Decoding definitions of major bleeding in retrospective observational studies on direct oral anticoagulants for atrial fibrillation: a targeted literature review

Abstract

Abstract Background Major bleeding is an adverse event associated with using direct oral anticoagulants (DOACs) for treatment of atrial fibrillation (AF). Incidence rates of major bleeding vary notably between randomised controlled trials (RCTs) and real-world evidence (RWE) studies. This may be attributable to inconsistent major bleeding definitions across RWE studies. RCTs commonly use the International Society on Thrombosis and Haemostasis (ISTH) major bleeding definition, including fatal bleeding, critical area/organ bleeding, bleeding requiring a blood transfusion of 2 or more units, and/or reductions in haemoglobin. These parameters may be absent in RWE studies, leading to utilisation of proxy definitions. Thus, the components used to define major bleeding in RWE studies remain unclear. Purpose To examine major bleeding definitions via a targeted literature review of retrospective observational studies on patients with AF receiving DOACs in routine clinical practice. Methods Studies of patients with AF receiving DOACs that were published between 2012 and 2022 were identified using keywords and MeSH terms in PubMed. Retrospective observational studies were screened for inclusion by title and abstract, and subsequently by full text. Articles were assessed for the types of databases and DOAC therapies studied. Components of major bleeding definitions were recorded and compared across database types. Results Of 328 articles screened, 18 retrospective observational studies were included in this analysis. The most common data sources were claims data (n = 10), followed by electronic medical records (n = 4), and registry data (n = 4). Across studies, major bleeding incidence rates varied from 0.59 to 11.4 events per 100 person-years. Of the 6 components of major bleeding definitions identified, requiring hospitalisation (n = 12) was the most frequent (Figure), though is not essential per the ISTH definition. No articles used fatal bleeding as a component to define major bleeding. Approximately half of all articles used the component involvement of critical organ or site (n = 7). Requiring blood transfusion was used in 6 articles and only 4 used reduction in haemoglobin or haematocrit levels (none of these involved claims data). Conclusions This targeted literature review of retrospective database studies of routine clinical practice decodes proxy use of ISTH major bleeding definitions. Claims data were the most common data source but lacked components describing fatal bleeding and haemoglobin levels. Further research is vital to assess the variation in reported major bleeding rates in RWE studies using proxy definitions in place of the ISTH definition.Figure