By Michael H. Crawford, MD, Editor

SYNOPSIS: A randomized, controlled trial of rivaroxaban vs. warfarin in patients with atrial fibrillation and a bioprosthetic valve revealed rivaroxaban is noninferior to warfarin for the prevention of major cardiovascular events and the avoidance of major bleeding events over 12 months.

SOURCE: Guimarães HP, Lopes RD, de Barros E Silva PGM, et al. Rivaroxaban in patients with atrial fibrillation and a bioprosthetic mitral valve. N Engl J Med 2020;383:2117-2126.

Rivaroxaban (RO) has been shown to be noninferior to warfarin for preventing stroke in patients with atrial fibrillation (AF). However, patients with prosthetic heart valves were excluded from these trials. Based on favorable outcomes in small pilot studies and subgroup analyses of larger observational studies, Guimarães et al conducted a multicenter, randomized, controlled trial of RO vs. warfarin in patients with AF or flutter and a bioprosthetic mitral valve in 49 sites in Brazil.

Inclusion criteria were adults with paroxysmal, persistent, or permanent AF and a bioprosthetic mitral valve who were either receiving or about to receive an oral anticoagulant. Exclusion criteria included extremely high risk of bleeding, transient AF after surgery, and contraindications to the study drugs. The RO dose was 20 mg daily (or 15 mg daily if estimated creatinine clearance was 30-49 mL/min/1.73m2). Warfarin was dose-adjusted to maintain an INR of 2-3, measured at least monthly. Follow-up visits were conducted at one, three, six, nine, and 12 months. Stroke and bleeding risk were assessed by the CHA2DS2-VASc and HAS-BLED scores. The primary outcome was a composite of death, major cardiovascular events, or major bleeding at 12 months. The primary safety outcomes were bleeding events classified by the ROCKET AF trial criteria and the BARC criteria.

Between 2016 and 2019, the authors randomized 1,005 patients, and only six were lost to follow-up. The mean age was 59 years, 62% were Black or multiracial, and 96% had AF. The mean CHA2DS2-VASc score was 2.6, the mean HAS-BLED score was 1.6, and 62% of patients underwent mitral valve replacement one to 10 years before study inclusion.

Study drug discontinuation was observed in 10% of the RO group and 7% of the warfarin group. In the intention to treat analysis, the time to the primary outcome was 348 days in the RO group and 340 days in the warfarin group (P < 0.001 for noninferiority). In the as-treated analysis, it was 350 days with RO and 340 days with warfarin. In the subgroup with valve replacement less than three months before enrollment, the results favored RO, with a time to the primary endpoint of 349 days vs. 314 days in the warfarin group.

The primary outcome event occurred in 6.4% of the RO group and 19% of the warfarin group (HR, 0.31; 95% CI, 0.12-0.79). The secondary outcome of death from cardiovascular causes or thromboembolic events was 3.4% with RO and 5.1% with warfarin (HR, 0.65; 95% CI, 0.35-1.2). Total stroke was 0.6% with RO and 2.4% with warfarin (HR, 0.25; 95% CI, 0.07-0.88). Valve thrombosis was 1% with RO and 0.6% with warfarin. Major bleeding occurred in 1.4% of the RO group and 2.6% of the warfarin group (HR, 0.54; 95% CI, 0.21-1.35). There were no intracranial bleeds with RO and five in the warfarin group. The authors concluded that in patients with AF and a bioprosthetic mitral valve, RO was noninferior to warfarin for preventing death, major cardiovascular outcomes, or major bleeding at one year.


The failure of the direct oral anticoagulants (DOAC) to prevent thromboembolic events in patients with mechanical prosthetic heart valves vs. warfarin was a great disappointment. The theory as to why is mechanical valves activate components of the clotting cascade that are not activated by stagnant blood in atria or other perturbations of natural tissue. However, with transcatheter valves increasingly used instead of surgery for failed bioprostheses, the longitudinal durability of mechanical valves is becoming less important. Many patients who would have received a mechanical valve because of their young age or other reasons now are receiving bioprostheses.

Since bioprostheses are more similar to native valves, it makes sense to test whether DOACs would provide equivalent protection from thrombotic events as warfarin. Small subgroups of the ARISTOTLE trial (apixaban) and the ENGAGE AF trial (edoxaban) of AF patients had bioprosthetic valves placed more than three months before enrollment and showed favorable results. The same was true for a small pilot study of dabigatran in patients with bioprosthetic valves.1-3 Also, a larger trial of edoxaban in patients with bioprosthetic valves with or without AF showed zero deaths, thromboembolic events, or intracranial hemorrhages on edoxaban vs. 3.7% in warfarin patients, with no difference in major bleeds over three months.4

Encouraged by these preliminary findings, Guimarães et al found patients with AF and a bioprosthetic mitral valve were free of major cardiovascular or bleeding events for 7.4 days longer than those on warfarin at one year. This result was statistically noninferior for RO, but not superior to warfarin. Also, secondary outcomes such as stroke (0.6% RO vs. 2.4% warfarin), bleeding, and valve thrombosis were similar. However, the incidence of these events were low, so interpret them with caution. In addition, Guimarães et al included patients who underwent mitral valve replacement within three months and found superior results for RO. Again, this was a smaller subgroup.

There were limitations to this study. It was open label, so the reporting of events could have been biased. However, a blinded, independent panel adjudicated outcomes. Also, it is challenging to blind a study with warfarin since the INR monitoring is hard to keep secret. Still, a trial with less frequent endpoints and smaller subgroup analyses are of limited validity. Finally, the authors correctly noted their results do not apply to those with bioprosthetic aortic valves, mechanical valves, or native valve mitral stenosis. Currently, in these groups, warfarin remains the recommendation.


  1. Avezum A, Lopes RD, Schulte PJ, et al. Apixaban in comparison with warfarin in patients with atrial fibrillation and valvular heart disease: Findings from the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) Trial. Circulation 2015;132:624-632.
  2. Carnicelli AP, De Caterina R, Halperin JL et al. Edoxaban for the prevention of thromboembolism in patients with atrial fibrillation and bioprosthetic valves. Circulation 2017;135:1273-1275.
  3. Durães AR, de Souza Roriz P, de Almeida Nunes B, et al. Dabigatran versus warfarin after bioprosthesis valve replacement for the management of atrial fibrillation postoperatively: DAWA pilot study. Drugs R D 2016;16:149-154.
  4. Hong G-R. 2020 ACC Scientific Sessions, March 28-30, 2020.