Abstract
BACKGROUND: Trazodone has emerged as a potentially safer alternative to atypical antipsychotic medications for managing delirium-related symptoms in older adults after hospital admission. This study aimed to compare safety outcomes in older adults who were prescribed trazodone versus atypical antipsychotic medication after hospital discharge. METHODS: This population-based cohort study was done using US Medicare data from Jan 1, 2013, to Dec 31, 2022, and Optum de-identified Clinformatics Data Mart Database data from Jan 1, 2013, to Aug 31, 2025, with a target trial emulation framework. Patients aged at least 65 years without psychiatric disorders who were prescribed a new trazodone or atypical antipsychotic medication (quetiapine, risperidone, or olanzapine) prescription within 30 days of hospital discharge with a claims-based delirium diagnosis were included. We applied propensity score overlap weighting to adjust for 162 covariates. The primary outcome was all-cause rehospitalisation. Secondary outcomes included rehospitalisation for delirium, falls (including fall-related emergency department visits), pneumonia, urinary tract infections, stroke, and all-cause mortality. Treatment-outcome associations were estimated using the Fine-Gray model for non-fatal outcomes (presented as subdistribution hazard ratio [SDHR]) and the Cox model for all-cause mortality (presented as hazard ratio [HR]). Estimates from the two databases were pooled using inverse-variance weighting. FINDINGS: 11 678 trazodone and 29 590 atypical antipsychotic medication users were included. Compared with antipsychotic medications, trazodone was associated with lower risks of rehospitalisation (SDHR 0·95 [95% CI 0·91-0·98]), rehospitalisation for delirium (SDHR 0·80 [0·75-0·86]), urinary tract infection (SDHR 0·84 [0·72-0·99]), and all-cause mortality (HR 0·84 [0·79-0·90]); no significant difference was observed for falls (SDHR 0·93 [0·85-1·02]), pneumonia (SDHR 0·96 [0·79-1·16]), and stroke (SDHR 0·92 [0·77-1·11]). The reduced risk of rehospitalisation with trazodone was consistent when compared with risperidone (SDHR 0·89 [0·84-0·95]), and olanzapine (SDHR 0·88 [0·82-0·94]) but no difference was seen with quetiapine (SDHR 0·97 [0·93-1·01]). INTERPRETATION: These results revealed reduced risks of rehospitalisation and all-cause mortality associated with trazodone versus atypical antipsychotic medications. Trazodone is a potentially safer alternative to antipsychotic medications in the management of delirium-related symptoms after hospital discharge. FUNDING: National Institute on Aging.