Dietary determinants of plasma lipids and lipoproteins: the Caerphilly Study.

Associations between nutrient intakes and plasma lipids were examined in a representative sample of 653 middle-aged men in Caerphilly, South Wales. Nutrient intakes were calculated from 7-d weighed intake records and were expressed both as g/d and as a percentage of total energy intake. In multivariate analyses, after taking the effects of potential confounding factors (age, body mass index and smoking habit) into account, the percentage of variance in the plasma lipid concentrations which could be explained solely by the dietary variables was very small, ranging from 1 to 7 per cent. Associations which were statistically significant included a positive association between saturated fat intake and both total cholesterol and low density lipoprotein cholesterol; a negative association between carbohydrate intake and both total cholesterol and high density lipoprotein cholesterol (HDL-C); a positive association between magnesium intake and an HDL-C subfraction (HDL2-C); a positive association between alcohol intake and each of the following: HDL-C, its subfractions and triglyceride.