Lead Investigator: Chris Gale, University of Leeds
Title of Research Proposal: Efficacy and safety of edoxaban versus warfarin in patients with atrial fibrillation and frailty insights from the ENGAGE AF-TIMI 48 trial.
Vivli Data Request: 4540
Funding Source: Employment Contracts – Analysis will be performed by researchers funded by the University of Leeds
Potential Conflicts of Interest: None
Summary of the Proposed Research:
Atrial fibrillation (AF) is common in older people, and is associated with an increased risk of both mortality and stroke. Frailty is a condition in which there is a decline in biological reserves and deterioration in physiological mechanisms, which render the person vulnerable to a range of additional adverse outcomes. Frailty provides an insight into biological age and is more useful than chronological age in predicting adverse events including death. Frailty therefore provides a clinically useful framework in approaching individualised care planning, including the relative risks and benefits associated with a clinical intervention.
Patients with AF commonly have concomitant frailty. However, there is a lack of evidence to guide the management of AF in people who also have frailty.(1) Whilst there is evidence to support the use of Apixaban for stroke prophylaxis in AF for patients with multi-morbidity, there are currently no studies that use a validated method of defining frailty in order to report clinical outcomes associated with direct oral anticoagulation for patients with AF by frailty category. It is therefore unknown whether the risks and benefits associated with anticoagulation differ by frailty category. We would like to investigate this question using data from the ENGAGE AF-TIMI 48 trial.
Statistical Analysis Plan:
Key inclusion criteria:
These will match the original trial. A sensitivity analysis will be performed excluding patients aged under 65 years of age.
Patients will be included if they were in the original study, aged 21 years or older, and had AF documented by means of an electrical tracing within the 12 months preceding randomization, a score of 2 or higher on the CHADS2 risk assessment, and anticoagulation therapy planned for the duration of the trial.
Key exclusion criteria:
These will match the original trial, and include AF due to a reversible disorder; an estimated creatinine clearance of less than 30 ml per minute; a high risk of bleeding; use of dual antiplatelet therapy; moderate-to-severe mitral stenosis; other indications for anticoagulation therapy; acute coronary syndromes, coronary revascularization, or stroke within 30 days before randomization; and an inability to adhere to study procedures.
Baseline characteristics for the cohort will be reported by frailty category and treatment arm. Kaplan-Meier survival curves for primary and secondary clinical outcomes will be constructed and compared between frailty category using the Log Rank test. A Cox proportional hazards model will be used to assess the association between frailty groups and clinical outcomes if the proportional hazard assumption is satisfied. The association between clinical outcomes and frailty category for each treatment arm will be tested in unadjusted models, and minimally adjusted models (adjusted for sex, age, race and region), in order to preserve the association between frailty and outcome.
The reported efficacy and safety analyses of the ENGAGE AF-TIMI 48 trial will be replicated, using data from patients who underwent randomization and received at least one dose of the study drug during the treatment period (modified intention-to-treat population). A Cox proportional hazard model will be used to assess whether the efficacy and safety between treatment arms are similar across frailty category. This will be achieved by including treatment arms, frailty index and their interaction in the Cox model, further adjusted for the two randomisation stratification factors used in the original trial. If an interaction between treatment arm and frailty index is found to be significant, subgroup analysis will be used quantify the difference.
The extent and pattern of missingness will be assessed for variables with missing data. It would be unusual to observe a high proportion of missingness in clinical trial data. To minimize the bias caused by a small proportion of missing data, multiple imputation with chained equations will be used to produce 5 imputed datasets. Pooled modelling estimates and accompanying 95% confidence interval will be generated according to Rubin’s Rules. Cox proportional hazards models will be applied to multiply imputed datasets, and the same analyses will be performed on the complete dataset as a sensitivity analysis to assess the influence of the missing data.
A Phase 3, Randomized, Double-Blind, Double-Dummy, Parallel Group, Multi-Center, Multi-National Study for Evaluation of Efficacy and Safety of Edoxaban (DU-176b) Versus Warfarin In Subjects With Atrial Fibrillation – Effective Anticoagulation With Factor Xa Next Generation in Atrial Fibrillation (ENGAGE – AF TIMI – 48)
Sponsor: Daiichi Sankyo, Inc.
Study ID: NCT00781391
Sponsor ID: DU176b-C-U301
Wilkinson, C., Wu, J., Searle, S.D. et al. Clinical outcomes in patients with atrial fibrillation and frailty: insights from the ENGAGE AF-TIMI 48 trial. BMC Med 18, 401 (2020).