Use of direct oral anticoagulant in atrial fibrillation with gastrointestinal bleeding: Balancing gastrointestinal safety and cardiovascular risks

In January 2019, the American College of Cardiology, American Heart Association, and the Heart Rhythm Society joined hands to publish an update on guidelines for management of atrial fibrillation (AF).1 In the updated guideline, direct (non-warfarin) oral anticoagulants (DOACs) are considered “… as a group, superior to warfarin for preventing stroke and systemic embolism and were associated with lower risks of serious bleeding.” The preference of DOACs over warfarin is in fact based more on improved safety, in particular, a reduction in bleeding risk. This is an important shift in guideline recommendation and, together with the global aging population, will lead to an exponential rise in use of DOACs in the years to come. Although DOACs are reported to have improved safety, the annual incidence of gastrointestinal (GI) bleeding among users is still up to 3.2% with rivaroxaban, which confers the highest risk among DOACs,2 whereas apixaban and edoxaban are relatively safer with the former (apixaban) having the most favorable balance of efficacy for stroke prevention and safety.3 Factors including old age, concomitant use of antiplatelet agents, and hepatic and/or renal dysfunction also add to the increased risk for GI bleeding.4 So, when AF patients presenting with GI bleeding on DOACs undergo endoscopic procedures, how should one manage their anticoagulation? Most clinicians would stop anticoagulants in these emergency GI conditions. There are, however, very little data to guide how DOACs should be withheld and when to resume these agents for stroke prophylaxis after bleeding is controlled or when high-risk endoscopic procedures (e.g. endoscopic papillotomy, endoscopic polypectomy, and endoscopic mucosectomy or submucosal dissection) are finished. Neither American nor European Guidelines provide clear guidance on this. The Asian Pacific Association of Gastroenterology and the Asia Pacific Society for Digestive Endoscopy recommend resuming DOAC by day 3 once hemostasis is achieved without bridging therapy.5 This bold step is based on theoretical interpretation of pharmacokinetics of DOACs instead of clinical evidence. While we (gastroenterologists) are quite ready to interrupt DOAC to avoid bleeding during endoscopic procedures or rebleeding after achieving hemostasis, we must not forget that patients are taking these drugs for good reasons, that is, to prevent serious thromboembolic events. In a study where the anticoagulant effect of dabigatran was reversed by idarucizumab, 4.8% of subjects developed a thrombotic event within 30 days and 6.8% within 90 days.6 In a prospective cohort analysis of over 10 000 cases of upper GI bleeding from peptic ulcer disease, the majority of patients died of non-bleeding-related causes including multi-organ failure, pulmonary conditions, and cardiovascular or cerebrovascular deaths.7 For the same reason, when aspirin was discontinued in patients presenting with GI bleeding, compared to those in whom aspirin was resumed early, these patients suffered a much higher mortality rate from cardiovascular and cerebrovascular deaths.8 Besides, AF patients seem to be more willing to accept the risk of GI bleeding than the risk of stroke as a result of discontinuing anticoagulants.9 When choosing between continuing and interrupting NOAC in a GI bleeding patient, we have to consider both the GI risk (of continuing DOAC) and the thrombotic risk (of withholding DOAC). The concept of “net clinical benefit” emerges in a delicate situation of bleeding versus stroke prevention. The risk of GI bleeding is determined by the nature of the endoscopic procedure10, 11 or the endoscopic appearance of the ulcer or varix on presentation. The thrombotic risk, on the other hand, is determined by a number of variables summarized in the CHA2DS2-VASc score. Oral anticoagulants are usually prescribed for patients with CHA2DS2-VASc score ≥ 2 in men or ≥ 3 in women.1 For moderate to high risk of thromboembolism in AF, CHA2DS2-VASc offers little guidance. There are several notions worth discussing and debating. First, the anticoagulant effects of DOAC, even with renal impairment, diminish relatively quickly (within 12–24 h) as the half-life of these drugs is short.12 Temporary cessation of DOAC is probably sufficient without the need for reversal therapies, as charcoal, vitamin K, or bridging therapy offers no benefit. Second, mortality risk is usually higher with cardiovascular and cerebrovascular thrombotic events than with GI bleeding (except in massive bleeding leading to hypovolemic shock). Most importantly, there is no randomized study addressing the issue of timing to restart DOAC in the context of GI bleeding. Patients with renal or hepatic decompensation require dose adjustments when prescribed DOAC. The following practical recommendations are proposed for discussion (Table 1). Robust data from well-designed clinical trials are very much in need to validate this. Low thrombotic risk CHA2DS2-VASc score 0–1 High thrombotic risk CHA2DS2-VASc score ≥ 2 (male) or 3 (female) Lower GI bleeding risk • Low risk ulcer or varix (e.g. clean base or pigmented spot ulcer and small varix with no red wale sign) • Low endoscopic risk (e.g. diagnostic endoscopy, biliary stenting, capsule endoscopy, enteral stent, EUS without FNA, and APC) High GI bleeding risk • High-risk ulcer or varix (e.g. ulcer with protuberant vessel or adherent clot and large varix with cherry red spot) • High-risk endoscopic procedure (e.g. endoscopic hemostasis, polypectomy, sphincterotomy, PEG placement, and EMR and ESD)