Comprehensive evaluation of the triglyceride glucose index (TyG) and body roundness index (BRI) on cardiovascular disease risk prediction: a 9-year prospective cohort study in Chinese middle-aged and older adults

Figures

• 

  Fig. 1. The flowchart

• 

  Fig. 2. Incidence rate of incident CVD. The figures illustrates the incidence rates (%) of CVD: Panel A displays incidence rates stratified by TyG & BRI categories, in which quantile1(Q1) means TyG index < median & BRI < median, Q2 means TyG index < median & BRI ≥ median, Q3 means TyG index ≥ median & BRI < median, and Q4 means TyG index ≥ median & BRI ≥ median. Panel B displays incidence rates stratified by TyG-BRI quartile, in which Q1: < 26.782, Q2:26.782–34.328, Q3:34.328–44.062, Q4: ≥ 44.062. Panel C displays incidence rates stratified by TyG + BRI quartile, in which Q1: < − 1.375, Q2: − 1.375–1.182, Q3: − 1.182–0.953, Q4: ≥ − 0.953. TyG, triglyceride-glucose index; BRI, body roundness index; CVD, cardiovascular disease

• 

  Fig. 3. Nonlinear association of TyG, BRI, and composite indexes with CVD. The figure displays the dose–response relationships of (A) TyG index and (B) BRI levels, as well as the composite indexes including (C) TyG-BRI and (D) TyG + BRI with the risk of cardiovascular events. The solid lines represent adjusted HRs and the shaded areas indicate the 95% CI. All models were adjusted for age, gender, residence, educational level, exercise, marital status, smoking status, drinking status, SBP, DBP, sleep during, chronic diseases (depression, diabetes, cancer, liver disease, lung diseases, CKD), WBC, PLT, BUN, SCR, UA, HB. TyG, triglyceride-glucose index; BRI, body roundness index; CVD, cardiovascular disease

• 

  Fig. 4. Kaplan–Meier survival curve. Subjects were stratified into TyG < median&BRI < median, TyG < median&BRI ≥ median, TyG ≥ median&BRI < median, and TyG ≥ median& BRI ≥ median based on the composite indexes (A) TyG&hs-CRP. Subjects were stratified into quartiles (Q1-Q4) based on the composite indexes (B) TyG-BRI and (C) TyG + BRI. TyG ≥ median& BRI ≥ median and higher quartiles (Q4) consistently demonstrated significantly increased risk of incident cardiovascular events compared to TyG < median&BRI < median and lower quartiles (Q1) across these biomarkers. TyG, triglyceride-glucose index; BRI, body roundness index; CVD, cardiovascular disease

• 

  Fig. 5. Performance comparison of individual and composite indexes in predicting CVD. Receiver operating characteristic curves demonstrates the discriminatory capacity of TyG index, BRI, TyG&BRI, TyG-BRI, and TyG + BRI for predicting incident CVD. AUC values are reported for each index comparison. ROC, receiver operating characteristic; AUC, area under the curve; TyG, triglyceride-glucose index; BRI, body roundness index; CVD, cardiovascular disease

• 

  Fig. 6. Performance comparison of basic models without and with composite indexes in predicting CVD. Receiver operating characteristic curves demonstrates the discriminatory capacity of basic models versus models enhanced with TyG&BRI, TyG-BRI, and TyG + BRI for predicting incident CVD. The basic model included traditional risk factors: age, gender, exercise, educational level, smoking status, drinking status, DBP, CKD, diabetes, lung diseases, liver diseases, HB, PLT, HDL-c, LDL-c, UA, hs-CRP. AUC values are reported for each model comparison. ROC, receiver operating characteristic; AUC, area under the curve; TyG, triglyceride-glucose index; BRI, body roundness index; CVD, cardiovascular disease

• 

  Fig. 7. Temporal comparison of predictive performance for TyG, BRI, and the composite indexes. This figure shows the area under the receiver operating characteristic curve values for various indexes assessed over different time periods (2-year, 4-year, 7-year and 9-year follow-up). The x-axis represents different parameters including TyG, BRI, TyG&BRI, TyG-BRI, and TyG + BRI. The y-axis shows AUC values ranging from 0.52 to 0.60. The composite index TyG + BRI demonstrated superior predictive performance compared to individual indexes

• 

  Fig. 8. Mediating effects of TyG and BRI on their associations with CVD risk. Path diagrams illustrates (A) mediating effects of TyG on the BRI-CVD association and (B) mediating effects of BRI on the TyG-CVD association in no-adjustment models; Path diagrams illustrates (C) mediating effects of TyG on the BRI-CVD association and (D) mediating effects of BRI on the TyG-CVD association in fully adjusted models, which adjusted for age, gender, residence, educational level, exercise, marital status, smoking status, drinking status, SBP, DBP, sleep during, chronic diseases (depression, diabetes, cancer, liver disease, lung diseases, CKD), WBC, PLT, BUN, SCR, UA, HB. Values represent standardized path coefficients with proportion of mediation effect shown as percentages. TyG, triglyceride-glucose index; BRI, body roundness index; CVD, cardiovascular disease

• 

  Fig. 9. Cross-lagged path analysis for TyG and BRI. The cross-lagged model was adjusted for age, gender, residence, educational level, exercise, marital status, smoking status, drinking status, SBP, DBP, sleep during, chronic diseases (depression, diabetes, cancer, liver disease, lung diseases, CKD), WBC, PLT, BUN, SCR, UA, HB. TyG, triglyceride-glucose index; BRI, body roundness index

• 

  Fig. 10. Estimated weights with the WQS model. Bar graph displays estimated weights of individual components within the composite insulin resistance-visceral adiposity indexes from WQS regression models for predicting incident CVD. Higher weights indicate greater relative importance of the component in the composite index-CVD relationship. All models were adjusted for potential confounders including age, gender, residence, educational level, exercise, marital status, smoking status, drinking status, SBP, DBP, sleep during, chronic diseases (depression, diabetes, cancer, liver disease, lung diseases, CKD), WBC, PLT, BUN, SCR, UA, HB. WQS, weighted quantile sum; TyG, triglyceride-glucose index; BRI, body roundness index; CVD, cardiovascular disease