UNLABELLED: What is already known about this subject Fasting triglycerides above 1.17 mmol/L have been shown to be useful to select obese children and adolescents who may present impaired glucose tolerance in a Canadian cohort. Fasting plasma glucose is associated with the risk to present impaired glucose tolerance in several cohorts of obese children and adolescents. What this study adds When applied to Italian cohorts of obese children and adolescents, the triglycerides cut-off of 1.17 mmol/L has similar validity as in the Canadian cohort to select patients who may present impaired glucose tolerance. Fasting plasma glucose and fasting triglycerides can be combined to obtain an accurate criterion to select obese children and adolescents who may present impaired glucose tolerance. OBJECTIVES: We aimed to validate fasting triglycerides > 1.17 mmol L(-1) , a criterion recently proposed for selecting obese children at risk of impaired glucose tolerance (IGT), and to assess whether the accuracy of triglycerides (TG) can be improved by the use of other variables. METHODS: We studied an Italian cohort of 817 obese children and adolescents (8-18.4 years) who underwent clinical examination, fasting blood analysis and the oral glucose tolerance test (OGTT). The discriminative properties of TG > 1.17 mmol L(-1) were assessed and compared with those observed in a Canadian cohort from which this criterion was derived: 71.4 [57.8-85.1]% sensitivity and 64.1 [57.7-70.4]% specificity. The possible contribution of other variables was evaluated by assessing the net reclassification improvement (NRI), i.e., the net increase in the percentage of subjects correctly classified. RESULTS: Thirty-nine children (4.7%) had IGT. The 1.17 mmol L(-1) TG threshold showed 66.6 [51.8-81.4]% sensitivity and 68.2 [64.9-71.5]% specificity, thus successfully validated. Fasting plasma glucose (FPG) was independently associated with IGT (odds ratio = 3.86 [2.09-7.14], P < 0.001), besides TG. The bivariate criterion of TG ≥ 1.13 mmol L(-1) plus FPG ≥ 4.44 mmol L(-1) had a 69.2 [54.7-83.7]% sensitivity and a 78.2 [76.8-79.6]% specificity, thus displaying a 12.6% NRI (P < 0.001) compared with TG>1.17 mmol L(-1) . CONCLUSIONS: TG > 1.17 mmol L(-1) is a useful criterion to detect roughly 66% of obese children with IGT through OGTT performed in about 33% of all obese children. However, the 'TG≥1.13 mmol L(-1) plus FPG≥4.44 mmol L(-1) ' criterion improved discrimination accuracy, leading to the possibility of detecting even more than 66% of obese children with IGT though limiting OGTT to just 25% of all obese children.