P1256Resumption of anticoagulant treatment in patients with atrial fibrillation following gastrointestinal bleeding: a nationwide cohort study

Abstract

Abstract Introduction Following the first gastrointestinal bleeding in a patient with atrial fibrillation (AF) on oral anticoagulant treatment, apprehension of a reoccurrence may prevent clinicians from resuming anticoagulant treatment. However, risk of stroke and thromboembolism in these patients remains substantial. Limited literature exists on resumption patterns of vitamin K antagonists (VKA) and non-VKA oral anticoagulants (NOACs), as well as risks of outcomes associated with treatment compared to discontinuation. Purpose We investigated time trends of resumption of NOACs and VKA, and second, risks of recurrent gastrointestinal bleeding, stroke/thromboembolism and all-cause mortality associated with resumption of NOACs and VKA compared with non-resumption. Methods This study was based on a nationwide cohort study using Danish registries, from 1st January 2005 to 31st July 2017. All patients diagnosed with AF and receiving oral anticoagulants prior to admission to hospital with gastrointestinal bleeding were included. Follow-up began 90 days post discharge and patients were categorised in groups of non-resumption, resumption of NOAC, or resumption of VKA. Trends from 2005 to 2017 of resumption of oral anticoagulants were tested using Cochran-Armitage trend tests. Outcomes of interest included recurrent gastrointestinal bleeding, stroke/thromboembolism, and all-cause mortality, measured in terms of multiple-adjusted hazard ratios (HR) and standardised absolute risks. Results The study included 4842 patients with AF receiving oral anticoagulants, discharged from hospital after a gastrointestinal bleeding. Time trends showed a decline in non-resumption from 35.2% in 2006 to 16.5% in 2016. Within 90 days following discharge, 16.3% resumed NOACs, 57,9% resumed VKA, and 25,8% did not resume either anticoagulant treatment. Compared with non-resumption, resumption of was associated with lower hazard of all-cause mortality [HR 0.63 (0.54–0.75) and 0.68 (0.56–0.83), respectively] and lower, but not significant, hazard of stroke/thromboembolism [HR of 0.79 (0.52–1.19) and 0.78 (0.48–1.28), respectively]. There is similar hazard of recurrent gastrointestinal bleeding [HR 0.98 (0.75–1.28) and 0.82 (0.59–1.13), respectively]. The figure shows the standardised absolute risk of recurrent gastrointestinal bleeding and stroke/thromboembolism. Standardised Absolute Risks Conclusion(s) The number of patients not resuming anticoagulant treatment has approximately halved over 10 years between 2006 and 2016. Resumption of NOACs and VKA subsequent to gastrointestinal bleeding in patients with AF was associated with lower hazard of all-cause mortality. There is lower, but not significant, hazard of stroke/thromboembolism and similar hazard of recurrent gastrointestinal bleeding with resumption of VKA and NOACs compared to non-resumption, which suggests that apprehension of recurrent gastrointestinal bleeding is insufficient to justify discontinuation of treatment.