Abstract
Background
Insulin resistance (IR) heightens the danger of cardiovascular diseases in the participants with diabetes mellitus (DM) and prediabetes mellitus(preDM). The glucose-lipid metabolism index (GLMI) has emerged as a potential biomarker for IR. The association between GLMI and cardiovascular diseases (CVD) risk in prediabetic and diabetic participants remains unidentified. The purpose of this study was to investigate the relationship between GLMI and CVD prevalence in participants with diabetes and prediabetes.
Methods
Using a cross-sectional design with retrospective analysis of prevalent CVD and follow-up for mortality, this investigation analyzed data from the National Health and Nutrition Examination Survey spanning from 2001 to 2018. The primary cross-sectional analysis included 15,037 eligible participants to assess the association between GLMI and prevalent CVD. The GLMI was calculated as follows: ln [Total triglycerides × Fasting plasma glucose ∕2] ×Body mass index × Total cholesterol ∕ High density lipoprotein cholesterol. CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina. Multivariable survey-weighted logistic regression model was employed to assess the association between GLMI and CVD. Subgroup analyses were conducted to explore potential association factors. Multivariable survey-weighted logistic regression analysis was used to assess the association between GLMI and the CVD prevalence.
Results
Comparing with control(n = 10152) and preDM(n = 949) participants, the DM participants (n = 3936) showed the higher GLMI level (all P < 0.001) and prevalence of CVD (all P < 0.001). Furthermore, in a fully adjusted survey-weighted model, we observed higher GLMI tertiles were associated with increased CVD prevalence in preDM and DM participants with the lowest GLMI tertile as reference (the middle tertile: 3.12 95% CI: 1.97–4.28; the highest tertile: 5.84,95% CI: 3.05–7.63; all P < 0.001). In addition, GLMI showed positive trends with severe CVD outcomes, including CVD subtypes and mortality. (All P for trend < 0.05). Meanwhile, GLMI demonstrated superior discriminative performance for identifying prevalent CVD compared to TyG index and TG/HDL ratio. Furthermore, the GLMI tertiles–CVD association was significantly stronger in participants with hypertension than in those without hypertension (P for interaction = 0.038). For each 1- standard deviation (SD) increase in GLMI, the CVD prevalence was increased by 23% (OR 1.23, 95% CI 1.09–1.40) after adjustment for confounders.
Conclusions
GLMI showed a strong association with CVD and may be an important marker correlated with metabolic health and CVD burden.
Clinical trial number
Not applicable
Supplementary Information
The online version contains supplementary material available at 10.1186/s12902-026-02169-2.
Keywords: GLMI, CVD, Diabetes mellitus, Prediabetes mellitus
Introduction
Cardiovascular disease (CVD) stands as the foremost cause of death globally for more than three decades, accounting for over 18 million deaths annually [1]. Proactively identifying and screening participants at risk of CVD, along with implementing necessary interventions to manage these risk factors, are essential in reducing the incidence of cardiovascular events and mitigating their threat to human health [2]. Among them, participants with prediabetes and diabetes exhibit a markedly higher CVD incidence, attributable to mechanisms such as chronic hyperglycemia-induced endothelial dysfunction, accelerated atherosclerosis, metabolic disturbances and systemic inflammation [3].
Insulin resistance (IR) is defined by a reduced responsiveness of target tissues to normal circulating levels of insulin. Over time, it may result in various metabolic disturbances, including prediabetes, diabetes, dyslipidemia, and hypertension [4]. More significantly, IR is intricately linked to the development and progression of CVD [5]. Effective control of IR may mitigate the impact of metabolic diseases on CVD [6]. Understanding the role of IR in cardiovascular health and evaluating participants for IR offers valuable insights for predicting CVD prevalence and for developing targeted therapies and preventive measures.
While the hyperinsulinemic-euglycemic (HIEG) clamp is considered the gold standard for identifying IR, its clinical utility and feasibility in large-scale epidemiological investigations are limited due to its time-consuming and burdensome nature [7]. Similarly, the homeostasis model assessment for insulin resistance (HOMA-IR) is not suitable for large population-based cohort studies due to its cost and complexity [8]. The triglyceride and glucose (TyG) index, calculated from triglyceride and fasting glucose levels, serves as a valuable marker for assessing IR [9]. The TyG index offers a more cost-effective and accessible alternative. Its simplicity and ease of use render it particularly advantageous in large-scale screenings and in settings where advanced testing methods are impractical [10]. Moreover, the TyG index exhibits a close association with prediabetes and diabetes [3]. Although the TyG index can be used as a surrogate marker for IR, it is prudent to consider its limitations when assessing CVD risk due to its high susceptibility to diet and medication [11].Meanwhile, TyG index was calculated based on the combination of triglyceride and glucose. This single measure may not provide a complete picture of an participants’ metabolic status or cardiovascular health. The metabolic diversity of different participants may influence the performance of the TyG index, leading to an underestimation or overestimation of cardiovascular risk in some patients. Therefore, novel biomarkers for assessing the severity of IR and thus the risk of metabolic heart disease remain to be explored.
The glucose-lipid metabolism index (GLMI) extends the TyG index by incorporating additional metabolic parameters, including body mass index (BMI), total cholesterol (TC), and high-density lipoprotein (HDL) levels, offering a more comprehensive reflection of IR. Currently, GLMI has demonstrated superior predictive capability in pancreatic cancer patients with DM 12. However, the relationship between GLMI and CVD prevalence in prediabetes and diabetes populations remains unknown. We hypothesized that higher GLMI is independently and dose-dependently associated with greater odds of prevalent CVD among US adults with prediabetes or diabetes. Secondarily, we hypothesized that this association would be significantly modified by hypertension, with stronger effects observed in hypertensive individuals. Moreover, it also had acceptable power to identify CVD mortality and CVD subtypes, and would demonstrate superior discriminative capacity compared to conventional metabolic indices.
Methods
Study population
The data analyzed in this study were from the 2001 to 2018 National Health and Nutrition Examination Survey (NHANES) database. NHANES is a national survey of children and adults in the United States and the data were collected through personal structured interviews at home, health examinations at a mobile examination center, and specimen analyses in the laboratory. Mortality data were obtained through NHANES-linked National Death Index (NDI) records, with CVD mortality defined by ICD-10 codes I00-I78. This study consisted of two parts: a cross-sectional analysis examining the association between baseline GLMI and prevalent CVD, and a cohort analysis using linked mortality data to examine the association with CVD death.
Data collection and clinical definitions
Demographic data, physical examination results, laboratory blood test results, and medical history information were collected for all participants. We enrolled adult participants aged 18 years and older with Type 2 DM or preDM who underwent mobile examinations from NHANES 2001–2018, which provided comprehensive data for blood and physical measurements. The following groups of patients were excluded from the study: (1) participants without preDM and DM diagnosis; (2) participants without TG, FBG, BMI, HDL and TC; (3) participants with incomplete data on CVD status.
Type 2 DM was defined as two abnormal metabolic measures from the same examination cycle (glycohemoglobin (HbA1c) in plasma sample ≥ 6.5%[HPLC method], fasting blood glucose (FBG) in plasma sample ≥ 7mmol/L(126 mg/dL) [venous plasma, Roche Cobas analyzer], random blood glucose in plasma sample ≥ 11.1mmol/L), or use of insulin or other diabetes medication by themselves among the self-confessed type 2 diabetes participants. PreDM was defined as FBG from 5.6 mmol/L(100 mg/dL) to 6.9 mmol/L(125 mg/dL), 2–2 h plasma glucose during 75-g oral glucose tolerance test from 7.8 mmol/L (140 mg/dL)to 11.0 mmol/L(198 mg/dL), or HbA1c from 5.7 to 6.4%.
The dependent variables for this study were continuous systolic and diastolic blood pressure readings and categorical indication of hypertension (yes/no). Blood pressure was measured by trained technicians using mercury sphygmomanometers after 5 min seated rest, with appropriate cuff sizes and three consecutive readings averaged. SBP and DBP were calculated as the mean of 3 consecutive BP measurements. We defined hypertension as having a BP reading meeting the 2017 ACC/AHA guidelines definition (SBP ≥ 130 mm Hg or DBP ≥ 80 mm Hg) or self-reported current use of BP-lowering medication (regardless of BP reading). Participants were asked the question, “Have you ever been told by a doctor or health professional that you had hypertension, also called high blood pressure?” Those who responded yes were considered aware of their hypertension status.
Age, gender, race, education level, smoking status, and drinking status were recorded. Race was categorized into five groups: Mexican American, non-Hispanic White, non-Hispanic Black, other Hispanic, and other. Educational levels were categorized as less than high school, high school or equivalent, and college or above.
Weight and height were obtained using calibrated electronic scales and stadiometers with participants in light clothing without shoes. Self-reported questionnaires were used to collect weight data from participants at age 25 and 10 years prior to recruitment. BMI (Body Mass Index) was calculated by dividing weight in kilograms by height in meters squared.
During the household interview, adults aged 18 and up self-reported their smoking status. Participants who claimed to have smoked fewer than 100 cigarettes in their lives were labeled ‘never smokers’. Former smokers were participants who had smoked more than 100 cigarettes in their lives but had quit, while current smokers were those who were currently smoking.
Alcohol intake status was delineated into four distinct categories: former drinkers, indicating participants who abstained from alcohol consumption within the past year but had consumed a minimum of 12 drinks throughout their lifetime; mild drinkers, characterized by an average intake of no more than 1 drink per day for women and 2 drinks per day for men during the preceding year; moderate drinkers, denoting participants with an average intake of no more than 3 drinks per day for women and 4 drinks per day for men in the past year; and heavy drinkers, representing those who consumed an average of 4 or more drinks per day for women and 5 or more drinks per day for men throughout the previous year.
For the calculation of eGFR, patients’ serum creatinine was converted using the well-recognized CKD-EPI formula.
Laboratory assays used venous samples processed at CDC-certified facilities—glucose via hexokinase method (Roche Cobas 6000), lipids enzymatically (Roche Cobas) with LDL-C calculated via Friedewald equation, and HbA1c by HPLC (Tosoh G8). All adhering to CDC quality control standards (CV < 3.5%) with detailed procedures documented in NHANES Laboratory Manuals. These revisions ensure full technical reproducibility. Laboratory characteristics, including total cholesterol (TC), total triglycerides (TG), low-density lipoprotein (LDL), high-density lipoprotein (HDL), creatinine, and fasting plasma glucose (FPG), were obtained for all subjects. All lipid parameters were measured from fasting venous blood samples collected after a minimum 8-hour fast. This aligns with NHANES protocols and NCEP ATP III guidelines for standardized lipid assessment.
Prevalent CVD was defined as a self-reported physician diagnosis of any of the following: coronary heart disease (CHD), stroke, congestive heart failure (CHF), or angina. Separately, CVD mortality was defined using linked NDI records according to ICD-10 codes I00-I78. In our study, analysis of CVD mortality—as the terminal manifestation of CVD—was included to holistically assess GLMI’s association with CVD prevalence. Initially, 90,277 participants were enrolled in the study. After excluding participants aged 18 years (N = 33168), those without data for diagnosing diabetes or prediabetes (N = 7302), and those with missing data on the TG (N = 2827), FBG (N = 2812), BMI (N = 3712), HDL(N = 4328), TC(N = 3946) and CVD conditions (N = 17145), our final analysis included 15,037 eligible participants (Fig. 1).
Fig. 1.
Flowchart of the sample selection from National Health and Nutrition Examination Survey (NHANES) 2001–2018. CVD indicates the composite endpoint (coronary heart disease, stroke, congestive heart failure, or angina)
Measurement of glucose-lipid metabolism index (GLMI)
GLMI was calculated as follows as described previously [12]: ln [TG (mg∕dL) × FPG (mg∕dL) ∕2] ×BMI × TC∕ HDL-c. The GLMI was classified as the lowest, middle, and highest groups according to tertiles ((T1 < 103.2, T2 103.2–219.9, and T3 > 219.9).
Statistical analysis
The statistical analyses in this study were conducted using R, version 4.3.1 (R Foundation). To account for NHANES’s complex survey design, we applied Stratification using the masked variance unit pseudo-stratum variable (SDMVSTRA), Clustering adjustment via the masked variance unit pseudo-primary sampling unit (SDMVPSU). In accordance with the NHANES analytical guidelines for multi-cycle data, we constructed a custom survey weight for the combined nine 2-year cycles (2001–2018) by re-scaling the dietary day one sample weight (WTDRD1) to its mean across cycles. The custom weight was calculated as: weight_custom = WTDRD1 / 9. The survey design was specified in R using the svydesign function from the survey package with the following parameters: ids = SDMVPSU (primary sampling units), strata = SDMVSTRA (strata), weights = weight_custom, and nest = TRUE.
The normality of continuous variables was assessed using Shapiro-Wilk test. All analyses accounted for the complex survey design of NHANES by incorporating sampling weights, strata, and clustering. Population-weighted descriptive statistics were computed for the entire study population and across GLMI tertiles. Continuous variables are presented as medians (interquartile range) or means (standard error), and categorical variables are presented as unweighted counts with weighted percentages.
Group comparisons for continuous and categorical variables across GLMI tertiles were performed using design-based tests. Specifically, differences in continuous variables were assessed using the design-based F-test from weighted linear regression models, and differences in categorical variables were assessed using the design-based Chi-squared test (Rao-Scott correction) from weighted logistic regression models. These methods provide valid inferences under the complex sampling design. Missing data were handled using complete-case analysis. The ability of GLMI to identify CVD was evaluated by receiver operating characteristic (ROC) analysis. Multivariable survey-weighted logistic regression models were used to assess the association between GLMI and prevalent CVD. For consistency in the analytical approach and because CVD mortality during follow-up was analyzed as a binary (yes/no) outcome, we also used multivariable survey-weighted logistic regression to assess its association with GLMI, as well as with CVD subtypes. Model 1 was adjusted for age, gender, and race. Model 2 was adjusted for Age, gender, race, education level, serum creatinine, LDL-C, eGFR, hypertension, smoking and alcohol consumption status, Waist circumference, HbA1c, Lipid-lowering medicines, Antihypertensive medicines and Antidiabetic medicines. In Model 1 and Model 2, we did not include the direct biochemical components used to calculate GLMI (TG, FPG, TC, HDL-C) or BMI, to avoid over-adjustment and to evaluate the independent association of the composite index. Restricted cubic splines were used with three knots at the 10th, 50th, and 90th percentiles to flexibly model the association between GLMI and the CVD prevalence. In addition, possible modifications of the association between GLMI (per increment) and CVD prevalence were assessed across subgroups (age, gender, BMI, hypertension) using adjusting P-values for multiple testing using false discovery rate (FDR) correction. Heterogeneity across subgroups was assessed using linear regression models and interactions between subgroups and GLMI were examined using likelihood ratio testing.
The C-statistic, integrated discrimination improvement (IDI), and net reclassification improvement (NRI) indices were calculated to compare the discrimination and classification abilities of GLMI and other metrics (as continuous variables [per 1 − SD change]) for CVD prevalence using multivariable survey-weighted logistic regression. All regression analyses assessing GLMI-CVD associations were restricted to preDM and DM participants. A P-value < 0.05 (two-tailed) was considered statistically significant.
Results
Characteristics of the subjects according to GLMI quartiles
The enrolled participants were stratified into three groups: the control group (n = 10,152), the preDM group (n = 949), and the DM group (n = 3,936). Regarding GLMI, the preDM group (124.2,70.43–203.0) and DM group (211.4,149.4-315.1) exhibited higher GLMI values than the control group (71.93, 35.23–127.7, all P < 0.001), with a significant difference between the preDM group and DM group (Fig. 2A). Additionally, CVD prevalence in the preDM group (24.8%, 95% CI: 22.3–25.3) and DM group (26.9%, 95% CI: 24.4–28.4) was significantly higher compared to the control group (14.9%, 95% CI: 11.4–16.7), and preDM group and DM group demonstrated similar CVD prevalence (Fig. 2B). Given the shared pathophysiological basis of dysglycemia and to enhance statistical power for the primary analysis, we combined the preDM and DM participants into a single ‘dysglycemia’ group for initial assessments of the GLMI-CVD association. The baseline information of the population was shown in Table 1. Furthermore, among the preDM and DM subjects, participants with CVD had a higher GLMI compared to those without CVD (P < 0.05, Fig. 2C).
Fig. 2.
Comparison of GLMI and CVD prevalence among the enrollment participants. (A) Comparison of the GLMI levels between the control group, PreDM group and DM group. (B) Comparison of CVD prevalence among the three groups. (C) Comparison of the GLMI levels between the non-CVD group and CVD group among the PreDM and DM participants. ***P < 0.001. CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina
Table 1.
The characteristics of enrolled participants of control group and PreDM/DM group
| Control | PreDM and DM | P value | |
|---|---|---|---|
| n = 10,152 | n = 4885 | ||
| Age(year) | 53(38,60) | 58(39,65) | 0.174 |
| Gender(n,%) | |||
| male | 5686(56.0) | 2596(53.1) | 0.166 |
| Female | 4466(44) | 2289(46.9) | |
| BMI (Kg/m2) | 27(24,32) | 32(29,37) | < 0.001 |
| Race(n,%) | |||
| Hispanic | 2548(25.1) | 1167(23.9) | 0.077 |
| White | 2503(24.7) | 1133(23.2) | |
| Black | 2398(23.6) | 1129(23.11) | |
| Others | 2703(26.6) | 1456(29.80) | |
| Education, n (%) | < 0.001 | ||
| Less than high school | 3736(36.80) | 1647(33.71) | |
| High school or equivalent | 2894(28.50) | 1506(30.82) | |
| College or above | 3522(34.70) | 1732(35.43) | |
| Smoke, n (%) | < 0.001 | ||
| Never | 4267(42.0) | 1568(32.1) | |
| Former | 3198(31.5) | 1339(27.4) | |
| Now | 2687(26.5) | 1978(40.49) | |
| Drink, n (%) | 0.041 | ||
| Never | 3340(32.9) | 2792(57.1) | |
| Mild-Moderate | 6812(67.1) | 2093(42.9) | |
| BMI (Kg/m2) | < 0.001 | ||
| < 23.9 | 3807(37.5) | 1301(26.6) | |
| 24–27.9 | 2102(20.7) | 2069(42.3) | |
| ≥ 28 | 4243(41.8) | 1515(31.0) | |
| hypertension(n,%) | 4213(41.5) | 3157(64.6) | < 0.001 |
| eGFR (mL/min/1.73 m2) | 102.(8.03) | 96(6.86) | < 0.001 |
| TG(mg/dl) | 142(6.95) | 158(7.55) | < 0.001 |
| TC (mg/dl) | 204(10.58) | 244(12.72) | < 0.001 |
| HDL-C (mg/dL) | 66(5.42) | 48(4.77) | < 0.001 |
| Fast glucose (mg/dL) | 87(8.35) | 165(13.24) | < 0.001 |
| CVD(n,%) | 619(6.10) | 1075(22.0) | 0.032 |
| Waist circumference | 79.15(5.32) | 94.46(9.47) | 0.043 |
| SBP | 102.7(11.46) | 131.6(14.46) | < 0.001 |
| DBP | 72.3(7.36) | 90.6(8.62) | < 0.001 |
| HbA1c | 4.87(0.21) | 6.40(0.32) | < 0.001 |
| Lipid-lowering medicines | 1075(10.6) | 2548(58.2) | < 0.001 |
| Antihypertensive medicine | 3330(32.9) | 2575(52.7) | < 0.001 |
BMI, Body Mass Index; SBP, Systolic Blood Pressure; DBP, Diastolic Blood Pressure; FBG, Fasting Blood Glucose; HbA1c, glycosylated hemoglobin A1c; HOMA-IR, homeostatic model assessment of insulin resistance; TG, Total Triglyceride; TC, Total Cholesterol; HDL, High-Density Lipoprotein; LDL, Low-Density Lipoprotein; Hs-CRP, High sensitive C-Reactive Protein; DM, Diabetes Mellitus; PreDM, Prediabetes Mellitus; CVD, Cardiovascular Disease; eGFR, estimated Glomerular Filtration Rate
Based on these results, preDM and DM participants were categorized into three groups according to the tertiles of baseline GLMI levels for subsequent analysis (T1 < 103.2, T2: 103.2–219.9, and T3 > 219.9, Table 2). Compared to subjects in the lowest GLMI tertile, those in the middle and highest tertiles exhibited significantly higher BMI, FPG, TG, LDL, and hypertension prevalence. Conversely, serum eGFR and HDL levels were significantly lower in the middle and highest GLMI tertiles compared to the lowest GLMI tertile group. Compared to the lowest GLMI tertile group, participants in the middle and highest tertiles were more likely to be male, have lower educational levels, and be smokers and drinkers. Notably, CVD prevalence increased significantly across ascending GLMI tertiles.
Table 2.
The characteristics of enrolled participants with tertile stratification according to the GLMI
| 1st Tertile | 2nd Tertile | 3rd Tertile | P value | |
|---|---|---|---|---|
| Age(year) | 53(39,62) | 57(42,68) | 62(45,80) | < 0.001 |
| Gender(n,%) | ||||
| male | 662(40.7) | 828(50.9) | 1106(67.9) | < 0.001 |
| Female | 965(59.3) | 799(49.1) | 522(32.1) | |
| BMI (Kg/m2) | 30.2(26.6,38.9) | 33.4(30.1,41.7) | 34.4(30.7,42.2) | < 0.001 |
| Race(n,%) | ||||
| Hispanic | 507(31.2) | 631(38.8) | 531(32.6) | 0.081 |
| White | 278(17.1) | 388(23.8) | 468(28.7) | |
| Black | 393(24.1) | 426(26.2) | 310(19.0) | |
| Others | 449(27.6) | 182(11.2) | 319(19.7) | |
| Education, n (%) | < 0.001 | |||
| Less than high school | 439(27.0) | 489(30.0) | 719(44.2) | |
| High school or equivalent | 585(36.0) | 497(30.5) | 424(26.0) | |
| College or above | 605(37.0) | 641(39.5) | 485(29.8) | |
| Smoke, n (%) | < 0.001 | |||
| Never | 666(39.3) | 525(31.6) | 377(22.7) | |
| Former | 603(36.3) | 354(21.3) | 382(23.0) | |
| Now | 358(20.7) | 783(47.1) | 903(54.3) | |
| Drink, n (%) | < 0.001 | |||
| Never | 1046(64.3) | 944(58.0) | 802(49.3) | |
| Mild-Moderate | 581(35.8) | 783(48.1) | 825(50.8) | |
| BMI (Kg/m2) | < 0.001 | |||
| < 23.9 | 940(57.8) | 702(43.1) | 659(40.5) | |
| 24–27.9 | 327(20.1) | 393(24.2) | 349(21.5) | |
| ≥ 28 | 360(22.1) | 532(32.7) | 620(38.0) | |
| hypertension(n,%) | 945(58.0) | 1059(65.1) | 1153(70.8) | < 0.001 |
| PreDM(n,%) | 377(23.2) | 302(18.6) | 273(16.8) | 0.048 |
| DM(n,%) | 1250(76.8) | 1325(81.4) | 1354(83.2) | |
| eGFR (mL/min/1.73 m2) | 98.7 (8.9) | 97.1 (6.8) | 92.14 (5.6) | < 0.001 |
| TG(mg/dl) | 88(7.25) | 144(11.9) | 242(20.4) | < 0.001 |
| TC (mg/dl) | 219(15.5) | 236(18.8) | 277(20.5) | < 0.001 |
| HDL-C (mg/dL) | 53.97 (4.71) | 48.84 (4.53) | 41.28 (3.48) | < 0.001 |
| Fast glucose (mg/dL) | 157.84 (12.53) | 174.28 (13.11) | 192.28 (13.42) | < 0.001 |
| CVD(n,%) | 242(14.9) | 363(22.3) | 470(28.9) | 0.032 |
| Waist circumference | 82.1(9.1) | 97.8(7.4) | 103.4(9.4) | 0.013 |
| SBP | 124.9(10.6) | 132.0(12.9) |
136.7 (13. 1) |
< 0.001 |
| DBP | 86.4(14.9) | 90.4(11.3) | 95.21(13.0) | < 0.001 |
| HbA1c | 5.72(0.15) | 6.34(0.28) | 7.04(0.32) | < 0.001 |
| Lipid-lowering medicines | 824(50.4) | 779(47.9) | 945(56.4) | < 0.001 |
| Antihypertensive medicine | 901(55.3) | 1017(62.5) | 1111(68.1) | < 0.001 |
| Antidiabetic medicines | 715(43.7) | 894(52.2) | 966(59.1) | < 0.001 |
Note: CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina
DM, diabetes mellitus; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG fasting blood glucose; HbA1c, glycosylated hemoglobin A1c; HOMA-IR, homeostatic model assessment of insulin resistance total triglyceride; TC, total cholesterol; HDL, high-density lipoprotein; LDL, low-density lipoprotein; Hs-CRP, high sensitive-CRP
CVD prevalence of subjects within the different tertiles of GLMI
Table 3 shows that the odds ratios were 1.47(95% CI 0.99–2.04) in the middle tertile and 1.76 (95% CI 1.12 − 2.53) in the highest tertile, if CVD prevalence of the lowest tertile was one in an unadjusted model. Moreover, the odds ratios for significant CVD using the lowest tertile as the reference were 2.59(95% CI 1.75–3.15) in the middle tertile and 3.17(95% CI 2.30–3.70) in the highest tertile (unadjusted model) following age, gender, and race adjustment. Following more potential confounders adjustment, a positive association between the GLMI tertiles and the CVD prevalence remained consistent (middle : 3.12, 95% CI 1.97–4.28; highest : 5.84 3.05–7.63), indicating that a GLMI in the middle and highest tertile was an independent associated factor for increased CVD prevalence in our subjects. To investigate whether the strength of this association differed by glucose tolerance status, we subsequently performed stratified analyses in the preDM and DM groups separately. Our analysis revealed a notably stronger association in the DM group. Specifically, each standard deviation increase in GLMI was associated with an 20% higher odds of CVD (OR = 1.20, 95% CI 1.10–1.45) in PreDM group, compared to a 114% increase (OR = 2.24, 95% CI 1.88,2.43) in DM group (Table S1-S2). As is shown in Fig. 3, we used restricted cubic splines to flexibly model and found the association between GLMI and the odds of prevalent CVD was overall significant (P for overall association < 0.001) and monotonic. The formal test indicated no significant deviation from linearity (P for nonlinearity = 0.82). Visual inspection of the spline curve suggested a gradual increase in odds across the GLMI range, with no statistically evident threshold. Although the slope appeared to increase at higher GLMI levels, this pattern was not statistically distinguished from a linear trend and should be considered exploratory.
Table 3.
Unadjusted and adjusted odds ratios with 95% confidence intervals of odds of prevalent CVD by tertiles of GLMI
| Unadjusted | Model 1 | Model 2 | |||||
|---|---|---|---|---|---|---|---|
| GLMI | OR(95% CI) | P value | OR(95% CI) | P value | OR(95% CI) | P value | |
| CVD | |||||||
| Continuous per unit increase | 1.24(1.02,1.39) | < 0.001 | 1.30(1.07, 1.41) | < 0.001 | 1.45(1.14,1.88) | < 0.001 | |
| Categorical | |||||||
| 1st Tertile(GLMI < 103.2) | 1 | 1 | 1 | ||||
| 2nd Tertile (103.2–219.9) | 1.47 (0.99–2.04) | < 0.001 | 2.59 (1.75–3.15) | 0.011 | 3.12 (1.97–4.28) | < 0.001 | |
| 3rd Tertile (> 219.9) | 1.76 (1.12–2.53) | < 0.001 | 3.17 (2.30–3.70) | < 0.001 | 5.84 (3.05–7.63) | < 0.001 | |
| P for trend | < 0.001 | < 0.001 | < 0.001 | ||||
Note: CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina
Model 1: Age, gender, and race were adjusted
Model 2: Age, gender, race, education level, serum creatinine, LDL-C, eGFR, hypertension, smoking and alcohol consumption status, Waist circumference, HbA1c, Lipid-lowering medicines, Antihypertensive medicines and Antidiabetic medicines were adjusted. Notably, GLMI components (TG, FPG, BMI, TC, HDL-C) were not adjusted for
Fig. 3.
Dose-response relation between GLMI and CVD prevalence in PreDM and DM patient. CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina. The smooth curve was fitted using a restricted cubic spline with 3 knots
The dose-response relationship was further examined using restricted cubic splines (Fig. 3). The association between GLMI and the odds of prevalent CVD was overall significant (P for overall association < 0.001) and monotonic. The formal test indicated no significant deviation from linearity (P for nonlinearity = 0.82). Visual inspection of the spline curve suggested a gradual increase in odds across the GLMI range, with no statistically evident threshold. Although the slope appeared to increase at higher GLMI levels (e.g., beyond approximately 125), this pattern was not statistically distinguished from a linear trend and should be considered exploratory.
The ROC curve analysis of GLMI cut-off values indicating CVD in males and females is shown in Figure S1. The cut-off value with the highest Youden index was a SIRI of 169.53 with a sensitivity of 77.7% in males, whereas the SIRI cut-off value was 147.55 (sensitivity 70.54%) in females.
Associations of GLMI with CVD severity: subtypes and mortality
We performed multivariable survey-weighted logistic regression to assess associations between GLMI and CVD mortality/CVD subtypes. Figure 4 illustrates the relationship between the GLMI and the oddsof CVD mortality(A), CHF(B) and CHD(C), stroke(D) and angina(E). After adjusting for Age, gender, race, education level, serum creatinine, eGFR, hypertension, smoking and alcohol consumption status, Waist circumference, HbA1c, Lipid-lowering medicines, antihypertensive medicines and antidiabetic medicines, the results showed that GLMI exhibited significant positive correlations with the odds of CVD mortality, CHF, CHD, stroke and angina (All P for trend < 0.05).
Fig. 4.
Forest plots show the ORs and 95% confidence intervals per 1-SD increase in GLMI for (A) CVD mortality, (B) congestive heart failure (CHF), (C) coronary heart disease (CHD), (D) stroke and (E) angina. Analyses were adjusted for the full set of covariates specified in Model 2 (see Methods). CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina
Stratified analysis of CVD prevalence and GLMI according to potential effect modifier
Stratified analyses were conducted to further evaluate the association between GLMI and CVD prevalence across various subgroups. Age, gender, BMI, and hypertension were used as stratification variables to determine the trend of effect sizes (Table 4). The association between GLMI and CVD prevalence in PreDM and DM participants was not significantly modified by a series of variables, including age, and gender. However, a significant interaction between GLMI and CVD prevalence was observed in relation to hypertension (P = 0.034). Among participants with hypertension, higher GLMI was significantly associated with an increased CVD prevalence (OR: 2.39 [95% CI 1.49–3.84]). Conversely, in participants without hypertension, higher GLMI had no significant effect on CVD prevalence (OR: 0.93 [95% CI 0.72–1.36]). Subgroups analysis results demonstrated similar effect modification by hypertension was observed in both PreDM and DM cohorts (Table S3 and S4).
Table 4.
Association between GLMI and CVD in various subgroups
| Characteristics | OR (95% CI) | P-value | P for Interaction | q-value (FDR-adjusted) |
|---|---|---|---|---|
| Age | ||||
| < 65 | 1.19(0.98,1.69) | < 0.001 | 0.135 | 0.344 |
| > 65 | 2.45(1.97,3.74) | < 0.001 | ||
| Gender | ||||
| Female | 1.88(1.56,3.09) | 0.032 | 0.509 | 0.815 |
| Male | 2.44(1.63,3.88) | < 0.001 | ||
| Hypertension | ||||
| NO | 0.87(0.66,1.66) | 0.066 | 0.026 | 0.038 |
| Yes | 2.59(1.59,3.47) | < 0.001 |
Note: CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina
Adjusted for Age, gender, race, education level, serum creatinine, LDL-C, eGFR, hypertension, smoking and alcohol consumption status, Waist circumference, HbA1c, Lipid-lowering medicines, Antihypertensive medicines and Antidiabetic medicines, if not stratified
Comparison of the discriminative performance for CVD among GLMI, TyG index and TG/HDL ratio
Elevated TyG index and TG/HDL ratio were found to apparently correlate with higher CVD risk [3, 13]. In our study, GLMI was found to have a significant correlation with both the TyG index and the TG/HDL ratio (all P < 0.001). Higher TyG and TG/HDL ratios were associated with increased odds ratios (ORs) for CVD prevalence. Indicators of discriminatory ability, including the C-statistic, IDI, and NRI, demonstrated that GLMI offered superior value in risk discrimination for CVD, which point a fresh insight into the significance of GLMI in CVD prevalence assessment (Table 5).
Table 5.
Comparison of the predictive performance for CVD among different indices in PreDM and DM participants
| Metric | GLMI (95% CI) | TyG Index (95% CI) | TG/HDL Ratio (95% CI) |
|---|---|---|---|
| OR per 1-SD | 1.230 (1.090–1.400) | 1.170 (1.030–1.330) | 1.150 (1.010–1.310) |
| C-statistic | 0.724 (0.701–0.747) | 0.681 (0.657–0.705) | 0.692 (0.668–0.716) |
| ΔC-statistic vs. GLMI | Ref | -0.043 (-0.064-0.022) | -0.032(-0.049-0.015) |
| IDI, % | Ref | -4.7 (-6.4 - -3.0) | -3.7 (-5.3 - -2.1) |
| Continuous NRI, % | Ref | -11.9 (-15.3 - -8.5) | -11.3 (-14.7 - -7.9) |
|
Hosmer-Lemeshow χ² (P-value) |
8.24 (0.408) | 14.37 (0.073) | 18.92 (0.015) |
| Calibration slope | 0.98 (0.92–1.04) | 0.87 (0.80–0.94) | 0.82 (0.75–0.89) |
Note: CVD was defined as a composite endpoint including self-reported physician diagnosis of coronary heart disease, congestive heart failure, stroke, or angina
Models adjusted for age, gender, race, education level, serum creatinine, eGFR, hypertension, smoking and alcohol consumption status, Waist circumference, HbA1c, Lipid-lowering medicines, Antihypertensive medicines and Antidiabetic medicines
All metrics derived from 1000 bootstrap replicates incorporating NHANES survey design
Calibration metrics based on 10 risk deciles
Net benefit calculated at 20% risk threshold (statin initiation point)
Discussion
This cross-sectional study with mortality follow-up is the first to investigate the relationship between GLMI and CVD prevalence in a nationally representative sample of U.S. adults with PreDM or DM. It was demonstrated that higher GLMI was observed in prediabetic and diabetic participants significantly and positively associated with the increased CVD prevalence. This association remained significant even after adjustment for potential confounding factors. Notably, GLMI exhibited significant positive correlations CVD mortality and subtypes. Moreover, our results indicate the GLMI–CVD association was significantly stronger in individuals with hypertension. Furthermore, the odds of prevalent CVD was increased by 23% with each 1-SD increase in GLMI.
IR is an important feature of metabolic syndrome, which includes diabetes, obesity, and dyslipidemia. Moreover, it is also a risk factor for CVD [14]. Specially, IR has been demonstrated that its close association with the development of CVD in patients with prediabetes and diabetes [15]. Mechanistically, hyperinsulinemia is responsible for the development of CVD via impaired adipose tissue function, inflammatory pathways overactivation, endothelial dysfunction, excessive oxidative stress, and abnormal RAAS axis. The hyperinsulinemic euglycemic clamp is considered the gold standard for IR, but its complexity and high operational costs make it impractical for widespread clinical use. Therefore, various different IR surrogates are widely used in clinical research. HOMA-IR is the most widely used surrogate marker, but its calculation requires the measurement of fasting insulin concentration [16], which increased the cost during the hospitalizations. The TyG index and TyG-BMI index have also received attention due to its simplicity and ease of use [17, 18]. The TyG index has been closely linked to the CVD incidence, including coronary artery calcification, subclinical atherosclerosis, and arterial stiffness [19–24]. However, there are still several observations that have failed to support the association between the TyG index and recurrent cardiovascular events [25, 26]. Cardiovascular diseases are dynamic in nature, yet most studies rely solely on the TyG index measurement at the baseline. In most studies, TG and FBG were examined only at baseline, regardless of their changes over time, which may lead to potential regression dilution bias. Therefore, the use of the TyG index at baseline as a biomarker to predict outcomes of CVD may be less robust. Meanwhile, nutrition data are missing from most studies, so we were unable to adjust for dietary habits when evaluating the diagnostic or predictive value of the TyG index in CVD. Combining the TyG index with BMI, TC, and HDL-C enhances the comparability and consistency of the GLMI, enabling a more comprehensive evaluation of IR comparing with TyG index.
In our study, subjects within the highest GLMI tertile exhibited a remarkable range of PreDM and DM-associated characteristics, such as a high BMI and disorders of glucose and lipid metabolism, in addition to elder age, higher proportion of males, poorer education status, higher proportion of smoke and drink status. Importantly, the results revealed a significant association between higher baseline GLMI tertile levels and increased CVD prevalence. The robust relationship between GLMI and CVD prevalence demonstrating a near 6fold higher adjusted odds among those in the highest GLMI tertile persisted after comprehensive adjustment for confounders. Importantly, this association remained significant after extensive adjustment for confounders while deliberately excluding the index‘s own components from the model, underscoring that GLMI captures cardiovascular risk information beyond and distinct from its individual metabolic parts. Actually, GLMI simultaneously captures insulin resistance through triglyceride-glucose interactions, adiposity burden via BMI, and atherogenic dyslipidemia reflected in the TC/HDL-C ratio. This tripartite structure likely underpins its superior discriminative capacity over conventional single-dimension indices, which overlooks obesity-driven inflammation and lipid partitioning dynamics. Subgroup analysis revealed a notably stronger association in the DM group compared with PreDM group. This enhanced association in diabetes likely reflects the compounded pathophysiological burden of advanced disease, including progressive β-cell dysfunction, more severe insulin resistance, and the cumulative effect of glucolipotoxicity on vascular endothelium. Clinically, this finding underscores GLMI’s particular value in diabetes management, where it may help identify patients with significant residual cardiovascular risk despite standard glycemic control, potentially guiding more aggressive lipid management and cardioprotective therapy.
Furthermore, the dose-response relationship was examined in detail using restricted cubic splines. This analysis confirmed a strong, positive, and monotonic association between GLMI and the odds of prevalent CVD. Critically, the formal test indicated no significant deviation from linearity (P for nonlinearity = 0.82), supporting that the relationship is approximately linear across the observed range of GLMI. Visually, the curve suggests a gradual increase in risk across most of the GLMI distribution. While visual inspection suggests the slope might increase at very high GLMI levels, this apparent change was not statistically significant and thus does not represent a validated threshold or inflection point. Therefore, the spline analysis robustly supports the primary finding of a continuous, linear increase in cardiovascular risk with higher GLMI levels in this population.
In practice, patients with higher GLMI levels may warrant more aggressive management of their composite metabolic risk, including intensified lifestyle interventions and consideration of cardioprotective medications, irrespective of individual risk factors meeting treatment thresholds. Our study also suggested GLMI exhibited significant positive correlations or trends of positive correlation with CVD mortality, CHF, CHD, stroke and angina. It indicated that the utility of the GLMI in the monitoring of odds of prevalent CVD related to IR.
The significant interaction between IR and hypertension stems from a synergistic, vicious cycle where metabolic and hemodynamic pathologies amplify one another [27]. Hypertension exacerbates the adverse effects of IR by increasing shear stress on the endothelium, promoting vascular remodeling, and enhancing oxidative stress. Conversely, IR-induced endothelial dysfunction and arterial stiffness contribute to the development and progression of hypertension [28],. With the development of hypertension and IR, the CVD risk was enhanced [29, 30]. Accordingly, in the current study, there was a significant positive association between GLMI level and the prevalence of CVD among patients with hypertension.
In addition, we compared the discriminative value for CVD prevalence between GLMI and other IR metrics including TyG index and TG/HDL ratio. In the present study, we found that GLMI had superior discriminatory ability comparing with TyG index and TG/HDL ratio in risk discrimination for CVD. This further emphasizes that GLMI captures a broader metabolic context associated with CVD prevalence.
The primary analysis of this study is cross-sectional; therefore, while a significant association between GLMI and prevalent CVD was observed, the design does not allow for the establishment of causality or temporal sequence, and the findings may be susceptible to reverse causation. Although we controlled for several confounders during the course of the study, there may still be possible residual confounders that could potentially affect the results. Also, because data on GLMI and related covariates were collected only at baseline, they failed to reflect dynamic changes in these indicators during follow-up, which may have led to an underestimation of the strength of the association. In addition, the lack of information on key confounders such as family history, detailed drug use, the presence of diabetes-related complications and duration of diabetes or prediabetes in the NHANES database may also impose some limitations on the accuracy of the study results. Finally, it should be noted that this study was conducted primarily in the United States, and the applicability of its findings to populations of other races or regions requires further validation.
Conclusion
In conclusion, the GLMI is positively associated with odds of prevalent CVD in participants with PreDM and DM, indicating its potential as a marker associated with the presence and burden of CVD in clinical practice. Considering the increasing prevalence of CVD worldwide in participants with PreDM and DM, the GLMI might be a promising and easy-to-measure serum marker for the identification of individuals with elevated CVD risk in public health settings. Large prospective studies are necessary to determine whether GLMI is associated with incident CVD and to explore the association of GLMI with CVD risk in individuals with normal glycemia.
Supplementary Information
Below is the link to the electronic supplementary material.
Author contributions
Ruishan Liu: Methodology, Formal Analysis, Investigation, Data Curation, Writing – Original Draft, Writing – Review & Editing, Visualization Peiyao Ma: Conceptualization ,Formal Analysis , Project Administration, Investigation, Writing – Review & Editing Qiang Zhao: Methodology, Data Curation, Visualization, Project Administration Kun Wang: Investigation, Visualization Dongbin Xiao: Methodology, Data Curation, Visualization, Investigation, Project Administration Fei He: Methodology, Formal Analysis, Investigation, Shenke Kong: Methodology, Conceptualization ,Formal Analysis, Project Administration, Investigation, Writing – Review & Editing, Visualization
Funding
This research was not funded by any funds.
Data availability
Publicly available datasets were analyzed in this study. These data are accessible at the following website: www.cdc.gov/nchs/nhanes/.
Declarations
Ethics approval and consent to participate
The National Center for Health Statistics and the Ethics Review Board approved the protocol for NHANES, and all participants provided written informed consent. The authors have disclosed no conflicts of interest. This study was conducted in accordance with the Declaration of Helsinki.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Peiyao Ma, Email: mayao0036@163.com.
Shenke Kong, Email: 18883368582@163.com.
References
- 1.Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, Barengo NC, Beaton AZ, Benjamin EJ, Benziger CP, et al. Global burden of cardiovascular diseases and risk Factors, 1990–2019: update from the GBD 2019 study. J Am Coll Cardiol. 2020;76:2982–3021. 10.1016/j.jacc.2020.11.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lloyd-Jones DM, Hong Y, Labarthe D, Mozaffarian D, Appel LJ, Van Horn L, Greenlund K, Daniels S, Nichol G, Tomaselli GF, et al. Defining and setting National goals for cardiovascular health promotion and disease reduction: the American heart association’s strategic impact goal through 2020 and beyond. Circulation. 2010;121:586–613. 10.1161/circulationaha.109.192703. [DOI] [PubMed] [Google Scholar]
- 3.Liu C, Liang D. The association between the triglyceride-glucose index and the risk of cardiovascular disease in US population aged ≤ 65 years with prediabetes or diabetes: a population-based study. Cardiovasc Diabetol. 2024;23:168. 10.1186/s12933-024-02261-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dang K, Wang X, Hu J, Zhang Y, Cheng L, Qi X, Liu L, Ming Z, Tao X, Li Y. The association between triglyceride-glucose index and its combination with obesity indicators and cardiovascular disease: NHANES 2003–2018. Cardiovasc Diabetol. 2024;23:8. 10.1186/s12933-023-02115-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Banerjee D, Biggs ML, Mercer L, Mukamal K, Kaplan R, Barzilay J, Kuller L, Kizer JR, Djousse L, Tracy R, et al. Insulin resistance and risk of incident heart failure: cardiovascular health study. Circulation Heart Fail. 2013;6:364–70. 10.1161/circheartfailure.112.000022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tian X, Chen S, Wang P, Xu Q, Zhang Y, Luo Y, Wu S, Wang A. Insulin resistance mediates obesity-related risk of cardiovascular disease: a prospective cohort study. Cardiovasc Diabetol. 2022;21:289. 10.1186/s12933-022-01729-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bonora E, Targher G, Alberiche M, Bonadonna RC, Saggiani F, Zenere MB, Monauni T, Muggeo M. Homeostasis model assessment closely mirrors the glucose clamp technique in the assessment of insulin sensitivity: studies in subjects with various degrees of glucose tolerance and insulin sensitivity. Diabetes Care. 2000;23:57–63. 10.2337/diacare.23.1.57. [DOI] [PubMed] [Google Scholar]
- 8.Li H, Zuo Y, Qian F, Chen S, Tian X, Wang P, Li X, Guo X, Wu S, Wang A. Triglyceride-glucose index variability and incident cardiovascular disease: a prospective cohort study. Cardiovasc Diabetol. 2022;21:105. 10.1186/s12933-022-01541-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Simental-Mendía LE, Rodríguez-Morán M, Guerrero-Romero F. The product of fasting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects. Metab Syndr Relat Disord. 2008;6:299–304. 10.1089/met.2008.0034. [DOI] [PubMed] [Google Scholar]
- 10.Pacini G, Mari A. Methods for clinical assessment of insulin sensitivity and beta-cell function. Best Pract Res Clin Endocrinol Metab. 2003;17:305–22. 10.1016/s1521-690x(03)00042-3. [DOI] [PubMed] [Google Scholar]
- 11.Reaven GM. The insulin resistance syndrome: definition and dietary approaches to treatment. Annu Rev Nutr. 2005;25:391–406. 10.1146/annurev.nutr.24.012003.132155. [DOI] [PubMed] [Google Scholar]
- 12.Wang H, Ruan S, Wu Z, Yan Q, Chen Y, Cui J, Zhang Z, Huang S, Hou B, Zhang C. Prognostic significance of glucose-lipid metabolic index in pancreatic cancer patients with diabetes mellitus. Cancer Med. 2024;13:e7108. 10.1002/cam4.7108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Che B, Zhong C, Zhang R, Pu L, Zhao T, Zhang Y, Han L. Triglyceride-glucose index and triglyceride to high-density lipoprotein cholesterol ratio as potential cardiovascular disease risk factors: an analysis of UK biobank data. Cardiovasc Diabetol. 2023;22:34. 10.1186/s12933-023-01762-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang T, Li M, Zeng T, Hu R, Xu Y, Xu M, Zhao Z, Chen Y, Wang S, Lin H, et al. Association between insulin resistance and cardiovascular disease risk varies according to glucose tolerance status: A nationwide prospective cohort study. Diabetes Care. 2022;45:1863–72. 10.2337/dc22-0202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hill MA, Yang Y, Zhang L, Sun Z, Jia G, Parrish AR, Sowers JR. Insulin resistance, cardiovascular stiffening and cardiovascular disease. Metabolism. 2021;119:154766. 10.1016/j.metabol.2021.154766. [DOI] [PubMed] [Google Scholar]
- 16.Tobin AM, Hackett CB, Rogers S, Collins P, Richards HL, O’Shea D, Kirby B. Body mass index, waist circumference and HOMA-IR correlate with the psoriasis area and severity index in patients with psoriasis receiving phototherapy. Br J Dermatol. 2014;171:436–8. 10.1111/bjd.12914. [DOI] [PubMed] [Google Scholar]
- 17.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;95:3347–51. 10.1210/jc.2010-0288. [DOI] [PubMed] [Google Scholar]
- 18.Huang X, He J, Wu G, Peng Z, Yang B, Ye L. TyG-BMI and hypertension in normoglycemia subjects in japan: A cross-sectional study. Diab Vasc Dis Res. 2023;20:14791641231173617. 10.1177/14791641231173617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wu S, Xu L, Wu M, Chen S, Wang Y, Tian Y. Association between triglyceride-glucose index and risk of arterial stiffness: a cohort study. Cardiovasc Diabetol. 2021;20:146. 10.1186/s12933-021-01342-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Park K, Ahn CW, Lee SB, Kang S, Nam JS, Lee BK, Kim JH, Park JS. Elevated TyG index predicts progression of coronary artery calcification. Diabetes Care. 2019;42:1569–73. 10.2337/dc18-1920. [DOI] [PubMed] [Google Scholar]
- 21.Lu YW, Chang CC, Chou RH, Tsai YL, Liu LK, Chen LK, Huang PH, Lin SJ. Gender difference in the association between TyG index and subclinical atherosclerosis: results from the I-Lan longitudinal aging study. Cardiovasc Diabetol. 2021;20:206. 10.1186/s12933-021-01391-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Liu X, Tan Z, Huang Y, Zhao H, Liu M, Yu P, Ma J, Zhao Y, Zhu W, Wang J. Relationship between the triglyceride-glucose index and risk of cardiovascular diseases and mortality in the general population: a systematic review and meta-analysis. Cardiovasc Diabetol. 2022;21:124. 10.1186/s12933-022-01546-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ding X, Wang X, Wu J, Zhang M, Cui M. Triglyceride-glucose index and the incidence of atherosclerotic cardiovascular diseases: a meta-analysis of cohort studies. Cardiovasc Diabetol. 2021;20:76. 10.1186/s12933-021-01268-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Alizargar J, Bai CH, Hsieh NC, Wu SV. Use of the triglyceride-glucose index (TyG) in cardiovascular disease patients. Cardiovasc Diabetol. 2020;19:8. 10.1186/s12933-019-0982-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Cho YR, Ann SH, Won KB, Park GM, Kim YG, Yang DH, Kang JW, Lim TH, Kim HK, Choe J, et al. Association between insulin resistance, hyperglycemia, and coronary artery disease according to the presence of diabetes. Sci Rep. 2019;9:6129. 10.1038/s41598-019-42700-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sánchez-Íñigo L, Navarro-González D, Fernández-Montero A, Pastrana-Delgado J, Martínez JA. The TyG index May predict the development of cardiovascular events. Eur J Clin Invest. 2016;46:189–97. 10.1111/eci.12583. [DOI] [PubMed] [Google Scholar]
- 27.Jia G, Sowers JR. Hypertension in Diabetes: An Update of Basic Mechanisms and Clinical Disease. Hypertension (Dallas, Tex: 1979). 2021;78:1197–1205. 10.1161/hypertensionaha.121.17981 [DOI] [PMC free article] [PubMed]
- 28.Artunc F, Schleicher E, Weigert C, Fritsche A, Stefan N, Häring HU. The impact of insulin resistance on the kidney and vasculature. Nat Rev Nephrol. 2016;12:721–37. 10.1038/nrneph.2016.145. [DOI] [PubMed] [Google Scholar]
- 29.Arnlöv J, Pencina MJ, Nam BH, Meigs JB, Fox CS, Levy D, D’Agostino RB, Vasan RS. Relations of insulin sensitivity to longitudinal blood pressure tracking: variations with baseline age, body mass index, and blood pressure. Circulation. 2005;112:1719–27. 10.1161/circulationaha.105.535039. [DOI] [PubMed] [Google Scholar]
- 30.Zhou Z, Liu Q, Zheng M, Zuo Z, Zhang G, Shi R, Wu T. Comparative study on the predictive value of TG/HDL-C, TyG and TyG-BMI indices for 5-year mortality in critically ill patients with chronic heart failure: a retrospective study. Cardiovasc Diabetol. 2024;23:213. 10.1186/s12933-024-02308-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Publicly available datasets were analyzed in this study. These data are accessible at the following website: www.cdc.gov/nchs/nhanes/.