Based on reports that the use of nonsteroidal anti-inflammatory drugs (NSAIDs) may reduce the risk for Alzheimer's disease (AD), we studied the cross-sectional relation between NSAID use and the risk for AD in a population-based study of disease and disability in older people. After controlling for age, education, gender, and use of benzodiazepines, we found a relative risk (RR) for AD of 0.38 (0.15 to 0.95) when comparing NSAID users (n = 365) to NSAID non-users (n = 5,893). To address confounding by indication or contraindication, we compared NSAID users with a subset of NSAID non-users who were using topical medication for ear, eye, or dermatologic conditions (n = 365). In this comparison, the adjusted RR for AD was 0.54 (0.16 to 1.78). These findings are compatible with a possible protective effect of NSAIDs on the risk for AD.