Clinical Outcomes and Management Associated With Major Bleeding in Patients With Atrial Fibrillation Treated With Apixaban or Warfarin
Insights From the Aristotle Trial
Claes Held; Elaine M. Hylek; John H. Alexander; Michael Hanna; Renato D. Lopes; Daniel M. Wojdyla; Laine Thomas; Hussein Al-Khalidi; Marco Alings; Dennis Xavier; Jack Ansell; Shinya Goto; Witold Ruzyllo; Ma°rten Rosenqvist; Freek W. A. Verheugt; Jun Zhu; Christopher B. Granger; Lars Wallentin
Eur Heart J. 2015;36(20):1264-1272.
Abstract and Introduction
Abstract
Aim In the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) trial, apixaban compared with warfarin reduced the risk of stroke, major bleed, and death in patients with atrial fibrillation. In this ancillary study, we evaluated clinical consequences of major bleeds, as well as management and treatment effects of warfarin vs. apixaban.
Methods and results Major International Society on Thrombosis and Haemostasis bleeding was defined as overt bleeding accompanied by a decrease in haemoglobin (Hb) of ≥2 g/dL or transfusion of ≥2 units of packed red cells, occurring at a critical site or resulting in death. Time to event [death, ischaemic stroke, or myocardial infarction (MI)] was evaluated by Cox regression models. The excess risk associated with bleeding was evaluated by separate time-dependent indicators for intracranial (ICH) and non-intracranial haemorrhage. Major bleeding occurred in 848 individuals (4.7%), of whom 126 (14.9%) died within 30 days. Of 176 patients with an ICH, 76 (43.2%) died, and of the 695 patients with major non-ICH, 64 (9.2%) died within 30 days of the bleeding. The risk of death, ischaemic stroke, or MI was increased roughly 12-fold after a major non-ICH bleeding event within 30 days. Corresponding risk of death following an ICH was markedly increased, with HR 121.5 (95% CI 91.3–161.8) as was stroke or MI with HR 21.95 (95% CI 9.88–48.81), respectively. Among patients with major bleeds, 20.8% received vitamin K and/or related medications (fresh frozen plasma, coagulation factors, factor VIIa) to stop bleeding within 3 days, and 37% received blood transfusion. There was no interaction between apixaban and warfarin and major bleeding on the risk of death, stroke, or MI.
Conclusion Major bleeding was associated with substantially increased risk of death, ischaemic stroke, or MI, especially following ICH, and this risk was similarly elevated regardless of treatment with apixaban or warfarin. These results underscore the importance of preventing bleeding in anti-coagulated patients.
Clinical perspective Major bleedings in patients with atrial fibrillation treated on anti-coagulation therapy was associated with substantially increased risk of death, ischaemic stroke, or MI, especially following intracranial haemorrhage. This risk was similarly elevated regardless of treatment with apixaban or warfarin. The results underscore the importance of preventing bleeding in anti-coagulated patients.
Introduction
Atrial fibrillation (AF) is an important risk factor for stroke. Anti-coagulation is highly effective for reducing thrombo-embolic complications. Bleeding complications during anti-coagulation are, however, common and associated with increased risk of subsequent death and thrombotic events. Warfarin is known to be associated with bleeding and medication related emergent hospitalizations that relate to a narrow therapeutic window and difficulties in adjusting the dose due to a variable dose response.[1,2] The direct acting oral anti-coagulants that have emerged in recent years as alternatives to warfarin are proven to reduce these events (similar to or more than warfarin) with a similar or lower risk of bleeding complications.[3–6] In the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) trial, apixaban compared with warfarin significantly reduced the risk of stroke, major bleeding, and death in patients with AF.[3] Bleeding among patients on anti-coagulant therapy often leads to cessation of anti-coagulant treatment, which may increase the risk of subsequent thrombotic events. Termination of warfarin after a gastrointestinal (GI) bleed was shown in one study to increase the risk of death and thrombotic events.[7] In a real-world registry, patients who experienced serious bleeding had a three-fold increase in the risk of thrombotic events even though the majority of patients did not permanently stop oral anti-coagulation.[8] We have in a recent report presented data on the location and characteristics of bleeding, and the 30-day mortality associated with the first major bleed in the ARISTOTLE trial.[9] In the present study, we evaluate the clinical consequences of major bleeds, including risk of death, ischaemic stroke, and myocardial infarction (MI) in the 30 days preceded by a major bleed. We also report the consequences of a major bleed in terms of management, such as therapeutic actions to stop the bleeding and to decrease subsequent bleeding risk, change in anti-thrombotic therapy, or study drug discontinuation and also treatment effects of apixaban vs. warfarin on clinical outcomes.
Methods
The ARISTOTLE trial design has been reported previously.[10] In summary, it was a double-blind, double-dummy, randomized clinical trial enrolling 18 201 patients with AF and at least 1 CHADS2 risk factor for stroke or systemic embolism. Patients were randomized to warfarin (n = 9081) or apixaban (n = 9120). The primary endpoint was stroke or systemic embolism. The primary safety outcome was bleeding according to the International Society on Thrombosis and Haemostasis (ISTH) criteria. The median length of follow-up was 1.8 years. Patients with AF and at least one risk factor for stroke were randomized to either dose-adjusted warfarin or apixaban 5 mg twice daily. A reduced dose of apixaban 2.5 mg twice daily was designated for participants with two or more of the following criteria: an age of at least 80 years, a body weight of no >60 kg, or a serum creatinine level of 1.5 mg/dL (133 μmol/L) or more. To enhance the quality of warfarin management, a dosing algorithm was provided and a program implemented to provide regular feedback to sites regarding their level of international normalized ratio (INR) control.
Bleeding Definitions
This study captured bleeding events occurring between randomization and efficacy censoring date, in all randomized patients. Major bleeding was defined according to the criteria of the ISTH as clinically overt bleeding accompanied by a decrease in the Hb level of at least 2 g/dL or transfusion of at least two units of packed red cells, occurring at a critical site (intra-cranial, intra-ocular, intra-spinal, intra-articular, intra-muscular with compartment syndrome, pericardial, retroperitoneal) or resulting in death. No time restrictions were applied to this definition but the Hb drop occurred most often within 24–48 h. Laboratory and transfusion data coupled with clinical event details were used to identify and adjudicate potential bleeding events. Routine collection of Hb occurred every 3 months. Location of bleeding was extracted from the case report form. Additional source documents were collected when necessary. The primary safety outcomes were adjudicated on the basis of pre-specified criteria by a clinical events committee whose members were not aware of study-group assignments.
We evaluated the consequences of recent major bleeding (within 30 days) on efficacy endpoints, including death and thrombotic events (stroke or MI). The 30-day window was selected in order to focus on efficacy events that were potentially attributable to circumstances surrounding the bleed, rather than underlying comorbidities which would dominate long-term differences. In addition, changes in anti-thrombotic therapy, study drug discontinuation, and transfusions after the bleeding were evaluated.
Statistical Analyses
Demographic and clinical characteristics were presented for patients with at least one ISTH major bleeding and patients without major bleedings. Continuous variables were summarized as medians and quartiles and categorical variables as frequencies and percentages.
In the full randomized population, multivariable Cox regression models were used to evaluate the increase in efficacy events within 30 days of ISTH major bleeding. Specifically, separate time-dependent indicators were included for non-intra-cranial and intra-cranial bleeding, taking a value of 1 during the 30-day window following a bleed and 0 otherwise. Therefore, we assessed the elevation in risk of events during the immediate 30 days after a bleed. In patients with multiple ISTH major bleedings, only the first event was evaluated. An interaction term with randomized treatment was added to evaluate whether the risk associated with recent bleeding differed in apixaban- vs. warfarin-treated patients. These models were adjusted for a pre-specified set of clinical and demographic factors including type of AF, age, sex, region, history of MI, congestive heart failure, stroke/TIA or systemic embolism, diabetes, time with AF, hypertension, history of bleeding, prior vitamin K antagonist (VKA) use, CHADS2 score, history of vascular disease, creatinine clearance, and baseline medications including aspirin, statins, ACE inhibitors/ARBs, and amiodarone. This analysis was repeated in sensitivity analyses, where ISTH major bleeding was replaced by other definitions of bleeding (i.e. any bleeding). Among patients who experienced major bleeding, information about change in anti-thrombotic therapy, study drug discontinuation and resumption, and transfusion were summarized as frequencies and percentages. Also in this subgroup, the cumulative incidence of all-cause death and thrombotic events after major bleeding events were presented as Kaplan–Meier curves. All analyses