Statistical Analysis
Data analysis was conducted with RevMan version 5.4.1 and Stata version
14.0 software. Dichotomous outcomes were reported as risk ratios (RRs)
with 95% confidence intervals (95% CIs). For continuous variables,
data were presented as mean differences (MDs) with 95% CIs.
Heterogeneity was investigated by the χ2 test andI2 statistics; if there was no significant
heterogeneity (P≥0.10, I2 ≤50%), a
fixed-effects model was used; if not, a random-effects
model
was used. Sensitivity analyses
were performed for outcomes with significant heterogeneity by excluding
one study at one time. Publication bias and sensitivity analysis were
evaluated via Stata version 14.0 software. We performed a subgroup
analysis based on the time of the first dose of antibiotic drugs.