The recently shown association between apolipoprotein E (APOE) genotype and depressive illness has been challenged by subsequent studies. However, controversial results may derive from the different diagnostic criteria used for depression and from the small numbers of depressed patients included in the studies. We examined the association between depression and the genetic polymorphism of APOE in a large sample of depressed patients, Alzheimer's disease (AD) patients, and healthy controls following clear definitions for late-life depression. The cumulative incidence of depression depending on the age at onset of the first episode was examined by survival analysis. Our data do not disconfirm the hypothesis of depression sharing some common pathophysiologic features with AD, however, it seems very unlikely that the APOE genotype will elucidate the assumed common mechanisms.