There were 9 studies in total (3,056 infants): 2 RCTs (310 infants), 2 prospective cohort studies (547 infants) and 5 retrospective cohort studies (2,199 infants).
No pooled risk difference for any single outcome was significantly different from 0%, indicating no difference between the trial of labour and no trial of labour groups.
When all the study types were combined, the pooled risk difference between the trial of labour and no trial of labour groups for any injury (intracranial bleeding, brachial plexus injury, clavicular fracture, facial nerve injury) was 0.89% (95% confidence interval, CI: 0.02, 1.75), and for any injury or death was 1.10% (95% CI: 0.27, 1.93).
Restricting attention to the RCTs resulted in non significant differences for both any injury and any injury or death.
When the study identified as being an outlier was excluded from the analysis, the risk difference was smaller for both the risk of any injury, and any injury or death, but was still statistically significant (0.41%, 95% CI: 0.11, 0.72; and 0.76%, 95% CI: 0.35, 1.18, respectively).