BACKGROUND: There is uncertainty as to the optimal approach for screening and diagnosis of gestational diabetes mellitus (GDM). Based on systematic reviews published in 2003 and 2008, the U.S. Preventive Services Task Force concluded that there was insufficient evidence upon which to make a recommendation regarding routine screening of all pregnant women. OBJECTIVES: (1) Identify properties of screening tests for GDM, (2) evaluate benefits and harms of screening for GDM, (3) assess the effects of different screening and diagnostic thresholds on outcomes for mothers and their offspring, and (4) determine the benefits and harms of treatment for a diagnosis of GDM. DATA SOURCES: We searched 15 electronic databases from 1995 to May 2012, including MEDLINE and Cochrane Central Register of Controlled Trials (which contains the Cochrane Pregnancy and Childbirth Group registry); gray literature; Web sites of relevant organizations; trial registries; and reference lists. METHODS: Two reviewers independently conducted study selection and quality assessment. One reviewer extracted data, and a second reviewer verified the data. We included published randomized and nonrandomized controlled trials and prospective and retrospective cohort studies that compared any screening or diagnostic test with any other screening or diagnostic test; any screening with no screening; women who met various thresholds for GDM with those who did not meet various criteria, where women in both groups did not receive treatment; any treatment for GDM with no treatment. We conducted a descriptive analysis for all studies and meta-analyses when appropriate. Key outcomes included preeclampsia, maternal weight gain, birth injury, shoulder dystocia, neonatal hypoglycemia, macrosomia, and long-term metabolic outcomes for the child and mother. RESULTS: The search identified 14,398 citations and included 97 studies (6 randomized controlled trials, 63 prospective cohort studies, and 28 retrospective cohort studies). Prevalence of GDM varied across studies and diagnostic criteria: American Diabetes Association (75 g) 2 to 19 percent; Carpenter and Coustan 3.6 to 38 percent; National Diabetes Data Group 1.4 to 50 percent; and World Health Organization 2 to 24.5 percent. Lack of a gold standard for the diagnosis of GDM and little evidence about the accuracy of screening strategies for GDM remain problematic. The 50 g oral glucose challenge test with a glucose threshold of 130 mg/dL versus 140 mg/dL improves sensitivity and reduces specificity. Both thresholds have high negative predictive values (NPV) but variable positive predictive values (PPVs) across a range of prevalence. There was limited evidence for the screening of GDM diagnosed less than 24 weeks' gestation (three studies). One study compared the International Association of Diabetes in Pregnancy Study Groups' (IADPSG) diagnostic criteria with a two-step strategy. Sensitivity was 82 percent, specificity was 94 percent. Only two studies examined the effects on health outcomes from screening for GDM. One retrospective cohort study (n=1,000) showed more cesarean deliveries in the screened group. A survey within a prospective cohort study (n=93) found the same incidence of macrosomia (≥4.3 kg) in screened and unscreened groups (7 percent each group). Thirty-eight studies examined health outcomes for women who met different criteria for GDM and did not undergo treatment. Methodologically strong studies showed a continuous positive relationship between increasing glucose levels and the incidence of primary cesarean section and macrosomia. One of these studies also found significantly fewer cases of preeclampsia, cesarean section, shoulder dystocia and/or birth injury, clinical neonatal hypoglycemia, and hyperbilirubinemia for women without GDM compared with those meeting IADPSG criteria. Among the other studies, fewer cases of preeclampsia were observed for women with no GDM and women who were false positive versus those meeting Carpenter and Coustan criteria. For maternal weight gain, few comparisons showed differences. For fetal birth trauma, single studies showed no differences for women with Carpenter and Coustan GDM and World Health Organization impaired glucose tolerance versus women without GDM. Women diagnosed based on National Diabetes Data Group GDM had more fetal birth trauma compared with women without GDM. Fewer cases of macrosomia were seen in the group without GDM compared with Carpenter and Coustan GDM, Carpenter and Coustan 1 abnormal oral glucose tolerance test, National Diabetes Data Group GDM, National Diabetes Data Group false positives, and World Health Organization impaired glucose tolerance. Fewer cases of neonatal hypoglycemia were found among patient groups without GDM compared with those meeting Carpenter and Coustan criteria. There was more childhood obesity for Carpenter and Coustan GDM versus patient groups with no GDM. Eleven studies compared diet modification, glucose monitoring, and insulin as needed with no treatment. Moderate evidence showed fewer cases of preeclampsia in the treated group. The evidence was insufficient for maternal weight gain and birth injury. Moderate evidence found less shoulder dystocia with treatment for GDM. Low evidence showed no difference for neonatal hypoglycemia between treated and untreated GDM. Moderate evidence showed benefits of treatment for reduction of macrosomia (>4,000 g). There was insufficient evidence for long-term metabolic outcomes among offspring. Five studies provided data on harms of treating GDM. No difference was found for cesarean delivery, induction of labor, small for gestational age, or admission to a neonatal intensive care unit. There were significantly more prenatal visits among those treated. CONCLUSIONS: While evidence supports a positive association with increasing plasma glucose on a 75 g or 100 g oral glucose tolerance test and macrosomia and primary cesarean section, clear thresholds for increased risk were not found. The 50 g oral glucose challenge test has high NPV but variable PPV. Treatment of GDM results in less preeclampsia and macrosomia. Current evidence does not show that treatment of GDM has an effect on neonatal hypoglycemia or future poor metabolic outcomes. There is little evidence of short-term harm from treating GDM other than an increased demand for services. Research is needed on the long-term metabolic outcome for offspring as a result of GDM and its treatment, and the "real world" effects of GDM treatment on use of care.