Systemic inflammation-based refinement of triglyceride glucose index for predicting new-onset cardiovascular disease: A nationwide multicenter cohort study

Background

Triglyceride glucose (TyG) index is a surrogate marker of insulin resistance, which has been related to cardiovascular disease (CVD) risk. However, the predictive rationale for incorporating triglyceride and glucose is not well established. Furthermore, whether inflammatory information adds prognostic value to TyG remains largely unknown.

Methods

This cohort study involved 7456 participants without pre-existing CVD. CVD events were defined as physician-diagnosed coronary heart disease (CHD) and/or stroke during follow-up. The predictive abilities of TyG, triglycerides, glucose, C-reactive protein (CRP), a TyC index (triglycerides [mg/dL]+60 ×CRP [mg/L]) and a TyG-C index (TyG+CRP [mg/L]) were compared. Using multivariable Cox models, hazard ratios (HR) of any and type-specific CVD according to TyG-C tertiles (T) were estimated.

Results

The median age of participants was 57.0 years. There were 3448 (46.2 %) men and 4008 (53.8 %) women. During the follow-up of up to 6 years, 1470 CVD, 1114 CHD and 487 stroke incidents occurred. TyG showed comparable discrimination to triglycerides in predicting new-onset CVD and CHD, and similar performance to glucose in predicting stroke (all P > 0.05). TyG-C was superior to TyG and other investigated indices for predicting CVD (Harrell’s C-index [95 %CI]=0.557 [0.541–0.573], all P < 0.05). TyG-C tertiles were monotonically associated with CVD risk (HR [95 %CI]: T2 vs T1, 1.200 [1.050–1.372]; T3 vs T1, 1.282 [1.121–1.468], P trend<0.001). Similar patterns were observed for CHD and stroke. These associations attenuated with age (all P interaction<0.001).

Conclusion

TyG does not outperform its components in predicting new-onset CVD. TyG-C might serve as a better biomarker of CVD risk.