Abstract
Background:
The triglyceride-glucose (TyG) index is a recognized surrogate marker of insulin resistance but lacks integration of high-density lipoprotein cholesterol (HDL-C) and adiposity measures, which are pivotal in the pathogenesis of diabetic kidney disease (DKD). The novel triglyceride-high-density lipoprotein-glucose-body (TyHGB) index, combining TG/HDL-C ratio, fasting blood glucose (FBG), and body mass index (BMI), may offer a more comprehensive metabolic profile. This study aimed to evaluate the associative value of TyHGB for DKD in type 2 diabetes mellitus (T2DM) patients.
Methods:
A retrospective cross-sectional analysis of 1,382 adults with T2DM was conducted. We employed multivariable logistic regression, restricted cubic spline (RCS) analysis, and subgroup analyses to assess the independent and non-linear association of the TyHGB index with DKD. Receiver operating characteristic (ROC) curves, net reclassification improvement (NRI), and integrated discrimination improvement (IDI) were used to evaluate and compare its associative performance against the TyG index.
Results:
Among the participants, 286 (20.7%) were diagnosed with DKD. After full adjustment for demographic, clinical, and biochemical confounders, TyHGB was independently associated with DKD (OR = 1.11, 95%CI:1.05-1.17, p<0.001). RCS analysis revealed a significant non-linear relationship, with a sharp increase in DKD risk beyond a TyHGB threshold of 8.74. The TyHGB index demonstrated superior discriminative ability (AUC = 0.775, 95% CI: 0.747-0.803) compared to the TyG index (AUC = 0.644, p<0.001). Incorporating TyHGB into a baseline clinical model significantly improved risk association (AUC increased from 0.715 to 0.788, p<0.001) and provided substantial reclassification improvement (NRI = 0.647, IDI = 0.067).
Conclusion:
The TyHGB index exhibits a robust, independent, and non-linear association with DKD risk in T2DM patients and outperforms the established TyG index. As a readily accessible composite metric, it holds significant promise as a superior tool for early identification and risk stratification of DKD in clinical practice.
1 Introduction
The escalating global prevalence of Type 2 diabetes mellitus (T2DM) presents a paramount challenge to public health systems worldwide. According to International Diabetes Federation estimates, approximately 537 million adults were living with diabetes in 2021, with projections indicating this figure will rise to 783 million by 2045 (). A particularly concerning complication is diabetic kidney disease (DKD), which affects approximately 40% of individuals with diabetes () and represents the principal cause of chronic kidney disease and end-stage renal disease globally (). Beyond substantially elevating cardiovascular event risks and all-cause mortality (, ), DKD imposes devastating financial burdens on healthcare systems and society at large (). Consequently, early identification of high-risk individuals and implementation of timely interventions hold critical importance for both clinical practice and public health initiatives.
DKD arises from multifactorial pathogenesis, characterized by a complex interplay among metabolic imbalances, hemodynamic changes, chronic inflammatory states, and oxidative stress (, ). Insulin resistance (IR) serves as a fundamental pathophysiological mechanism in both T2DM development and DKD progression, promoting renal injury through processes including glomerular hyperfiltration, endothelial dysfunction, and enhanced fibrogenesis (). Although gold-standard IR assessment methods like the hyperinsulinemic-euglycemic clamp offer high accuracy (), their technical complexity and substantial costs limit routine clinical application. This limitation has motivated the development of straightforward surrogate biomarkers derived from commonly measured clinical parameters.
Among emerging biomarkers, the triglyceride-glucose (TyG) index—calculated as Ln[fasting triglycerides (TG, mg/dL)×fasting blood glucose (FBG, mg/dL)/2]—has gained recognition as a reliable and readily accessible IR indicator (). Numerous studies have documented strong associations between the TyG index and various metabolic disorders, including non-alcoholic fatty liver disease (NAFLD) and cardiovascular diseases (, ). Despite these advantages, the TyG index exhibits certain limitations. Its primary focus on glucose and triglyceride metabolism potentially overlooks other crucial pathophysiological pathways contributing to DKD. For instance, high-density lipoprotein cholesterol (HDL-C) demonstrates anti-inflammatory, antioxidant, and endothelial-protective properties, with low HDL-C levels independently correlating with albuminuria development and renal function decline (, ). Additionally, obesity—typically quantified by body mass index (BMI)—represents a well-established DKD risk factor that promotes renal damage through mechanisms such as hyperfiltration, leptin resistance, and chronic inflammation (). The conventional TyG index fails to incorporate these two critical dimensions of HDL-C metabolism and adiposity, thereby constraining its comprehensive risk stratification potential.
To address these limitations, researchers recently proposed a novel composite index: the triglyceride high-density cholesterol-glucose body index (TyHGB). The TyHGB index is defined by the formula: TG/HDL-C + 0.7×FBG (mmol/L) + 0.1×BMI (kg/m²). This formula, including the specific coefficients (0.7 and 0.1) for FBG and BMI, was originally developed and validated by Xu et al. () through multivariate linear regression analysis, aiming to optimize the assessment of insulin resistance in the context of gestational diabetes mellitus. This single metric thereby encapsulates the key dimensions of dyslipidemia, dysglycemia, and adiposity. Initially validated for assessing insulin resistance in gestational diabetes mellitus, TyHGB has subsequently demonstrated superior performance over the TyG index in associating with NAFLD across diverse populations (, ). Since all required components are routinely measured in standard clinical practice, TyHGB represents a cost-effective and readily implementable tool.
Given the shared metabolic underpinnings of NAFLD and DKD—particularly regarding IR, dyslipidemia, and chronic inflammation—we hypothesize that TyHGB may similarly serve as a powerful marker for DKD in T2DM patients. However, no investigation to date has examined the relationship between TyHGB and DKD. Therefore, this study aimed to comprehensively evaluate TyHGB’s associative value for DKD in a large T2DM patient cohort, while comparing its performance against the established TyG index. To date, no established clinical cut-off points for TyHGB exist, and its association with DKD remains unexplored.
2 Materials and methods
2.1 Study design and population
We conducted a retrospective cross-sectional analysis using de-identified electronic health records sourced from a network of community health centers. The investigation included 1,382 adults diagnosed with type 2 diabetes mellitus who received medical services between 2023 and 2024. Data collection encompassed five community health centers affiliated with the Longhua District People’s Hospital of Shenzhen and one center under the Eighth Affiliated Hospital of Sun Yat-sen University (Futian).
Eligibility criteria required participants to be: (1) aged 18 years or older, and (2) diagnosed with T2DM according to American Diabetes Association standards (). Exclusion criteria encompassed: (1) diagnosis of type 1 diabetes, gestational diabetes, or other specific diabetes types; (2) pre-existing chronic kidney disease from non-diabetic etiologies (e.g., glomerulonephritis, polycystic kidney disease); (3) end-stage renal disease requiring renal replacement therapy; (4) acute diabetic complications (including ketoacidosis or hyperosmolar state) at data collection; (5) severe systemic conditions such as heart failure (NYHA Class III/IV), advanced liver cirrhosis, autoimmune diseases, hematological malignancies, or active cancer; (6) active infectious diseases, systemic inflammatory conditions, or those on immunosuppressants; and (7) incomplete data necessary for TyHGB index calculation or DKD diagnosis.
2.2 Ethics statement
The study protocol obtained approval from the Medical Ethics Committee of Longhua District People’s Hospital of Shenzhen (Approval No.: Longhua District People’s Hospital (No [2025].087)). Formal data access authorization was secured through a written agreement with the Data Governance Committee of the Eighth Affiliated Hospital of Sun Yat-sen University (Futian) (Approval No.:SYSU8H-DGC- [2025]DA-013). All data handling complied strictly with institutional security protocols. All data handling complied with institutional security protocols and relevant data protection regulations, including the Protection of Personal Information Act (POPIA) where applicable. The research adhered fully to ethical principles outlined in the Declaration of Helsinki and relevant international guidelines governing human subjects research and retrospective data analysis. The ethics committees waived individual informed consent requirements due to the use of de-identified retrospective data.
It is hereby clarified that this ethical approval (2025) pertains specifically to the retrospective analysis of pre-existing data from the 2023–2024 period.
2.3 Data collection and variable definitions
The electronic health records provided information across three domains:
Demographics and History: Sex, age, T2DM duration, hypertension history, cardiovascular disease, stroke, smoking status, and alcohol consumption.
Anthropometric Measures: Body mass index (BMI), computed as weight in kilograms divided by height in meters squared (kg/m²).
Laboratory Parameters: Fasting blood glucose (FBG), glycated hemoglobin (HbA1c), total bilirubin (TBIL), alanine aminotransferase (ALT), aspartate aminotransferase (AST), triglycerides (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), blood urea nitrogen (BUN), serum creatinine (SCr), uric acid (UA), white blood cell count (WBC), and platelet count (PLT).
We calculated the novel metabolic index, the triglyceride high-density cholesterol-glucose body index (TyHGB), according to the following formula: TyHGB = TG (mmol/L)/HDL-C (mmol/L) + 0.7×FBG (mmol/L) + 0.1×BMI (kg/m²). For comparative purposes, we computed the established Triglyceride-Glucose (TyG) index as: TyG = Ln [TG (mg/dL)×FBG (mg/dL)/2]. Note that the original formulas for TyHGB and TyG indices were developed using different units systems (mmol/L and mg/dL, respectively). In our calculations, all laboratory values were recorded in mmol/L. To compute the TyG index according to its standard formula, TG and FBG values were converted to mg/dL using the standard conversion factors (TG: 1 mmol/L = 88.57 mg/dL; FBG: 1 mmol/L = 18 mg/dL).
2.4 Outcome definition
The primary outcome was diabetic kidney disease (DKD) presence (). The diagnosis required meeting one of two criteria: (1) a urinary albumin-to-creatinine ratio (UACR) of ≥30 mg/g, confirmed on at least two of three tests over a 3–6 month period without other identifiable causes; or (2) a persistently reduced estimated glomerular filtration rate (eGFR) below 60 mL/min/1.73 m² for over three months, as determined by the CKD-EPI equation. Based on these criteria, we categorized participants into DKD and non-DKD groups.
2.5 Statistical analysis
All statistical analyses were carried out with R software (version 4.3.1; R Foundation for Statistical Computing). A two-sided p-value < 0.05 indicated statistical significance. Normality of continuous variables was assessed using the Shapiro-Wilk test.
Baseline Characteristics: For continuous variables, normally distributed data are summarized as mean ± standard deviation, while non-normally distributed data are reported as median with interquartile range. Group comparisons were performed using the Student’s t-test and the Mann-Whitney U test, respectively. We expressed categorical variables as frequencies (percentages) and compared them via Chi-square tests.
Multivariable Logistic Regression: To quantify the independent association between TyHGB and DKD, we employed multivariable logistic regression with three sequentially adjusted models: Model 1 accounted for basic demographics (age and sex); Model 2 added lifestyle factors and comorbidities (hypertension, cardiovascular disease, stroke, smoking status, alcohol consumption, and BMI); Model 3 represented the primary fully adjusted model, which incorporated HbA1c in addition to Model 2 variables. Renal function markers (serum creatinine and blood urea nitrogen) were intentionally excluded from this primary model to avoid over-adjustment bias, as they are on the causal pathway of the outcome. To assess the robustness of our findings, we conducted a sensitivity analysis (Model 3a) that additionally adjusted for serum creatinine and blood urea nitrogen. Results appear as odds ratios (ORs) with corresponding 95% confidence intervals (CIs).
Restricted Cubic Spline (RCS) Analysis: We explored potential non-linear relationships between continuous TyHGB values and DKD risk by fitting multivariable logistic regression models incorporating RCS with four knots (at the 5th, 35th, 65th, and 95th percentiles). The likelihood ratio test assessed non-linearity significance.
Subgroup Analysis and Interaction: To evaluate association robustness across patient strata, we conducted subgroup analyses based on sex, BMI (<24 vs≥24 kg/m²), and hypertension status. After introducing interaction terms between TyHGB and subgroup variables into multivariable models, we used likelihood ratio tests to determine interaction significance.
Predictive Performance and Model Comparison: Using receiver operating characteristic (ROC) curve analysis, we evaluated and compared the associative abilities of TyHGB, the TyG index, and other individual indicators for DKD. We calculated the area under the ROC curve (AUC) and employed DeLong’s test for pairwise comparisons. Additionally, we constructed three nested association models:
Model A (Baseline): Incorporated traditional risk factors (hypertension, smoking, T2DM duration, HbA1c, age, total cholesterol, and serum creatinine).
Model B (Baseline + TyG): Enhanced the baseline model with the TyG index.
Model C (Baseline + TyHGB): Enhanced the baseline model with the TyHGB index.
The discrimination performance of these nested models was internally validated using the bootstrap method with 1000 re-samplings. Model discrimination capabilities were then compared using the AUC values derived from this validation.
Reclassification Improvement: We quantitatively assessed the incremental associative value of adding TyHGB to both the baseline model and the TyG-enhanced model using continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI).
3 Results
3.1 Study population characteristics
Our retrospective cross-sectional analysis comprised 1,382 adults diagnosed with type 2 diabetes mellitus. Within this population, 286 patients (20.7%) received a diagnosis of diabetic kidney disease. Table 1 presents the baseline characteristics according to DKD status. The DKD group demonstrated significantly longer diabetes duration and elevated concentrations of multiple metabolic parameters compared to the non-DKD group. Specifically, we observed higher values for body mass index, fasting blood glucose, glycated hemoglobin, liver enzymes (ALT and AST), triglycerides, serum creatinine, uric acid, white blood cell count, and both TyG and TyHGB indices (all p<0.05). Conversely, high-density lipoprotein cholesterol levels were markedly reduced in DKD patients (p<0.05). Several parameters showed no significant intergroup differences, including age, cardiovascular disease or stroke prevalence, total bilirubin, total cholesterol, low-density lipoprotein cholesterol, blood urea nitrogen, and platelet count (all p > 0.05).
Table 1
| Variable | Overall1 | Non-DKD1 | DKD1 | P Value2 |
|---|---|---|---|---|
| Sex | <0.001 | |||
| 0 | 538 (39%) | 453 (41%) | 85 (30%) | |
| 1 | 843 (61%) | 642 (59%) | 201 (70%) | |
| Age (years) | 57.29 (9.83) | 57.16 (9.57) | 57.79 (10.78) | 0.6 |
| Duration (years) | 7.31 (5.89) | 7.05 (5.66) | 8.30 (6.59) | 0.013 |
| Hypertension | <0.001 | |||
| 0 | 723 (52%) | 607 (55%) | 116 (41%) | |
| 1 | 658 (48%) | 488 (45%) | 170 (59%) | |
| Cardiovascular Disease | 0.5 | |||
| 0 | 1,250 (91%) | 994 (91%) | 256 (90%) | |
| 1 | 131 (9.5%) | 101 (9.2%) | 30 (10%) | |
| Stroke | 0.4 | |||
| 0 | 1,303 (94%) | 1,036 (95%) | 267 (93%) | |
| 1 | 78 (5.6%) | 59 (5.4%) | 19 (6.6%) | |
| Smoking | <0.001 | |||
| 0 | 957 (69%) | 790 (72%) | 167 (58%) | |
| 1 | 424 (31%) | 305 (28%) | 119 (42%) | |
| Alcohol | 0.008 | |||
| 0 | 989 (72%) | 802 (73%) | 187 (65%) | |
| 1 | 391 (28%) | 292 (27%) | 99 (35%) | |
| Unknown | 1 | 1 | 0 | |
| BMI (kg/m²) | 24.98 (3.24) | 24.77 (3.16) | 25.77 (3.44) | <0.001 |
| FBG (mmol/L) | 7.49 (2.24) | 7.33 (2.10) | 8.13 (2.63) | <0.001 |
| HbA1c (%) | 7.15 (1.43) | 7.04 (1.37) | 7.55 (1.59) | <0.001 |
| TBIL (μmol/L) | 13.08 (4.85) | 13.15 (4.87) | 12.82 (4.79) | 0.3 |
| ALT (U/L) | 24.99 (15.19) | 24.38 (14.47) | 27.32 (17.51) | 0.015 |
| AST (U/L) | 22.16 (8.95) | 21.74 (7.82) | 23.73 (12.25) | 0.012 |
| TG (mmol/L) | 1.81 (1.74) | 1.67 (1.46) | 2.33 (2.46) | <0.001 |
| TC (mmol/L) | 4.77 (1.16) | 4.74 (1.11) | 4.92 (1.33) | 0.2 |
| HDL (mmol/L) | 1.28 (0.39) | 1.30 (0.40) | 1.21 (0.32) | <0.001 |
| LDL (mmol/L) | 3.02 (0.97) | 3.01 (0.97) | 3.06 (0.97) | 0.4 |
| BUN (mmol/L) | 6.01 (2.74) | 5.91 (2.62) | 6.36 (3.14) | 0.2 |
| SCr (μmol/L) | 73.70 (37.38) | 70.41 (18.86) | 86.29 (72.11) | <0.001 |
| UA (μmol/L) | 349.41 (90.60) | 343.54 (88.81) | 371.85 (93.99) | <0.001 |
| WBC (×109/L) | 6.39 (1.76) | 6.26 (1.70) | 6.90 (1.89) | <0.001 |
| PLT (×109/L) | 230.66 (61.77) | 230.59 (62.62) | 230.94 (58.47) | 0.7 |
| TyG Index | 9.04 (0.70) | 8.96 (0.67) | 9.32 (0.75) | <0.001 |
| TyHGB Index | 9.36 (2.65) | 9.08 (2.43) | 10.40 (3.15) | <0.001 |
Baseline characteristics of study participants.
1n (%); Mean (SD).
2Pearson's Chi-squared test; Wilcoxon rank sum test.
3.2 Multivariable analysis of TyHGB-DKD association
We employed multivariable logistic regression to examine the independent relationship between TyHGB and DKD (Figure 1). This association maintained statistical significance across progressively adjusted models:
Figure 1
Model 1 (demographic adjustment): OR = 1.18, 95% CI 1.12-1.24, p < 0.001.
Model 2 (lifestyle and comorbidity adjustment): OR = 1.16, 95% CI 1.10-1.21, p < 0.001.
Model 3 (primary fully adjusted model, excluding renal function markers): OR = 1.11, 95% CI 1.05-1.17, p < 0.001.
Sensitivity Analysis: When renal function markers (serum creatinine and blood urea nitrogen) were added to the fully adjusted model (Model 3a), the association between TyHGB and DKD remained virtually unchanged and highly significant (OR = 1.11, 95% CI: 1.05-1.17, p < 0.001), confirming the robustness of our primary finding.
3.3 Non-linear dose-response relationship
Restricted cubic spline analysis identified a significant non-linear pattern in the TyHGB-DKD relationship (p = 0.002). As illustrated in Figure 2, DKD risk exhibited gradual increase until reaching a TyHGB threshold of approximately 8.74, beyond which risk escalation accelerated substantially with increasing TyHGB values. The overall association demonstrated high significance (p < 0.001), supporting a robust dose-response relationship. This non-linear configuration persisted following comprehensive adjustment for demographic, clinical, and biochemical covariates.
Figure 2
3.4 Subgroup analyses and effect modification
Stratified analyses evaluated the consistency of TyHGB-DKD associations across patient subgroups (Figure 3). Hypertension status significantly influenced this relationship (p for interaction = 0.037), with hypertensive patients showing stronger positive associations (OR = 1.213, 95% CI 1.040-1.415, p = 0.014) compared to normotensive individuals (OR = 0.835, 95% CI 0.702-0.992, p = 0.040). In contrast, neither sex (p = 0.307) nor obesity status (p = 0.990) demonstrated significant effect modification. The direction of association remained consistent across all subgroups, though statistical significance was not achieved in sex- or BMI-stratified analyses.
Figure 3
3.5 Associative performance assessment
Receiver operating characteristic analysis evaluated the discriminative capacity of TyHGB against established biomarkers (Figure 4, Table 2). TyHGB achieved superior performance (AUC = 0.775, 95% CI: 0.747-0.803) compared to both the TyG index (AUC = 0.644, 95% CI: 0.608-0.681; p < 0.001) and serum creatinine (AUC = 0.625, 95% CI: 0.587-0.662; p < 0.001), establishing its advantage as a standalone marker.
Figure 4
Table 2
| Analysis type | Metric | Value | X95..CI...Components | P.value |
|---|---|---|---|---|
| Single Marker | TyHGB AUC | 0.775 | 0.747-0.803 | Reference |
| TyG AUC | 0.644 | 0.608-0.681 | <0.001 | |
| SCr AUC | 0.625 | 0.587-0.662 | <0.001 | |
| Nested Model | Model A AUC | 0.715 | 0.682-0.749 | Reference |
| Model B AUC | 0.728 | 0.696-0.761 | 0.060 | |
| Model C AUC | 0.788 | 0.759-0.817 | <0.001 | |
| Improvement | NRI (Model C vs A) | 0.647 | Events: 0.252, Non-events: 0.395 | N/A* |
| IDI (Model C vs A) | 0.067 | ΔEvents: 0.053, ΔNon-events: -0.014 | N/A* |
Comprehensive statistical analysis results.
*P-values for NRI and IDI require bootstrap estimation for statistical significance.
Model A: Hypertension + Smoking + Duration + HbA1c +Age + TC + SCr.
Model B: Model A + TyG, Model C: Model A + TyHGB.
3.6 Model enhancement and reclassification
Nested model construction evaluated TyHGB’s incremental associative value (Figure 5):
Figure 5
Baseline model: AUC = 0.715 (0.682-0.749).
Baseline + TyG: AUC = 0.728 (0.696-0.761).
Baseline + TyHGB: AUC = 0.788 (0.759-0.817).
TyHGB incorporation yielded significant improvement over both the baseline model (ΔAUC = 0.073, p < 0.001) and the TyG-enhanced model (ΔAUC = 0.060, p < 0.001). Furthermore, TyHGB addition produced substantial reclassification improvement (continuous NRI = 0.647) and enhanced integrated discrimination (IDI = 0.067), confirming its value for risk stratification.
4 Discussion
This community-based investigation establishes a robust and independent relationship between TyHGB—a novel composite metabolic index—and diabetic kidney disease prevalence in adults with type 2 diabetes. The association remained statistically significant following extensive adjustment for demographic, clinical, and biochemical variables. We identified a non-linear dose-response pattern, with DKD risk accelerating substantially beyond a TyHGB threshold of approximately 8.7. However, this threshold should be considered exploratory and requires validation in independent, prospective cohorts. As an independent marker, TyHGB demonstrated enhanced discriminative ability for DKD compared to both the TyG index and serum creatinine, achieving an AUC of 0.775. Incorporation of TyHGB into conventional risk association models significantly improved both discrimination and reclassification accuracy. Collectively, these findings position TyHGB as a practical, integrative metabolic indicator that delivers meaningful incremental prognostic value for DKD risk assessment.
The complex pathogenesis of DKD involves multiple interconnected mechanisms, with insulin resistance, dyslipidemia, chronic inflammation, and oxidative stress representing key pathophysiological contributors (, ). The triglyceride-glucose index has emerged as a well-validated IR surrogate that correlates with DKD risk across multiple epidemiological studies (, ). Complementary evidence from Jiang et al. confirmed that both TyG and TyG-BMI independently associate with DKD in newly diagnosed type 2 diabetes patients, though with limited diagnostic accuracy (AUC = 0.57) (). Supporting these observations, Li et al. documented positive correlations between TyG index values and albuminuria severity (). However, the TyG index predominantly reflects glucose and triglyceride metabolism, potentially overlooking other critical pathways involving anti-inflammatory lipid regulation and adiposity-related mechanisms.
Our results reinforce the conceptual framework that composite indices capturing broader metabolic dysregulation may offer superior associative capability. TyHGB advances beyond conventional TyG metrics by incorporating the TG/HDL-C ratio—an established marker of atherogenic dyslipidemia and IR—together with fasting glucose and BMI measurements. This integrated approach enables TyHGB to represent the complex interplay among dyslipidemia, glucotoxicity, and adiposity. The inclusion of HDL-C is particularly relevant given its documented anti-inflammatory, antioxidant, and endothelial-protective functions, with low levels independently associated with renal functional decline (, ). Similarly, BMI serves as an indicator for obesity-mediated renal injury, which operates through glomerular hyperfiltration, leptin resistance, and chronic inflammatory pathways (). Thus, TyHGB provides a more comprehensive assessment of the multifaceted metabolic disturbances driving DKD progression.
Our restricted cubic spline analysis identified a significant non-linear relationship between TyHGB values and DKD risk. The observed threshold effect at TyHGB = 8.74 suggests a critical transition point in metabolic homeostasis, beyond which renal injury risk increases substantially. This pattern aligns with established concepts of threshold effects in microvascular complication pathogenesis. Previous investigations have documented similar non-linear associations for other metabolic indices, including the U-shaped relationship between TyG index and diabetic retinopathy reported by Zhou et al. (). Our findings extend this paradigm to the TyHGB-DKD relationship, emphasizing the importance of evaluating non-linearity in metabolic risk assessment.
Stratified analyses provided clinically important insights regarding effect modification. The strengthened TyHGB-DKD association in hypertensive individuals implies that metabolic dysregulation and hemodynamic stress may synergistically accelerate renal damage through multiple complementary pathways in the context of hypertension, thereby enhancing the TyHGB-DKD relationship. Chronic low-grade inflammation and upregulation of pro-inflammatory cytokines (e.g., IL-6, TNF-α) in the comorbidity of hypertension and insulin resistance amplify the glomerular and interstitial inflammatory injury triggered by dysglycemia, dyslipidemia, and obesity (). Concurrently, oxidative stress and lipid peroxidation can damage the glomerular basement membrane and podocytes, promoting proteinuria and the deterioration of filtration function (). Furthermore, sympathetic nervous system overactivation and hemodynamic alterations (glomerular hypertension, hyperfiltration) exacerbate these metabolically induced pathological processes through sustained mechanical stress and microcirculatory disturbances (). These interactions collectively manifest in shared pathways such as RAAS activation, mitochondrial dysfunction, and endothelial impairment, forming a pathogenic network that drives DKD progression. Conversely, the consistent associations observed across sex and BMI categories reinforce TyHGB’s generalizability as a risk marker, demonstrating its utility beyond specific demographic or adiposity subgroups.
TyHGB’s superior associative performance (AUC = 0.775) relative to TyG index (AUC = 0.644) and serum creatinine (AUC = 0.625) underscores its clinical potential. These observations are consistent with emerging research evaluating TyHGB in other contexts, including its excellent associative accuracy for gestational diabetes mellitus (AUC 0.707-0.859) as demonstrated by Xu et al. (). The conceptual approach of enhancing TyG through additional parameters follows established methodology, exemplified by Er and Li et al.’s development of TyG-BMI as a superior IR marker (, ); our work successfully translates and validates this approach specifically for DKD risk assessment.
The incorporation of TyHGB into baseline clinical models produced significant improvement in discriminant capacity (AUC increase from 0.715 to 0.788) and reclassification accuracy (NRI = 0.647; IDI = 0.067). These findings indicate that TyHGB captures unique risk information not fully represented by conventional factors, supporting its potential integration into clinical risk scores and electronic health record-based screening protocols.
Furthermore, our findings proved robust in a sensitivity analysis that included renal function markers in the adjustment model, mitigating concerns about over-adjustment bias. While our study benefits from real-world community data, comprehensive confounder adjustment, and sophisticated non-linearity testing that identified an actionable clinical threshold, several limitations merit consideration. The retrospective single-center design may affect generalizability, though multi-site data sourcing enhances real-world applicability. Potential unmeasured confounders, including dietary patterns, physical activity levels, and specific medications (e.g., SGLT2 inhibitors, GLP-1 receptor agonists), could influence observed relationships. Additionally, DKD diagnosis relied on UACR and eGFR criteria without biopsy confirmation, though this approach aligns with current clinical practice guidelines (). Future multi-center, prospective, and multi-ethnic studies, with serial TyHGB measurements, are needed to validate our findings and elucidate the causal mechanisms.
5 Conclusion
The triglyceride high-density cholesterol-glucose body index (TyHGB)—a readily calculable composite of TG/HDL-C ratio, fasting glucose, and BMI—demonstrates independent non-linear association with diabetic kidney disease in type 2 diabetes patients. It significantly enhances risk discrimination beyond conventional factors and established indices. TyHGB’s practical derivation and robust performance support its potential role in DKD screening and stratification protocols, warranting additional prospective validation and exploration within integrated assessment frameworks.
Statements
Data availability statement
The datasets presented in this article are not readily available because they constitute electronic health records sourced from a network of community health centers, containing sensitive patient information. The data were used under license for the current retrospective study in a de-identified format, in compliance with the ethical approvals and data access agreements granted by the Medical Ethics Committee of Longhua District People’s Hospital of Shenzhen and the Data Governance Committee of the Eighth Affiliated Hospital of Sun Yat-sen University (Futian). Requests to access these datasets should be directed to the corresponding author and will require a formal data sharing agreement and subsequent permission from the aforementioned ethics and data governance committees. Requests to access these datasets should be directed to the corresponding author, Email: 54574963@qq.com.
Ethics statement
The studies involving humans were approved by the Medical Ethics Committee of Longhua District People’s Hospital of Shenzhen. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin due to the use of de-identified retrospective data.
Author contributions
JY: Formal Analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing. BX: Conceptualization, Data curation, Project administration, Resources, Software, Writing – review & editing. ZD: Data curation, Formal Analysis, Investigation, Writing – review & editing. ZZ: Project administration, Resources, Software, Writing – original draft. HZ: Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This study was supported by Medical research project of Shenzhen Longhua Medical Association: 2024LHMA10.
Acknowledgments
We gratefully acknowledge the data support provided by the Department of General Medicine at Longhua District People’s Hospital of Shenzhen and the Department of General Medicine at the Eighth Affiliated Hospital of Sun Yat-sen University. We also extend our sincere gratitude to all the patients whose data were used in this retrospective study.
Conflict of interest
The authors declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1
SunHSaeediPKarurangaSPinkepankMOgurtsovaKDuncanBBet al. IDF Diabetes Atlas: Global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045. Diabetes Res Clin Pract. (2022) 183:109119. doi: 10.1016/j.diabres.2021.109119
2
AlicicRZRooneyMTTuttleKR. Diabetic kidney disease: challenges, progress, and possibilities. Clin J Am Soc Nephrol. (2017) 12:2032–45. doi: 10.2215/CJN.11491116
3
GuptaSDominguezMGolestanehL. Diabetic kidney disease: an update. Med Clin North Am. (2023) 107:689–705. doi: 10.1016/j.mcna.2023.03.004
4
JinQLauESHLukAOTamCHTOzakiRLimCKPet al. Circulating metabolomic markers linking diabetic kidney disease and incident cardiovascular disease in type 2 diabetes: analyses from the Hong Kong Diabetes Biobank. Diabetologia. (2024) 67:837–49. doi: 10.1007/s00125-024-06108-5
5
MinutoloRSimeonVDeNicola LChiodiniPGalieroRRinaldiLet al. Sex-difference of multifactorial intervention on cardiovascular and mortality risk in DKD: post-hoc analysis of a randomised clinical trial. Cardiovasc Diabetol. (2024) 23:285. doi: 10.1186/s12933-024-02371-3
6
HuRZhaoZXieLMaZWuWLiS. Global, regional, and national burden of chronic kidney disease due to diabetes mellitus type 2 from 1990 to 2021, with projections to 2036: a systematic analysis for the Global Burden of Disease Study 2021. Front Med (Lausanne). (2025) 12:1531811. doi: 10.3389/fmed.2025.1531811
7
EidSSasKMAbcouwerSFFeldmanELGardnerTWPennathurSet al. New insights into the mechanisms of diabetic complications: role of lipids and lipid metabolism. Diabetologia. (2019) 62:1539–49. doi: 10.1007/s00125-019-4959-1
8
HouGDongYJiangYZhaoWZhouLCaoSet al. Immune inflammation and metabolic interactions in the pathogenesis of diabetic nephropathy. Front Endocrinol (Lausanne). (2025) 16:1602594. doi: 10.3389/fendo.2025.1602594
9
AgarwalR. Pathogenesis of diabetic nephropathy. ADA Clin Compendia. (2021) 2021:2–7. doi: 10.2337/db20211-2
10
Domingues-HajjPMSGomesPMMagalhãesPKRMacielLMZ. Assessment of cardiometabolic risk factors and insulin sensitivity by hyperinsulinemic-euglycemic clamp in resistance to thyroid hormone β Syndrome. Thyroid. (2024) 34:1038–46. doi: 10.1089/thy.2024.0132
11
El-SehrawyAAMAKhachatryanLGKubaevARekhaMMRabSOKaurMet al. Triglyceride-glucose index: a potent predictor of metabolic risk factors and eating behavior patterns among obese individuals. BMC Endocr Disord. (2025) 25:71. doi: 10.1186/s12902-025-01887-3
12
ZhangQLiuZZhangJYangSLiuL. Association between TyG related parameters and metabolic dysfunction associated fatty liver disease among nondiabetic individuals. Sci Rep. (2025) 15:4566. doi: 10.1038/s41598-024-84917-9
13
ZhengXZhangWYangFWangLYuBLiangB. Evaluative performance of TyG-ABSI versus traditional indices in relation to cardiovascular disease and mortality: evidence from the U.S. NHANES. Cardiovasc Diabetol. (2025) 24:344. doi: 10.1186/s12933-025-02902-6
14
LiuHYaoXWangLLiuJLiXFuXet al. The causal relationship between 5 serum lipid parameters and diabetic nephropathy: a Mendelian randomization study. Front Endocrinol (Lausanne). (2024) 15:1358358. doi: 10.3389/fendo.2024.1358358
15
WangYLiS. Lipid metabolism disorders and albuminuria risk: insights from National Health and Nutrition Examination Survey 2001–2018 and Mendelian randomization analyses. Ren Fail. (2024) 46:2420841. doi: 10.1080/0886022X.2024.2420841
16
JiangWWangJShenXLuWWangYLiWet al. Establishment and validation of a risk prediction model for early diabetic kidney disease based on a systematic review and meta-analysis of 20 cohorts. Diabetes Care. (2020) 43:925–33. doi: 10.2337/dc19-1897
17
XuZLiCWangNSongG. Association between the new TyG indicator-TyHGB and gestational diabetes mellitus: results from the case-control and prospective cohort studies. J Transl Med. (2025) 23:190. doi: 10.1186/s12967-025-06166-2
18
ZhaoEWangRChenYXieHGaoYDongBet al. Evaluation of a new TyG indicator, TyHGB, in predicting non-alcoholic fatty liver disease: evidence from two independent populations. Lipids Health Dis. (2025) 24:342. doi: 10.1186/s12944-025-02742-z
19
SongSZhangYQiaoXDuoYXuJPengZet al. HOMA-IR as a risk factor of gestational diabetes mellitus and a novel simple surrogate index in early pregnancy. Int J Gynaecol Obstet. (2022) 157:694–701. doi: 10.1002/ijgo.13905
20
DaviesMJArodaVRCollinsBSGabbayRAGreenJMaruthurNMet al. Management of hyperglycemia in type 2 diabetes, 2022. A consensus report by the american diabetes association (ADA) and the european association for the study of diabetes (EASD). Diabetes Care. (2022) 45:2753–86. doi: 10.2337/dci22-0034
21
GuoS. Diagnosis and differential diagnosis of diabetic kidney disease. BMJ-British Med J. (2017) 16:750–2. doi: 10.3760/cma.j.issn.1671-7368.2017.10.002
22
JungCYYooTH. Pathophysiologic mechanisms and potential biomarkers in diabetic kidney disease. Diabetes Metab J. (2022) 46:181–97. doi: 10.4093/dmj.2021.0329
23
PennoGSoliniAOrsiEBonoraEFondelliCTrevisanRet al. Insulin resistance, diabetic kidney disease, and all-cause mortality in individuals with type 2 diabetes: a prospective cohort study. BMC Med. (2021) 19:66. doi: 10.1186/s12916-021-01936-3
24
LvLZhouYChenXGongLWuJLuoWet al. Relationship between the tyG index and diabetic kidney disease in patients with type-2 diabetes mellitus. Diabetes Metab Syndr Obes. (2021) 14:3299–306. doi: 10.2147/DMSO.S318255
25
LiuLXiaRSongXZhangBHeWZhouXet al. Association between the triglyceride-glucose index and diabetic nephropathy in patients with type 2 diabetes: A cross-sectional study. J Diabetes Investig. (2021) 12:557–65. doi: 10.1111/jdi.13371
26
JiangYLaiX. Association between the triglyceride glucose index, triglyceride-glucose body mass index and diabetic kidney disease in adults with newly diagnosed type 2 diabetes. Front Med (Lausanne). (2024) 11:1328601. doi: 10.3389/fmed.2024.1328601
27
LiXWangY. Associations of the TyG index with albuminuria and chronic kidney disease in patients with type 2 diabetes. PloS One. (2024) 19:e0312374. doi: 10.1371/journal.pone.0312374
28
RussoGTDeCosmo SViazziFPacilliACerielloAGenoveseSet al. AMD-Annals Study Group. Plasma Triglycerides and HDL-C Levels Predict the Development of Diabetic Kidney Disease in Subjects With Type 2 Diabetes: The AMD Annals Initiative. Diabetes Care. (2016) 39(12):2278–87. doi: 10.2337/dc16-1246
29
LeeSHKimHSParkYMKwonHSYoonKHHanKet al. HDL-Cholesterol, Its Variability, and the Risk of Diabetes: A Nationwide Population-Based Study. J Clin Endocrinol Metab. (2019) 104(11):5633–4. doi: 10.1210/jc.2019-01080
30
JiangZWangYZhaoXCuiHHanMRenXet al. Obesity and chronic kidney disease. Am J Physiol Endocrinol Metab. (2023) 324:E24–41. doi: 10.1152/ajpendo.00179.2022
31
ZhouYLuQZhangMYangLShenX. The U-shape relationship between triglyceride-glucose index and the risk of diabetic retinopathy among the US population. J Pers Med. (2023) 13:495. doi: 10.3390/jpm13030495
32
YuCZhouWZhouXZhuLWangTBaoHet al. Association Between Triglyceride Glucose Index and Chronic Kidney Disease in Normal-Weight Chinese Adults With Hypertension. J Clin Hypertens (Greenwich). (2024) 26(12):1433–40. doi: 10.1111/jch.14913.121.16498
33
FengAYinRXuRZhangBYangL. An update on renal tubular injury as related to glycolipid metabolism in diabetic kidney disease. Front Pharmacol. (2025) 16:1559026. doi: 10.3389/fphar.2025.1559026
34
WangHFanKNingYLiYQiXWangZet al. Hypertension and BMI as mediators of type 2 diabetes-induced CKD: insights from an integrative multi-database study. Diabetes Metab Syndr Obes. (2025) 18:3343–61. doi: 10.2147/DMSO.S538554
35
ErLKWuSChouHHHsuLATengMSSunYCet al. Triglyceride glucose-body mass index is a simple and clinically useful surrogate marker for insulin resistance in nondiabetic individuals. PloS One. (2016) 11:e0149731. doi: 10.1371/journal.pone.0149731
36
LiWShenCKongWZhouXFanHZhangYet al. Association between the triglyceride glucose-body mass index and future cardiovascular disease risk in a population with Cardiovascular-Kidney-Metabolic syndrome stage 0-3: a nationwide prospective cohort study. Cardiovasc Diabetol. (2024) 23:292. doi: 10.1186/s12933-024-02352-6
37
Kidney Disease: Improving Global Outcomes (KDIGO) Diabetes Work Group. KDIGO 2022 clinical practice guideline for diabetes management in chronic kidney disease. Kidney Int. (2022) 102:S1–S127. doi: 10.1016/j.kint.2022.06.008
Summary
Keywords
associative, diabetic kidney disease, insulin resistance, TyHGB index, type 2 diabetes
Citation
Yang J, Xie B, Deng Z, Zhang Z and Zhou H (2026) The triglyceride-high-density lipoprotein-glucose-body index: a superior novel biomarker for diabetic kidney disease in type 2 diabetes. Front. Endocrinol. 16:1749826. doi: 10.3389/fendo.2025.1749826
Received
19 November 2025
Revised
16 December 2025
Accepted
19 December 2025
Published
12 January 2026
Volume
16 - 2025
Edited by
Hsiang-Chen Wang, National Chung Cheng University, Taiwan
Reviewed by
Bo Jiang, Rice University, United States
Haskly Mokoena, University of Limpopo, South Africa
Updates
Copyright
© 2026 Yang, Xie, Deng, Zhang and Zhou.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Hairong Zhou, 54574963@qq.com
† These authors have contributed equally to this work
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.