Waist circumference partially mediates the non-linear association between TyHGB and MASLD in non-diabetic Japanese adults
Background The triglyceride high-density cholesterol-glucose body index (TyHGB) integrates atherogenic dyslipidemia, glucose metabolism, and generalized obesity, aligning closely with the pathophysiology of metabolic dysfunction-associated steatotic liver disease (MASLD). However, whether its association with MASLD follows a non-linear pattern and to what extent waist circumference (WC)-a surrogate for visceral adiposity-mediates this relationship remains unclear. Methods This cross-sectional study included 13,682 non-diabetic Japanese adults from a community-based cohort. Multivariable logistic regression, generalized additive models, and two-piecewise linear regression were used to evaluate the independent and non-linear association between TyHGB and prevalent MASLD. Mediation analysis with 1,000 bootstrap resamples quantified the mediating effect of WC and its heterogeneity across metabolic subgroups. Results The overall prevalence of MASLD was 15.25%. Each 1-unit increase in TyHGB was associated with a 1.35-fold higher odds of MASLD (OR = 1.35, 95% CI: 1.31–1.40, P < 0.001). A non-linear inflection point was identified at TyHGB = 9.70; below this threshold, the association was markedly stronger (OR=2.02, 95% CI: 1.88–2.17), whereas above it the risk increment plateaued (OR = 1.04, P = 0.090). WC partially mediated 46.4% of the TyHGB–MASLD association, with a substantially larger mediated proportion in non-obese individuals (42.67%) than in obese participants (7.83%). Conclusion In non-diabetic Japanese adults, TyHGB is independently and non-linearly associated with MASLD, with WC partially mediating this relationship. The marked attenuation of WC mediation among obese individuals suggests a relative shift towards non-WC-mediated pathways under high metabolic burden. TyHGB may serve as a practical composite surrogate marker for MASLD risk stratification in community settings, particularly for non-obese individuals in whom WC captures a larger share of the risk pathway and who are frequently missed by BMI-based assessment.