Maternal obesity is associated with high birthweight, obesity and premature mortality in adult offspring, probably as a result of maternal hyperglycaemia and insulin resistance. We present the results of a trial designed to test the hypothesis that metformin will improve insulin sensitivity in obese pregnant women, thereby reducing the incidence of high-birthweight babies.
To determine the efficacy of metformin (up to 2500 mg daily) given to obese pregnant women in reducing the gestational age-, parity- and sex-adjusted birthweight centile of the baby.
Double-blind, placebo-controlled, randomised controlled trial with embedded substudies.
Fifteen NHS hospitals in the UK.
Pregnant women aged ≥ 16 years with a singleton fetus and a body mass index of ≥ 30 kg/m2.
Metformin tablets (or placebo) administered between 12 and 16 weeks’ gestation until delivery of the baby.
Main outcome measures
The primary outcome measure was z-score corresponding to the gestational age-, parity- and sex-adjusted birthweight centile of live-born babies delivered at ≥ 24 weeks’ gestation. The main secondary outcome was maternal insulin resistance at 36 weeks’ gestation. Embedded substudies were included to assess the effect of metformin on insulin sensitivity using the hyperinsulinaemic–euglycaemic clamp; endothelial function; maternal and fetal fat distribution using magnetic resonance imaging; placental expression of 11β-hydroxysteroid dehydrogenase types 1 and 2 and glucocorticoid receptor; and myometrial contractility and glycogen storage.
We randomised 449 women to either placebo (n = 223) or metformin (n = 226), of whom 434 were included in the final intention-to-treat analysis. Mean birthweight at delivery was 3463 g [standard deviation (SD) 660 g] in the placebo group and 3462 g (SD 548 g) in the metformin group. The estimated effect size of metformin on the primary outcome was non-significant [adjusted mean difference in z-score –0.029, 95% confidence interval (CI) –0.217 to 0.158; p = 0.7597]. There was no evidence of a reduction in the main secondary outcome of homeostatic model assessment – insulin resistance (HOMA-IR) at 36 weeks’ gestation (mean HOMA-IR 5.98 and 6.30 molar units in the placebo and metformin groups, respectively; adjusted mean ratio 0.974, 95% CI 0.865 to 1.097). Metformin had no effect on the combined adverse outcome of miscarriage, termination of pregnancy, stillbirth or neonatal death. Subjects taking metformin demonstrated increased insulin sensitivity [glucose disposal per unit plasma insulin difference between means during high-dose insulin 0.02 mg/kg, 95% CI 0.001 to 0.03 mg/kg (fat-free mass)/minute/µIU/l; p = 0.04] compared with those taking placebo and enhanced endogenous glucose production [difference between means 0.54 mg/kg, 95% CI 0.08 to 1.00 mg/kg (fat-free mass)/minute; p = 0.02]. There were no differences in endothelial function, maternal or fetal body fat distribution, placental expression of 11β-hydroxysteroid dehydrogenase types 1 and 2 and glucocorticoid receptor, or myometrial contractility and glycogen storage.
Metformin has no clinically significant effect on birthweight centile in obese pregnant women. Follow-up studies of the children born to participants in the trial are required to determine whether or not there are any longer-term benefits or harms of maternal metformin for offspring weight, fat mass or metabolism.