Pattern and Associated Factors of Benzodiazepine Discontinuation Among Older Adults Following Hospitalization

Prolonged benzodiazepine (BZD) use following hospital discharge may heighten the risk of BZD dependence and serious adverse drug events in older adults. However, there is limited understanding of BZD discontinuation pattern and associated factors in this population.

This retrospective cohort study was conducted using Optum CDM from January 2004 to February 2025. We included patients aged ≥ 65 years who newly initiated a BZD prescription within 30 days of hospitalization. We excluded patients with anxiety, mental illnesses, alcohol abuse, seizure disorder, and those receiving hospice care. BZD discontinuation was defined as a gap exceeding 15 days following the end of a BZD prescription. Discontinuation rates were estimated by the Kaplan-Meier method, and the association between patient characteristics and BZD discontinuation was analyzed using Cox proportional hazards models in the primary analysis. Inverse probability of censoring weight was applied to adjust for the competing risk of death.

A total of 33,449 patients were included (mean age [SD]: 73.1 [5.8] years; 51.7% men). The IPW-weighted BZD discontinuation rate by 30, 60, and 90 days was 53.3% (95% CI: 52.7%-53.8%), 86.7% (86.3%-87.1%), and 92.6% (92.3%-93.0%), respectively. The 30-day discontinuation rate increased from 31.7% (29.5%-33.9%) in 2004 to 71.1% (68.7%-73.5%) in 2024. During the study period, an annual 4% increase in the discontinuation rate was observed. Risk factors associated with lower BZD discontinuation were insomnia (hazard ratio [HR]: 0.66 [0.63-0.69]), moderate-to-severe frailty (HR: 0.82 [0.75-0.90]), and new uses of antidepressants (HR: 0.80 [0.76-0.85]) and atypical antipsychotics (HR: 0.90 [0.82-0.99]) following hospitalization and before the first BZD prescription.

Discontinuation rates of BZDs following hospitalization in older adults without a psychiatry history have increased over time. However, challenges remain in discontinuing BZDs, particularly among individuals with insomnia and moderate-to-severe frailty, which may be target populations for developing effective deprescribing strategies in the future.