Anti-inflammatory drugs have been suggested as a possible treatment for Alzheimer's disease (AD). The association of immune proteins and immune-competent microglial cells with senile plaques (SP) in both AD and normal aging suggests that these drugs may be able to modify the course of AD, either by interfering with SP formation or by suppressing the inflammation associated with SP. We compared postmortem brain tissue from elderly, nondemented, arthritic patients with a history of chronic nonsteroidal anti-inflammatory drug (NSAID) use (n = 32, aged 77 +/- 7 years) and nondemented control subjects with no history of arthritis or other condition that might promote the regular use of NSAIDs (n = 34, aged 77 +/- 6 years). In both the NSAID-treated group and control subjects, 59% of patients had some SP. There was no difference between the two groups in the mean number of plaques or in the number of specific SP subtypes (diffuse or neuritic). The degree of neurofibrillary pathology was also similar. Activated microglia were identified using CR3/43, an anti-MHC class II antibody. Both patient age and the presence of SP correlated positively with the number of CR3/43+ microglia (p < 0.02), whereas NSAID use was associated with less microglial activation (p < 0.01). Control patients with SP had almost three times the number of activated microglia as NSAID-treated patients with SP (11 versus 4 cells/mm2, p < 0.02). These results suggest that if NSAID use is effective in treating AD, the mechanism is more likely to be through the suppression of microglial activity than by inhibiting the formation of SP or neurofibrillary tangles.