Liu MJ, Xiang SM, Hu XQ. Triglyceride glucose-body mass index is associated with cardiovascular outcomes and overall mortality in type-2 diabetes mellitus patients. World J Diabetes 2025; 16(8): 108839 [DOI: 10.4239/wjd.v16.i8.108839]
Corresponding Author of This Article
Xin-Qun Hu, MD, Chief Pharmacist, Professor, Department of Cardiovascular Medicine, The Second Xiangya Hospital, Central South University, No. 87 Xiangya Road, Changsha 410011, Hunan Province, China. huxinqun@csu.edu.cn
Research Domain of This Article
Endocrinology & Metabolism
Article-Type of This Article
Retrospective Cohort Study
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Mao-Jun Liu, Xin-Qun Hu, Department of Cardiovascular Medicine, The Second Xiangya Hospital, Central South University, Changsha 410011, Hunan Province, China
Sun-Min Xiang, Department of Hospital Infection Control, Xingsha District of Hunan Provincial People's Hospital (Changsha County People's Hospital), Changsha 410100, Hunan Province, China
Author contributions: Liu MJ and Xiang SM contribute equally to this study as co-first authors; the study was designed by Hu XQ and Liu MJ; the manuscript was prepared by Hu XQ, Liu MJ and Xiang SM; the draft was revised by Hu XQ, Liu MJ and Xiang SM; all authors read and approved the final manuscript.
Institutional review board statement: The study was exempt from ethical review and approval, as no additional institutional review board approval was necessary for the secondary analysis.
Informed consent statement: The data for our study are derived from the Action to Control Cardiovascular Risk in Diabetes (ACCORD). All participants completed informed consent forms before participating in the study.
Conflict-of-interest statement: The authors declare that they have no competing interests.
STROBE statement: The authors have read the STROBE Statement—checklist of items, and the manuscript was prepared and revised according to the STROBE Statement—checklist of items.
Data sharing statement: The original data are available from the NHLBI BioLINCC. Some or all of the data sets generated and/or analyzed in the current study are not publicly available, but may be obtained from the corresponding author upon reasonable request at huxinqun@csu.edu.cn.
Open Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Xin-Qun Hu, MD, Chief Pharmacist, Professor, Department of Cardiovascular Medicine, The Second Xiangya Hospital, Central South University, No. 87 Xiangya Road, Changsha 410011, Hunan Province, China. huxinqun@csu.edu.cn
Received: April 24, 2025 Revised: May 31, 2025 Accepted: July 4, 2025 Published online: August 15, 2025 Processing time: 112 Days and 15.2 Hours
Abstract
BACKGROUND
Although recent studies have reported the reliability of triglyceride-glucose index (TyG) as a biomarker for evaluation of insulin resistance, the research exploring the association of TyG-body mass index (BMI) with major cardiovascular outcomes and all-cause mortality among the type 2 diabetes mellitus (T2DM) population remains scarce.
AIM
To ascertain whether TyG-BMI correlates with cardiovascular outcomes and mortality risk in individuals diagnosed with T2DM.
METHODS
Data were drawn from the ACCORD trial and its subsequent follow-up (ACCORDION), which together included 10190 participants. To evaluate the associations of TyG-BMI with future cardiovascular outcomes and overall mortality risk among T2DM patients, Cox’s proportional-hazards model was employed. Non-linear associations and thresholds were investigated by performing restricted cubic spline regression, fitting smooth curves, and conducting piecewise regression. Additionally, we examined whether the above findings were robust through interaction and subgroup analyses. The robustness of results was further verified by a series of sensitivity assessments. Through the area under the receiver operating characteristic curve estimation, the incremental predictive power of TyG-BMI beyond conventional risk factors was also evaluated.
RESULTS
Applying multivariable Cox regression analysis, we identified a significant correlation of TyG-BMI with both cardiovascular outcomes and overall mortality. Specifically, after adjusting for multiple confounders, the cardiovascular mortality, congestive heart failure (CHF) and overall mortality risks in the highest TyG-BMI quartile were 1.74, 2.65 and 1.42 times greater, respectively, when compared to the lowest quartile. The analysis revealed that TyG-BMI was linked nonlinearly to both types of mortality, while a two-stage linear regression discerned inflection points at 252.77 for cardiovascular mortality and 245.90 for overall mortality. TyG-BMI was found to correlate positively with future CHF. Additionally, introducing TyG-BMI into conventional prediction models resulted in a substantial improvement of their predictability for cardiovascular outcomes and overall mortality when compared to incorporating TyG or BMI alone.
CONCLUSION
Among T2DM individuals, a higher TyG-BMI is linked strongly to elevated risks of cardiovascular outcomes and overall mortality, highlighting its utility as a predictive marker for these risks.
Core Tip: This study uncovers a novel association between triglyceride-glucose index (TyG)-body mass index (BMI) and both cardiovascular outcomes and overall mortality in patients with type 2 diabetes mellitus (T2DM), based on the high-risk Action to Control Cardiovascular Risk in Diabetes cohort. It also marks the first identification of TyG-BMI as a robust predictor of future congestive heart failure, providing fresh perspectives on the risk stratification for complex T2DM patients.
Citation: Liu MJ, Xiang SM, Hu XQ. Triglyceride glucose-body mass index is associated with cardiovascular outcomes and overall mortality in type-2 diabetes mellitus patients. World J Diabetes 2025; 16(8): 108839
The past few decades have witnessed the emergence of type 2 diabetes mellitus (T2DM) as a major public health concern worldwide, with its prevalence rising alongside aging populations. T2DM is now among the world’s most rapidly expanding chronic diseases[1,2]. The International Diabetes Federation’s 2021 report states that 537 million people globally live with T2DM, a figure projected to rise by 46% to 783 million by 2045[3]. T2DM imposes a considerable social and economic burden on patients and their families, with its economic impact estimated at more than double that for individuals without diabetes[4]. T2DM is intricately associated with cardiovascular disease (CVD), placing diabetics at a markedly elevated risk of developing CVD compared to their non-diabetic counterparts. CVD is responsible for nearly half of all fatalities among individuals with diabetes[5-7]. Despite progress in diabetes management that has led to a reduction in the incidence of CVD, it continues to be the primary cause of mortality among diabetic individuals, underscoring the critical nature of prevention efforts. Identifying cardiovascular risk factors early in this population seems essential, which helps prevent the diabetes progression and lower the T2DM-related mortality.
A condition known as insulin resistance (IR) is marked by diminished insulin sensitivity and response[8-10]. Higher levels of IR are significantly linked to CVD risk in both diabetic and non-diabetic individuals. In those with diabetes, IR contributes to the development of CVD[11,12]. Hence, long-term IR monitoring and management may be beneficial in lowering cardiovascular mortality within the T2DM population. While hyperinsulinemic-euglycemic clamp (HEC) is widely considered the gold standard for evaluating IR, its complexity and high cost restrict its application in small-scale studies and broader clinical settings. Despite the frequent employment of the Homeostasis Model Assessment for Insulin Resistance (HOMA-IR) for assessing IR, it has limited accuracy in patients undergoing insulin therapy[13,14]. Recently, triglyceride-glucose index (TyG) has gained widespread acceptance as an IR assessment tool[15]. Guerrero-Romero et al[16] reported a high association between the TyG and the HEC standard for diagnosing IR, demonstrating strong sensitivity and specificity in assessing individual IR cases. Recent studies further corroborate the clinical utility of TyG in forecasting cardiovascular risk across diabetic and non-diabetic individuals[17-21].
TyG index, computed using the formula ln [fasting triglycerides (mg/dL) × fasting glucose (mg/dL)/2], is well-acknowledged as a surrogate marker for IR. TyG-body mass index (BMI), the product of TyG index and BMI (a standard anthropometric measure of general adiposity) incorporating both metabolic and anthropometric parameters, has recently been proposed as a novel and practical indicator for predicting metabolic disorders and cardiovascular risk. Evidence has suggested the particular effectiveness of TyG-BMI in diagnosing IR[22-25]. Nevertheless, its prognostic value for cardiovascular events and overall mortality in T2DM individuals remains understudied and debated. Existing research has predominantly focused on the general T2DM population, which often consists of younger individuals with relatively fewer comorbidities. In contrast, the ACCORD study cohort provides a unique opportunity to assess TyG-BMI’s predictive capacity within a higher-risk demographic. This cohort includes middle-aged and elderly individuals with T2DM, characterized by poor glycemic control, elevated cardiovascular risk, and a high prevalence of complex comorbidities such as heart failure and proteinuria. These distinct clinical features allow for a more comprehensive exploration of TyG-BMI’s relevance in real-world, high-risk scenarios. Leveraging data from the ACCORD trial and its ACCORDION follow-up, the present study attempts to evaluate the correlation of TyG-BMI with multiple cardiovascular outcomes, including cardiovascular mortality, congestive heart failure (CHF), and overall mortality. Our findings may offer important insights into the clinical utility of TyG-BMI in risk stratification and long-term prognosis for individuals with advanced-stage T2DM.
MATERIALS AND METHODS
Study design and participants
The present study employed a post-hoc analysis, drawing data from the ACCORD/ACCORDION Trial (ClinicalTrials.gov No. NCT00000620). As a randomized, multicenter, dual 2 × 2 factorial trial, the principles, design, and key findings of the ACCORD have been published in detail previously[26,27]. The trial enrolled 10251 T2DM patients, whose average age was 62.2 years and median value of glycated hemoglobin (HbA1c) 8.1%. Among these participants, some had a documented CVD history, while others were classified as having an elevated cardiovascular risk. The trial was conducted between June 2001 and June 2009 across 77 United States and Canadian sites, with an average treatment and follow-up duration of approximately 5 years. From 2011 to 2014, participants who remained in the ACCORDION trial were tracked for an average period of 3.5 years via telephone follow-ups and periodic clinic visits. The Morbidity and Mortality Subcommittee systematically assessed cardiovascular and all-cause mortality for each patient, based on the medical data gathered during follow-up visits.
Data collection and results
Data on demographic and clinical factors were collected, including age, sex, education, ethnicity, treatments, medical records, health checkups, laboratory test results, and past medication use. Out of the 10251 participants, 61 lacked baseline TyG-BMI data, leading to a final analysis sample of 10190 individuals (Figure 1). The respective formulas for calculating TyG and TyG-BMI were: TyG = ln[triglycerides (TG; mg/dL) × fasting blood glucose (FBG; mg/dL)/2]; BMI = weight/height2 (kg/m2); TyG-BMI = ln[TG (mg/dL) × FBG (mg/dL)/2] × weight/height2 (kg/m2).
Figure 1 Study participant selection from the ACCORD/ACCORDION study.
TyG: Triglyceride glucose; BMI: Body mass index.
The primary and secondary endpoints of this study were the severity of cardiovascular outcomes (including cardiovascular mortality and CHF), and the overall mortality.
Statistical analysis
SPSS 26.0 (IBM, Armonk, NY, United States), R software (The R Foundation, Vienna, Austria), combined with EmpowerStats (X&Y Solutions, Inc., Boston, MA, United States) were employed to statistically analyze the data in this study. According to the variable distributions, baseline traits were presented as means ± SD, frequencies and percentages, or medians and interquartile ranges. Continuous variables were analyzed via the Kruskal-Wallis test or ANOVA, while categorical data were assessed using the χ2 test. Kaplan-Meier plots were also utilized for evaluating cumulative risks for both cardiovascular and overall mortality. Log-rank test was conducted to investigate the inter-group differences.
By applying the Cox’s proportional-hazards models, the correlations of TyG-BMI with cardiovascular outcomes and overall mortality were assessed. Prior to performing the multivariate adjusted Cox proportional hazards regression analysis, an association assessment between all collected variables and future cardiovascular mortality rates was conducted through univariate analysis to identify possible confounding factors. Every covariate was included based on a sound rationale or clinical significance. Variables deemed statistically significant in the univariate analysis (P < 0.05, Supplementary Table 1) were subsequently incorporated into the multivariate analysis. Additionally, we also incorporated variables that were closely related to CVD mortality and clinically significant, along with variables exhibiting estimated effect changes greater than 10% (even if P > 0.05 in the univariate procedure), into the multivariate analysis. Throughout each analysis, the potential confounding factors related to cardiovascular mortality were controlled for by employing one unadjusted model in conjunction with three varying degrees of stepwise adjusted models. Model 1 (minimally adjusted): Considerations included factors such as sex, race, age, educational level, history of CVD, previous hyperlipidemia, previous hypertension, as well as diastolic blood pressures (DBPs), systolic blood pressures (SBPs). Model 2 (primarily adjusted): Building upon model 1, further covariates, including diabetes duration, proteinuria, status of smoking, cardiac failure, HbA1c, estimated glomerular filtration rate (eGFR), total cholesterol (TC), and low-density lipoprotein-cholesterol (LDL-C), were incorporated. Model 3 (fully adjusted): Expanded upon model 2, additional covariates encompassing the use of beta-blockers, calcium channel blockers (CCBs), thiazolidinediones, biguanides, insulin, and aspirin were introduced. Moreover, adjustments were made for the use of cholesterol absorption blockers and statins. To ensure the analytical accuracy, the adjusted variables were subjected to multicollinearity tests. A variance inflation factor (VIF) exceeding 5 or a tolerance level below 0.1 typically indicated the presence of collinearity issues. Our results indicated that the VIF of each variable remained below 5, with a tolerance exceeding 0.1 (Supplementary Table 2), suggesting no significant multicollinearity. Additionally, nonlinear correlations were explored and thresholds were determined by conducting restricted cubic spline (RCS) regression, fitting smooth curves, and performing piecewise regression. The robustness of the results was further evaluated through grouping and interaction analysis. Finally, the additional predictive value of TyG-BMI beyond conventional risk factors was examined with the area under the receiver-operating characteristic (ROC) curve (AUC).
A range of sensitivity analyses were conducted to ascertain the robustness of our findings. Initially, we performed multiple imputation to address missing data and reduce the potential bias associated with incomplete covariate information. Subsequently, participants with a baseline history of CVD were excluded to minimize confounding effects from pre-existing vascular conditions. We also removed those with a history of heart failure at baseline, given its strong association with adverse cardiovascular outcomes. Lastly, to mitigate reverse causality and early mortality bias, we eliminated participants who either experienced cardiovascular events or passed away within the initial 3-year period of follow-up.
RESULTS
Stratification of baseline characteristics by TyG-BMI quartile
The baseline profile comprised 10190 patients, with 61.47% being males 62.81 ± 6.64 years in mean age. The mean baseline levels were as follows: TC: 183.30 ± 41.85, LDL-C: 104.90 ± 33.91, TG: 190.06 ± 148.27, fasting plasma glucose (FPG): 175.19 ± 56.18, TyG: 9.49 ± 0.73, and TyG-BMI: 306.18 ± 58.74. Depending on the quartiles of baseline TyG-BMI, we stratified the patients as follows: Quartile 1 (234.32 ± 21.93), Quartile 2 (282.91 ± 11.31), Quartile 3 (322.94 ± 12.37), and Quartile 4 (384.54 ± 30.17). Table 1 outlines the baseline clinical, and demographic traits, along with laboratory indicators for the four groups. We identified significant differences among the groups in the following aspects: Age, BMI, sex, race, living conditions, educational level, proteinuria, cardiac failure, depression, alcohol consumption, status of smoking, diabetes duration, previous hypertension, DBP, HR, FPG, HbA1c, TC, TG, high density lipoprotein, glomerular filtration rate, CCB, angiotensin receptor blocker, angiotensin converting enzyme inhibitors, beta-blockers, thiazolidinediones, insulins, biguanides, statins, sulfonylureas, and aspirin.
Table 1 Baseline characteristics of participants across triglyceride glucose-body mass index quartiles.
Q1 (n = 2548; 131.06-263.12)
Q2 (n = 2547; 263.12-302.56)
Q3 (n = 2547; 302.56-345.60)
Q4 (n = 2548; 345.60-536.15)
P value
TyG-BMI
234.32 ± 21.93
282.91 ± 11.31
322.94 ± 12.37
384.54 ± 30.17
< 0.001
Age (years)
64.32 ± 7.20
63.44 ± 6.57
62.37 ± 6.41
61.13 ± 5.89
< 0.001
Sex
< 0.001
Male
1697 (66.60)
1684 (66.12)
1545 (60.66)
1338 (52.51)
Female
851 (33.40)
863 (33.88)
1002 (39.34)
1210 (47.49)
Race
< 0.001
White
1197 (46.98)
1533 (60.19)
1704 (66.90)
1933 (75.86)
Non-white
1351 (53.02)
1014 (39.81)
843 (33.10)
615 (24.14)
Education
< 0.001
Less than high school graduate
423 (16.62)
418 (16.41)
359 (14.11)
301 (11.82)
High school grad (or GED)
689 (27.07)
640 (25.13)
683 (26.84)
678 (26.63)
Some college or technical school
698 (27.43)
807 (31.68)
872 (34.26)
964 (37.86)
College graduate or more
735 (28.88)
682 (26.78)
631 (24.79)
603 (23.68)
CVD history
885 (34.73)
929 (36.47)
905 (35.53)
863 (33.87)
0.246
Duration of diabetes (years)
12.12 ± 8.16
10.96 ± 7.64
10.19 ± 7.19
9.93 ± 7.15
< 0.001
Previous hypertension
1837 (72.10)
1901 (74.64)
1943 (76.29)
1998 (78.41)
< 0.001
Previous hyperlipidemia
1742 (68.37)
1792 (70.36)
1776 (69.73)
1816 (71.27)
0.144
Proteinuria
433 (16.99)
472 (18.53)
525 (20.62)
597 (23.43)
< 0.001
Heart failure
93 (3.65)
113 (4.44)
102 (4.01)
179 (7.03)
< 0.001
Depression
394 (15.46)
545 (21.40)
612 (24.05)
861 (33.79)
< 0.001
Living alone
2082 (81.71)
2033 (79.82)
2025 (79.57)
1985 (77.90)
0.009
Smoking
< 0.001
Yes
1394 (54.71)
1512 (59.36)
1532 (60.15)
1495 (58.67)
No
1154 (45.29)
1035 (40.64)
1015 (39.85)
1053 (41.33)
Alcohol
< 0.001
Yes
630 (24.73)
664 (26.08)
623 (24.49)
515 (20.22)
No
1918 (75.27)
1882 (73.92)
1921 (75.51)
2032 (79.78)
BMI (kg/m2)
26.08 ± 2.47
30.21 ± 2.23
33.76 ± 2.56
38.86 ± 3.30
< 0.001
SBP (mmHg)
136.45 ± 17.19
136.34 ± 17.01
136.47 ± 16.97
136.13 ± 17.28
0.905
DBP (mmHg)
72.71 ± 10.55
74.41 ± 10.53
75.72 ± 10.42
76.70 ± 10.73
< 0.001
Heart rate (bpm)
71.07 ± 11.54
71.71 ± 11.68
73.40 ± 11.70
74.52 ± 11.78
< 0.001
FPG (mg/dL)
157.67 ± 54.17
170.58 ± 53.78
178.59 ± 53.77
193.92 ± 56.75
< 0.001
HbA1c (%)
8.21 ± 1.07
8.26 ± 1.05
8.29 ± 1.01
8.45 ± 1.09
< 0.001
TC (mg/dL)
174.85 ± 38.80
180.92 ± 40.15
183.96 ± 41.12
193.46 ± 44.91
< 0.001
TG (mg/dL)
124.30 ± 73.92
170.07 ± 99.22
202.26 ± 129.90
263.60 ± 213.27
< 0.001
VLDL-C (mg/dL)
24.67 ± 13.68
33.44 ± 18.32
39.08 ± 22.08
48.92 ± 32.34
< 0.001
LDL-C (mg/dL)
104.50 ± 32.86
105.44 ± 33.44
104.23 ± 33.84
105.43 ± 35.45
0.558
HDL-C (mg/dL)
45.68 ± 12.99
42.04 ± 11.33
40.65 ± 10.81
39.12 ± 10.17
< 0.001
eGFR (mL/min/1.73 m2)
91.07 ± 29.02
90.32 ± 28.28
91.68 ± 25.43
91.07 ± 25.62
0.041
Medications
ARB/ACEI
1672 (65.62)
1762 (69.18)
1794 (70.44)
1834 (71.98)
< 0.001
CCB
472 (18.52)
452 (17.75)
486 (19.08)
541 (21.23)
0.012
β-blockers
658 (25.89)
729 (28.70)
815 (32.09)
862 (33.92)
< 0.001
Biguanides
1561 (61.26)
1653 (64.90)
1666 (65.44)
1638 (64.29)
0.009
Thiazolidinediones
496 (19.47)
512 (20.10)
590 (23.17)
648 (25.43)
< 0.001
Sulfonylureas
1413 (55.46)
1374 (53.95)
1353 (53.14)
1305 (51.22)
0.023
Insulins
818 (32.10)
853 (33.49)
902 (35.41)
988 (38.78)
< 0.001
Statins
1642 (64.75)
1656 (65.33)
1636 (64.41)
1531 (60.28)
< 0.001
Aspirin
1351 (53.29)
1375 (54.26)
1451 (57.22)
1371 (54.02)
0.027
Cholesterol absorption inhibitors
45 (1.78)
58 (2.29)
50 (1.97)
54 (2.13)
0.608
Association of TyG-BMI with cardiovascular outcomes and overall mortality among T2DM individuals
During a median follow-up period of 8.82 years, 662 patients with T2DM (6.50%) died from cardiovascular causes, while 690 (6.77%) patients died from CHF, contributing to an overall mortality count of 1942 (19.06%). Cumulative cardiovascular and overall mortality risks were determined by the Kaplan-Meier survival analysis, which revealed a link between elevated TyG-BMI and heightened cumulative risks for both parameters in the future (cardiovascular mortality: P = 0.0016; CHF: P < 0.0001; overall mortality: P = 0.019; Figure 2). Utilizing three multivariate regression models, the correlation of TyG-BMI with the incidence of outcome events was analyzed (Table 2). Model 1 was adjusted for gender, race, age, educational level, history of CVD, previous hypertension, previous hyperlipidemia, SBP, and DBP. Building upon Model 1, Model 2 further adjusted for the diabetes duration, proteinuria, cardiac failure, status of smoking, HbA1c, eGFR, TC, and LDL-C. Model 3, expanded upon Model 2, additionally adjusted for the use of such medications as CCB, beta-blockers, thiazolidinediones, cholesterol absorption blockers, insulins, biguanides, statins, and aspirin.
Figure 2 Kaplan-Meier curves for cardiovascular outcomes and total mortality with triglyceride glucose-body mass index.
A: Cardiovascular mortality; B: Congestive heart failure; C: Total mortality. TyG: Triglyceride glucose; BMI: Body mass index; CHF: Congestive heart failure.
Table 2 Risk of cardiovascular outcomes and total mortality based on triglyceride glucose, body mass index and triglyceride glucose-body mass index.
Outcome
Non-adjusted
Model 1
Model 2
Model 3
HR (95%CI)
P value
HR (95%CI)
P value
HR (95%CI)
P value
HR (95%CI)
P value
Cardiovascular mortality
TyG-BMI
Q1
Reference
Reference
Reference
Reference
Q2
1.03 (0.82, 1.30)
0.7775
1.06 (0.84, 1.34)
0.6247
1.05 (0.83, 1.33)
0.6576
1.04 (0.82, 1.32)
0.7537
Q3
1.23 (0.98, 1.53)
0.0710
1.39 (1.10, 1.74)
0.0050
1.39 (1.10, 1.75)
0.0056
1.36 (1.07, 1.71)
0.0108
Q4
1.45 (1.17, 1.80)
0.0007
1.93 (1.53, 2.42)
< 0.0001
1.81 (1.42, 2.31)
< 0.0001
1.74 (1.36, 2.22)
< 0.0001
Per 1 SD
1.14 (1.05, 1.23)
0.0009
1.27 (1.17, 1.38)
< 0.0001
1.24 (1.13, 1.35)
< 0.0001
1.22 (1.11, 1.33)
< 0.0001
P for trend
0.0002
< 0.0001
< 0.0001
< 0.0001
TyG
Q1
Reference
Reference
Reference
Reference
Q2
0.90 (0.71, 1.14)
0.3720
0.89 (0.71, 1.13)
0.3514
0.90 (0.71, 1.14)
0.3819
0.89 (0.70, 1.13)
0.3308
Q3
1.19 (0.96, 1.48)
0.1156
1.21 (0.96, 1.51)
0.1015
1.17 (0.93, 1.49)
0.1859
1.19 (0.93, 1.51)
0.1608
Q4
1.37 (1.11, 1.70)
0.0033
1.46 (1.17, 1.82)
0.0009
1.48 (1.12, 1.96)
0.0058
1.43 (1.08, 1.89)
0.0130
Per 1 SD
1.15 (1.07, 1.24)
0.0003
1.19 (1.09, 1.28)
< 0.0001
1.23 (1.09, 1.39)
0.0008
1.20 (1.07, 1.36)
0.0024
P for trend
0.0003
< 0.0001
0.0027
0.0046
BMI
Q1
Reference
Reference
Reference
Reference
Q2
1.02 (0.82, 1.28)
0.8488
1.03 (0.82, 1.30)
0.7766
1.04 (0.83, 1.31)
0.7508
1.02 (0.81, 1.29)
0.8433
Q3
1.15 (0.92, 1.43)
0.2107
1.30 (1.04, 1.63)
0.0201
1.30 (1.04, 1.64)
0.0213
1.29 (1.03, 1.63)
0.0266
Q4
1.31 (1.06, 1.62)
0.0143
1.69 (1.34, 2.11)
< 0.0001
1.61 (1.28, 2.02)
< 0.0001
1.56 (1.23, 1.97)
0.0002
Per 1 SD
1.08 (1.00, 1.17)
0.0403
1.20 (1.11, 1.31)
< 0.0001
1.18 (1.09, 1.29)
< 0.0001
1.17 (1.07, 1.27)
0.0003
P for trend
0.0073
< 0.0001
< 0.0001
< 0.0001
CHF
TyG-BMI
Q1
Reference
Reference
Reference
Reference
Q2
1.29 (1.02, 1.64)
0.0331
1.34 (1.05, 1.70)
0.0176
1.29 (1.01, 1.65)
0.0376
1.29 (1.01, 1.66)
0.0393
Q3
1.46 (1.16, 1.84)
0.0015
1.77 (1.40, 2.25)
< 0.0001
1.75 (1.37, 2.23)
< 0.0001
1.71 (1.34, 2.19)
< 0.0001
Q4
2.15 (1.73, 2.68)
< 0.0001
3.03 (2.40, 3.83)
< 0.0001
2.70 (2.12, 3.45)
< 0.0001
2.65 (2.07, 3.40)
< 0.0001
Per 1 SD
1.32 (1.23, 1.42)
< 0.0001
1.52 (1.41, 1.65)
< 0.0001
1.46 (1.34, 1.58)
< 0.0001
1.45 (1.33, 1.58)
< 0.0001
P for trend
< 0.0001
< 0.0001
< 0.0001
< 0.0001
TyG
Q1
Reference
Reference
Reference
Reference
Q2
1.22 (0.97, 1.52)
0.0855
1.21 (0.97, 1.52)
0.0902
1.22 (0.97, 1.53)
0.0829
1.23 (0.98, 1.54)
0.0788
Q3
1.13 (0.90, 1.42)
0.2804
1.16 (0.92, 1.47)
0.1966
1.10 (0.86, 1.40)
0.4420
1.12 (0.88, 1.43)
0.3721
Q4
1.56 (1.26, 1.93)
< 0.0001
1.76 (1.41, 2.20)
< 0.0001
1.62 (1.24, 2.13)
0.0004
1.61 (1.22, 2.12)
0.0007
Per 1 SD
1.18 (1.09, 1.27)
< 0.0001
1.25 (1.15, 1.35)
< 0.0001
1.26 (1.12, 1.42)
0.0001
1.26 (1.12, 1.41)
0.0002
P for trend
0.0002
< 0.0001
0.0035
0.0042
BMI
Q1
Reference
Reference
Reference
Reference
Q2
1.39 (1.10, 1.75)
0.0054
1.44 (1.14, 1.82)
0.0021
1.37 (1.09, 1.74)
0.0082
1.40 (1.10, 1.77)
0.0062
Q3
1.32 (1.05, 1.67)
0.0194
1.54 (1.21, 1.95)
0.0004
1.53 (1.20, 1.94)
0.0006
1.53 (1.20, 1.95)
0.0006
Q4
2.00 (1.61, 2.49)
< 0.0001
2.67 (2.12, 3.35)
< 0.0001
2.43 (1.92, 3.07)
< 0.0001
2.41 (1.89, 3.06)
< 0.0001
Per 1 SD
1.27 (1.18, 1.37)
< 0.0001
1.44 (1.33, 1.56)
< 0.0001
1.39 (1.28, 1.50)
< 0.0001
1.39 (1.28, 1.51)
< 0.0001
P for trend
< 0.0001
< 0.0001
< 0.0001
< 0.0001
Total mortality
TyG-BMI
Q1
Reference
Reference
Reference
Reference
Q2
1.04 (0.91, 1.18)
0.5610
1.09 (0.96, 1.24)
0.1740
1.07 (0.94, 1.22)
0.3110
1.06 (0.93, 1.21)
0.3553
Q3
1.01 (0.89, 1.15)
0.8545
1.18 (1.04, 1.35)
0.0131
1.16 (1.01, 1.32)
0.0337
1.13 (0.99, 1.30)
0.0708
Q4
1.19 (1.05, 1.35)
0.0059
1.63 (1.42, 1.86)
< 0.0001
1.48 (1.29, 1.70)
< 0.0001
1.42 (1.23, 1.64)
< 0.0001
Per 1 SD
1.05 (1.00, 1.10)
0.0316
1.19 (1.13, 1.25)
< 0.0001
1.14 (1.08, 1.20)
< 0.0001
1.12 (1.07, 1.18)
< 0.0001
P for trend
0.0123
< 0.0001
< 0.0001
< 0.0001
TyG
Q1
Reference
Reference
Reference
Reference
Q2
0.90 (0.79, 1.03)
0.1319
0.91 (0.80, 1.04)
0.1749
0.91 (0.80, 1.04)
0.1570
0.91 (0.80, 1.04)
0.1817
Q3
1.03 (0.91, 1.17)
0.6588
1.06 (0.93, 1.21)
0.3577
1.00 (0.87, 1.15)
0.9752
1.03 (0.89, 1.18)
0.7184
Q4
1.18 (1.05, 1.34)
0.0076
1.31 (1.15, 1.49)
< 0.0001
1.21 (1.03, 1.41)
0.0226
1.21 (1.03, 1.42)
0.0212
Per 1 SD
1.07 (1.02, 1.12)
0.0034
1.12 (1.07, 1.18)
< 0.0001
1.08 (1.01, 1.15)
0.0315
1.08 (1.01, 1.15)
0.0338
P for trend
0.0014
< 0.0001
0.0270
0.0185
BMI
Q1
Reference
Reference
Reference
Reference
Q2
1.03 (0.91, 1.17)
0.6020
1.08 (0.95, 1.22)
0.2541
1.06 (0.93, 1.21)
0.3746
1.05 (0.92, 1.20)
0.4544
Q3
0.96 (0.84, 1.09)
0.5209
1.10 (0.96, 1.25)
0.1550
1.09 (0.96, 1.25)
0.1869
1.07 (0.94, 1.23)
0.2924
Q4
1.11 (0.98, 1.25)
0.1122
1.47 (1.29, 1.68)
< 0.0001
1.39 (1.22, 1.59)
< 0.0001
1.34 (1.17, 1.53)
< 0.0001
Per 1 SD
1.02 (0.98, 1.07)
0.3059
1.15 (1.09, 1.20)
< 0.0001
1.12 (1.07, 1.18)
< 0.0001
1.11 (1.05, 1.16)
< 0.0001
P for trend
0.2515
< 0.0001
< 0.0001
< 0.0001
With TyG-BMI considered a continuous variable and after a comprehensive confounder adjustment, each 1-SD rise in TyG-BMI corresponded to a 22% elevated cardiovascular death risk (HR: 1.22, 95%CI: 1.11-1.33, P < 0.0001), a 45% higher risk of CHF (HR: 1.45, 95%CI: 1.33-1.58), and a 12% elevated risk of all-cause mortality (HR: 1.12, 95%CI: 1.07-1.18, P < 0.0001). In Model 3, following the potential confounder adjustment, TyG-BMI still correlated positively with the cumulative cardiovascular death risk (TyG-BMI Q2: HR: 1.04, 95%CI: 0.82-1.32, P = 0.7537; TyG-BMI Q3: HR: 1.36, 95%CI: 1.07-1.71, P = 0.0108; TyG-BMI Q4: HR: 1.74, 95%CI: 1.36-2.22, P < 0.0001). Moreover, baseline TyG-BMI exhibited significant associations with CHF (TyG-BMI Q2: HR: 1.29, 95%CI: 1.01-1.66, P = 0.0393; TyG-BMI Q3: HR: 1.71, 95%CI: 1.34-2.19, P < 0.0001; TyG-BMI Q4: HR: 2.65, 95%CI: 2.07-3.40, P < 0.0001) and overall mortality (TyG-BMI Q2: HR: 1.06, 95%CI: 0.93-1.21, P = 0.3553; TyG-BMI Q3: HR: 1.13, 95%CI: 0.99-1.30, P = 0.0708; TyG-BMI Q4: HR: 1.42, 95%CI: 1.23-1.64, P < 0.0001). An upward trend regarding the cardiovascular mortality, CHF and overall mortality risks was noted as TyG-BMI increased (cardiovascular mortality: P-trend < 0.0001; CHF: P-trend < 0.0001; overall mortality: P-trend < 0.0001). We also found that compared to TyG or BMI alone, TyG-BMI displayed more robust correlations with and hazard ratios concerning future cardiovascular outcomes and overall mortality risk in T2DM patients.
The study then evaluated a possible nonlinear association of TyG-BMI with cardiovascular and overall mortality risks among T2DM patients by the RCS analysis (cardiovascular mortality: P-overall < 0.001, P-nonlinear = 0.021; CHF: P-overall < 0.001, P-nonlinear > 0.050; overall mortality: P-overall < 0.001, P-nonlinear = 0.025; Figure 3). By fitting a smooth with a Cox’s model, a non-linear correlation of TyG-BMI with both mortality outcomes was confirmed. Log-likelihood test findings further supported this observation (P < 0.05). Additionally, a two-stage Cox’s model was utilized for evaluating the effects on either side of the inflection point along with their confidence intervals. Results suggested that beyond the inflection point, every 10-unit reduction in TyG-BMI corresponded to a 5% lowered cardiovascular mortality risk (HR: 1.05, 95%CI: 1.03-1.06). As for overall mortality, the HR values were 0.96 (95%CI: 0.92-1.00) and 1.03 (95%CI: 1.02-1.04) before and after the inflection point, respectively (Figure 3, Table 3).
Figure 3 Restricted cubic spline and smooth curve fitting analyses were performed to assess the relationship between triglyceride glucose-body mass index and the risk of cardiovascular outcomes and total mortality.
A and B: Cardiovascular mortality; C and D: Congestive heart failure; E and F: Total mortality. The hazard ratios shown are adjusted for Model 3, including sex, race, age, education, history of cardiovascular disease, previous hypertension, previous hyperlipidemia, diastolic blood pressure, systolic blood pressure, duration of diabetes, proteinuria, heart failure, smoking status, glycated hemoglobin, total cholesterol, low density lipoprotein-cholesterol, estimated glomerular filtration rate, calcium channel blocker, beta-blockers, biguanides, thiazolidinediones, insulin, aspirin, cholesterol absorption inhibitors. TyG: Triglyceride glucose; BMI: Body mass index; CHF: Congestive heart failure; CVD: Cardiovascular disease.
Table 3 Analysis of the threshold effects of triglyceride glucose-body mass index levels on cardiovascular and total mortality.
Outcomes
One linear-regression model
Inflection point (K)
< K, effect 1
> K, effect 2
P value for LRT
Cardiovascular mortality
1.03 (1.02, 1.05), P < 0.0001
252.77
0.95 (0.89, 1.01), P = 0.1161
1.05 (1.03, 1.06), P < 0.0001
0.012
Total mortality
1.02 (1.01, 1.03), P < 0.0001
245.90
0.96 (0.92, 1.00), P = 0.0302
1.03 (1.02, 1.04), P < 0.0001
0.002
Subgroup and interaction analyses
For the robustness examination of our findings and the exploration into association between TyG-BMI and outcome events, interaction and subgroup analyses were undertaken across varying age (< 65 years and ≥ 65 years), sex, race, CVD history, cardiac failure, prior hyperlipidemia, prior hypertension, diabetes duration (< 10 years and ≥ 10 years), and treatment subgroups (intensive vs standard glucose control). Further stratified analyses were also conducted based on insulin usage and HbA1c (< 8.1% vs ≥ 8.1%). Results indicated a significant interaction between TyG-BMI and the cardiovascular and overall mortality risks. Notably, in T2DM patients with HbA1c levels below 8.1%, TyG-BMI demonstrated a heightened predictive power for both mortality types (Figure 4).
Figure 4 Subgroup and interaction analyses were conducted to examine the relationship between triglyceride glucose-body mass index and the risks of cardiovascular outcomes and total mortality.
A: Cardiovascular mortality; B: Congestive heart failure; C: Total mortality. CVD: Cardiovascular disease; HbA1c: Glycated hemoglobin.
Sensitivity analysis
For the robustness verification of our results, we conducted a range of sensitivity analyses. Through multiple imputation, missing data were handled to minimize the potential impact of data incompleteness. The results of multivariable Cox regression analysis, conducted following the elimination of subjects with missing values, were consistent with those obtained after multiple imputation (Supplementary Table 3), reinforcing the reliability of our conclusions. Following the elimination of subjects with a baseline CVD history, the final study cohort comprised 6608 individuals. The correlation of TyG-BMI with cardiovascular outcomes and overall mortality remained largely consistent with the primary analysis after adjusting for the entire covariates (Supplementary Table 4). Similarly, when subjects with a baseline cardiac failure history were eliminated, the resulting cohort comprised 9703 individuals, and the above correlation remained comparable to the main findings (Supplementary Table 5). Furthermore, after omitting subjects from the initial 3-year follow-up period, a total of 9496 individuals were included, and the above correlation persisted (Supplementary Table 6). Collectively, these analytical results further validated the robustness of our findings.
The added predictive power of TyG-BMI for cardiovascular outcomes and overall mortality among T2DM individuals
Through the ROC curve approach, we assessed TyG-BMI’s predictive value for future cardiovascular and overall mortality risks. Results showed an increase in the AUC for cardiovascular outcomes and overall mortality after incorporating TyG-BMI levels into conventional models, which accounted for factors like age, sex, educational background, diabetes duration, smoking status, history of hypertension/hyperlipidemia, SBP, DBP, proteinuria, depression, FPG, HbA1c, and LDL-C. Suggestively, the introduction of TyG-BMI into the classic models substantially enhances predictive precision for future cardiovascular outcomes and overall mortality among T2DM individuals. The base model incorporating TyG-BMI exhibited superior predictive power to those considering TyG or BMI alone (Figure 5).
Figure 5 Receiver-operating characteristic analysis for cardiovascular outcomes and total mortality.
A: Cardiovascular mortality; B: Congestive heart failure; C: Total mortality. The conventional model + triglyceride glucose (TyG)-body mass index (BMI) included clinically recognized cardiovascular risk factors (sex, age, education, smoking status, duration of diabetes, previous hypertension, previous hyperlipidemia, diastolic blood pressure, systolic blood pressure, proteinuria, depression, fasting plasma glucose, glycated hemoglobin, low density lipoprotein-cholesterol) along with TyG-BMI levels, while the conventional model included only the standard cardiovascular risk factors. TyG: Triglyceride glucose; BMI: Body mass index; CHF: Congestive heart failure; AUC: Area under the receiver-operating characteristic curve.
DISCUSSION
This study conducted an initial investigation into the link of TyG-BMI to both cardiovascular outcomes and overall mortality among the T2DM population. The ACCORD/ACCORDION trial is a broad, population-based cohort study. As far as we are aware, this is the first research within this cohort to assess the correlations of TyG-BMI with cardiovascular outcomes and overall mortality. Our results indicate that TyG-BMI serves as a valid predictor for both overall and cardiovascular mortality risks in T2DM individuals. After comprehensive adjustment for confounding factors, we found that TyG-BMI elevations were independently linked to heightened risks of cardiovascular outcomes and overall mortality for this population. These findings may contribute to novel understandings about the prevention and management of cardiovascular and overall mortality in T2DM patients, serving as a reference for developing targeted strategies.
CVD continues to pose a major global health challenge, placing substantial strain on healthcare systems and families worldwide[28,29]. Diabetes has been acknowledged as a salient risk factor for CVD, with its poor outcomes being strongly linked to cardiovascular complications[30,31]. The median follow-up in our study spanned 8.82 years, with 6.50% of T2DM patients dying from cardiovascular causes, 6.77% dying from CHF, and 19.06% from any cause. The elevated incidence rates observed, relative to prior studies, may stem from the fact that the majority of ACCORD and ACCORDION participants were older, at a high cardiovascular risk, and exhibited poor glycemic control.
Prior research has shown that IR is pivotal in the development of atherosclerosis and recognized it as a major risk factor for CVD. Numerous investigations have emphasized the link between IR and CVD progression, underscoring the predictive value of IR for cardiovascular outcomes[10,11,32,33]. While the HEC is generally considered the gold standard for evaluating IR, its complexity and cost hinder its broader clinical application. Hence, it is crucial to identify a dependable and simpler surrogate indicator for the IR assessment.
Recently, TyG-BMI, a new and practical indicator combining TG, FBG, and BMI, has been recognized for its strong diagnostic value in detecting IR[22-25]. Its predictive power for clinical outcomes has been substantiated. For example, evidence suggests that TyG-BMI, either independently or in combination with sedentary behavior and physical activity, is associated with the risks of all-cause mortality and myocardial infarction[34]. In heart failure patients, a decreased TyG-BMI strongly correlated with an elevated 1-year mortality risk[35]. Among patients sustaining severe acute myocardial infarction, TyG-BMI was significantly correlated with overall mortality[36]. Among hypertensive patients, the incorporation of TyG-BMI into the basic predictive model significantly improved the cardiovascular mortality predictability[37]. As suggested by these results, TyG-BMI serves as a useful prognostic biomarker.
In our current study, the TyG-BMI index demonstrated a similar predictability for both cardiovascular and overall mortality among diabetic subjects, establishing a connection between them among T2DM individuals within the ACCORD trial cohort for the first time. In a cohort comprising 10090 patients with T2DM, our findings indicated that a higher baseline TyG-BMI correlated with greater future risks of both cardiovascular and overall mortality. Such an association remained statistically noteworthy even following the key confounder adjustment. Patients whose baseline TyG-BMI fell within the highest quartile exhibited a 1.74-fold higher cardiovascular mortality risk, a 2.65-fold elevated CHF risk, and a 1.42-fold higher overall mortality risk, respectively, in contrast to those within the lowest quartile. The RCS analysis, along with smooth-curve fitting, unveiled a nonlinear correlation of baseline TyG-BMI with future risks of cardiovascular and overall mortality in T2DM patients, whereas a positive linear association of TyG-BMI with CHF risk. The identified inflection points for cardiovascular and overall mortality were 252.77 and 245.90, respectively. These results are consistent with prior research. For instance, Xiao et al[38], drawing on the National Health and Nutrition Examination Survey data, also reported a U-shaped correlation of TyG-BMI with both mortality types among diabetic individuals. They identified the inflection points at 279.67 for overall mortality, whereas at 270.19 for cardiovascular mortality, both significantly exceeding the corresponding thresholds identified in our analysis (245.90 and 252.77, respectively). This discrepancy may reflect the unique characteristics of the ACCORD cohort, which consisted of older adults with poor glycemic control and high cardiovascular risk. Our findings suggest that in high-risk, late-stage T2DM sufferers, more stringent control of TyG-BMI may be necessary to reduce the long-term mortality risk. This further highlights the clinical importance of TyG-BMI as a tailored risk stratification tool for complex diabetic cohorts. The implications of this result are significant in clinical practice, offering a reference point for the cardiovascular mortality prevention among T2DM patients. Decreasing BMI, TG, and FPG through dietary and lifestyle modifications may contribute to lowered cardiovascular and overall mortality risks, particularly when TyG-BMI levels are maintained below 245.90. Additional stratified and interaction analyses provided results that largely corroborated the initial findings. These analyses indicated that TyG-BMI showed stronger predictive power for both cardiovascular and overall mortality in T2DM individuals with HbA1c levels below 8.1%. This finding may reflect distinct pathophysiological mechanisms operating across different glycemic control statuses. In individuals with poorly managed diabetes (HbA1c ≥ 8.1%), the direct vascular toxicity from sustained hyperglycemia—manifesting as endothelial dysfunction, oxidative stress, and pro-inflammatory states—may act as a more dominant driver of adverse cardiovascular outcomes, potentially attenuating the independent contribution of TyG-BMI. Conversely, in patients with relatively better glycemic control (HbA1c < 8.1%), where overt glucose toxicity is less pronounced, IR may become a more central driver of cardiovascular risk. In this scenario, TyG-BMI potentially serves as a more sensitive and integrative biomarker of metabolic disturbance, reflecting both dyslipidemia and IR burden. The above findings lend support to the assumption that TyG-BMI is particularly useful in identifying high-risk individuals with subclinical metabolic dysfunction, even in the absence of severe hyperglycemia, highlighting its potential clinical utility in early risk stratification. We also conducted a range of sensitivity analyses, yielding results that agreed with the main findings, thus confirming the robustness of our results.
As implied by the results of this study, TyG-BMI is a vital predictor for both cardiovascular outcomes and overall mortality among T2DM patients. Reducing TyG-BMI levels may substantially lower the risks associated with these outcomes and mortality for this population. Regular TyG-BMI assessment during the initial stages of diabetes diagnosis and treatment, coupled with effective weight management, is vital for minimizing cardiovascular comorbidities and mortality among patients with T2DM. Moreover, incorporating TyG-BMI into the conventional risk models greatly improves their prediction accuracy for cardiovascular and overall mortality risks.
The primary strength of this research lies in its pioneering evaluation within the ACCORD cohort of TyG-BMI’s predictive value for future cardiovascular outcomes and overall mortality among T2DM patients. Following adjustments of cardiovascular risk factors, an extensive, long-term follow-up was conducted on a substantial cohort of T2DM patients for cardiovascular and an all-cause mortality evaluation, aiming to objectively assess TyG-BMI’s potential as a biomarker for mortality risk prediction. This study introduces several important innovations. To our best knowledge, we are the first to reveal a significant positive association of TyG-BMI with the CHF risk among T2DM individuals, which broadens the predictive utility of this index beyond atherosclerotic cardiovascular outcomes to encompass heart failure-related endpoints. Furthermore, utilizing data from the ACCORD trial (a cohort characterized by older age, poor glycemic control, and high baseline cardiovascular risk), we are the first to uncover a U-shaped correlation of TyG-BMI with both cardiovascular and overall mortality, alongside a clinically relevant inflection point. This non-linear pattern may assist in refining risk stratification within the advanced-stage T2DM patients. In addition, our analyses validated that TyG-BMI has superior predictive power for adverse outcomes compared to either TyG or BMI alone, underscoring the importance of combining metabolic and anthropometric markers. These results provide a fresh perspective on the prognostic relevance of TyG-BMI, highlighting its potential clinical utility in high-risk diabetic cohorts with complex comorbidities.
Nevertheless, our study does presents certain limitations. Firstly, it is a secondary analysis of the data from the ACCORD and ACCORDION trials. Despite the multivariate Cox adjustment of potential confounders, some residual or overlooked confounders, such as lifelong dietary habits and lifestyle changes, might still persist. Secondly, being observational in nature, this study cannot confirm a causal link between baseline TyG-BMI and cardiovascular outcomes or overall mortality risks in T2DM patients. Consequently, future studies should include carefully designed, long-term prospective interventions with propensity score matching to further confirm our results.
Thirdly, the ACCORD and ACCORDION trial cohorts comprised T2DM patients at elevated cardiovascular risk, all of whom were exclusively from the United States and Canada. Variations in metabolic profiles, lipid levels, and cardiovascular risks exist across ethnicities, and healthcare systems in the United States and Canada differ from those in other regions regarding medical access and CVD management practices. Moreover, most participants in this study were older, had poorly controlled blood glucose, and were at a high risk for CVD, which may restrict the applicability of our findings to younger or better-controlled patients. Therefore, caution is warranted when generalizing these conclusions to other groups, and further research in more diverse cohorts is needed to confirm these findings. Lastly, relying solely on data mining to form conclusions could impact the generalizability. To make our results more generalizable and robust, we intend to collect cohort data from diverse subgroups and settings for further analyses. This strategy aims to reduce data bias and improve the credibility of our findings. Finally, the unavailability of fasting insulin data in the ACCORD database prevented us from comparing TyG-BMI with other established IR indices, such as HOMA-IR. This limitation hinders our ability to fully position TyG-BMI among conventional IR markers. To comprehensively evaluate the relative predictive value and clinical utility of TyG-BMI, future studies should incorporate direct measures of IR and compare them with this composite index.
CONCLUSION
Conclusively, this study has revealed that variations in TyG-BMI are independently linked to both cardiovascular outcomes and overall mortality among T2DM individuals. TyG-BMI exhibits U-shaped correlations with both cardiovascular and overall mortality risks, whereas a positive linear association with CHF risk, establishing TyG-BMI as a predictor for these significant outcomes. Our results emphasize the clinical value of TyG-BMI as an important biomarker for evaluating cardiovascular risk. The implications of this study are substantial, since the incorporation of TyG-BMI into standard evaluations may enable early intervention, thus lowering cardiovascular outcomes and overall mortality. Additionally, by concentrating on modifiable TyG-BMI factors, clinicians can develop more personalized treatment approaches. This research provides valuable insights into managing T2DM patients clinically and assessing their risks of cardiovascular and overall mortality.
ACKNOWLEDGEMENTS
The authors express their sincere gratitude to the ACCORD/ACCORDION study group and NHLBI BioLINCC. The views presented in this article are solely those of the authors and do not necessarily represent the perspectives of the ACCORD/ACCORDION study authors or NHLBI BioLINCC.
Footnotes
Provenance and peer review: Unsolicited article; Externally peer reviewed.
Sun H, Saeedi P, Karuranga S, Pinkepank M, Ogurtsova K, Duncan BB, Stein C, Basit A, Chan JCN, Claude Mbanya J, Pavkov ME, Ramachandaran A, Wild SH, James S, Herman WH, Zhang P, Bommer C, Kuo S, Boyko EJ, Magliano DJ. Erratum to "IDF Diabetes Atlas: Global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045" [Diabetes Res. Clin. Pract. 183 (2022) 109119].Diabetes Res Clin Pract. 2023;204:110945.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 64][Reference Citation Analysis (0)]
Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala MG, Hernández-González SO, Jacques-Camarena O, Rodríguez-Morán M. The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp.J Clin Endocrinol Metab. 2010;95:3347-3351.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 607][Cited by in RCA: 1171][Article Influence: 78.1][Reference Citation Analysis (0)]
Lopez-Jaramillo P, Gomez-Arbelaez D, Martinez-Bello D, Abat MEM, Alhabib KF, Avezum Á, Barbarash O, Chifamba J, Diaz ML, Gulec S, Ismail N, Iqbal R, Kelishadi R, Khatib R, Lanas F, Levitt NS, Li Y, Mohan V, Mony PK, Poirier P, Rosengren A, Soman B, Wang C, Wang Y, Yeates K, Yusuf R, Yusufali A, Zatonska K, Rangarajan S, Yusuf S. Association of the triglyceride glucose index as a measure of insulin resistance with mortality and cardiovascular disease in populations from five continents (PURE study): a prospective cohort study.Lancet Healthy Longev. 2023;4:e23-e33.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 160][Reference Citation Analysis (0)]
Ramírez-Vélez R, Pérez-Sousa MÁ, González-Ruíz K, Cano-Gutierrez CA, Schmidt-RioValle J, Correa-Rodríguez M, Izquierdo M, Romero-García JA, Campos-Rodríguez AY, Triana-Reina HR, González-Jiménez E. Obesity- and Lipid-Related Parameters in the Identification of Older Adults with a High Risk of Prediabetes According to the American Diabetes Association: An Analysis of the 2015 Health, Well-Being, and Aging Study.Nutrients. 2019;11:2654.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 24][Cited by in RCA: 70][Article Influence: 11.7][Reference Citation Analysis (0)]
Marx N, Federici M, Schütt K, Müller-Wieland D, Ajjan RA, Antunes MJ, Christodorescu RM, Crawford C, Di Angelantonio E, Eliasson B, Espinola-Klein C, Fauchier L, Halle M, Herrington WG, Kautzky-Willer A, Lambrinou E, Lesiak M, Lettino M, McGuire DK, Mullens W, Rocca B, Sattar N; ESC Scientific Document Group. 2023 ESC Guidelines for the management of cardiovascular disease in patients with diabetes.Eur Heart J. 2023;44:4043-4140.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 57][Cited by in RCA: 619][Article Influence: 309.5][Reference Citation Analysis (0)]
Ahmad A, Lim LL, Morieri ML, Tam CH, Cheng F, Chikowore T, Dudenhöffer-Pfeifer M, Fitipaldi H, Huang C, Kanbour S, Sarkar S, Koivula RW, Motala AA, Tye SC, Yu G, Zhang Y, Provenzano M, Sherifali D, de Souza RJ, Tobias DK; ADA/EASD PMDI, Gomez MF, Ma RCW, Mathioudakis N. Precision prognostics for cardiovascular disease in Type 2 diabetes: a systematic review and meta-analysis.Commun Med (Lond). 2024;4:11.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 5][Cited by in RCA: 20][Article Influence: 20.0][Reference Citation Analysis (0)]
Luo C, Li Q, Wang Z, Duan S, Ma Q. Association between triglyceride glucose-body mass index and all-cause mortality in critically ill patients with acute myocardial infarction: retrospective analysis of the MIMIC-IV database.Front Nutr. 2024;11:1399969.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 5][Reference Citation Analysis (0)]
Li C, Zhang Z, Luo X, Xiao Y, Tu T, Liu C, Liu Q, Wang C, Dai Y, Zhang Z, Zheng C, Lin J. The triglyceride-glucose index and its obesity-related derivatives as predictors of all-cause and cardiovascular mortality in hypertensive patients: insights from NHANES data with machine learning analysis.Cardiovasc Diabetol. 2025;24:47.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 6][Cited by in RCA: 21][Article Influence: 21.0][Reference Citation Analysis (0)]
Xiao S, Zhang Q, Yang HY, Tong JY, Yang RQ. The association between triglyceride glucose-body mass index and all-cause and cardiovascular mortality in diabetes patients: a retrospective study from NHANES database.Sci Rep. 2024;14:13884.
[RCA] [PubMed] [DOI] [Full Text][Cited by in RCA: 16][Reference Citation Analysis (0)]
