a b s t r a c tThe role of plasma retinol and carotenoids in coronary heart disease (CHD) remains unclear. The PRIME Study prospectively evaluated these in France and Northern Ireland in 9758 men aged 50-59 years who were free of CHD at baseline. After five years' follow-up 150 incident cases of CHD (non-fatal myocardial infarction and fatal CHD) were compared with 285 controls matched for age, date of blood collection and study centre. Geometric means of major carotenoids did not differ significantly between cases and controls (P > 0.05), whereas the absolute and lipid-standardized plasma retinol levels were 9% lower in cases than controls in both countries (P < 0.002), without correlation with carotenoids. After adjusting for risk factors, the relative risks (RRs) of CHD in the first four quintiles of retinol distribution in controls (≤601, −683, −760, and −846 g/l) were 2.65 (P = 0.0009), 1.70, 1.03, and 1.12 (all P > 0.05) respectively, relative to the top quintile (retinol ≥846 g/l; linear trend P = 0.0001). The 10th percentile of lipid-standardized retinol (≤544 g/l) predicted an RR of 4.7 (P < 0.001). The risk associated with low retinol was comparable to strong risk factors (e.g. HDL-cholesterol, Interleukin-6) and behaved additively.In conclusion, plasma retinol levels of <601 g/l in a fifth of middle-aged European men place them at an approximately threefold RR of developing CHD. Thus the intake of vitamin A might be too low in middle-aged men. These findings must be confirmed.