Background and purpose: Several observational studies have compared the effect of the non-vitamin K antagonist oral anticoagulants to each other in patients with atrial fibrillation. However, confounding by indication is a major problem when comparing non-vitamin K antagonist oral anticoagulant treatments in some of these studies. This meta-analysis was conducted to compare the effectiveness and safety between non-vitamin K antagonist oral anticoagulant and non-vitamin K antagonist oral anticoagulant by only including the propensity score matching studies.
Methods: We systematically searched the PubMed and Ovid databases until May 2020 to identify relevant observational studies. Hazard ratios (HRs) and 95% CIs of the reported outcomes were collected and then pooled by a random-effects model complemented with an inverse variance heterogeneity or quality effects model.
Results: A total of 17 retrospective cohort studies were included in this meta-analysis. Compared with dabigatran use, the use of rivaroxaban was significantly associated with increased risks of stroke or systemic embolism (HR, 1.16 [95% CI, 1.05-1.29]) and major bleeding (HR, 1.32 [95% CI, 1.24-1.41]), whereas the use of apixaban was associated with a reduced risk of major bleeding (HR, 0.78 [95% CI, 0.67-0.90]) but not stroke or systemic embolism (HR, 0.84 [95% CI, 0.56-1.28]). Compared with rivaroxaban use, the use of apixaban was associated with a decreased risk of major bleeding (HR, 0.63 [95% CI, 0.54-0.73]) but not stroke or systemic embolism (HR, 0.83 [95% CI, 0.67-1.04]). Reanalyses with the inverse variance heterogeneity or quality effects model produced similar results as the random-effects model.
Conclusions: Current observational comparisons with propensity score matching methods suggest that apixaban might be a better choice compared with dabigatran or rivaroxaban for stroke prevention in atrial fibrillation patients.
Keywords: anticoagulants; apixaban; atrial fibrillation; propensity score; safety.