Background: To address gaps in knowledge of the longitudinal trajectory of dysglycemia in children with obesity, this study aimed to: (1) describe the changes in glycemic status over 2 years; (2) establish a predictive model for development of prediabetes among children with euglycemia; and (3) evaluate the influence of change in body mass index (BMI) z-score on glycemic status. Methods: Children aged 5 to 17 years entered this prospective, longitudinal study at the time of entry to a weight management program. Measures included a 75-g oral glucose tolerance test (OGTT), fasting blood glucose, hemoglobin A1c (HbA1c), lipid profile, liver enzymes and anthropometric measures at baseline, 1 and 2years. Cox proportional hazard was used to build a predictive model for prediabetes. Results: The cohort included 270 children, mean age: 11.6 +/- 2.7years and BMI z-score: 3.1. The baseline prevalence of prediabetes, based upon elevated 2-hour glucose in OGTT or HbA1c, was 100/270 (37.0%). Among children with prediabetes at baseline, 53 (53.0%) continued to have prediabetes over the following 2 years, 15 (15.0%) were euglycemic at 1year and had prediabetes at 2years, 20 (20.0%) became euglycemic and remained so. Change in BMI z-score predicted dysglycemic status at 2years. Among those euglycemic at baseline, the incidence of prediabetes was 14 (8.2%) after 1 year, 20 (12.8%) at 2 years. Predictors of incident prediabetes were baseline BMI z-score; hazard ratio (HR): 1.72, 95th confidence interval (CI: 1.08, 2.74) and baseline HbA1c HR: 1.26, 95th CI (1.02-1.56) when controlling for age, family history of diabetes and sex. Conclusion: Prediabetes presents significant morbidity in children with obesity. Family-based lifestyle interventions might delay prediabetes progression.
