Body Mass Index Outperforms Triglyceride-Glucose Index in Predicting MAFLD in Type 2 Diabetes: A Cross-Sectional Mediation and Interaction Analysis

Abstract

Background The interplay between body mass index (BMI) and triglyceride-glucose (TyG) has yet to be systematically investigated on metabolic dysfunction-associated fatty liver disease (MAFLD) in type 2 diabetes mellitus (T2DM). This knowledge gap regarding the specific nature of their combined effects limits our ability to develop precise risk stratification algorithms and targeted intervention strategies. Purpose This study aimed to investigate the independent and synergistic roles of BMI and TyG in MAFLD development among T2DM patients. Methods A cross-sectional study was conducted among 816 T2DM patients recruited from the Affiliated Hospital of Nanjing University of Chinese Medicine. MAFLD was diagnosed via Fibroscan. Dose-response relationships were assessed using restricted cubic splines (RCS). Multiplicative and additive interactions were evaluated using logistic regression models. Causal mediation analysis quantified the proportion of the BMI-MAFLD association mediated by TyG. Predictive performance was compared using receiver operating characteristic (ROC) curves. Results Among 816 participants, 528 (64.7%) had MAFLD. RCS confirmed linear positive associations of BMI and TyG with MAFLD. Compared to those with BMI <24 kg/m2, patients with BMI ≥28 kg/m2 had an adjusted odds ratio (aOR) of 6.94 (95% CI: 2.59–18.6). The highest TyG quartile had an aOR of 3.23 (95% CI: 1.08–9.64) vs the lowest. Mediation analysis suggested that TyG explained 7.5% of the observed association between BMI and MAFLD. Significant multiplicative interaction was observed (P < 0.05), whereas additive interaction was not significant. ROC analysis demonstrated that TyG had inferior discriminative ability (AUC: 0.679) compared to BMI (AUC: 0.784), with combined models (AUC: 0.796) providing minimal incremental value over BMI alone in identifying MAFLD. Conclusion In T2DM, BMI exhibited a stronger association with MAFLD risk than TyG, with TyG accounting for a portion of this association but providing limited incremental improvement in discriminative ability.