Screening, diagnosis, and treatment of gestational diabetes mellitus (GDM) are common practice, despite controversy regarding benefits. A review of the literature from 1950 to 2006 revealed 3 randomized controlled trials evaluated the treatment of GDM but 2 of these studies lacked power to detect a difference in outcomes. The single trial with sufficient power showed a 67% lower rate of serious perinatal complication (a composite of shoulder dystocia, nerve injury, fracture, and death) and a 53% lower rate of macrosomia with treatment of GDM. There are no well-designed studies evaluating screening or diagnostic strategies. Treatment of GDM may improve some neonatal and obstetric outcomes, but there is limited evidence useful for determining the best screening method or diagnostic test, strategy, and criteria. Ongoing studies may provide some evidence to guide future research and clinical practice.