Antipsychotic medications have a modest effect on the neuropsychiatric symptoms of dementia, but product labels warn of the excess risk of death and morbidity associated with their use in older patients. As such, these agents should not be the first choice for the treatment of behavioral and psychotic symptoms of dementia. Nevertheless, a trial of these agents may be indicated in instances in which the severity of symptoms is extreme, or symptoms do not respond to nonpharmacologic methods or other medications. Most clinical trials in which the efficacy of these agents was studied for the treatment of neuropsychiatric symptoms were of short duration, and thus may not be representative of true efficacy over the long term. There is no evidence to suggest differences in effectiveness between atypical and conventional antipsychotics; therefore, the choice of an antipsychotic for neuropsychiatric symptoms often relies on side effect profile and individual patient circumstances. Extrapyramidal symptoms and QTc prolongation are concerns with conventional antipsychotic agents. The incidence of cerebrovascular events with either atypical or conventional antipsychotics appears increased compared with placebo. A discussion of the risk-benefit ratio of antipsychotics with the patient's family and/or caregivers should precede the decision to use these agents.