Use of Gastroprotective Agents and Risk of Dabigatran Associated Gastrointestinal Bleeding: A Population-Based Retrospective Cohort Study

Free Paper Session: Risk factors and management of upper GI bleedingDabigatran, a direct thrombin inhibitor, is the first new oral anticoagulant available as an alternative to warfarin. Despite its convenience and superiority over warfarin in the prevention of stroke and thromboembolism, recent studies suggested an increase risk of gastrointestinal bleeding (GIB) in patients treated with dabigatran when compared to warfarin. These data were however largely derived from clinical trials in selected patient population. AIMS & METHODS: This study determined the risk of dabigatran associated GIB and the role of gastroprotective agents, including proton pump inhibitors (PPIs) and H2-receptor antagonists (H2RAs), in preventing dabigatran related GIB in a population-based retrospective cohort study. Data were extracted from the central database of the Hong Kong Hospital Authority, which is the provider of all public medical services to 7 million populations. We identified all patients who were newly prescribed with dabigatran between Jan 2010 and Dec 2013. The primary endpoint is the onset of clinical GIB. Multivariate analysis was used to characterize the risk of GIB after adjusting for baseline patient’s characteristics, medical illnesses and concurrent medications. RESULTS: 5,041 patients, who were newly prescribed dabigatran, were included in the analysis. Among them, 222 (4.4%) patients developed GIB with a median time to bleeding of 97 (IQR 262) days. Patients who were aged ≥75 years (OR 1.83; 95% CI, 1.36 to 2.47), had prior ischemic stroke, transient ischemic attack or systemic embolic events (OR 1.62; 1.19 to 2.2), and a prior history of peptic ulcer or GIB (OR 2.48; 1.81 to 3.39) were found to have higher risks of GIB. Concurrent use of gastroprotective agents (OR 0.61; 0.44-0.84; log rank test P = 0.018) or statin (OR: 0.58; 0.43-0.78) reduced the likelihood of GIB. Subcategory analysis showed that the use of either PPIs (OR 0.70; 0.51-0.98) or H2RAs (OR 0.67; 0.50-0.90) significantly lowered the bleeding risk. The risk reduction by gastroprotective agents was significant only in patients with prior history of ulcers or GIB (OR 0.24; 0.14 to 0.43) but not in patients with no prior history (OR 0.83; 0.56 to 1.21). CONCLUSION: The risk of GIB associated with dabigatran use in real life clinical settings is 4.4%. The use of gastroprotective agents significantly reduced the risk of dabigatran related GIB, particularly in high-risk patients with prior history of peptic ulcer or GIB