Ways to reduce saturated fat intake.

Current dietary guidelines emphasize a reduced intake of total fat. This means that (saturated) fatty acids should be replaced by carbohydrates. But are carbohydrates the optimal replacement? High-carbohydrate diets do indeed lower plasma LDL-cholesterol levels. However, both sugars and complex carbohydrates lower HDL-cholesterol and increase fasting triacylglycerol levels (Mensink & Katan, 1987, 1992) and these effects last as long as the low-fat-high carbohydrate diet is eaten (Ernst et al. 1980). Low HDL-cholesterol levels are associated with increased risk of CHD (Gordon et al. 1989; Kinosian et al. 1994), and many factors that reduce HDL-cholesterol levels, such as smoking, obesity and lack of physical exercise increase the incidence of CHD. Although we are not sure that diet-induced decreases in HDL increase CHD risk, diets that lower HDL-cholesterol levels should be viewed with caution (Katan et al. 1997). An alternative to high-carbohydrate diets is to replace saturated fat with cis-unsaturated oils (oils rich in the n-9 cis-monounsaturated oleic or the n-6 cis-polyunsaturated linoleic acid, and some n-3 fatty acids) and leave the intake of total fat the same. Replacement of saturated fatty acids with cis-unsaturated fatty acids reduces LDL-cholesterol, with little decrease in HDL. Cis-unsaturated fatty acids considerably lower the total : HDL-cholesterol ratio (the strongest predictor of CHD in observational studies; Kinosian et al. 1994), while carbohydrates do not improve this ratio. These effects have been well established by controlled dietary trials (Mensink & Katan, 1992) and they are confirmed by the study of Williams et al. (1999) in the present issue of the British Journal of Nutrition. Low-fat, high-carbohydrate diets are thought to provide benefits that offset their unfavourable effects on HDLcholesterol and triacylglycerol levels. One of these benefits is weight loss. Indeed, populations that habitually consume low-fat diets are often quite lean (Seidell, 1998). However, controlled trials show that in affluent populations restriction of energy from fat and replacement by carbohydrates results in only modest reductions in body weight (Willett, 1998). These reductions are too small to compensate for the HDLlowering effect of high-carbohydrate diets (Leenen et al. 1993). Also, the prevalence of obesity in the USA has increased considerably over the past 25 years, despite a concomitant decline in the intake of energy from fat. Thus, restricting fat intake does not always result in weight reduction. This may hold particularly for Western societies, where the typical low-fat food items available in the supermarket are often high in energy. We may be fooling ourselves if we think that foods high in carbohydrates, protein and fat replacers will stop the oncoming worldwide epidemic of obesity. Another proposed benefit of high-carbohydrate diets is a reduction of cancer risk. However, the evidence that a high fat intake causes cancer is weak (World Cancer Research Fund/American Institute of Cancer Research, 1997; Department of Health, 1998). A high intake of n-3 fatty acids from fish may, if anything, offer some protection (de Deckere et al. 1998), and intake of oleic acid and of linoleic acid is not associated with cancer risk in human subjects (Gerber, 1997; Zock & Katan, 1998). Thus, cisunsaturated fatty acids seem a suitable and safe replacement for saturated fatty acids to reduce CHD risk, in particular for people who are close to ideal body weight. Overweight people should reduce energy intake by lowering consumption of saturated fat and not replacing this with other nutrients; they should also eat less sugar and other refined carbohydrates. Do different cis-unsaturated fatty acids have different effects on risk? The n-6 cis-polyunsaturated fatty acid linoleic acid may be somewhat more effective in improving the plasma lipid profile (Mensink & Katan, 1992; Clarke et al. 1997) than n-9 cis-monounsaturated oleic acid, but the impact of this difference on CHD risk is probably small. The n-3 cisunsaturated fatty acids from fish (mainly eicosapentaenoic and docosahexaenoic acids) and plant oils (a-linolenic acid) may give additional protection, in particular against fatal CHD (Albert et al. 1998; Hu et al. 1999). These observations need to be tested in future studies but they do suggest that a diet high in a variety of n-3, n-6 and n-9 cis-unsaturated fatty acids might provide the optimum for reduction of CHD risk.