Background: The association between the triglyceride-glucose (TyG) index and mortality in cardiovascular disease (CVD) patients with diabetes or pre-diabetes remains unclear. This study aimed to investigate the relationship between baseline TyG index and all-cause and cardiovascular (CV) mortality in CVD patients with diabetes or pre-diabetes among American adults.. Methods: This study enrolled 1072 CVD patients with diabetes or pre-diabetes from the National Health and Nutrition Examination Survey (2001–2018). Mortality outcomes were determined by linking to National Death Index (NDI) records up to December 31, 2019. Multivariate Cox proportional hazards models were constructed to analyze explore the associations between baseline TyG index and mortality. Non-linear correlations were explored using restricted cubic splines, and a two-piecewise Cox proportional hazards model for both sides of the inflection point was constructed. Results: During 7541 person-years of follow-up, a total of 461 all-cause deaths and 154 CVD-related deaths were recorded. The restricted cubic splines revealed a U-shaped association between the baseline TyG index with all-cause and CVD mortality in CVD patients with diabetes or pre-diabetes. Specifically, baseline TyG index lower than the threshold values (TyG index < 9.05 in all-cause mortality and < 8.84 in CVD mortality) was negatively associated with mortality (HR 0.47, 95% CI = 0.27–0.81 for all-cause mortality and HR 0.25, 95% CI = 0.07–0.89 for CVD mortality). In contrast, baseline TyG index higher than the threshold values (TyG index > 9.05 in all-cause mortality and > 8.84 in CVD mortality) was positively associated with mortality (HR 1.42, 95% CI = 1.02–1.99 for all-cause mortality and HR 1.77, 95% CI = 1.08–2.91 for CVD mortality). Conclusions: A U-shaped association was observed between the baseline TyG index with CVD and all-cause mortality in CVD patients with diabetes or pre-diabetes in a American population. The thresholds of 8.84 and 9.05 for CVD and all-cause mortality, respectively.
Keywords: Triglyceride-glucose index; Mortality; Cardiovascular Disease; NHANES
Qin Zhang and Shucai Xiao contributed equally to this study.
With the accelerated aging process, the morbidity and mortality of cardiovascular disease (CVD) continue to rise, and the high incidence of CVD is becoming an important public health problem [[
The aim of our study was to discover whether the TyG index has a prognostic value for the risk of all-cause and CVD mortality in in CVD patients with diabetes or pre-diabetes.
The National Health and Nutrition Examination Survey (NHANES) is a crucial research program that aims to evaluate the health and nutritional status of both adults and children residing in the United States. The Centers for Disease Control and Prevention (CDC) is responsible for furnishing health statistics for the nation, and the protocols of NHANES have been duly approved by the Research Ethics Review Board of NCHS. To ensure the protection of the participants' rights, NHANES has obtained informed written consent from all the individuals involved in the study. Moreover, the datasets generated and analyzed in the current study are readily available on the official NHANES website (https://
Graph: Fig. 1Flow chart of study participants
Information on various demographic and health-related factors was gathered, including age, sex, race/ethnicity, education level, family income, smoking status, disease status, and medication use, from NHANES household interviews. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Race/ethnicity was categorized as White, Black, Mexican, or Other, while education level was classified as less than high school, high school or equivalent, or college or above. Household income and poverty rate are divided into 0–1.0, 1.0–3.0, or > 3.0. Smoking status was recorded as never smoker, former smoker, or current smoker. Drinking status was categorized into heavy drinker (defined as consuming ≥ 3 drinks per day for females, ≥ 4 drinks per day for males, or binge drinking [≥ 4 drinks on the same occasion for females, ≥ 5 drinks on the same occasion for males] on 5 or more days per month), moderate drinker (defined as consuming ≥ 2 drinks per day for females, ≥ 3 drinks per day for males, or binge drinking ≥ 2 days per month), mild drinker (not meeting the above criteria), nondrinker, or a history of daily binge drinking. Clinical indicators such as fasting glucose, HbA1c, triglycerides (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) were measured in the NHANES laboratory.
The TyG index was calculated by TyG index = Ln [fasting TG (mg/dL) × fasting glucose (mg/dL)/2]. The measurement of triglycerides and fasting glucose were measured through enzymatic assays on Roche Modular P and Roche Cobas 6000 chemistry analyzers, respectively. The hexokinase-mediated reaction was utilized on Roche/Hitachi Cobas C 501 chemistry analyzers for measuring fasting glucose. The participants were classifed into four groups (Q1, Q2, Q3, Q4) by the quartiles of TyG index, and the Q1 group was used as the reference group.
In order to ascertain the mortality status in the follow-up population, we employed the NHANES public-use linked mortality file as of December 31, 2019. This file was linked with the National Death Index (NDI) by the National Center for Health Statistics (NCHS) via a probability matching algorithm. Moreover, we used the International Statistical Classification of Diseases, 10th Revision (ICD-10) to determine disease-specific deaths, with the NCHS classifying heart diseases (054–064), malignant neoplasms (019–043), and all other causes (010) for our study [[
The statistical analysis was performed using R software (version 4.2.1; https://
Table 1 showed the baseline characteristics of the cohort study participants (n = 1072) stratified by quartiles of the TyG index. The average age of the participants was 66.54 years, and 59.33% of them were male. Average TyG index in the enrolled patients was 9.03 ± 0.03. According to the quartiles of the TyG index, the laboratory characteristics at baseline are shown in Table 2. Participants with a higher TyG index were more likely to be younger, Mexican and obese, compared with participants in the lowest quartile. Additionally, significant differences in Biochemical indicators were observed between the groups, with participants in the highest quartile showing significantly higher levels of HbA1c, LDL-C, TC, TG, GGT, ALT, FINS, and FPG, compared with those in the first quartile.
Table 1 Baseline characteristics according to the TyG index quartiles
Characteristics Quartiles of TyG index overall Q1(6.80–8.56) Q2(8.56–8.96) Q3(8.96–9.40) Q4(9.40-12.55) 1072 268(25.00) 267(24.90) 269(25.10) 268(25.00) 66.54(0.45) 69.03(0.74) 66.89(0.83) 67.09(0.81) 63.49(0.86) <0.001 Male 436(40.78) 104(43.45) 108(42.98) 113(40.67) 111(36.38) 0.53 Female 636(59.22) 164(56.55) 159(57.02) 156(59.33) 157(63.62) <0.001 <24 119(10.93) 48(19.17) 30(11.91) 23(7.91) 18(5.95) ≥24 953(89.07) 220(80.83) 237(88.09) 246(92.09) 250(94.05) 0.002 Black 218(11.88) 83(18.95) 53(11.90) 43(8.84) 39(8.92) Mexican 126(4.83) 18(3.08) 30(4.66) 33(4.51) 45(6.81) White 590(74.22) 132(67.29) 142(71.72) 165(80.88) 151(75.92) Other 138(9.07) 35(10.68) 42(11.72) 28(5.76) 33(8.35) 0.48 High school grad or equivalent 275(28.83) 62(25.48) 65(28.50) 76(29.15) 72(31.71) Less than high school 363(24.97) 79(22.68) 88(23.27) 90(25.53) 106(28.08) Some college or above 434(46.20) 127(51.84) 114(48.24) 103(45.31) 90(40.21) 0.70 Former 344(30.12) 85(28.23) 86(30.26) 86(29.61) 87(32.12) Heavy 105(9.27) 21(7.88) 30(11.77) 23(6.78) 31(10.41) Mild 365(39.04) 101(39.71) 88(37.73) 90(41.26) 86(37.57) Moderate 85(7.39) 22(10.43) 20(6.72) 16(5.08) 27(7.85) Never 173(14.17) 39(13.75) 43(13.51) 54(17.27) 37(12.05) 0.52 Former 434(42.60) 109(42.28) 103(42.82) 111(40.44) 111(44.82) Never 446(39.11) 122(44.39) 110(35.84) 110(41.33) 104(35.69) Now 192(18.30) 37(13.32) 54(21.34) 48(18.23) 53(19.49) 2.54(0.06) 2.50(0.13) 2.46(0.14) 2.60(0.12) 2.61(0.14) 0.86 0.80 No 181(17.34) 38(14.73) 54(19.07) 41(17.22) 48(17.92) Yes 891(82.66) 230(85.27) 213(80.93) 228(82.78) 220(82.08) 0.17 No 941(86.90) 235(87.84) 228(83.20) 231(84.57) 247(92.23) Yes 131(13.10) 33(12.16) 39(16.80) 38(15.43) 21(7.77) 744(66.16) 161(55.56) 168(57.13) 189(66.52) 226(84.09) <0.001 328(33.84) 107(44.44) 99(42.87) 80(33.48) 42(15.91) <0.001
Date are presented as mean (SD) or n (%);
Table 2 Baseline levels of laboratory characteristics according to the TyG index quartiles
Quartiles of TyG index Q1(6.80–8.56) Q2(8.56–8.96) Q3(8.97–9.40) Q4(9.40-12.55) 6.06(0.07) 6.11(0.10) 6.44(0.08) 7.52(0.15) <0.001 69.17(1.69) 74.08(1.74) 72.56(1.59) 75.24(1.70) 0.10 2.33(0.07) 2.62(0.06) 2.83(0.09) 2.66(0.09) < 0.001 1.53(0.04) 1.29(0.03) 1.20(0.02) 1.04(0.02) < 0.001 4.23(0.08) 4.50(0.07) 4.85(0.09) 5.14(0.10) < 0.001 0.80(0.02) 1.28(0.02) 1.79(0.03) 3.40(0.16) < 0.001 107.07(6.57) 92.32(2.65) 93.91(2.47) 94.67(2.78) 0.25 149.69(3.05) 141.46(2.69) 139.17(2.11) 134.52(2.81) 0.01 353.11(7.55) 363.86(6.64) 375.61(7.85) 371.84(8.92) 0.22 40.98(0.25) 41.50(0.23) 41.19(0.24) 41.18(0.23) 0.43 74.09(4.40) 106.28(9.51) 111.25(7.45) 143.24(9.77) < 0.001 6.10(0.07) 6.58(0.12) 7.20(0.12) 9.75(0.27) < 0.001 4.15(0.04) 4.14(0.03) 4.17(0.03) 4.18(0.03) 0.67 15.09(0.51) 15.69(0.47) 14.95(0.36) 15.19(0.41) 0.61 139.71(0.23) 139.46(0.21) 139.32(0.21) 138.47(0.28) 0.01 6.67(0.23) 6.18(0.20) 6.20(0.21) 6.69(0.24) 0.12 27.50(2.36) 26.02(0.69) 26.36(1.16) 25.58(0.58) 0.79 22.28(1.04) 24.62(0.96) 25.75(1.28) 27.08(0.80) 0.004 13.20(0.45) 12.96(0.49) 12.63(0.30) 12.11(0.39) 0.31 29.63(2.08) 39.31(4.13) 39.54(5.50) 43.76(3.35) 0.001
Date are presented as mean (SD) or n (%);
Table 3 presents the occurrence of 461 all-cause deaths and 154 CVD-related deaths during the follow-up period. We constructed three Cox regression models to investigate the independent association between TyG index levels and mortality risk. After adjusting for age, gender, race, BMI, tobacco use, alcohol use, education, hypertension, and family income-poverty ratio in Model 3, the multivariate-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) from lowest to highest TyG index quartile (6.80–8.56, 8.56–8.96, 8.96–9.40, and 9.40-12.55) were 1.00 (reference), 0.65 (0.45, 0.93), 0.84 (0.62, 1.14), and 1.3 (0.73, 1.26), respectively, for all-cause mortality (P = 0.623); 1.00 (reference), 0.49 (0.25, 0.97), 1.10 (0.63, 1.90), and 1.30 (0.73, 2.33), respectively, for CVD mortality (P = 0.068).
Table 3 HRs (95% CIs) for mortality according to the TyG index quartiles
Quartiles of TyG index Q1(6.80–8.56) Q2(8.56–8.96) Q3(8.96–9.40) Q4(9.40-12.55) Number of deaths 119 104 122 116 Model 1 1 0.61(0.42,0.89) 0.01 0.77(0.57,1.03) 0.08 0.76(0.57,1.01) 0.06 0.29 h(95%CI) P-value Model 2 1 0.66(0.46,0.95) 0.02 0.86(0.64,1.17) 0.34 1.00(0.75,1.32) 0.99 0.46 h(95%CI) P-value Model 3 1 0.63(0.44,0.91) 0.01 0.84(0.62,1.13) 0.25 1.01(0.77,1.32) 0.94 0.36 h(95%CI) P-value Number of deaths(%) 40 24 47 43 Model 1 1 0.35(0.17,0.73) 0.01 0.84(0.50,1.40) 0.05 0.88(0.51,1.51) 0.64 0.49 h(95% CI) P-value Model 2 1 0.37(0.18,0.75) 0.01 0.96(0.57,1.62) 0.89 1.12(0.65,1.96) 0.68 0.14 h(95% CI) P-value Model 3 1 0.37(0.18,0.78) 0.01 1.13(0.65,1.96) 0.67 1.28(0.71,2.30) 0.41 0.04 h(95% CI) P-value
Model 1: Non-adjusted Model 2: Adjusted for age, race and gender Model 3: Adjusted for age, gender, race, BMI, tobacco use, alcohol use, education, hypertension, COPD, family income-poverty ratio. HR: Hazard ratio; CI: Confidence interval
Due to previous multivariate analysis indicated a non-linear relationship between baseline TyG index and the risk of all-cause and CVD mortality, we employed a Cox proportional hazards regression models with restricted cubic splines and smooth curve fitting (penalized spline method) to further investigate the correlation. Interestingly, the adjusted smoothed plots displayed U-shaped associations between TyG index and all-cause (Fig. 2A) and CVD mortality (Fig. 2B). We fitted the association between baseline TyG index and mortality using the standard Cox proportional hazards regression models and the the two-piecewise Cox proportional hazards regression models. Based on the two-piecewise Cox proportional hazards regression models, we identified the inflection points for all-cause and CVD mortality as 9.05 and 8.84, respectively (both P values for log-likelihood ratio < 0.05) (Table 4). After adjusting for age, gender, race, BMI, tobacco use, alcohol use, education, hypertension, and family income-poverty ratio, the risk of all-cause and CVD mortality decreased by approximately 53% (HR 0.47, 95% CI = 0.27–0.81) and 75% (HR 0.25, 95% CI = 0.07–0.89), respectively, with each unit increase in the TyG index up to the inflection points. Furthermore, the risk of all-cause and CVD mortality decreased to the lowest level as the baseline TyG index increased up to the threshold values (Table 4; Fig. 2). Conversely, the baseline TyG index was significantly and positively associated with the risk of all-cause and CVD mortality when it exceeded 9.05 (HR 1.42, 95% CI = 1.02–1.99) and 8.84 (HR 1.77, 95% CI = 1.08–2.91), respectively. We studied the population of diabetes and pre-diabetes separately. The results showed that there was still a U-shaped relationship between the TyG index and the all-cause mortality and CVD mortality in CVD patients with diabetes (both P for non-linear < 0.01). The TyG index and the all-cause mortality in CVD patients with pre-diabetes approximate a U-shaped relationship (P for non-linear = 0.28), while there was an approximate linear relationship with CVD mortality (P for non-linear = 0.95) Fig. 3.
Graph: Fig. 2Association between TyG index and all-cause (A) and CVD mortality (B) in CVD patients with diabetes or pre-diabetes. Each hazard ratio was computed with a TyG index level of A 9.05 and B 8.84 as the reference. Adjusted for age, gender, race, BMI, tobacco use, alcohol use, education, hypertension, family income-poverty ratio. The solid line and red area represent the estimated values and their corresponding 95% CIs, respectively (TyG index: triglyceride-glucose index; CVD: cardiovascular disease)
Table 4 Threshold effect analysis of TyG index on all-cause and CVD mortality in CVD patients with diabetes or pre-diabetes
Adjusted HR (95% CI), Total 1.12(0.92,1.35) 0.25 Fitting by two-piecewise Cox proportional risk model Inflection point 9.05 TyG index < 9.05 0.47(0.27,0.81) 0.01 TyG index ≥ 9.05 1.42(1.02,1.99) 0.04 < 0.001 Total 1.53(1.08,2.18) 0.02 Fitting by two-piecewise Cox proportional risk model Inflection point 8.84 TyG index < 8.84 0.25(0.07,0.89) 0.03 TyG index ≥ 8.84 1.77(1.08,2.91) 0.02 < 0.001
Cox proportional hazards models were used to estimate HR and 95% CI. Adjusted for age, gender, race, BMI, tobacco use, alcohol use, education, hypertension, COPD, family income-poverty ratio. HR: Hazard ratio; CI: Confidence interval
Graph: Fig. 3Association between TyG index and all-cause (A) and CVD mortality (B) in CVD patients with diabetes. Each hazard ratio was computed with a TyG index level of A 9.08 and B 9.08 as the reference. Association between TyG index and all-cause (C) and CVD mortality (D) in CVD patients with pre-diabetes. Each hazard ratio was computed with a TyG index level of A 8.97 and B 8.85 as the reference. Adjusted for age, gender, race, BMI, tobacco use, alcohol use, education, hypertension, family income-poverty ratio. The solid line and red area represent the estimated values and their corresponding 95% CIs, respectively (TyG index: triglyceride-glucose index; CVD: cardiovascular disease)
The survival advantage of a higher TyG index (≥ 9.05 for all-cause mortality and ≥ 8.84 for CVD mortality) compared to a lower TyG index (< 9.05 for all-cause mortality and <8.84 for CVD mortality9.05) among CVD patients with diabetes or pre-diabetes was consistent across various subgroups based on age, gender, race, and BMI, as depicted in Tables 5 and 6. There was no significant interaction between the baseline TyG index and stratified variables.
Table 5 Stratified analyses of the associations between TyG and All mortality
All-cause mortality HR(95% CI) P-value TyG index < 9.05 >=9.05 1 1.18(0.95,1.48) 0.13 0.53 Female 1 1.26(0.83,1.91) 0.28 Male 1 1.89(1.16,3.07) 0.01 0.12 < 60 1 1.29(0.69,2.44) 0.42 >=60 1 0.98(0.78,1.24) 0.86 0.99 <24 1 0.94(0.56,1.57) 0.82 >=24 1 1.20(0.95,1.51) 0.13 0.78 Black 1 1.51(0.98,2.31) 0.06 Mexican 1 0.93(0.43,2.00) 0.85 Other 1 0.90(0.34, 2.38) 0.83 White 1 1.16(0.89,1.50) 0.27 0.59 T2DM 1 1.15(0.90,1.46) 0.27 Pre-dm 1 1.04(0.64,1.64) 0.85
Table 6 Stratified analyses of the associations between TyG and CVD mortality
CVD mortality HR(95% CI) P-value TyG index < 8.84 >=8.84 1 1.70(1.16,2.50) 0.01 0.61 Female 1 1.55(0.77, 3.12) 0.22 Male 1 1.89(1.16,3.07) 0.01 0.39 < 60 1 2.12(0.45,9.98) 0.34 >=60 1 1.18(0.76,1.81) 0.46 0.14 <24 1 0.56(0.17,1.85) 0.34 >=24 1 2.10(1.30,3.38) 0.002 0.74 Black 1 1.91(0.93, 3.93) 0.08 Mexican 1 2.93(0.70,12.23) 0.14 Other 1 9.05(0.69, 118.85) 0.09 White 1 1.59(0.97,2.91) 0.06 0.31 T2DM 1 1.33(0.87,2.01) 0.18 Pre-dm 1 2.85(1.28,6.33) 0.01
To our knowledge, this is the first study to revealed a U-shaped association between the baseline TyG index with all-cause and CVD mortality among CVD patients with diabetes or pre-diabetes. We revealed a turning point (9.05 in all-cause mortality and 8.84 in CVD mortality) using threshold effect analysis. Our study shows that the TyG index is a strong predictor of all-cause and cardiovascular mortality among CVD patients with diabetes or pre-diabetes.
Previous clinical studies have also explored the association between the TyG index with CVD morbidity and mortality in various patient groups and the general population. A study conducted by Zhai [[
While the exact biological mechanisms underlying the correlation between the TyG index and mortality remain unclear, possible key pathways may be related to IR, a state of reduced sensitivity and responsiveness to insulin action. IR as a risk factor for CVD, can leads to the development of CVD in both the general population and diabetes patients and predicts cardiovascular prognosis in patients with CVD [[
Interestingly, lower levels of baseline TyG index (TyG index < 9.05 for all-cause mortality and < 8.84 for CVD mortality) significantly altered the correlation between TyG index and risk of all-cause and CVD mortality. After adjusting for confounding factors, each unit increase in baseline TyG index decreases the risk of all-cause and CVD mortality by almost 53% and 75%, respectively, in participants whose baseline TyG index is below the threshold. There were evidences that extremely low levels of TG or FPG are associated with adverse effects on health and may contribute to the development of diseases [[
The results of our study indicate that the TyG index is a valuable tool for predicting the risk of all-cause and CVD mortality in CVD patients with diabetes or pre-diabetes, and that the association between the TyG index and mortality is non-linear. Therefore, measuring the TyG index may be beneficial in assessing the risk and predicting the prognosis of this patient population. Future studies should explore whether interventions targeting the TyG index could lead to improved clinical outcomes in these patients.
The authors express their gratitude to the participants and staff of the NHANES for their invaluable contributions to this study.
The study design was conceived by QZ, XJ, SX, and YS. QZ and SX organized the data, conducted the analyses, and wrote and edited the manuscript. QZ and YS contributed to the interpretation of the results, revision, and finalization of the manuscript. All authors have reviewed and approved the final version of the manuscript.
The National Key R&D Program of China, Synthetic Biology Research (No. 2019YFA0904500). The National Natural Science Foundation of China (No. 82160170).
The datasets that were used and evaluated in this study can be obtained from the corresponding author upon making a reasonable request.
The National Center for Health Statistics and Ethics Review Board approved the protocol for NHANES, and all participants provided written informed consent. The authors have disclosed no conflicts of interest.
The authors declare no competing interests.
• ACS
- Acute coronary syndrome
• ALT
- Alanine aminotransferase
• BMI
- Body mass index
• CDC
- The Centers for Disease Control and Prevention
• CHD
- Coronary heart disease
• CHF
- Congestive heart failure
• CI
- Confidence interval
• CV
- Cardiovascular
• CVD
- Cardiovascular disease
• DBP
- Diastolic blood pressure
• FBG
- Fasting blood glucose
• FINS
- Fasting insulin
• GGT
- γ-glutamyl transpeptadase
• HbA1c
- Glycosylated hemoglobin A1c
• HDL-C
- High-density lipoprotein cholesterol
• HF
- Heart failure
• HOMA-IR
- Homeostasis model assessment of insulin resistance
• HR
- Hazard ratio
• IR
- Insulin resistance
• LDL-C
- Low-density lipoprotein cholesterol
• MI
- Myocardial Infarction
• NCHS
- National Center for Health Statistics
• NDI
- National Death Index
• NHANES
- National Health and Nutrition Examination Survey
• SBP
- Systolic blood pressure
• SD
- Standard deviation
• T2DM
- Type 2 diabetes mellitus
• TC
- Total cholesterol
• TG
- Triglyceride
• TyG
- Triglyceride-glucose index
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
By Qin Zhang; Shucai Xiao; Xiaojuan Jiao and Yunfeng Shen
Reported by Author; Author; Author; Author