Temporal Trend of Non-Invasive Method Capacity for Early Detection of Metabolic Syndrome in Children and Adolescents: A Bayesian Multilevel Analysis of Pseudo-Panel Data

Abstract

Background: The aim of this study was to examine the ability of Noninvasive methods to early predictions of metabolic syndrome (MetS) among children and adolescents from 2003 to 2016. Methods: This was a repeated cross-sectional study based on 24,409 Iranian children and adolescents. The variables included anthropometric measures, serum lipid profiles, hypertension, and MetS. The receiver operating characteristic regression and Bayesian multilevel modeling conducted on data to comparison the power of anthropometric measures to early prediction of cardiometabolic risk factors. Results: The tri-ponderal body shape index (TBSI) in females and waist circumference (WC) percentile in males yielded a greater ability to predict lipid profiles and hypertension than the rest of anthropometric factors. The TBSI (β = 6.24, 95% credible interval [95% Crl] 3.9–8.7) followed by the WC percentile (β = 4.43, 95% Crl 3.5–5.4) were considered the better predictors of MetS compared with the body mass index (BMI), tri-ponderal mass index (TMI), WC, waist-to-height ratio, and WC to height⁵ in adolescents. The TBSI with Youden index J (JI) = 0.85 was significantly more accurate than the BMI (JI = 0.73), and TMI (JI = 0.7) for classifying individuals with MetS and in healthy groups. The predictability of early MetS was consistent for both TBSI and WC components throughout the study period. Conclusions: The TBSI including, both BMI and WC components, predicts MetS and cardiometabolic risk factors more accurately than BMI or WC alone in females. The TBSI ability was higher than other anthropometric factors for screening MetS and cardiometabolic risk factors among adolescents.