Apixaban causes half the bleeding of rivaroxaban for blood clots, head-to-head trial finds

The Ottawa Hospital

Two blood thinners dominate the treatment of venous thrombosis. Both are pills. Both prevent clots effectively. Both are recommended in clinical guidelines. Until now, doctors choosing between them have largely been guessing which is safer, relying on indirect comparisons across separate trials that used different patient populations and different study designs.

That guessing is over. The COBRRA trial, the first randomized controlled trial to directly compare apixaban and rivaroxaban in venous thrombosis patients, has produced a clear result: rivaroxaban causes more than double the rate of clinically relevant bleeding, with no compensating advantage in clot prevention.

7.1 percent versus 3.3 percent

The trial enrolled 2,760 participants from 32 sites across Canada, Australia, and Ireland. Patients with venous thrombosis, blood clots in the veins of the legs or lungs, were randomly assigned to receive either apixaban or rivaroxaban for the standard three-month treatment course.

After three months, 7.1 percent of participants receiving rivaroxaban experienced clinically relevant bleeding, compared to 3.3 percent of those receiving apixaban. The risk of recurrent blood clots was not significantly different between the two drugs.

Lead author Lana Castellucci, senior scientist and thrombosis physician at The Ottawa Hospital and professor at the University of Ottawa, described the results as highly anticipated evidence that should change clinical practice and provide peace of mind to patients living with the dual fear of clot recurrence and bleeding.

Why this trial was needed

Venous thrombosis is the third leading cause of cardiovascular death after heart attack and stroke, and the most common preventable cause of death in hospitalized patients. Standard treatment involves three months of anticoagulation therapy to prevent further clots. Both apixaban and rivaroxaban belong to the same drug class, direct oral anticoagulants, and have been individually shown to be effective.

But the two drugs were approved based on separate trials against different comparators. No one had put them against each other in a randomized trial until COBRRA. Observational studies had hinted at a bleeding difference, but observational data cannot account for all the confounding variables that randomization eliminates.

Real-world relevance across three countries

Senior author Marc Rodger, Physician-in-Chief at the McGill University Health Centre, emphasized that this was an academically led, government-funded trial, not an industry-sponsored study with proprietary interests in the outcome.

Vivien Chen, Thrombosis Lead Haematologist at Concord Hospital and professor at the University of Sydney, highlighted the global relevance of the multinational design. Patients across Canada, Australia, and Ireland were treated according to routine clinical practice, making the findings applicable to diverse healthcare systems and patient populations.

What the trial cannot resolve

The trial measured outcomes over a three-month treatment period. Some patients require longer-term anticoagulation, and whether the bleeding difference persists or widens over extended treatment is unknown.

The trial population had venous thrombosis specifically. Whether the same bleeding differential applies to patients taking these drugs for other indications, such as atrial fibrillation or post-surgical prophylaxis, cannot be inferred from these data.

The study was not designed to identify which patient subgroups might see the largest or smallest differences between the two drugs. Age, kidney function, body weight, and concurrent medications could all modify the bleeding risk, but subgroup analyses in a trial of this size would be underpowered.

Patients already stable on rivaroxaban should not switch medications based on this single study without consulting their physician. The trial establishes a population-level difference that should inform new prescribing decisions, not necessarily trigger changes in patients who are tolerating their current therapy.

A straightforward answer to a straightforward question

Clinical medicine rarely produces results this clean. Two drugs, same indication, same patient population, one causes half the bleeding with equal effectiveness. The COBRRA trial gives physicians prescribing anticoagulants for venous thrombosis something they have lacked until now: direct comparative evidence to guide their choice.