Flowchart shows subject participation. ITT represents intent-to-treat.
Least squares mean (SEM) change from baseline in glycated hemoglobin A1c (HbA1c) (A), fasting plasma glucose (FPG) (B), and fructosamine (C) levels in subjects (intent-to-treat population without last observation carried forward imputation) receiving colesevelam hydrochloride, 3.75 g/d, or placebo for 16 weeks. *P < .05 vs placebo. †P < .01 vs placebo. ‡P < .001 vs placebo. To convert glucose to millimoles per liter, multiply by 0.0555; HbA1c to a proportion of total hemoglobin, multiply by 0.01.
Least squares mean percent change from baseline in lipid and lipoprotein levels in subjects (intent-to-treat population with last observation carried forward imputation) receiving colesevelam hydrochloride, 3.75 g/d, or placebo for 16 weeks. Triglyceride levels (TG) are reported as median rather than mean. The numbers above the bracketed pairs of bars refer to the differences between the mean or median change in the group receiving placebo and the mean or median change in the group receiving colesevelam. LDL-C indicates low-density lipoprotein cholesterol; non–HDL-C, non–high-density lipoprotein cholesterol; TG, triglycerides; TC, total cholesterol; TG, triglycerides; HDL-C, high-density lipoprotein cholesterol; ApoA-I, apolipoprotein A-I; ApoB, apolipoprotein B. *P < .001. †P < .05.
Least squares mean change from baseline in lipid and lipoprotein ratios in subjects (intent-to-treat population with last observation carried forward imputation) receiving colesevelam hydrochloride, 3.75 g/d, or placebo for 16 weeks. The numbers above the bracketed pairs of bars refer to the differences between the mean or median change in the group receiving placebo and the mean or median change in the group receiving colesevelam. TC indicates total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; non–HDL-C, non–high-density lipoprotein cholesterol; ApoB, apolipoprotein B; and ApoA-I, apolipoprotein A-I. *P < .001. †P = .004.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Goldberg RB, Fonseca VA, Truitt KE, Jones MR. Efficacy and Safety of Colesevelam in Patients With Type 2 Diabetes Mellitus and Inadequate Glycemic Control Receiving Insulin-Based Therapy. Arch Intern Med. 2008;168(14):1531–1540. doi:10.1001/archinte.168.14.1531
Copyright 2008 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2008
Poor glycemic control is a risk factor for microvascular complications in patients with type 2 diabetes mellitus. Achieving glycemic control safely with insulin therapy can be challenging.
A prospective, 16-week, multicenter, randomized, double-blind, placebo-controlled, parallel-group study conducted at 50 sites in the United States and 1 site in Mexico between August 12, 2004, and December 28, 2005. Subjects had type 2 diabetes mellitus that was not adequately controlled (glycated hemoglobin level, 7.5%-9.5%, inclusive) receiving insulin therapy alone or in combination with oral antidiabetes agents. In total 287 subjects (52% men; mean age, 57 years; with a mean baseline glycated hemoglobin level of 8.3%) were randomized: 147 to receive colesevelam hydrochloride, 3.75 g/d, and 140 to receive placebo.
Using the least squares method, the mean (SE) change in glycated hemoglobin level from baseline to week 16 was −0.41% (0.07%) for the colesevelam-treated group and 0.09% (0.07%) for the placebo group (treatment difference, −0.50% [0.09%]; 95% confidence interval, −0.68% to −0.32%; P < .001). Consistent reductions in fasting plasma glucose and fructosamine levels, glycemic-control response rate, and lipid control measures were observed with colesevelam. As expected, the colesevelam-treated group had a 12.8% reduction in low-density lipoprotein cholesterol concentration relative to placebo (P < .001). Of recipients of colesevelam and placebo, respectively, 30 and 26 discontinued the study prematurely; 7 and 9 withdrew because of protocol-specified hyperglycemia, and 10 and 4 withdrew because of adverse events. Both treatments were generally well tolerated.
Colesevelam treatment seems to be safe and effective for improving glycemic control and lipid management in patients with type 2 diabetes mellitus receiving insulin-based therapy, and it may provide a novel treatment for improving dual cardiovascular risk factors.
clinicaltrials.gov Identifier: NCT00151749
Type 2 diabetes mellitus (T2DM) is a well-known cardiovascular risk factor and a common cause of morbidity and mortality. It is well established that a direct relationship exists between the degree of glycemic control and risk of microvascular complications.1-3 The goal of antidiabetes therapy is to maintain plasma glucose levels as close to normal as possible. The American Diabetes Association recommends a glycated hemoglobin A1c (HbA1c) goal of less than 7.0% (to convert to a proportion of total hemoglobin, multiply by 0.01), the level at which clinical trials have demonstrated fewer long-term microvascular and macrovascular complications.4 However, data from the 1999-2000 National Health and Nutrition Examination Survey suggest that only 37% of patients with T2DM achieve this goal.5
In addition to glycemic control, another major risk factor for cardiovascular disease is elevated serum cholesterol concentrations, in particular, elevated low-density lipoprotein cholesterol (LDL-C) concentrations.6 The American Diabetes Association recommends a primary LDL-C goal in patients with T2DM to be less than 100 mg/dL (to convert to millimoles per liter, multiply by 0.0259) in individuals without overt cardiovascular disease and less than 70 mg/dL in those with overt cardiovascular disease.4 However, when Kennedy et al7 evaluated the feasibility of patients attaining these lipid goals, only 49.4% had LDL-C concentrations less than 100 mg/dL, and of 191 patients considered at very high risk, only 30 (15.7%) achieved the optional LDL-C goal of less than 70 mg/dL. Kennedy et al calculated that 25% of patients require more than 2 lipid-lowering drugs at maximal doses to attain this goal.
Colesevelam hydrochloride (Welchol; Daiichi Sankyo Inc, Parsippany, New Jersey) is a specifically engineered bile acid sequestrant that reduces LDL-C concentrations in patients with hypercholesterolemia.8 In a previous study, the bile acid sequestrant cholestyramine decreased levels of HbA1c in patients with T2DM.9 Similar observations were noted with colesevelam treatment in a post hoc analysis of data from patients with T2DM included in a primary lipid trial.10 In addition, recent evidence has revealed a link between bile acid and glucose homeostasis, which suggests that bile acid sequestrants may improve glycemic control via alterations in the bile acid pool.11 To prospectively evaluate the glucose-lowering effect of colesevelam treatment, a 12-week double-blind, placebo-controlled, pilot study was conducted in patients with T2DM inadequately controlled with oral antidiabetes agents (metformin, sulfonylurea, or both).12 The addition of colesevelam treatment significantly reduced the levels of HbA1c (by 0.50%; P = .007) and LDL-C (by 11.7%; P = .007) compared with placebo. The present study evaluated the glucose-lowering effects of colesevelam treatment in subjects with T2DM not adequately controlled with a stable regimen of insulin alone or insulin in combination with oral antidiabetes agents.
This was a prospective, 16-week, multicenter, randomized, double-blind, placebo-controlled, parallel-group study conducted at 50 sites in the United States and 1 site in Mexico between August 12, 2004, and December 28, 2005. The study protocol was conducted in compliance with institutional review board regulations, good clinical practice guidelines, and the fourth amendment to the Declaration of Helsinki. All subjects provided written informed consent before participation. An assessment of race/ethnicity was obtained from each subject. Figure 1 shows subject disposition throughout the trial.
The study enrolled subjects aged 18 to 75 years with T2DM not adequately controlled (baseline HbA1c level, 7.5%-9.5%, inclusive) with insulin alone or in combination with oral antidiabetes agents (a biguanide [metformin hydrochloride], a biguanide-sulfonylurea combination [metformin-glibenclamide], a sulfonylurea [glibenclamide, glimepiride, or glipizide], a thiazolidinedione [pioglitazone hydrochloride or rosiglitazone maleate], or a meglitinide [nateglinide or repaglinide]). Subjects were receiving a stable (±10%) dose of insulin (30-200 U/d) for 6 weeks or longer before screening (those receiving oral antidiabetes agents were required to receive a stable dose for 90 days or longer before screening) and had C-peptide levels greater than 0.5 ng/mL (to convert to nanomoles per liter, multiply by 0.331), LDL-C concentration of 60 mg/dL or greater, and triglyceride levels of 500 mg/dL or less (to convert to millimoles per liter, multiply by 0.0113). All subjects were to have been prescribed a diet accepted by the American Diabetes Association, although no specific, protocol-directed dietary evaluation or dietary recommendations were made during the trial.
Exclusion criteria included a body mass index (calculated as weight in kilograms divided by height in meters squared) of greater than 45; uncontrolled hypertension (systolic blood pressure >160 mm Hg, diastolic blood pressure >95 mm Hg, or both); acute coronary syndrome (eg, myocardial infarction or unstable angina), coronary intervention (coronary artery bypass grafting, percutaneous transluminal coronary angioplasty, or similar procedure), or transient ischemic attack within 3 months of screening; or a history of pancreatitis, ketoacidosis, type 1 diabetes mellitus, intestinal motility disorders, severe peripheral vascular disease, dysphasia or other swallowing disorders, or AIDS or human immunodeficiency virus infection. Oral corticosteroids, cholestyramine resin, and colestipol hydrochloride were prohibited medications. Subjects were withdrawn from the study if the HbA1c level increased to 10.0% or greater or fasting plasma glucose (FPG) levels increased to 260 mg/dL or greater (to convert to millimoles per liter, multiply by 0.0555) during the double-blind treatment period. The management, reporting, and actions taken in response to hypoglycemia (FPG <50 mg/dL with or without symptoms or FPG <70 mg/dL with symptoms) were left to the medical judgment of the blinded study investigators (R.B.G. and V.A.F.). Discharge from the study was considered only after repeated hypoglycemic episodes and if glucose levels could not be controlled with dietary changes or reductions in insulin dosage.
After a 1-week screening period, subjects entered a single-blind, 2-week placebo run-in period during which they took 6 placebo tablets per day, composed of magnesium stearate and microcrystalline cellulose, with a commercially supplied film-coating mixture. After the placebo run-in period, subjects were randomized 1:1 to receive colesevelam hydrochloride, 3.75 g/d (6 × 625-mg tablets) or placebo for the 16-week double-blind treatment period. Colesevelam hydrochloride was administered either as 1.875 g twice daily with the noon and evening meals or 3.75 g once daily with the evening meal; subjects chose their preferred dosing regimen. However, the dosing regimen initiated during the placebo run-in period was to be maintained throughout the double-blind study period. Subjects continued prestudy insulin and oral antidiabetes agent dosages throughout the clinical trial. Investigators were instructed not to vary the insulin dosage by more than ±10% of the initial dosage.
The primary efficacy measure was the mean change in HbA1c level from baseline to week 16, with subjects analyzed on an intent-to-treat basis, using a last observation carried forward (LOCF) approach. Secondary efficacy measures included the mean change in FPG, fructosamine, and HbA1c levels from baseline to weeks 4, 8, and 16; the arbitrary predefined assessment (based on efficacy data for HbA1c and FPG levels from the pilot study with colesevelam12) of a glycemic-control response: a reduction in the FPG level of 30 mg/dL or greater or a reduction in the HbA1c level of 0.7% or greater from baseline by week 16; mean change in C-peptide levels from baseline to week 16; mean change and mean percent change in concentrations of total cholesterol (TC), LDL-C, high-density lipoprotein cholesterol (HDL-C), non–HDL-C, triglycerides, and apolipoprotein A-I (ApoA-I) and apolipoprotein B (ApoB) levels and in ratios of TC/HDL-C, LDL-C/HDL-C, non–HDL-C/HDL-C, and ApoB/ApoA-I from baseline to week 16; and median change and median percent change in levels of high-sensitivity C-reactive protein and triglycerides from baseline to week 16.
All standard blood samples were obtained under fasting conditions, and tests were performed by a certified laboratory (Medical Research Laboratories International, Highland Heights, Kentucky). Blood samples were collected at weeks −3 (screening), 0 (randomization baseline), and 4, 8, and 16 or at an early termination visit, if applicable. The TC concentration and triglyceride levels were measured using enzyme assays. High-density lipoprotein cholesterol concentration was measured using a cholesterol oxidase assay of the supernate from the precipitate of non-HDL lipoproteins with heparin and manganese chloride. Low-density lipoprotein cholesterol concentration was calculated using the Friedewald equation in subjects with triglyceride levels of 400 mg/dL or lower or measured directly by the Lipid Research Clinic beta-quantification method13 in subjects with triglyceride levels higher than 400 mg/dL at screening (week −3). Apolipoprotein A-I and ApoB, and high-sensitivity C-reactive protein levels were quantitated using immunonephelometry.
Safety assessments included changes in vital signs, findings at physical examinations, treatment-emergent adverse events, and clinical laboratory test results. Urine samples were collected for analysis of kidney function at weeks −3 (screening), 0 (randomization baseline), 8, and 16 or at an early termination visit, if applicable. A 24-hour urine sample was not collected as part of the study. Compliance with the medication regimen was evaluated by tablet counts at each visit.
The total intent-to-treat population included all randomized subjects who took at least 1 dose of randomized study medication and had a baseline measurement and 1 or more postbaseline HbA1c or FPG measurements. Imputation of missing data was by LOCF. We present analyses of both LOCF and non-LOCF sets. Sensitivity analysis was performed based on the intent-to-treat and per-protocol populations for both week 16 completers (non-LOCF) and week 16 LOCF. For follow-up analyses, 2 mutually exclusive subject cohorts were defined: the insulin monotherapy cohort and the insulin combination therapy cohort. Subjects who were receiving only insulin in addition to study medication during the trial comprised the monotherapy cohort; the insulin combination therapy cohort included all subjects who received an oral antidiabetes agents in addition to insulin therapy. The safety population included all randomized subjects who received at least1 dose of randomized study medication. All statistical tests were considered significant at P < .05 (2-sided).
The study required a total of 260 randomized subjects, 130 in each group, and had 81% to more than 95% power to detect a difference of 0.58% to 0.80% reduction in mean HbA1c level from baseline between colesevelam treatment and placebo (with a 2-sided type I error at .05), assuming a common SD of 1.5% or less and a maximum dropout rate of 15%. The power in terms of percentage was calculated on the basis of the change in HbA1c units (not percentage obtained by dividing change by baseline) because the sample size calculation and power estimation were based on the primary efficacy analysis.
Differences between treatment groups in age, weight, height, FPG and HbA1c levels, and body mass index at baseline were assessed using an analysis of variance model with treatment as a factor. Sex, race/ethnicity, and concomitant antidiabetes medication categories were tested using the Fisher exact test or the Fisher-Freeman-Halton test.14
An analysis of covariance (ANCOVA) model with treatment and concomitant antidiabetes medication status (insulin alone or insulin in combination with oral antidiabetes agents) as fixed effects and baseline HbA1c level as a covariate was used to evaluate the treatment effect. The normality assumption of the efficacy data was examined before fitting the ANCOVA models. When a significant departure from normality was observed, a nonparametric equivalent of ANCOVA (rank ANCOVA) was applied.
The treatment effect was estimated using least squares (LS) means and standard error of the mean, 2-tailed 95% confidence intervals, and the 2-sided P value. Secondary efficacy parameters were similarly compared unless otherwise noted.
Median changes and median percent changes in high-sensitivity C-reactive protein and triglyceride levels from baseline to week 16 with and without LOCF imputation were analyzed using a nonparametric ANCOVA model. The treatment difference was estimated using the estimator of Hodges and Lehmann,15 and a 2-tailed 95% confidence interval for the treatment difference was obtained using the method of Moses et al.16
The glycemic-control response rate was tabulated and compared between treatment groups using the Pearson χ2 test. The statistical analysis plan was finalized before data unblinding. Commercially available software (SAS Institute Inc, Cary, North Carolina) was used to analyze all of the efficacy and safety data.
The study was conducted between August 12, 2004, and December 28, 2005. Of 785 subjects screened, 287 met study criteria for randomization (Figure 1). Seven subjects (4.8%) in the colesevelam-treated group and 9 (6.4%) in the placebo group were withdrawn because of protocol-specified criteria for hyperglycemia. Ten subjects (6.8%) in the colesevelam-treated group and 4 (2.9%) in the placebo group withdrew during the randomized treatment period because of an adverse event. Thirteen subjects (8.8%) in the colesevelam-treated group and 14 (10.0%) in the placebo group withdrew because of other reasons. One hundred seventeen subjects (79.6%) in the colesevelam-treated group and 114 (81.4%) in the placebo group completed the study.
Baseline demographic characteristics did not differ between the 2 groups (Table 1). Participants were primarily overweight white individuals with a mean age of 57 years. A total of 116 subjects (40.4%) were receiving insulin alone, and 171 subjects (59.6%) were receiving a combination of insulin and an oral antidiabetes agent. In the cohort receiving insulin alone, mean baseline FPG levels were similar for subjects in the colesevelam-treated and placebo groups (164.8 vs 166.8 mg/dL; P = .84). However, in the cohort receiving a combination of insulin and an oral antidiabetes agent, the mean baseline FPG level was significantly higher in the colesevelam-treated group compared with the placebo group (164.7 vs 142.9 mg/dL; P = .007).
Overall compliance with randomized study medication was 92.7% in the colesevelam-treated group and 94.5% in the placebo group. The mean (SD) change in daily insulin dose for the intent-to-treat population was not significant (−0.3 [7.73] U in the colesevelam-treated group and 1.4 [10.39] U in the placebo group). Throughout the course of the study, the daily insulin dosage was changed in 21 subjects in the colesevelam-treated group and 22 subjects in the placebo group.
Treatment with colesevelam for 16 weeks significantly reduced the primary end point of the mean HbA1c level compared with placebo. Least squares mean change in the HbA1c level from baseline to week 16 (LOCF) was −0.41% (0.07%) in the colesevelam-treated group and 0.09% (0.07%) in the placebo group, resulting in a LS mean treatment difference of −0.50% (P < .001; Table 2). A treatment difference in HbA1c level (based on observed data only, ie, without LOCF) was apparent after 4 weeks and was maintained throughout the study (Figure 2A). The similarity in the treatment differences based on week 16 completers (non-LOCF) and week 16 LOCF revealed that no significant effect was noted from dropouts in this efficacy analysis. In addition, the reduction in HbA1c level in the total subject population was consistent across the cohorts receiving insulin alone or those receiving insulin in combination with oral antidiabetes agents (Table 2).
The total subject population was stratified into 2 subgroups to evaluate the effect of baseline HbA1c level on the response to the addition of colesevelam treatment: subjects with an HbA1c level of 8.0% or less at baseline and subjects with an HbA1c level greater than 8.0% at baseline. In the subgroup with an HbA1c level of 8.0% or less at baseline, the HbA1c level decreased in the colesevelam-treated group by 0.27% and increased by 0.11% in the placebo group, resulting in an LS mean treatment difference of −0.38% (P < .0007). A greater effect was observed in the subgroup with an HbA1c level greater than 8.0% at baseline; the mean change in HbA1c level in the colesevelam-treated group was −0.50% compared with 0.07% in the placebo group (LS mean treatment difference, −0.57%; P < .001).
A numerically greater reduction in FPG level from baseline to week 16 (LOCF) occurred in the colesevelam-treated group compared with the placebo group, although this was not significant (LS mean treatment difference, −14.6 mg/dL; P = .08); however, colesevelam treatment, compared with placebo, significantly reduced the FPG level at weeks 4, 8, and 16 (−15.1, −17.2, and −23.6 mg/dL, respectively) (Table 3 and Figure 2B). Colesevelam significantly reduced mean fructosamine levels from baseline to weeks 4, 8, and 16 compared with placebo (Figure 2C) (LS mean treatment difference, −21.7 μmol/L [P < .001] at week 16 LOCF).
In the total intent-to-treat population, 70 subjects in the colesevelam-treated group (48.6%) and 43 subjects in the placebo group (31.6%) had a glycemic-control response, that is, a reduction in the FPG level of 30 mg/dL or higher or a reduction in the HbA1c level of 0.7% or greater from baseline, by week 16 (LOCF; P = .004). More than twice as many subjects in the colesevelam-treated group had a reduction in the HbA1c level of 0.7% or greater compared with those in the placebo group (34.7% vs 14.0%; P < .001). However, no significant difference was noted in the percentage of subjects achieving a reduction in FPG level of 30 mg/dL or higher between the colesevelam-treated and placebo groups at week 16.
Mean change from baseline in C-peptide levels was similar in both groups. No significant LS mean treatment difference was evident at week 16 LOCF (P = .65).
Treatment with colesevelam resulted in a significantly greater percentage reduction in LDL-C concentration than did placebo at week 16 LOCF (P < .001; Table 4 and Figure 3). A significant increase in triglyceride levels was observed in the colesevelam-treated group compared with the placebo group. The LS median percent change and median change in triglyceride levels from baseline to week 16 LOCF for the colesevelam-treated and placebo groups were 22.7% vs 0.3% and 32.0 mg/dL vs −1.3 mg/dL (Figure 3), respectively (P < .001 for both).
Treatment with colesevelam significantly reduced ApoB levels by 5.3% compared with placebo (P = .04) but did not result in a significant increase in ApoA-I after 16 weeks (Figure 3). Colesevelam treatment resulted in a significant decrease in LDL-C/HDL-C and ApoB/ApoA-I ratios but not in the TC/HDL-C or non–HDL-C/HDL-C ratios (Figure 4).
The LS median treatment difference in change in high-sensitivity C-reactive protein levels between the colesevelam-treated and placebo groups was not significant at week 16 LOCF (−0.40 mg/L; P = .13). This difference represented changes of −3.0% in the colesevelam-treated group and 8.7% in the placebo group at week 16 (P = .07 for the treatment difference). However, the treatment difference was significant at 16 weeks in the cohort of subjects who completed the study (−0.60 mg/L [P = .01] and −18.6% [P = .003]).
Overall, colesevelam treatment was safe and well tolerated by subjects with T2DM receiving a regimen that included insulin either alone or in combination with oral antidiabetes agents. Treatment-emergent adverse events (hereinafter referred to as “adverse events”) occurred in 92 colesevelam and 82 placebo recipients during the randomized trial period (Table 5); most were mild to moderately severe. Fourteen subjects withdrew from the study because of an adverse event, 10 in the colesevelam-treated group and 4 in the placebo group (One discontinued for an adverse event that began during the run-in period and was not considered “treatment emergent.”).
Twenty-four subjects in the colesevelam-treated group and 13 in the placebo group experienced an adverse event considered by the blinded investigator (R.B.G. and V.A.F.) to be definitely, probably, or possibly related to study medication (Table 5). Seven subjects withdrew because of a drug-related adverse event, 5 in the colesevelam-treated group and 2 in the placebo group. In the colesevelam-treated group, the most frequently reported (incidence >1%) drug-related adverse events were constipation (6.8%), dyspepsia (3.4%), hypoglycemia (3.4%), flatulence (2.0%), and nausea (1.4%), and in the placebo group, the most frequently reported adverse event was hypoglycemia (5.7%). The incidence of hypoglycemia (as assessed by self-report) was similar in the treatment groups. No subject was removed from the study because of hypoglycemia.
No drug-related serious adverse events occurred in either group (Table 5). Two subjects in the colesevelam-treated group and 1 in the placebo group withdrew from the study because of a serious adverse event, including pneumonitis and pneumonia (1 subject each) in the colesevelam-treated group and staphylococcal infection (1 subject) in the placebo group.
Mean changes in weight, vital signs, and safety laboratory factors during randomized treatment were similar for colesevelam-treated and placebo groups. No significant weight gain was noted in either group by study end; weight modestly increased by 0.6 (SD, 3.29) kg in subjects receiving colesevelam and 0.2 (SD, 2.44) kg in subjects receiving placebo by week 16.
The progressive nature of T2DM makes management of the disease difficult, and approximately two-thirds of individuals with T2DM fail to achieve HbA1c levels less than 7.0%.17 Most patients who begin therapy with oral antidiabetes agents eventually require add-on insulin therapy to maintain glycemic control,18-21 and many patients who require insulin will benefit from combination therapy with oral antidiabetes agents because of poor glycemic control with insulin therapy alone.17,22,23 The major new finding in this study is that the addition of colesevelam to established insulin therapy in subjects with T2DM leads to improvement in glycemic control as evidenced by a significant mean reduction in HbA1c level (−0.50%) compared with placebo and a significant increase in the glycemic-control response rate after 16 weeks. The improvement in glycemic control was rapid because both the FPG and fructosamine levels were significantly reduced compared with those in the placebo control group by week 4 of treatment. The overall glycemic effect with colesevelam treatment is similar to that reported when oral antidiabetes agents, including vildagliptin (−0.30%) or pioglitazone (−0.55%), were added to insulin therapy.17,24 This mean 0.50% reduction in the HbA1c level may be clinically beneficial by bringing it closer to the glycemic goal of an HbA1c level less than 7.0% recommended by the American Diabetes Association.4 Furthermore, as the United Kingdom Prospective Diabetes Study demonstrated, every 1.0% reduction in HbA1c level yields a 37% reduction in microvascular complications and a 14% reduction in all-cause mortality and myocardial infarctions.25
To determine the true effect of colesevelam treatment without confounders, subjects were intentionally included who were receiving stable doses of insulin and oral antidiabetes agents, and investigators were asked not to make changes except for safety reasons. The mean dosage of insulin was unchanged during the study, which indicates that the glycemic improvement was the result of the added colesevelam. In practice, patients frequently continue to receive fixed dosages of insulin for long periods despite poor glycemic control.26 The reason for this clinical inertia is multifactorial and relates in part to both physician-patient barriers to increasing the dosage of insulin. In this respect, the ability to improve glycemic control without changing insulin dosages may be an advantage. The improved glycemic control was achieved without an evident increase in reported episodes of hypoglycemia or weight gain. Most studies of the addition of oral antidiabetes therapy to insulin therapy have demonstrated either increased rates of hypoglycemia or weight gain.17,24,27,28 In addition, the glucose-lowering effect with colesevelam in the insulin combination therapy cohort occurred regardless of whether insulin was combined with a biguanide, sulfonylurea, or thiazolidinedione.
The exact mechanisms through which colesevelam exerts an effect on glycemic control are not well understood. It is possible that bile acid sequestrants act in the gastrointestinal tract to either reduce the amount of glucose absorbed or change the time course of glucose absorption.9 Alternatively, colesevelam may influence glucose metabolism by binding to bile acids, thereby disrupting the enterohepatic pathway of bile acid metabolism and altering the makeup of the circulating bile acid pool.29 Certain bile acids are endogenous ligands of the farnesoid X receptor (FXR), a member of the nuclear receptor superfamily of ligand-activated transcription factors.11,30 The FXR has established functions in bile acid and cholesterol metabolism and has also been implicated in glucose metabolism.11,30,31 The FXR regulates various downstream receptors including the fibroblast growth factor 15/19 (FGF15/19) in the intestine and the small heterodimer partner (SHP) in the liver. Evidence suggests that altering the FXR-FGF15/19-SHP pathway can improve glycemic control; both FGF15 knockout mice and SHP knockout mice have reduced glucose levels.32 Furthermore, cholestyramine has been shown to reduce FGF15/19.33 Similarly, colesevelam may improve glycemic control via alteration to the FXR-FGF15/19-SHP pathway.
In addition to effects on glycemic control factors, colesevelam treatment for 16 weeks significantly reduced TC and LDL-C concentrations compared with placebo. The efficacy of colesevelam treatment on LDL-C concentration in patients with T2DM in this study (−12.8%) was consistent with LDL-C lowering in patients without diabetes in a previous study with colesevelam (−15.0%).34 Other than statins, bile acid sequestrants are the only class of LDL-C–lowering drugs that have been demonstrated in a controlled clinical trial to reduce coronary heart disease events. Cholestyramine treatment in the Lipid Research Clinics Coronary Primary Prevention Trial was associated with an approximate 1% reduction in coronary events for each 1% lowering of the concentration of LDL-C.35 Individuals with T2DM are considered to have a coronary heart disease risk equivalent,4 and almost 75% do not achieve LDL-C concentrations lower than 100 mg/dL.36 Thus, long-term treatment with colesevelam in subjects with T2DM would be expected to contribute to reducing the elevated risk of coronary heart disease.37 Colesevelam significantly reduced ApoB levels and LDL-C/HDL-C and ApoB/ApoA-I ratios. Median triglyceride levels were significantly increased by week 16 in subjects receiving colesevelam, consistent with data from other studies with bile acid sequestrants. The clinical implications of modest triglyceride level increases are unknown, especially inasmuch as bile acid sequestrants reduce cardiovascular events through documented reductions in LDL-C concentration despite elevations in triglyceride levels.35
The mechanism by which bile acid sequestrants lower LDL-C concentrations is thought to involve inactivation of FXR, which leads to upregulation of hepatic bile acid synthesis and LDL receptor activity, increasing fractional LDL clearance from plasma. In addition, reduced FXR expression seems to be accompanied by increased hepatic triglyceride production and reduced plasma clearance, explaining the tendency for triglyceride levels to rise with this treatment.11 The use of bile acid sequestrants in the treatment of both hyperglycemia and elevated LDL-C concentrations thus comprises a novel therapeutic approach for T2DM.
In the present trial, colesevelam was weight neutral and well tolerated, and the overall results of this study suggest a favorable safety profile in subjects with T2DM who are receiving insulin therapy. The most common adverse event was gastrointestinal disturbance, and no patient had a serious adverse event related to hypoglycemia.
Treatment with colesevelam seems to be safe and efficacious when used in combination with insulin for improving glycemic control and lipid management in patients with T2DM, which may lead to a greater percentage of patients meeting glycemic goals and achieving a further reduction in LDL-C concentrations, 2 critical factors in the management of T2DM. The capability of improving both glycemic control and primary lipid management in individuals with T2DM with 1 medication constitutes an advantage over currently available single-use agents.
Correspondence: Ronald B. Goldberg, MD, Lipid Disorders Clinic, Division of Endocrinology Diabetes and Metabolism, and Diabetes Research Institute, University of Miami Miller School of Medicine, 900 NW 17th St, Miami, FL 33136 (email@example.com).
Accepted for Publication: January 13, 2008.
Author Contributions: Dr Goldberg had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Truitt and Jones. Acquisition of data: Goldberg and Truitt. Analysis and interpretation of data: Goldberg, Fonseca, and Truitt. Drafting of the manuscript: Fonseca and Truitt. Critical revision of the manuscript for important intellectual content: Goldberg, Fonseca, Truitt, and Jones. Obtained funding: Truitt. Administrative, technical, and material support: Truitt and Jones. Study supervision: Goldberg and Truitt.
Financial Disclosure: Dr Goldberg has been a speaker and received honoraria from Eli Lilly, Takeda, Pfizer, Merck, Merck/Schering Plough, KOS, AstraZeneca, and Abbott; has received research grants from Novo Nordisk, Pfizer, Merck, KOS, AstraZeneca, and Daiichi Sankyo; and has been a consultant and received honoraria from Eli Lilly/Takeda, Pfizer, Merck, Merck/Schering Plough, AstraZeneca, and Abbott. Dr Fonseca has received support from the American Diabetes Association; has served as a paid consultant to Daiichi Sankyo; and served as an investigator and consultant to Glaxo Smith Kline, Novartis, Takeda, AstraZeneca, Pfizer, Sanofi-Aventis, Eli Lilly, Novo Nordisk, Minimed, and Viaject. Drs Truitt and Jones are employees of and hold stock options in Daiichi Sankyo.
Funding/Support: This study was supported in part by grants 5M01RR05096 and RR-00827 from the National Institutes of Health (Dr Fonseca) and by Daiichi Sankyo.
Role of the Sponsor: Daiichi Sankyo designed and conducted the study and had a role in interpretation of the data and in preparation and approval of the manuscript.
Additional Contributions: Karen Stauffer, PhD, provided editorial assistance.