Predicting progression to diabetes in islet autoantibody positive children. (June 2018)

• Record Type: Journal Article
• Authors: Steck, Andrea K.; Dong, Fran; Frohnert, Brigitte I.; Waugh, Kathleen; Hoffman, Michelle; Norris, Jill M.; Rewers, Marian J.
• Is Part Of: Journal of autoimmunity. Volume 90(2018)
• Page Start: 59
• Page End: 63
• Publication Date: 2018-06
• Subjects: Islet autoimmunity -- Progression to diabetes -- Type 1 diabetes -- Predictors for diabetes -- Oral glucose tolerance test; Autoimmunity -- Periodicals; Autoimmune diseases -- Periodicals; Autoantibodies -- Periodicals; Autoimmune Diseases -- Periodicals; Auto-immunité -- Périodiques; Maladies auto-immunes -- Périodiques; Electronic journals; 616.978005
• DOI: 10.1016/j.jaut.2018.01.006
• Languages: English
• ISSNs: 0896-8411

Abstract

Abstract: While full oral glucose tolerance test (OGTT) helps improve prediction, it requires intravenous access with 6 sample collections for glucose and C-peptide. The objective of this study was to explore less costly and less time-consuming options. All children being prospectively followed by the Diabetes Autoimmunity Study in the Young (DAISY) who had a complete baseline OGTT and at least one confirmed islet autoantibody (Ab+) were included in this study (n = 68). Of 68 Ab+ subjects with a baseline OGTT, 25 developed diabetes after a mean follow-up 5.7 yrs, at a mean age of 12.4 yrs. Univariate proportional hazards (PH) models suggested that age at seroconversion, number of Ab+, IA-2A levels, HbA1c and metabolic variables from the OGTT predicted progression to diabetes, while HLA DR3/4, BMI, levels of IAA or GADA did not. Five multivariate PH predictive models were similar (p = 0.32). All five models included age at seroconversion, number of Ab+, IA-2A levels and HbA1c, and in addition included: model 1 - 1 h glucose and 1 h C-peptide; model 2 - 2 h glucose and 2 h C-peptide; model 3 - glucose sum and C-peptide sum; model 4 - glucose AUC and C-peptide AUC; and model 5: index 60. A model containing age at seroconversion, number of Ab+, IA-2A levels, HbA1c, 1 h glucose and 1 h C-peptide was as predictive for type 1 diabetes progression as models including all sum or AUC values for glucose and C-peptide from full OGTT. The performance of this model should be confirmed in an independent population of Ab+ children. Highlights: Islet antibody (Ab) positive children need surveillance for diabetes development. HLA DR3/4, BMI, IAA or GADA levels do not seem to predict progression to diabetes. Age, Ab number, IA-2A levels, HbA1c and OGTT variables predict diabetes progression. A 1 h glucose/C-peptide model is as predictive as glucose/C-peptide AUC or sum models. Surveillance with 1 h OGTT seems accurate, less costly and less time-consuming.