The aim of this study was to determine whether plasma and red cell fatty acid levels were associated with cardiovascular disease, and whether any association was independent of other major risk factors. Over 1100 subjects were examined in a random sample survey of the general population (the Edinburgh Artery Study). Fatty acids were measured in three plasma fractions (triglyceride, cholesteryl ester and phospholipid) and in red cell phospholipids. Fatty acid levels in groups with cardiovascular disease (myocardial infarction (MI), angina and lower limb disease) were compared with a no disease group. In the cholesteryl ester and phospholipid fractions there were significantly lower levels of eicosapentaenoic acid in the MI group on univariate analysis (p<0.05), but not when adjusted for age, sex, smoking and systolic blood pressure using logistic regression. In the red cell fraction, alpha-linolenic acid was significantly lower in those with stroke (p<0.01) and lower limb disease (p<0.05). Linoleic acid was significantly raised in the triglyceride fraction in those with MI, probably reflecting recent dietary changes. There were significant increases in dihomo-gamma-linolenic acid in the phospholipid and red cell fractions in those with MI, and in the phospholipid fraction in the stroke group. These results do not support the hypothesis that n-6 fatty acids are protective against cardiovascular disease, although there may be some beneficial effects of the n-3 fatty acid, alpha-linolenic acid. Results from cross-sectional surveys must, however, be interpreted with caution because the presence of disease may affect dietary intake.