The predictive value of combined insulin resistance and β-cell secretion in Yemeni school-aged children for type 2 diabetes mellitus.

The present study aimed to determine the predictive power of the diabetic markers and metabolic syndrome factors in School-aged children for developing Type 2 DM. In this cross-sectional study, 1288 students aged 12-13 were recruited from public schools in the capital city of Sana'a. Anthropometric measurements and blood pressure were recorded and body mass index (BMI) was calculated. Fasting venous blood (5 ml) was collected for biochemical analysis including FBG, HbA1c, insulin, and lipid profile. Insulin resistance (HOMA-IR) and β-cell function (HOMA-β) were calculated. Our results showed that neither insulin, HOMA-IR nor HOMA-β individually were good predictors for Type 2 DM as assessed by the ROC curve with AUC < 0.75. However, the ROC curve of combined HOMA-IR and HOMA-β (Model 1) gave a superior AUC of 0.998 (p = 2.7 × 10) and predicted 140 (10.9%) children to develop Type 2 DM. This model picked up all impaired fasting glucose (IFG), 74% of metabolic glucose, and 71% of metabolic syndrome (MetS) groups. On the other hand, the ROC curve for metabolic syndrome (Model 2) gave an AUC of 0.751 (p = 0.003) and predicted a higher number of 416 (32.3%) children to develop prediabetes and Type 2 DM. This model picked up 75% of IFG, 71% of MetS, 82% of those having two factors of MetS, and 72% of obesity groups. Moreover, the 53 children common between the two models include 75% of IFG and 43% of MetS groups. Therefore, the combined HOMA-IR and HOMA-β model in children proved to be a good predictor for Type 2 DM development, whereas the MetS model predicts the development of prediabetes and Type 2 DM.