Therapeutic Advances in Neurological Disorders (Sep 2016)

Novel oral anticoagulants for the secondary prevention of cerebral ischemia: a network meta-analysis

  • Aristeidis H. Katsanos,
  • Dimitris Mavridis,
  • John Parissis,
  • Spyridon Deftereos,
  • Alexandra Frogoudaki,
  • Agathi-Rosa Vrettou,
  • Ignatios Ikonomidis,
  • Maria Chondrogianni,
  • Apostolos Safouris,
  • Angeliki Filippatou,
  • Konstantinos Voumvourakis,
  • Nikos Triantafyllou,
  • John Ellul,
  • Theodore Karapanayiotides,
  • Sotirios Giannopoulos,
  • Anne W. Alexandrov,
  • Andrei V. Alexandrov,
  • Georgios Tsivgoulis

DOI
https://doi.org/10.1177/1756285616659411
Journal volume & issue
Vol. 9

Abstract

Read online

Background: Novel oral anticoagulants (NOACs) have shown to be both safe and effective for ischemic stroke prevention in patients with nonvalvular atrial fibrillation (NVAF). We conducted a network meta-analysis (NMA) using published data from secondary prevention subgroups of different phase III randomized clinical trials (RCTs) comparing individual NOACs with warfarin. Methods: Eligible studies were identified by searching MEDLINE and SCOPUS and the Cochrane Central Register of Controlled Trials databases. First, we conducted a pairwise meta-analysis for each pairwise comparison, and then we performed NMA to combine direct and indirect evidence for any given pair of treatments. The comparative effects of all NOACs against warfarin were ranked with the surface under the cumulative ranking (SUCRA) curve for each outcome. Results: We identified four RCTs (including 15,240 patients) comparing individual NOACs (apixaban, dabigatran, rivaroxaban) with warfarin. Using indirect evidence, dabigatran was related to a significantly lower risk of hemorrhagic stroke compared with rivaroxaban [risk ratio (RR) 0.28; 95% confidence interval (CI) 0.11–0.75], while rivaroxaban was associated with a significantly lower risk of major gastrointestinal bleeding compared with dabigatran (RR 0.14; 95% CI 0.03–0.74). We also performed clustered ranking plot for the primary efficacy and safety endpoints to identify the treatment with the probably best benefit-to-risk ratio profile. Conclusions: The three NOACs showed differences in terms of safety and efficacy for secondary stroke prevention in NVAF. Our findings can serve only as hypothesis generation and require independent confirmation in head-to-head RCTs, owing to the sparse available evidence and increased uncertainty in both indirect effect estimates and ranking of treatments.