Gestational diabetes (GD) is defined by a hyperglycemia discovered during pregnancy, leading to fetal and maternal complications which may be prevented by reaching very strict glycaemia targets. Prevalence depends on patient's ethnic group and is about 6 to 14%. This prevalence is increasing due to increased GD risk factors (obesity, pregnancy over 35) and also because criteria of screening have been strengthened after the results of last studies. Usual treatment is diet and in case of failure insulin therapy with multiple injections which may lead to hypoglycemia and weight gain and is very difficult to manage for patients. Some studies have shown the comparable effect of metformin and insulin in about 50% of GD obese patients. The aim of our study is to evaluate efficacy of metformin, outcomes in mother and fetus and baby of metformin. In case of metformin failure, insulin will be added in order to obtain glycaemia in desired goals.Oxidative stress will be assessed in mother blood, baby umbilical cord blood, baby umbilical cord and placenta in 90 women and the oxidative stress compared between insulin and metformin alone treated patients.

Inclusion Criteria:

Women.

• Age more than 18 and less than 40 years.
• Unique spontaneous pregnancy.-BMI less than 30kg/m2 before pregnancy.
• 24 to 30 weeks of amenorrhea or 22 to 28 weeks of gestation.
• Gestational diabetes.
• Social Security affiliated subject.- Patient able to understand and signed informed consent.

Exclusion Criteria:

• Contraindications to metformin.
• Metformin treatment prior to protocol inclusion.
• Multiple pregnancies.
• Diabetes diagnosed prior to pregnancy.
• High blood pressure prior to pregnancy.
• Pregnancy hepatic complication
• High blood pressure prior to study inclusion.
• Pre or eclampsia.- Premature membranes rupture.