Benzodiazepine or Antipsychotic Use and Mortality Risk Among Patients With Dementia in Hospice Care.

Abstract

Benzodiazepines and antipsychotics are frequently prescribed for symptom management in patients with Alzheimer disease and related dementias (ADRD) receiving hospice care, yet evidence guiding their safety and effectiveness is limited. To evaluate the association between incident benzodiazepine or antipsychotic use and 180-day mortality among nursing home residents with ADRD enrolled in hospice. This retrospective case-control study used national Medicare claims and Minimum Data Set assessments from July 1, 2014, to September 30, 2018, from US nursing homes providing long-term care. Residents with ADRD newly enrolled in hospice who had no benzodiazepine or antipsychotic use in the prior 6 months were included. Incident users were matched 1:1 to nonusers on enrollment timing, age, sex, comorbidity, cognitive function, and baseline central nervous system medication use. Follow-up was 180 days from hospice enrollment. Analyses were performed from March 2023 to December 2024. Initiation of benzodiazepine or antipsychotic use during hospice, with no fills of the respective medication class in the prior 6 months. The primary outcome was all-cause mortality within 180 days of hospice enrollment. Hazard ratios (HRs) were estimated using Cox proportional hazards regression models adjusting for demographic, clinical, and hospice-level characteristics. Sensitivity analyses included propensity score-weighted models and models accounting for cumulative exposure. Among 139 103 participants (105 372 [75.8%] female; mean [SD] age, 87.6 [7.7] years), 26 872 matched pairs were included in the benzodiazepine cohort and 10 240 in the antipsychotic cohort. For the 53 859 participants in matched cohorts, the mean (SD) age was 89.0 (6.4) years, and 45 116 (83.8%) were female. Initiation of benzodiazepine use was associated with higher 180-day mortality compared with nonuse (HR, 1.41; 95% CI, 1.38-1.44), as was initiation of antipsychotic use (HR, 1.16; 95% CI, 1.12-1.20). These findings were consistent across propensity score-weighted models and with greater cumulative exposure. In this national case-control study of nursing home residents with ADRD receiving hospice care, initiation of benzodiazepine or antipsychotic use was associated with increased 180-day mortality. While these medications may provide symptom relief in appropriate clinical scenarios, their use is associated with substantial risks. These findings highlight the need for careful prescribing decisions and the development of dementia-specific hospice prescribing guidelines.