Effect of Apixaban on All-Cause Death in Patients with Atrial Fibrillation: a Meta-Analysis Based on Imputed Placebo Effect.

Journal Article (Journal Article)

PURPOSE: Vitamin K antagonists (VKAs) are the standard of care for stroke prevention in patients with atrial fibrillation (AF); therefore, there is not equipoise when comparing newer oral anticoagulants with placebo in this setting. METHODS: To explore the effect of apixaban on mortality in patients with AF, we performed a meta-analysis of apixaban versus placebo using a putative placebo analysis based on randomized controlled clinical trials that compared warfarin, aspirin, and no antithrombotic control. We used data from two prospective randomized controlled trials for our comparison of apixaban versus warfarin (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) and apixaban versus aspirin (Apixaban Versus Acetylsalicylic Acid to Prevent Stroke in Atrial Fibrillation Patients Who Have Failed or Are Unsuitable for Vitamin K Antagonist Treatment). Using meta-analysis approaches, we indirectly compared apixaban with an imputed placebo with respect to the risk of death in patients with AF. We used results from meta-analyses of randomized trials as our reference for the comparison between warfarin and placebo/no treatment, and aspirin and placebo/no treatment. RESULTS: In these meta-analyses, a lower rate of death was seen both with warfarin (odds ratio [OR] 0.74, 95% confidence interval [CI] 0.57-0.97) and aspirin (OR 0.86, 95% CI 0.69-1.07) versus placebo/no treatment. Using data from ARISTOTLE and AVERROES, apixaban reduced the risk of death by 34% (95% CI 12-50%; p = 0.004) and 33% (95% CI 6-52%; p = 0.02), respectively, when compared with an imputed placebo. The pooled reduction in all-cause death with apixaban compared with an imputed placebo was 34% (95% CI 18-47%; p = 0.0002). CONCLUSIONS: In patients with AF, indirect comparisons suggest that apixaban reduces all-cause death by approximately one third compared with an imputed placebo.

Full Text

Duke Authors

Cited Authors

  • Guimarães, PO; Lopes, RD; Wojdyla, DM; Abdul-Rahim, AH; Connolly, SJ; Flaker, GC; Wang, J; Hanna, M; Granger, CB; Wallentin, L; Lees, KR; Alexander, JH; McMurray, JJV

Published Date

  • June 2017

Published In

Volume / Issue

  • 31 / 3

Start / End Page

  • 295 - 301

PubMed ID

  • 28516318

Electronic International Standard Serial Number (EISSN)

  • 1573-7241

Digital Object Identifier (DOI)

  • 10.1007/s10557-017-6728-z


  • eng

Conference Location

  • United States