Abstract 44: Childhood Predictors of Type 2 Diabetes in Young Adulthood to Middle-age: The International Childhood Cardiovascular Cohort (i3c) Consortium

• Published in: Circulation (New York, N.Y.) N.Y.), 2020-03, Vol.141 (Suppl_1)
• Main Authors: Hu, Tian, Jacobs, David R, Sinaiko, Alan R, Bazzano, Lydia A, Burns, Trudy L, Daniels, Stephen R, Dwyer, Terence, Juonala, Markus, Prineas, Ronald J, Raitakari, Olli T, Urbina, Elaine M, Venn, Alison, Woo, Jessica G, Steinberger, Julia
• Format: Article
• Language: English
• ISSN: 0009-7322; 1524-4539
• DOI: 10.1161/circ.141.suppl_1.44

Abstract only Introduction: Although childhood obesity is related to incident adult type 2 diabetes (T2DM), whether other childhood risk factors relate to T2DM is not clear and no childhood risk score has been established to help with prediction of T2DM. Methods: We examined childhood risk factors in The International Childhood Cardiovascular Cohort (i3C) Consortium with a goal of developing a risk score for predicting adult T2DM. The pooled datasets included childhood measurements during the 1970s to 1990s. Participants were re-contacted at mean age 40 yr and completed a health questionnaire including self-report of adult T2DM (occurrence age, medication use). Results: N = 4,425 i3C participants with measurements during ages 3-19 yr. 225 (5.1%) reported onset of T2DM between ages 20-55 (mean 38.4) years (82% reported use of a valid antidiabetic medication). After stepwise regression, age-sex standardized deviates (mean of all repeated measures during ages 3-19 yr) of body mass index (BMI), glucose and ln(insulin) were associated with incident T2DM; the unadjusted risk curves visibly steepened at about 0.75 standard deviates (Figure, top panel). In multivariable models, BMI, glucose and ln(insulin) in a single model were associated with adult T2DM after adjustment for sex, race, cohort, and individual mean childhood age and calendar year. Prediction of incident T2DM was directly related to the number of the three risk factors with levels above +0.75 standard deviates (0: 3.2% (96/2966), 1: 6.2% (63/1,012), 2: 14.3% (56/392), 3:18.2% (10/55); the corresponding adjusted HR (95%CI) for the counts were 1 (reference), 1.8 (1.3, 2.4), 4.7 (3.3, 6.5), and 6.7 (3.5, 13.1), respectively). The +0.75 standard deviates in natural units are given in the Figure (bottom panel), e.g. about 92 mg/dL for glucose at age 12-14. Conclusion: The study suggests that a childhood risk score composed of BMI, glucose and insulin is a strong predictor of adult T2DM. Provisionally suggested cutpoints are lower than currently defined normal.