In reply
We appreciate the comments from Correia and Lopes. We acknowledge in the “Comment” section of our article1 that the Randomized Evaluation of Long Term Anticoagulant Therapy (RE-LY) Trial had a large impact on our study, comprising 59% of the cohort and 74% of the events. In all of the 7 studies, myocardial infarction (MI) and acute coronary syndrome (ACS) were secondary safety outcomes, and limitations in interpreting secondary outcomes are true across the studies. Given the scarcity of MI and ACS events and the imbalance between trial arms, performing a Mantel-Haenszel fixed-effects model for meta-analysis is a better method than a random-effects model.