Predicting progression to type 1 diabetes from ages 3 to 6 in islet autoantibody positive TEDDY children

Area of Science:

• Endocrinology
• Genetics
• Precision Medicine

Background:

• Type 1 diabetes (T1D) prediction in young children remains a significant clinical challenge.
• Early identification of children at high risk for T1D is crucial for timely intervention.
• The Environmental Determinants of Diabetes in the Young (TEDDY) study provides a valuable cohort for investigating T1D progression.

Purpose of the Study:

• To develop and validate a risk prediction algorithm for T1D progression in children with high-risk human leukocyte antigen (HLA) genes.
• To identify key clinical, immunologic, metabolic, and genetic predictors of T1D development.
• To compare the performance of a logistic regression model against simpler prediction methods.

Main Methods:

• Logistic regression and 4-fold cross-validation were employed to analyze 38 candidate predictors.
• The study focused on TEDDY participants with at least one persistent autoantibody at age 3, with T1D by age 6 as the primary endpoint.
• The developed model was compared to predictions based on multiple autoantibody status and insulinoma-associated-2 autoantibodies (IA-2A) presence.

Main Results:

• A total of 363 subjects with autoantibodies at age 3 were analyzed; 21% progressed to T1D by age 6.
• Five significant predictors were identified: IA-2A status, hemoglobin A1c, BMI Z-score, SNP rs12708716_G, and a combined autoantibody/fasting insulin marker.
• The logistic regression model achieved an AUC of 0.80, outperforming simpler predictors, though differences in performance metrics were marginal.

Conclusions:

• Precision medicine approaches can enhance T1D progression prediction in children over a 3-year timeframe.
• The developed multifaceted model offers a preliminary improvement over existing prediction tools.
• Further research and additional predictive tools are necessary to optimize T1D risk assessment.