Efficacy of Psyllium in Reducing Serum Cholesterol Levels in Hypercholesterolemic Patients on High- or Low-Fat Diets

Methods

Participants

The University of Cincinnati Medical Center and the Washington University School of Medicine enrolled 135 patients with primary hypercholesterolemia (total serum cholesterol 5.7 mmol/L [220 mg/dL]) for study. Analysis of an initial 7-day diet record determined the continued eligibility of patients for study. Only patients meeting dietary criteria ( 40% of calories derived from fat or 30% of calories derived from fat) were asked to participate in the study. Reasons for withdrawal from the study included a weight change of more than 5% of initial body weight (3 participants); noncompliance with study medication (7 participants) even though no adverse events were reported; diet instability (4 participants); withdrawal of consent to participate (2 participants); and stomach cramps (1 participant). Data collected on these persons were not used in the analysis. A total of 118 participants, 68 men and 50 women, successfully completed the trial; 54 were from the University of Cincinnati and 64 were from Washington University. All participants were between 21 and 70 years old, had a stable total serum cholesterol level at baseline (that is, measurements at weeks 2 and 1 were within 0.78 mmol/L [30 mg/dL] of each other), and had triglyceride levels 4.5 mmol/L (400 mg/dL). Written informed consent forms (approved by the institutional review boards at both study sites) were obtained from all participants.

Test Articles and Doses

The active test article in our study was psyllium, which was given as orange-flavored, sugar-free Metamucil. The placebo test article matched the active product but contained an inert bulk fiber purified from microcrystalline cellulose (Avicel, FMC Corp., Philadelphia, Pennsylvania) in the place of psyllium. Both test articles were provided to participants in identical foil packets. Each dose provided 5.1 g of psyllium or matching placebo. Participants were instructed to take the test article before breakfast and dinner (5.1 g twice daily) in 240 mL (8 oz) of water or fruit juice. Adherence to the regimen was monitored by interviews and by counts of returned packets at each clinic visit.

Study Design

We conducted a double-blind, placebo-controlled, parallel study in which psyllium or placebo was given in combination with either a high-fat or a low-fat diet. The study lasted 16 weeks and was done in two phases. The first 8 weeks (week 8 to 0) consisted of a dietary lead-in phase. After the initial visit and the analysis of a 1-week food record, which reflected established individual dietary habits, participants were instructed to maintain their established diet (target diet) and were monitored for adherence to the target diet and stability of lipid concentrations. After 8 weeks, participants who met study entry criteria were randomly assigned to treatment [psyllium or placebo] within their target diet, were administered the test article at week 0, and continued taking the test article throughout the 8-week treatment phase. Participants were evaluated during the treatment phase at weeks 2, 4, 6, and 8 and were required to maintain their target diet throughout the treatment phase of the study. We evaluated four groups: 1) patients on a high-fat diet who received psyllium, 2) patients on a high-fat diet who received placebo, 3) patients on a low-fat diet who received psyllium, and 4) patients on a low-fat diet who received placebo.

Dietary Evaluation

One-week food records were collected throughout the study: Four were obtained during the dietary lead-in phase and four were obtained during the treatment phase. Records for weeks 7, 4, 0, 4, and 8 were analyzed using the Minnesota Nutrition Data System software (Nutrition Coordinating Center, Minneapolis, Minnesota), and dietary compliance was assessed at weeks 2, 2, and 6 using the Food Record Rating system [26]. Four specified days (Saturday, Sunday, Monday, and Wednesday) of the 1-week food record were analyzed, using the Nutrition Data System, for total calories; total fat; saturated, polyunsaturated, and mono-unsaturated fats; polyunsaturated to saturated fat ratio; cholesterol intake; and dietary fiber intake.

Experimental Measurements

Fasting blood samples were obtained for lipid profiles nine times during the 16-week study (at weeks 8, 6, 4, 2, 1, 2, 4, 6, and 8). Lipid profiles consisted of total serum cholesterol, high-density lipoprotein cholesterol, and triglyceride determinations done by Smith-Kline Beecham Laboratories using standard laboratory methods [27]. Total serum cholesterol was measured using a cholesterol oxidase-esterase mixture and a colorimetric assay [28], and high-density lipoprotein cholesterol was isolated from serum after precipitating LDL cholesterol and very-low-density lipoprotein cholesterol with phosphotungstate-magnesium chloride. Serum triglyceride concentrations were quantified using a lipase-glycerol kinase enzyme reaction [29]. Serum concentrations of LDL cholesterol were calculated using the Friedewald equation [30]. Apolipoprotein (apo) A-I and apo B concentrations were measured and quantification was done at the visit before initiation of treatment (week 1)and again at the final visit (week 8). Radioimmunoassays (Washington University Lipid Research Core Laboratory) were used to measure concentrations of apo A-I and apo B in plasma [31, 32]. A complete physical examination, clinical chemistry and hematologic studies, and urinalysis were done at weeks 6 and 8 to assess the safety of the treatments.

Statistical Methods

Mean baseline lipid, lipoprotein, and apolipoprotein levels were calculated using the average values at weeks 2 and 1 (before random assignment to treatment). Mean post-treatment levels were calculated using the average of the final two visits (weeks 6 and 8). Baseline dietary data were collected at week 0, and post-treatment values were gathered at week 8. The percentage change between baseline and post-treatment values for each group was calculated using the percentage difference between baseline and post-treatment means for each participant.

At each time point, we compared psyllium and placebo recipients on high-fat diets; psyllium and placebo recipients on low-fat diets; and psyllium recipients on high- and low-fat diets. A two-way analysis of variance was used to test for treatment, site, and treatment-by-site effects on the final percentage difference between baseline and final post-treatment lipid levels. No treatment-by-site interactions were noted.

Paired t-tests were used to compare baseline and post-treatment values for each variable within each treatment group. Because treatment effects were observed for both diet groups, the lipid and apolipoprotein data were pooled by diet and a comparison of psyllium and placebo recipients was done along with the paired t-tests mentioned above.

Study participants were divided, using NCEP guidelines for LDL cholesterol levels, into the following coronary heart disease risk categories both at baseline and after treatment: group I, less than 3.4 mmol/L (130 mg/dL); group II, 3.4 to 4.1 mmol/L (130 to 159 mg/dL); group III, 4.1 to 4.9 mmol/L (160 to 189 mg/dL); and group IV, 4.9 mmol/L (190 mg/dL) or more.

We evaluated the movement of participants among cholesterol categories from baseline to the final post-treatment evaluation, regardless of baseline starting values for a given treatment (Table 1). We used a marginal homogeneity test for matched pairs [33] to assess movement among cholesterol categories. The test uses baseline and post-treatment patient totals in each cholesterol category to calculate the test statistic. The null hypothesis assumes that there is no movement among LDL cholesterol categories. Small P values provide evidence to contradict this hypothesis. The test statistic and exact P value were calculated using StatXact [34].

Results

Baseline Data

The number of participants assigned to each treatment regimen is presented by sex and race in Table 2, which also includes data on age and body mass index (weight [kg]/height² [meters]). Persons consuming the high-fat diets were younger and a had higher mean body mass index than patients consuming the low-fat diet, regardless of whether they were assigned to psyllium or placebo.

Safety Evaluation

No statistically significant or clinically meaningful changes were observed in serum chemistry, hematologic, or urinalysis results when psyllium recipients were compared with placebo recipients. Psyllium and placebo recipients on the low-fat diet lost an average of 1.2 lb (0.8% of body weight) and 1.8 lb (1.1% of body weight), respectively; these decreases in weight were significant when compared with baseline values (P < 0.05) but were not significant when compared with the corresponding values in patients on the high-fat diet. Therefore, treatment effects cannot be attributed to weight loss. Adverse events were predominantly gastrointestinal in nature and were mild, transient, and self-limited.

Diet Comparison

No significant changes in percentage of calories derived from fat were found when baseline and post-treatment values were compared. Participants on high-fat diets maintained a fat intake that accounted for more than 40% of their calories, whereas those on low-fat diets maintained a fat intake that accounted for less than 27% of their calories (see Table 2). No difference in the total energy intake was observed between the two diet groups [data not shown]. Total dietary fiber and soluble fiber intakes (excluding psyllium or cellulose supplement) were monitored in all groups. No significant differences were observed among groups regarding fiber intake (Table 2). The differences in cholesterol intake, saturated fat intake, and polyunsaturated to saturated fat ratio between the high-fat and low fat diet groups were consistent with the components of the diet and therefore make the high-fat diet generalizable to the average Western diet and the low-fat diet generalizable to the American Heart Association Step I diet that is prescribed for hypercholesterolemic patients.

Unexpected decreases in cholesterol intake were observed in both psyllium groups. The mean reduction in dietary cholesterol intake from baseline for psyllium recipients in the high- and low-fat diet groups was 115 mg and 29 mg, respectively. To investigate whether changes in LDL cholesterol and total cholesterol levels can be attributed to psyllium in light of these changes in dietary cholesterol, we made correlations between dietary cholesterol intake and the LDL cholesterol and total cholesterol levels for each treatment regimen-site combination (that is, psyllium plus high-fat diet, psyllium plus low-fat diet, placebo plus high-fat diet, and placebo plus low-fat diet for both site 1 and site 2). Marginally significant positive correlations were noted between dietary cholesterol and lipid changes (LDL cholesterol, 0.6 [P = 0.09] and total cholesterol, 0.7 [P = 0.02]) among psyllium recipients in the high-fat diet group at site 1 (n = 9). Although the correlations between these same variables were not significant in the remaining high-fat diet groups, strong trends were observed among psyllium recipients in the high-fat diet group at site 2 and among placebo recipients in the high-fat diet group at site 1; correlations in the placebo group were stronger than in the psyllium group. Despite the positive correlations in three of four site groups of patients on the high-fat diet, the psyllium groups differed only marginally from placebo groups regarding reduction from baseline in LDL cholesterol (P = 0.06) and differed greatly regarding reduction in total cholesterol (P = 0.009). This finding indicates that psyllium had an effect on the reduction of LDL cholesterol and total cholesterol levels that was not accounted for by the decrease in dietary cholesterol intake.

Changes in Lipoprotein and Apolipoprotein Levels

Mean lipoprotein and apolipoprotein values and the percentage change from baseline to the final post-treatment determination for each group are presented in Table 3. Psyllium recipients on high-fat and low-fat diets showed significant decreases in total cholesterol of 0.4 mmol/L (15.1 mg/dL) (P < 0.01) and 0.3 mmol/L (10.5 mg/dL) (P = 0.001), respectively, and decreases in LDL cholesterol of 0.3 mmol/L (13.1 mg/dL) (P < 0.01) and 0.3 mmol/L (11.2 mg/dL) (P = 0.0001), respectively, when post-treatment and baseline levels were compared. No significant reductions from baseline were observed in either placebo group.

The mean percentage change in LDL cholesterol level for each group is shown in Figure 1. The difference in percentage change from baseline between psyllium recipients on high-fat and low-fat diets was not significant (P > 0.2) for either total cholesterol or LDL cholesterol. The difference in percentage change from baseline when psyllium and placebo recipients were compared according to dietary group was significant for both total cholesterol (P = 0.01 and P = 0.05 for high-fat and low-fat diet groups, respectively) and for LDL cholesterol (P = 0.05 and P = 0.04, respectively).

Percentage change in total cholesterol and LDL cholesterol concentrations may be conservatively estimated by subtracting the mean percentage change observed in the placebo groups from the mean percentage change in the corresponding psyllium groups: Percentage changes were 3.4% and 5.1%,respectively, for the low-fat diet group, and 6.7% and 7.4%,respectively, for the high-fat diet group. However, because the decrease in cholesterol level occurred equivalently between diet groups, the overall reduction was determined by pooling participants by treatment alone, without regard to diet: Percentage changes in total cholesterol and LDL cholesterol levels were 4.5%(P < 0.01) and 5.8%(P < 0.01) when psyllium recipients (n = 59) were compared with placebo recipients (n = 59).

Apolipoprotein B levels decreased in each of the four groups, but no group showed a significant decrease in single-measurement comparisons of baseline and post-treatment values. Significant reductions in post-treatment apo B levels relative to baseline levels were observed in the overall analysis for both the placebo (P = 0.05) and psyllium (P = 0.03) groups (Table 4). However, no difference was observed between the two groups (P > 0.2). We expect that the single-value measurements at weeks 0 and 8 added to the overall error and thus contributed to this outcome.

Although the difference was not statistically significant, triglyceride levels increased by 27.4% among placebo recipients on high-fat diets and decreased by 2.8% among psyllium recipients on high-fat diets. Triglyceride changes did not differ among low-fat diet groups. No significant changes in high-density lipoprotein cholesterol concentrations were observed between any groups compared. Apolipoprotein A-I levels significantly increased (6.4%; P = 0.03) from the post-diet lead-in baseline value only in psyllium recipients on the low-fat diet, suggesting a difference in the overall comparison between placebo and psyllium recipients (see Table 4). Again, only single measurements were obtained at each time point.

Changes in LDL Cholesterol Classification

Participants in each treatment group were categorized according to the NCEP LDL cholesterol classification at baseline and at the end of the study: group I, less than 3.4 mmol/L (130 mg/dL); group II, 3.4 to 4.1 mmol/L (130 to 159 mg/dL); group III, 4.1 to 4.9 mmol/L (160 to 189 mg/dL); and group IV, 4.9 mmol/L (190 mg/dL) or more. The percentage of participants in each treatment group who changed LDL cholesterol classification is shown in Figure 2. Psyllium recipients on the low-fat diet showed a highly significant decrease (P < 0.0001) in LDL cholesterol classification. Results were as follows: Of the 13 participants in group IV ( 4.9 mmol/L [190 mg/dL]) at baseline, 3 showed post-treatment reductions in LDL cholesterol level sufficient to move into group III (4.1 to 4.9 mmol/L [160 to 189 mg/dL]); four had reductions large enough to move into group II (3.4 to 4.1 mmol/L [130 to 159 mg/dL]). Of the 15 patients who began the study in group III, 5 showed reductions in LDL cholesterol levels sufficient to move them into group II. Of the 13 participants initially in group II, 3 achieved LDL cholesterol reductions sufficient to move into group I. Psyllium recipients following a high-fat diet also showed significant movement within LDL cholesterol classifications (P = 0.04), although the findings were not as dramatic as in psyllium recipients on low-fat diets. No significant changes in LDL cholesterol status occurred in either placebo group (see Figure 2).

The overall analysis of NCEP LDL cholesterol classification is shown in Figure 2 and Table 1. Regardless of dietary regimen, 39% of psyllium recipients showed a significant change in LDL cholesterol classification (P > 0.0001) compared with 20.3% of placebo recipients (P > 0.2).

Discussion

Hypercholesterolemic patients on low-fat and high-fat diets who received psyllium (5.1 g twice a day) showed decreases in total cholesterol of 4.2% and 5.8%, respectively, and decreases in LDL cholesterol of 6.4% and 7.2%, respectively. Although the mean percentage reductions in total cholesterol and LDL cholesterol were equivalent in the two psyllium-treated diet groups, the combination of a low-fat diet and psyllium was the most successful in improving NCEP LDL cholesterol classification. Because of its large sample size, careful dietary control, and inclusion of a high-fat diet group, our study represents the most careful evaluation yet of the cholesterol-lowering properties of psyllium in the presence of both high-fat and low-fat diets.

Accounting for parallel placebo differences from baseline, we found that psyllium appeared to reduce total cholesterol and LDL cholesterol by 3.4% and 5.1%, respectively, in the low-fat diet group and by 6.7% and 7.4%, respectively, in the high-fat diet group. When participants were analyzed without regard to diet group, we found that psyllium decreased total cholesterol and LDL cholesterol levels by 4.5% (P < 0.01) and 5.8% (P < 0.01), respectively, when accounting for the placebo changes. These reductions are generally consistent with findings of previous studies in which participants maintained low-fat diets (Table 5). However, many were particularly responsive to psyllium and achieved more substantial and thus more clinically useful cholesterol lowering than the mean values indicate. Although any cholesterol reduction could be viewed as leading to reduced cardiovascular risk [35], the average mild reductions we observed may be particularly valuable when combined with reductions resulting from other lipid-lowering agents. Further studies are needed to explore this issue.

Overall, reductions in total cholesterol ranging from 4% to 10% and LDL cholesterol reductions ranging from 6% to 20% have been observed [14-20, 25]. For example, Bell and colleagues [15] studied 75 patients, 40 of whom received 10.2 g of psyllium per day while controlled on a low-fat diet ( 30% of calories derived from fat). These investigators found decreases in total cholesterol and LDL cholesterol levels of 4.8% and 8.2%, respectively. In studies by Neal and colleagues [16], Levin and colleagues [17], and Anderson and colleagues [18], psyllium dosages of 10.2 to 20.4 g/d consistently decreased total cholesterol levels by approximately 4% to 7% and LDL cholesterol levels by approximately 6% to 9% relative to placebo, without effecting changes in high-density lipoprotein cholesterol or triglyceride levels.

In contrast, Anderson and colleagues [14] randomly assigned 26 men on high-fat diet (approximately 40% of calories derived from fat) to receive either psyllium, 10.2 g/d, or placebo and found that psyllium recipients had reductions in total cholesterol and LDL cholesterol levels of 14.8% and 20.2%, respectively. These reductions in LDL cholesterol levels were substantially greater than those observed in previous studies, suggesting that the effect of psyllium may differ according to diet. The equivalent LDL cholesterol reduction between groups in our study differs from aforementioned data. The results of our study, which included an 8-week dietary lead-in period, continuous evaluation of 1-week food records, distinctively disparate fat consumptions between groups, and relatively tight confidence intervals (see Table 5), are therefore more representative of psyllium's relative effect on patients following low-fat and high-fat diets. The estimation of equivalent cholesterol reductions using a soluble fiber, regardless of dietary fat intake, provides insights into the mechanism whereby psyllium lowers plasma cholesterol levels.

Two major hypotheses have been frequently postulated for lipid-lowering action of psyllium [36-38]. First, soluble fiber may physically entrap bile acids, resulting in increased fecal loss similar to that seen with the bile acid sequestrants. Second, soluble fiber produces a decrease in cholesterol absorption and bile acid reabsorption by physically disrupting the intraluminal formation of micelles [36-38]. In both mechanisms, the interruption of enterohepatic circulation leads to increased conversion of cholesterol into newly synthesized bile acids in the liver. This reduction in cholesterol synthesis leads to an up-regulation of the LDL receptor and, in turn, to an enhanced uptake of LDL cholesterol from the plasma. Thus, reductions occur in serum LDL cholesterol and therefore in total cholesterol. One could speculate that the second mechanism, involving cholesterol gastrointestinal absorption, might lead to a discrepancy in the psyllium effect between high-fat and low-fat diets. The similar results for cholesterol lowering in patients on high-fat and low-fat diets in our study are consistent with the first hypothesis and support the recent findings of Everson and colleagues [21], who provided metabolic data on in vivo sterol balance indicating a psyllium effect related to the sequestration of bile acids.

Another hypothesis was presented in a well-publicized report on oat bran. The investigators suggested that foods high in saturated fat and cholesterol are replaced by foods containing high fiber and that soluble fiber has no direct action on cholesterol [22]. In contrast, Davidson and colleagues [23] showed that -glucans in oat bran produced a significant reduction in LDL cholesterol without changing the dietary fat content. Anderson and colleagues [24], in a metabolic study of patients maintained on an isocaloric diet with 41% of the calories derived from fat, compared the hypocholesterolemic effects of oat bran with those of wheat bran. Significant reductions in both total cholesterol and LDL cholesterol were observed in the patients who received oat bran compared with those who received wheat bran, and again, no change was made in the fat composition of the diet. Our study, which included a sizeable population, a placebo cohort, and careful dietary monitoring for fat content, further supports the direct role of fiber and adequately dispels any notion that fat substitution accounts for the primary cholesterol-lowering effect of psyllium.

However, variations in the intake of dietary cholesterol may have had a confounding influence on the absolute difference in LDL cholesterol levels observed among groups. Psyllium recipients on a high-fat diet showed a mean reduction in dietary cholesterol of 115 mg per 2430 kcal(47 mg per 1000 kcal), whereas psyllium recipients on a low-fat diet showed a mean reduction in dietary cholesterol of 29 mg per 1600 kcal(18 mg per 1000 kcal). A decrease in dietary cholesterol of 100 mg per 1000 kcal is reported to result in a 10 mg/dL reduction in plasma total cholesterol [39], which translates into an estimated 4 mg/dL change in plasma total cholesterol in the high-fat diet group and no change in the low-fat diet group. Thus, the decrease in dietary cholesterol observed in the groups taking psyllium cannot entirely explain the reduction in total plasma cholesterol (12 to 16 mg/dL). Furthermore, analyses at each study site showed the maintenance of LDL cholesterol reduction when dietary cholesterol remained unchanged during the study period. Thus, it is very likely that modifications in dietary cholesterol can explain only a small part, at best, of the LDL cholesterol changes observed with psyllium supplementation.

We have established that persons on a low-fat diet have a greater tendency to reach NCEP goals (see Figure 2) when they also receive psyllium. When our study participants were assessed for movement among NCEP LDL cholesterol categories, those who received psyllium showed a significant trend toward moving into subgroups associated with lower cholesterol values [P < 0.0001], whereas those who received placebo showed no change (P > 0.2) (see Figure 2). Alternatively, when psyllium recipients were distributed among the same risk subgroups, those on a low-fat diet showed a more significant trend toward moving into lower-risk categories [P < 0.0001] than did those on a high-fat diet (P = 0.04). For example, among psyllium recipients starting the study with an LDL cholesterol level above the 4.1 mmol/L (160 mg/dL) cut point, 39.3% (11 of 28) of those on a low-fat diet achieved an LDL cholesterol level of less than 4.1 mmol/L (160 mg/dL) compared with 35.3% (6 of 17) of those on a high-fat diet. Corresponding values among placebo recipients were 10.0% (6 of 30) and 15.4% (2 of 13) in the low-fat diet and high-fat diet groups, respectively. Psyllium recipients on the low-fat diet appeared to achieve the greatest benefit of the groups studied. Regardless of dietary regimen, psyllium recipients (37.8%) showed decreases in LDL cholesterol to less than the 4.1 mmol/L (160 mg/dL) level compared with 8 of 43 placebo recipients (18.6%).

We have shown that psyllium, in the form of Metamucil, a nonsystemic, nonconstipating fiber supplement, produces modest but significant improvement in total serum cholesterol and LDL cholesterol levels in persons on either high-fat or low-fat diets. Study participants on a low-fat diet who received psyllium showed a greater tendency to achieve LDL cholesterol reductions below the appropriate NCEP classification cut point. Psyllium therapy alone will produce, on average, a small degree of cholesterol lowering and may be adequate treatment only for particularly responsive persons. However, its use in combination with a low-fat diet should reduce the number of hypercholesterolemic persons who are placed on drug therapy and reduce doses when drug therapy is required.