A systematic review and Bayesian network meta-analysis of risk of intracranial hemorrhage with direct oral anticoagulants

: Essentials Risk of intracranial hemorrhage (ICH) may differ between direct oral anticoagulants (DOACs). We compared the risk of ICH between DOACs using network meta-analysis. Dabigatran 110 mg and 150 mg were safer than rivaroxaban on Bayesian analysis. Dabigatran 110 mg ranked as the safest DOAC while rivaroxaban ranked last. SUMMARY: Background The comparative risk of intracranial hemorrhage (ICH) among direct oral anticoagulants (DOACs) (dabigatran, rivaroxaban, apixaban and edoxaban) remains unclear. Objective To determine the difference in risk of ICH between DOACs Methods Seventeen randomized controlled trials (RCTs) were selected using PubMed/MEDLINE, EMBASE and CENTRAL (Inception, 31 December 2017). Estimates were reported as odds ratio (OR) with 95% credible interval (CR.I) in Bayesian network meta-analysis (NMA), and OR with 95% confidence interval (CI) in traditional meta-analyses. Relative ranking probability of each group was generated based on surface under the cumulative ranking curve (SUCRA). Results In NMA of 116 618 patients from 17 RCTs (apixaban = 19 495 patients, rivaroxaban = 14 157 patients, dabigatran = 16 074 patients, edoxaban = 11 652 patients, and comparator = 55 315 patients), all DOACs were safer than warfarin for risk of ICH. Dabigatran 110 mg ranked as the safest drug (SUCRA, 0.85) and reduced the risk of ICH by 56% compared to rivaroxaban (OR, 0.44; 95% Cr.I, 0.22-0.82). Pairwise meta-analysis validated these findings, showing that DOACs were safer than warfarin (OR, 0.46; 95% CI, 0.35-0.59). Subgroup analysis showed that the benefit was present when DOACs were used in non-valvular atrial fibrillation (NVAF) (OR, 0.51; 95% CI, 0.38-0.68) or venous thromboembolism (VTE) (OR, 0.32; 95% CI, 0.18-0.58). Conclusion Dabigatran 110 mg may be the safest choice among any anticoagulant regarding risk of ICH. Both dabigatran 110 mg and 150 mg were safer than rivaroxaban.