Forty-four epidemiologic studies on tardive dyskinesia were evaluated as to whether they provided information on diagnostic criteria, objective scale and assessment, interobserver reliability, period of observation, and specific interhospital coordination. Studies which met these standards were reviewed for data on class of neuroleptic therapy, dose, duration, continuity of treatment, extrapyramidal toxicity, spontaneous dyskinesias, other drugs and treatment modalities, age and sex. A higher prevalence of tardive dyskinesia has been consistently noted in the elderly and in females. No other predisposing factors for tardive dyskinesia have been conclusively demonstrated thus far. Prevalence of tardive dyskinesia is estimated at 24-56% in chronic neuroleptic users.