In a cross-sectional population study of 1132 unselected Eastern Finnish men aged 54 years, serum selenium concentration had a weak positive association with plasma HDL cholesterol (standardised partial regression coefficient, beta = 0.061, P = 0.019) and a fairly strong inverse relationship (beta = -0.223, P less than 0.001) with the extent of ADP-induced platelet aggregation. Neither plasma ascorbate concentration nor alpha-tocopherol to total cholesterol ratio had any association with plasma lipoproteins, platelet aggregability or prevalent ischaemic heart disease (IHD). When a covariance-correction was applied, men with ischaemic ECG findings at exercise had a lower mean serum selenium than others (81.5 micrograms/l vs. 85.9 micrograms/l, P less than 0.01 for difference). This difference was equally large for men with neither symptoms nor previous diagnosis of IHD.