Objective: To investigate in a longitudinal cohort of people with Alzheimer's disease whether taking antipsychotics is associated with more rapid cognitive deterioration. Method: From a sample of 224 people with Alzheimer's disease recruited as epidemiologically representative, those taking antipsychotic drugs for more than 6 months were compared with those who were not, in terms of change in three measures of cognition. The effects of potential mediators and confounders (demographic factors, neuropsychiatric symptoms, cognitive severity and cholinesterase inhibitors) were also examined. Results: No significant difference was observed in cognitive decline between those taking antipsychotics (atypical or any) and others on any measure of cognition. The only predictor of more cognitive decline was greater baseline cognitive severity (B = 3.3, 95% confidence interval 0.6 to 6