Huang_2026_Pharmacoepidemiol.Drug.Saf_35_e70426

PURPOSE: Limited evidence suggests donepezil might pose an increased risk of malignant arrhythmias. We examined whether donepezil initiation was associated with a higher risk of hospitalization for malignant arrhythmia compared to other acetylcholinesterase inhibitors. METHODS: Population-based cohort study using linked administrative databases from seven Canadian provinces (BC, AB, SK, MB, ON, PEI, NL). We included individuals aged <= 66 who were new users of donepezil, galantamine, or rivastigmine from April 1st, 2011, to January 31st, 2019. We used multivariable Cox proportional hazards regression to estimate the risk of hospitalization for malignant arrhythmia among new donepezil users compared to new users of galantamine or rivastigmine. RESULTS: We identified 162 616 subjects (mean age 82, 40% male). Most (127 103; 78%) were treated with donepezil, while 25 597 (15.7%) received galantamine, (n = 6541; 4.0%) received oral rivastigmine, and (n = 3375; 2.1%) received transdermal rivastigmine. The median age was 82 years, and 59.8% were women. Baseline characteristics were generally well balanced. During a median follow-up of 386 days (maximum 2616 days), we identified 90 hospitalizations for malignant arrhythmias, including 58 in the donepezil group (23 per 100 000 person-years) and 32 for other AChEIs (44 per 100 000 person-years). Cumulative incidence curves showed a lower risk among donepezil initiators (Log-Rank p = 0.0026; Wilcoxon p = 0.0008). Crude regression suggested a reduced hazard (HR 0.52; 95% CI 0.34-0.80), which persisted after adjustment (aHR 0.55; 95% CI 0.36-0.85). In drug-specific analyses, donepezil initiators had a lower hazard relative to galantamine (aHR 0.49; 95% CI 0.31-0.77) but not rivastigmine (aHR 0.90; 95% CI 0.36-2.27). Sensitivity analyses-including alternative diagnosis field definitions, restricted follow-up, modified exclusion criteria, and propensity score adjustment-yielded consistent findings. No clear dose-response pattern was observed for any AChEI. CONCLUSION: In this population-based cohort study, initiation of donepezil was not associated with an increased risk of malignant arrhythmias compared to initiation of galantamine or rivastigmine.