Statistical analysis

The researchers used specialized RevMan 5.4 software (Cochrane Training) to crunch the numbers from the studies, with pooled risk ratio (RR) of 95% confidence intervals (Cis) applied to analyze the random- or fixed-effect models. The significant outcome of the two-sided statistical tests was determined with a p-value < 0.05. Heterogeneity was assessed using inconsistency index statistic (I²) to assess heterogeneity, and its level was reflected by the value of the I² statistic. The trials were deemed heterogeneous when I² > 50% and p < 0.05, and random-effects models were used. If the studies seemed low heterogeneity, the fixed-effects model was used. Publication bias analysis was performed minimally when the study contained more than 10 trials.