Decline in cognitive function in the elderly is common and represents a major clinical and public health concern. Aspirin may reduce the decline in cognitive function by influencing multi-infarct dementia, but data are sparse. The East Boston Senior Health Project is a population-based cohort study that enrolled 3,809 community-dwelling residents aged 65 years and older in 1982-1983 and followed them with home visits every 3 years until 1988-1989. Trained interviewers assessed cognitive function by using the Short Portable Mental Status Questionnaire and assessed medication use, including over-the-counter drugs. Response to the Short Portable Mental Status Questionnaire was scored as high, medium, or low, and decline was defined as transition to a lower category. Participants who used drugs containing aspirin in the 2 weeks prior to the interview were classified as aspirin users. Multiple logistic regression was used to obtain adjusted odds ratios and their 95% confidence intervals for decline of cognitive function. The estimating equation approach was used to adjust the standard errors for repeated measurements. Aspirin users had an odds ratio for cognitive decline of 0.97 (95% confidence interval 0.82-1.15). Low frequency of aspirin use (less than daily) was associated with an odds ratio of 0.87 (95% confidence interval 0.69-1.09). Although no substantial effect was observed, the data are also compatible with a modest benefit of aspirin, especially with intermittent use, on decline of cognitive function. Concern about small residual biases from self-selection or confounding suggests that randomized trials will be necessary to provide definitive data on this question.