The incidence of death and risk factors for mortality following major bleeding in patients anticoagulated with factor Xa inhibitors: an observational study in the UK clinical setting (AXIOM UK)

Abstract

Abstract Background Major bleeding (MB) in association with anticoagulant use results in significant mortality. Mortality can be directly attributed to bleeding, or may follow bleeding complications such acute kidney injury, myocardial infarction, stroke and other cardiovascular events. Most real-world data relating to bleeding mortality is derived from cohorts receiving vitamin-K antagonists. We investigated all-cause mortality (ACM) after MB in a large cohort anticoagulated with Factor Xa inhibitors (FXai), e.g. direct oral anticoagulants (DOAC) rivaroxaban, apixaban and low-molecular-weight heparins (LMWH). Purpose To describe incidence rates of and risk factors for ACM following MB in patients receiving FXai. To investigate MB related ACM by bleeding site, treatment indication, treatment subgroups and individual and cumulative time periods. Methods Retrospective cohort analysis of new users of FXai, between 2012-2020, for therapeutic indications including atrial fibrillation (AF), venous thromboembolism, bioprosthetic cardiac valves or adult congenital heart disease was conducted. MB was defined by the International Society on Thrombosis and Haemostasis (ISTH) criteria. Incidence rates were reported per 100 person-years. Sub-distribution hazard ratios with 95% confidence intervals (95% CI) were estimated from Fine and Gray regression models accounting for competing risks. Results In 240,357 new users of FXai, 88% used an oral FXai. The most common FXai indication was AF (72%). There were 10,163 patients (4.2%) hospitalised for bleeding events during FXai use. Of these, 3,873 MB events (38%) fulfilled the ISTH definition of MB and 1312 died during 3,542 person years of follow-up (Table 1). ACM incidence rates varied by time, indication, bleeding subtype, and drug type. ACM occurred in 926 patients (23.9%) in the first month. Mortality rates were highest in the first month 607 (95% CI 568-647) per 100 person years, the rates then decreased over the year but remained elevated throughout the first year compared with subsequent years. Those receiving LMWH had higher ACM than those receiving a DOAC, but this was not significant in the adjusted risk factor analysis. Risk factors for ACM following bleeding included age ≥80, and a history of current smoking, myocardial infarction, congestive heart failure, peripheral artery disease, Stage 4 or 5 kidney disease, metastatic cancer, and intracranial haemorrhage. Age<60, bleeding following major trauma, and a history of anaemia were associated with lower risk of ACM (Table 2). Conclusions There is a significant incidence of ACM following MB among FXai users. The incidence rate of ACM in patients treated with FXai was highest in the first month after FXai initiation, declined during the first year and then plateaued. Multiple risk factors associated with ACM have been recognised and these could guide urgent therapies such as antidotes.