Data synthesis
We performed statistical analyses using Review Manager 5 Software
(version 5.3) and Stata (version 15.1). For dichotomous data, we
calculated risk ratio (RR) with 95% confidence interval (CI) by using
the Mantel-Haenszel approach. For continuous outcomes, we calculated
mean differences (MD) with 95% confidence intervals using an Inverse
Variance method. A P value of < 0.05 was considered
statistically significant. Summary effects estimates were calculated
using DerSimonian and Laird random-effects models.16We performed subgroup analysis on the basis of lubricants type, parity,
overall risk of bias, birthweight and ethnicities when sufficient data
for each subgroup. An interaction analysis (P for interaction)
was performed to evaluate the difference between subgroups. Statistical
heterogeneity was evaluated using the I2 statistic and
Cochran Chi-square test. Publication bias was assessed visually by
funnel plots and quantitatively with Begg’s test for
asymmetry.17 Sensitivity analysis was also undertaken
to examine the effect of a single study on the overall effect size and
the robustness of results with omitting one study at a time.