Results from two small studies, involving a total of only 174 cases, have suggested that the increased risk of coronary heart disease conferred by cigarette smoking is substantially affected by genotype at the apolipoprotein E (APOE) epsilon2/epsilon3/epsilon4 polymorphism. We have established APOE genotypes in 4484 patients with acute myocardial infarction diagnosed before the age of 55 years for male and 65 years for female patients, and in 5757 controls with no history of cardiovascular disease. On average, the hazard ratio for myocardial infarction was 1.17 (95% CI 1.09-1.25; p<0.00001) per stepwise change from epsilon3/2 to epsilon3/3 to epsilon3/4 genotype. Among individuals in this study with known cigarette smoking status, the hazard ratio for myocardial infarction in smokers versus non-smokers was 4.6 (4.2-5.1). There was, however, no significant difference between the smoker/non-smoker hazard ratios for those with different APOE genotypes (chi2(2)=0.69; p=0.7). When differences in risk between different genotypes are not extreme (as with this APOE polymorphism), reliable assessment of hypothesised gene-environment interactions will often require the study of many thousands of disease cases.