The management of medication use in the LGIB setting requires considering the risks of ongoing/recurrent bleeding and thromboembolic events (Figure 1). Cessation of these agents can be considered in patients on antithrombotic agents with life-threatening or serious bleeding. Although there are few data to guide the timing of the resumption of antithrombotic agents, current guidelines recommend resumption as soon as hemostasis is achieved[13,14]. A multidisciplinary approach involving cardiology, neurology, hematology, and gastroenterology is necessary, particularly for managing patients taking dual antiplatelet agents or anticoagulants.
Figure 1 Recommendation for the management of medication based on current studies.
1During the first 30 d following coronary stenting and during the first 90 d following acute coronary syndrome; 2The influence of short-term discontinuation has not been determined; 3Aspirin should be continued; 4Resumption reduces cardiovascular events but may increase rebleeding; 5The influence of long-term discontinuation has not been determined; 6Changing to apixaban, or reducing the dose of dabigatran to 110 mg b.i.d may reduce rebleeding in GIB patients taking warfarin, dabigatran (150 mg b.i.d) or rivaroxaban. NSAIDs: Nonsteroidal anti-inflammatory drug; DOAC: Direct-acting oral anticoagulant; PT-INR: Prothrombin time-international normalized ratio.
Antiplatelet agents increase the risk of both event and recurrence of LGIB[11,79,80,83]. The risk of LGIB with antiplatelet agents use is approximately three times that for UGIB[84,85], probably because LGIB lacks prophylactic measures, such as H. pylori eradication and PPIs.
Available data on the influence of discontinuing aspirin in the GIB setting are as follows. A retrospective cohort study of patients with LGIB showed that the rate of cardiovascular events was significantly higher in those who discontinued aspirin (37%) than in those who continued the drug (23%), while the rate of recurrent LGIB was lower in the former cohort (7%) than in the latter cohort (19%) within 5 years. In an RCT of peptic-ulcer bleeding, 60 d mortality was significantly higher in patients who discontinued aspirin after endoscopic therapy than in those who continued; the rate of rebleeding was not different between the groups. Based on this evidence, aspirin for secondary prophylaxis in patients with established cardiovascular disease should not be interrupted to prevent thrombotic events in the LGIB setting. However, aspirin as the primary prophylaxis for patients who are not at high risk of cardiovascular events had little effect (0.07% absolute risk reduction) and should be discontinued after LGIB.
The influence of short-term drug interruption in single antiplatelet users (aspirin or other antiplatelet agents) has not been determined. No difference in in-hospital rebleeding was observed in a retrospective study comparing patients who had their antiplatelet drug stopped for < 5 d with those who continued it throughout their admission. In that study, cardiovascular events were too few to allow meaningful comparison.
There are some data that can guide the management of dual antiplatelet therapy. The risk of myocardial infarction and death after discontinuing dual antiplatelet therapy is high during the first 30 d following coronary stenting and during the first 90 d following acute coronary syndrome. Such patients are advised to continue dual therapy. In contrast, discontinuing the second non-aspirin antiplatelet agent for up to 7 d is allowed for patients with more distant coronary stenting or coronary syndrome, because it seems to carry a relatively low risk as long as aspirin is continued.
Anticoagulants are classified into warfarin and DOACs. Two types of DOACs are currently available: thrombin inhibitors (dabigatran) and coagulation factor Xa inhibitors (rivaroxaban, apixaban, and edoxaban). Current endoscopic and LGIB guidelines do not discuss the role of a heparin bridge sufficiently, nor management of DOACs in the acute GIB setting[13,14]. Evidence is mainly based on studies of UGIB or all types of GIB.
Prothrombin time-international normalized ratio and the reverse method: Guidelines[13,14] recommend INR < 2.5 as being reasonable for endoscopy in the acute GIB setting, based on reports that a moderate elevation in INR does not increase the risk of rebleeding following endoscopic therapy for nonvariceal UGIB[91-93]. Guidelines also recommend using reversal agents before endoscopy for patients with an INR > 2.5, but the evidence for this is not well-established. Indeed, some retrospective studies found that a higher INR does not increase the rebleeding rate in LGIB or all types of GIB. Thus, an elevated INR appears not to carry a risk of rebleeding. However, an elevated INR at onset has been reported to be a predictor of thromboembolism within 90 d of endoscopy for all GIB (INR > 2.5, OR: 7.9), and of mortality for nonvariceal UGIB (INR > 1.5, OR: 5.6). This is presumably because INR is an indicator of underlying comorbid diseases. In a study of all types of GIB, other factors related to anticoagulant management, such as the difference in onset and pre-endoscopic INR, reversal agent use, and anticoagulant interruption, were associated with thromboembolism. Therefore, it might be unnecessary to actively reduce the INR. Rather, early endoscopy without using a reversal agent or interrupting anticoagulant therapy may be warranted for acute GIB.
Reversal of the anticoagulant effect should be considered for ongoing severe bleeding via intravenous vitamin K, fresh frozen plasma, or prothrombin complex concentrate (PCC) for warfarin users[97,98], and via oral charcoal, hemodialysis, idarucizumab, or PCC for DOAC users[99-102]. Oral charcoal is considered if a DOAC was taken within 2 h. Hemodialysis or idarucizumab is considered for dabigatran users. The effect of PCC on bleeding of DOAC users has not been established.
Heparin Bridge: Previous reports suggest that a heparin bridge might be ineffective in the acute GIB setting. A heparin bridge did not significantly alter the risk of rebleeding or thromboembolism in a recent retrospective study of patients with GIB. In an RCT of warfarin users undergoing invasive procedures, the heparin bridge group suffered from more major bleeding than the non-bridged group, without a difference in the thromboembolism rate during the periprocedural period. Furthermore, a similar result was found in a prospective observational study of DOAC users undergoing interventional procedures.
Resumption of anticoagulants: A meta-analysis concluded that resuming anticoagulants reduces the rate of thrombotic events in patients with disrupted use of anticoagulants due to GIB (HR: 0.68, 95%CI: 0.52-0.88), and mortality (HR: 0.76, 95%CI: 0.66-0.88), without significantly increasing the rebleeding rate (HR: 1.20, 95%CI: 0.97-1.48). This result was consistent with other reports[106-108]. Studies that compared warfarin and DOAC users reported that the rate of thrombotic events was similar between the two groups, during the 90 d after GIB and during the anticoagulant-interrupted period. A retrospective cohort study on DOAC users reported that the rate of thromboembolism within 90 d of GIB did not differ between those who resumed DOAC and those who did not. In that study, a history of venous thromboembolism was associated with thromboembolism events (HR: 3.30, 95%CI: 1.29-7.38).
The optimal duration before restarting anticoagulants after an episode of GIB remains uncertain. In a retrospective cohort study, the HRs of rebleeding, thromboembolism, and mortality in patients who resumed warfarin within 7 d were 3.27 (95%CI: 1.82-5.91), 0.76 (95%CI: 0.37-1.59), and 0.56 (95%CI: 0.33-0.93), respectively, compared with patients who resumed warfarin after 1 mo.
The bleeding risk of individual anticoagulants should be considered, when resuming anticoagulants in patients with high-risk GIB. Changing to apixaban, or reducing the dose of dabigatran to 110 mg b.i.d may reduce rebleeding in GIB patients taking warfarin, dabigatran (150 mg b.i.d) or rivaroxaban[111-115]. The HAS-BLED is a scoring system to evaluate bleeding risk among anticoagulants users. However, the main outcome of the score is composite bleeding events, including intracerebral hemorrhage and GIB. One study focused specifically on the risk of acute GIB in anticoagulant users, and developed a new scoring model for acute GIB risk based on five factors (no PPI use, chronic kidney disease, chronic obstructive pulmonary disease, history of peptic ulcer disease, and liver cirrhosis). The c-statistic of the new score (0.65) was superior to that of the HAS-BLED score (0.57) for predicting acute GIB[117,118]. The utility of these scoring systems for predicting re-bleeding, and the strategy of changing anticoagulants would be important topics of study.