Childhood obesity increases the risk of type 2 diabetes (T2DM) and AAP guidelines recommend diabetes screening in youth with obesity. There is lack of consensus on the best screening measure due to lack of data sets inclusive of children with severe obesity with enough power to explore these differences. We examined use of fasting blood glucose(FBG) and hemoglobin A1c(HbA1c) for T2DM screening among a large nationally representative sample of obesity treatment-seeking youth.


Data were collected prospectively from 31 multicomponent pediatric weight management programs in POWER (the Pediatric Obesity Weight Evaluation Registry). Demographics and weight status as percentage of 95th percentile for body mass index (%BMIp95) for patients with FBG or HbA1c measurements were summarized. The weighted kappa statistic and Bowker’s test of symmetry for matched data was used to test agreement.


This cohort of 3852 children [median age 11.8 years (IQR 9.6,14.3), 46% males, 72% with severe obesity [median %BMIp95 131.5 (IQR 118,147)] underwent screening for T2DM. Tests obtained were FBG in 2512(65%), HbA1c in 3341(87%), and both in 2001(51.9%) patients. 9% had prediabetes (PDM) with FBG (100-125 mg/dL) and 24% had PDM with HbA1c (5.7-6.4%), while 0.8% had DM by FBG criteria (≥126 mg/dL) and 1.6% had DM by HbA1c criteria (≥6.5%). There was significant discordance between screening tests for PDM and DM. Of those with PDM based on criteria for HbA1c, 333/386(86%) had normal FBG; based on FBG criteria, 125/184(68%) had normal HbA1c. For those with T2DM based on criteria for HbA1c, 10/24(42%) had normal FBG; in those with FBG criteria, 9/17(53%) had normal HbA1c[Kappa=0.11 (95% CI 0.06, 0.16), p<0.0001].


There is heterogeneity in tests used to screen for T2DM in youth with obesity. The diagnosis of PDM and DM in these youth is influenced by the screening test used. Further studies are needed to understand the predictive capability of these tests in development of cardiometabolic risk.