Effect on lipoprotein profile of replacing butter wit margarine in a low fat diet: randomised crossover study with hypercholesterolaemic subjects

Introduction

Dietary recommendations aimed at reducing the risk of coronary heart disease require an appreciable reduction in saturated fatty acids from present intake in most Western countries.1 2 A major justification for this recommendation is the association between saturated fatty acids and total and low density lipoprotein cholesterol.3 n-6 Polyunsaturated fatty acids and n-9 monounsaturated fatty acids are widely recommended as replacements for saturated fatty acids.4 It has recently become apparent that high intakes of n-6 polyunsaturated fatty acids are associated with decreased concentrations of high density lipoprotein cholesterol5 and that trans forms of unsaturated fatty acids may increase the risk of coronary heart disease6, possibly by raising total and low density lipoprotein cholesterol and Lp(a) lipoprotein and reducing high density lipoprotein cholesterol.7 8 Health professionals and the public have become uncertain about implementing nutrient recommendations and about the optimal sources of dietary fat.

Margarines rich in unsaturated fatty acids are often recommended as replacement fats for butter, which is relatively high in saturated fatty acids. Most margarines contain at least some trans unsaturated fatty acids,8 and this has led to the suggestion that their use might be inappropriate in diets aimed at reducing the lipoprotein mediated risk of coronary heart disease. To assess the effects on lipid and lipoprotein concentrations, we compared butter and a margarine high in monounsaturates used as a source of “hard” fat in the context of an overall reduced fat diet.

Subjects and methods

Fifty six subjects with total cholesterol concentrations in the range 5.5-7.9 mmol/l were recruited by newspaper advertisements and from those who had participated in previous studies. People with familial or secondary hyperlipidaemia or taking drugs known to influence lipid metabolism were excluded; thus most subjects had polygenic hyperlipidaemia. All subjects gave informed consent on the understanding they could withdraw at any time from the study, which was approved by the Ethics Committee (Otago) of the Southern Regional Health Authority. Forty nine subjects (35 women, 14 men), mean age 46.8 (SD 11.6) years, completed the 21 week protocol. Initial mean body mass index was 26.4 (3.8) kg/m2. At the time of recruitment mean concentrations of total cholesterol, low density lipoprotein cholesterol, and high density lipoprotein cholesterol were 6.5 (0.6), 4.2 (0.7), and 1.3 (0.3) mmol/l.

During a four week run in period participants continued their usual diet and completed a four day food record for assessment of usual food and nutrient intakes. They also completed questionnaires concerning past, present, and family medical history as well as use of medications. At the end of the run in period they were randomised to groups who were asked to follow either the butter diet or the margarine diet for six weeks. After the first experimental period, followed by a five week washout period (during which they were asked to revert to their usual diet), they then followed the alternative diet for a final six week period. Two blood samples were collected after an overnight fast during the last week of the run in period, the last week of the washout period, and the last week of each six week experimental period. Blood specimens were separated by centrifugation at 1500 g. Lipid and lipoprotein concentrations were measured in each blood sample, and apolipoproteins A I and B, Lp(a) lipoprotein, and triglyceride fatty acids were measured in one sample during each period.

The experimental diets were individually prescribed and based on the energy intake calculated from the self selected diets reported during the run in period. They were designed so that protein, carbohydrate, and fat provided approximately 19%, 55%, and 26% of total energy, respectively. The two experimental diets differed only in the source of “hard fat” used for baking, cooking, and spreading. Total saturated fatty acids were calculated to provide 12% and 6% of total energy in the butter and margarine diets respectively and total polyunsaturated fatty acids 3% and 7% (see table 3). Table 1 shows the fatty acid composition of butter and margarine (a standard brand widely available in New Zealand and comparable with margarines made from canola oil available in the United Kingdom, North America, and most other countries). The only other nutrient that differed as a consequence of the differing fat source was cholesterol (32 mg/MJ in the butter diet and 23 mg/MJ in the margarine diet).

Detailed dietary information and reinforcement were provided during personal and telephone interviews initially and at regular periods throughout the study. Individual instruction booklets containing practical details which translated the nutrient requirements of the experiment into advice about food were prepared for each participant. Butter and margarine were provided for the participants and their families. Compliance was measured by four day food records completed during the last week of each experimental period and by examining the fatty acid profile of plasma triglyceride. Participants were encouraged to continue their usual pattern of physical activity and all other aspects of their lifestyle throughout the study.

Cholesterol concentration in plasma and lipoprotein fractions was measured enzymatically with Boehringer kits and calibrators, and triglyceride was measured enzymatically with Roche Diagnostic kits on a Cobas Fara analyser. Coefficient of variation was 1.6% for cholesterol and 3.4% for triglycerides in the Royal Australasian College of Pathologists' quality assurance programme. After precipitation of lipoproteins containing apolipoprotein B, high density lipoprotein cholesterol was measured in the supernatant with phosphotungstate-magnesium chloride solution.9 Low density lipoprotein cholesterol was calculated by using the Friedewald formula.10 Apolipoproteins A I and B were measured by immunoturbimetry with Boehringer kits (coefficient of variation 2.6% and 6% respectively). Lp(a) lipoprotein was determined by two site radioimmunoassay with a commercial kit provided by Pharmacia (Uppsala, Sweden; coefficient of variation 4%). Plasma lipids were extracted according to the procedure of Bligh and Dyer11 and triglycerides were separated by thin layer chromatography. Triglyceride fatty acids were analysed by gas liquid chromatography.12 The precision of the fatty acid measurement was determined by repeated analysis of a pooled plasma sample. The coefficient of variation for all fatty acids in the pooled sample ranged from 1% to 4%.

For all variables, measurements made during the last week of each period were not significantly different, so the mean of the two values was used to provide a more reliable measure of values before intervention (end of run in, end of washout) and after intervention. Differences between the changes in measurements before and after intervention on the two diets were compared by analysis of variance with repeated measures. As part of this analysis, the effect of the diet sequence on the relative treatment responses was tested as a diet by sequence interaction term. When no significant effects of sequence on the dietary responses were observed, the data for the butter and margarine periods were pooled over both dietary sequences. The difference between the changes are represented as means with 95% confidence intervals. Statistics for Lp(a) lipoprotein were computed with log transformed data.

Results

Body weight and blood pressure were unchanged throughout the study. Table 2 shows concentrations of plasma lipids, lipoproteins, apolipoproteins, and Lp(a) lipoprotein. Total and low density lipoprotein cholesterol and apolipoprotein B showed an appreciably greater decrease during the margarine diet than the butter diet. The sequence of dietary interventions had no significant effects on these differences between diets (P=0.416, P=0.525, and P=0.153, respectively). All other measurements were similar on the two diets.

The reported nutrient intakes based on the four day diet records are contrasted with the prescribed diets in table 3. Participants consumed similar amounts of foods from the major food groups during the two experimental periods. On the two diets an average 17 g of butter or margarine was used per day for cooking and spreading. Plasma triglyceride fatty acid composition on the two diets is shown for 20 randomly selected participants in table 4. As expected, proportions of myristic acid (C14:0) and palmitic acid (C16:0) were significantly lower and proportions of linoleic acid (C18:2) were significantly higher in plasma triglycerides of subjects during the margarine diet than the butter diet.

Discussion

We compared the effects of butter and of margarine consisting of predominantly unsaturated fatty acids on lipid and lipoprotein concentrations in a group of moderately hypercholesterolaemic subjects. This approach has practical application, having been carried out in an outpatient setting rather than a metabolic ward, and in addition permits conclusions concerning the effects of various fatty acids consumed in the context of a low fat diet.