Genotypes with the apolipoprotein epsilon4 allele are predictors of coronary heart disease mortality in a longitudinal study of elderly Finnish men.

Earlier we reported that allelic variation in the gene coding for apolipoprotein (apoE is a significant predictor of variation in the risk of coronary heart disease (CHD) death in a longitudinal study of elderly Finnish men. Here we address the question: which of the apoE genotypes confers the risk information in these men, and whether such information persists after other CHD risk factors are considered? We followed two cohorts of elderly Finnish men aged 65 to 84 years, one in Eastern (n = 281) and the other in the Southwestern (n = 344) Finland for 5 years during which 26 (9.3%) of the men from the Eastern cohort and 40 (11.6%) of the men in the Southwestern cohort died from CHD. Baseline high density lipoprotein (HDL) cholesterol and (HDL cholesterol)2 in the Eastern cohort and age, and total and HDL cholesterol and smoking status in the Southwestern cohort were significant predictors of CHD death (P < 0.05). The apoE genotypes were significant predictors in the Southwestern cohort at P = 0.02 and in the Eastern cohort at P = 0.18. In multivariable models, information about apoE genotypes improved the prediction at P = 0.1O level of statistical significance in both cohorts. When genotypes were considered separately, the episilon2/4 combined with the epsilon4/4 in the Eastern cohort (odds ratio = 7.69, 95% CI = 1.67-35.52) and the epsilon 3/4 in the Southwestern cohort (odds ratio = 2.44, 95% CI = 1.165.10) had sigificantly greater odds of CHD death compared to the common F3/3 genotype. We conclude that apoE genotypes confer risk information about CHD death in two cohorts of elderly Finnish men in a longitudinal study, and this information persists after adjustment for other CHD risk factors. Because different genotypes were predictors in these two cohorts, we further conclude that the utility of a particular genotype as a predictor of CHD death in other populations may depend on the distribution of risk factor profiles at baseline, geographically defined environmental exposures, the CHD mortality history, and the evolutionary history of background genotypes in the population considered.