Mortality in elderly patients with dementia markedly and progressively increases with extended use of antipsychotics, according to the first long-term controlled study of risk in this population. Earlier evidence of this risk was from short-term trials not exceeding 14 weeks.

Clive Ballard, MD, King's College, London, and colleagues1 recently reported on the dementia antipsychotic withdrawal trial (DART-AD), a 12-month, placebo-controlled study of 165 patients with Alzheimer disease, with a follow-up of 54 months. The mean age of patients was 85 years. Participants had to have received at least 3 months of antipsychotic medication for behavioral or psychiatric symptoms. Patients were randomized to continue antipsychotic medication treatment or to switch to placebo. The antipsychotics were those most commonly prescribed for this population in the United Kingdom—thioridazine, chlorpromazine, trifluoperazine, haloperidol, and risperidone.

The cumulative probability of survival for the patients who received antipsychotic medication for 1 year was 70%, compared with 77% in patients who received placebo. After 2 years, those who continued to receive antipsychotic medication had a 46% probability of survival, compared with 71% for those not being treated with antipsychotics. Mortality after 3 years of antipsychotic treatment was about double that without medication. "The results further highlight the need to seek less harmful alternatives for the long-term treatment of behavioral symptoms in Alzheimer's patients," Ballard commented in a press release from the Alzheimer's Research Trust in the United Kingdom, which funded the study.

An FDA advisory on using second-generation antipsychotics in this population was issued in 2005, following findings across 17 trials of a 1% to 2% increased mortality above an approximate 2.6% rate in patients not receiving these medications.2 An earlier FDA advisory had warned of increased risk of stroke with 3 second-generation antipsychotics based on a meta-analysis of 11 trials with risperidone or olanzapine and 3 with aripiprazole.3 Ballard and colleagues reported that atypical antipsychotics other than risperidone had not been included in their study and that treatment with the other agents could have had a different outcome. In addition, they indicated that the findings of this long-term study do not detract from the documented value of short-term antipsychotic treatment of severe neuropsychiatric manifestations of Alzheimer disease.

The DART-AD investigators warn that "the accumulating safety concerns, including the substantial increase in long-term mortality, emphasize the urgent need to put an end to unnecessary and prolonged prescribing."