OBJECTIVE: Saturated fats have adverse effects on health. To investigate which is more bene®cial for energy replacement, we compare the effects of polyunsaturated fatty acid and carbohydrate intake on obesity and metabolic variables (fasting triglycerides, HDL-cholesterol, LDL-cholesterol and 2 h glucose). Further, because the optimum diet may differ according to glucose tolerance, we examine the same associations in glucose tolerant and intolerant groups. Finally, we test the effect of macronutrient intake on the presence or absence of the metabolic syndrome. DESIGN: Cross-sectional analysis. SUBJECTS: A total of 4497 men and 1865 women aged 39 ± 62 in the Whitehall II study. RESULTS: In men, higher intakes of both polyunsaturated fats and carbohydrates were linked to lower waist ± hip ratio, triglycerides and LDL-cholesterol. Higher carbohydrate intake alone was linked to decreased body mass index (for 10 g higher carbohydrate intake, À0.12 kgam 2 , P`0.0001) and lower HDL-cholesterol (À0.01 mmolal, P`0.01). In normoglycaemic men, higher carbohydrate intakes were associated with higher 2 h insulin and glucose levels (0.25 pmolal, P`0.05 and 0.01 mmolal, P 0.001, respectively). Dietary effects among women were similar, the exception being a positive association of polyunsaturated fat intake with body mass index and waist ± hip ratio (0.47 kgam 2 , P`0.05 and 0.006, P`0.05, respectively). Dietary components, with the exceptions of cholesterol and protein in men, were unrelated to prevalence of the metabolic syndrome, and adjustment for differences in macronutrient intake did not account for the strong inverse association between socioeconomic position and the metabolic syndrome. CONCLUSION: Our observational data provide evidence that both polyunsaturated fatty acids and carbohydrates offer small metabolic bene®ts with few adverse effects compared with saturated fats.