Lipoprotein Lipid Response to the National Cholesterol Education Program Step II Diet by Hypercholesterolemic and Combined Hyperlipidemic Women and Men

Methods

Study Design

The study, shown schematically in Fig 1⇓, consisted of recruitment, which involved two fasting lipid measurements, an orientation, and randomization; a baseline evaluation; dietary instruction; and 2 years of follow-up involving six visits. The current analyses are limited to the first 6 months of follow-up, the diet therapy trial period recommended by the NCEP. During recruitment, subjects with an elevated LDL cholesterol at the initial blood draw and without other exclusionary criteria were invited to a second eligibility blood draw. Those with an elevated LDL cholesterol after the second blood draw were invited to an orientation session conducted by a study dietitian who described the study and taught participants how to keep food records. Study participants signed an informed consent approved by the University of Washington Human Subjects Institutional Review Board. Hypercholesterolemia was defined as an elevated LDL cholesterol at both screening evaluations; combined hyperlipidemia further required triglycerides to be at or above the age- and sex-specific 75th percentile value at one or both evaluations.²⁴ A stratified random sampling scheme was used to allocate half of the participants to an immediate and half to a 6-month-delayed intervention group. The delayed intervention group was included to demonstrate that plasma lipid changes were due to the dietary intervention and not to regression to the mean, to concurrent changes in the general population or the workforce from which subjects were drawn, or to lifestyle changes stimulated by participating in study screening and enrollment. The baseline evaluation consisted of assessing lifestyle habits and medical history via questionnaire, assessing dietary intake via food record, and analyzing plasma lipids. At this visit, participants were informed of their randomization group. Subjects in the delayed group had a baseline visit, then a brief visit without a dietitian interview or a food record 3 months later. Subjects in the delayed group were encouraged but not required to postpone dietary changes. Six months after their baseline visit and just before starting dietary instruction, delayed subjects had a second baseline visit with the same assessments as at the first visit.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Schematic design of the beFIT study through 6 months after start of dietary instruction.

Diet instruction consisted of eight weekly classes of nutrition information and behavior modification to achieve and maintain an NCEP step II diet (<30% of calories from total fat, <7% from saturated fat, and <200 mg/d dietary cholesterol).²⁵ Classes included lecture, discussion, and in-class and at-home activities. Subjects were assigned a food group pattern for their calorie needs and tallied their intake of featured food groups each week. Activities included selecting from menus, rating foods by their labels, stocking “quick to fix” meals, and planning for difficult situations (usually social events). Spouses were invited to attend. Subjects were required to attend a minimum of five classes. The immediate group received dietary instruction soon after the baseline assessment, and classes for the delayed group started ≈6 months later.

All participants had individual follow-up visits 1 and 2 at 3 and 6 months after the start of diet instruction, which consisted of a fasting blood draw, weight measurement, a lifestyle and medical history questionnaire, a 4-day food record, and counseling with a dietitian (Fig 1⇑).

Results

An initial cholesterol screening was performed on 9770 employees. After the first screening visit, 1974 subjects (20.2%) had an elevated LDL cholesterol and no other exclusionary conditions and were eligible for a second blood draw; 1549 (79%) had this blood test. After the second screening test, 1138 (73.5%) were eligible for the study, and 692 (61%), 334 women and 358 men, consented and were randomized to the immediate and delayed groups. Before the 6-month visit, 238 subjects dropped out of or were excluded from the study: 53 for failure to attend a baseline visit or at least 5 classes, 10 for pregnancy, 9 for high thyroid-stimulating hormone, 107 because of time and logistical constraints, 11 due to problems associated with the diet, 9 for taking lipid-altering medications, 16 to seek other therapy, and 23 for other reasons. Of the dropouts, 55% were women, 53% were combined hyperlipidemic, and 54% were randomized to the delayed intervention group. Forty-five of the continuing subjects missed all or part of their 6-month visit. At baseline, these 45 subjects had higher body mass indexes and reported higher dietary cholesterol intake but were similar in age, lipoprotein lipids, and other nutrients compared with subjects with 6-month data. The following analyses include 409 subjects with complete lipid and dietary data at baseline and 6 months after intervention.

Effect of the Diet in Intervention Versus Control Groups

Subjects randomized to the immediate (n=217) and delayed (n=192) groups were similar in age, anthropometric measures, blood pressure, and demographic characteristics. At visit 0, the only baseline observation for the immediate group and the first baseline observation for the delayed group (see Fig 1⇑), the two groups reported similar dietary intakes of calories: total fat, 34% kcal; saturated fat, 12% kcal; dietary cholesterol, 258 and 272 mg/d; and carbohydrates, 47% and 48% kcal. At visit 2, 6 months after dietary intervention began, the immediate group reported significant reductions in calories, total and saturated fat (to 25.2% and 7.6% kcal), and dietary cholesterol (to 171 mg/d) and an increase in carbohydrates to 55.2%. Weight decreased 2.7 kg. After the 6-month control period in the delayed group, the only significant dietary change was a decrease of 129 in calories. Weight did not change. Six months after dietary intervention began, the delayed group reported changes in nutrient intakes similar to those made by the immediate group, and weight loss (2.5 kg) was similar. These findings demonstrate that minimal dietary change occurred in the delayed group during the control period and significant changes occurred after the dietary intervention.

The plasma lipoproteins were similar between the immediate and delayed groups at visit 0. Six months after diet instruction, the immediate group had significant decreases in total (0.41 mmol/L, 16 mg/dL), LDL (0.36 mmol/L, 14 mg/dL) (Fig 2⇓), and HDL (0.05 mmol/L, 2 mg/dL) cholesterol. No significant changes were observed for the delayed group after the 6-month control period in total (−0.08 mmol/L, −3 mg/dL), LDL (−0.03 mmol/L, −1 mg/dL) (Fig 2⇓), and HDL (0 mmol/L) cholesterol. Six months after their diet instruction, the delayed group had significantly lower total, LDL (Fig 2⇓), and HDL cholesterol, and the changes, −0.47, −0.49, and −0.02 mmol/L (−18, −19, and −2 mg/dL), respectively, were similar to those in the immediate group. Since the delayed group did not have significant changes in diet or lipids during the control period and had changes similar to those of the immediate intervention group 6 months after diet instruction, the two groups are combined in the remaining analyses. Visit 0 is used for baseline for the immediate and visit D2 is used for the delayed group.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Mean and 95% CI of LDL cholesterol at baseline and 3 and 6 months after diet instruction and during the control period for the immediate (⋄) and delayed (▪) intervention groups, respectively.

Effect of Dietary Intervention by Sex and Hyperlipidemia

Characteristics of women and men with hypercholesterolemia or combined hyperlipidemia are shown in Table 1⇓. Body mass index was higher in combined hyperlipidemic women and men compared with their hypercholesterolemic counterparts and higher in combined hyperlipidemic women than men. A greater percentage of men than women were married and had college degrees. The majority of subjects were white, and few smoked cigarettes. Approximately one third of the women were postmenopausal.

View this table:

• View inline
• View popup

Table 1.

Preintervention Characteristics of Women and Men With Hypercholesterolemia or Combined Hyperlipidemia

Nutrient intakes and weight at baseline and 6 months after intervention are shown in Table 2⇓. At baseline, women reported lower calories and alcohol. Women and men in both the hypercholesterolemia and combined hyperlipidemia groups significantly lowered their intakes of calories; total, saturated, monounsaturated, and polyunsaturated fats (% kcal); and dietary cholesterol. Carbohydrate and fiber intakes were increased in all groups. There were no differences in the amount of change between women and men or between hypercholesterolemic and combined hyperlipidemic groups. Each group experienced significant weight loss. Hypercholesterolemic men lost more weight on average than did hypercholesterolemic women, 3.4 versus 1.6 kg, despite reporting similar reductions in caloric intake. Weight loss was intermediate in combined hyperlipidemic women and men, 2 and 3 kg, respectively.

View this table:

• View inline
• View popup

Table 2.

Nutrient Intake From 4-Day Food Records and Body Weight for Women and Men With Hypercholesterolemia or Combined Hyperlipidemia Before and 6 Months After Starting the NCEP Step II Diet

Lipoprotein lipid values before subjects began the NCEP step II diet are shown in Table 3⇓. Total cholesterol was different between the hypercholesterolemic and combined hyperlipidemic subjects of each sex. LDL cholesterol was not different by sex or lipid disorder. HDL cholesterol was lower in combined hyperlipidemic women and men compared with hypercholesterolemic subjects. Women in both lipid disorder groups had higher HDL cholesterol than men in the respective groups. By definition, triglycerides were higher in the combined hyperlipidemic groups.

View this table:

• View inline
• View popup

Table 3.

Lipoprotein Lipids in Women and Men With Hypercholesterolemia or Combined Hyperlipidemia Before and 6 Months After Starting the NCEP Step II Diet

After 6 months of dietary intervention, total and LDL cholesterol were significantly lower in all groups (P<.001) (Table 3⇑). The LDL cholesterol reduction was the least in hypercholesterolemic women, intermediate in combined hyperlipidemic women and men, and greatest in hypercholesterolemic men (Fig 3⇓). The amount of change was not different for total or LDL cholesterol when compared between women and men with the same lipid disorder, between women with different lipid disorders, or between men with different lipid disorders. HDL cholesterol was significantly decreased in hypercholesterolemic and combined hyperlipidemic women (Table 3⇑). The reduction was significantly different for hypercholesterolemic women, 0.11 mmol/L (−4.3 mg/dL) (−6.4%), compared with both hypercholesterolemic men, −0.03 mmol/L (−1.0 mg/dL) (−1.3%), and combined hyperlipidemic men, −0.04 mmol/L (−1.4 mg/dL) (−2.7%) (Fig 3⇓). No significant changes within persons or among the four groups were observed in total triglycerides.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Mean and 95% CI for LDL (▪) and HDL (⋄) cholesterol 6 months after diet instruction in hypercholesterolemic (HC) and combined hyperlipidemic (CHL) women and men. *Statistically significant reduction from baseline, P<.01.

Effect of Diet on LDL Cholesterol Goal

Recommendations of the NCEP Adult Treatment Panel II² for LDL cholesterol lowering are based on history of CHD and other risk factors. The positive risk factors include age ≥45 years (men), age ≥55 years or premature menopause (women), family history of premature CHD, smoking, hypertension, diabetes, and HDL cholesterol <35 mg/dL (0.91 mmol/L); HDL cholesterol >60 mg/dL (1.55 mmol/L) is considered a negative risk factor. The LDL cholesterol target values are <100 mg/dL (2.59 mmol/L) with a personal history of CHD, <130 mg/dL (3.36 mmol/L) with two or more risk factors, and <160 mg/dL (4.14 mmol/L) with fewer than two risk factors. Drug therapy may be considered if LDL cholesterol values are 30 mg/dL (0.78 mmol/L) above these goals after diet therapy. Fig 4⇓ presents the percentages of beFIT subjects, at baseline and 6 months after diet instruction, whose LDL cholesterol was within their goal, within 30 mg/dL of their goal, or >30 mg/dL above their goal. At visit 0, 30% and 29% of hypercholesterolemic and combined hyperlipidemic women, respectively, and 37% and 27% of hypercholesterolemic and combined hyperlipidemic men had LDL cholesterol values >30 mg/dL above their goal. Six months after diet instruction, 19% and 17% of women and 17% and 16% of men fell into this category and were candidates for drug therapy. At visit 0, 27% and 43% of hypercholesterolemic and combined hyperlipidemic women and 50% and 52% of hypercholesterolemic and combined hyperlipidemic men, respectively, were within 30 mg/dL of their goal. After diet, 23% and 32% of women and 31% and 32% of men were in this category, with continued recommendations for dietary restrictions. After 6 months of diet, >50% of each group had LDL cholesterol values in the desirable range based on the NCEP goals.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Percentage of subjects in each of the NCEP LDL cholesterol goal categories at baseline and 6 months after diet instruction in hypercholesterolemic (HC) and combined hyperlipidemic (CHL) women and men.

Discussion

The beFIT was the first large randomized field trial of at least 6 months' duration in moderately hypercholesterolemic women and men to test the efficacy of the NCEP step II diet (<30% kcal total fat, <7% kcal saturated fat, and <200 mg/d dietary cholesterol). The results presented answer four questions. First, teaching this NCEP diet to free-living subjects is a viable strategy for dietary intervention. Second, fat and cholesterol intakes and LDL cholesterol were significantly decreased after 6 months of dietary intervention. Third, LDL cholesterol reductions were similar between women and men, but HDL decreased more in women. Fourth, subjects with hypercholesterolemia only and those with combined hyperlipidemia had similar lipoprotein changes.

Our experience demonstrates that teaching dietary and behavioral modifications to free-living hypercholesterolemic persons can be an effective therapeutic strategy. All participants had LDL cholesterol elevations and thus were candidates for diet therapy. Subjects in the immediate intervention group made dietary changes and significantly lowered their LDL cholesterol. Subjects randomized to the delayed group had similarly elevated LDL cholesterol levels, which did not change until participation in the diet intervention. The lack of change in lipids during the control period demonstrates that postintervention lipid changes were not due to regression to the mean or to changes caused by other than the dietary intervention.

LDL cholesterol reductions of 7.6% to 8.8% after 6 months of dietary changes in this study were statistically significant. The NCEP estimates 8% to 14% reductions² on the basis of metabolic ward studies. The largest attainable results may be the 16% and 18% reductions reported by Schaefer et al,¹⁰ when food was provided and adherence was 100%.

The differences between beFIT and previous studies may be attributed to differences in subject selection, duration of the intervention, dietary compliance, weight change, or initial level of fat intake. Subjects in this study were recruited at their workplace. They were not repeat participants of cholesterol-lowering studies or patients at lipid clinics, were not familiar with dietary modifications to lower cholesterol, and were not selected to maximize compliance as reported elsewhere.¹² Subjects were not excluded from beFIT on the basis of elevated triglycerides, smoking, alcohol consumption, menopausal status, or hormone use as done in other studies. They were neither selected or excluded for existing CHD.

Our results are similar to or slightly better than responses seen in shorter-term studies of moderately hypercholesterolemic subjects who were taught NCEP^{12 13} or AHA^{9 32 33} diets and self-selected their foods. In those studies, the minimum LDL cholesterol response, −3.5%, was at 6 weeks in subjects with mild hypercholesterolemia.³² The reductions in beFIT were most similar to the 8% reductions in a 4-month study of hypercholesterolemic men.¹³ The greatest response, −12.4% after 3 months, was reported for men after coronary artery bypass graft surgery, who were apparently very motivated, in that they were taught the NCEP step I but achieved the step II diet goals.³³

Assessing dietary intake of free-living subjects who self-select their foods is known to be problematic; biased reporting is especially a concern in an intervention study like this. By use of the doubly labeled water technique, a precise measure to evaluate calorie expenditure, food records have been found to underestimate intake by ≈20%.³⁴ Underreporting is more common among women and the overweight.^{34 35} Calculation of nutrients such as fatty acids is made more difficult by the variety of products available and the increased use of convenience and restaurant foods. We used 4-day food records to document dietary intake and included procedures to maximize their accuracy. At baseline, beFIT subjects reported consuming 34% total fat and 11.5% saturated fat, matching the average American adult intake reported in NHANES III of 34% total fat and 12% saturated fat.³⁶ At 6 months, mean intake in beFIT subjects was 25% kcal from total fat, 7% to 8% from saturated fat, and <200 mg/d dietary cholesterol, suggesting that saturated fat is the most difficult to achieve of the step II diet criteria. The Keys equation predicts that the dietary change reported would lower total cholesterol 0.26 to 0.31 mmol/L (10 to 12 mg/dL), whereas actual reduction was 0.36 to 0.47 mmol/L (14 to 18 mg/dL).³⁷ The additional change may be due to weight lost during the 6 months, which averaged 2 to 3 kg, and perhaps the small increase in dietary fiber.

The recommendations of diet therapy for hypercholesterolemia have been extended to women even in the absence of comprehensive dietary efficacy studies in hyperlipidemic women. In normolipidemic women, several studies evaluating the effects of <30% kcal total fat diets have found decreases of 7% to 14% in total cholesterol,^{16 17 18} of 5% to 13% in LDL cholesterol, and of 8% and 12% in HDL cholesterol^{17 18} and both decreases¹⁶ and increases^{17 18} in triglycerides. In hyperlipidemic women, several studies have reported inconsistent lipid responses to <30% fat diets,^{9 10 19 20} including decreases of 5% to 9% in total cholesterol,^{9 19 20} of 6% to 13% in LDL cholesterol,^{9 10 20} and of 4% to 15% in HDL cholesterol.^{9 10 19 20} Triglycerides were reported to be unchanged,⁹ increased 30%,¹⁹ or decreased 11%.²⁰ The lipoprotein cholesterol reductions in beFIT women were in the ranges reported in the above studies. Further analyses and studies will be necessary to determine whether lipoprotein response to diet is related to menopausal status or hormone use. Approximately one third of the women in our study were postmenopausal, and 15% and 17% of the premenopausal and postmenopausal women, respectively, used hormones.

In this study, women and men had similar total and LDL cholesterol responses as they have in most other studies. Geil et al⁹ reported no sex difference between 99 men and 63 women in response to the AHA step 1 diet (with higher total and saturated fat intakes than the beFIT). Mensink and Katan^{7 8} reported similar total and LDL cholesterol responses by normolipidemic women and men to low-saturated-fat diets enriched with monounsaturated and polyunsaturated fats. Ferro-Luzzi et al⁶ found that switching from a high-monounsaturated- to high-saturated-fat diet resulted in 15% and 19% increases in total and LDL cholesterol for both normolipidemic women and men.⁶ On the other hand, only the metabolic ward study by Schaefer et al¹⁰ has compared responses by women and men to the NCEP step II diet. They reported a greater LDL cholesterol response in 12 men (−20.7%) than in 12 women (−13.4%) and similar HDL cholesterol changes (−14.5% and −15.3%, respectively).

We observed a differential response by sex for HDL cholesterol. These sex differences were not due to differential triglyceride responses, weight loss, or changes in exercise. Both hypercholesterolemic and combined hyperlipidemic women had statistically significant reductions (−6.4% and −4.5%), whereas there was no significant change in men. The reductions were statistically significantly greater in hypercholesterolemic women than in hypercholesterolemic or combined hyperlipidemic men. Since beFIT was a study of the step II diet, changes in both total fat and fat composition (P/S ratio) occurred. Differential HDL cholesterol changes have been reported for women and men in studies in which both total fat and fat composition were modified. A meta-analysis reported a differential HDL cholesterol response by normolipidemic women and men in crossover studies of low-P/S-ratio to high-P/S-ratio diets.³⁸ Mensink and Katan^{7 8} found that men had greater decreases of HDL cholesterol than women in response to low-saturated-fat diets with different P/S ratios. Clifton and Nestel³⁹ added a 31-g-fat (56% saturated) and 650-mg-cholesterol supplement to a 25% fat diet and found greater HDL₂ but not HDL or HDL₃ cholesterol increases in women than in men. Ferro-Luzzi⁶ found that switching from a high-monounsaturated- to high-saturated-fat diet resulted in no change in men and a significant 19% increase in HDL cholesterol for women, which was reversed when they switched back to their usual “Mediterranean” diet. The similar LDL but differential HDL cholesterol changes between women and men in response to saturated fat feeding and restriction seen in the Ferro-Luzzi⁶ and beFIT studies suggest that women's metabolism might be more sensitive to dietary saturated fat, responding with more efficient LDL cholesterol removal. The significance of the greater HDL cholesterol decrease in women is not clear. Low HDL cholesterol is an independent CVD risk factor in women consuming Western diets.^{40 41 42} Whether alternative dietary regimens that might maintain higher HDL cholesterol levels would offer long-term advantages is unknown.

It is important to know whether response to cholesterol-lowering diets is different in persons with combined hyperlipidemia, because they may be at greater CVD risk than hypercholesterolemic persons with normal triglyceride levels.^{21 22 23} The Helsinki Heart Study²³ showed that the combined elevation of cholesterol and triglyceride is a useful marker for CHD risk. In beFIT, there was no difference in LDL cholesterol lowering between hypercholesterolemic and combined hyperlipidemic women and men with similar reported dietary intakes. This contrasts with our previous work, which found that combined hyperlipidemic men had less LDL cholesterol response than hypercholesterolemic men (−3% to −7% versus −5% to −13%) at reported intakes of <30% kcal total fat and 7% to 8% kcal saturated fat.¹¹ These apparently conflicting results may be explained by differences in subject characteristics; beFIT men were younger, weighed more, and had a lower baseline fat intake, higher triglyceride, and lower HDL cholesterol. Another important reason for comparing responses in hypercholesterolemic and combined hyperlipidemic subjects is that low-fat diets with increased carbohydrates potentially induce a triglyceride increase^{7 43 44 45} and HDL cholesterol decrease,⁴⁶ which could be exacerbated when triglyceride is already elevated. Both of these conditions are associated with CVD mortality,⁴⁷ CHD,^{23 48} and coronary artery disease.^{49 50 51} Carbohydrate intake increased moderately, from ≈48% to 56% kcal, with no statistically significant effect on triglycerides. This finding concurs with our previous work in free-living combined hyperlipidemic men indicating that carbohydrate intake <60% of calories is not associated with induced hypertriglyceridemia.⁵² We do not extrapolate these findings to more extreme diets or to more severely hypertriglyceridemic persons.

On average, subjects achieved total fat and cholesterol goals for the step II diet; however, the mean intake of saturated fat was not <7% but just above 7%. To determine how many subjects achieved the step I and II diets and whether there was differential LDL cholesterol response, subjects were classified as achieving the step II or step I or not achieving at least a step I diet on the basis of the 4-day food record 6 months after diet instruction. Forty-one percent of subjects achieved a step II diet, 30% achieved a step I, and 29% did not meet step I criteria. The LDL cholesterol reductions in each group were 9.9%, 7.9%, and 6.2%, respectively. These findings indicate that a majority of individuals taught a step II diet can maintain total fat <30% kcal and saturated fat <10% kcal, meeting at least the step I criteria. More importantly, subjects taught a step II diet, regardless of how compliant, experience on average some LDL cholesterol reduction, and the greater the dietary restriction the greater the LDL cholesterol reduction.

This was an intent-to-treat analysis and did not adjust for compliance, weight loss, or lifestyle changes. Subjects were free-living and self-selected their foods and represent a cross section of the general population that would be likely to seek general clinical care for hyperlipidemia. These subjects received more dietary instruction and support than usually provided. Consequently, the amount of LDL lowering achieved may represent about the best that can be expected through community-based or intensive clinical programs. Through modest LDL reductions averaging 8%, the majority of subjects attained the desirable range for LDL cholesterol defined by the NCEP guidelines. Fewer women were defined as being above their LDL cholesterol goal, perhaps because age is less likely to be a risk factor for women than for men (55 versus 45 years), and women's HDL cholesterol is less likely to be below the 35-mg/dL (0.91 mmol/L) cutoff point defined as a risk factor. Nevertheless, the diet intervention was effective in that <20% of individuals in each sex and lipid disorder group remained in the range for drug therapy 6 months after intervention.

In summary, hypercholesterolemic and combined hyperlipidemic women and men taught the NCEP step II diet and followed for 6 months reported similar dietary changes and achieved similar LDL cholesterol reductions. However, HDL cholesterol was lowered in women but not in men. There was no adverse effect on triglycerides. The explanation and significance of the HDL decrease is unknown, and more work is needed to determine the factors responsible, whether the decrease is sustained, and whether it is deleterious. We conclude that teaching the NCEP step II diet can effectively reduce both LDL cholesterol and CVD risk and that the diet is appropriate for moderately hypercholesterolemic and combined hyperlipidemic women and men.

Selected Abbreviations and Acronyms