Recent studies suggest that the use of nonsteroidal anti-inflammatory drugs (NSAIDs) may reduce the risk for Alzheimer's disease (AD). We investigated the relation of NSAID use over a 10-year period and the risk for incident AD using a nested case-control design in the population-based Rotterdam Study. The study was performed in 306 subjects; 74 Alzheimer patients diagnosed according to NINCDS-ADRDRA criteria and 232 age and sex-matched controls. NSAID use was abstracted from general practitioners' medical records and expressed as cumulative prescription days. The relative risk for AD associated with long-term use (> or = 2 months) was 0.95 (95% CI: 0.46-1.99) as compared to nonusers, after controlling for possible confounders. In a separate examination, subjects who had more than 6 months of prescription days had a reduced relative risk for AD (RR = 0.74 (95% CI: 0.20-2.72). In an age-stratified analysis the effect in long-term users was evident in those aged 85 and under; 0.53 (95% CI: 0.15-1.77). All risk estimates were lower when the last 2 years of exposure were excluded from the analyses. Our point estimates in subjects younger than 85 years and in subjects using NSAIDs for 6 months or more are consistent with the hypothesis that long-term use of NSAIDs reduces the risk for AD. However, overall there was no association between NSAID use and the risk for incident AD.