- Click here for explanation of classification scheme.

To assess the changes in the electrophoretic characteristics of LDL particles in response to a diet very low in saturated fat and incorporating simultaneously viscous fiber, soybean protein, plant sterols and almonds in hyperlipidemic subjects.

 

• Attending the Risk Factor Modification Center, St. Michael's Hospital, Ontario, Canada
• Taken part in previous dietary studies
• Experienced in following dietary protocols
• Previously had raised LDL cholesterol levels (>4.1mmol/L)
• No history of diabetes, renal disease or liver disease
• Not taking medications known to influence serum lipids.

   

   

   

 

No subjects had a history of diabetes, renal or liver disease and none were taking medications known to influence serum lipids.

Recruitment

Subjects were recruited from patients attending the Risk Factor Modification Center at St. Michael's Hospital in Ontario, Canada. 

Intervention

Subjects were followed on their own low-saturated fat therapeutic diets (followed National Cholesterol Education Program Step Two guidelines; <7% energy from saturated fat and <200 milligrams per day dietary cholesterol) for one week prior to the start of the study, and for an additional two weeks after the study on return to their low-saturated fat therapeutic diets. Dietary advice on low-saturated fat (<7% dietary calories) and low-cholesterol diets (<200 milligrams per day) had been reinforced on at least two occasions over the previous year, and at entry to the study. During the middle four weeks, subjects followed a combination diet in which all foods were provided at weekly clinic visits with the exception of fresh fruit and low-calorie vegetables (i.e., non-starch-containing vegetables) which subjects were instructed to obtain from their local stores. Subjects were provided with a seven-day rotating menu plan, including specified fruits and vegetables which they checked off each item as it was eaten and confirmed the weight of the foods. The same menu plan was used for all subjects but was modified to suit individual preferences, providing the goals for viscous fiber, soy protein, plant sterols, and almond consumption were met. Subjects were provided with self-taring electronic scales and asked to weigh all food items consumed during the study period. 

The aim of the combination diet was to provide 1g of plant sterols per 1,000kcals as an enriched margarine. The Unilever margarine contained approximately 46% sitosterol, 26% campesterol, 19% stigmasterol, 2.7% brassicasterol, 1.3% sitostanol, 0.8% campestanol, and 0.8% avenasterol, with the remainder made up of various other plant sterols. The Unilever margarine provided 12% plant sterol (w/w). Two grams of plant sterol was contained in 25g margarine, for 82kcal. In addition, the diet supplied 8.2g viscous fiber per 1,000kcal from oats, barley, and psyllium and 22.7g of soy protein per 1,000 kcal as soy milk or meat analogs. Raw or unblanched almonds also provided vegetable protein (2.9g per 1,000kcal). Emphasis was placed on eggplant and okra as additional sources of viscous fiber (0.55g per 1,000kcal and 0.67g per 1,000kcal, respectively). Thus, 200g of eggplant and 100g of okra were prescribed to be eaten on a 2,000kcal diet each day. 15g/4.2 MJ almonds were consumed as part of the diet. Diets were provided at a targeted intake to maintain body weight based on estimated caloric requirements.

Compliance was assessed from the completed weekly checklists and from the return of uneaten food items.

Statistical Analysis

Results were expressed as mean values with their standard deviations. Because no significant differences were found between weeks two and four, their mean value was used as the combination diet value. The significance of the differences between the baseline period (week zero), the combination diet (mean of weeks two and four) and the run-out period (week six) were assessed by the least square mean test with the Tukey multi-comparison adjustment using the PROC MIXED procedure. Values with a skewed distribution were log-normalized. There was no interaction between treatment and gender. Spearman rank correlation analysis was used to test for association between diet-induced changes in body weight and metabolic variables and changes in the various LDL electrophoretic characteristics. All analysis were conducted using the SAS software (version 8.2; SAS Institute Inc., Cary, NC, USA).

Timing of Measurements

Blood samples and body weights were obtained after 12-hour overnight fasts at weekly intervals and at week two of the washout. Seven-day weighed diet histories were obtained for the week prior to and for two weeks following the combination diet. Completed menu checklists were returned at weekly intervals during the four week combination diet period.

Dependent Variables

• Variable 1: Weight (kilograms)
• Variable 2: Total cholesterol (Lipid Research Clinics Protocols)
• Variable 3: LDL-cholesterol (calculated)
• Variable 4: HDL-cholesterol (Lipid Research Clinics Protocols)
• Variable 5: Triacylglycerol (Lipid Research Clinics Protocols)
• Variable 6: Apo B (measured by nephelometry)
• Variable 7: LDL particle size phenotype (obtained by non-denaturing polyacrylamide gradient [2-26%] gel electrophoresis from serum stored at -70^oC. LDL size was extrapolated from the relative migration of four plasma standards of known diameter. The estimated diameter for the major peak in each scan was identified as the LDL-peak particle diameter [LDL=PPD]. An integrated LDL diameter was also computed. This integrated LDL particle size corresponded to the weighed mean size of all LDL subclasses in one individual. Analysis of pooled plasma standards revealed that measurements of LDL-PPD and LDL mean particle size were highly reproducible, with inter-assay CV <0.6%. The relative proportion of LDL having diameter <25.5nm. The absolute concentration of cholesterol among particles with a diameter <25.5nm [LDL-cholesterol<25.5nm] was estimated by multiplying the total plasma LDL-cholesterol levels by the relative proportion of LDL with a diameter <25.5nm as described previously. A similar approach was used to estimate the relative and absolute concentration of cholesterol in particles with a diameter >26.0nm [LDL%>26.0nm and LDL-cholesterol>26.0nm respectively]. The CV for the measurements of LDL%<25.5nm and LDL%>26.0nm were 12.0 and 9.3% respectively. 

Independent Variables

• Combination diet as described earlier (measured by seven-day weighed histories, menu checklists, food returned at visits)
• Low-saturated fat (<7% dietary calories) and low-cholesterol diets (<200 milligrams per day) (recorded diets).

 

 

Initial N

13 subjects (seven men and six post-menopausal women)

Attrition (final N)

12 (six men and six post-menopausal women, one subject completed only three weeks and withdrew because of dyspepsia associated with Helicobacter pylori infection requiring antibiotic therapy)

Age

65±3 years (median 64 years; range 43-84)

Ethnicity

Not reported

Anthropometrics

BMI 25.6±0.9kg/m² (median 26.1kg/m²; range 20.6-30.7kg/m²); baseline LDL cholesterol 4.22±0.11mmol/L (median 4.27mmol/L; range 3.51-4.99mmol/L)

Location

St. Michael's Hospital, Ontario, Canada

Summary of Findings

 

	Baseline (week zero)		Mean Treatment (weeks two to four)		Run-out (week six)		Statistical Significance of Effect: P
Variable	Mean	SD	Mean	SD	Mean	SD	Treatment vs. Baseline	Run-out vs. Baseline
Weight (kg)	70.0	13.7	69.5	13.1	68.7	13.0	0.02	<0.0001
Total Cholesterol (mmol/l)a	6.37	0.80	4.97	0.76	5.95	0.82	<0.0001	<0.05
LDL-Cholesterol (mmol/l)	4.20	0.42	2.95	0.63	3.82	0.76	<0.0001	0.04
HDL-Cholesterol (mmol/l)	1.38	0.41	1.33	0.43	1.34	0.44	0.44	0.55
Triacyglycerol (mmol/l)b	1.87	1.31	1.52	0.76	1.72	1.01	0.04	0.34
Apo B (g/l)	1.31	0.17	1.00	0.19	1.26	0.20	<0.0001	0.25
LDL particle size phenotype
LDL-peak particle diameter	2.55	5.3	255.4	4.3	256.3	4.7	0.98	0.33
LDL-integrated size	257.0	4.0	255.7	2.9	256.0	3.9	<0.01	0.18
LDL >26.0 (%)c	37.8	13.4	33.9	8.0	38.2	10.0	<0.05	0.85
LDL 25.5-26.0 (%)c	19.1	3.1	19.6	3.8	20.4	5.1	0.35	0.35
LDL <25.5 (%)c	43.1	15.8	46.5	11.3	41.5	14.1	0.05	0.63
LDL-Chol 26.0 (mmol/l)	1.55	0.49	0.99	0.24	1.45	0.45	<0.0001	0.43
LDL-Chol 25.5-26.0 (mmol/l)	0.80	0.13	0.57	0.14	0.78	0.27	<0.0001	0.81
LDL-Chol 25.5 (mmol/l)	1.85	0.82	1.39	0.56	1.59	0.68	<0.01	<0.05

Other Findings

Compliance in terms of energy intake was good in most subjects, with a mean value of 92.5 (SD 2.9%) of the prescribed energy consumed. Mean energy intake was greater on combination diet compared with the baseline and run-out values. Subjects lost body weight at an average rate of -10 (SD 0.05) kilograms per week on the experimental combination diet and -0.20 (SD 0.05) kilograms per week during the run-out phase.  

Significant reductions in blood lipids were seen during the experimental diet compared with the baseline, including reductions from baseline in LDL-Chol levels [30.0 (SE 3.0)%, P<0.0001) and the total Chol:HDL-Chol ratio [18.6 (SE 3.1)%, P<0.0001). 

The combination diet had no significant effect on LDL-PPD. On the other hand, the LDL integrated size, which reflects the whole distribution of LDL based on all subclasses in a given individual, was significantly reduced (P<1.01) following the combination diet, suggesting a shift in the distribution of LDL from larger to smaller species. This was confirmed by densitometric analysis of the relative proportion of small and large LDL particles. The combination diet was associated with a reduction in LDL%>26.0nm (from 37.8 to 33.9%, P<0.05) and with a diametrical increase in LDL%<25.5nm (from 43.1 to 46.5%, P=0.05). Both LDL%<25.5nm and LDL%<26.0nm came back to baseline values during the run-out phase of the study.

By comparison with baseline levels, the experimental diet reduced the absolute concentrations of all three distinct LDL subclasses, including LDL-Chol<25.5nm. Thus, the reduction in LDL-Chol>26.0nm, LDL-Chol25.5-26.0nm and LDL-Chol<25.5nm averaged -0.57 (SD 0.37)mmol/l (P<0.0001), -0.23 (SD 0.09)mmol/l (P<0.0001) and -0.45 (SD 0.51)mmol/l (P<0.01) respectively (total LDL-Chol reduction 1.25mmol/l, P<0.001). There was no correlation between diet-induced changes in body weight and changes in any of the LDL electrophoretic characteristics measured in the present study. However, multivariate adjustment for diet-induced variation in body weight attenuated to some extent the impact of the combination diet on LDL%<25.5nm and LDL%26.0nm (P=0.07 and P=0.08 respectively). Diet-induced changes in plasma TG levels were associated with simultaneous changes in LDL-PPD (r -0.52, P=0.08), but not with changes in the relative proportion of small (r 0.38, P=0.22) or large LDL (r -0.46, P=0.13).

Sub-group analysis based on the median distribution of LDL-Chol<25.5nm at baseline (1.72mmol/l) indicated that the group with LDL-Chol<25.5nm >1.72mmol/l at baseline (N=6) comprising a greater proportion of men, had higher plasma LDL-Chol (+0.54mmol/l) and ApoB (+23g/l) levels and smaller LDL-PPD (-7.4 Å) compared with subjects with LDL-Chol<25.5nm <1.72mmol/l at baseline (N=6); they also tended to have higher TG levels (+1.27mmol/l) and lower HDL-Chol levels (-0.35mmol/l). Baseline LDL-Chol<25.5nm levels appeared to be a significant determinant of the experimental diet effect on LDL size phenotype. Subjects with increased LDL-Chol<25.5nm levels at baseline experienced a significant reduction (P<0.001 in LDL-Chol<25.5nm concentrations and showed an increase in LDL-PPD following the experimental diet. Total plasma LDL-Chol levels were also reduced significantly in this group of subjects (-31.4%, P<0.01). The further increase in LDL-PPD during the washout phase in subjects with LDL-Chol<25.5nm levels <1.72mmol/l may be explained by the fact that these subjects showed a greater decrease in body weight (run-out vs. baseline) compared with those with LDL-Chol<25.5nm <1.72mmol/l at baseline (1.53 vs. 0.96kg respectively). On the other hand, in subjects with relatively lower LDL-Chol<25.5nm levels at baseline, the diet had virtually no impact on LDL-Chol<25.5nm levels and LDL-PPD was reduced by 0.3nm. However, the magnitude of the diet-induced reduction in total plasma LDL-Chol levels (-28.6%, P<0.0001) was not attenuated in this subgroup of subjects compared with individuals with high levels of LDL-Chol<25.5nm at baseline.

 

A dietary combination of foods that have been shown to induce marked reduction in plasma LDL-Chol levels may also favorably alter important aspects of the LDL particle size phenotype, including LDL-Chol<25.5nm levels, particularly in subjects with a sub-optimal phenotype at baseline. These results need to be replicated in larger samples of subjects and for a longer time period in order to address effectiveness of the diet over the longer-term, when availability of foods and compliance are major issues. Nevertheless, the magnitude of the reduction in plasma LDL-Chol levels, which essentially matched that achieved using pharmacological therapy, combined with the favorable changes in small dense LDL-Chol levels (LDL-Chol<25.5nm), identify this diet as a potentially potent dietary strategy to reduce the risk of CVD in high-risk individuals.

Government:

Canadian Federal Government

Study did not examine independent effect of almonds, only effect of almonds as part of a combination diet that also included enriched margarine, oats, barley, psyllium, and soybean protein. 

The authors point out in the discussion that "the impact of almonds or other nuts on LDL size is unclear."