NOAC non-inferior to warfarin in AF patients with bioprosthetic mitral valve

Warfarin In Patients With Bioprosthetic Mitral Valves And Atrial Fibrillation Or Flutter: Primary Results From The RIVER Randomized Trial

News - Nov. 15, 2020

Presented during the AHA Scientific Sessions 2020 by Otavio Berwanger, Research Institute HCor and Albert Einstein Hospital, Sao Paulo, Brazil

Introduction and methods

Long-term anticoagulant therapy is required in patients with atrial fibrillation (AF) and a bioprosthetic mitral valve, but optimal strategy is unclear. In clinical practice, most patients use warfarin, because evidence with NOACs is limited. Subgroup analyses with AF patients and a mitral bioprosthetic valve from pivotal trials included less than 200 patients. For example the ARISTOTLE trial with 31 patients and the ENGAGE-TIMI 48 with 131 patients. Patients with bioprosthetic valves were excluded from the ROCKET-AF trial.

The RIVER trial enrolled 1005 patients with AF or flutter and a bioprosthetic mitral valve at 49 sites in Brazil, who were randomized to rivaroxaban (20 mg daily) or warfarin (INR target 2.0-3.0). Follow-up was 12 months. The RIVER trial was designed as a non-inferiority trial. The primary endpoint was all-cause mortality, major CV events or major bleeding.

Main results

  • The difference in restricted mean survival time (RMST), the event free time until 12 months, between the rivaroxaban group and the warfarin group was 7.4 days (95%CI: -1.4 to 16.3), meaning that the event free time in the rivaroxaban group was 7.4 days longer than that of the warfarin group. With a non-inferiority margin of -8 days, P for non-inferiority was <0.001.
  • No difference for secondary efficacy endpoints was observed between the two groups, with the exception of total stroke. Total stroke was reduced in the rivaroxaban group comped to the warfarin group (HR 0.25, 95%CI: 0.07 to 0.88, P=0.03). However, event rate was low, confidence interval wide and there was no correction for multiple testing.
  • Bleeding events (major, clinically relevant non major, minor and total) were not different between the two groups.
  • A subgroup of patients in whom mitral valve implantation was less than 3 months ago (~20% of the study population) showed a RMST difference of 35.2 days (95%CI: 8.6 to 61.7) with rivaroxaban compared to warfarin (HR 0.31, 95%CI: 0.12 to 0.79, Psuperiority=0.01).


The RIVER trial demonstrated that use of rivaroxaban is non-inferior to warfarin in AF patients with a bioprosthetic mitral valve. The authors suggest that these results can potentially inform practice and rivaroxaban may be an attractive alternative for this patient population.


The discussant Elaine Hylek, MD (Boston University School of Medicine, Boston, MA, USA) compared the patients in RIVER with those in the ROCKET AF trial. Those in RIVER a low risk population; they were younger, less previous stroke patients and CHA₂DS₂VASC score was lower. A lower risk of AF in this population could better isolate the thrombotic risk from the bioprosthetic valve, said Hylek, which was a strength in the design of the study. However, these findings may not necessarily be extrapolated to an older population for whom bioprosthetic valves would be preferred.

Two other points that were raised by prof. Hylek was the low number of patients enrolled within 2-90 days post bio valve surgery (risk-laden period) and a systemic assessment of valve thrombosis/dysfunction (by exit study echocardiography) could fortify the results.

- Our reporting is based on the information provided during AHA Scientific Sessions 2020 -

The findings of this study were simultaneously published in N Eng J Med

Facebook Comments


We’re glad to see you’re enjoying PACE-CME…
but how about a more personalized experience?

Register for free