Impact of the Triglyceride-Glucose index on all-cause and cardiovascular mortalities across different metabolic health and obesity statuses in US adults

Abstract

Background: Data for the study cohort were sourced from the National Health and Nutrition Examination Survey (1999-2018). Study participants were classified as obese (BMI ≥ 30 kg/m²) or nonobese (BMI < 30 kg/m²) then further categorized as metabolically healthy or unhealthy on the basis of metabolic syndrome criteria, resulting in four groups: metabolically healthy obese (MHO), metabolically unhealthy obese (MUO), metabolically healthy nonobese (MHNO), and metabolically unhealthy nonobese (MUNO). Complex sampling statistical methods were employed for descriptive analysis. The associations between the TyG index and mortality, including all-cause and cardiovascular mortalities, were examined by using multivariable Cox regression and restricted cubic splines (RCS). The reliability of the results was confirmed through multiple sensitivity analyses.

Methods: Data for the study cohort were sourced from the National Health and Nutrition Examination Survey (1999-2018). Study participants were classified as obese (BMI ≥ 30 kg/m²) or nonobese (BMI < 30 kg/m²) then further categorized as metabolically healthy or unhealthy on the basis of metabolic syndrome criteria, resulting in four groups: metabolically healthy obese (MHO), metabolically unhealthy obese (MUO), metabolically healthy nonobese (MHNO), and metabolically unhealthy nonobese (MUNO). Complex sampling statistical methods were employed for descriptive analysis. The associations between the TyG index and mortality, including all-cause and cardiovascular mortalities, were examined by using multivariable Cox regression and restricted cubic splines (RCS). The reliability of the results was confirmed through multiple sensitivity analyses.

Results: A total of 16 179 participants were included, with a median follow-up of 129 months. Over this follow-up period, 1875 participants (11.59%) died from all causes, including 568 (3.51%) who died due to cardiovascular diseases. After adjustment for confounding variables, the TyG index significantly predicted mortality in the overall and metabolically unhealthy populations: for each one standard deviation increase in the TyG index, all-cause mortality increased by 1.42 times (95% confidence interval [CI]: 1.27, 1.58) in the overall population, by 1.62 times (95% CI: 1.36, 1.93) in the MUNO group, and by 1.47 times (95% CI: 1.26, 1.71) in the MUO group. Cardiovascular mortality in the overall population increased by 1.52 times (95% CI: 1.27, 1.82), that in the MUNO group increased by 2.01 times (95% CI: 1.49, 2.72), and that in the MUO group increased by 1.47 times (95% CI: 1.14, 1.88). No significant association was found in the metabolically healthy populations regardless of obesity status. RCS and sensitivity analyses further confirmed and visualized these conclusions.

Conclusions: The TyG index is positively correlated with mortality risk in the overall and metabolically unhealthy populations but not in the metabolically healthy populations. This finding indicates that the predictive value of the TyG index for mortality differs across populations, highlighting the necessity of accounting for metabolic status when the TyG index is used for prognostic evaluation.

Fig. 1

Study Flow

Fig. 2

RCS Analysis of the Relationship Between the TyG Index and All-Cause Mortality. Hazard ratios were adjusted for gender, age, race or ethnicity, educational level, smoking status, alcohol use, family income-to-poverty ratio, estimated glomerular filtration rate, total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, hypertension status, high cholesterol status, use of antihypercholesterolemia medications, use of antihypertensive medications, history of coronary heart disease, myocardial infarction, heart failure, and stroke

Fig. 3

RCS Analysis of the Relationship Between the TyG Index and Cardiovascular Mortality. Hazard ratios were adjusted for gender, age, race or ethnicity, educational level, smoking status, alcohol use, family income-to-poverty ratio, estimated glomerular filtration rate, total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, hypertension status, high cholesterol status, use of antihypercholesterolemia medications, use of antihypertensive medications, history of coronary heart disease, myocardial infarction, heart failure, and stroke