In Western populations, the level of serum cholesterol increases with age, whereas in certain developing countries this increase is absent. In order to determine which factors are related to this increase, the authors investigated 99 men and 70 women whose serum cholesterol, habitual food intake, and body mass index were measured between 1974 and 1979 when they were students (baseline values) and again in 1985 (follow-up measurement). Serum cholesterol had increased by 0.59 mmol/l (14%) in men, and by 0.34 mmol/l (7%) in women. The effect of changes in habitual food intake on serum cholesterol was quantified as the Keys score. The mean increase of the Keys score corresponded to a change in cholesterol of 0.12 mmol/l for both men and women. Body mass index had increased by 0.9 kg/m2 in men and by 0.5 kg/m2 in women. Regression analysis showed that in men change in body mass index partly explained the change in serum cholesterol (r = 0.20). In women none of the independent variables could explain changes in serum cholesterol. In order to determine what the effect of ageing was independent of changes in body mass index and dietary changes, 34 volunteers who had participated in the follow-up measurements and whose change in body mass index had been less than 2 kg/m2, were prescribed a diet for three weeks. This diet had the same composition as their habitual diet at baseline measurement. It lowered serum cholesterol levels by 0.1 mmol/l over the three weeks of the trial, and thus could not abolish the rise of serum cholesterol with age.(ABSTRACT TRUNCATED AT 250 WORDS)