BACKGROUND: Gestational diabetes mellitus (GDM) is carbohydrate intolerance resulting in hyperglycaemia with onset or first recognition during pregnancy. If untreated, perinatal morbidity and mortality may be increased. Accurate diagnosis allows appropriate treatment. Use of different tests and different criteria will influence which women are diagnosed with GDM. OBJECTIVES: To evaluate and compare different testing strategies for diagnosis of gestational diabetes mellitus to improve maternal and infant health while assessing their impact on healthcare service costs. SEARCH METHODS: We searched the Cochrane Pregnancy and Childbirth Group's Trials Register (31 October 2014) and reference lists of retrieved studies. SELECTION CRITERIA: We included randomised trials if they evaluated tests carried out to diagnose GDM. We excluded studies that used a quasi-random model. DATA COLLECTION AND ANALYSIS: Two review authors independently assessed trials for inclusion and risk of bias, extracted data and checked them for accuracy. MAIN RESULTS: We identified six small trials, including 694 women. These trials were assessed as having varying risk of bias, with few outcomes reported. We prespecified six outcomes to be assessed for quality using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach; data for only one outcome (diagnosis of gestational diabetes) were available for assessment. One trial compared three different methods of delivering glucose: a candy bar (39 women), a 50-gram glucose polymer drink (40 women) and a 50-gram glucose monomer drink (43 women). We have reported results reported by this trial as separate comparisons. 75-gram oral glucose tolerance test (OGTT) versus 100-gram OGTT (one trial, 248 women): Women given the 75-gram OGTT had a higher relative risk of being diagnosed with GDM (risk ratio (RR) 2.55, 95% confidence interval (CI) 0.96 to 6.75). This difference was borderline in terms of statistical significance, and evidence was considered to be of very low quality when assessed by GRADE. No data were reported for the following additional outcomes prespecified for assessment in GRADE: caesarean section, macrosomia > 4.5 kg or however defined in the trial, long-term type 2 diabetes maternal, long-term type 2 diabetes infant and economic costs. Candy bar versus 50-gram glucose monomer drink (one trial, 60 women): More women receiving the candy bar, rather than glucose monomer, preferred the taste of the candy bar (RR 0.60, 95% CI 0.42 to 0.86). Infant outcomes were not reported. 50-gram glucose polymer drink versus 50-gram glucose monomer drink (three trials, 239 women): Mean difference (MD) in gestation at birth was -0.80 weeks (one trial, 100 women; 95% CI -1.69 to 0.09). Total side effects were less common with the glucose polymer drink (one trial, 63 women; RR 0.21, 95% CI 0.07 to 0.59), and no clear difference in taste acceptability was reported (one trial, 63 women; RR 0.99, 95% CI 0.76 to 1.29). Significantly fewer women reported nausea following the 50-gram glucose polymer drink compared with the 50-gram glucose monomer drink (one trial, 66 women; RR 0.29, 95% CI 0.11 to 0.78). No other measures of maternal morbidity or outcomes for the infant were reported. 50-gram glucose food versus 50-gram glucose drink (one trial, 30 women): Women receiving glucose in their food, rather than as a drink, reported fewer side effects (RR 0.08, 95% CI 0.01 to 0.56). No clear difference was noted in the number of women requiring further testing (RR 0.14, 95% CI 0.01 to 2.55). No other measures of maternal morbidity or outcome were reported for the infant. 75-gram oral glucose tolerance test (OGTT) World Health Organization (WHO) criteria versus 75-gram OGTT American Diabetes Association (ADA) criteria (one trial, 116 women): No clear differences in included outcomes were observed between women who received the 75-gram OGTT and were diagnosed using criteria based on WHO (1999) recommendations and women who received the 75-gram OGTT and were diagnosed using criteria recommended by the ADA (1979). Outcomes measured included diagnosis of gestational diabetes (RR 1.47, 95% CI 0.66 to 3.25), caesarean birth (RR 1.07, 95% CI 0.85 to 1.35), macrosomia defined as > 90th percentile by ultrasound or birthweight equal to or exceeding 4000 g (RR 0.73, 95% CI 0.19 to 2.79), stillbirth (RR 0.49, 95% CI 0.02 to 11.68) and instrumental birth (RR 0.21, 95% CI 0.01 to 3.94). AUTHORS' CONCLUSIONS: Evidence is insufficient to permit assessment of which strategy is best for diagnosing GDM.