RESUMO
BACKGROUND: Gestational diabetes that is not properly controlled with diet has been commonly treated with insulin. In recent years, several studies have published that metformin can lead to, at least, similar obstetrical and perinatal outcomes as insulin. Nevertheless, not all clinical guidelines endorse its use, and clinical practice is heterogeneous. OBJECTIVE: This study aimed to test whether metformin could achieve the same glycemic control as insulin and similar obstetrical and perinatal results, with a good safety profile, in women with gestational diabetes that is not properly controlled with lifestyle changes. STUDY DESIGN: The metformin for gestational diabetes study was a multicenter, open-label, parallel arms, randomized clinical trial performed at 2 hospitals in Málaga (Spain), enrolling women with gestational diabetes who needed pharmacologic treatment. Women at the age of 18 to 45 years, in the second or third trimesters of pregnancy, were randomized to receive metformin or insulin (detemir or aspart). The main outcomes were (1) glycemic control (mean glycemia, preprandial and postprandial) and hypoglycemic episodes and (2) obstetrical and perinatal outcomes and complications (hypertensive disorders, type of labor, prematurity, macrosomia, large for gestational age, neonatal care unit admissions, respiratory distress syndrome, hypoglycemia, jaundice). Outcomes were analyzed on an intention-to-treat basis. RESULTS: Between October 2016 and June 2019, 200 women were randomized, 100 to the insulin-treated group and 100 to the metformin-treated group. Mean fasting and postprandial glycemia did not differ between groups, but postprandial glycemia was significantly better after lunch or dinner in the metformin-treated-group. Hypoglycemic episodes were significantly more common in the insulin-treated group (55.9% vs 17.7% on metformin; odds ratio, 6.118; 95% confidence interval, 3.134-11.944; P=.000). Women treated with metformin gained less weight from the enrollment to the prepartum visit (36-37 gestational weeks) (1.35±3.21 vs 3.87±3.50 kg; P=.000). Labor inductions (45.7% [metformin] vs 62.5% [insulin]; odds ratio, 0.506; 95% confidence interval, 0.283-0.903; P=.029) and cesarean deliveries (27.6% [metformin] vs 52.6% [insulin]; odds ratio, 0.345; 95% confidence interval, 0.187-0.625; P=.001) were significantly lower in the metformin-treated group. Mean birthweight, macrosomia, and large for gestational age and babies' complications were not different between treatment groups. The lower cesarean delivery rate for women treated with metformin was not associated with macrosomia, large or small for gestational age, or other complications of pregnancy. CONCLUSION: Metformin treatment was associated with a better postprandial glycemic control than insulin for some meals, a lower risk of hypoglycemic episodes, less maternal weight gain, and a low rate of failure as an isolated treatment. Most obstetrical and perinatal outcomes were similar between groups.
