We evaluated the association between the phenotypes or alleles of apolipoprotein E and hypertension in Japanese-Americans living in Hawaii or Los Angeles. The study consisted of 932 nondiabetic participants aged 40-79 years. Of these participants, 315 were hypertensive, while the remainder were normotensive. The prevalence of hypertension was higher in subjects with the e2 allele than in those without. Using a multivariable prediction model that included age, serum glucose, insulin, lipids, and body mass index, we found the epsilon2 allele to be associated with hypertension in men (odds ratio (OR), 1.53; 95% confidence interval (CI), 1.05-2.24) and for both sexes combined (OR, 1.35; 95% CI, 1.05-1.73). In a follow-up study, 37 of the 352 (11%) normotensive subjects had developed hypertension 4 years after the baseline examination. The frequencies of the apolipoprotein E phenotype E3/2 and the epsilon2 allele were higher in the subjects who became hypertensive than in those who remained normotensive. The E3/2 phenotype was associated with incidence of hypertension after adjustment for age, sex, baseline body mass index, and systolic blood pressure (OR, 1.75; 95% CI, 1.01-2.90). When serum lipids, glucose, and insulin were included in this analysis, the E3/2 phenotype tended to be associated with the incidence of hypertension. In conclusion, the epsilon2 allele of apolipoprotein E was associated with hypertension in Japanese-Americans. This association may be mediated via high levels of serum triglycerides that in turn are mediated in part by the epsilon2 allele. Nonlipid mechanisms may also be responsible for the effect of the epsilon2 allele on hypertension.