TyG and TyHGB indices predict type 2 diabetes onset in elderly Chinese cohort

Photo by Navy Medicine / Unsplash

Frontiers in Medicine Published April 14, 2026 Medically reviewed April 25, 2026 DOI ↗ By Dr. Amelia Tan, PhD · Internal Medicine & Chronic Disease

Key Takeaway

Consider TyG and TyHGB as potential monitoring indicators for T2D risk in elderly Chinese patients, but interpret cautiously as associations.

This prospective cohort study enrolled 18,251 community-dwelling elderly individuals in Shenzhen, China, with follow-up until the end of 2022. It assessed six novel obesity- and lipid-related indices—BMI, WC, VAI, LAP, TyG, and TyHGB—comparing their associations with different fasting blood glucose statuses (normal and elevated FBG) to predict the primary outcome of type 2 diabetes onset. Secondary outcomes included predictive performance (AUC), optimal cutoff values, and linear versus nonlinear associations.

Main results showed positive associations between all indices and T2D risk. Per one-standard-deviation increase, adjusted hazard ratios were: BMI 1.17 (95% CI: 1.11–1.23), WC 1.23 (95% CI: 1.16–1.30), VAI 1.30 (95% CI: 1.24–1.37), LAP 1.21 (95% CI: 1.17–1.26), TyG 1.76 (95% CI: 1.67–1.86), and TyHGB 1.37 (95% CI: 1.33–1.42). TyG and TyHGB exhibited better predictive performance, with AUCs of 0.706 (95% CI: 0.691–0.721) and 0.737 (95% CI: 0.724–0.750), respectively (P<0.001). Overall, 1,350 participants (7.40%) were diagnosed with T2D during the study.

Safety and tolerability were not reported. Key limitations include the observational design, which precludes causal inferences, and lack of reported funding or conflicts. The study population was specific to elderly Chinese individuals, limiting generalizability. Practice relevance is restrained: TyG and TyHGB may serve as valuable indicators for monitoring and preventing T2D in similar settings, but clinicians should interpret these as associations rather than proven causal factors, with further research needed for validation.

Study Details

Study typeCohort

EvidenceLevel 3

PublishedMar 2026

View Original Abstract ↓

BackgroundAlthough VAI, LAP, TyG and TyHGB are considered alternative indicators of abdominal fat deposition, their longitudinal association and strength of these indices with T2D remain unclear. The study aimed to evaluate the association between the indices and onset T2D in the elderly population across different FBG statuses and compare their predictive performance in risk assessment.MethodsData were from the BaHLS, a cohort study of community-dwelling elderly individuals in Shenzhen, China. The study examined the associations between six novel obesity- and lipid-related indices (including BMI, WC, VAI, LAP, TyG and TyHGB) and the onset of T2D across different FBG statuses, including normal and elevated FBG. A multivariate Cox proportional hazards model and a GAM were employed to assess the longitudinal associations between each index and T2D. ROC curves analysis and AUC were utilized to evaluate the predictive performance of indices and determine their optimal cutoff values.ResultsA total of 18,251 participants were enrolled in the study, with 1,350 (7.40%) participants were diagnosed with T2D by end of 2022. In the overall elderly cohort, the adjusted HRs of T2D per one-standard-deviation increase were: BMI 1.17 (95% CI: 1.11–1.23), WC 1.23 (95% CI: 1.16–1.30), VAI 1.30 (95% CI: 1.24–1.37), LAP 1.21 (1.17–1.26), TyG 1.76 (95% CI: 1.67–1.86), and TyHGB 1.37 (95% CI: 1.33–1.42). Except for BMI and WC, the remaining indices were independent risk factors for the onset of T2D in both normal and elevated FBG individuals. These indices showed significant associations with T2D across different subgroups. Regarding predictive performance, all six indices demonstrated predictive ability for the onset of T2D across all participants (AUC > 0.5, P< 0.001). Specifically, TyG (AUC = 0.706, 95% CI: 0.691–0.721) and TyHGB (AUC = 0.737, 95% CI: 0.724–0.750) exhibited better predictive performance than the other indices. RCS analysis revealed that BMI and WC exhibited linear associations with the risk of T2D across all participants, whereas VAI, LAP, TyG, and TyHGB demonstrated nonlinear relationships.ConclusionThe study demonstrated that 6 obesity- and lipid- indices are positively associated with the incidence of onset T2D, and TyG and TyHGB demonstrated superior predictive performance for the onset of T2D in the elderly. Suggesting that TyG and TyHGB should be served as a valuable indicator for monitoring and preventing T2D.