To investigate whether mortality risk is influenced by apolipoprotein E (AP
OE) genotype and whether the risk differs by ethnicity, we compared the mor
tality risk in 2,112 individuals greater than or equal to 65 years of age r
esiding in northern Manhattan in New York. Mortality risks associated with
the APOE genotype, adjusted for sex, high-density lipoprotein (HDL), low-de
nsity lipoprotein (LDL), and triglycerides, differed significantly by ethni
c group. Among Caucasian and Hispanics, the E2/E3 genotype was associated w
ith the lowest mortality risk in the multivariate Cox proportional hazards
modeling, adjusted for lipid levels, whereas mortality risk did not differ
substantially between the E4/E3 and E3/E3 genotypes. Among African-American
s, the E2/E3 genotype was not associated with the lowest mortality risk, bu
t the E4/E3 genotype was. Adjustment for heart disease, diabetes, and strok
e reduced mortality risk associated with each genotype by about 50% for all
ethnic groups, but the patterns remained the same. Although we cannot rule
out the possibility of a healthy survival bias, our analyses designed to e
xamine healthy survival by comparing risk of mortality in groups who were y
ounger or older at entry do not support this possibility. Our findings sugg
est that the APOE genotype is associated with mortality and that the genoty
pic risks differ by ethnic group. Nearly 50% of the mortality risk associat
ed with the APOE genotype appears to act through major chronic diseases, bu
t those diseases only partially explain the mechanism by which the genotypi
c risk acts. To better understand the observed ethnic differences in mortal
ity risk by genotype, a detailed prospective study is needed to examine the
relationships among APOE, other candidate genes, health conditions, and ev
entual death. (C) 2001 Wiley-Liss, Inc.