To compare the risk of major medical events in nursing home residents newly initiated on conventional or atypical antipsychotic medications (APMs).DESIGN:
Cohort study, using linked Medicaid, Medicare, Minimum Data Set, and Online Survey Certification and Reporting data. Propensity score-adjusted proportional hazards models were used to compare risks for medical events at a class and individual drug level.SETTING:
Nursing homes in 45 U.S. states.PARTICIPANTS:
Eighty-three thousand nine hundred fifty-nine Medicaid-eligible residents aged 65 and older who initiated APM treatment after nursing home admission in 2001 to 2005.MEASUREMENTS:
Hospitalization for myocardial infarction, cerebrovascular events, serious bacterial infections, and hip fracture within 180 days of treatment initiation.RESULTS:
Risks of bacterial infections (hazard ratio (HR) = 1.25, 95% confidence interval (CI) = 1.05–1.49) and possibly myocardial infarction (HR = 1.23, 95% CI = 0.81–1.86) and hip fracture (HR = 1.29, 95% CI = 0.95–1.76) were higher, and risks of cerebrovascular events (HR = 0.82, 95% CI = 0.65–1.02) were lower in participants initiating conventional APMs than in those initiating atypical APMs. Little variation existed between individual atypical APMs, except for a somewhat lower risk of cerebrovascular events with olanzapine (HR = 0.91, 95% CI = 0.81–1.02) and quetiapine (HR = 0.89, 95% CI = 0.79–1.02) and a lower risk of bacterial infections (HR = 0.83, 95% CI = 0.73–0.94) and possibly a higher risk of hip fracture (HR = 1.17, 95% CI = 0.96–1.43) with quetiapine than with risperidone. Dose-response relationships were observed for all events (HR = 1.12, 95% CI = 1.05–1.19 for high vs low dose for all events combined).CONCLUSION:
These associations underscore the importance of carefully selecting the specific APM and dose and monitoring their safety, especially in nursing home residents who have an array of medical illnesses and are undergoing complex medication regimens.