- Click here for explanation of classification scheme.

To determine the effectiveness of consuming a combination of cholesterol-lowering foods under real-world conditions and with published data from the same participants who had undergone four-week metabolic studies to compare the same dietary portfolio with the effects of a statin. 

High LDL-cholesterol (>4.1mmol/L).

 

History of cardiovascular disease, diabetes or renal or liver disease.

Recruitment

Newspaper advertisements.

Design

Single phase one-year open-label study of a self-selected (ad libitum) dietary portfolio of cholesterol-lowering foods.

Intervention

Subjects followed a low-fat diet (mean macronutrient profiles close to current ATP III guidelines) incorporating four primary components of the dietary portfolio: 1.0g plant sterols/1,000kcal from a plant sterol ester-enriched margarine; ~10g viscous fibers/1,000kcal from oats, barley, psyllium, okra, and eggplant; 22.5g soy protein/1,000kcal from soy milk and soy meat analogues; and 23g whole almonds/1,000kcal. They were also asked to consume additional sources of plant protein and fiber in the form of dried legumes and to eat the recommended five to 10 servings of fruit and vegetables. Advice on eating a vegetarian diet without the dairy, eggs, or meat was given. Egg products were to consist of egg substitute and liquid egg whites. Any meat or dairy foods consumed were to be low-saturated-fat options. They purchased the foods from supermarkets and health food stores except for the bread and margarine. These were provided to all but four of the participants. 

Statistical Analysis

The results are expressed as means ± SEs. An intent-to-treat with last-observation-carried-forward analysis was performed for those subjects who did not complete the study. All statistical tests were performed at the P<0.05 significance level. The effect of change over time from baseline was assessed with the use of a generalized linear mixed model, with spatial power co-variance to adjust for unequal sampling periods. The repeated-measures analysis had percentage change from week zero as the response variable and week as the covariate predictor, with participant ID as the sole class variable. An analysis was also carried out with the use of only the 55 participants who completed the one-year study.  

For multiple pair-wise comparisons between the self-selected, metabolic portfolio, and statin treatments, the significance of the differences in percentage reduction from baseline was assessed by least-squares means with Tukey-Kramer adjustment. ANOVA F-test was performed to assess the effect of potential confounders, including sex and recruitment cohort, on LDL-cholesterol change. The subjects were divided into three groups which were defined on the basis of their LDL-cholesterol reductions from baseline to 52 weeks (group one, >20%, group two; 10-20%; and group three, <10%). The groupings represented the following: 1) The <10% LDL-cholesterol reductions that may be seen after fiber supplementation or the use of low-saturated-fat and low-cholesterol diets; 2) the 10-20% LDL-cholesterol reductions that may be observed at the high end of the range for highly effective dietary cholesterol reductions and the minimum reduction considered therapeutically significant in the pre-statin and early statin periods; and 3) the >20% LDL-cholesterol reduction that is typical of the effect seen in most of the large statin trials. 

Linear associations between mean LDL-cholesterol reduction and mean compliance measures were tested by using Pearson's correlation. Compliance was assessed for the four portfolio components (soy protein, plant sterols, viscous fibers, and almonds), where the prescribed amount represented 100%. Total compliance was the sum of the four individual compliances given equal weighting. 

Timing of Measurements

Weeks zero, two, four, eight, 12, 24, 32, 42, and 52 and one random blood sample between month four and five.

Dependent Variables

• Blood lipids were obtained after a 12-hour overnight fast. Serum was analyzed for total cholesterol, triacylglycerol, and HDL cholesterol according to a specific protocol after dextran sulfate-magnesium chloride precipitation. LDL cholesterol was calculated in mmol/L by the method of Friedewald.  

Independent Variables

The subjects kept a seven-day food record the week prior to each of the visits. The dietitian reviewed the records at each visit. The consumption of the four primary components of the diet were estimated from the records and expressed as a percentage of the amount recommended for that person's energy requirement. If a person felt unable to consume the full amount of one component, an attempt was made to ensure that increased amounts of the other three components were consumed. The diets were analyzed for macronutirients, fatty acids, cholesterol, and fiber with the use of a computer program based on US Department of Agriculture data. 

Control Variables

 

 

Initial N

31 men and 35 post-menopausal women

Attrition (final N)

55 (not described)

Age

59.3±1.1 year (range: 32-86)

Other relevant demographics

Anthropometrics
(e.g., were groups same or different on important measures)

• Four of the participants had untreated hypertension
• Eight subjects were taking anti-hypertensive medications at a constant dose throughout the study, four changed dosage during the study and one starting taking anti hypertensive medication at week 12 
• Eight women and one man were taking stable doses of thyroxine and had normal thyroid stimulating hormone concentrations 
• One female was taking a stable dose of estrogen replacement therapy.

Location

Toronto, Canada

 

 

Variables	Pre-study Measures and Confidence Intervals	Week 52 Measures and Confidence Intervals	Change	Statistical Significance of Difference*
Total Cholesterol	6.75±0.10	6.05±0.11	-0.70±0.09	P=<0.0001
LDL-C	4.48±0.09	3.87±0.10	-0.61±0.09	P=<0.0001
HDL-C	1.24±0.04	1.28±0.05	0.04±0.02	P=<0.0256
Triacylglycerol	2.29±0.14	1.98±0.11	-0.31±0.11	P=<0.0051
Total: HDL-C	5.76±0.16	5.03±0.16	-0.73±0.10	P=<0.0001
LDL-C: HDL-C	3.82±0.12	3.23±0.12	-0.59±0.08	P=<0.0001

 

*Significance of change was determined with Student's paired T-test

Other Findings

Compliance was satisfactory for almonds (mean ± SE: 78.8±3.2%) and for the plant sterol-enriched margarine (67.1±3.2%), whereas it was less adequate for viscous fiber (55.1±2.9%) and for soy protein (51.0±3.0%). At one year, only two subjects were following a vegan diet and five were following a ovolactovegetarian diet. Total protein intake remained similar to pre-study values (18.9%±0.5% of energy at pre-study vs. 20.0±0.4% at one y). Animal protein was reduced from 11.2±0.6% of total calories at pre-study to 5.6±0.5% (P=<0.001). There was a small significant weight loss over the year (-0.7±0.3kg; N=66: P=0.036). 

Near maximum reductions in LDL-cholesterol and the ratio of total to HDL-C were seen at 12 weeks and were sustained to one year. Mean LDL cholesterol was reduced at 12 weeks by 14.0±1.6% (P<0.001) and at one year by 12.8±2.0% (P<0.001). There were no significant differences in LDL-C reductions between the sexes or two recruitment cohorts. At one year, 31.8% of the participants showed LDL-C reductions >20%; 27% a 10-20% reduction and 40.9% a reduction <10%. 

Mean total compliance over one year and LDL-C change were significantly correlated (r=-0.42, P<0.001; N=65). Only two of the 26 participants who had a compliance <55% achieved an LDL-C reduction at one year that was >20%. Associations between compliance with individual dietary components and the change in LDL-C were also seen for consumption of soy (r=-0.52, P<0.001), fiber (r=-0.39, P=0.001), and almonds (r=-0.33, P-0.008).

A significant proportion of motivated persons can achieve a sustained and clinically meaningful reduction in LDL-cholesterol of >20% by dietary means. This degree of reduction was not significantly different from the response to a first-generation statin under metabolically controlled setting.