Acarbose and Prandial Insulin for the Treatment of Gestational Diabetes Mellitus.

Phase III study. Prospective, multicenter, non-inferiority, randomized, open-labelled and controlled study with two arms. 1. In the 37 participating hospitals: selection of women with GDM who have unmet post prandial glycemic targets between 14 and 37 (+6 days) weeks of amenorrhea after at least 7 days of dietary and lifestyle measures. They may be treated with basal insulin to control pre prandial glucose values. 2. Explanation of protocol, with signature of consent in case of acceptation. 3. Randomization . Experimental group: The women will receive acarbose with a progressive increase of dose according to post prandial glucose values and digestive tolerance, with a maximal dose of 3 x 100 mg / day. The progressive titration of acarbose reduces gastro-intestinal side effects. Patients who have not reached the glycemic targets at this highest tolerated dose for at least one meal will receive instead prandial insulin therapy for each meal, whereas acarbose will be stopped. Failure to reach post-prandial target will be defined as 3 or more post-prandial glycaemic values ≥ 1.20 g/L for a given meal in a week (3 values out of 7) after the two weeks of dose adjustment. · Control group: The women will receive prandial insulin according to usual practice (routine care according to French recommendations): before each meal, with dose titration according to post prandial values. Basal insulin may be necessary in both arms to control pre-prandial glucose values. At delivery: \- Maternal blood samples : 14 ml of blood will be collected at the same time as the sample routinely collected just before delivery for irregular agglutinin test measurement, when the women are perfused. \- Cord fluid : 7 ml will be collected at the same as cord fluid pH is routinely measured just after delivery. There will be 5 aliquots to prepare. The aliquots previously labelled and stowed in the specific boxes for the study will be stored locally and will be transported to the "Centre de Ressources Biologiques"(CRB) of the Jean Verdier Hospital. 4. Routine monitoring of the women with GDM in both arms, up to delivery. No use of other oral hypoglycemic agents during pregnancy. 5. Last consultation three months after delivery.