Five RCTs (2,643 patients) and six retrospective cohort studies (3,110 patients) were included in the review. Risk of bias was high for one RCT, low for one RCT and unclear for three RCTs. All cohort studies were judged as high quality (quality scores ranged from 7 to 9).
Compared with no treatment, treatment was significantly associated with fewer cases of pre-eclampsia (RR 0.62, 95% CI 0.43 to 0.89; three RCTs), shoulder dystocia (RR 0.42, 95% CI 0.23 to 0.77; three RCTs), and macrosomia (birthweight above 4,000g) (RR 0.50, 95% CI 0.35 to 0.71; five RCTs). Significant heterogeneity was observed only for the outcome of macrosomia (Ι²=50%) and not for the other two outcomes.
Two RCTs reported that treatment was significantly associated with less weight gain compared with no treatment; there was no difference for this outcome between the two groups in other two RCTs.
Two RCTs showed no difference in birth injury between the treatment and no treatment groups but one cohort study showed fewer cases with treatment than no treatment.
No difference between the treatment and no treatment groups was found for neonatal hypoglycaemia, cesarean delivery and admission to a neonatal intensive care unit and small-for-gestational-age neonates. There was insufficient evidence on long-term metabolic outcomes among infants.
There was no evidence from included studies for long-term maternal outcomes including type 2 diabetes mellitus, obesity and hypertension.