Major Bleeding Mostly Early, Not Severe With Rivaroxaban Plus Aspirin in COMPASS

Most instances of major bleeding among patients with CAD or PAD who take rivaroxaban plus aspirin occur within the first year of treatment, according to a new analysis from the COMPASS randomized trial. On top of that, the advantage of combination therapy over aspirin alone in preventing CV events does not taper off during year two.

John W. Eikelboom, MBBS (Hamilton Health Sciences, Hamilton, Canada), outline their findings in a paper published online ahead of the September 24, 2019, issue of the Journal of the American College of Cardiology.

COMPASS, released at the European Society of Cardiology Congress 2017 and simultaneously published in the New England Journal of Medicine, showed that using the combination of a 2.5-mg, twice-daily dose of rivaroxaban (Xarelto; Bayer/Janssen) and low-dose aspirin reduced the composite risk of CV death, MI, or stroke compared with placebo plus aspirin in patients with stable CAD, PAD, or both (4.1% vs 5.4%; HR 0.76; 95% CI 0.66-0.86).

Rivaroxaban/aspirin also increased major bleeding (3.1% vs 1.9%; HR 1.70; 95% CI 1.40-2.05), although an analysis looking at the net clinical benefit—which incorporated risks of the most serious bleeds—found the combination to be superior.

On the strength of those results—COMPASS had been halted early due to “overwhelming efficacy”—the US Food and Drug Administration approved rivaroxaban for CV event reduction in CAD and PAD patients in October 2018.

Commenting on the latest COMPASS report for TCTMD, Usman Baber, MD (Icahn School of Medicine at Mount Sinai, New York, NY), said: “This analysis, and the COMPASS trial in aggregate, to me is consistent with this evolving paradigm that identifying the patient populations who are truly enriched with manifest atherothrombotic disease tend to derive the greatest benefit.”

PEGASUS-TIMI 54, the DAPT Study, and THEMIS-PCI, among others, all point to the idea that, when considering the pros and cons of antithrombotic therapy, the balance hinges on who’s being treated, Baber said.

Bleeding Under the Microscope

In this analysis of COMPASS data, Eikelboom and his co-investigators sought to better understand the location, severity, timing, and management of bleeding that occurred during the trial. Mean follow-up was 23 months for the 18,278 patients randomized.

As previously reported, the most common site of major bleeding was the GI tract (1.5% with rivaroxaban vs 0.7% with aspirin alone), followed by intracranial, skin or injection site, eye, nasal, urinary, respiratory, and genital tract bleeding.

For both the rivaroxaban combo and aspirin alone, the researchers note in JACC, GI bleeds were evenly divided among three locations: gastric or duodenal, colonic or rectal, and from an unknown GI site.

Around three-quarters of events categorized as “major bleeding” were, in terms of investigator-reported intensity, mild (“usually transient in nature and generally not interfering with normal activities”) or moderate (“sufficiently discomforting to interfere with normal activities”), they say.

Patients who experienced major bleeding were managed with conventional therapies. Red blood cell transfusions were given to 30.2% of those in the rivaroxaban group and 25.9% of the aspirin-alone group. Use of platelets, clotting factors, or other hemostatic agents was rare and equally distributed between the treatment arms.

Most of the discrepancy in bleeding events occurred within the first year after randomization but adverse event and mortality rates stayed steady from year one to year two.

Timing of Events: COMPASS

“Among persons with chronic CAD or PAD, most of the increase in bleeding with the combination of rivaroxaban and aspirin compared with aspirin alone was GI [and] occurred in the first year after starting treatment. . . . In contrast, the reduction in CV death, stroke, or MI continued to accrue after the first year, suggesting accumulating net clinical benefit over time,” Eikelboom et al conclude.

How Much Does Bleeding Matter?

Christoph Bode, MD, Christoph B. Olivier, MD, and Daniel Duerschmied, MD (all from the University of Freiburg, Germany), propose in an accompanying editorial that perhaps bleeding isn’t the “bad boy” it is often made out to be.

Most of the events we prevent and most of the adverse outcomes we cause are nonfatal in nature. Usman Baber

“As long as an antithrombotic drug without bleeding side effects is not available, however, bleeding can also be considered as evidence of efficacy,” they write. “Adverse cardiovascular outcome that is associated with bleeding is often rather a consequence of discontinuation of antithrombotic medication than of the bleeding event itself. Bleeding is not always clinically significant.”

Importantly, the editorialists say, bleeding “appears to be front-loaded, a phenomenon that might be accessible to other therapeutic interventions reducing bleeding risk.” They acknowledge, however, that patients with repeat episodes of severe bleeding need to be watched carefully and may need de-escalation of antithrombotic therapy.

They conclude that the latest COMPASS analysis “is reassuring and promises a true net clinical benefit for high-risk patients with stable atherosclerotic disease. Bleeding is never intended, but rarely is it prohibitive.”

Baber, though, cautioned against being too casual about bleeding events. “In fact, numerous studies have shown that the relationship between the occurrence of bleeding and subsequent death is quite strong. It’s significant, and in fact it approximates even that of an MI,” he stressed.

This isn’t to say that bleeding is the cause of death in these patients. “It may be that the occurrence of bleeding identifies a higher risk or leads to discontinuation of therapy that in turn causes an adverse outcome,” Baber explained. “We don’t understand all the mechanisms, but it is clear that bleeding is an event that has a high prognostic impact and should be avoided.” This, he said, is the inspiration behind trials exploring whether the duration of dual antiplatelet therapy can safely be shortened.

The notion that “if bleeding doesn’t kill you it doesn’t matter” for Baber is a “gross oversimplification.”

“Most MIs don’t kill us,” he continued, “so if one assumes that the threshold of relevance is whether or not the event is mortal, we basically wouldn’t be treating anybody. Most of the events we prevent and most of the adverse outcomes we cause are nonfatal in nature.”

Another ingredient in decision-making here is patient preference, Baber advised, noting, “Patients have very different interpretations of how much bleeding they are willing to incur or tolerate to prevent events. [In practice] that can certainly alter the calculus of the risk/benefit equation.”

As of yet, Baber said, the rivaroxaban/aspirin combo hasn’t been that widely adopted for CAD and PAD patients, in part because starting any new therapy involves an up-front risk. “And this analysis supports that—that’s where all the bleeding occurred. I think that makes the acceptance, perhaps, on the part of either clinicians or patients a little more difficult. It’s not that they don’t benefit . . . but it’s always a little more challenging to initiate patients who are otherwise doing well on an additional therapy,” he commented.