Predicting Insulin Resistance in a Pediatric Population With Obesity

Area of Science:

• Pediatric Endocrinology
• Metabolic Disorders
• Obesity Research

Background:

• Insulin resistance (IR) is a growing concern in obese children and adolescents, necessitating early diagnosis to prevent long-term health issues.
• Identifying predictive factors for IR is crucial for timely intervention and management in pediatric populations.

Purpose of the Study:

• To identify predictors of insulin resistance (IR) in children and adolescents.
• To develop and validate a multivariate model for accurate IR prediction in this demographic.

Main Methods:

• A cross-sectional analysis of demographic, clinical, and biochemical data from 1423 participants (aged 3-17 years) was performed.
• Multivariate regression models were developed and validated using distinct patient cohorts.
• Model performance was assessed using area under the curve (AUC), sensitivity, specificity, and negative predictive values.

Main Results:

• Models utilizing demographic and clinical variables showed good discriminative ability for IR (AUC: 0.834-0.868).
• Adding fasting insulin significantly enhanced predictive power (AUC: 0.996), with validated models demonstrating excellent discrimination (AUC: 0.978).
• High negative predictive values were consistently observed across models, indicating strong ability to rule out IR.

Conclusions:

• Demographic and clinical variables effectively identify children and adolescents at moderate to high risk of IR.
• Fasting insulin evaluation is recommended for individuals identified by these predictive models.
• The developed models offer a valuable tool for early IR detection in pediatric obesity.