The authors examined whether the presence of extrapyramidal side effects (EPS), measured before neuroleptic treatment was initiated, could be used to predict the development and severity of neuroleptic-induced parkinsonism (NIP) in Alzheimer's disease (AD). Twenty-four newly medicated probable AD patients were assessed with a battery of measures of extrapyramidal motor function. Assessments were made before neuroleptic therapy, and 3 and 9 months after treatment. Posttreatment clinical findings revealed that 66.7% of the AD patients developed NIP. Patients who developed NIP exhibited more severe pretreatment bradykinesia on instrument-derived measures. These findings suggest that a substantial proportion of AD patients treated with neuroleptics develop significant EPS and that the risk for EPS can be estimated before intervention by use of instruments measuring motor function.