(AL) Real World Effects of Oral Anticoagulants in AF: A Cohort Study with Validation Against Results from a Randomised Controlled Trial

Publication Number: 328

Background: Direct oral anticoagulants (DOACs) have been widely adopted as an alternative to warfarin in prevention of stroke in patients with atrial fibrillation (AF). Treatment guidelines for AF are based on results of randomised controlled trials (RCT) for DOACs. There is uncertainty on the applicability of RCT results to patient groups excluded from the trials.

Objectives: To use eligibility criteria and baseline patient characteristics from ARISTOTLE, the pivotal RCT for the DOAC apixaban, to construct a trial-analogous cohort of patients in the UK Clinical Practice Research Datalink (CPRD). To compare results in this trial-analogous cohort with the RCT to validate non-interventional methods for assessing treatment effectiveness in stroke prevention in AF.

Methods: ARISTOTLE eligibility criteria were applied to AF patients in CPRD Aurum prescribed apixaban or warfarin between 1 Jan 2013 to 31 Jan 2020 (418,705 people; 144,495 HES-linked and trial-eligible). Apixaban patients were matched to the RCT apixaban arm on key baseline characteristics creating a trial-analogous apixaban arm; this arm was then matched 1:1 to warfarin CPRD patients using propensity score matching yielding a trial-analogous cohort of 8603 apixaban and 8603 warfarin patients. A Cox proportional hazards model was used to analyse the ARISTOTLE efficacy outcomes of (1) stroke or systemic embolism and (2) death from any cause and results compared with the trial. Time in therapeutic range (TTR) was calculated using Rosendaal’s method and treatment effectiveness further evaluated based on TTR.

Results: Results in the CPRD trial-analogous cohort [HR (95%CI)] were consistent with the trial results in confirming non-inferiority of apixaban compared with warfarin for stroke or systemic embolism [CPRD 0.90 (0.75-1.09) vs trial 0.79 (0.66-0.95)] and death from any cause [CPRD 0.77 (0.70-0.86) vs trial 0.89 (0.80-0.99)].
TTR was higher in the CPRD cohort than in the trial (mean 71.9% vs 62.2%); analysis by TTR subgroup suggested superiority of apixaban vs warfarin in those with TTR < 75% [Stroke/SE 0.67 (0.50-0.90), all-cause death 0.62 (0.53-0.72)] whereas apixaban was associated with increased hazards of both outcomes in those with well-controlled warfarin therapy (TTR ≥ 75%) [Stroke/SE 1.62 (1.19-2.19), all-cause death 1.55 (1.30-1.84)].

Conclusions: Analysis of this non-interventional data generated results comparable to ARISTOTLE. The weaker treatment benefit observed overall in our cohort may be linked to the higher proportion of patients with well-controlled warfarin compared with the trial. The methodological template developed can be used to determine risks and benefits of AF treatments in people in CPRD who would have been excluded from ARISTOTLE.