Abstract
BACKGROUND: Atrial fibrillation is one of the most common arrhythmias, but the optimal drug choice for a rhythm-control strategy remains uncertain. METHODS: This article reports on a retrospective cohort claims database study conducted using the Truven Health Market Scan Commercial Claims and Encounters and Medicare Supplemental databases. Patients with a new diagnosis of atrial fibrillation, and a discharge date between 2011 and 2015, were included. The exposure variables of interest were a discharge prescription for amiodarone or dronedarone. The average treatment effect for the composite of total mortality or a repeat cardiovascular (CV)-related hospitalization was the primary outcome. Sensitivity analyses with other treatment effect metrics were performed. Baseline covariate imbalances between the groups were adjusted using propensity-score methods with inverse probability weighting. RESULTS: A total of 1735 patients were discharged on amiodarone, and 338 were discharged on dronedarone, with a median follow-up time of 357 days. A total of 43 (12.7%) CV-related hospitalizations occurred in the dronedarone group, and 146 (8.4%) occurred in the amiodarone group (risk difference 4.3%, 95% confidence interval [CI] 0.4%-8.3%, P = 0.02). A total of 4 (1.2%) deaths occurred in the dronedarone group, and 31 (1.8%) deaths occurred with amiodarone (risk difference -0.6%, 95% CI -2.1%-0.9%, P = 0.6). After adjusting for baseline covariates, the dronedarone hazard ratio for the composite endpoint was 1.47 (95% CI 1.01-2.12). This result was generally robust to sensitivity analyses. CONCLUSION: In this incident cohort of patients hospitalized for atrial fibrillation, compared to those discharged on amiodarone, patients who received a dronedarone discharge prescription had an increase in the composite endpoint of recurrent CV-related hospitalization and death, over a median 1-year follow-up period.