Methods

Top Methods Results Discussion Author & Article Info References

Patients

We studied 21 patients (20 men and 1 woman) in the metabolic ward of the Department of Veterans Affairs Medical Center, Dallas, Texas. Eleven patients were white, 7 were African-American, and 3 were Mexican-American. All participants had NIDDM as defined by the criteria of the National Diabetes Data Group [8]. Their ages ranged from 41 to 74 years (mean ±SD, 61.8 ±8.4 years). Their mean (±SD) body mass index was 28.4 ±2.5 kg/m² (range, 23.5 to 31.9 kg/m²). The protocol for this study was approved by the appropriate institutional review board, and each patient gave informed consent.

Twelve patients were receiving glyburide (2.5 to 15 mg/d) and 9 were receiving insulin therapy (24 to 168 units/d) to control their glycemia. Patients maintaining stable and "good" glycemic control for at least 1 month before the study were considered eligible to participate. The criteria defining "good" glycemic control included a mean plasma glucose concentration, measured at 3, 7, and 11 a.m. and at 4 and 8 p.m., ranging from 3.9 mmol/L to 7.8 mmol/L (70 to 140 mg/dL), and a glycosylated hemoglobin concentration ranging from 4% to 10% (normal range, 4% to 8%). At entry into the study, all patients had fasting LDL cholesterol levels greater than 3.34 mmol/L (130 mg/dL) and plasma triglyceride levels less than 3.4 mmol/L [300 mg/dL]. We divided patients into two groups: 1) those with fasting plasma triglyceride levels less than 1.7 mmol/L [less than 150 mg/dL, normotriglyceridemic] and 2) those with triglyceride levels between 1.7 and 3.4 mmol/L (150 to 300 mg/dL, borderline hypertriglyceridemic). Each group had a separate randomization schedule, and we planned to recruit about an equal number of participants in the two groups. Therapy with lipid-lowering medications, if any, was discontinued at least 1 month before the baseline period. No patient was taking ß-adrenergic blocking agents or diuretics. Other concomitant therapy included aspirin, dipyridamole, calcium channel blocking agents, and amitriptyline; the dose and frequency of administration of these drugs were held constant throughout the study. No patient had thyroidal, renal, or hepatic disease.

We advised patients to follow an isocaloric diet according to the recommendations of the American Diabetes Association throughout the study [9]. We estimated the daily energy intake needed for weight maintenance by determining basal energy expenditure using the Harris-Benedict equations and multiplying the basal energy expenditure by an activity factor [10]. The diet consisted of 15% of total energy from proteins, 55% from carbohydrates, and 30% from fats (< 10% saturated fats). Cholesterol content of the diet was less than 300 mg/d. We advised patients not to drink alcohol during the study period.

Study Design

We studied all patients during three hospitalizations, each lasting 5 days. Before starting the drug trial, patients were hospitalized in the metabolic unit for 5 days, the baseline period, and glycemic control was assessed, adjustments were made in the dose of glyburide or insulin, if needed, and energy intake needed to maintain constant body weight was estimated. During the hospital stay, routine screening laboratory tests were obtained. Plasma glucose was determined at 3, 7, and 11 a.m. and at 4 and 8 p.m. daily. For 3 consecutive days, 24-hour urine output was collected to determine glucose and creatinine concentrations. Blood was also collected once to determine glycosylated hemoglobin values. Fasting blood was obtained daily to determine plasma lipid and lipoprotein levels.

After the baseline period, patients were given cholestyramine (Questran Light; Bristol-Myers, Evansville, Indiana) or identical placebo, two packets twice daily within one-half hour of the morning and evening meals for 6 weeks, in a double-blind, randomized, crossover manner. Each 5-g packet of the active drug preparation contained 4 g of anhydrous cholestyramine resin. In the placebo, cholestyramine was replaced with cellulose, gelatin, and silicon dioxide. Sucrose contents of the active drug preparation and the placebo were identical. Patients were given 150 packets of cholestyramine or placebo during each study period. Compliance of the patients with medication was assessed by counting the remaining number of packets at the completion of each study period. If a patient could not tolerate the full dose, the clinical staff could reduce the dose to 3 packets per day after consulting with us. We allowed no further reductions in the dose.

During the study periods, patients were followed in the out-patient setting every 2 weeks and chemistry profiles were obtained and alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, -glutamyltransferase, lactate dehydrogenase, and glycosylated hemoglobin levels were determined. An inquiry was made about any side effects of medication, particularly constipation and stomach upset. All the patients were advised to take 100 mg of sodium docusate (Colace; Mead Johnson & Co., Evansville, Indiana), a stool softener, once or twice daily as needed for constipation. For stomach upsets, patients were instructed to take two teaspoonfuls of antacid (Maalox suspension; Rorer Consumer Pharmaceuticals, Fort Washington, Pennsylvania) with meals and at night. All other medications were taken at least 1 hour before consuming cholestyramine or placebo to avoid interference in absorption. The dose of insulin or glyburide was not changed during the study periods.

After 37 days of each study period, patients were admitted again to the metabolic unit for 5 days. During this period, all the studies mentioned in the baseline period were repeated.

Biochemical Analyses

Plasma glucose concentrations were assayed using the glucose oxidase method (Beckman Glucose Analyzer; Beckman Instruments, Fullerton, California). Quantitative analyses of glycosylated hemoglobin were done with kits for agar-gel electrophoresis (Helena Laboratories, Beaumont, Texas). Serum chemistry profile and urinary glucose and creatinine concentrations were measured using Hitachi 747 (Boehringer-Mannheim/Hitachi, Indianapolis, Indiana).

Fasting plasma samples were analyzed for total cholesterol, triglyceride, and lipoprotein cholesterol levels according to the Lipid Research Clinic procedures [11], except that cholesterol and triglyceride levels were measured enzymatically using kits from Boehringer-Mannheim (Indianapolis, Indiana) and Sigma Diagnostics (St. Louis, Missouri), respectively. Briefly, very-low- density lipoproteins (VLDL) (density <1.006 kg/L) were removed by preparative ultracentrifugation, and cholesterol was measured in the VLDL subfraction and the infranatant. Plasma high-density lipoprotein (HDL) cholesterol was measured in the supernatant after lipoproteins containing apolipoprotein B were precipitated by heparin-manganese. Cholesterol in the LDL fraction was taken as the difference between the cholesterol contents of the 1.006 kg/L infranatant and the HDL fraction.

Statistical Analysis

We did repeated-measure analysis of variance to assess the effect of the order in which patients received placebo or cholestyramine and the overall effect of cholestyramine therapy [12]. We used the Friedman test for data not consistent with the hypothesis of normality. Using two grouping factors in the statistical model, we compared responses of patients with normal triglyceride levels with responses of patients with borderline hypertriglyceridemia and the responses of patients receiving insulin therapy with those of patients receiving glyburide therapy. Further, we assessed effects of baseline and other possible confounding variables using analysis of variance and covariance models. We did multiple comparisons using a two-tailed, paired t-test with the Bonferroni correction. We used the Wilcoxon signed-rank test for data not consistent with the hypothesis of normality. All the results are expressed as means. We used the McNemar test to compare frequency of side effects of medication. All analyses were done using CLINFO (BBN Software Product Corp., Cambridge, Massachusetts), BMDP (BMDP Statistical Software Inc., Los Angeles, California), and SYSSTAT (SYSTAT Inc., Evanston, Illinois) software packages.

Results

Top Methods Results Discussion Author & Article Info References

Cholestyramine was well tolerated and participant compliance was excellent. Two patients who completed the placebo period first dropped out within the first 2 weeks of starting cholestyramine therapy because of the side effects of the medication. Both the patients had severe constipation and nausea and reported having angina when straining at stools. In another patient, the dose of cholestyramine was reduced to three packets per day. A fourth patient took only one packet twice daily during the cholestyramine period without our knowledge. However, we did not exclude the data on these patients from analysis. Incidence of constipation and use of sodium docusate was greater with cholestyramine therapy. Six of the 21 patients had constipation during both the placebo and cholestyramine periods, but 8 patients reported constipation only during the cholestyramine period (P = 0.005). Six patients used sodium docusate (1 to 3 tablets/d) during both the placebo and cholestyramine periods, and 7 patients used the drug (1 or 2 tablets/d) only during the cholestyramine period (P = 0.008). Other side effects of cholestyramine therapy were limited to flare-up of preexisting heartburn (1 patient), nausea (2 patients), and transient increase in serum transaminase levels (1 patient). Compared with placebo, cholestyramine therapy increased serum alkaline phosphatase values (mean values, 78 compared with 86 units; P = 0.02); however, values did not increase beyond the normal range in any of the patients. We observed no change in other hepatic enzymes.

The order in which the patients received cholestyramine or placebo had no effect on the results. Further, baseline variables such as age, body mass index, and parameters of glycemic control did not affect the outcome. Values for all variables were similar during the baseline and placebo periods (Tables 1 and 2). Further, no differences were observed in the response to cholestyramine therapy between patients receiving glyburide and those receiving insulin or between normotriglyceridemic and borderline hypertriglyceridemic patients; therefore, we pooled the data of all patients for all metabolic variables. The results for each patient are shown in Figures 1, 2, and 3 and are summarized for all the patients in Tables 1 and 2.

View larger version (19K): [in this window] [in a new window]   Figure 1. Plasma levels of total cholesterol, low-density lipoprotein, and high-density lipoprotein measured when patients received placebo and cholestyramine. Each symbol represents the mean of five determinations. \#9679; =patients with normotriglyceridemia (fasting plasma triglyc-erides <1.7 mmol/L [150 mg/dL]), and cir; = patients with borderline hypertriglyceridemia [fasting plasma triglyceride level of 1.7 to 3.4 mmol/L (150 to 300 mg/dL)]. To convert cholesterol values to mg/dL, multiply by 38.674.

View larger version (22K): [in this window] [in a new window]   Figure 2. Plasma levels of triglycerides and very-low-density lipoprotein cholesterol measured when patients received placebo and cholestyramine. Each symbol represents the mean of five determinations. \#9679; =patients with normotriglyceridemia (fasting plasma triglycerides <1.7 mmol/L [150 mg/dL]), and cir; = patients with borderline hypertriglyceridemia [fasting plasma triglyceride level of 1.7 to 3.4 mmol/L (150 to 300 mg/dL)]. To convert cholesterol and triglycerides values to mg/dL, multiply by 38.674 and 88.574, respectively.

View larger version (19K): [in this window] [in a new window]   Figure 3. Levels of plasma glucose, urinary glucose, and blood glycosylated hemoglobin measured when patients received placebo and cholestyramine. Each symbol for plasma glucose represents the mean of five daily determinations at 3, 7 and 11 a.m. and at 4 and 8 p.m. for 5 consecutive days. Each symbol for urinary glucose represents the mean of three consecutive 24-hour collections. Each symbol for glycosylated hemoglobin represents the mean of two determinations on days 28 and 38. \#9679; =patients with normotriglyceridemia (fasting plasma triglycerides <1.7 mmol/L [150 mg/dL]), and cir; = patients with borderline hypertriglyceridemia [fasting plasma triglyceride level of 1.7 to 3.4 mmol/L (150 to 300 mg/dL)]. To convert plasma glucose values to mg/dL, multiply by 18.

Compared with placebo, cholestyramine therapy resulted in an 18% decrease in plasma total cholesterol (95% CI, 14% to 22%; P < 0.001) and a 28% decrease in plasma LDL cholesterol (CI, 21% to 35%; P < 0.001) Table 2, Figure 1. We observed an average LDL cholesterol reduction of 1.13 mmol/L (44 mg/dL) with cholestyramine therapy. Cholestyramine therapy caused an overall 13.5% increase in plasma triglyceride concentrations (CI, 1% to 26%; P = 0.02) (Figure 2). Marked hypertriglyceridemia (plasma triglyceride levels >5.6 mmol/L or 500 mg/dL) developed in none of the participants with cholestyramine therapy. The increase in plasma triglyceride concentrations was not accompanied by an increase in plasma VLDL cholesterol concentrations, which essentially remained unchanged (P > 0.2). High-density lipoprotein cholesterol levels did not change statistically (Figure 1).

The dose of glyburide or insulin was not changed during the placebo and the cholestyramine periods in accordance with the protocol (Table 1). Body weight remained constant during the study. Cholestyramine therapy was associated with improved glycemic control. Compared with placebo, cholestyramine therapy decreased mean plasma glucose concentrations by 13% (CI, 5% to 21%; P = 0.003) and urinary glucose excretion decreased from a median value of 0.58 g/d (range, 0.09 to 20.14 g/d) to 0.18 g/d (range, 0.04 to 5.44 g/d), a median change of 0.22 g/d (CI, 0.11 to 1.79 g/d; P < 0.001) (Figure 3). We observed no change in the mean urinary creatinine excretion, which was 1.504 g/d compared with 1.444 g/d during the placebo and cholestyramine therapy periods, respectively. Mean glycosylated hemoglobin concentrations tended to be lower by 0.5% (CI, 0.2% to –1.3%;P = 0.17) with cholestyramine therapy compared with placebo.

Discussion

Top Methods Results Discussion Author & Article Info References

The efficacy of bile acid sequestrants to treat patients with primary hyperlipoproteinemias has been known for the last three decades [13, 14]. In nondiabetic persons with primary hypercholesterolemia, reductions in total and LDL cholesterol levels using cholestyramine therapy have been shown to decrease the incidence of coronary heart disease [6]. A recent report provided additional evidence of the safety of long-term therapy with cholestyramine [15]. Despite the extensive literature on bile acid sequestrants used in persons without diabetes, adequate data on their efficacy in treating dyslipidemias in persons with diabetes mellitus are generally lacking. Two reports are available on the use of colestipol in patients with hypercholesterolemia and diabetes mellitus [16, 17]. These studies, however, were not systematic and well controlled and did not measure various lipoprotein fractions [16, 17].

One reason for previously not using bile acid sequestrants in patients with NIDDM is an increased frequency of elevated triglyceride levels in such patients. Of note, however, is that the prevalence of hypertriglyceridemia in patients with NIDDM is not nearly as high as many believe. For example, epidemiologic studies of patients with NIDDM report a prevalence of hypertriglyceridemia (>2.8 mmol/L or >250 mg/dL) of only 19.2% (the World Health Organization Multinational Study [18]), of 11.3% (the Diabetes Intervention Study, GDR [19]), of approximately 17% to 19% or less (the Framingham Heart Study [1]), and of approximately 23.5% (San Antonio Heart Study [2]). The Prospective Cardiovascular Munster Study from Munster, West Germany [3], showed that 40.4% of male patients and 61.9% of female patients with NIDDM have borderline or high cholesterol levels (>5.2 mmol/L or >200 mg/dL) in the absence of hypertriglyceridemia defined as a fasting plasma triglyceride level less than 2.26 mmol/L (200 mg/dL). The San Antonio Heart Study [2] also noted elevated LDL cholesterol levels (>3.34 mmol/L) in 43.3% of patients with NIDDM. Therefore, patients in our study having LDL cholesterol levels greater than 3.34 mmol/L and triglyceride levels less than 3.4 mmol/L are typical of most patients with NIDDM.

Our study showed that cholestyramine therapy is surprisingly efficacious in decreasing LDL cholesterol levels in patients with NIDDM. Low-density lipoprotein cholesterol levels were decreased by an average of 28% with cholestyramine therapy. A similar reduction (28%) in LDL cholesterol levels was observed with lovastatin therapy (20 mg twice daily) in our previous study, although different patients participated in the two studies [20]. Previous studies [21, 22] indicate that statins generally are considerably more efficacious in LDL cholesterol lowering than are bile acid sequestrants. The finding of similar efficacy of these two drugs in the same laboratory was unexpected, and the results suggest that bile acid sequestrants may have more utility in treating dyslipidemia in patients with NIDDM than previously suspected.

It is interesting to speculate why cholestyramine produced a greater reduction than expected in LDL cholesterol levels in patients with NIDDM. One possibility is that compared with other drugs, bile acid sequestrants produce a greater depletion of cholesterol from LDL particles. This tendency has been noted previously [23, 24]. Further, several investigators have reported that LDL particles in patients with normal triglyceride levels and NIDDM are abnormally enriched in cholesterol [25-28], and bile acid sequestrants might be particularly effective in these patients. Further, investigators have proposed that cholesterol enrichment of VLDL and LDL may interfere with reverse cholesterol transport and thereby promote atherogenesis [29]. If this mechanism pertains, use of bile acid sequestrants could be particularly advantageous in patients with NIDDM.

A modest 13% increase (CI, 1% to 26%) in plasma triglyceride levels occurred with cholestyramine therapy in our study. However, a marked hypertriglyceridemia (fasting plasma triglycerides >5.6 mmol/L or 500 mg/dL) did not develop in any of the patients. Anecdotal cases of bile acid sequestrant-induced extreme hypertriglyceridemia have been reported by others [7, 16, 30]; such a response could lead to acute pancreatitis. However, such extreme increases in triglyceride levels were observed only in patients who already had marked hypertriglyceridemia and, in some cases, poorly controlled diabetes mellitus. In well-controlled trials, such as the Lipid Research Clinic Coronary Primary Prevention Trial [6] and the study by Levy and colleagues [31], only a modest 5% to 6% increase in fasting plasma triglyceride levels was reported with cholestyramine therapy in persons without diabetes. This modest increase in triglycerides did not offset the benefit of cholesterol lowering in persons without diabetes [6], and presumably it would not in patients with NIDDM. Our patients were carefully selected to have fasting plasma triglyceride levels less than 3.4 mmol/L (300 mg/dL). Further, at the time of entry all the patients had "good" glycemic control. Patients with NIDDM with fasting plasma triglyceride levels more than 3.4 mmol/L (or preferably >2.8 mmol/L) probably should not be treated with bile acid sequestrants. To further minimize the risk for severe hypertriglyceridemia, it may be necessary to confirm more than once that fasting plasma triglyceride levels are consistently less than 3.4 mmol/L before starting the drug therapy and to follow plasma triglyceride levels carefully during bile acid sequestrant therapy.

Although cholestyramine therapy increased plasma triglyceride levels, concentrations of VLDL cholesterol remained unchanged, suggesting a change in VLDL particle composition. Other investigators have also described an increase in VLDL triglyceride-to-cholesterol ratio and an increase in VLDL particle size with bile acid sequestrant therapy [32, 33]. Kinetic studies show that the increase in VLDL triglyceride content is primarily due to increased hepatic synthesis of triglycerides [33]. Some investigators suggested that the bile acid sequestrant-induced accentuation of hypertriglyceridemia and change in VLDL particle composition may be transitory. However, persistent but modest increases in plasma triglyceride levels have been documented clearly in long-term, placebo-controlled studies [6].

An unexpected finding of our study was improvement in glycemic control with cholestyramine therapy. Lowering of plasma glucose levels was observed in both insulin-treated and glyburide-treated patients. In the study by Duntsch [16], fasting blood glucose levels were not substantially affected by colestipol therapy. The study, however, was not placebo controlled, and the details of the dose of oral hypoglycemic drugs and of insulin were not provided. In the other study, Bandisode and Boshell [17] compared three participants given placebo with five patients treated with colestipol. The authors concluded that colestipol therapy did not affect glycemic control. Specific values of plasma glucose, however, were not provided.

How cholestyramine therapy decreased plasma glucose levels is not known. A few possibilities can be considered. First, bile acid sequestrants might cause fat malabsorption [34], which may lead to weight reduction and thus improvement in glycemic control. In our study, fecal fat was not measured. In a previous study [34], however, no increase in fecal fat excretion was noted in two participants given cholestyramine, 15 g/d, a dose similar to that used in our study. In our study, we observed no weight loss. Another theoretical possibility is that bile acid sequestrants reduce postprandial glycemic excursions by slowing the absorption of carbohydrates. Further investigations are needed to study the precise mechanisms of cholestyramine-induced improvement in glycemic control.

Cholestyramine therapy was well tolerated by our patients. As expected, constipation and gastrointestinal distress were the primary side effects. Other minor side effects included heartburn, nausea, and transient increase in serum aminotransferase levels. Constipation was easily managed with stool softeners in six of the eight patients. Two patients dropped out because of severe constipation, nausea, and angina during straining at stools during the first 2 weeks of cholestyramine therapy. Clinical experience suggests that tolerance to bile acid sequestrants may improve with time and also by gradually increasing the dose.

In conclusion, our short-term study shows that cholestyramine therapy effectively decreases LDL cholesterol levels in patients with normal triglyceride levels who have NIDDM and elevated LDL cholesterol levels. The view that aggressive treatment of dyslipidemia in patients with NIDDM is justified because of their unusually high risk for coronary heart disease is becoming more popular. According to the National Cholesterol Education Program [5], the primary target of lipid-lowering therapy is LDL cholesterol, which should be at least 2.6 to 3.4 mmol/L (100 to 130 mg/dL) in patients with NIDDM, or even lower according to some investigators. The mean LDL cholesterol level achieved with cholestyramine therapy was 3 mmol/L (115 mg/dL), which may be an acceptable response for most patients with NIDDM. The bile acid sequestrants thus appear to be superior to nicotinic acid therapy, which often worsens hyperglycemia in patients with NIDDM. They appear to be comparable to statins for LDL lowering and thus may be useful either alone or in combination with statins in those with very high LDL cholesterol levels. The decision whether to choose statins or bile acid sequestrants in such patients with NIDDM may be based on individual tolerance to the medications. Because bile acid sequestrants are not systemically absorbed, they may be preferred to statins in patients with diabetic nephropathy and renal insufficiency and in those with liver function abnormalities. However, bile acid sequestrants probably are contraindicated in patients with NIDDM who have autonomic neuropathy and constipation but may relieve diarrhea of diabetic autonomic neuropathy. Bile acid sequestrants, however, cannot be recommended for patients with NIDDM and hypercholesterolemia whose fasting plasma triglyceride levels exceed 3.4 mmol/L or 300 mg/dL, for those in poor glycemic control, and for morbidly obese patients.

Finally, in carefully selected men with NIDDM and high LDL cholesterol and normal triglyceride levels, cholestyramine therapy effectively reduces LDL levels and may also improve glycemic control. The long-term efficacy of cholestyramine therapy in patients with NIDDM, however, needs further evaluation.