Health Care Resource Use and Costs of Rivaroxaban Versus Warfarin Among Nonvalvular Atrial Fibrillation Polypharmacy Patients With Obesity.

Background: The economic burden of nonvalvular atrial fibrillation (NVAF) is substantial. Many patients with NVAF are obese and manage other health conditions requiring multiple medications. This real-world study compared health care resource use (HRU) and costs for rivaroxaban and warfarin in patients with NVAF who had polypharmacy and obesity.

Methods And Results: We used health care claims databases (Merative MarketScan commercial and Medicare supplemental claims) to identify patients initiating the direct oral anticoagulant rivaroxaban or warfarin with ≥1 diagnostic claim for atrial fibrillation, presence of polypharmacy (based on 3 categories for the number of concurrent medications: 1-4, 5-9, ≥10), and obesity. Cohorts were balanced for demographic and baseline characteristics using propensity score matching. All-cause and NVAF-related HRU rates and costs were compared between treatments using rate ratios and adjusted mean differences per patient per year. Eligible patients totaled 95 875, with 19 990 patients in each treatment cohort following propensity score matching. All-cause HRU rates were significantly lower with rivaroxaban versus warfarin, and hospital stays were reduced by 3.1 days with rivaroxaban. Mean (95% CI) all-cause total medical and total health care costs per patient per year were significantly reduced with rivaroxaban versus warfarin (-$4499 [-$5660 to -$3305] and -$1627 [-$2790 to -$438], respectively). NVAF-related HRU was reduced with rivaroxaban versus warfarin, but total NVAF-related medical costs were not significantly different between treatment groups ($144 [-$756 to $1079] per patient per year). Subgroup and sensitivity analysis results were generally consistent with the main analysis.

Conclusions: Among patients with NVAF, polypharmacy, and obesity, rivaroxaban was associated with a reduction in HRU and all-cause costs compared with warfarin.