Association between novel triglyceride-glucose (TyG) related indices and MAFLD risk in rural adults from Xinjiang

Combined indices of triglyceride-glucose (TyG) and obesity indicators (e.g., BMI, waist circumference) show good predictive value for metabolic dysfunction-associated fatty liver disease (MAFLD). However, whether combining TyG with abdominal obesity indicators—Chinese visceral adiposity index (CVAI), body roundness index (BRI), waist-to-hip ratio (WHR), and lipid accumulation product (LAP)—offers improved predictive performance remains unclear. This prospective cohort study included 10,703 adult Uyghur residents aged ≥ 18 years. Kaplan-Meier curves estimated MAFLD cumulative incidence. Cox proportional hazards models assessed associations between TyG-related indices and new-onset MAFLD. Restricted cubic spline analysis explored non-linear relationships. Subgroup analyses and receiver operating characteristic curves were performed, with pairwise AUCs compared using the DeLong test and incremental predictive value evaluated by net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Sensitivity analyses yielded consistent results. During median follow-up of 5.79 years, 596 (5.57%) participants developed MAFLD. Cumulative incidence increased progressively across quartiles of each index (log-rank P < 0.001). After multivariable adjustment (Model 4), each one-standard-deviation increase in TyG-CVAI, TyG-BRI, TyG-WHR, and TyG-LAP was associated with 60%, 29%, 12%, and 6% increased MAFLD risk, respectively. Participants in the highest TyG-CVAI quartile had 493% higher risk versus the lowest (HR = 5.93; 95%CI: 4.04–8.71). Non-linear associations were significant (P for nonlinearity < 0.05). TyG-related indices showed positive associations in most subgroups. TyG-BMI demonstrated the best predictive efficacy overall and in males (AUC = 0.714 and 0.701), while TyG-CVAI showed the highest AUC in females (0.747 vs. 0.723 for TyG-BMI; DeLong test P = 0.007) with significant incremental predictive value (NRI = 0.637, IDI = 0.022, both P < 0.001). Novel TyG-related indices are independently associated with MAFLD risk. TyG-CVAI showed a modestly higher discriminative ability in females compared to traditional indices, suggesting population-specific clinical potential for MAFLD risk stratification.