Real-world vs. randomized trial outcomes in similar populations of rivaroxaban-treated patients with non-valvular atrial fibrillation in ROCKETAF and XANTUS

Aims Based on Phase III data, non-vitamin K antagonist oral anticoagulants are recommended for stroke prevention in patients with atrial fibrillation. To determine whether trial outcomes translate into similar event rates in unselected patients, this analysis compared outcomes from the real-world XANTUS study with those from the Phase III ROCKET AF study. Methods and results Individual patient data from 4020 XANTUS patients were re-weighted to match the proportion of selected baseline characteristics in 7061 rivaroxaban-treated patients from ROCKET AF, using the matching-adjusted indirect comparison (MAIC) method. For the primary analysis, CHADS2 scores and gender were selected as relevant variables. Adjusted annualized incidence rates for XANTUS were calculated and compared with incidence rates from ROCKET AF-the ratio of these rates ('MAIC ratio') was used as a relative effect estimate. Rates of major bleeding [3.10%/year vs. 3.60%/year; MAIC ratio 0.86; 95% confidence interval (CI) 0.67-1.12] and stroke/non-central nervous system systemic embolism (1.54%/year vs. 1.70%/year; MAIC ratio 0.91; 95% CI 0.62-1.32) were similar between XANTUS and ROCKET AF. The rate of all-cause death was higher in XANTUS (3.22%/year vs. 1.87%/year; MAIC ratio 1.72; 95% CI 1.31-2.27), but the rates of vascular death were similar (1.83%/year vs. 1.53%/year; MAIC ratio 1.19; 95% CI 0.84-1.70). Sensitivity analyses weighted by different baseline characteristics supported these results. Conclusion The low rates of major bleeding and stroke in XANTUS were consistent with results from ROCKET AF. All-cause death, but not vascular death, was higher in XANTUS, as expected in an unselected real-world population.