We have investigated the interrelationships of plasma PAI-1 activity, the PAI-1 4G/5G polymorphism and risk of myocardial infarction (MI) in the ECTIM study, a case-control study of MI based in Belfast, Lille, Strasbourg and Toulouse. Mean PAI-1 levels in cases were similar across all centres but in controls, levels in the French centres were significantly higher. Only in Belfast were PAIl1 levels higher in cases (11.7 AU/ml) than controls (10.5 AU/ml). The PAI-1 4G allele frequency was similar in cases and controls (0.55 and 0.54). In all groups, 4G homozygotes had the highest mean plasma PAI-1 level (4G4G vs 5G5G; cases overall: 14.2 vs 12.1AU/ml; controls overall: 15.0 vs 12.6AU/ml), with the heterozygotes generally intermediate. The data from Belfast are consistent with the literature implicating PAI-1 level as an MI risk factor. In ECTIM, the PAI-1 4G/5G polymorphism is not a genetic risk factor for MI but is associated with PAI-1 activity. Thus homozygosity for the 4G allele may predispose to elevated PAI-1 and impaired fibrinolysis, perhaps requiring interaction with other genetic or environmental factors to influence MI risk.