Comparison of risk of cerebrovascular events in an elderly VA population with dementia between antipsychotic and nonantipsychotic users.

INTRODUCTION: The credibility of an increased risk of cerebrovascular events (CVEs) in elderly patients with dementia being treated with second-generation antipsychotics (SGAs) is debatable. Although early published and unpublished data indicated a risk, much of the subsequent literature has not supported this initial finding. Previously published studies were flawed in part because they lacked a control group and did not stratify by dementia subtype. This study examined the risk of a CVE in patients diagnosed with Alzheimer or vascular dementia while being treated with SGA, first-generation antipsychotics, or no antipsychotic medication.

METHODS: Data from 14,029 patients aged 65 years and older were evaluated using patient information from Veterans Administration and Medicare databases. Patients who received care for dementia were categorized according to dementia subtype (vascular or Alzheimer) and antipsychotic use during an 18-month period. Patients were observed until they were admitted to a hospital for a CVE, stopped taking or switched antipsychotics, died, or until the 18-month observation period ended.

RESULTS: Overall, CVE risk did not differ whether patients were receiving a first-generation antipsychotic, SGA, or no antipsychotic therapy. However, patients with vascular dementia had an increased risk in hospitalization for a CVE. There was no increase in risk of a CVE for patients treated with quetiapine, olanzapine, or risperidone relative to haloperidol, or for patients receiving olanzapine or risperidone relative to quetiapine. Stratified subgroup analyses demonstrated no difference in risk for CVE-related admission patients with Alzheimer dementia among individual agents. However, patients with vascular dementia receiving risperidone, but not olanzapine or quetiapine, were found to be at decreased risk for CVE admission relative to haloperidol.

CONCLUSIONS: This study found no increase in overall risk for CVE-related hospital admission in patients treated with antipsychotic medications.