Background: This study aimed to compare the efficacy and safety of insulin versus oral hypoglycemic drugs for gestational diabetes treatment, using evidence based medicine approach.

Materials and Methods: This secondary study was conducted by selecting a comprehensive systematic review through searching the Cochrane library. To calculate the number needed to treat (NNT), outcomes such as cesarean section, macrosomia, birth weight more than 90th percentile, operative vaginal delivery, labor induction, post partum hemorrhage, and shoulder dystocia has been considered while neonatal hypoglycemia, NICU hospitalization and fetal distress were considered for calculating number needed to harm (NNH). To calculate the likelihood of helped versus harmed (LHH), macrosomia and hypoglysmia were selected. Data were analyzed with Wilson score test.

Results: The NNT for above outcomes were 3, -14, -5, 16, -3, 31, and 32, respectively. Among them cesarean section was significance. Calculated NNH for side effects were 6, 15, and -17 which only was significance for neonatal hypoglycemia (p<0.05). The LHH was estimated -0.43.

Conclusion: Considering LHH, it seems that the benefit of insulin in management of gestational diabetes is more than metformin. But with respect to issues such as low sample size of trials, lack of assessment the maternal hypoglysemia as complication of insulin therapy, the findings should be considered cautiously.