While a complete understanding of the pathogenesis of Alzheimer's disease (AD) remains elusive, many conclusions can be drawn from the numerous epidemiological studies undertaken to date. Prevalence and incidence estimates show consistency, following a roughly exponential pattern with a doubling of both parameters roughly every five years after age 65. Roughly 7% of the population aged 65 and over has AD. The clinical course of the disease is reasonably well established and mortality rates rise with increasing levels of cognitive deficit. Four risk factors for AD are firmly established: increasing age, the presence of the apolipoproteinE-epsilon4 allele, familial aggregation of cases, and Down's syndrome. Numerous other associations have been shown in some studies, but not in others. For example, women generally appear at higher risk than men, as do people with lower levels of education; depression is probably prodromal; head injury is an established risk factor, and may interact with the apoE gene; several occupational exposures appear hazardous, and exposure to aluminum in the water supply confers excess risk. Hypertension and other vascular symptoms appear to predispose to AD, which is now seen as nosologically closer to vascular dementia than was previously believed. Several apparently protective factors have been identified, although preventive trials based on these have so far shown minimal effectiveness. The use of non-steroidal anti-inflammatory drugs to treat arthritis is associated with a reduced risk of AD, as is estrogen use by post-menopausal women. Physical activity appears beneficial, as does a diet with high levels of vitamins B6, B12 and folate. while red wine in moderate quantities appears protective. This review concludes with a discussion of the strengths and limitations of current epidemiological methods for studying Alzheimer's disease.