Emerging Biomarkers in Pediatric NAFLD: Beyond BMI

In a prospective sample of 199 children, the triglyceride‑to‑HDL‑cholesterol ratio and the triglyceride‑glucose (TyG) index showed markedly higher sensitivity and specificity for non‑alcoholic fatty liver disease than standard anthropometrics and HOMA‑IR.

Where BMI and waist circumference summarize body size, TG/HDL‑C and TyG capture dyslipidemia–insulin‑resistance signals that align more closely with hepatic steatosis. In this cohort, anthropometrics and HOMA‑IR produced lower AUCs and narrower discriminatory margins, reflecting their limited ability to detect early metabolic perturbations that may precede visible changes in body habitus.

Cutoffs were explicit and ROC‑derived: TG/HDL‑C ≥2.40 achieved an AUC of 0.936 (81.8% sensitivity, 95.9% specificity), and TyG ≥3.66 produced an AUC of 0.912 (81.8% sensitivity, 91.8% specificity). These thresholds were established using the Youden index in 199 prospectively enrolled children (150 with obesity, 49 healthy controls; mean age ~11.8 years; obesity defined as BMI ≥95th percentile). The authors note these values reflect the specific cohort and analytic approach. The ROC cutoffs offer clear operating points but require cautious interpretation given sample size, single‑center recruitment, and limited ethnic diversity.

HOMA‑IR underperformed in this analysis (AUC 0.705), with lower combined sensitivity and specificity at its optimal cutoff. The authors attribute this to pediatric‑specific biological and assay limitations—pubertal insulin dynamics, intra‑assay variability in insulin measurement, and lower signal‑to‑noise for early hepatic fat—whereas lipid–glucose indices integrate fasting triglyceride and glucose signals that may better reflect hepatic lipid handling. Waist circumference and BMI remain useful markers of adiposity but had weaker standalone ROC performance for NAFLD detection.

Clinically, these findings suggest lipid–glucose indices are pragmatic, low‑cost additions to laboratory panels for risk stratification among at‑risk obese children, provided calculation is standardized, fasting conditions are consistent, and assays are reliable.

The authors call for external validation in larger, multi‑center and multi‑ethnic cohorts with explicit assessment of pubertal stage and imaging‑based standards to clarify generalizability and downstream cardiometabolic risk implications. Further validation is needed before routine pediatric screening adoption.

Key Takeaways:

• In 199 children, TG/HDL‑C (AUC 0.936) and TyG (AUC 0.912) had superior NAFLD discrimination versus BMI, waist circumference, and HOMA‑IR.
• At‑risk obese pediatric patients (150 obese, 49 controls; mean age ~11.8 years) showed the strongest differentiation with these indices.
• ROC‑derived cutoffs (TG/HDL‑C ≥2.40; TyG ≥3.66) provide defined operating points for validation studies and potential lab‑panel augmentation pending multicenter confirmation.