OBJECTIVES: This study compared treatment patterns among NVAF patients initiating apixaban or warfarin by race/ethnicity [White, African American (AA), or Hispanic].

METHODS: NVAF patients aged ≥18 years who initiated apixaban or warfarin following their first NVAF diagnosis were identified using the Komodo Healthcare Map data between 01Jan2019 to 31Dec2022. Patients were enrolled in a US commercial health plan. The first treatment initiation date was designated as the index date. Patients had ≥6 months of pre-index and post-index continuous enrollment. Follow-up was defined as index date until earliest date of disenrollment, discontinuation of index treatment, death, or study end. In the follow-up period, treatment patterns (discontinuation, switch) were compared in the overall population and by race/ethnicity. Inverse-probability treatment weighting (IPTW) balanced the treatment groups on relevant demographic and clinical characteristics. Incidence rates (IR) for outcomes were calculated per 100 person-years. Cox proportional hazards models were used to evaluate the adjusted risk of outcomes reported as hazard ratios (HR) with 95% confidence interval (CI).

RESULTS: The final population included 195,420 NVAF patients, with 62.5% being White, 8.5% being AA, 8.3% being Hispanic, and 20.8% being another race/unknown. Post IPTW, the IR for discontinuation and switch were significantly lower for apixaban than warfarin (IR_{discontinuation}: 55.7 vs 75.3, p<0.0001; IR_switch: 6.3 vs 31.5, p<0.0001). The HR for discontinuation and switch were also significantly lower for apixaban than warfarin (HR_{discontinuation}: 0.76 [95%CI=0.75-0.77]; HR_Switch: 0.21 [95%CI=0.20-0.22]). When stratified by race and ethnicity, the trends were similar among all racial and ethnic groups.

CONCLUSIONS: Apixaban was associated with lower risk of discontinuation and switch compared to warfarin among all NVAF patients and stratified among White, AA, and Hispanic patients.