“…We stated in our article that we tested heterogeneity between studies by I 2 statistic with I 2 ≥ 50% indicating heterogeneity, and if no significant heterogeneity existed, a fixed effects model was adopted, otherwise a random effects model was used. For this question, we recalculated the data (Table 1) with a random effects model and verified that the outcomes in our article were correct [1], so we do not doubt the statistical methods in our study.…”