'Typical' antipsychotic agents can lead to a variety of extrapyramidal symptoms (EPS), including parkinsonism. The efficacy of a number of atypical antipsychotics in reducing psychosis without a detrimental effect on motor function has been studied in the group of patients most vulnerable to EPS, those who already have parkinsonian symptoms. Multiple open-label studies with clozapine strongly suggested that at low doses the drug was an effective antipsychotic and did not impair motor function. This was confirmed by two double-blind, placebo-controlled studies. A disadvantage of clozapine is that it can cause agranulocytosis and therefore patients require ongoing hematological monitoring. Studies with both risperidone and olanzapine have produced conflicting results, with some patients showing an overall improvement and others exhibiting severe deterioration of motor function. As with clozapine, multiple open-label studies with quetiapine have consistently demonstrated that it improves psychosis without impairing motor function. Double-blind trials are yet to be performed: however, the existing data, coupled with the lack of required blood monitoring, have led some experts to recommend quetiapine as the drug of choice for treatment of drug-induced psychosis in patients with parkinsonism. The atypical antipsychotics have also been tested in the largest group of EPS-vulnerable patients, the demented elderly. Results from a number of trials are described here. These data are more difficult to interpret as the number of variables is far greater than for the population with parkinsonism. However, the evidence to date indicates a generally low incidence of tardive dyskinesia with atypical antipsychotics. (C) 2002 Elsevier Science Ltd. All rights reserved.