Six cohort studies were included (637 women who had versions plus 36,556 controls). In three studies the controls were matched to cases.
Compared with cephalic-presenting pregnancies, successful version was associated with a significantly increased risk for all-cause Caesarean section (27.6% versus 12.5%; RR 2.04, 95% CI: 1.43, 2.91), Caesarean section for dystocia (RR 2.19, 95% CI: 1.38, 3.48) and Caesarean section for foetal distress (RR 2.10, 95% CI: 1.18, 3.75).
The increased risk of Caesarean section was statistically significant for studies with and without matched controls (the data were not presented).
The funnel plot showed no evidence for publication bias (P=0.263).
Statistically significant heterogeneity was found and attributed to one study with a high rate of Caesarean section in the control group. After excluding this study, the increased risk of Caesarean section was still statistically significant for all-cause Caesarean section (RR 2.62, 95% CI: 2.27, 3.03), Caesarean section for dystocia (RR 2.80, 95% CI: 2.20, 3.56) and Caesarean section for foetal distress (RR 2.80, 95% CI: 2.01, 3.90). Heterogeneity was no longer statistically significant for any of the meta-analyses (P>0.32).