RIVER: Rivaroxaban Rivals Warfarin in Atrial Fib With Bioprosthetic Mitral Valves

November 15, 2020

The direct oral anticoagulants (DOACs) took cardiology by storm a decade ago when trials showed they were at least as safe and effective as warfarin for several indications, especially reducing stroke risk in atrial fibrillation (AF). But the trials that established the drugs in AF largely excluded patients with valve disease, leaving them without a good evidence base in support of the drugs.

Now that evidence base is a bit firmer. In the RIVER trial, treatment with the factor Xa inhibitor rivaroxaban (Xarelto, Bayer/Janssen) was noninferior to warfarin for a primary endpoint that included death, cardiovascular events, stroke, and bleeding over 1 year in more than 1000 patients with AF and bioprosthetic mitral valves.

Unlike warfarin, rivaroxaban doesn't require anticoagulation monitoring for dosage titration, has more consistent effects, and so "represents an attractive alternative for this population," said Otavio Berwanger, MD, PhD, when presenting the RIVER results November 14 during the  American Heart Association (AHA) Scientific Sessions 2020 virtual meeting.

Patients with AF and bioprosthetic mitral valves were largely absent from the major nonvalvular AF DOAC trials, and "most of our recommendations in patients with valve disease are based on expert opinion," Berwanger said. For example, ARISTOTLE with apixaban (Eliquis, Pfizer/Bristol-Myers Squibb) and ENGAGE-TIMI 48 with edoxaban (Lixiana, Daiichi-Sankyo) between them included only about 160 patients with bioprosthetic mitral valves. There were none in the major rivaroxaban trial, ROCKET-AF.

RIVER "sets the stage to better study this population," said Berwanger, from Research Institute HCor and Albert Einstein Hospital, Sao Paulo, Brazil, who is also senior author on the trial's same-day publication in the New England Journal of Medicine.

Of note, women made up about 60% of the population in RIVER, which was conducted at 49 centers in Brazil; the entire population's median age was 59 years.

As a result, they were probably at much lower risk for stroke compared with the older patients in the three nonvalvular AF trials, observed Elaine M. Hylek, MD, MPH, invited discussant following the Berwanger presentation. For example, she said, the mean age in ROCKET-AF was 73 years.

That means "we may not be able to extrapolate the efficacy and safety from this particular study to older patients in whom a bioprosthetic valve would likely be preferred," said Hylek, from Boston University School of Medicine in Massachusetts.

The RIVER population was indeed "pretty young for patients getting bioprosthetic valves," agreed Athena Poppas, MD, Brown University School of Medicine, Providence, Rhode Island, in an interview. That cautions against applying their experience to, for example, North American or European populations with AF and prosthetic mitral valves. The generally older age of the latter patients may put them at higher bleeding risk than the patients in RIVER.

Also, it's possible that many patients in the Brazilian trial received their bioprosthesis because of rheumatic valvular disease. Such an inflammatory heart process would likely elevate their risk for stroke, observed Poppas, who was not part of RIVER. "Their atria can be abnormal and at higher risk of thrombosis." That would also complicate comparisons with other populations.    

Rheumatic heart disease drove the need for a prosthetic mitral valve in about 20% of patients in RIVER, Berlanger told theheart.org | Medscape Cardiology, "so we believe most of its patients are comparable to patients in North American and Europe."

RIVER randomly assigned 1005 adults with AF and bioprosthetic mitral valves to receive rivaroxaban, 20 mg/day (or 15 mg/day in patients with renal dysfunction), or warfarin titrated to a target international  normalized ratio of 2.0 to 3.0 on an open-label basis.  Randomization occurred at least 48 hours after the implantation surgery; it was within 3 months of surgery in 18.8% of the cohort and at least 1 year afterward in about 63%.

Patients assigned to rivaroxaban or to warfarin fared similarly over 12 months for the primary endpoint of all-cause mortality, stroke, transient ischemic attack, major bleeding, valve thrombosis, non-neurologic systemic embolism, or hospitalization for heart failure.

The protocol called for the primary outcome to be interpreted with respect to restricted mean survival time (RMST)—that is, average time free from a primary outcome event, which reflects the area under the survival curve— RIVER documentation explains. The time to such an event on warfarin was subtracted from that for rivaroxaban, yielding an RMST that, when positive, indicated a rivaroxaban benefit.

The mean time until a primary event by intention-to-treat analysis was 347.5 days for the 500 patients assigned to rivaroxaban and 340.1 days for the 505 receiving warfarin, for a rivaroxaban RMST of 7.4 days at 1 year (P < .001 for noninferiority; P = .10 for superiority).

In two sensitivity analyses, the rivaroxaban RMST for outcomes was 9.6 days on a per-protocol basis and 10.5 days on an as-treated basis. Both were significant for noninferiority at P < .001 and also for superiority at P = .01, Berlanger reported.

Those analyses therefore "suggest that rivaroxaban was not only noninferior, but also superior to warfarin for the primary endpoint," he proposed.

Table. Selected Secondary Outcomes at 1 Year in RIVER, Rivaroxaban vs Warfarin

Endpoint Hazard Ratio (95% CI) P Value
Death from any cause 1.01 (0.54 - 1.87) .98
Cardiovascular death or thromboembolic events 0.65 (0.35 - 1.20) .17
Any stroke 0.25 (0.07 - 0.88) .03
Major bleeding 0.54 (0.21 - 1.35) .18
Valve thrombosis 1.68 (0.40 - 7.01) .48


In a subgroup analysis of the 189 patients randomly assigned within 3 months of bioprosthetic valve surgery (18.8% of the population), the 94 in the rivaroxaban group showed a significant advantage over the 95 warfarin recipients for the primary endpoint, at a hazard ratio of 0.31 (95% CI, 0.12 - 0.79; P = .01) and an RMST of 35.2 days.

"Definitely there's an interesting signal there that deserves to be further tested," Berwanger said. "No trials in the past have included such a population," he told theheart.org | Medscape Cardiology.

The rivaroxaban noninferiority finding in RIVER's patients may well be a DOAC class effect, Poppas speculated, much as the drugs collectively showed at least similar efficacy and safety in nonvalvular AF. That especially applies to the other factor Xa inhibitors, apixaban and edoxaban, she said, but potentially also the direct thrombin inhibitor dabigatran (Pradaxa, Boehringer-Ingelheim).

RIVER was funded by Bayer and the Brazilian Ministry of Health. Berwanger discloses receiving research grants from AstraZeneca, Bayer, Amgen, Novartis, Pfizer, Boehringer-Ingelheim, and Servier. Hylek discloses receiving research support from Abbott, Bristol-Myers Squibb, Janssen, Boehringer-Ingelheim, Squibb, Medtronic, and CryoLife; consulting for Bayer, Boehringer-Ingelheim, BMS/Pfizer, and Janssen; and receiving honoraria from Boehringer-Ingelheim and BMS/Pfizer. Poppas had no relevant disclosures.

American Heart Association (AHA) Scientific Sessions 2020. LBS.03. Presented November 14, 2020.

N Engl J Med. November 14, 2020. Abstract

Follow Steve Stiles on Twitter: @SteveStiles2. For more from theheart.org | Medscape Cardiology, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.