Management of severe bleeding in patients treated with direct oral anticoagulants: an observational registry analysis

The currently available guidelines are based off a limited evidence pool, and are therefore influenced by clinician preference. The investigators aimed to obtain more data on the current management strategies of patients taking NOACs who present with significant bleeding, describing outcomes up to day 30 after presentation. Their secondary objectives were to describe the sites, types and severity of bleeding events, and the clinical and biological characteristics of the patients.

To that end, they analysed a large, prospectively-maintained, multicentre registry covering 36 centres in France and Belgium. The registry enrols patients treated with NOACs; for this report, only patients being treated for atrial fibrillation, or deep vein thrombosis that was not a result of surgery were included. Demographic data were collected, alongside clinical, laboratory and treatment details. The bleeding management was assessed by collecting the rates and numbers of infusions of red blood cell, fresh frozen plasma, prothrombin complex concentrate (PCC), factor VIII inhibitor bypass activator, recombinant activated factor VII concentrate, fibrinogen concentrate or tranexamic acid. The clinical outcomes were assessed 30 days after admission, bleeding events were divided into major (fatal, in a critical location, with a haemoglobin drop ≥20 g/dL, or requiring at least a 2-unit transfusion), or non-major clinically relevant (requiring medical intervention, hospitalisation, or if in hospital already, escalation of care). The patients were also assessed for suspected major cerebral and cardiovascular events (MACCE).

A total of 732 patients were treated with NOACs for significant bleeding across the 36 units; this included 535 from emergency departments, 156 from intensive care departments, and 39 from medical or surgical wards. The most common NOAC used was rivaroxaban, accounting for 64% of patients, followed by dabigatran (28%) and apixaban (7%). Most patients were elderly, with 42% over 80 years; 69% had moderate or severe renal dysfunction. Bleeding events were predominantly major (74%), and most commonly spontaneous (79%) rather than trauma-related. In terms of management, 94% of cases had coagulation tests performed, while serum anticoagulant levels were measured in 62% of cases. In patients with plasma concentrations less than 50 ng/mL the bleeding was most commonly intracerebral haemorrhage (42%), followed by gastrointestinal (GI) bleeding (27%) with the remainder being other sites. Treatment was more frequently with PCC when bleeding occurred in a critical organ versus other sites (57% of patients with intracerebral or spinal bleeding received PCC versus 28% with bleeding at other sites). Higher plasma concentrations of NOAC led to higher likelihood of PCC administration. The adequacy of haemostasis achieved was not measured objectively, but was assessed by local investigators: 44% rated haemostasis as totally adequate, 37% partially adequate, and 19% not adequate.

After 30 days, MACCEs had occurred in 7.4% of cases, there was an overall case-related mortality of 13.5% (95% CI 11.0–16.2%). Cause of death was related to central nervous system bleeding in 42% of cases. GI bleeding events were mostly in patients with high CHA2DS2Vasc scores (reflecting high stroke risk, and suggesting anticoagulation), the 30-day mortality was 11.9%.

This study aimed to describe how bleeding in patients on NOACs is currently being managed, and to describe characteristics of the population affected. The authors note that the population was largely elderly, and that renal impairment was common – they suggest that this may be due to either a slow deterioration of renal function, an acute kidney injury associated with the bleeding, or off-label use of NOACs in patients with poor renal function. When this report was compiled, no specific reversal agents were available, and reversal strategies were largely reliant on non-specific agents. Plasma concentration of NOACs has yet to be related to severity of bleeding events, although there was evidence that more PCC was used when concentrations were higher. The analysis of mortality did not comprise enough patients to enable meaningful analysis, although the data were consistent with the larger pivotal studies.

This report provides a useful picture of the current state of practice in a European setting, although its usefulness in comparing different NOAC-reversal strategies during severe bleeding episodes is limited.