Skip to content
Frontiers
Search

Original Research

Frontiers in Endocrinology ·

Volume 17 - 2026 | https://doi.org/10.3389/fendo.2026.1883785

Associations between a series of composite glucose-lipid indices and moderately increased albuminuria in patients with type 2 diabetes mellitus: a cross-sectional study

  • Yaru Li1,2
  • Zhenyu Pan3
  • Xinbo Zhao2
  • Guanqi Gao2
  • Dan Fu2
  • Baolan Ji2*corresponding author
  1. 1School of Clinical Medicine, Shandong Second Medical University, Weifang, Shandong, China
  2. 2Department of Endocrinology, Linyi People’s Hospital, Shandong Second Medical University, Linyi, Shandong, China
  3. 3Teaching and Research Section of Diagnostics, Shandong Medical College, Linyi, Shandong, China
Article metrics0 viewsView details
PDFCiteMetrics

Abstract

Diabetic kidney disease (DKD) represents one of the most significant microvascular complications associated with type 2 diabetes mellitus (T2DM). Moderately increased albuminuria, defined as a urinary albumin-to-creatinine ratio (UACR) of 30–300 mg/g, is widely recognized as an early indicator of renal injury in patients with T2DM. Previous studies have confirmed the role of glucolipotoxicity in the progression of DKD. Recently, a series of composite indices originating from blood glucose and lipid levels have been proposed, covering triglyceride-glucose index (TyG), triglyceride-glucose index to high-density lipoprotein cholesterol ratio (TyG/HDL-c), triglyceride glucose (TyG)-atherogenic index of plasma (AIP) index (TyG-AIP), cholesterol-high-density lipoprotein-glucose index (CHG) and triglyceride-glycated hemoglobin index (TyH-i). This study aimed to investigate the association between the aforementioned composite glucose-lipid indices and moderately increased albuminuria in patients with T2DM. The cross-sectional analysis included 1743 patients with T2DM. The clinical and laboratory data were obtained and related indices were calculated. Patients were categorized into the moderately increased albuminuria group (UACR 30–300 mg/g) and the normal to mildly increased albuminuria group (UACR < 30 mg/g). This study encompassed 342 participants with moderately increased albuminuria. The composite glucose-lipid indices levels were elevated in the moderately increased albuminuria group (all P < 0.05). Univariate analysis revealed that these indices all exhibited a positive relationship with moderately increased albuminuria (all P < 0.05). Logistic regression analysis showed that TyH-i was the only composite glucose-lipid index retained in the logistic model (OR: 1.789; 95% CI: 1.276 - 2.510). In separate adjusted models, TyH-i remained independently associated with moderately increased albuminuria (OR: 2.279; 95% CI: 1.671 - 3.110). Among numerous composite glucose-lipid indices, TyH-i remained independently associated with moderately increased albuminuria after adjustment. However, its discriminatory performance was limited and did not differ significantly from that of the comparator indices.

Keywords

Introduction

The worldwide concern about type 2 diabetes mellitus (T2DM) is escalating, posing deepening challenges for health systems and communities across the globe (1). Inappropriately controlled T2DM has been shown to result in manifold complications, such as vascular, neurological, and multi-organ issues (2). Among the microvascular complications of T2DM, diabetic kidney disease (DKD) is highly prevalent (3). On a global scale, end-stage renal disease (ESRD) is most frequently attributed to DKD (4). Research indicated that roughly 30-40% of diabetic patients are inclined to develop DKD (5). An albumin-to-creatinine ratio (UACR) level exceeding 30 mg/g but remaining below 300 mg/g defines moderately increased albuminuria (historically termed microalbuminuria), which is indicative of early-stage DKD (6, 7). Moreover, research has verified that moderately increased albuminuria is a predisposing factor for developing both ESRD and cardiovascular diseases (CVD) (8). Thus, early identification and prevention of moderately increased albuminuria are critical, which can arrest the initiation and worsening of DKD.

The progression of DKD is driven by glucolipotoxicity. Sustained hyperglycemia disrupts metabolic mediators such as reactive oxygen species and advanced glycation end products from an overall perspective, while it induces aberrant renal hemodynamics and glomerular hyperfiltration at the renal level (9). In addition, renal lipotoxicity, a key driver of DKD that arises from disordered lipid uptake, oxidation, efflux, and synthesis within renal cells, involves accumulated toxic lipid intermediates that disrupt mitochondria, exacerbate oxidative and endoplasmic reticulum stress, trigger inflammation and fibrosis, and eventually lead to renal function decline (10). Therefore, simple and readily available composite indices integrating blood glucose and lipid parameters may help capture the combined effects of glucolipotoxicity, whereas a single traditional glucose or lipid parameter may not fully reflect these synergistic effects.

Recent advances have enabled the development of novel composite indices that combine blood glucose and lipid parameters, thereby providing a more comprehensive assessment of glucolipotoxicity. The triglyceride-glucose index (TyG) is a classical sign of glucolipotoxicity (11), constructed by fasting plasma glucose (FPG) and triglycerides (TG) and is associated with DKD (12). Subsequently, the researchers put forward a series of derivative indices of the TyG index. The triglyceride-glucose index to high-density lipoprotein cholesterol ratio (TyG/HDL-c) has been shown to be significantly associated with diabetes in metabolic-associated fatty liver disease (13). The triglyceride glucose (TyG)-atherogenic index of plasma (AIP) index (TyG-AIP) correlates with elevated stroke risk, particularly among individuals with dysglycemia (14). Furthermore, the triglyceride-glycated hemoglobin index (TyH-i), a novel composite index analogous to the TyG index that integrates glycated hemoglobin (HbA1c) and TG, has demonstrated clinical utility in cohorts of patients with T2DM and DKD (15, 16). Moreover, the cholesterol-high-density lipoprotein-glucose index (CHG) also serves as a novel biomarker that is akin to the TyG index, demonstrates a marked correlation with DKD (17).

Currently, most studies have focused on single or a limited set of composite glucose-lipid indices, and forceful systematic evidence linking these indices to T2DM and its complications remains scarce. In addition, considering that the pathophysiological mechanisms of DKD differ between the early stage and the middle to late stages (18). Therefore, we employed a suite of complementary statistical approaches to explore the association between the composite glucose-lipid indices and moderately increased albuminuria in patients with T2DM.

Methods

Research methods and sample characteristics

For this study, participants aged 18 to 90 years diagnosed with T2DM were enrolled at Linyi People’s Hospital’s Department of Endocrinology during the time frame of March 2020 to February 2023. All patients were diagnosed in accordance with the 1999 World Health Organization (WHO) diagnostic criteria. The criteria listed below were used for exclusion: (1) patients suffering from different forms of diabetes; (2) individuals with UACR exceeding 300 mg/g; (3) patients diagnosed with severe liver (beyond two-fold of the normal upper threshold) or kidney involvement (eGFR < 60 mL/min/1.73m2); (4) patients have urinary tract infections; (5) medical records that were not complete, including patients with missing UACR data and missing data for calculating composite glucose-lipid indices. Ultimately, our study included 1743 patients with T2DM.

General conditions and physical assessments

The participants’ baseline characteristics, including sex, smoking and drinking habits, age, duration of diabetes, height and weight were documented. Information regarding the use of sodium-glucose cotransporter 2 (SGLT2) inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), statins, fibrates, metformin, and insulin was collected. Following a sufficient period of rest, a standard electronic device was utilized to gather the systolic blood pressure (SBP) and diastolic blood pressure (DBP). Bioelectrical impedance analysis (HDS-2000; Omron, Kyoto, Japan) served for the assessment of the visceral fat area (VFA) and subcutaneous fat area (SFA).

Laboratory measurements

Early morning venous samples were obtained and submitted to biochemical profiling using an automated system (Cobas c 702, Roche, Germany) after a minimum of 8 hours of sustained fasting. The total cholesterol (TC), TG, HDL-c, low-density lipoprotein cholesterol (LDL-c), aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT), uric acid (UA), serum creatinine (Scr), hemoglobin (Hb), FPG and HbA1c (high-performance liquid chromatography) were tested. Fasting serum insulin (FINS) levels were determined via a fully automated machine (Aptio Automation, Siemens, USA). UACR was obtained through an autoanalyzer (Beckman Coulter AU5821).

Diagnostic criteria and assessment of diabetic retinopathy

Diagnose DR in accordance with established clinical criteria (19).

Parameter calculations

Homeostasis model assessment index (HOMA-IR) = FPG (mmol/L) * FINS (µU/mL)/22.5 (20);

TyG index = Ln [TG (mg/dL) * FPG (mg/dL)/2] (21);

TyG/HDL-c = TyG index/HDL-c (mmol/L) (22);

Atherogenic index of plasma (AIP) = log10 [TG (mg/dl)/HDL-c (mg/dl)] (23);

TyG-AIP =TyG * [AIP - min (AIP)]/[max (AIP) - min (AIP)] (23);

CHG index = Ln [TC (mg/dL) * FPG (mg/dL)/2 * HDL-c (mg/dL)] (24);

"TyH-i = Ln [HbA1c (%) * TG (mg/dL)/2] (15);

Body mass index (BMI) = weight (kg)/height (m)2 (25);

eGFR = 175 * Scr (mg/dL) - 1.234 * age - 0.179 × (0.79, if female) (26).

Statistical methods

SPSS 25.0 (SPSS Inc., Chicago, USA), R software (version 4.6.1, R Foundation for Statistical Computing, Vienna, Austria), and SPSSAU (https://spssau.com) were utilized to perform the data analysis. Candidate variables were first selected based on clinical knowledge and prior literature. Non-normally distributed variables were presented as median and interquartile range and compared using the Mann-Whitney U test, normally distributed variables were presented as mean ± SD and compared using independent samples t-tests. Categorical variables were presented as percentages and compared using the chi-square test. Univariate analysis was then performed on these candidate variables, and variables with P < 0.05 were entered into binary stepwise forward logistic regression for further selection. The Hosmer-Lemeshow test was employed to assess whether the model exhibits a lack of fit.

Correlation coefficients and variance inflation factors (VIF) were used to examine multicollinearity. Separate multivariable models, each including one indicator from the multicollinearity set on top of common confounders, were constructed for sensitivity analysis. The reference model for Net Reclassification Improvement (NRI) and Integrated Discrimination Improvement (IDI) included only common confounders. Continuous NRI was employed. Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were computed for each model for descriptive purposes. Restricted cubic spline (RCS) analysis was used as an exploratory method to examine the potential nonlinear association between TyH-i and moderately increased albuminuria. ROC curves and DeLong tests were used to assess and compare discriminatory performance. LASSO regression was used as an exploratory penalized variable-selection approach. Two-tailed P < 0.05 was considered statistically significant.

Results

Baseline clinical and biochemical characteristics

Table 1 displays the clinical and biochemical characteristics of participants. Compared with the normal to mildly increased albuminuria group, the moderately increased albuminuria group had elevated levels of age, duration of diabetes, BMI, VFA, SBP, DBP, TC, TG, FPG, FINS, HbA1c, HOMA-IR, GGT, UA, the percentage of ACEIs or ARBs, the percentage of insulin, UACR, Scr, TyG index, TyG/HDL-C, TyG-AIP, CHG index, TyH-i, and the percentage of DR (all P < 0.05). Furthermore, HDL-c, eGFR, Hb, and the percentage of metformin were significantly lower in the moderately increased albuminuria group (all P < 0.05). However, sex, the percentage of smoking and drinking, SFA, LDL-c, AST, ALT, the percentage of SGLT2 inhibitors, the percentage of GLP-1 receptor agonists, the percentage of statins, and the percentage of fibrates did not differ significantly between the two groups (all P > 0.05).

Table 1. Clinical and biochemical characteristics by presence of moderately increased albuminuria.

VariablesAIINormal to mildly increased albuminuria groupModerately increased albuminuria groupP
Numbers17431401342
Sex (Females, n, %)896 (51.4%)720 (51.4%)176 (51.5%)0.981
Smoking (%)337 (19.3%)267 (19.1%)70 (20.5%)0.554
Drinking (%)323 (18.5%)258 (18.4%)65 (19.0%)0.801
Age (years)57.3 ± 12.456.9 ± 12.258.8 ± 13.40.009
Duration ofdiabetes (years)6.0 (1.5 - 10.0)6.0 (1.0 - 10.0)8.0 (3.0 - 14.0)< 0.001
BMI (kg/m2)25.66 ± 3.5425.56 ± 3.4026.09 ± 4.030.013
VFA (cm2)91.00 (68.00 - 120.00)90.00 (67.00 - 119.00)98.00 (71.00 - 127.00)0.005
SFA (cm2)184.00 (142.50 - 229.00)182.00 (142.00 - 227.75)191.00 (143.50 - 238.50)0.085
SBP (mmHg)129.7 ± 18.3128.5 ± 17.3134.5 ± 21.1< 0.001
DBP (mmHg)80.8 ± 11.480.4 ± 10.782.3 ± 13.90.009
TC (mmol/L)4.75 ± 1.164.72 ± 1.144.87 ± 1.240.036
TG (mmol/L)1.37 (0.97 - 2.03)1.35 (0.95 - 1.97)1.50 (1.02 - 2.22)0.004
HDL-c (mmol/L)1.16 ± 0.321.17 ± 0.321.13 ± 0.310.038
LDL-c (mmol/L)3.03 ± 0.983.02 ± 0.963.09 ± 1.060.229
FPG (mmol/L)8.86 ± 3.548.71 ± 3.459.45 ± 3.840.001
FINS (µU/mL)15.96 (9.43 - 22.11)15.43 (8.78 - 21.48)18.14 (11.86 - 24.48)< 0.001
HbA1c (%)9.18 ± 2.029.08 ± 2.009.61 ± 2.07< 0.001
HOMA-IR5.83 (3.07 - 9.35)5.42 (2.90 - 8.96)7.30 (4.47 - 11.20)< 0.001
AST (U/L)17.20 (14.00 - 21.70)17.30 (14.00 - 21.70)17.10 (14.00 - 21.93)0.857
ALT (U/L)18.20 (13.60 - 25.70)18.20 (13.60 - 25.80)18.10 (13.15 - 25.25)0.556
GGT (U/L)21.00 (16.00 - 31.08)21.00 (15.00 - 30.88)23.70 (16.38 - 35.25)0.005
UA (μmol/L)280.61 ± 86.12276.11 ± 82.88299.06 ± 96.21< 0.001
UACR (mg/g)8.90 (5.10 - 22.80)7.10 (4.60 - 12.15)67.60 (43.20 - 121.28)< 0.001
Scr (µmol/L)61.74 ± 13.7560.83 ± 12.9465.46 ± 16.14< 0.001
eGFR(mL/min/1.73 m2)124.56 ± 29.98126.40 ± 29.36116.99 ± 31.32< 0.001
Hb (g/L)143.07 ± 17.63143.88 ± 16.66139.79 ± 20.83< 0.001
SGLT2 inhibitors, n (%)92 (5.3%)71 (5.1%)21 (6.1%)0.421
GLP-1 receptor agonists, n (%)381 (21.9%)304 (21.7%)77 (22.5%)0.729
ACEIs or ARBs, n (%)302 (17.3%)214 (15.3%)88 (25.7%)< 0.001
Statins, n (%)1253 (71.9%)993 (70.9%)260 (76.0%)0.050
Fibrates, n (%)65 (3.7%)48 (3.4%)17 (5.0%)0.174
Metformin, n (%)1140 (65.4%)944 (67.4%)196 (57.3%)0.001
Insulin, n (%)1405 (80.6%)1104 (78.8%)301 (88.0%)< 0.001
TyG index9.15 ± 0.749.12 ± 0.739.29 ± 0.79< 0.001
TyG/HDL-c8.45 ± 2.518.35 ± 2.468.85 ± 2.670.001
TyG-AIP4.59 ± 1.744.52 ± 1.714.89 ± 1.83< 0.001
CHG index5.72 ± 0.535.69 ± 0.515.83 ± 0.58< 0.001
TyH-i6.34 ± 0.626.31 ± 0.616.47 ± 0.65< 0.001
DR626 (35.9%)462 (33.0%)164 (48.0%)< 0.001

The distribution of the variables was normally approximated by a mean ± standard deviation, and comparisons between the two groups were conducted using an independent samples t-test. Variables with abnormally distributed data were expressed as the median (25th percentile to 75th percentile), and comparisons between the two groups were conducted using the Mann–Whitney U-test. Categorical variables were expressed as percentages and subjected to a chi-square test for comparison. Statistical differences were defined by a P value less than 0.05 (two-tailed).

Normal to mildly increased albuminuria group, UACR< 30 mg/g; Moderately increased albuminuria group, UACR 30-300 mg/g; BMI, body mass index; VFA, visceral fat area; SFA, subcutaneous fat area; SBP, systolic blood pressure; DBP, diastolic blood pressure; TC, total cholesterol; TG, triglyceride; HDL-c, high-density lipoprotein cholesterol; LDL-c, low-density lipoprotein cholesterol; FPG, fasting plasma glucose; FINS, fasting serum insulin; HbA1c, glycosylated hemoglobin; HOMA-IR, homeostatic model assessment of insulin resistance;

AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; UA, uric acid; UACR, urinary albumin to creatinine ratio; Scr, serum creatinine; eGFR, estimated glomerular filtration rate; Hb, hemoglobin; S GLT2 inhibitors, sodium glucose co-transporter 2 inhibitors; GLP-1 receptor agonists, glucagon-like peptide-1 receptor agonists; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; TyG index, triglyceride–glucose index; TyG/HDL-c, triglyceride-glucose index to high-density lipoprotein cholesterol ratio; TyG-AIP, triglyceride glucose (TyG)-atherogenic index of plasma (AIP) index; CHG index, cholesterol-high-density lipoprotein-glucose index; TyH-i, triglyceride-glycated hemoglobin index; DR, diabetic retinopathy.

Univariate analysis

Table 2 presents the univariate analysis of factors associated with moderately increased albuminuria. Age, duration of diabetes, VFA, SBP, TC, TG, FPG, FINS, HbA1c, HOMA-IR, GGT, UA, ACEIs or ARBs, insulin, TyG index, TyG/HDL-c, TyG-AIP, CHG index, TyH-i, and DR were positively correlated with moderately increased albuminuria (all P < 0.05). HDL-c, eGFR, Hb, and metformin were negatively correlated with moderately increased albuminuria (all P < 0.05). Notably, no significant correlations with moderately increased albuminuria were shown for sex, smoking, drinking, BMI, SFA, DBP, LDL-c, AST, ALT, SGLT2 inhibitors, GLP-1 receptor agonists, statins, and fibrates (all P > 0.05). As shown in Supplementary Appendix Figure 1, a range of -0.5 to 1.0 was observed for the Spearman correlations among all variables. Correlation values spanned from a moderate negative association (-0.5) to a perfect positive association (1.0).

Table 2. Univariate correlation analysis for moderately increased albuminuria.

VariablesCorrelation coefficientP
Sex-0.0010.981
Smoking0.0140.554
Drinking0.0060.801
Age0.0580.016
Duration of diabetes0.113<0.001
BMI0.0400.092
VFA0.0690.005
SFA0.0430.085
SBP0.115<0.001
DBP0.0470.050
TC0.0490.041
TG0.0700.004
HDL-c-0.0580.015
LDL-c0.0250.298
FPG0.086<0.001
FINS0.118<0.001
HbA1c0.106<0.001
HOMA-IR0.156<0.001
AST0.0040.857
ALT-0.0140.556
GGT0.0670.005
UA0.102<0.001
eGFR-0.130<0.001
Hb-0.0820.001
SGLT2 inhibitors0.0190.421
GLP-1 receptor agonists0.0080.729
ACEIs or ARBs0.110<0.001
Statins0.0470.050
Fibrates0.0330.174
Metformin-0.0830.001
Insulin0.095<0.001
TyG index0.093<0.001
TyG/HDL-c0.0760.002
TyG-AIP0.0800.001
CHG index0.100<0.001
TyH-i0.099<0.001
DR0.124<0.001

To identify the factors associated with moderately increased albuminuria, a univariate correlation analysis was conducted.

Moderately increased albuminuria, UACR 30-300 mg/g; BMI, body mass index; VFA, visceral fat area; SFA, subcutaneous fat area; SBP, systolic blood pressure; DBP, diastolic blood pressure; TC, total cholesterol; TG, triglyceride; HDL-c, high-density lipoprotein cholesterol; LDL-c, low-density lipoprotein cholesterol; FPG, fasting plasma glucose; FINS, fasting serum insulin; HbA1c, glycosylated hemoglobin; HOMA-IR, homeostatic model assessment of insulin resistance; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; UA, uric acid; eGFR, estimated glomerular filtration rate; Hb, hemoglobin; SGLT2 inhibitors, sodium glucose co-transporter 2 inhibitors; GLP-1 receptor agonists, glucagon-like peptide-1 receptor agonists; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; TyG index, triglyceride–glucose index; TyG/HDL-c, triglyceride-glucose index to high-density lipoprotein cholesterol ratio; TyG-AIP, triglyceride glucose (TyG)-atherogenic index of plasma (AIP) index; CHG index, cholesterol-high-density lipoprotein-glucose index; TyH-i, triglyceride-glycated hemoglobin index; DR, diabetic retinopathy.

Multivariate analysis

Multivariate logistic regression analysis

In the present study, moderately increased albuminuria was recognized as the dependent variable. Variables with statistical significance in the univariate correlation analysis were included in the regression model. Our study adjusted for age, duration of diabetes, VFA, SBP, TC, TG, HDL-c, FPG, FINS, HbA1c, HOMA-IR, GGT, UA, eGFR, Hb, ACEIs or ARBs, insulin, metformin, TyG, TyG/HDL-c, TyG-AIP, CHG, TyH-i and DR. Table 3 presents the outcomes of binary stepwise forward logistic regression analysis, which revealed the independent correlation of moderately increased albuminuria. Results showed that TyH-i (OR: 1.789, 95% CI: 1.276 - 2.510), DR (OR: 1.872, 95% CI: 1.289 - 2.721), ACEIs or ARBs (OR: 1.791, 95% CI: 1.149 - 2.792), HbA1c (OR: 1.237, 95% CI: 1.112 - 1.376), duration of diabetes (OR: 1.042, 95% CI: 1.015 - 1.069), SBP (OR: 1.015, 95% CI: 1.005 - 1.026), eGFR (OR: 0.989, 95% CI: 0.983 - 0.996), Hb (OR: 0.987, 95% CI: 0.976 - 0.998) and metformin (OR: 0.590, 95% CI: 0.403 - 0.863) were associated with moderately increased albuminuria. Figure 1 provides a forest plot depicting the independent association of moderately increased albuminuria identified through binary stepwise forward logistic regression analysis. The Hosmer-Lemeshow test result showed that P equals 0.411.

Table 3. The independent variables for moderately increased albuminuria.

VariablesPOR95.0% CI for OR
DR0.0011.8721.289 - 2.721
ACEIs or ARBs0.0101.7911.149 - 2.792
TyH-i0.0011.7891.276 - 2.510
HbA1c<0.0011.2371.112 - 1.376
Duration of diabetes0.0021.0421.015 - 1.069
SBP0.0051.0151.005 - 1.026
eGFR0.0020.9890.983 - 0.996
Hb0.0180.9870.976 - 0.998
Metformin0.0070.5900.403 - 0.863

The independent variables for moderately increased albuminuria were evaluated using a logistic regression analysis.

DR, diabetic retinopathy; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; TyH-i, triglyceride-glycated hemoglobin index; HbA1c, glycosylated hemoglobin; SBP, systolic blood pressure; eGFR, estimated glomerular filtration rate; Hb, hemoglobin.

Forest plot of factors independently associated with moderately increased albuminuria. The independent variables for moderately increased albuminuria were evaluated using a logistic regression analysis. The blue solid dots represent the OR of each variable for moderately increased albuminuria and the horizontal black lines correspond to the 95% CI of OR values. The vertical line at X = 1.0 is the line of no effect. DR, diabetic retinopathy; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; TyH-i, triglyceride-glycated hemoglobin index; HbA1c, glycosylated hemoglobin; SBP, systolic blood pressure; eGFR, estimated glomerular filtration rate; Hb, hemoglobin.
Figure 1. Forest plot of factors independently associated with moderately increased albuminuria. The independent variables for moderately increased albuminuria were evaluated using a logistic regression analysis. The blue solid dots represent the OR of each variable for moderately increased albuminuria and the horizontal black lines correspond to the 95% CI of OR values. The vertical line at X = 1.0 is the line of no effect. DR, diabetic retinopathy; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; TyH-i, triglyceride-glycated hemoglobin index; HbA1c, glycosylated hemoglobin; SBP, systolic blood pressure; eGFR, estimated glomerular filtration rate; Hb, hemoglobin.

Model performance metrics and RCS analysis

Due to the similarities among the components of the indices, our study examined multicollinearity by assessing the correlation coefficients and VIF among these variables. The results revealed that there existed a covariance problem between FPG, HbA1c, TG, TyG index, TyG/HDL-c, TyG-AIP, CHG index, and TyH-i (as illustrated in Supplementary Appendix Tables 1, 2). Therefore, we constructed separate multivariable models, each including only one index alongside the relevant clinical covariates. TyH-i remained independently associated with moderately increased albuminuria (OR 2.279, 95% CI 1.671 - 3.110; Table 4). NRI and IDI were calculated as exploratory reclassification metrics (as shown in Table 4), and the corresponding P values and 95% CI were provided in Supplementary Appendix Tables 5, 6. AIC and BIC values for each model were also summarized in Table 4. The AUC values of the composite glucose-lipid indices for identifying moderately increased albuminuria were all below 0.60 (Supplementary Appendix Table 3; Supplementary Appendix Figure 3), and DeLong tests showed no statistically significant differences between indices (all P > 0.05; Supplementary Appendix Table 4). In the RCS analysis (as shown in Figure 2), 4 knots were used, and the reference value was set at the median of TyH-i (6.310). The overall model fit test was significant (P < 0.001), whereas the nonlinearity test was not significant (P = 0.109; Table 5).

RCS analysis of TyH-i with moderately increased albuminuria.The likelihood ratio test was used to assess the overall nonlinear association between TyH-i and moderately increased albuminuria. Moderately increased albuminuria, UACR 30-300 mg/g; TyH-i, triglyceride-glycated hemoglobin index.
Figure 2. RCS analysis of TyH-i with moderately increased albuminuria.The likelihood ratio test was used to assess the overall nonlinear association between TyH-i and moderately increased albuminuria. Moderately increased albuminuria, UACR 30-300 mg/g; TyH-i, triglyceride-glycated hemoglobin index.

Table 4. Sensitivity analysis in different variables for moderately increased albuminuria.

ModelsVariablesOR95% CI for ORAICBICNRIIDI
Model 1HbA1c1.3461.217-1.488773.093826.23232.503%*0.035*
Model 2TyH-i2.2791.671-3.110784.559828.03725.762% *0.015*
Model 3CHG index2.5781.784-3.725785.764829.24222.429%*0.017*
Model 4TyG index1.9191.481-2.487788.567827.21411.580%0.009
Model 5FPG1.0891.036-1.144790.210843.34924.516%*0.017*
Model 6TyG-AIP1.2391.118-1.372793.173836.65113.223%*0.010*
Model 7TyG/HDL-c1.1241.050-1.204797.986841.4648.095%0.003
Model 8TG1.3021.126-1.506798.284836.931-5.010%-0.003

Model: Common confounding variables: age, duration of diabetes, VFA, SBP, TC, HDL-c, FINS, GGT, UA, eGFR, HOMA-IR, Hb, ACEIs or ARBs, Metformin, Insulin, DR.

Model 1: Common confounding variables + HbA1c.

Model 2: Common confounding variables + TyH-i.

Model 3: Common confounding variables + CHG index.

Model 4: Common confounding variables + TyG index.

Model 5: Common confounding variables + FPG.

Model 6: Common confounding variables + TyG-AIP.

Model 7: Common confounding variables + TyG/HDL-c.

Model 8: Common confounding variables + TG.

Models 1–8 incorporated only one indicator from the multicollinearity set on top of common confounding variables. The reference model for NRI and IDI includes only common confounding variables. Continuous NRI was employed. * means that the NRI/IDI metrics exhibited statistically significant differences when compared with the reference model.

TG, triglyceride; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin; TyG index, triglyceride–glucose index; TyG/HDL-c, triglyceride-glucose index to high-density lipoprotein cholesterol ratio; TyG-AIP, triglyceride glucose (TyG)-atherogenic index of plasma (AIP) index; CHG index, cholesterol-high-density lipoprotein-glucose index; TyH-i, triglyceride-glycated hemoglobin index; AIC, Akaike information criterion; BIC, Bayesian information criterion; NRI, Net reclassification improvement; IDI, Integrated Discrimination Improvement. * Indicates P < 0.05.

Table 5. Overall nonlinear relationship test of the model for moderately increased albuminuria.

Model-2 Log-likelihoodχ²dfP
Intercept-only model1545.125
Final model1540.7004.42620.109

The likelihood ratio test was used to assess the overall nonlinear association between TyH-i and the moderately increased albuminuria. P-value ≥ 0.05 indicates no statistically significant evidence for nonlinearity, nonlinearity cannot be entirely ruled out.

Moderately increased albuminuria: UACR 30-300 mg/g; TyH-i, triglyceride-glycated hemoglobin index; df, degrees of freedom.

Exploratory LASSO regression

In LASSO regression analysis, ten-fold cross-validation was adopted to determine the optimal penalty parameter λ (lambda.1se = 0.02452612). Eleven variables with non-zero coefficients were selected: duration of diabetes, SBP, HbA1c, UA, eGFR, Hb, ACEIs or ARBs, metformin, CHG index, TyH-i, and DR. Supplementary Appendix Figure 2A shows the LASSO coefficient-profile plot of all patients, regression coefficients of most variables gradually shrank to zero with increasing penalty parameter lambda. Supplementary Appendix Figure 2B shows the 10-fold cross-validation curve used to select the penalty parameter λ.

Discussion

Using a cross-sectional design, we explored the association between numerous composite glucose-lipid indices and moderately increased albuminuria. Univariate correlation analysis revealed a significant connection between these indices and moderately increased albuminuria in patients with T2DM. TyH-i was independently associated with moderately increased albuminuria following adjustment. Nevertheless, its discriminatory performance was limited, with no significant difference relative to other comparator indices.

A previous study confirmed that, after adjusting for confounding factors such as age, gender, and BMI, the results of logistic regression indicated that the TyG index remained significantly associated with moderately increased albuminuria (27). In this study, we observed that the TyG index was significantly elevated in the moderately increased albuminuria group but did not enter the final logistic regression model. Moreover, a study reported a statistically outstanding association between the CHG index and DKD (17). However, we did not find a significant association between the CHG index and moderately increased albuminuria. Additionally, no published studies have analyzed the association between the TyG/HDL-c and TyG-AIP and moderately increased albuminuria. Similarly, we also found that these two indices were significantly elevated in the moderately increased albuminuria group, yet they did not enter the regression model in the end.

Notably, logistic regression analysis identified only TyH-i entering the logistic model among composite glucose-lipid indices (OR = 1.789, 95% CI = 1.276 - 2.510). The Hosmer-Lemeshow test yielded P = 0.411. This result indicated that there was insufficient evidence of lack of fit. In the separate model adjustment, TyH-i remained independently associated with moderately increased albuminuria (OR 2.279; 95% CI 1.671 - 3.110). Recently, Ma et al. analyzed the correlation between TyH-i and DKD by enrolling study participants with varying degrees of albuminuria, their findings indicated that TyH-i exhibited a stronger correlation with DKD than the TyG index (16). This discovery further supports our conclusion. And in our study, the OR value of TyH-i exhibited a discrepancy between the full variables model and the separate model. This discrepancy may be attributed to the multicollinearity among indices. In the full regression model incorporating all candidate variables, TyH-i was introduced into the model simultaneously with HbA1c, TG, FPG, and other composite glucose-lipid indices, leading to variance overlap among variables, which may result in the attenuation of the effect estimate of TyH-i. By contrast, each separate model incorporated only one index from the multicollinearity set on top of common confounders, which avoided the effect competition caused by multicollinearity, and the obtained effect estimates could better reflect their independent associations. The direction of association and statistical significance remained consistent across different models, indicating that the observed association possessed a certain degree of robustness. In addition, the RCS analysis did not identify a statistically significant nonlinear association (P = 0.109).

Nonetheless, the mechanism linking TyH-i and moderately increased albuminuria remains unclear. It may encompass the following aspects. HbA1c was a critical component of TyH-i, and HbA1c could reflect the historical blood glucose control level and glycemic variability. Although FPG also served as a reliable indicator of blood glucose control (28), evidence from studies indicated a stronger association between HbA1c and DKD than FPG (29). Multiple studies demonstrated that even after subsequent improvement in glycemic levels, prior poor glycemic control could induce a sustained effect on the risk of diabetic complications known as the metabolic memory effect (30, 31), and this phenomenon, together with glycemic variability (32), triggered persistent oxidative stress, chronic inflammation, and epigenetic reprogramming to promote progressive kidney injury. HbA1c was a reliable indicator of metabolic memory and glycemic fluctuations (33). In addition, elevated TG levels represented a significant risk factor for the development of DKD (34). Emerging evidence indicated that dyslipidemia in early-stage DKD was characterized by elevated TG levels (35). This may be mediated by TG accumulation, which activated monocytes to produce inflammatory mediators, transforming growth factor beta and reactive oxygen species, leading to renal cell injury (36). However, based on the current data, we were unable to distinguish whether the relationship between TyH-i and moderately increased albuminuria was derived from the mathematical structure of the index or from a genuine biological effect.

Additionally, logistic regression analysis identified several other factors associated with moderately increased albuminuria. Undoubtedly, prolonged diabetes duration increased the risk of moderately increased albuminuria (37). As the first clinical expression of DKD, moderately increased albuminuria is also known to anticipate the onset of DR (38). HbA1c and SBP were directly positively correlated with moderately increased albuminuria, which aligns with the previous study, this finding points to a potential involvement of metabolic disturbances in the progression of moderately increased albuminuria (7). A potential explanation for the observed negative correlation between Hb and moderately increased albuminuria may be attributable to impaired kidney function affecting erythropoietin production (39). According to the guidelines, patients with high risk factors for renal injury are more likely to be prescribed ACEIs or ARBs for treatment (40). The connection between metformin and moderately increased albuminuria may be attributed to metformin effectively modulating blood glucose concentrations without promoting weight gain, thereby establishing its status as a guideline endorsed first line pharmacologic intervention for T2DM (41).

However, the AUC indicated that the predictive performance of the composite glucose-lipid indices for moderately increased albuminuria was limited. The DeLong test further revealed no statistically significant difference in discriminatory accuracy among these indices. These results demonstrated that a statistically significant independent association may not guarantee satisfactory discriminatory capacity for disease prediction (42). In addition, the LASSO regression was used as an exploratory penalized variable-selection approach. Although TyH-i remained independently associated with moderately increased albuminuria in multivariable models, this observational finding required further external validation.

Further, the retrospective, single center research design may introduce selection bias, and residual confounding factors such as lifestyle and physical exercise may have an impact on the results. At the same time, constrained by the inherent structural limitations of composite indices, these indices were retrospectively derived from metabolic parameters based on routine clinical test results, and no clinical consensus has yet been established regarding their calculation methodologies. Another, based on cross-sectional studies, reverse causality cannot be fully excluded, as incipient renal injury indicated by moderately increased albuminuria may disturb composite glucose-lipid indices homeostasis and further elevate composite indices. In this study, moderately increased albuminuria was defined as a single UACR of 30–300 mg/g, and persistent albuminuria could not be confirmed. Intra-individual variability in UACR may result in outcome misclassification. Such misclassification could distort the observed associations, reduce the accuracy of risk stratification, and introduce bias in effect estimation, thereby compromising the interpretability and reliability of the results. Therefore, although these composite glucose-lipid indices showed an independent correlation with moderately increased albuminuria, they were not suitable for standalone clinical diagnosis.

Limitations

Several limitations should be acknowledged. Primarily, defining moderately increased albuminuria based on a single UACR measurement was a major limitation of this study. A single measurement may misclassify transient albuminuria and bias risk stratification. Future studies should adopt repeated UACR measurements to mitigate misclassification. Second, owing to the cross-sectional design, causal relationships between the composite glucose-lipid indices and moderately increased albuminuria could not be established, and cut-off values for predictive purposes could not be derived. Third, variable selection in this study was exploratory without an a priori statistical analysis plan. Relying solely on univariate correlation analysis for variable selection may compromise the precision of effect estimation. Fourth, findings from some analyses require careful interpretation. The NRI and IDI were exploratory and not based on formal model comparisons. AIC and BIC reflected goodness-of-fit rather than discriminative ability. LASSO was used solely for variable screening, not model validation. RCS analysis revealed no significant nonlinearity and was reported descriptively. Calibration was assessed exclusively via the Hosmer-Lemeshow test, a non-significant P value from this test did not confirm adequate calibration. Future studies should prospectively define analytical protocols, enhance calibration assessment, incorporate bootstrapped internal validation, and standardize exploratory methods to robustly evaluate the TyH-i’s predictive performance. In addition, analyses were restricted to participants with normal-to-mild and moderately increased albuminuria. Future multicenter large-scale studies are needed to explore the generalizability of these findings to the wider DKD population and conduct subgroup analyses, such as gender, duration of diabetes, age, and obesity status.

Conclusion

In conclusion, among various composite glucose-lipid indices, TyH-i remained independently associated with moderately increased albuminuria after adjustment. However, its discriminatory performance was limited and did not differ significantly from that of the comparator indices. Future multicenter, longitudinal or interventional studies are needed to rigorously validate the association between TyH-i and moderately increased albuminuria.

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.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fendo.2026.1883785/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Bo Cao, The Second Affiliated Hospital of Xi’an Jiaotong University, China

Reviewed by: De Jin, Hangzhou Hospital of Traditional Chinese Medicine Affiliated to Zhejiang Chinese Medical University, China

Tiago Sampaio, Federal University of Ceara, Brazil

Statements

Conflict of interest

The author(s) 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.

Funding

The author(s) declared that financial support was received for this work and/or its publication. Support for this work came from the Key Research and Development Program of Linyi City (2024YX004).

Data availability

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics approval

The studies involving humans were approved by The Human Ethics Committee of Linyi People’s Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

YL: Conceptualization, Data curation, Formal analysis, Writing – original draft. ZP: Investigation, Writing – original draft. XZ: Data curation, Writing – review & editing. GG: Visualization, Writing – review & editing. DF: Software, Writing – review & editing. BJ: Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – review & editing.

References

  1. 1 Wilk A, Wilkosz N, Rugiel M, Saluga M, Kawon K, Stepanenko T. Next-generation diabetes diagnostics: spectrochemical staging of red blood cells using vibrational spectroscopy. J Adv Res. (2026) 88: 415–436. doi: doi:10.1016/j.jare.2026.01.052
  2. 2 C MN, Edward S, Grace A, N BS. Assessment of prevalence and risk factors of diabetes distress among patients with type 2 diabetes mellitus in an urban area of Chengalpattu District, India: A cross-sectional study. Cureus. (2025) 17: e98974. doi: doi:10.7759/cureus.98974
  3. 3 Atak Tel BM, Khalid A, Aktas G. Serum C-reactive protein and C-reactive protein derived markers in diabetic microvascular complications. World J Diabetes. (2025) 16: 113237. doi: doi:10.4239/wjd.v16.i12.113237
  4. 4 Yu C, Zhang H, Ni F, Jin L, Ying Z, Fu H. Causal relationships of lipid metabolism in diabetic nephropathy risk: A two-sample Mendelian randomization study. Nutr Metab Cardiovasc Dis. (2026) 36: 104528. doi: doi:10.1016/j.numecd.2025.104528
  5. 5 Kato M, Natarajan R. Epigenetics and epigenomics in diabetic kidney disease and metabolic memory. Nat Rev Nephrol. (2019) 15: 327–45. doi: doi:10.1038/s41581-019-0135-6
  6. 6 Qi L, Kang N, Chen X, Li Z, Deng C, Chen S. Predictive value of plasma atherogenic index for microalbuminuria in newly diagnosed patients with type 2 diabetes mellitus. Diabetes Metab Syndr Obes. (2022) 15: 1245–52. doi: doi:10.2147/dmso.s359811
  7. 7 Zhao H, Ma X, Shi S, Wang X, Sheng J, Ban B. Association between FT4/FT3 ratio and microalbuminuria in euthyroid patients with type 2 diabetes mellitus. Sci Rep. (2025) 15: 7324. doi: doi:10.1038/s41598-025-92046-0
  8. 8 Wester CW, Shepherd BE, Wudil UJ, Musa BM, Ingles DJ, Prigmore HL. Etiology of persistent microalbuminuria in Nigeria (P_MICRO study): protocol and study design. BMC Infect Dis. (2022) 22: 591. doi: doi:10.1186/s12879-022-07531-y
  9. 9 Guedes M, Pecoits-Filho R. Can we cure diabetic kidney disease? Present and future perspectives from a nephrologist's point of view. J Intern Med. (2022) 291: 165–80. doi: doi:10.1111/joim.13424
  10. 10 Zaitoon H, Feigin E, Abdul-Ghani M, Clarke GD, Tuttle KR, DeFronzo RA. Renal lipotoxicity: mechanisms and therapeutic perspectives. Diabetes Care. (2026). doi: doi:10.2337/dci26-0036
  11. 11 Bukke SPN, Pathange BBR, Nelluri KDD, Yadesa TM, Kamepalli S, Suvarna K. Association of triglyceride glucose index with clinical outcomes in ischemic stroke: a retrospective study. BMC Neurol. (2024) 24: 371. doi: doi:10.1186/s12883-024-03873-z
  12. 12 Yan H, Zhou Q, Wang Y, Tu Y, Zhao Y, Yu J. Associations between cardiometabolic indices and the risk of diabetic kidney disease in patients with type 2 diabetes. Cardiovasc Diabetol. (2024) 23: 142. doi: doi:10.1186/s12933-024-02228-9
  13. 13 Sun T, Liu J. Study on the correlation between triglyceride glucose index, triglyceride glucose index to high-density lipoprotein cholesterol ratio, and the risk of diabetes in nonalcoholic fatty liver disease. Front Endocrinol (Lausanne). (2025) 16: 1594548. doi: doi:10.3389/fendo.2025.1594548
  14. 14 Zou X, Zhou C, Zhou R, Zhu T, Zhao J, Qing W. The joint association of the combined triglyceride-glucose index and atherogenic index of plasma with hypertension on stroke risk across different glycemic status: a prospective cohort study. Cardiovasc Diabetol. (2026) 25: 141. doi: doi:10.1186/s12933-026-03080-9
  15. 15 Cao C, Zan Y, Han Y, Hu H, Long J, Wang Y. Triglyceride-glycated hemoglobin index as a superior predictor of type 2 diabetes risk in a large-scale retrospective cohort study. Sci Rep. (2025) 15: 20898. doi: doi:10.1038/s41598-025-05786-4
  16. 16 Ma J, Jin K, Wang Y, Wang X, Lou X, Liu X. Relationship between the TyGH index (based on triglycerides and HbA1c) and diabetic kidney disease in patients with type 2 diabetes: A real-world study. Diabetes Metab Syndr Obes. (2025) 18: 3955–65. doi: doi:10.2147/dmso.s542653
  17. 17 Cui W, Jiang S, Yu T, Yang Z, Zhou J, Liu L. The cholesterol, high-density lipoprotein, and glucose index and its derived indices predict diabetic kidney disease development and progression: evidence from two nationwide cohorts and a biopsy-proven clinical cohort. Cardiovasc Diabetol. (2026) 25: 163. doi: doi:10.1186/s12933-026-03167-3
  18. 18 Feng YL, Chen H, Chen DQ, Vaziri ND, Su W, Ma SX. Activated NF-κB/Nrf2 and Wnt/β-catenin pathways are associated with lipid metabolism in CKD patients with microalbuminuria and macroalbuminuria. Biochim Biophys Acta Mol Basis Dis. (2019) 1865: 2317–32. doi: doi:10.1016/j.bbadis.2019.05.010
  19. 19 Hao Z, Cui S, Zhu Y, Shao H, Huang X, Jiang X. Application of non-mydriatic fundus examination and artificial intelligence to promote the screening of diabetic retinopathy in the endocrine clinic: an observational study of T2DM patients in Tianjin, China. Ther Adv Chronic Dis. (2020) 11: 2040622320942415. doi: doi:10.1177/2040622320942415
  20. 20 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. (1985) 28: 412–9. doi: doi:10.1007/bf00280883
  21. 21 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. doi: doi:10.1089/met.2008.0034
  22. 22 He S, Xie L, Xie G, Jian G, Jiang K, Lu Z. Independent and combined associations of high-density lipoprotein cholesterol-modified triglyceride-glucose index with all-cause and cardiovascular mortality in patients with acute decompensated heart failure. Front Endocrinol (Lausanne). (2025) 16: 1629066. doi: doi:10.3389/fendo.2025.1629066
  23. 23 Chen X, Zhao X, Fei H, Zhang Y, Pang J, Liu Z. Combined association of triglyceride-glucose index and the atherogenic index of plasma with the incidence of cardiovascular diseases among middle-aged and older population. Cardiovasc Diabetol. (2025) 25: 19. doi: doi:10.1186/s12933-025-03038-3
  24. 24 Zhang Y, Liu B, Zhu Y, Xie Y, Du Y, Xiong P. Associations of cumulative exposure and dynamic trajectories of cholesterol-HDL-glucose (CHG) index with cardiovascular disease in middle-aged and older Chinese adults: a longitudinal analysis. Cardiovasc Diabetol. (2026) 25: 88. doi: doi:10.1186/s12933-026-03137-9
  25. 25 Zhang X, Yang XL, Liu S, Ye YS, Yang QT, Xia LN. Prevalence of hypertension among type 2 diabetes mellitus patients in China: a systematic review and meta-analysis. Int Health. (2024) 16: 144–51. doi: doi:10.1093/inthealth/ihad047
  26. 26 Ma YC, Zuo L, Chen JH, Luo Q, Yu XQ, Li Y. Modified glomerular filtration rate estimating equation for Chinese patients with chronic kidney disease. J Am Soc Nephrol. (2006) 17: 2937–44. doi: doi:10.1681/asn.2006040368
  27. 27 Pan Y, Zhong S, Zhou K, Tian Z, Chen F, Liu Z. Association between diabetes complications and the triglyceride-glucose index in hospitalized patients with type 2 diabetes. J Diabetes Res. (2021) 2021: 8757996. doi: doi:10.1155/2021/8757996
  28. 28 Abdul Murad NA, Abdullah N, Kamaruddin MA, Abd Jalal N, Ismail N, Yusof NAM. Discordance between fasting plasma glucose (FPG) and HbA1c in diagnosing diabetes and pre-diabetes in the Malaysian Cohort. J ASEAN Fed Endocr Soc. (2021) 36: 127–32. doi: doi:10.15605/jafes.036.02.02
  29. 29 Zhang M, Zhu S, Hu D, Fu W, Hu H, Xu Y. Developing and validating a clinlabomics-based machine-learning model for early detection of occult diabetic kidney disease: implications for primary care screening. Front Endocrinol (Lausanne). (2026) 17: 1835866. doi: doi:10.3389/fendo.2026.1835866
  30. 30 Miller RG, Orchard TJ. Understanding metabolic memory: a tale of two studies. Diabetes. (2020) 69: 291–9. doi: doi:10.2337/db19-0514
  31. 31 Lachin JM, Nathan DM. Understanding metabolic memory: the prolonged influence of glycemia during the Diabetes Control and Complications Trial (DCCT) on future risks of complications during the Study of the Epidemiology of Diabetes Interventions and Complications (EDIC). Diabetes Care. (2021) 44: 2216–24. doi: doi:10.2337/dc20-3097
  32. 32 Valente T, Arbex AK. Glycemic variability, oxidative stress, and impact on complications related to type 2 diabetes mellitus. Curr Diabetes Rev. (2021) 17: e071620183816. doi: doi:10.2174/1573399816666200716201550
  33. 33 Oniki K, Shigaki T, Kajiwara-Morita A, Shigetome K, Yoshida A, Jinnouchi H. Quantitative assessment of metabolic memory and its prediction of renal function decline in patients with type 2 diabetes: a retrospective observational study. Diabetes Metab Syndr. (2025) 19: 103225. doi: doi:10.1016/j.dsx.2025.103225
  34. 34 Li J, Shi L, Zhao G, Sun F, Nie Z, Ge Z. High triglyceride levels increase the risk of diabetic microvascular complications: a cross-sectional study. Lipids Health Dis. (2023) 22: 109. doi: doi:10.1186/s12944-023-01873-5
  35. 35 Tu QM, Jin HM, Yang XH. Lipid abnormality in diabetic kidney disease and potential treatment advancements. Front Endocrinol (Lausanne). (2025) 16: 1503711. doi: doi:10.3389/fendo.2025.1503711
  36. 36 Wang C, Wang L, Liang K, Yan F, Hou X, Liu F. Poor control of plasma triglycerides is associated with early decline of estimated glomerular filtration rates in new-onset type 2 diabetes in China: results from a 3-year follow-up study. J Diabetes Res. (2020) 2020: 3613041. doi: doi:10.1155/2020/3613041
  37. 37 Kundu D, Roy A, Mandal T, Bandyopadhyay U, Ghosh E, Ray D. Relation of microalbuminuria to glycosylated hemoglobin and duration of type 2 diabetes. Niger J Clin Pract. (2013) 16: 216–20. doi: doi:10.4103/1119-3077.110159
  38. 38 Lv X, Ran X, Chen X, Luo T, Hu J, Wang Y. Early-onset type 2 diabetes: a high-risk factor for proliferative diabetic retinopathy (PDR) in patients with microalbuminuria. Med (Baltimore). (2020) 99: e20189. doi: doi:10.1097/MD.0000000000020189
  39. 39 Antoniadou C, Gavriilidis E, Ritis K, Tsilingiris D. Anemia in diabetes mellitus: pathogenetic aspects and the value of early erythropoietin therapy. Metabol Open. (2025) 25: 100344. doi: doi:10.1016/j.metop.2024.100344
  40. 40 Kidney Disease: Improving Global Outcomes (KDIGO) Diabetes Work Group (2020). KDIGO 2020 Clinical Practice Guideline for Diabetes Management in Chronic Kidney Disease. Kidney Int. 98 (4S): S1–S115. doi: doi:10.1016/j.kint.2020.06.019
  41. 41 Ahmad E, Sargeant JA, Zaccardi F, Khunti K, Webb DR, Davies MJ. Where does metformin stand in modern day management of type 2 diabetes ? Pharm (Basel). (2020) 13: 427. doi: doi:10.3390/ph13120427
  42. 42 Jiang S, Rosner BA, Colditz GA. Statistically significant association does not imply improvement in prediction of clinical outcomes. Cancer Prev Res (Phila). (2025) 18: 727–33. doi: doi:10.1158/1940-6207.capr-25-0056

Citation

Li, Y., et al. (2026). Associations between a series of composite glucose-lipid indices and moderately increased albuminuria in patients with type 2 diabetes mellitus: a cross-sectional study. Frontiers in Endocrinology, Volume 17. https://doi.org/
Received
Accepted
Published
Volume
17 - 2026

Copyright

Copyright © 2026 Li, Pan, Zhao, Gao, Fu and Ji.

License: https://creativecommons.org/licenses/by/4.0/

*Correspondence: Baolan Ji

Article metrics

Views
0
Downloads
0
Citations
0
Shares
0

Related articles

Systematic ReviewPrognostic value of multiple biomarkers for kidney function decline in type 2 diabetes with relatively preserved kidney function: a systematic review and exploratory meta-analysisJie Hu, Xinyu Pan, Shuwu Wei, Junping WeiOriginal ResearchAssociations of the triglyceride-glucose index and intracranial arterial stenosis with poor functional outcome after ischemic stroke: a retrospective observational study of 712 patientsBaoyi Liao, Yongzhe Gu, Nuo Ma, Aiping Jin, Shengqi YaoOriginal ResearchMetformin and atorvastatin ameliorate MASH via the AMPK/NF-κB pathwayBin Zhang, Na Li, Fang Liu, Guorui Wang, Quanwei Chen, Yiming Sun, Ni Yang, Hao Zhang, Jianjun Han