Introduction Rivaroxaban, apixaban and dabigatran compile the novel oral anticoagulants (NOAC). Since 2012, NOACs are used as first line treatment for atrial fibrillation, pulmonary embolism and deep venous thrombosis at Russells Hall Hospital (RHH). The trial data regarding the risk of gastrointestinal (GI) bleeding has been conflicting, although a meta-analysis found an increased risk of GI bleeding with NOACs as opposed to warfarin (2.3% vs. 1.3%).1However, little is known about its incidence in the real world setting. We aimed to demonstrate the incidence of GI bleeding, endoscopic findings, time of onset of bleeding and the need for intervention in patients taking NOACs.
Method We performed a retrospective review of all patients at RHH who received NOACs. These patients were identified from the anticoagulation database and cross-referenced with the GI database and patient notes. Basic demographic, clinical and laboratory data and endoscopic findings were collated.
Results A total of 2487 patients were identified to be on NOACs of which 61 were found to have an episode of GI bleeding, giving an incidence rate of 2.45%. 54 of 2334 patients were on rivaroxaban, 3 of 77 on apixaban and 4 of 76 on dabigatran with an incidence of GI bleeding of 2.3%, 3.9% and 5.3% respectively. There was a greater incidence found with females than males (55% vs. 45%) with the median age 80 years (range 53–95). The 30-day mortality rate was 11.5% (n = 11). The cause of death were not directly due to GI bleeding.
Of these 61 patients, 30 (49.2%) had an upper GI bleed (UGIB) and 31 (50.8%) had a lower GI bleed (LGIB). The median time from starting NOAC to having a GI bleed was 6.7 months (range 2 days-19.7 months). Endoscopy results were varied with the most common findings being haemorrhoids (15.7%), diverticular disease (14.6%), normal findings (12.4%), gastritis (7.9%) and peptic ulcer disease (7.9%). 7 patients (11.5%) required endoscopic intervention and 21 patients (34.4%) required the use of blood products. Although 12.9% of patients with LGIB required endoscopic intervention versus 10% of UGIB, patients with UGIB were more likely to be inpatients (97% vs. 35%), with a greater drop in haemoglobin (median Hb: 87 g/L versus 105 g/L) and the increased use of blood products (53% vs. 16%).
Conclusion The risk of GI bleeding with NOACs is greater than warfarin, with rivaroxaban being the safest of the three drugs in our cohort and is similar to the published data. Further studies are required to confirm these findings.
Disclosure of interest None Declared.
Capodanno D, Capranzano P, Giacchi G, Calvi V, Tamburino C. Novel oral anticoagulants versus warfarin in non-valvular atrial fibrillation: a meta-analysis of 50,578 patients. Int J Cardiol. 2013;167(4):1237–41