Patients frequently have bloating and constipation while receiving bile acid sequestrants, and the "sandy" or "gritty" texture of these agents is also a substantial impediment to their use. We became aware of two anecdotal accounts of patients who improved the tolerability and efficacy of bile acid sequestrants by mixing them with psyllium at about the same time that Anderson and colleagues [4] reported that psyllium mucilloid had a beneficial effect on hyperlipidemia. We therefore conducted this study to determine the efficacy and tolerability of a combination of psyllium and bile acid sequestrant resin. We assessed quality of life using a disease-specific questionnaire that had previously been validated in a pilot study.

We studied 105 evaluable patients who had persistently elevated cholesterol levels while following the National Cholesterol Education Program Step Two diet [5]. Patients were treated three times daily before meals for 10 weeks with 5 g of cellulose placebo; 2.5 g of colestipol and 2.5 g of psyllium; 5 g of colestipol alone; or 5 g of psyllium mucilloid alone.

Methods

Top Methods Results Discussion Author & Article Info References

Protocol

We used a double-blind, randomized, parallel-group study design. Patients qualified for the study on the basis of cholesterol levels obtained at a screening visit (total cholesterol level, more than 6 mmol/L and less than 8 mmol/L; and triglyceride level less than 3 mmol/L). Patients were excluded if they had diabetes mellitus, a history of cholelithiasis without cholecystectomy, chronic diarrhea or constipation, or uncontrolled thyroid disease. They were also excluded if they required warfarin therapy because of anticipated drug binding by bile acid sequestrants. Patients gave written informed consent to a protocol approved by the University of Western Ontario Standing Committee on Ethics of Human Research. They then had a 1-month period of dietary revision before the start of the study; this included a refresher visit with the dietitian, who gave them new copies of their written instructions. Each patient then made a baseline visit at the beginning of the trial and additional visits at weeks 4 and 10 for the measurement of body weight, blood pressure, heart rate (Critikon Dinamapp Vital Signs Monitor 8100, Tampa, Florida), fasting lipid levels, and apolipoprotein concentrations. Routine chemistry laboratory tests (for levels of fasting glucose, serum albumin, total protein, alkaline phosphatase, aspartate aminotransferase, bilirubin, urea, creatinine, calcium, chloride, carbon dioxide, potassium, and sodium) were done and a hematology profile (hemoglobin level; hematocrit; leukocyte, lymphocyte, monocyte, and platelet counts; and erythrocyte indices) was obtained at baseline and at week 10. Quality-of-life instruments were completed by the patients at baseline and at weeks 4 and 10.

Study Population

Study patients were primary prevention patients recruited from the Atherosclerosis Prevention Clinic at Victoria Hospital, which is a referral clinic for patients with vascular disease, multiple risk factors for atherosclerosis, or both. All patients had been followed for at least 1 year while maintaining a fat-modified diet for the treatment of hyperlipidemia under the supervision of the attending physician and a registered dietitian. The patients had not been previously treated with bile acid sequestrant resins or with hypolipidemic drugs. Patients who participated in the pilot study were not included in the main trial.

Dietary Control

All patients had been previously counseled, as part of their treatment program in the clinic, to follow a diet that corresponded to the National Cholesterol Education Program Step Two diet [5]. They had been followed in the clinic under a dietitian's supervision for at least 1 year before the beginning of the study. At the beginning of the dietary run-in period, each patient made another visit to the dietitian. At this visit, dietary history was reviewed and instructions were repeated, and patients were given new copies of printed dietary instructions. Compliance with diet was assessed by measuring body weight and by a review of dietary intake (3-day recall) with study personnel at each visit. The groups did not differ at baseline or during the study.

Randomization, Dispensing, and Package Counts

The study was done in the Victoria Hospital Clinical Trials center. Patients were randomly assigned to treatment groups by using a table of random numbers; this was done in the pharmacy. Treatment packages were labeled as treatment A, B, C, or D, and both patients and investigators were blinded to treatment assignments until all data analysis had been done. The four treatment packs were prepared by the Victoria Hospital Pharmacy. Compliance was measured by counting returned packages. Patients were given a 5-week supply of packages at each monthly visit; this was done to increase the accuracy of package counts by ensuring that some packages would always be left over to return. Patients were instructed to take the contents of three packages per day before meals, stirred into fruit juice or water. If they could not take all three packages because of adverse effects, they were instructed to take as many as they could tolerate.

Treatments

The packages each contained 5 g of powder, which was either colestipol (Colestid, Upjohn Corporation, Toronto, Ontario, Canada); 55.74% psyllium mucilloid (Metamucil, Searle Canada, Inc., Oakville, Ontario, Canada); 2.5 g of psyllium combined with 2.5 g of colestipol; or cellulose (Avicel, Lot 1415; FMC, Philadelphia, Pennsylvania). Although the full recommended dosage of colestipol is 30 g/d, taking one package three times daily seemed more likely to approximate usual practice. Similarly, although the usual dose of psyllium is lower when the indication is constipation, taking psyllium three times daily with meals is common practice for bile acid sequestration; it is done to take advantage of the emptying of the gall bladder after meals. The other treatments were also given in 5-g doses three times daily with meals so that the regimens and the amount of powder in each package would be the same for all patients. All treatments were unflavored and were provided in small brown paper envelopes. All packages given to a patient contained the treatment to which that patient had been assigned; that is, patients randomly assigned to receive psyllium alone, colestipol alone, or the combination therapy did not receive any placebo packages. The four treatments differed somewhat in color and texture, but differences in flavor were minimal because the predominant flavor was that of the juice in which the powder was mixed. Because each patient received only one kind of treatment, the patients had no way to compare their treatment with the others.

Quality-of-Life Instruments

A quality-of-life instrument was designed specifically for this study according to the principles of disease-specific measures for clinical trials [6]. The instrument was pretested [7], and its psychometric properties were assessed in a pilot study.

The internal consistency reliabilities of the scales were assessed by using Cronbach alphas [8] separately for each visit and for combined data. Internal consistency reliability was high; Cronbach alphas were above 0.9. Analysis of covariance (controlling for social variables such as age, sex, education, and occupation [PSSPC+, Chicago, Illinois]) was used to compare treatments [9]. Further details on the quality-of-life instrument are available from the authors.

Lipid and Apolipoprotein Measurements

While patients reclined in a blood-sampling chair, fasting blood samples (a 12- to 14-hour fast) were collected from an arm vein into tubes containing EDTA-Na₂ at a final concentration of 0.15%. Plasma was obtained by centrifugation at 1000 g for 25 minutes at 4 °C and was assayed for total cholesterol and triglyceride levels using enzymatic methods (Boehringer Mannheim, Montreal, Quebec, Canada; cholesterol: CHOD-PAP; triglycerides: triglycerides without free glycerol). A 5-mL aliquot of plasma from each sample was subjected to ultracentrifugation through a 1.006 g/mL density solution containing 0.195 M of sodium chloride, 1 mM of TRIS, pH 7.4, 1 mM of EDTA, 10 µ M of phenylmethane sulfonyl fluoride, 3 mM of NaN₃, and 0.10 mM of merthiolate in a Beckman 50.3 Ti rotor (Beckman Instruments, Mississauga, Ontario, Canada) for 18 hours at 40 000 rotations per minute at 12 °C using a Beckman L8 centrifuge (Beckman Instruments) [10]. Very-low-density lipoprotein (VLDL) cholesterol was collected by tube slicing, and the infranatant was carefully recovered and adjusted to 5 ml. Aliquots were obtained for triglyceride and cholesterol levels and for apolipoprotein B and A-I concentrations, and they were stored at –20 °C until assayed.

To rule out type III hyperlipidemia, the VLDL from the first sample obtained from each patient was assayed for apolipoprotein E phenotype by using analytical isoelectric focusing gel electrophoresis [11]. High-density lipoprotein cholesterol was measured after the precipitation of plasma with dextran sulfate magnesium chloride according to the method of Warnick and coworkers [12]. Low-density lipoprotein cholesterol (LDL) was measured by subtracting the HDL value from the infranatant cholesterol value. Very-low-density lipoprotein cholesterol and triglyceride levels were measured as the difference between total plasma and infranatant values. The intra- and interassay coefficients of variation were 2.2% and 3.0%, respectively, for total cholesterol and 2.2% and 3.0%, respectively, for total and infranatant triglycerides. Apolipoproteins B and A-I were measured by immunoturbidimetric assay using reagents from Boehringer Mannheim. Apolipoproteins B and A-I were measured in the plasma infranatant after ultracentrifugation was done as described above. These values were taken to represent LDL and intermediate-density lipoprotein apolipoprotein B and total apolipoprotein A-I, respectively. The intra- and interassay coefficients of variation were 4.5% and 8.0%, respectively, for apolipoprotein B and 3.8% and 7.2%, respectively, for apolipoprotein A-I. All assays were standardized to plasma secondary standards, for which the target values were assigned at the Lipid Research Laboratory, University of Toronto, St. Michael's Hospital, Toronto, Canada, which is standardized by the Centers for Disease Control and Prevention.

Purchase Cost of Treatments

To compare prices, we used the Best Available Price in the 1993 edition of the Ontario Drug Benefit formulary (the most recent edition available). We calculated the purchase cost of a 1-month supply of the doses of treatments used in this study, provided as single-dose packages of Metamucil and Colestid.

Statistical Methods

The sample size was developed and the quality-of-life instrument was validated in a pilot study done in 25 patients. The pilot study indicated that we would need 104 patients to achieve a 90% power to show a 15% reduction in LDL at a significance level of 0.05. Because this study was designed as an efficacy study rather than as a management study, and because cases with missing values were deleted from the analysis by the statistical techniques used, dropouts were replaced to achieve 104 evaluable patients. Patients not completing the study were not included in the analysis because their lipid and lipoprotein levels were not measured after they had dropped out.

Analysis of blood pressure, heart rate, lipoprotein levels, and routine laboratory tests was done using repeated-measures analysis of variance (multivariate analysis of variance, SPSS PC+, Chicago, Illinois), with post hoc analysis using the Tukey test for significant differences between treatments and between visits [9]. Results are presented as the mean ±SD or as the mean with 95% CIs.

Results

Top Methods Results Discussion Author & Article Info References

Table 1 shows the baseline characteristics of the patients in the four treatment groups.

Age and Sex Distribution

Mean age was 55 ± 10 years (range, 27 to 70 years); 43 patients were men and 62 were women. Mean age did not differ significantly between the sexes, and no significant differences in age or sex were seen among the treatment groups.

Weight and Diet

The baseline values of the groups did not differ significantly, and weight did not change significantly in any group during the study. Mean changes in weight were –0.9 ± 2.04 kg for placebo; 0.19 ± 1.8 kg for combination therapy; –0.2 ± 2.9 kg for colestipol alone; and 0.02 ± 1.63 kg for psyllium alone. The absence of any differences in weight and the interviews with the dietitian indicated that the groups did not differ in compliance with the recommended diet.

Blood Pressure and Heart Rate

The groups did not differ at baseline for blood pressure and heart rate; however, diastolic blood pressure was significantly lower at week 10 in the group receiving colestipol, having decreased from 134 ± 13/82 ± 8 mm Hg to 129 ± 17/79 ± 9 mm Hg (P = 0.025). Pressures in the other groups were similar at baseline (range, 130 ± 13/83 ± 7 to 133 ± 15/82 ± 9) and did not change more than 1 mm Hg on average, with the exception of the diastolic pressure of patients receiving psyllium, which decreased from 82 ± 11 to 78 ± 19. Heart rate ranged from 65 ± 14 to 68 ± 12 beats/min at baseline, and mean changes were only 2 to 3 beats/min and were similar for all groups.

Routine Laboratory Tests

Only glucose levels showed a significant interaction between group and visit (F₃,100 equals 3.00; P = 0.034). Subsequent post hoc analysis using the Tukey HSD test showed changes from baseline to be significantly different in the group receiving combination therapy from those in the group receiving psyllium alone; blood glucose levels increased 0.41 mmol/L with combination therapy and decreased 0.14 mmol/L with psyllium alone (P < 0.05).

Lipoproteins

Colestipol alone and combination therapy were both more effective than psyllium alone in lowering LDL (Table 2). No statistically significant differences were seen with respect to individual lipid values, such as total cholesterol level, among the groups receiving combination therapy, psyllium alone, or colestipol alone. Psyllium alone did not have any statistically significant reductions in lipid levels or ratios. Combination therapy was significantly more effective than either colestipol or psyllium alone in reducing the ratio of total cholesterol to HDL, even though the dose of each component in the combination was only half that given in the single treatments. As shown in Figure 1, combination therapy had a more favorable effect on the lipid profile (the direction and magnitude of change of LDL, triglyceride, and HDL levels).

View larger version (20K): [in this window] [in a new window]   Figure 1. Change from baseline to end of week 10 in mmol/L for lipids and g/L for apolipoproteins with 95% CIs for the change. APOA1 equals apolipoprotein A-I; APOB1 equals apolipoprotein B; HDL equals high-density lipoprotein; LDL equals low-density lipoprotein; TCHOL equals total cholesterol; TRIGLY equals triglyceride; VLDL equals very-low-density lipoprotein.* P < 0.05 compared with placebo; **P < 0.001 compared with placebo and P < 0.05 for combination compared with colestipol (multivariate analysis of variance with post hoc Tukey HSD tests for comparisons of significant differences).

As a summary estimate of the effect of these treatments on the overall lipid profile, we examined the effect on the ratio of total cholesterol to HDL [13]. Combination therapy reduced this ratio by 18.2% (95% CI, 12.3% to 24%); colestipol alone reduced it by 10.6% (CI, 2.0% to 15.4%); psyllium alone reduced it by 6.1% (CI, 1.5% to 10.6%); and the mean change with placebo was a reduction by 0.1% (CI, –4.8% to 7%). Multivariate analysis of variance showed significant differences between group and test (P = 0.0002); subsequent post hoc Tukey HSD tests showed changes to be significant (combination therapy compared with placebo, P < 0.001; colestipol compared with placebo, P < 0.05; combination therapy compared with colestipol, P < 0.05).

Compliance

Percentage compliance (the percentage of those packages patients were expected to consume that actually were consumed) ranged from 92.5% to 96.6%, and no statistically significant differences were seen among treatment groups. Compliance tended to decrease slightly from week 4 to week 10, and it tended to be lower in the group receiving psyllium alone, but compliance remained above 90% in all groups through week 10.

Quality of Life: Comparisons among Treatments

In general, quality of life was best with combination therapy and worst with placebo and with psyllium alone; quality of life with colestipol was intermediate. Details of the quality-of-life study are available from the authors.

Serious Adverse Events

One patient had a severe allergic reaction to psyllium (angioedema affecting the face [14]). One patient had a myocardial infarction, one was hospitalized with a bowel obstruction attributed to adhesions from previous bowel surgery; one had a right-hemisphere brain infarction, and one had chest pain attributed to myocardial ischemia. The allergic reaction was attributed to psyllium; the other adverse events were thought to be unrelated to treatment.

Sixteen patients dropped out of the study because of adverse effects of treatment. Five had nausea, 2 had diarrhea, 3 had bloating, and 2 had abdominal cramps (some patients had more than one event). One patient dropped out because she had a family history of cancer and was worried about taking a study drug; 1 dropped out because she felt she could not be compliant; and 1 dropped out because she was preoccupied with preexisting chronic back pain and "didn't want another thing to worry about."

Eleven of the patients who dropped out were women and 5 were men; no statistically significant differences were seen among the treatment groups with respect to the frequency of adverse events leading to dropouts. Table 3 shows the frequency of gastrointestinal symptoms throughout the study (evaluated at week 10) in the four treatment groups; these results are for patients who completed the study.

Purchase Cost

Surprisingly, a 1-year supply of resin costs $984, or 1.5 times the cost of a 1-year supply of HMG CoA reductase inhibitor ($650 per year for lovastatin, 20 mg/d). In contrast, psyllium costs about $142 per year when purchased in the more expensive single-dose packages; the purchase cost of the combination of psyllium and colestipol is approximately $412 per year. Savings are even greater if psyllium is purchased in larger quantities and dispensed as a half teaspoon mixed with colestipol.

Discussion

Top Methods Results Discussion Author & Article Info References

We show that half the usual dose of colestipol combined with an equal amount of psyllium had effects on individual lipid values that did not differ from those achieved with colestipol alone or psyllium alone. However, the combination was more effective in improving the cardiovascular risk profile (ratio of total cholesterol to HDL) and was tolerated as well as or better than twice as much of either component given alone.

It is interesting, and perhaps surprising, that a reduction in the dosage of colestipol from 15 to 7.5 g per day could be associated with an increased reduction in LDL. The reverse might be expected, because Superko and colleagues [15] showed that 5 g of colestipol per day achieved 51% of the LDL reduction seen with 15 g/d. To further assess this issue, we calculated a dose-response curve using the data of Superko and colleagues, and we estimated from it that 7.5 g of colestipol per day should achieve 75% of the response seen with 15 g/d. Thus (if changes in LDL are proportionate to changes in the ratio of total cholesterol to LDL), colestipol would reduce this ratio by 13.7% and psyllium would reduce it by another 4.5%, so that combination therapy would reduce this ratio by a total of 18.2%. Because psyllium alone, given at twice the dose used in the combination therapy, achieved an 8% reduction, an additive effect of the combination therapy seems likely. A similar effect might be expected with cholestyramine, because in the LRC-CPPT, patients following a low-fat diet who achieved 50% compliance with their cholestyramine regimen had only a 9% reduction in LDL; those who were fully compliant had a 15% reduction [1].

The results of our trial were consistent in magnitude and direction with the results of the pilot study, which also showed significantly better efficacy and quality of life with the combination therapy.

Importance of Dietary Fat Intake on Effectiveness of Bile Acid Sequestration

It was surprising that psyllium alone had little effect, because Anderson and coworkers [4] showed psyllium to have a substantial effect on lipids using the same control. We suspect that the difference between our results and those of Anderson and coworkers may be related to the diets to which additional treatments were added. The National Cholesterol Education Program Step One diet aims to reduce dietary cholesterol to less than 300 mg/d, to reduce total dietary fat to fewer than 30% of calories, and to reduce saturated fats to fewer than 10% of calories. The Step Two diet aims to reduce dietary cholesterol to less than 200 mg daily, to reduce total dietary fat to fewer than 30% of calories, and to reduce saturated fats to fewer than 7% of calories [5].

Thus, differences in dietary fat intake may have caused the reduction of LDL with psyllium alone to be smaller in our study than in the study by Anderson and colleagues [4]. They found a 20.2% reduction of LDL in patients who continued their usual diet (< 300 mg of dietary cholesterol per day and 40% of calories from fat), whereas our patients were instructed to follow a more restrictive Step Two diet. In patients following a Step One diet, psyllium has been found to produce more modest results, reducing LDL by 8.6% [16], 8.8% [17], 8.6% [18], and 6.4% [19].

Our finding of a significant reduction in the ratio of total cholesterol to HDL and of levels of total cholesterol and LDL level to below values achieved by a Step Two diet suggests that bile acid sequestration deserves more widespread use. This is supported by Hunninghake and colleagues [20], who found that a Step Two diet reduced LDL by only 5%.

It is possible that, as dietary fat intake becomes very low, the benefit of bile acid sequestration may diminish. Jenkins and colleagues [21] recently showed that a high intake of soluble fiber (> 24 g/1000 kcal), but not insoluble fiber, reduced LDL by a mean of only 4.8%, to below levels achieved by a restrictive diet containing less than 25 mg cholesterol per 1000 kcal and fewer than 4% of total calories from saturated fat.

Blood Pressure and Glucose

It is possible that the changes detected in blood pressure and glucose may have been chance findings. We have reported them because they are consistent with related findings in the literature and may provide leads for further study.

The reduction of diastolic blood pressure by colestipol raises the issue of the mechanism by which vegetarian diets lower blood pressure [22, 23]. Although some studies indicate that dietary fiber reduces blood pressure, the confounding effect of weight loss is problematic. Because our patients did not lose weight, a study of the effects on blood pressure of soluble compared with insoluble fiber, similar to the study by Jenkins and colleagues [21], might be of interest.

The reduction of serum glucose levels by psyllium appears to be similar to the reduction of these levels by guar gum, another source of soluble fiber [24]. The possibility that psyllium reduces peaks of blood sugar and thereby increases insulin sensitivity merits further study.

Possible Role of the Combination Regimen

Taken together, the studies of Hunninghake and colleagues [20] and Jenkins and colleagues [21] seem to place the efficacy of our most effective treatment regimen (a Step Two diet and therapy with a low-dose combination of psyllium and colestipol) between that of a very restrictive diet and that of treatment with a combination of moderate diet and HMG CoA reductase inhibitor.

This finding may be clinically important if patients unable to tolerate 15 g of colestipol daily because of adverse gastrointestinal side effects can be "salvaged" by halving the colestipol dose and adding psyllium.

Adverse Effects

It was surprising that psyllium alone, 5 g three times daily, was associated with adverse effects that were as bothersome as those associated with colestipol alone, 5 g three times daily. It appears that the reduction of the dose of both agents in the combination therapy may have accounted for some of the reduction in adverse effects. This was in addition to the reduction of constipation (a common effect of colestid) by the addition of psyllium, which was expected. The high proportion of patients who had adverse effects with the cellulose placebo was also surprising. We had chosen this placebo on the basis of the work of Anderson and colleagues [4], but our experience suggests that a different control should be used in future studies.

Purchase Cost Reduction

The possibility of improving—or at least maintaining—the efficacy of bile acid sequestrant resins while halving purchase costs by reducing the dose of these resins and adding psyllium is an important implication of this study and a strong reason to suggest combination therapy to patients. For many patients, combination therapy would provide a significant financial benefit in addition to benefits in efficacy and quality of life.

It should be noted that the purchase cost savings described above are relative to bile acid sequestrants alone. In the context of current practice, comparison with HMG CoA reductase inhibitors might be more relevant. Schrott and colleagues [25] recently reported that a combination of low-dose lovastatin with colestipol was less costly ($21 per percentage point of cholesterol reduction) than high-dose lovastatin alone ($28 per percentage point of cholesterol reduction). Using a combination of psyllium and colestipol would further reduce the purchase cost, which suggests that this therapy would be more cost-effective than therapy with high-dose HMG CoA reductase inhibitors alone. Further studies should be done to test this hypothesis.

Study Limitations

The number of patients in our study, the study duration, and the study setting (a randomized controlled clinical trial in a tertiary care center) limit the generalizability of our results. We did not address the relative cost-effectiveness of the treatments; to do so, we would have had to consider long-term compliance, effectiveness, and adverse effects. Although compliance in this study was better than 90% for 10 weeks, long-term compliance in a practice setting would probably be lower. Our observations about "purchase cost" are thus relevant only to purchase costs and not to cost-effectiveness. We have used the term purchase cost to emphasize this limitation. In addition, we have no data on the combination therapy studied here in patients receiving other drugs, such as HMG CoA reductase inhibitors. Thus, we can only speculate on the relation of our findings to those of Schrott and colleagues [25], who used colestipol with lovastatin, with respect to the possible benefit of using the combination of colestipol and psyllium in conjunction with other drugs. Additionally, because patients who had been previously treated were not included in the study, we can only speculate that patients who cannot tolerate colestipol alone may be able to tolerate combination therapy. Further studies should be done to explore these issues.

Conclusions

We show that it is possible to maintain efficacy for lipid profile improvement while maintaining or improving quality of life and compliance by mixing psyllium with half the usual dose of a bile acid sequestrant resin. A substantial reduction in the purchase cost of improving the ratio of total cholesterol to HDL, compared with the cost of colestipol alone, justifies consideration of this strategy. By increasing the acceptability of bile acid sequestration, the combination of low-dose colestipol and psyllium may permit the wider use of bile acid sequestrant resins. Our findings may have implications for both clinical practice and public policy.