The role of multiple regression analysis in prediction of insulin resistance in overweight and obese children

Abstract

Introduction. Overweight and obesity is a global epidemic among children of all age groups. Obese children are at increased risk of insulin resistance, cardiovascular disease (including arterial hypertension), as well as bone fractures and psychological problems. In this regard, insulin resistance has become one of the most serious health concerns in overweight and obese children. The objective of the study was to investigate the specifics for carbohydrate metabolism in overweight and obese children, to identify the key factors for insulin resistance and to develop a regression analysis-based prognostic model to predict its occurrence. Material and methods. In 90 obese and 20 overweight children aged between 10-17 years, anthropometric measurements, history data collection and laboratory investigations were performed. Multiple regression analysis has been used to develop a mathematical model for prediction of insulin resistance. Results. Such variables as weight, body mass index, waist and hip circumferences, abdominal type obesity, family history, duration of breastfeeding (if any), birth weight, sedentary lifestyle, leptin and adiponectin levels and dyslipidemia were closely related to fasting glucose levels and insulin/insulin resistance indices. Conclusions. Abdominal obesity, male gender, family history of abnormal carbohydrate metabolism, insulin levels, duration of breastfeeding and plasma leptin levels have been defined as main predictors of insulin resistance in overweight and obese children and were included in regression equation for the index of insulin resistance using the method of multiple regression analysis.