Prediction of Insulin Resistance Based on Anthropometric and Clinical Variables in Children with Overweight or Obesity at a Tertiary Center in Northeast Mexico Academic Article in Scopus uri icon

abstract

  • Copyright © 2022, Mary Ann Liebert, Inc.Background: This study provides a clinical model to identify children with insulin resistance (IR) in health care units where laboratory tests are not readily available. Methods: A retrospective study of Mexican children aged 2-16 years at an obesity (OB) clinic. A receiver operating characteristic (ROC) curve was used to assess the accuracy of the proposed model consisting of clinical parameters and to establish the cutoff value for the variables (439 children). A second cohort of children with similar characteristics served as the cohort for the validation of the model (577 children). Results: To determine the best model for predicting IR, we performed a multivariate logistic regression analysis, which showed that waist circumference, acanthosis nigricans, and pubertal status are independent predictors of IR, and when integrated, their predictive power increases. Based on this model, we constructed a simplified equation. The predictive tool was constructed using an ROC curve, with an area under the curve of 0.849. A cutoff value of 7.68 was selected based on the Youden Index, with sensitivity and specificity of 78.3% and 83.3%, respectively. Incorporating metabolic laboratory determinations with a cutoff value of 20.64 improved the sensitivity to 94.9%. Conclusions: We developed a simple and affordable method of identifying IR in children with overweight or OB based on anthropometric variables and routine blood tests for metabolic indicators, such as glucose and triglycerides, which can be implemented in underserved sites.

publication date

  • April 1, 2022