Background Childhood and adolescent obesity has emerged as a global public health challenge. Metabolically unhealthy obesity (MUO) is closely associated with an increased risk of cardiovascular and metabolic diseases. There is currently a paucity of systematic research on screening indicators for MUO in children and adolescents.

Objective To compare the discriminative ability of different anthropometric and derived indicators for MUO versus metabolically healthy obesity (metabolically healthy obesity, MHO) in children and adolescents, and to provide a basis for early screening.

Methods This was a cross-sectional study. Clinical data were collected from obese children and adolescents who participated in the "Epidemiology and Surveillance System for Childhood Obesity-Related Metabolic Diseases" project from February to December 2023 in Changping District, Beijing. Anthropometric indicators (height, weight, waist circumference, hip circumference, body fat percentage, etc.), laboratory parameters (fasting blood glucose, blood lipids, insulin, etc.), and derived indicators (triglyceride-glucose index TyG, triglyceride to high-density lipoprotein cholesterol ratio TG/HDL-C, cardiometabolic index CMI, visceral adiposity index VAI, homeostasis model assessment of insulin resistance HOMA-IR, etc.) were recorded and compared between groups. Firth penalized logistic regression was used to analyze the independent association between each indicator and MUO, and receiver operating characteristic (ROC) curves were used to evaluate their discriminative ability. A parallel diagnostic test was further conducted to explore the net reclassification improvement (NRI) of composite indicators.

Results A total of 100 obese children and adolescents were enrolled, including 83 in the MUO group and 17 in the MHO group, with no significant differences in gender or age between the two groups. Compared with the MHO group, the MUO group had significantly higher levels of body fat percentage, triglycerides, fasting blood glucose, TG/HDL ratio, CMI, VAI, HOMA-IR, and TyG index (P < 0.05), and lower HDL levels (P < 0.05). After adjusting for age and sex, the TyG index (OR=11.146, 95% CI: 2.793 - 44.487, P < 0.001), TG/HDL ratio (OR=7.180, 95% CI: 1.678 - 30.718, P=0.003), CMI (OR=18.325, 95% CI: 1.648 - 203.797, P=0.008), TyG_BMI (OR=1.039, 95% CI: 1.015 - 1.064, P < 0.001), and VAI (OR=2.944, 95% CI: 1.115 - 7.778, P=0.017) were independently associated with MUO. ROC curve analysis showed that the TyG index had the best discriminative ability for MUO, with an AUC of 0.79 (95% CI: 0.69-0.89) and an optimal cutoff value of 8.47. The parallel analysis showed that the TyG+BMI combination had the best diagnosis performance and NRI (0.11 95% CI: 0.05-0.17).

Conclusion The TyG index, TG/HDL ratio, CMI, TyG_BMI, and VAI are independently associated with MUO. Among these, the TyG index exhibits the best discriminative ability, and its combination with BMI can further improve the discriminative performance, showing potential application value in the early screening of MUO in children and adolescents.