Article Text

other Versions

Randomised controlled trial
Should metformin be preferred over insulin therapy in the management of gestational diabetes (GDM)?
  1. Shannon D Sullivan
  1. Department of Endocrinology & Metabolism, Medstar Washington Hospital Center, Washington, DC, USA
  1. Correspondence to: Dr Shannon D Sullivan
    Department of Endocrinology & Metabolism, Medstar Washington Hospital Center, 110 Irving Street, NW Suite 2A-72, Washington, DC, 20010, USA; Shannon.D.Sullivan{at}

Statistics from

Commentary on: Barrett HL, Gatford KL, Houda CM, et al. Maternal and neonatal circulating markers of metabolic and cardiovascular risk in the Metformin in Gestational Diabetes (MIG) trial: Responses to maternal metformin versus insulin treatment. Diabetes Care 2013;36:529–36.


In this ancillary study of the Metformin in Gestational Diabetes (MiG) Trial, a prospective, randomised, multicentre trial in New Zealand and Australia, the investigators compare several circulating markers of metabolic risk in women with gestational diabetes (GDM) and in the cord blood of their offspring, to determine the differences between those treated with metformin or insulin during pregnancy.


In 236 metformin-treated and 242 insulin-treated women, fasting glucose, lipids, triglycerides (TGs), high-density lipoprotein, low-density lipoprotein and C reactive protein (CRP) were measured at randomisation (20–33 weeks’ gestation), at 36 weeks and at 6–8 weeks postpartum, as well as in cord blood. Treatment-group differences in maternal glucose, lipids and CRP levels over the course of pregnancy, as well as the relationships between maternal and cord blood levels and birth weight, were analysed after adjustment for maternal age, ethnicity, parity, smoking and body mass index (BMI).


Women in both groups had similar baseline glucose, lipids and CRP levels, and similar time from randomisation to delivery. Of note, women treated with insulin gained significantly more weight than women treated with metformin (1.9±3.4 kg vs 0.4±2.9 kg at 36 weeks, p<0.001). Metformin-treated women saw greater increases in TG levels at 36 weeks, however, by 6–8 weeks postpartum, TG levels equalised. The percentage change in fasting glucose from randomisation to the postpartum was greater in metformin-treated women (−2.9% vs −0.7%, p=0.04). Changes in all other maternal markers, cord blood markers and birth weight of offspring did not differ between treatment groups. At randomisation and at 36 weeks, birth weight was positively correlated with maternal glucose (r=0.19, p<0.001) and birth weight centile was positively correlated with maternal TG levels (r=0.17, p=0.001). Birth weight was positively correlated with cord blood c-peptide and negatively correlated with cord blood glucose and TG levels. Maternal BMI at randomisation was the strongest predictor of infant birth weight.


Gestational diabetes increases maternal and fetal morbidity, in part by increasing rates of large-for-gestational-age (LGA) infants and by heightening future cardiometabolic risk for mother and child. GDM has traditionally been treated with insulin, and recently, the use of the insulin-sensitising agent metformin has increased. One advantage of metformin is that it tends to promote weight loss (or is weight neutral), whereas insulin is known to promote weight gain, important considerations in overweight or obese women, given that maternal BMI is one of the strongest predictors of LGA neonates.

In this study, maternal and neonatal glucose, lipids and CRP, and infant birth weights were compared between women with GDM treated with metformin versus insulin during pregnancy. Not surprisingly, metformin more favourably affected maternal weight gain compared to insulin therapy. Interestingly, metformin-treated women also saw larger drops in fasting glucose over the course of pregnancy compared to insulin-treated women. This more favourable effect on glycaemic control may have been due to less weight gain, better medication compliance or improved insulin sensitivity in metformin-treated compared to insulin-treated women. On the other hand, metformin therapy resulted in a greater rise in maternal triglyceride levels at 36 weeks’ gestation; however, this resolved postpartum and did not affect cord blood TG levels. Overall, the metabolically favourable effects of metformin may be advantageous during GDM pregnancies, especially when considering that, first, infant birth weight was positively associated with maternal fasting glucose levels and, second, higher birth weight was most strongly correlated with higher maternal BMI. Of note, the women in the MiG Trial were obese at baseline (average BMI of ∼35 kg/m2), thus a treatment like metformin that minimises weight gain and improves insulin sensitivity may be particularly advantageous.

Taken together, these data provide further support for the use of metformin during GDM pregnancies. Indeed, the use of metformin, compared to subcutaneous insulin injections, will improve insulin sensitivity, minimise weight gain and undoubtedly be easier to comply with, potentially leading to improved pregnancy outcomes. Additionally, as obesity rates rise, more women are being diagnosed with polycystic ovary syndrome (PCOS) or prediabetes, for which metformin is a popular and effective treatment option. Women with PCOS and prediabetes face higher rates of metabolic syndrome and increased risk of developing GDM compared to women without these disorders. For them, continuing metformin therapy throughout pregnancy is a suitable option for the prevention and treatment of GDM.

View Abstract


  • Competing interests None.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.