Major Finding: After 2003, rates of conventional antipsychotic use among outpatients with dementia were less than 2%.
Data Source: National data from the Department of Veterans Affairs on 254,564 outpatients with dementia.
Disclosures: Neither Dr. Kales nor Dr. Zivin disclosed any conflicts.
SAVANNAH, GA. — Black box warnings on antipsychotics led to a decrease in their use to treat patients with dementia, without a commensurate increase in use of antianxiety or other psychotropic agents but with a rise in the use of anticonvulsant therapies, some of which appear to be riskier than the antipsychotics in older patients.
In fact, valproate appears to have a higher mortality than the antipsychotics studied, Dr. Helen C. Kales and Kara Zivin, Ph.D., of the University of Michigan and the VA Healthcare System in Ann Arbor, said in reporting the preliminary findings from an ongoing National Institutes of Health–funded study at the annual meeting of the American Association for Geriatric Psychiatry.
Antipsychotics once were widely prescribed as an off-label treatment for dementia, but in 2005, the Food and Drug Administration issued a black box warning that the use of atypical antipsychotics in the treatment of behavioral disorders in elderly patients with dementia was associated with increased mortality. A similar warning for conventional antipsychotics was issued in 2008.
To see how the warnings affected practice patterns, Dr. Kales and Dr. Zivin looked at national data from 254,564 Department of Veterans Affairs outpatients with dementia.
They found that a decline in use of atypical antipsychotics began in 2003, coinciding with the release of data from randomized, controlled trials about cerebrovascular events. The decline accelerated after the black box warning.
For conventional antipsychotics, the major decline in use came during the 1990s with the introduction of atypical antipsychotics. After 2003, the rates of conventional antipsychotic use among outpatients with dementia were less than 2%.
No significant compensatory increase was found in the use of antianxiety or antidepressant agents, Dr. Zivin said. In an interview, Dr. Kales speculated that possibly those drugs are not viewed as substitute therapies for the types of behaviors for which antipsychotics are prescribed.
The research did turn up a small but statistically significant increase in anticonvulsant use among outpatients with dementia.
The researchers also examined mortality for 13,857 elderly patients with any diagnosis and 3,954 patients with dementia who had been prescribed individual antipsychotics (haloperidol, olanzapine, quetiapine, and risperidone) and individual anticonvulsants (valproate and carbamazepine).
The researchers examined records for elderly patients in general as well as for those with dementia. As in prior studies, they found that haloperidol was associated with higher mortality than were the atypicals.
What was unexpected was the finding that valproate was associated with higher mortality than all of the antipsychotics, including haloperidol.
The populations of older patients receiving haloperidol and valproate differed somewhat from those receiving atypicals. The haloperidol cohort included more African Americans, and more subjects with schizophrenia and medical comorbidities. The valproate cohort was younger and a lower proportion had dementia. Neither accounting for selection bias nor excluding for bipolar disorders changed the results, Dr. Kales reported.
More research on the association between mortality and antipsychotics is needed, Dr. Kales said in an interview. “Is it a direct medication effect or related to the pathophysiology underlying the need for use?” She plans to use an electronic medical record search engine to look for other variables in such studies, including the severity of cognitive impairment and the type of behavioral disturbance.
Disclosures: The study was funded by the National Institutes of Health. Neither Dr. Kales nor Dr. Zivin disclosed any relevant conflicts of interest.