients (1.1 times the upper limit of normal) and colesevelam-treated patients (1.8 times the upper limit of normal), both before (P = 0.96) and after (P = 0.27) treatment. Also, serum levels of alkaline phosphatase and aminotransferases remained unchanged and were comparable for both groups.

Figure 2.

Absolute changes [Spearman correlation for the placebo, P = 0.43 (r = −0.21); Spearman correlation for colesevelam, P = 0.41 (r = −0.23)] and proportional changes [Spearman correlation for the placebo, P = 0.42 (r = −0.22); Spearman correlation for colesevelam, P = 0.26 (r = −0.31)] in morning pruritus VAS scores and in serum bile acid levels for individuals treated with colesevelam and placebo (day 0 versus day 21).

Individual VAS Scores.

The individual pruritus VAS scores during the study are shown in Fig. 3. The positive change in the mean morning pruritus VAS scores during the study period was statistically significant for the colesevelam group (P = 0.01) but not for the placebo group (P = 0.37). However, the difference between the colesevelam and placebo groups was not statistically different (P = 0.18).

Figure 3.

Individual pruritus VAS scores: (A) morning (interpolation line for the placebo, P = 0.37; interpolation line for colesevelam, P = 0.01; colesevelam versus placebo, P = 0.18) and (B) evening (interpolation line for the placebo, P = 0.03; interpolation line for colesevelam, P = 0.01; colesevelam versus placebo, P = 0.70).

The mean evening VAS scores decreased significantly in both the colesevelam group (P = 0.01) and the placebo group (P = 0.03); however, there was no statistical difference between the two groups (P = 0.70). As shown in Fig. 4, in the majority of patients, the pruritus VAS score decreased. Overall, the magnitude of a positive response (a decrease in the pruritus VAS score) was comparable for patients treated with colesevelam and patients treated with placebo.

Figure 4.

Individual proportional differences in VAS scores for pruritus in the morning and in the evening. The gray bars represent participants treated with colesevelam (n = 17), and the black bars represent participants treated with placebo (n = 18).

Dermatological Assessment.

The dermatological assessment, based on photographs taken before and after treatment, showed no significant differences between the two groups at entry with respect to the nature and severity of scratch lesions (P = 1.0). Also, no significant changes were noted within or between the groups at the end of the trial (P = 0.35).

Quality-of-Life Scores.

No statistically significant changes were found with respect to the domains of physical functioning (P = 0.67), role physical functioning (P = 0.50), bodily pain (P = 1.00), general health (P = 0.48), vitality (P = 0.90), social functioning (P = 0.37), emotional functioning (P = 0.17), and mental health (P = 0.26) of the Short Form 36 questionnaire in the colesevelam group before and after treatment. Both treatment groups were comparable before and after treatment. Liver Disease Symptom Index 2.0 revealed no significant differences either.


This trial shows that colesevelam significantly lowers serum bile acid levels in cholestatic individuals with pruritus but is not more effective than a placebo in alleviating pruritus. Although patients treated with colesevelam experienced less pruritus, this was also the case for patients treated with a placebo, and there was no difference in the number of participants with a significant response or in the magnitude of symptom reduction between the two groups. We were also unable to demonstrate a clear relation between a decrease in serum bile acid levels and a reduction in pruritus. Because the majority of the patients in the study population did not respond or responded insufficiently to other treatments, our results do not address the efficacy of colesevelam as a first-line therapy.

This is the first randomized controlled trial to assess the efficacy of colesevelam in cholestatic pruritus and the second trial to evaluate bile acid sequestrants for this indication. In an earlier open study with eight participants, which was published only in abstract form, colesevelam had a beneficial effect on cholestatic pruritus in five of eight patients, whereas side effects were absent.15 The present study only confirms the observations of that study and other studies: colesevelam is well tolerated and seems free of major adverse treatment effects.12, 14

There are several potential explanations for the observed lack of efficacy of colesevelam. First, this drug may not interfere with the mechanisms eventually resulting in the perception of pruritus. Because serum bile acid levels decreased in all but one person treated with colesevelam, the present study may suggest that (intestinal) bile acids are not of key importance in the pathogenesis of pruritus.2, 18 During colesevelam treatment, serum bile acid levels decreased by nearly 50%. However, in the majority of cases, these levels remained markedly elevated. Therefore, an alternative explanation for the negative results of the trial may be that the decrease in serum bile acid levels was not enough to have an impact on the severity of pruritus. The negative trial result could also be related to insufficient compliance. However, all participants were highly motivated to participate in this study, and the reported intake of the study medication was 100%. The observed and expected effect of colesevelam (but not the placebo) on serum bile acid levels also indicates that noncompliance was highly unlikely. The trial may also have been too small to detect beneficial treatment effects. However, the basis of the power size calculation seems reasonable, and we were able to recruit and analyze the required number of patients. Our inability to document a beneficial effect of colesevelam could also be related to the selected endpoints (particularly the percentage of responders with at least a 40% decrease in the severity of pruritus over a 3-week period). The decrease in pruritus scores was relatively low versus the reported response rate of approximately 60% for bile acid sequestrants.6, 15 However, our analysis, as illustrated in Fig. 4, shows that choosing other cutoff levels would not have changed the results in any way. The trial may also have been too short in order to detect an effect on pruritus. We chose a 3-week treatment duration because we felt uncomfortable about withholding treatment from patients for more than 6 weeks (the time for washing out cholestyramine and for the trial). Moreover, a longer treatment, possibly with a placebo, was expected to be a major obstacle to participation and thus would have had a negative impact on recruitment. Also, nasobiliary drainage has an immediate effect on pruritus,19, 21 and this suggests that an effect of colesevelam on the enterohepatic circulation of an undefined pruritogen would at least have become apparent within 3 weeks. The majority of our patients experienced severe pruritus for which they had already been treated with agents such as cholestyramine, naltrexone, and rifampicin. The majority of the included patients possibly may have suffered from truly refractory pruritus, and the results may have been different in populations with other characteristics, such as patients with less severe pruritus or previously untreated patients. We cannot completely exclude this possibility.

It is intriguing that the results of the present trial are negative, whereas cholestyramine, a drug with a markedly weaker bile acid–binding capacity, is generally believed to be an effective drug and is usually prescribed as a first treatment option. In theory, cholestyramine could have effects other than those of colesevelam, such as more potent binding of intestinal or biliary constituents undergoing an enterohepatic circulation (other than bile acids) involved in the pathogenesis of pruritus. Currently, there are no data to support or refute this possibility. The rationale for the clinical use of cholestyramine is mainly based on empirical experience. The only double-blind, randomized trial with cholestyramine was reported by Di Padova et al.6 more than 25 years ago. This study, which had only 10 participants, found a significant beneficial effect of cholestyramine versus a placebo (P = 0.01). A positive linear relationship between itching and serum bile acids was also demonstrated. Other studies reporting beneficial effects of cholestyramine were not placebo-controlled.5, 20 The scientific basis for the use of cholestyramine as a treatment for cholestatic pruritus is therefore weak at best.21 The present study was the first adequately powered trial evaluating anion-exchange resins in cholestatic pruritus. The number of included patients was higher than that reported by any other comparable trials.10, 11 Furthermore, several complementary methods were used to assess treatment effects, and all analyses revealed consistent results. A weak aspect of the study was the unequal distribution of liver disease etiologies in the treatment groups. This imbalance in etiology was also reflected in the gender distribution. It seems unlikely, however, that these features have significantly influenced the main results of the trial.

In conclusion, this randomized, placebo-controlled trial shows that colesevelam is not more effective than a placebo in the treatment of cholestatic pruritus.