Colesevelam Hydrochloride (Cholestagel) a New, Potent Bile Acid Sequestrant Associated with a Low Incidence of Gastrointestinal Side Effects

ORIGINAL INVESTIGATION Colesevelam Hydrochloride (Cholestagel) A New, Potent Bile Acid Sequestrant Associated With a Low Incidence of Gastrointestinal Side Effects

Michael H. Davidson, MD; Maureen A. Dillon; Bruce Gordon, MD; Peter Jones, MD; Julie Samuels, MD; Stuart Weiss, MD; Jonathon Isaacsohn, MD; Phillip Toth, MD; Steven K. Burke, MD

Objectives: To compare colesevelam hydrochloride mg/dL) (19.1%) in the 3.75-g/d colesevelam treatment (Cholestagel), a nonabsorbed hydrogel with bile acid– group. Low-density lipoprotein cholesterol concentra- sequestering properties, with placebo for its lipid- tions at the end of treatment were significantly reduced lowering efficacy, its effects on laboratory and clinical from baseline levels in the 3.0- and 3.75-g/d colesevelam safety parameters, and the incidence of adverse events. treatment groups (P = .01 and PϽ.001, respectively). To- tal cholesterol levels demonstrated a similar response to Methods: Following diet and placebo lead-in periods, colesevelam treatment, with an 8.1% decrease from base- placebo or colesevelam was administered at 4 dosages (1.5, line in the 3.75-g/d treatment group (P<.001). High- 2.25, 3.0, or 3.75 g/d) for 6 weeks with morning and density lipoprotein cholesterol levels rose significantly evening meals to men and women with hypercholester- in the 3.0- and 3.75-g/d colesevelam treatment groups, olemia (low-density lipoprotein cholesterol level Ͼ4.14 by 11.2% (P = .006) and 8.1% (P = .02), respectively. mmol/L [Ͼ160 mg/dL]). Patients returned to the clinic Median triglyceride levels did not change from baseline, every 2 weeks throughout the treatment period for lipid nor were there any significant differences between parameter measurements and adverse event assess- treatment groups. The incidence of adverse events was ments. Samples were collected for serum chemistry pro- similar among all groups. files, hematologic studies, coagulation studies, and vitamin level assessment at baseline and after 6 weeks of Conclusions: Colesevelam therapy is effective for low- treatment. ering low-density lipoprotein cholesterol concentrations in persons with moderate hypercholesterolemia. It Results: Among the 149 patients randomized, 137 com- lacks the constipating effect of other bile acid seques- pleted the study. Low-density lipoprotein cholesterol con- trants, demonstrating the potential for increased com- centrations decreased in a dosage-dependent manner pliance. by 0.11 mmol/L (4.2 mg/dL) (1.8%) in the 1.5-g/d colesevelam treatment group and up to 1.01 mmol/L (39 Arch Intern Med. 1999;159:1893-1900

LEVATED SERUM concentra- Bile acid sequestrants interrupt the From the Chicago Center for tions of cholesterol, particu- enterohepatic circulation of bile acids by Clinical Research, Chicago, Ill larly low-density lipoprotein binding with them in the intestine to form (Dr Davidson); GelTex (LDL) cholesterol, constitute an insoluble complex that is excreted in Pharmaceuticals Inc, Waltham, a risk factor for the develop- the feces.10 The depletion of the bile acid Mass (Ms Dillon and ment of coronary heart disease (CHD).1-3 pool alters hepatic cholesterol homeosta- Dr Burke); the Rogosin E A direct correlation exists between hyper- sis, producing an increase in surface- Institute, New York, NY cholesterolemia and cardiovascular disease active LDL receptors and decreased se- (Dr Gordon); the Baylor Lipid 4,5 11 Clinic, Houston, Tex mortality and morbidity. Clinical trials rum cholesterol levels. The safety and (Dr Jones); National Clinical have demonstrated that aggressive reduction lipid-lowering efficacy of cholestyramine Research, Richmond, Va of lipid levels reduces this risk in patients resin and colestipol hydrochloride, 2 of the (Dr Samuels); the San Diego with and without CHD.1,6,7 In fact, each 1% most commonly prescribed bile acid se- Endocrine and Medical Clinic, reduction of LDL cholesterol lowers the risk questrants, which have been in use for over San Diego, Calif (Dr Weiss); of CHD by 2%.1,8 Diet and lifestyle interven- 20 years, have been demonstrated repeat- the Metabolic and tionsaregenerallythefirststepstakentoward edly. Bile acid sequestrants alone reduce Atherosclerosis Research reducing lipid levels but often have small and LDL cholesterol concentrations by 10% to Center, Cincinnati, Ohio inconsistent effects. Therefore, pharmaco- 30%.1,12 Combination therapy of bile acid (Dr Isaacsohn); and the Midwest Institute for Clinical logical measures are necessary for those sequestrants with niacin or lovastatin pro- Research Inc, Indianapolis, Ind patients who continue to have elevated lipid duces even larger LDL cholesterol reduc- 9 6,13 (Dr Toth). levels following diet and lifestyle changes. tions of up to 60%.

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1893

Downloaded From: https://jamanetwork.com/ on 10/01/2021 PATIENTS AND METHODS determine patient eligibility for the trial, (2) diet/washout (6 weeks for patients currently receiving lipid-lowering PATIENTS medication and 4 weeks for others) to normalize the effect of diet on lipid profiles and to allow for the washout of other Men and women aged 18 years or older who had elevated lipid-lowering drugs, (3) placebo run-in (4 weeks) to as- LDL cholesterol levels were screened for participation in sess compliance prior to randomization, and (4) drug treat- the study at 1 of 6 clinical research centers. Individuals were ment (6 weeks) to compare the safety and efficacy of ineligible if they had tendinous xanthomas, thyroid dis- colesevelam therapy with placebo. ease, clinically significant liver or renal disease, vasculitis, At the time of screening (week −8), all lipid-lowering human immunodeficiency virus infection, poorly con- medications were discontinued, and patients were in- trolled diabetes, poorly controlled hypertension (systolic structed to follow a National Cholesterol Education Pro- blood pressure Ͼ160 mm Hg or diastolic blood pressure gram (NCEP) Step 1 diet (total fat intake Յ30% of total Ͼ105 mm Hg), unstable cardiac disease, recent myocar- energy, saturated fat intake Յ10% of energy, and choles- dial infarction or cardiac bypass surgery (within 2 months terol intake Ͻ300 mg/d) for 4 weeks. Participants were pro- of screening), any evidence of active malignant neoplasm, vided with copies of the educational booklets, Step by Step: any clinically significant unstable medical condition, or any Eating to Lower Your High Blood Cholesterol15 and The 2D serum chemistry or hematologic abnormalities at screen- Food Portion Visual.16 Nutrient intake was estimated using ing. Patients were also excluded from participation if they the University of Minnesota Nutrition Coordination Cen- had a history of dysphagia, swallowing disorder, or motil- ter’s Nutrition Data System, version 2.8 (Minneapolis, Minn, ity disorder of the intestines. Individuals who used probu- 1996). Twenty-four–hour dietary recalls were centrally ad- col in the prior year, used fibrates in the month before ministered by telephone by the Nutrient Analysis Center screening, or had a recent history of ethanol or drug use at the Chicago Center for Clinical Research on 2 noncon- were also ineligible. Pregnant women and nursing moth- secutive days within the last 2 weeks of both the diet lead-in ers were excluded from participation, and women of child- period and the treatment period. bearing potential were required to undergo a urine preg- After the NCEP Step 1 diet stabilization period, pa- nancy test and to be adequately protected against pregnancy. tients returned to the clinic (week −4) for measurement of Patients could not take lipid-lowering medications con- fasting lipid profiles. Placebo was administered (single- currently. Patients were asked to avoid intentional changes blind) to be taken as 5 capsules twice per day with meals in diet during the study, such as fasting or binge eating; if for 4 weeks. Patients returned to the clinic after 2 weeks of they were taking fiber supplements (ie, cellulose, methyl- the placebo run-in period (week −2) and again at the end of cellulose, psyllium, polycarbophil, or bran), patients were the placebo run-in period (week 0) for lipid profile mea- also asked to maintain constant intake levels throughout surements, treatment compliance review, and safety assess- the study. ment. Criteria for entrance into the double-blind treatment This study was conducted in accordance with the US period required at least 80% compliance with placebo in the Code of Federal Regulations for clinical studies (21 CFR) final 2 weeks of the placebo run-in period and mean (weeks and the Declaration of Helsinki. Before the study started, −4 and −2) LDL cholesterol (4.14 mmol/L [160 mg/dL]) the investigators at each site forwarded copies of the pro- and TG (3.39 mmol/L [300 mg/dL]) levels. Furthermore, tocol and consent form of the study to an appropriately con- the difference between the 2 lipid measurements at weeks stituted institutional review board that reviewed and ap- −4 and −2 had to be 12% or more. If the difference was proved the informed consent and the protocol. Signed greater than 12%, a third measurement was obtained and informed consent was obtained from each patient prior to the mean of the 3 measurements was used to determine entering the study. eligibility. At week 0, the study drug was dispensed to patients STUDY DESIGN meeting all lipid and nonlipid criteria in a double-blind fashion according to a computer-generated randomiza- This was a multicenter, randomized, double-blind, placebo- tion scheme. Patients were randomized to receive placebo controlled study divided into 4 periods: (1) screening to or colesevelam hydrochloride in dosages of 0.75 g twice

Despite their proven efficacy and the lack of sys- currently available bile acid sequestrants would be a temic effects caused by nonabsorption in the gastroin- useful option for hypercholesterolemia treatment. testinal tract, the bile acid sequestrants that are cur- Each 400-mg capsule of Cholestagel (GelTex Phar- rently available have a high drug discontinuance rate maceuticals Inc, Waltham, Mass) contains 375 mg of an- caused, in part, by gastrointestinal side effects.14 Consti- hydrous colesevelam hydrochloride and 1% magnesium pation occurs in up to 39% of patients taking currently stearate. Colesevelam hydrochloride is a nonabsorbed poly- available bile acid sequestrants; in severe cases, it may mer (water-absorbing hydrogel) that has been specifi- lead to fecal impaction.1,14 Other gastrointestinal side ef- cally engineered to bind to bile acids. It is a polyallyl- fects, such as gas, bloating, flatulence, and cramping, also amine cross-linked with epichlorohydrin and alkylated with occur in a large percentage of patients who are treated 1-bromodecane and 6-bromohexyltrimethylammonium with currently available bile acid sequestrants.1 There- bromide (Figure 1). fore, an effective cholesterol-lowering agent with a non- Colesevelam has been found to have a high affinity systemic mechanism of action and greater tolerability than for both trihydroxy and dihydroxy bile acids in the in-

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1894

Downloaded From: https://jamanetwork.com/ on 10/01/2021 per day (1.5 g/d), 1.125 g twice per day (2.25 g/d), 1.5 g end of treatment (week 6). At each clinic visit, the pa- twice per day (3.0 g/d), or 1.875 g twice per day (3.75 g/d) tient’s weight, pulse, and blood pressure were measured and for 6 weeks. Colesevelam was supplied as tasteless, hard adverse events were assessed. Additionally, prothrombin gelatin capsules containing 375 mg of colesevelam hydro- time, partial thromboplastin time, and vitamin A, vitamin chloride. Placebo capsules identical in appearance to the E, and estradiol levels in women were measured at the time colesevelam capsules contained 350 mg of microcrystal- of randomization and at the end of the treatment period. line cellulose. The study drug was administered as 5 cap- Thyroid-stimulating hormone levels were assessed at the sules (placebo alone, colesevelam alone, or placebo and time of screening. colesevelam to arrive at the exact dose) taken with the morning meal and 5 capsules taken with the evening meal. If a STATISTICAL METHODS patient missed a meal, he or she was instructed to take the study drug with a snack. Missed doses were not to be made The statistical analyses presented are for the intent-to- up by doubling the next dose. Patients were also in- treat population. Intent-to-treat patients were all those who structed to take the study drug with a sufficient amount of were randomized, took at least 1 dose of study medica- liquid to ensure that the capsules cleared the oral cavity tion, and had at least 1 postbaseline efficacy evaluation. Miss- and esophagus. ing data were handled using a carry forward approach. All During the 6-week treatment phase, patients re- tests for main effects were 2-sided and were conducted at turned to the clinic at weeks 2, 4, and 6. At each visit, vital the ␣ = .05 level of significance, while tests for interaction signs were taken, fasting lipid profiles were measured, and effects were conducted at the ␣ = .10 level of significance. patients were questioned regarding adverse events and con- The comparability of the groups at baseline was assessed comitant medication use. Compliance with study medica- using analysis of variance (ANOVA) for continuous vari- tion was assessed by counting the number of pills re- ables and the Fisher exact test for categorical variables. turned. Additionally, at week 6, physical examinations were Analyses of lipid data were performed using all fast- performed and samples were collected for serum chemis- ing blood draws. The mean percentage change and the mean try profiles, hematologic studies, coagulation studies, and change from baseline (week 0) to end point (week 6) in vitamin level assessments. cholesterol concentrations were analyzed using a paired t test. The difference in percentage change and change across LIPID ANALYSES groups were analyzed using a 2-way ANOVA with factors for treatment, center, and treatment-by-center interac- Total and HDL cholesterol and TG levels were measured tion. A covariance model was tested for both change and in fasting (no food or drink other than water or nonsweet- percentage change in LDL cholesterol levels with covari- ened clear liquids for a minimum of 9 hours) serum, ac- ates of baseline LDL cholesterol levels, change in total fat cording to previously described procedures and the Stan- in the diet, change in body mass index (weight [kg]/ dardization Program of the Centers for Disease Control and [height (m)]2), and age. Since the distribution for changes Prevention and the National Heart, Lung, and Blood Insti- in TG levels was not normal, a Wilcoxon signed rank test tute.17 All laboratory analyses were performed by Smith- was used to test the change from baseline, and a Kruskal- Kline Beecham Clinical Laboratories (Van Nuys, Calif). Low- Wallis test was used to test between-group comparisons. density lipoprotein cholesterol levels (milligrams per Safety evaluations were performed on all patients who deciliter) were calculated using the equation described by received study treatment after randomization (safety popu- Friedewald et al.18 lation). The percentage of patients with adverse events was compared across treatment groups using the Fisher exact SAFETY MONITORING test. Changes in hematologic and chemical parameters from week 0 to week 6 were analyzed using a paired t test. Be- A complete physical examination and clinical laboratory tween-group changes in laboratory parameters were com- assessment, including serum chemistry profiles and hema- pared using a Kruskal-Wallis test, and changes in vital sign tologic tests, were completed during the initial screening variables were compared using ANOVA with factors for (week −8), at the time of randomization (day 0), and at the treatment, center, and treatment-by-center interaction.

testine, leading to increased fecal bile acid excretion. Pre- The study was also designed to investigate the effects of liminary in vivo and in vitro studies suggest that colesevelam therapy on high-density lipoprotein (HDL) colesevelam has superior bile acid binding efficacy com- cholesterol and triglyceride (TG) concentrations. pared with cholestyramine (GelTex Pharmaceuticals Inc, unpublished data, December 1995 and October 1997). RESULTS Furthermore, there is minimal potential for toxic effects, since colesevelam is not absorbed in the gastroin- Among the 275 persons screened, 149 were random- testinal tract. Its water-retaining characteristic creates a ized into 5 treatment groups of similar sizes (colesevelam soft, gelatinous-like material that minimizes the poten- therapy administered at dosages of 1.5 g/d [n = 30], 2.25 tial for gastrointestinal irritation. g/d [n = 30], 3.0 g/d [n = 31], or 3.75 g/d [n = 29] or pla- This double-blind, placebo-controlled study was con- cebo [n = 29]). Treatment groups were comparable with ducted to assess the safety and efficacy of colesevelam respect to demographic characteristics (Table 1). The therapy for lowering total and LDL cholesterol levels in mean age of all groups combined was 56 years (range, patients with moderate primary hypercholesterolemia. 31-80 years), with a distribution of 44% men and 56%

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1895

Downloaded From: https://jamanetwork.com/ on 10/01/2021 women. Eighty-two percent of the patients were white. after randomization [n = 1]). Overall compliance was more Baseline (week 0) lipid levels did not differ across treat- than 93% for all treatment groups. ment groups. The baseline mean ± SD total cholesterol level of all groups combined was 7.29 ± 0.87 mmol/L DIET (282 ± 34 mg/dL); LDL cholesterol, 5.13 ± 0.82 mmol/L (198 ± 32 mg/dL); HDL cholesterol, 1.32 ± 0.32 mmol/L Dietary intake of total energy and selected nutrients did (51 ± 12 mg/dL); and the baseline median TG, 1.78 not differ among treatment groups at baseline or at the end mmol/L (159 mg/dL). of treatment. On average, dietary compliance decreased Among the 149 patients randomized, 137 (92%) after week 0, as evidenced by an 11% increase in total fat completed the study. The dropout rates were similar intake, from 45 to 50 g/d, and a 23% increase in saturated among treatment groups: 0 in the placebo group, 2 in the fat intake, from 13 to 16 g/d. Mean energy intake increased 1.5-g/d colesevelam group (withdrawal of consent after 4%, from 6439 to 6720 kJ/d (1533 to 1600 kcal/d). These randomization [n = 1] and lost to follow-up [n = 1]), 3 changes were similar across all treatment groups. in the 2.25-g/d colesevelam group (adverse event [n = 1], lost to follow-up [n = 1], and other [n = 1]), 5 in the 3.0- SERUM LIPIDS g/d colesevelam group (adverse event [n = 2], withdrawal of consent after randomization [n = 2], and non- Low-density lipoprotein cholesterol concentrations compliance [n = 1]), and 2 in the 3.75-g/d colesevelam decreased in a dosage-dependent fashion in response to group (adverse event [n = 1] and withdrawal of consent colesevelam treatment. Low-density lipoprotein cholesterol concentrations at week 6 were reduced from baseline by 0.11 mmol/L (4.2 mg/dL) to 1.01 mmol/L (39 mg/dL) for the lowest to highest colesevelam dosages (Table 2). In the 2 groups receiving the highest dosages of colesevelam therapy, this represented a significant change from baseline (P = .01 and PϽ.001 for the NH2•nHCl A NH•nHCl NH•nHCl NH•nHCl 3.0-g/d and 3.75-g/d colesevelam therapy groups, respec-

(CH2)6 (CH2)9 tively). Low-density lipoprotein cholesterol concentra- + NMe3 CH3 – tions were reduced by 1.8%, 4.9%, 9.0%, and 19.1% in Cl D OH the 1.5-g/d, 2.25-g/d, 3.0-g/d, and 3.75-g/d colesevelam Cl– + NMe3 CH3 groups, respectively (Figure 2). The percentage of LDL (CH2)6 (CH2)9 cholesterol level reduction was statistically different from

NH2•nHCl NH•nHCl NH•nHCl NH•nHCl baseline for the 2 groups receiving the highest dosages of colesevelam (P = .01 and PϽ.001 for the 3.0-g/d and 3.75-g/d colesevelam groups, respectively). Individual LDL cholesterol level responses for all subjects are shown in A BDEGFigure 3. A = Primary Amines The greatest reduction in LDL cholesterol levels oc- B = Cross-linked Amines curred rapidly within the first 2 weeks of the treatment D = Quarternary Ammonium Alkylated Amines Figure 4 E = Decyalkylated Amines period ( ). During the second 2 weeks of treat- n = Fraction of Protonated Amines ment, the 1.5-g/d and 3.75-g/d colesevelam groups ex- G = Extended Polymeric Network hibited additional modest decreases, but the 2.25-g/d and 3.0-g/d colesevelam groups demonstrated relatively con- Figure 1. Chemical structure of colesevelam hydrochloride (Cholestagel). stant effects. During the final 2 weeks of treatment, there

Table 1. Baseline Patient Demographic Characteristics*

Colesevelam Hydrochloride, g/d Placebo Characteristics (n = 29) 1.5 (n = 30) 2.25 (n = 29) 3.0 (n = 30) 3.75 (n = 29) Age, y† 57.1 ± 11.6 54.1 ± 11.3 54.8 ± 11.7 56.3 ± 12.0 57.9 ± 8.4 Sex, No. (%) Male 13 (45) 13 (43) 18 (62) 10 (33) 11 (38) Female 16 (55) 17 (57) 11 (38) 20 (67) 18 (62) Race, No. (%) White 21 (72) 27 (90) 21 (72) 25 (83) 27 (93) Black 4 (14) 3 (10) 6 (21) 4 (13) 1 (3) Hispanic 1 (3) 0 (0) 1 (3) 1 (3) 0 (0) Asian 3 (10) 0 (0) 1 (3) 0 (0) 1 (3) Weight, kg† 73.9 ± 15.3 77.8 ± 12.9 79.8 ± 14.9 72.5 ± 14.5 80.5 ± 14.7 Body mass index, kg/m2† 26.7 ± 4.1 27.6 ± 3.8 27.9 ± 4.8 26.4 ± 4.3 28.2 ± 4.9

*There were no significant differences among treatment groups. †Mean ± SD.

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1896

Downloaded From: https://jamanetwork.com/ on 10/01/2021 Table 2. Lipid Levels at Baseline (Week 0) and End of Treatment (Week 6), With Change From Baseline to End of Treatment*

Colesevelam Hydrochloride, g/d

Time Placebo† 1.5 2.25 3.0 3.75 LDL Cholesterol Levels, mmol/L (mg/dL) (n = 28) (n = 26) (n = 26) (n = 25) (n = 24) Baseline 5.01 ± 0.65 (194 ± 25) 5.03 ± 0.99 (195 ± 38) 5.21 ± 0.78 (202 ± 30) 5.16 ± 1.02 (200 ± 39) 5.22 ± 0.66 (202 ± 26) Week 6 4.99 ± 0.75 (193 ± 29) 4.92 ± 1.04 (190 ± 40) 4.93 ± 0.94 (191 ± 37) 4.68 ± 1.34 (181 ± 52) 4.21 ± 0.70 (163 ± 27) Change‡ −0.02 ± 0.59 (−0.8 ± 23) −0.11 ± 0.52 (−4.2 ± 20) −0.28 ± 0.73 (−11 ± 28) −0.48 ± 0.90§ (−18 ± 35) −1.01 ± 0.61࿣ (−39 ± 24) Total Cholesterol Levels, mmol/L (mg/dL) (n = 29) (n = 26) (n = 26) (n = 25) (n = 27) Baseline 7.27 ± 0.85 (281 ± 33) 7.31 ± 0.99 (283 ± 38) 7.37 ± 0.80 (285 ± 31) 7.15 ± 1.03 (277 ± 40) 7.36 ± 0.67 (285 ± 26) Week 6 7.25 ± 0.97 (280 ± 38) 7.18 ± 1.07 (278 ± 41) 7.06 ± 0.96 (273 ± 37) 6.81 ± 1.37 (263 ± 53) 6.75 ± 0.90 (261 ± 35) Change −0.02 ± 0.58 (−0.7 ± 22) −0.12 ± 0.54 (−4.8 ± 21) −0.31 ± 0.74§ (−12 ± 29) −0.34 ± 0.85 (−13 ± 33) −0.61 ± 0.80࿣ (−24 ± 31) HDL Cholesterol Levels, mmol/L (mg/dL) (n = 29) (n = 26) (n = 27) (n = 25) (n = 27) Baseline 1.38 ± 0.37 (53 ± 14) 1.34 ± 0.26 (52 ± 10) 1.34 ± 0.31 (52 ± 12) 1.23 ± 0.31 (47 ± 12) 1.31 ± 0.32 (51 ± 12) Week 6 1.37 ± 0.39 (53 ± 15) 1.36 ± 0.26 (52 ± 10) 1.35 ± 0.29 (52 ± 11) 1.34 ± 0.30 (52 ± 12) 1.42 ± 0.44 (55 ± 17) Change −0.01 ± 0.21 (−0.5 ± 8.3) 0.01 ± 0.21 (0.5 ± 8.1) 0.01 ± 0.16 (0.5 ± 6.3) 0.11 ± 0.19¶ (4.4 ± 7.2) 0.11 ± 0.24§ (4.3 ± 9.1) Triglyceride Levels, mmol/L (mg/dL) (n = 29) (n = 26) (n = 26) (n = 25) (n = 27) Baseline 1.90 ± 0.87 (169 ± 78) 2.00 ± 0.85 (178 ± 75) 1.74 ± 0.84 (155 ± 74) 1.54 ± 0.78 (137 ± 70) 1.73 ± 0.99 (154 ± 88) Week 6 1.76 ± 1.07 (157 ± 95) 1.94 ± 0.81 (173 ± 72) 1.44 ± 0.85 (128 ± 75) 1.62 ± 0.74 (144 ± 66) 1.93 ± 2.11 (172 ± 188) Change 0.06 ± 0.65 (5 ± 58) 0.03 ± 0.71 (2.5 ± 63) −0.02 ± 0.64 (−2 ± 57) 0.02 ± 0.64 (2 ± 57) 0.16 ± 1.48 (14 ± 132)

*All treatment groups had similar lipid levels at baseline. Values for low-density lipoprotein (LDL) cholesterol, total cholesterol, and high-density lipoprotein (HDL) cholesterol levels are mean ± SD; for triglyceride levels, median ± SD. †Values for the placebo group did not change significantly from baseline for any lipid parameter. ‡Mean change from baseline was significantly different among treatment groups ( PϽ.001). §Significant change from baseline ( PϽ.05). ࿣Significant change from baseline ( PϽ.001). ¶Significant change from baseline ( PϽ.01). was a slight increase in LDL cholesterol levels in the 1.5-g/d, 2.25-g/d, and 3.0-g/d colesevelam groups. How- Placebo ever, in the 3.75-g/d colesevelam group, mean percent- Colesevelam Hydrochloride: 1.5 g/d age decreases in LDL cholesterol levels remained con- 2.25 g/d stant between the 2 intervals (19.2% decrease from 3.0 g/d week 0 to week 4 vs 19.1% decrease from week 0 to 15 3.75 g/d week 6). Using analysis of covariance, baseline LDL choles- 10 terol level and colesevelam treatment group were found to be predictive of the change in LDL cholesterol levels. 5 Higher baseline LDL cholesterol levels and high colesevelam dosage led to greater absolute reductions in 0 LDL cholesterol levels at the end of treatment. –5 Colesevelam treatment group was the only significant factor that predicted percentage change in LDL cholesterol Change From Baseline, % –10 levels. Other covariates (ie, sex, change of total fat in the diet, change of body mass index, and age) were not sig- –15 nificant in either the change or percentage change in LDL cholesterol levels. –20 Colesevelam treatment also reduced total choles- Total LDL HDL TG Cholesterol Cholesterol Cholesterol terol concentrations (Table 2). Decreases from baseline to week 6 in total cholesterol levels ranged from 0.12 Figure 2. Percentage change in total cholesterol, low-density lipoprotein mmol/L (4.8 mg/dL) to 0.61 mmol/L (24 mg/dL) across [LDL] cholesterol, high-density lipoprotein [HDL] cholesterol, and triglyceride (TG) levels from baseline (week 0) to the end of treatment (week the 4 colesevelam groups. These reductions from base- 6) in patients who received placebo or colesevelam hydrochloride therapy. line were significant in the 2.25-g/d and 3.75-g/d Cholesterol values are means; TG values are medians. Mean percentage colesevelam groups (P = .04 and P<.001, respectively) and changes in LDL cholesterol levels were significantly different among neared significance in the 3.0-g/d colesevelam group treatment groups (PϽ.001). (P = .06). The percentage reductions from baseline in total cholesterol levels were modest (highest, 8.1%). These the 3.0-g/d and 3.75-g/d colesevelam groups, respec- reductions were significant for the 2 groups receiving the tively) and approached significance in the 2.25-g/d highest dosages of colesevelam (P = .05 and P<.001 for colesevelam group (P = .05).

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1897

Downloaded From: https://jamanetwork.com/ on 10/01/2021 40 Table 3. Gastrointestinal Adverse Events* 30

20 Colesevelam Hydrochloride, g/d

10 Adverse Placebo 1.5 2.25 3.0 3.75 Event (n = 29) (n = 30) (n = 30) (n = 31) (n = 29) 0 Overall 3 (10) 3 (10) 6 (20) 7 (23) 8 (28) –10 Flatulence 2 (7) 0 (0) 3 (10) 3 (10) 3 (10) –20 Constipation 0 (0) 1 (3) 0 (0) 3 (10) 2 (7) Dyspepsia 1 (3) 1 (3) 1 (3) 1 (3) 2 (7) –30 Diarrhea† 0 (0) 0 (0) 3 (10) 0 (0) 0 (0)

–40 Nausea 0 (0) 0 (0) 1 (3) 0 (0) 1 (3) Anorexia 0 (0) 0 (0) 1 (3) 0 (0) 0 (0) Change From Baseline in LDL Cholesterol, % –50 Dysphagia 0 (0) 0 (0) 0 (0) 0 (0) 1 (3)

–60 Eructation 0 (0) 0 (0) 1 (3) 0 (0) 0 (0) Placebo 1.5 2.25 3.0 3.75 Mouth ulceration 0 (0) 1 (3) 0 (0) 0 (0) 0 (0) Tooth disorder 1 (3) 0 (0) 0 (0) 0 (0) 0 (0) Colesevelam Hydrochloride, g/d

Figure 3. Individual percentage changes in low-density lipoprotein (LDL) *Values are No. (%). cholesterol levels from baseline (week 0) to the end of treatment (week 6) in †Significantly different between treatment groups ( P = .03). patients who received placebo or colesevelam hydrochloride therapy (1.5, 2.25, 3.0, or 3.75 g/d). Solid circles indicate individual patients; horizontal lines, mean percentage changes in LDL cholesterol levels. [16/29]). The most common side effects were those that occurred in the body as a whole (ie, infection). The di-

Placebo gestive system was the area with the second most fre- 8 (309) Colesevelam Hydrochloride: quent reports of adverse events; flatulence and consti- 1.5 g/d pation were the most common complaints (Table 3). 7 (271) 2.25 g/d 3.0 g/d There were no statistically significant differences among 3.75 g/d treatment groups for the incidence of gastrointestinal ad- 6 (232) verse events, with the exception of diarrhea, which was reported by 3 patients in the 2.25-g/d colesevelam group

5 (193) only. Four (2.7%) of 149 patients left the study because of digestive system adverse events (1 from the 2.25-g/d colesevelam group because of diarrhea and gas, 2 from 4 (155) the 3.0-g/d colesevelam group because of constipation,

LDL Cholesterol, mmol/L (mg/dL) and 1 from the 3.75-g/d colesevelam group because of 3 (116) stomach burning). There were no clinically significant changes from base- 2 (77) line to the end of treatment in serum chemistry param- –4 –2 0 2 46 Week of Study eters. In general, the change from baseline for indicators of kidney and liver function did not differ among treat- Figure 4. Mean low-density lipoprotein (LDL) cholesterol concentrations in patients who received placebo or colesevelam hydrochloride therapy from ment groups, with the exception of a non–dose-related screening through the end of treatment. The diet period started at week −4, alkaline phosphatase response. Mean ± SD alkaline phos- and the treatment period started at week 0. phatase levels increased by 3.04 ± 5.93 U/L (P = .01) and 4.92 ± 6.44 U/L (P<.001) from baseline levels of High-density lipoprotein cholesterol levels did not dif- 67.37 ± 22.07 U/L and 70.35 ± 22.20 U/L for the 1.5- and fer across treatment groups at the end of treatment (Table 3.0-g/d colesevelam groups, respectively. In the group re- 2). However, HDL cholesterol levels significantly in- ceiving the highest dosage of colesevelam (3.75 g/d), ala- creased by approximately 0.10 mmol/L (4 mg/dL) in the nine aminotransferase and aspartate aminotransferase lev- 3.0-g/d and 3.75-g/d colesevelam groups (11.2%, P = .006, els were elevated by 6.81 ± 11.06 U/L (P<.001) and and 8.1%, P = .02, respectively). Median TG concentra- 3.41 ± 6.84 U/L (P = .02) from baseline levels of 18.52 ± 7.79 tions did not differ significantly across treatment groups, U/L and 18.14 ± 4.71 U/L, respectively, but remained within nor was there any significant change from baseline in TG the normal range. There were no clinically significant levels within any group (Table 2). changes noted for hematologic parameters, serum levels of vitamins A and E, prothrombin time, partial thrombo- SAFETY plastin time, estradiol levels, body weight, pulse, and systolic and diastolic blood pressure throughout the course All dosages of colesevelam were generally well toler- of the trial. Physical examination following treatment re- ated. There was no statistically significant difference across vealed little change from baseline assessment. treatment groups in the number of patients who experienced adverse events caused by colesevelam treatment COMMENT (placebo, 41% [12/29]; 1.5-g/d colesevelam, 40% [12/ 30]; 2.25-g/d colesevelam, 53% [16/30]; 3.0-g/d In this randomized, double-blind, placebo-controlled trial, colesevelam, 52% [16/31], and 3.75-g/d colesevelam, 55% colesevelam treatment administered at dosages of 1.5 g/d,

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1898

Downloaded From: https://jamanetwork.com/ on 10/01/2021 2.25 g/d, 3.0 g/d, and 3.75 g/d produced dosage-related data further suggests that compliance with bile acid se- reductions of total and LDL cholesterol concentrations questrants is likely to be even lower in nonresearch set- in patients with moderate hypercholesterolemia. tings.14 In the present study, overall compliance with Colesevelam treatment at 3.75 g/d resulted in a 19% de- colesevelam treatment was very high (93%), which re- crease in LDL cholesterol levels, which is similar to the flects its low incidence of intolerable side effects. Al- responses produced by standard dosages of other bile acid though this study period was shorter (6 weeks) than in sequestrants.19-24 The currently available bile acid seques- some previous reports, side effects have typically been trants, cholestyramine and colestipol hydrochloride, are found to be more severe in the early stages of trials, when generally administered as packets to reach dosages of 24 the gastrointestinal tract attempts to adjust to a nonab- or 30 g/d, respectively. Thus, colesevelam treatment ap- sorbable bile acid sequestrant.1,33 Thus, long-term com- pears to have greater potency than these other bile acid pliance with colesevelam therapy would be expected to sequestrants. remain high. Although the LDL cholesterol response was dosage- Laboratory parameters were largely unchanged dependent (ie, LDL cholesterol levels were reduced with except for statistically but not clinically significant in- each increase in the dosage of colesevelam), there seemed creases in alkaline phosphatase levels in the 1.5- and 3.0- to be increased efficacy per gram of colesevelam at the g/d colesevelam groups and increases in aspartate ami- highest dosage. This increased efficacy differs from the notransferase and alanine aminotransferase levels in the type of nonlinear response that is seen with most other 3.75-g/d colesevelam group. These alterations have been bile acid sequestrants. Typically, there is proportionally reported with use of other bile acid sequestrants1,7,32,35 and less additional LDL cholesterol lowering with increased are believed not to represent toxic effects but rather to dosages.25-28 This may be a chance finding owing to the be secondary to changes in lipid metabolism, such as in- relatively small sample size in the study, or it may be increased bile acid synthesis and excretion.36 Surpris- dicative of the recruitment of additional cholesterol- ingly, levels of vitamin E, which is carried in the LDL frac- lowering mechanisms at the highest dosage. Alter- tion, were not reduced along with LDL cholesterol levels. nately, the dosages used in the present study may not have Results from longer studies of colesevelam therapy will all fallen in a linear portion of the dosing curve. likely provide additional information regarding this ap- The response of other lipoprotein fractions to parent anomaly. colesevelam treatment was similar to that observed with In order to avoid the adverse events that lead to poor other bile acid sequestrants. Effects on HDL cholesterol compliance with bile acid sequestrant therapy, smaller dos- levels have varied among studies but generally exhibit a ages are often used in combination with 3-hydroxy-3- slight increase of 3% to 8%.20,21,23,29 This is consistent with methylglutaryl coenzyme A (HMG CoA) reductase in- the 8% increase in HDL cholesterol levels in the 3.75- hibitors, probucol, fibrates, or nicotinic acid, resulting in g/d colesevelam group observed in this study. There are LDL cholesterol level reductions as large as 60%.6,7,13,22,31,37,38 reports of TG level increases ranging from 5% to 20% with Although the increased complexity of the dosage regi- bile acid sequestrant therapy.19,21,23,24,30 These increases men may also reduce compliance, it is particularly effec- tend to be transient and are typically related to high base- tive because bile acid sequestrants and HMG CoA reduc- line TG levels or dysbetalipoproteinemia.25,31,32 In the cur- tase inhibitors have nonlinear dosage response curves.25 rent study, median TG levels did not differ across treat- Proportionally less additional LDL cholesterol level low- ment groups, although the 3.75-g/d colesevelam group ering is realized with the increased dosage of a single showed a trend toward higher TG levels at 6 weeks. drug.25-28 Combination therapy is an important strategy for Colesevelam treatment was well tolerated at all dos- optimizing the effectiveness of drug therapy to achieve LDL ages and lacked the constipating effect that is typical of cholesterol level goals.25 In fact, unpublished data from a other bile acid sequestrants. Gastrointestinal side ef- study comparing combination therapy of low-dose Cho- fects, particularly constipation, occur frequently in pa- lestagel and lovastatin with either drug therapy alone tients treated with cholestyramine and colestipol and lead showed an additive effect on LDL cholesterol level reduc- to poor compliance and a very high discontinuance tion with the combination regimen. rate.1,2,14,23,30,31,33,34 In the Lipid Research Clinics– Since publication of the LRC-CPPT results, and more Coronary Primary Prevention Trial (LRC-CPPT), more recently the Cholesterol Lowering Atherosclerosis Study6,37 than 1900 patients received cholestyramine therapy at a and the Familial Atherosclerosis Treatment Study,7 it is dosage of 24 g/d for a 7-year period. By the end of the clear that bile acid sequestrant therapy, alone and in com- first year of the study, 68% of these patients reported at bination with other agents, is effective for reducing cho- least 1 gastrointestinal side effect, 39% had constipa- lesterol levels as well as cardiac morbidity and mortality. tion, 32% had abdominal gas, and 27% had heartburn.1 Although the lipid-lowering effects of colesevelam therapy As a result of these adverse events, few patients in were relatively modest and less than those expected with the LRC-CPPT were compliant with the recommended HMG CoA reductase inhibitor therapy, they represent a dosage of 24 g/d of cholestyramine (average dose, ap- beneficial response in terms of CHD risk reduction. It is proximately 14 g/d in year 7). The mean postdiet reduc- estimated that approximately 52 million people have at tion in LDL cholesterol level for patients taking chol- least mild hypercholesterolemia and would benefit from estyramine at 12 to 16 g/d was only 16%; among patients lipid-lowering therapy. For the vast majority of persons who were more compliant with treatment (at least 20 g/d), with hypercholesterolemia, LDL cholesterol level reduc- LDL cholesterol levels were reduced by 28%.2 Compari- tions of 15% to 25% with diet and bile acid sequestrant son of health maintenance organization and clinical trial therapy would be adequate to reach their NCEP LDL cho-

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1899

Downloaded From: https://jamanetwork.com/ on 10/01/2021 lesterol level goal.39 In the LRC-CPPT, total and LDL cho- of pravastatin alone and with cholestyramine in hypercholesterolemia. Clin Phar- lesterol level reductions of 8% and 12%, respectively, be- macol Therap. 1990;48:201-207. 14. Andrade SE, Walker AM, Gottlieb LK, et al. Discontinuation of antihyperlipid- yond the effects of diet restrictions and relative to placebo emic drugs: do rates reported in clinical trials reflect rates in primary care set- produced a 19% decrease in risk for CHD death and/or non- tings? N Engl J Med. 1995;332:1125-1131. fatal myocardial infarction.1 15. American Heart Association. Step by Step: Eating to Lower Your High Blood Cho- Colesevelam therapy may offer an alternative to other lesterol. Bethesda, Md: National Institutes of Health; 1994. Publication NIH 94- bile acid sequestrants and systemic lipid-lowering drugs 2920. 16. Posner BM, Morgan JL. The 2D Food Portion Visual. Framingham, Mass: Nutri- for individuals requiring moderate (Յ20%) LDL choles- tion Consulting Enterprises; 1992. terol level reduction and has the potential for use in com- 17. Myers G, Cooper G, Winn C, Smith S. The Centers for Disease Control–National bination therapy. There is a large population of persons Heart Lung, and Blood Institute Lipid Standardization Program: an approach to with mild to moderate hypercholesterolemia for whom accurate and precise lipid measurements. Clin Lab Med. 1989;9:105-135. NCEP guidelines recommend bile acid sequestrants as first- 18. Friedewald W, Levy R, Fredrickson D. Estimation of the concentration of low den- 9,31 sity lipoprotein cholesterol in plasma, without use of the preparative ultracen- line therapy. However, primarily because of gastroin- trifuge. Clin Chem. 1972;18:499-502. testinal side effects, treatment is frequently discontinued, 19. Betteridge DJ, Bhatnager D, Bing RF, et al. Treatment of familial hypercholes- and the hypercholesterolemia either goes untreated or is terolaemia: United Kingdom lipid clinics study of pravastatin and cholestyra- treated with a systemic drug.14 Colesevelam therapy may mine. BMJ. 1992;304:1335-1338. achieve beneficial lipid reductions with very few of the un- 20. Brensike JF, Levy RI, Kelsey SF, et al. Effects of therapy with cholestyramine on progression of coronary arteriosclerosis: results of the NHLBI Type II Coronary pleasant gastrointestinal side effects associated with other Intervention Study. Circulation. 1984;69:313-324. bile acid sequestrants or the possible systemic toxic ef- 21. Farmer JA, Gotto AM Jr. Currently available hypolipidaemic drugs and future thera- fects experienced with absorbed lipid-altering agents. peutic developments. Baillieres Clin Endocrinol Metab. 1995;9:825-847. 22. Illingworth DR. How effective is drug therapy in heterozygous hypercholester- Accepted for publication January 26, 1999. olemia? Am J Cardiol. 1993;72:54D-58D. 23. The Lovastatin Study Group III. A multicenter comparison of lovastatin and cho- This study was funded by GelTex Pharmaceuticals Inc, lestyramine therapy for severe primary hypercholesterolemia. JAMA. 1988;260: Waltham, Mass. 359-366. We thank Mary R. Dicklin, PhD, for her assistance with 24. Wiklund O, Angelin B, Fager G, et al. Treatment of familial hypercholesterola- the preparation of the manuscript. emia: a controlled trial of the effects of pravastatin or cholestyramine therapy on Corresponding author: Michael H. Davidson, MD, Chi- lipoprotein and apolipoprotein levels. J Intern Med. 1990;228:241-247. 25. Schectman G, Hiatt J. Evaluation of the effectiveness of lipid-lowering therapy cago Center for Clinical Research, 515 N State St, Suite 2700, (bile acid sequestrants, niacin, psyllium, and lovastatin) for treating hypercho- Chicago, IL 60610 (e-mail: [email protected]). lesterolemia in veterans. Am J Cardiol. 1993;71:759-765. 26. Hunninghake DB, Probstfield JL, Crow LO, Isaacson SO. Effect of colestipol and clofibrate on plasma lipids and lipoproteins in type IIa hyperlipoproteinemia. REFERENCES Metabolism. 1981;30:605-609. 27. Superko HR, Greenland P, Manchester RA, et al. Effectiveness of low-dose colesti- 1. Lipid Research Clinics Program. The Lipid Research Clinics coronary primary pol therapy in patients with moderate hypercholesterolemia. Am J Cardiol. 1992; prevention trial results, I: reduction in the incidence of coronary heart disease. 70:135-140. JAMA. 1984;251:351-364. 28. Vecchio TJ, Linden CV, O’Connell MJ, Heilman J. Comparative efficacy of colesti- 2. Lipid Research Clinics Program. The Lipid Research Clinics coronary primary pol and clofibrate in type IIa hyperlipoproteinemia. Arch Intern Med. 1982;142: prevention trial results, II: the relationship of reduction in incidence of coronary 721-723. heart disease to cholesterol lowering. JAMA. 1984;251:365-374. 29. Lyons D, Webster J, Fowler G, Petrie JC. Colestipol at varying dosage interval in 3. National Cholesterol Education Program Expert Panel. Report of the National Cho- the treatment of moderate hypercholesterolaemia. Br J Clin Pharmacol. 1994; lesterol Education Program Expert Panel on Detection, Evaluation, and Treat- 37:59-62. ment of High Blood Cholesterol in Adults. Arch Intern Med. 1988;148:36-39. 30. Verzino EM, Kaplan B. Rational treatment for dyslipidemia. W V Med J. 1994; 4. Kannel WB, Castelli WP, Gordon T, McNamara PM. Serum cholesterol lipopro- 90:58-61. teins and the risk of coronary heart disease. Ann Intern Med. 1971;74:1-12. 31. Farmer JA, Gotto AM Jr. Choosing the right lipid-regulating agent: a guide to se- 5. Martin MJ, Hulley SY, Browner WS, Kuller LH, Wentworth D. Serum choles- lection. Drugs. 1996;52:649-661. terol, blood pressure, and mortality: implications from a cohort of 361,662 men. 32. Hunninghake DB, Stein EA, Fraser Bremner W, Greenland P, Demke DM, Oliph- Lancet. 1986;2:933-939. ant TH. Dose-response study of colestipol tablets in patients with moderate hy- 6. Blackenhorn DH, Nessim SA, Johnson RL, Sanmarco ME, Azen SP, Cashin- percholesterolemia. Am J Ther. 1995;2:180-189. Hemphill L. Beneficial effects of combined colestipol-niacin therapy on coro- 33. Cooper EE, Michel AM. Colestipol hydrochloride, a new hypolipidemic drug: a nary atherosclerosis and coronary venous bypass grafts. JAMA. 1987;257:3233- two-year study. South Med J. 1975;68:303-309. 3240. 34. Dorr AE, Gundersen K, Schneider JC Jr, Spencer TW, Martin WB. Colestipol 7. Brown G, Albers JJ, Fisher LD, et al. Regression of coronary artery disease as a hydrochloride in hypercholesterolemic patients: effect on serum cholesterol and result of intensive lipid-lowering therapy in men with high levels of apolipopro- mortality. J Chronic Dis. 1978;31:5-14. tein B. N Engl J Med. 1990;323:1289-1298. 35. Hoogerbrugge N, Mol MJTM, Van Dormaal JJ, et al. The efficacy and safety 8. Shepherd J, Cobbe ST, Ford I, et al. Prevention of coronary heart disease with pravas- of pravastatin, compared to and in combination with bile acid binding resins, in tatin in men with hypercholesterolemia. N Engl J Med. 1995;333:1301-1307. familial hypercholesterolaemia. J Intern Med. 1990;228:261-266. 9. National Cholesterol Education Program. Second report of the Expert Panel on 36. Probstfield JL, Statland BE, Gorman L, et al. Alterations in human serum alka- Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult line phosphatase and its isoenzymes by hypolipidemic agents: colestipol and clo- Treatment Panel II). Circulation. 1994;89:1333-1445. fibrate. Metabolism. 1983;32:818-821. 10. Grundy SM, Ahrens EH, Salen G. Interruption of the enterohepatic circulation of 37. Cashin-Hemphill L, Mack WJ, Pogoda JM, Sanmarco ME, Azen SP, Blanken- bile acids in man: comparative effects of cholestyramine and ileal exclusion on horn DH. Beneficial effects of colestipol-niacin on coronary atherosclerosis: a cholesterol metabolism. J Lab Clin Med. 1971;178:94-121. 4-year follow-up. JAMA. 1990;264:3013-3017. 11. Shepherd J, Packard CJ, Bicker S, Lawrie TDV, Morgan HG. Cholestyramine pro- 38. Kuo PT, Wilson AC, Kostis JB, Moreyra AE. Effects of combined probucol- motes receptor-mediated low-density-lipoprotein catabolism. N Engl J Med. 1980; colestipol treatment for familial hypercholesterolemia and coronary artery dis- 302:1219-1222. ease. Am J Cardiol. 1986;57:43H-48H. 12. Levine GN, Keaney JF, Vita JA. Cholesterol reduction in cardiovascular disease, 39. Sempos CT, Cleeman JI, Carroll MD, et al. Prevalence of high blood cholesterol clinical benefits and possible mechanisms of action. N Engl J Med. 1995;332: among US adults: an update based on guidelines from the second report of the 512-521. National Cholesterol Education Program Adult Treatment Panel. JAMA. 1993; 13. Pan HY, Devault AR, Swites BJ, et al. Pharmacokinetics and pharmacodynamics 269:3009-3014.

ARCH INTERN MED/ VOL 159, SEP 13, 1999 1900

Downloaded From: https://jamanetwork.com/ on 10/01/2021