MUNICH – The increased risk of major GI bleeding documented with dual-antiplatelet therapy using rivaroxaban and aspirin when compared with aspirin alone for vascular protection in the previously reported massive COMPASS trial may turn out to be a blessing in disguise.

That’s because a new secondary analysis of COMPASS (Cardiovascular Outcomes for People Using Anticoagulation Strategies), a randomized trial that included 27,395 patients with stable coronary or peripheral artery disease followed for a mean of 23 months at more than 600 centers in 33 countries, identified a strong association between early bleeding in rivaroxaban-treated patients and subsequent early diagnosis of new GI cancer. Ditto for rivaroxaban-associated genitourinary bleeding and subsequent GU cancer.

“Among COMPASS patients with vascular disease receiving long-term antithrombotic therapy, more than 1 in 5 new diagnoses of cancer are preceded by bleeding. GI and GU bleeding are powerful and relatively specific predictors of new GI and GU cancer diagnoses, respectively, and more than 75% of these cancers diagnosed after bleeding are diagnosed within 6 months of the bleed,” John W. Eikelboom, MD, reported at the annual congress of the European Society of Cardiology.

These findings strongly suggest that rivaroxaban (Xarelto), a direct-acting oral anticoagulant (DOAC), may be unmasking occult GI and GU cancers earlier than the malignancies would otherwise have declared themselves.

“Although overall cancer rates were similar in the three treatment groups [rivaroxaban at 2.5 mg twice daily plus aspirin at 100 mg/day, rivaroxaban monotherapy at 5 mg twice daily, or aspirin alone at 100 mg/day], the early increase in GI bleeding with rivaroxaban-based therapy resulted in earlier diagnosis of GI cancer in these patients. By reducing major cardiovascular events and mortality, the combination of rivaroxaban and aspirin already produces a clear net benefit, and by unmasking GI cancers at an earlier stage, the combination could potentially lead to the added benefit of improved GI cancer outcomes,” commented Dr. Eikelboom, a hematologist at McMaster University in Hamilton, Ont., and lead investigator of the previously published COMPASS trial (N Engl J Med. 2017 Oct 5;377[14]:1319-30).

This possibility that unmasking of occult GI cancer might result in improved survival deserves to be a high research priority in light of the enormous death toll caused by colorectal cancer. Planned longer-term follow-up of the COMPASS participants should be helpful in this regard, he added.

This new COMPASS finding effectively makes a silk purse out of a sow’s ear. When the primary outcomes of COMPASS were announced, many physicians reasoned that if the other commercially available DOACs also outperform warfarin but don’t pose a significantly increased threat of serious bleeding events, why not preferentially turn to them rather than rivaroxaban in patients with high HAS-BLED scores? But now rivaroxaban’s increased GI bleeding risk has begun to look like a potentially important advantage.

As previously reported, COMPASS participants on rivaroxaban plus aspirin had a 3.1% major bleeding rate as defined according to modified International Society on Thrombosis and Hemostasis criteria, for a statistically significant 70% increase in risk, compared with the 1.9% rate in patients on aspirin alone. Rivaroxaban monotherapy also was associated with increased bleeding risk. Of note, most of the excess in bleeding involved GI bleeding, and it was front-loaded during the first year of the trial. Also, reassuringly, there was no increased incidence of intracranial or fatal bleeding in patients on rivaroxaban.

A total of 1,082 patients were diagnosed with a new cancer during 23 months of follow-up. Nearly a quarter (23%) of the new GI cancers and 45% of the new GU cancers were diagnosed after bleeding from those sites.

The incidence of GI cancer diagnosed after GI bleeding was 7.8%, whereas patients with no prior GI bleeding had a mere 0.9% rate of newly diagnosed GI cancer during the study period. Thus, roughly 1 out of every 12 cases of GI bleeding was associated with diagnosis of a new GI cancer, and GI bleeding was associated with a 13-fold increased risk of subsequent GI cancer diagnosis. Put another way, the number of cases of GI bleeding occurring in patients on rivaroxaban that needed to be investigated in order to find one new GI cancer was 12.

Similarly, 13% of COMPASS participants who developed GU bleeding were subsequently diagnosed with a new GU cancer, compared with just 0.3% of subjects without GU bleeding. That translates into an 83.4-fold increased risk of GU cancer in patients with GU bleeding.

In contrast, the incidence of non-GI cancer following a GI bleed was 1.5%, while the rate was 1.0% in patients with no prior GI bleeding.

“That relationship was much weaker, with an odds ratio of 1.77, indicating that the relationship between GI bleeding and GI cancer is not only very strong but it’s rather specific,” according to Dr. Eikelboom.

Of GI cancers associated with a prior major GI bleed, 77% were diagnosed within 6 months following the bleed, as were nearly 89% of GU cancers. Another 9%-10% were diagnosed 6-12 months after the bleeding event.

“The implication for clinical practice is certainly that in patients who have GI or GU bleeding while receiving antithrombotic therapy, we should conduct a vigorous search for underlying cancer in the same organ system,” he concluded.

Discussant Lars C. Wallentin, MD, concurred. He called it a wake-up call for cardiologists to broaden their horizons and recognize that their elderly patients with vascular disease also are at substantial competing risk for major noncardiovascular diseases – and that his colleagues may have an important role to play in earlier cancer diagnoses.

He and his coinvestigators in the RE-LY (Randomized Evaluation of Long-Term Anticoagulant Therapy) study made a similar point in their investigation of 18,113 patients with atrial fibrillation on oral anticoagulation for stroke protection. In that randomized trial of dabigatran (Pradaxa) versus warfarin, roughly 1 in 12 major GI bleeding events was found to be related to an occult colorectal or gastric cancer (Clin Gastroenterol Hepatol. 2017 May;15[5]:682-90).

“The data are very consistent. The message to cardiologists is that no bleeding should be disregarded in patients on oral anticoagulation,” declared Dr. Wallentin, professor of cardiology at Uppsala (Sweden) University.

COMPASS was sponsored by Bayer. Dr. Eikelboom reported receiving research grants from that company and more than half a dozen others.

bjancin@mdedge.com

Rivaroxaban bonus: Early unmasking of occult GI cancers