Abstract
Background
Diabetic kidney disease (DKD) remains a leading cause of end-stage renal disease worldwide. Oxidative stress plays a crucial role in DKD pathogenesis, and 8-hydroxy-2’-deoxyguanosine (8-OHdG), a marker of DNA oxidative damage, may serve as an early predictive biomarker for renal function decline. This study investigated the predictive value of urinary 8-OHdG levels for estimated glomerular filtration rate (eGFR) decline in patients with type 2 diabetes mellitus (T2DM).
Methods
This retrospective cohort study analyzed data from 386 patients with T2DM and baseline eGFR ≥ 60 mL/min/1.73 m² who were followed between January 2020 and December 2024.Participants were followed for 36 months. Urinary 8-OHdG levels were measured by enzyme-linked immunosorbent assay at baseline. The primary outcome was rapid eGFR decline, defined as an annual eGFR reduction ≥ 3.0 mL/min/1.73 m²/year. Cox proportional hazards regression, receiver operating characteristic (ROC) analysis, and Kaplan-Meier survival analysis were performed.
Results
During follow-up, 128 patients (33.2%) experienced rapid eGFR decline. Baseline urinary 8-OHdG levels were significantly higher in patients with rapid decline compared to those without (18.7 ± 4.6 vs. 12.3 ± 3.8 ng/mg creatinine, P < 0.001). After adjusting for traditional risk factors, elevated urinary 8-OHdG (highest tertile) was independently associated with rapid eGFR decline (hazard ratio: 3.42, 95% confidence interval: 2.15–5.44, P < 0.001). The area under the ROC curve for urinary 8-OHdG was 0.786 (95% CI: 0.742–0.826), with optimal cutoff value of 15.2 ng/mg creatinine (sensitivity 78.1%, specificity 72.5%). Combined with clinical variables (age, HbA1c, baseline eGFR, and urinary albumin-to-creatinine ratio), the predictive model achieved an area under the curve of 0.873 (95% CI: 0.837–0.904).
Conclusions
Elevated urinary 8-OHdG levels independently predict rapid eGFR decline in T2DM patients. Incorporation of urinary 8-OHdG into risk stratification models may enhance early identification of patients at high risk for DKD progression and facilitate timely interventions.
Clinical trial number
Not applicable.
Keywords: 8-hydroxy-2'-deoxyguanosine, Diabetic kidney disease, Oxidative stress, Estimated glomerular filtration rate, Biomarker, Type 2 diabetes mellitus
Introduction
Type 2 diabetes mellitus (T2DM) has reached epidemic proportions globally, affecting over 537 million adults worldwide, with projections suggesting this number will exceed 783 million by 2045 [1]. Diabetic kidney disease (DKD) represents one of the most serious microvascular complications of T2DM, occurring in approximately 40% of diabetic patients and constituting the leading cause of end-stage renal disease (ESRD) in developed countries [2, 3]. The burden of DKD extends beyond renal outcomes, as it substantially increases cardiovascular morbidity and mortality risk [4]. Despite advances in glycemic control and renin-angiotensin-aldosterone system inhibition, the incidence of DKD-related ESRD continues to rise, highlighting the critical need for improved risk stratification and early intervention strategies [5, 6].
Current diagnostic approaches for DKD rely predominantly on albuminuria and estimated glomerular filtration rate (eGFR), both of which possess significant limitations [7, 8]. Albuminuria exhibits considerable variability and lacks specificity for diabetic renal injury, as substantial proportions of T2DM patients develop renal function decline without preceding albuminuria [9, 10]. Similarly, eGFR represents a late manifestation of kidney damage, often detected only after significant nephron loss has occurred [11]. Recent evidence suggests that approximately one-third of diabetic patients with declining renal function present with normoalbuminuria, emphasizing the inadequacy of traditional markers for early DKD detection [12]. Consequently, identification of novel biomarkers capable of predicting renal function deterioration before overt clinical manifestations becomes imperative for optimizing patient management and therapeutic interventions.
Oxidative stress has emerged as a fundamental pathogenic mechanism underlying DKD development and progression [13, 14]. Hyperglycemia-induced excessive reactive oxygen species (ROS) generation leads to oxidative damage of cellular macromolecules, including lipids, proteins, and nucleic acids [15]. Among various oxidative modifications, DNA damage represents a particularly critical event in DKD pathogenesis, as it triggers cellular dysfunction, inflammation, and apoptosis in renal tissues [16]. 8-Hydroxy-2’-deoxyguanosine (8-OHdG) serves as the most extensively studied and reliable biomarker of oxidative DNA damage, formed through hydroxyl radical-mediated oxidation of the guanine base in DNA [17]. Following DNA repair or cellular turnover, 8-OHdG is excreted into urine, making urinary 8-OHdG measurement a non-invasive approach for assessing systemic oxidative stress burden [18].
Accumulating evidence indicates that individuals with diabetes exhibit significantly elevated 8-OHdG levels compared to non-diabetic controls, with concentrations correlating positively with disease duration and glycemic control [19, 20]. Moreover, several cross-sectional and small-scale longitudinal studies have demonstrated associations between elevated 8-OHdG levels and the presence or severity of DKD [21, 22]. However, the prospective predictive value of urinary 8-OHdG for renal function decline in T2DM patients remains incompletely characterized. A recent case-control study reported that circulating 8-OHdG levels were higher in DKD patients than in those without renal complications, and each 1 ng/mL increase in serum 8-OHdG corresponded to an 11.5 mL/min/1.73 m² reduction in eGFR [23]. Additionally, plasma 8-OHdG concentrations predicted ESRD development in type 1 diabetes cohorts, though similar comprehensive analyses in T2DM populations are limited [24].
Recent systematic reviews have highlighted oxidative stress biomarkers as promising candidates for DKD risk stratification, yet emphasized the need for prospective validation studies with adequate sample sizes and follow-up durations [25, 26]. Recent position papers and comprehensive reviews have further underscored the importance of identifying novel biomarkers that can detect DKD at earlier stages and predict disease trajectory more accurately than conventional parameters [27, 28]. Furthermore, contemporary prediction models for diabetic kidney disease progression have primarily relied on conventional clinical parameters, achieving moderate discriminative performance with area under the curve (AUC) values typically ranging from 0.70 to 0.80 [29, 30]. Integration of novel biomarkers reflecting distinct pathophysiological mechanisms, particularly oxidative stress pathways, may substantially enhance predictive accuracy and clinical utility [31, 32]. The present study was designed to retrospectively evaluate the association between baseline urinary 8-OHdG levels and subsequent rapid eGFR decline in a well-characterized longitudinal cohort of T2DM patients using archived urine samples, and to assess whether incorporation of urinary 8-OHdG improves the performance of clinical prediction models for DKD progression.
Methods
Study design and population
This retrospective observational cohort study was conducted at The Xiangya Hospital of Central South University. Medical records and laboratory data were retrieved from the hospital electronic health system for patients treated between January 2020 and December 2024.The study protocol received approval from the Xiangya Hospital of Central South University ethics committee, and all participants provided written informed consent prior to enrollment. The study adhered to the principles outlined in the Declaration of Helsinki and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Eligible participants included adults aged 35–75 years with documented T2DM diagnosed according to American Diabetes Association criteria. The inclusion criteria were: (1) T2DM duration ≥ 1 year, (2) baseline eGFR ≥ 60 mL/min/1.73 m² calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation, (3) stable glycemic control defined as hemoglobin A1c (HbA1c) variation < 1.0% over the preceding 6 months, and (4) willingness to attend regular follow-up visits. The exclusion criteria encompassed: (1) type 1 diabetes mellitus or other specific types of diabetes, (2) acute kidney injury within 3 months prior to enrollment, (3) non-diabetic kidney disease confirmed by renal biopsy or clinical diagnosis (including glomerulonephritis, obstructive uropathy, or polycystic kidney disease), (4) active malignancy or history of chemotherapy within 2 years, (5) severe cardiovascular events (myocardial infarction, stroke) within 6 months, (6) chronic inflammatory diseases or immunosuppressive therapy, (7) pregnancy or lactation, and (8) inability to provide informed consent or comply with follow-up procedures.
Medical records of 492 potentially eligible patients were reviewed, of whom 106 were excluded based on predefined criteria: 34 with baseline eGFR < 60 mL/min/1.73 m², 28 with non-diabetic kidney disease, 22 with recent acute cardiovascular events, 14 with active malignancy, and 8 with insufficient follow-up data. The final cohort comprised 386 patients who met inclusion criteria and had complete baseline assessments and follow-up data available.
Clinical and laboratory assessments
Baseline clinical and laboratory data were retrospectively extracted from electronic medical records. Demographic data including age, sex, diabetes duration, smoking status, and medication history were retrieved from the medical records at the time of initial visit. Anthropometric measurements included body weight, height, waist circumference, and systolic/diastolic blood pressure (average of three measurements after 10 min of rest using calibrated electronic sphygmomanometers). Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared.
Fasting venous blood samples were obtained after overnight fasting (≥ 8 h) for measurement of glucose, HbA1c, total cholesterol, triglycerides, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and serum creatinine. HbA1c was quantified using high-performance liquid chromatography (Bio-Rad Variant II Turbo, Hercules, CA, USA), with coefficient of variation < 2.0%. Serum creatinine was measured by enzymatic method traceable to isotope dilution mass spectrometry reference standards. The eGFR was calculated using the CKD-EPI 2021 equation without race coefficient. First-morning void urine specimens were collected for measurement of urinary albumin-to-creatinine ratio (UACR) using immunoturbidimetric assay (Roche Diagnostics, Mannheim, Germany).
Urinary 8-OHdG measurement
Archived urine samples for 8-OHdG analysis that had been collected during routine clinical visits and stored at -80 °C were retrieved and analyzed in batch to minimize inter-assay variability.Urinary 8-OHdG concentrations were determined using a commercially available competitive enzyme-linked immunosorbent assay kit (ELISA; New 8-OHdG Check, Japan Institute for the Control of Aging, Shizuoka, Japan) according to the manufacturer’s instructions. Briefly, urine samples were diluted 1:10 with assay buffer, and duplicate measurements were performed for each sample. The detection limit was 0.5 ng/mL, and intra-assay and inter-assay coefficients of variation were 5.2% and 8.1%, respectively. Urinary 8-OHdG values were normalized to urinary creatinine concentration and expressed as nanograms per milligram of creatinine (ng/mg creatinine) to account for variations in urine concentration. Laboratory personnel performing 8-OHdG assays were blinded to clinical data and outcomes.
Follow-up and outcome assessment
Follow-up data for up to 36 months were retrospectively extracted from medical records. Laboratory results including fasting blood glucose, HbA1c, and serum creatinine from all available follow-up visits (typically conducted every 6 months as per standard clinical practice) were retrieved. Medication changes, hospitalizations, and adverse events were documented from electronic health records.The primary outcome was rapid eGFR decline, defined as an annual eGFR reduction ≥ 3.0 mL/min/1.73 m²/year, consistent with definitions used in previous landmark studies on diabetic kidney disease progression [29, 33]. The eGFR slope for each participant was calculated using linear regression of all eGFR measurements obtained during the follow-up period, requiring a minimum of three eGFR values for slope calculation. Secondary outcomes included: (1) development of macroalbuminuria (UACR ≥ 300 mg/g), (2) composite kidney outcome defined as ≥ 40% sustained eGFR decline from baseline or progression to ESRD (eGFR < 15 mL/min/1.73 m² or initiation of kidney replacement therapy), and (3) cardiovascular events (myocardial infarction, stroke, coronary revascularization, or cardiovascular death).
Statistical analysis
Sample size was determined based on the available patient cohort meeting inclusion criteria during the study period. Post-hoc power analysis confirmed adequate statistical power (> 80%) to detect hazard ratios of 2.0–3.0 for the association between elevated 8-OHdG levels and rapid eGFR decline, given the observed event rate of 33.2% and sample size of 372 patients with complete follow-up data. Continuous variables were expressed as mean ± standard deviation for normally distributed data or median with interquartile range for skewed distributions, assessed by Kolmogorov-Smirnov test. Categorical variables were presented as frequencies and percentages. Baseline characteristics were compared between patients with and without rapid eGFR decline using independent t-tests for normally distributed continuous variables, Mann-Whitney U tests for non-normally distributed continuous variables, and chi-square tests or Fisher’s exact tests for categorical variables. Participants were stratified into tertiles based on baseline urinary 8-OHdG concentrations for initial comparisons, with linear trends across tertiles evaluated using the Jonckheere-Terpstra test for continuous variables and the Cochran-Armitage test for categorical variables. Pearson or Spearman correlation coefficients were calculated to assess relationships between urinary 8-OHdG and clinical or biochemical parameters. Time-to-event analyses employed Kaplan-Meier survival curves with log-rank tests to compare cumulative incidence of rapid eGFR decline across urinary 8-OHdG tertiles. Cox proportional hazards regression models were constructed to estimate hazard ratios and 95% confidence intervals for the association between urinary 8-OHdG (analyzed as both continuous and categorical variable) and rapid eGFR decline, with the proportional hazards assumption verified using Schoenfeld residuals. Four sequential models were developed with progressively increasing adjustment: Model 1 was unadjusted, Model 2 adjusted for age, sex, diabetes duration, and BMI, Model 3 additionally adjusted for HbA1c, systolic blood pressure, and baseline eGFR, and Model 4 (fully adjusted) further included UACR, total cholesterol, smoking status, and use of renin-angiotensin system inhibitors. Restricted cubic spline analysis with four knots positioned at the 5th, 35th, 65th, and 95th percentiles of urinary 8-OHdG distribution was performed to examine potential non-linear associations. Subgroup analyses stratified by age, sex, diabetes duration, baseline HbA1c, and baseline eGFR were conducted, with interaction terms tested using likelihood ratio tests. Receiver operating characteristic curve analysis evaluated the discriminative performance of urinary 8-OHdG for predicting rapid eGFR decline, with the optimal cutoff value determined using Youden’s index. A multivariable prediction model incorporating urinary 8-OHdG and significant clinical variables was developed using stepwise backward selection with a retention threshold of P < 0.10. The net reclassification improvement and integrated discrimination improvement were calculated to quantify the incremental predictive value of adding urinary 8-OHdG to the clinical model. Internal validation was performed using bootstrap resampling with 1,000 iterations to assess model optimism and generate calibration plots. All statistical analyses were performed using R software version 4.3.1 with packages survival, survminer, pROC, and rms. Two-sided P values < 0.05 were considered statistically significant.
Results
Baseline characteristics
The study cohort comprised 386 patients with T2DM who met all inclusion criteria between January 2020 and December 2021. During the observation period extending up to 36 months, 14 patients (3.6%) had incomplete follow-up data and were excluded, leaving 372 participants with complete data for analysis. Among these, 128 patients (33.2%) experienced rapid eGFR decline, defined as an annual eGFR reduction ≥ 3.0 mL/min/1.73 m²/year.Table 1 presents baseline characteristics stratified by rapid eGFR decline status. Patients who developed rapid eGFR decline differed significantly from those who did not across multiple demographic and clinical parameters. These differences included older age, longer diabetes duration, poorer glycemic control, higher systolic blood pressure, and substantially higher baseline UACR. The two groups showed comparable baseline eGFR levels, though those who experienced rapid decline had slightly lower values. Medication use patterns were similar between groups, with the exception of renin-angiotensin system inhibitors, which were more commonly prescribed in patients who subsequently developed rapid decline.
Table 1.
Baseline characteristics stratified by rapid eGFR decline status
| Characteristic | Rapid eGFR Decline (n = 128) | No Rapid Decline (n = 244) | P Value |
|---|---|---|---|
| Age, years | 61.3 ± 8.9 | 57.4 ± 9.5 | < 0.001 |
| Male sex, n (%) | 76 (59.4) | 134 (54.9) | 0.407 |
| Diabetes duration, years | 10.1 ± 6.2 | 7.2 ± 5.1 | < 0.001 |
| Body mass index, kg/m² | 27.8 ± 4.6 | 26.9 ± 4.2 | 0.056 |
| Current smoker, n (%) | 28 (21.9) | 39 (16.0) | 0.161 |
| Systolic blood pressure, mmHg | 138.7 ± 15.3 | 132.4 ± 13.8 | < 0.001 |
| Diastolic blood pressure, mmHg | 84.2 ± 10.7 | 81.6 ± 9.8 | 0.018 |
| Fasting plasma glucose, mmol/L | 8.9 ± 2.4 | 8.1 ± 2.1 | 0.001 |
| HbA1c, % | 8.4 ± 1.6 | 7.6 ± 1.3 | < 0.001 |
| Total cholesterol, mmol/L | 4.8 ± 1.2 | 4.6 ± 1.1 | 0.108 |
| Triglycerides, mmol/L | 2.1 [1.4–3.2] | 1.8 [1.2–2.6] | 0.009 |
| HDL cholesterol, mmol/L | 1.1 ± 0.3 | 1.2 ± 0.3 | 0.012 |
| LDL cholesterol, mmol/L | 2.7 ± 0.9 | 2.6 ± 0.8 | 0.367 |
| Baseline eGFR, mL/min/1.73 m² | 87.6 ± 17.2 | 93.2 ± 19.2 | 0.004 |
| UACR, mg/g | 42.6 [18.7–98.4] | 14.2 [6.8–28.5] | < 0.001 |
| Urinary 8-OHdG, ng/mg Cr | 18.7 ± 4.6 | 12.3 ± 3.8 | < 0.001 |
| Medications, n (%) | |||
| Metformin | 82 (64.1) | 149 (61.1) | 0.572 |
| SGLT-2 inhibitors | 38 (29.7) | 71 (29.1) | 0.901 |
| DPP-4 inhibitors | 61 (47.7) | 108 (44.3) | 0.537 |
| GLP-1 receptor agonists | 24 (18.8) | 42 (17.2) | 0.709 |
| Insulin | 47 (36.7) | 68 (27.9) | 0.076 |
| RAS inhibitors | 76 (59.4) | 121 (49.6) | 0.074 |
| Statins | 89 (69.5) | 158 (64.8) | 0.357 |
Data are expressed as mean ± SD, median [interquartile range], or n (%). eGFR, estimated glomerular filtration rate; UACR, urinary albumin-to-creatinine ratio; 8-OHdG, 8-hydroxy-2’-deoxyguanosine; Cr, creatinine; HDL, high-density lipoprotein; LDL, low-density lipoprotein; SGLT-2, sodium-glucose cotransporter-2; DPP-4, dipeptidyl peptidase-4; GLP-1, glucagon-like peptide-1; RAS, renin-angiotensin system
Urinary 8-OHdG levels and correlation analyses
The median urinary 8-OHdG level in the entire cohort was 14.2 [10.8–18.6] ng/mg creatinine, with values ranging from 4.8 to 36.7 ng/mg creatinine. Urinary 8-OHdG concentrations were significantly higher in patients who developed rapid eGFR decline compared to those who did not (18.7 ± 4.6 vs. 12.3 ± 3.8 ng/mg creatinine, P < 0.001). When participants were stratified into tertiles based on urinary 8-OHdG levels (Tertile 1: <12.0, Tertile 2: 12.0-16.9, Tertile 3: ≥17.0 ng/mg creatinine), the proportion of patients experiencing rapid eGFR decline increased progressively across tertiles (17.1%, 31.5%, and 50.8%, respectively; P for trend < 0.001).Correlation analyses revealed significant associations between urinary 8-OHdG and several clinical and biochemical parameters (Table 2).
Table 2.
Correlation between urinary 8-OHdG and clinical/biochemical parameters
| Parameter | Correlation Coefficient (r) | P Value |
|---|---|---|
| Age, years | 0.32 | < 0.001 |
| Diabetes duration, years | 0.38 | < 0.001 |
| Body mass index, kg/m² | 0.14 | 0.006 |
| Systolic blood pressure, mmHg | 0.27 | < 0.001 |
| Diastolic blood pressure, mmHg | 0.18 | 0.001 |
| Fasting plasma glucose, mmol/L | 0.36 | < 0.001 |
| HbA1c, % | 0.44 | < 0.001 |
| Total cholesterol, mmol/L | 0.09 | 0.087 |
| Triglycerides, mmol/L | 0.23 | < 0.001 |
| HDL cholesterol, mmol/L | -0.22 | < 0.001 |
| LDL cholesterol, mmol/L | 0.12 | 0.021 |
| Baseline eGFR, mL/min/1.73 m² | -0.36 | < 0.001 |
| UACR, mg/g | 0.51 | < 0.001 |
Spearman correlation coefficients are presented for skewed variables (triglycerides, UACR). HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; eGFR, estimated glomerular filtration rate; UACR, urinary albumin-to-creatinine ratio
Association between urinary 8-OHdG and rapid eGFR decline
Kaplan-Meier analysis demonstrated significantly different cumulative incidence rates of rapid eGFR decline across urinary 8-OHdG tertiles (log-rank P < 0.001, Fig. 1). The 36-month cumulative incidence was 17.1%, 31.5%, and 50.8% for Tertiles 1, 2, and 3, respectively.
Fig. 1.
Kaplan-Meier curves for cumulative incidence of rapid eGFR decline stratified by urinary 8-OHdG tertiles
Cox proportional hazards regression analyses evaluating urinary 8-OHdG as a continuous variable revealed a strong association with rapid eGFR decline across all adjustment models (Table 3). In the unadjusted model, each 1-unit (ng/mg creatinine) increase in urinary 8-OHdG was associated with a 15% higher risk of rapid decline (HR: 1.15, 95% CI: 1.12–1.18, P < 0.001). After comprehensive adjustment for demographic, clinical, and biochemical variables (Model 4), the association remained robust (HR: 1.12, 95% CI: 1.09–1.16, P < 0.001). When analyzed categorically, patients in the highest 8-OHdG tertile demonstrated substantially elevated risk compared to the lowest tertile in the fully adjusted model (HR: 3.42, 95% CI: 2.15–5.44, P < 0.001).
Table 3.
Cox proportional hazards regression for association between urinary 8-OHdG and rapid eGFR decline
| Model | HR (95% CI) | P Value |
|---|---|---|
| Continuous (per 1 ng/mg Cr increase) | ||
| Model 1 (unadjusted) | 1.15 (1.12–1.18) | < 0.001 |
| Model 2 | 1.14 (1.11–1.17) | < 0.001 |
| Model 3 | 1.13 (1.10–1.16) | < 0.001 |
| Model 4 | 1.12 (1.09–1.16) | < 0.001 |
| Categorical (tertiles) | ||
| Tertile 1 (< 12.0 ng/mg Cr) | Reference | - |
| Tertile 2 (12.0-16.9 ng/mg Cr) | ||
| Model 1 | 2.12 (1.28–3.51) | 0.004 |
| Model 4 | 1.94 (1.15–3.27) | 0.013 |
| Tertile 3 (≥ 17.0 ng/mg Cr) | ||
| Model 1 | 4.89 (3.08–7.77) | < 0.001 |
| Model 4 | 3.42 (2.15–5.44) | < 0.001 |
Model 1: unadjusted. Model 2: adjusted for age, sex, diabetes duration, and body mass index. Model 3: Model 2 + HbA1c, systolic blood pressure, and baseline eGFR. Model 4 (fully adjusted): Model 3 + UACR, total cholesterol, smoking status, and RAS inhibitor use. HR, hazard ratio; CI, confidence interval; Cr, creatinine; eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c; UACR, urinary albumin-to-creatinine ratio; RAS, renin-angiotensin system
Restricted cubic spline analysis revealed a linear dose-response relationship between urinary 8-OHdG and rapid eGFR decline risk without evidence of threshold effects or plateaus (P for non-linearity = 0.327, Fig. 2). The risk increased progressively across the entire range of observed 8-OHdG values.
Fig. 2.
Dose-response relationship between urinary 8-OHdG and risk of rapid eGFR decline
Subgroup analyses
Subgroup analyses examining the association between urinary 8-OHdG (highest vs. lowest tertile) and rapid eGFR decline across predefined patient characteristics demonstrated consistent results (Fig. 3). The hazard ratios ranged from 2.87 to 4.16 across subgroups without significant interactions (all P for interaction > 0.10), indicating that the predictive value of urinary 8-OHdG was robust regardless of age, sex, diabetes duration, baseline glycemic control, or baseline renal function. Notably, the association appeared stronger in patients with diabetes duration ≥ 10 years (HR: 4.16, 95% CI: 2.34–7.39) compared to those with shorter duration (HR: 2.87, 95% CI: 1.42–5.81), although the interaction did not reach statistical significance (P = 0.184).
Fig. 3.
Subgroup analyses for association between urinary 8-OHdG and rapid eGFR decline
Predictive performance of urinary 8-OHdG
ROC curve analysis assessed the discriminative ability of urinary 8-OHdG for predicting rapid eGFR decline (Fig. 4). Urinary 8-OHdG alone yielded an AUC of 0.786 (95% CI: 0.742–0.826). The optimal cutoff value determined by Youden’s index was 15.2 ng/mg creatinine, corresponding to sensitivity of 78.1%, specificity of 72.5%, positive predictive value of 58.7%, and negative predictive value of 86.4%.
Fig. 4.
Receiver operating characteristic curves for prediction of rapid eGFR decline
A clinical prediction model incorporating traditional risk factors (age, sex, diabetes duration, HbA1c, systolic blood pressure, baseline eGFR, and UACR) achieved an AUC of 0.821 (95% CI: 0.779–0.858). Addition of urinary 8-OHdG to this clinical model significantly improved discriminative performance to 0.873 (95% CI: 0.837–0.904, P = 0.002 for comparison with clinical model alone). The continuous net reclassification improvement was 0.387 (95% CI: 0.244–0.531, P < 0.001), and the integrated discrimination improvement was 0.092 (95% CI: 0.058–0.126, P < 0.001), indicating substantial enhancement in risk stratification when urinary 8-OHdG was incorporated.
Internal validation using bootstrap resampling with 1,000 iterations yielded a bias-corrected AUC of 0.868 (95% CI: 0.831–0.901) for the combined model, indicating minimal optimism and robust calibration. Calibration plots demonstrated excellent agreement between predicted and observed probabilities of rapid eGFR decline across risk deciles (Hosmer-Lemeshow χ² = 7.32, P = 0.503, Fig. 5).
Fig. 5.
Calibration plot for the combined prediction model
Secondary outcomes
During follow-up, 46 patients (12.4%) developed macroalbuminuria (UACR ≥ 300 mg/g), and 38 patients (10.2%) experienced the composite kidney outcome (≥ 40% sustained eGFR decline or progression to ESRD). Elevated urinary 8-OHdG (highest tertile) was significantly associated with both macroalbuminuria development (HR: 3.68, 95% CI: 1.92–7.05, P < 0.001) and the composite kidney outcome (HR: 4.12, 95% CI: 2.04–8.33, P < 0.001) after full adjustment. Additionally, 32 patients (8.6%) experienced cardiovascular events, with a trend toward higher risk in the highest 8-OHdG tertile (HR: 1.95, 95% CI: 0.89–4.27, P = 0.095); however, this association did not reach statistical significance in the fully adjusted model. The relatively low number of cardiovascular events (n = 32) indicates that the study was underpowered for this secondary outcome, and the lack of statistical significance should be interpreted cautiously as the point estimate suggests a potentially meaningful association that warrants investigation in larger cohorts adequately powered for cardiovascular endpoints.
Discussion
This retrospective cohort study demonstrates that elevated urinary 8-OHdG levels measured at baseline are independently associated with rapid eGFR decline in patients with T2DM and preserved renal function at baseline. After comprehensive adjustment for established risk factors, patients in the highest tertile of urinary 8-OHdG exhibited a 3.4-fold increased risk of rapid eGFR decline compared to those in the lowest tertile. The dose-response relationship between 8-OHdG and renal function deterioration appeared linear across the observed range without threshold effects. Incorporation of urinary 8-OHdG into a clinical prediction model containing traditional risk factors significantly improved discriminative performance, with the combined model achieving an AUC of 0.873 and demonstrating robust internal validation. These findings suggest that urinary 8-OHdG may serve as a valuable predictive biomarker for early identification of T2DM patients at heightened risk for DKD progression, potentially enabling more intensive monitoring and timely therapeutic interventions.
The observed association between elevated 8-OHdG and accelerated renal function decline is biologically plausible and consistent with extensive mechanistic evidence implicating oxidative stress in DKD pathogenesis. Hyperglycemia-induced excessive ROS generation has been shown to activate multiple deleterious pathways in renal cells, including protein kinase C, advanced glycation end-product formation, polyol pathway flux, and hexosamine biosynthesis, all of which converge on oxidative stress as a central mediator [13, 14]. Oxidative DNA damage, as reflected by 8-OHdG formation, has been linked to cellular dysfunction through several mechanisms potentially relevant to DKD progression. These include activation of poly(ADP-ribose) polymerase leading to cellular energy depletion and subsequent apoptosis of podocytes and tubular epithelial cells [15], oxidative stress-induced mitochondrial DNA mutations that may compromise cellular respiration efficiency [16], and activation of inflammatory signaling cascades including nuclear factor-κB and mitogen-activated protein kinase pathways that amplify renal inflammation and fibrosis [25, 26]. However, it is important to emphasize that the observational and retrospective nature of our study precludes definitive causal inference. While elevated urinary 8-OHdG appears to be a robust predictive biomarker for renal function decline, whether it represents a direct mediator of kidney damage or merely reflects underlying metabolic derangements that independently promote DKD progression cannot be determined from our data.
The present findings align with and extend previous observations regarding 8-OHdG as a biomarker in diabetic complications. A recent case-control study by Granado-Casas et al. reported that serum 8-OHdG concentrations were significantly elevated in DKD patients compared to diabetic individuals without renal complications, with each 1 ng/mL increase associated with an 11.5 mL/min/1.73 m² reduction in eGFR [23]. Similarly, Sanchez et al. demonstrated in type 1 diabetes cohorts that plasma 8-OHdG predicted ESRD development and all-cause mortality during extended follow-up, with hazard ratios of approximately 2.0-2.5 per standard deviation increase [24]. Our study contributes novel insights by evaluating urinary rather than circulating 8-OHdG in a well-characterized T2DM population with comprehensive outcome ascertainment and covariate adjustment.Urinary measurement offers practical advantages including non-invasive sample collection, stability during storage, and direct reflection of renal oxidative stress burden, potentially enhancing clinical applicability.
The robust predictive performance observed in our study may reflect the multifactorial nature of oxidative stress in DKD pathogenesis. Urinary 8-OHdG integrates signals from multiple upstream metabolic derangements, including chronic hyperglycemia, dyslipidemia, hypertension, and inflammation, as evidenced by significant correlations with HbA1c, UACR, and diabetes duration in our correlation analyses. This integrative property distinguishes 8-OHdG from single-pathway biomarkers and may explain its incremental predictive value beyond traditional risk factors. The observation that the combined model incorporating 8-OHdG achieved superior discrimination compared to clinical variables alone supports the concept that multi-biomarker approaches reflecting distinct pathophysiological mechanisms enhance risk stratification accuracy, consistent with recent recommendations from systematic reviews of DKD biomarkers [31, 32].
Subgroup analyses revealed consistent associations across patient characteristics, suggesting broad applicability of urinary 8-OHdG measurement across diverse diabetic populations. The absence of significant interactions by age, sex, baseline glycemic control, or renal function indicates that 8-OHdG provides incremental prognostic information regardless of these traditional risk modifiers. This consistency strengthens the potential utility of 8-OHdG for personalized risk assessment and treatment decision-making. The trend toward stronger associations in patients with longer diabetes duration, while not statistically significant, might reflect cumulative oxidative burden exposure and warrants investigation in larger cohorts.
Several recent studies have explored oxidative stress biomarkers in diabetic kidney disease with variable findings. Goycheva et al. reported that markers of lipid peroxidation (malondialdehyde and 8-iso-prostaglandin F2α) correlated inversely with eGFR across DKD stages, with correlation coefficients ranging from − 0.85 to -0.92 [25]. Kovačević et al. demonstrated that plasma activities of antioxidant enzymes (glutathione peroxidase and superoxide dismutase) independently predicted annual eGFR decline in diabetic nephropathy patients [33]. These complementary findings suggest that assessment of oxidative stress through multiple biomarkers may capture different aspects of redox dysregulation. Future research should evaluate whether combining 8-OHdG with markers of lipid peroxidation, protein oxidation, or antioxidant capacity further enhances predictive accuracy.
The clinical implications of our findings merit careful consideration. First, measurement of urinary 8-OHdG could facilitate early identification of high-risk patients who might benefit from intensified therapeutic strategies, including stricter glycemic targets, combination renoprotective agents, or novel antioxidant interventions. The optimal cutoff value of 15.2 ng/mg creatinine identified in our study provides a potential threshold for risk stratification; however, this value was derived internally from the same dataset used for model development, which may result in overfitting and optimistic performance estimates. We caution against immediate clinical implementation of this specific cutoff value, and emphasize that rigorous external validation in independent, geographically and ethnically diverse cohorts is essential before this threshold can be recommended for routine clinical use. Second, urinary 8-OHdG might serve as a surrogate endpoint in clinical trials evaluating interventions targeting oxidative stress, potentially accelerating drug development by providing early signals of treatment efficacy before hard renal outcomes manifest. Third, longitudinal monitoring of urinary 8-OHdG changes could assess treatment response and guide therapeutic adjustments, analogous to serial HbA1c measurements for glycemic management.
Recent advances in diabetic kidney disease therapeutics underscore the importance of improved risk prediction. Sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 receptor agonists have demonstrated substantial cardiorenal benefits in diabetes, with effects partly mediated through oxidative stress reduction [26]. The ability to prospectively identify patients at highest risk for rapid progression could enable targeted deployment of these costly therapies to those most likely to benefit. Moreover, emerging therapeutic approaches directly targeting oxidative stress, including Nrf2 activators such as bardoxolone methyl, have shown promise in early-phase trials, though their clinical development requires better patient selection strategies [29]. Integration of urinary 8-OHdG into clinical algorithms might enhance precision medicine approaches by matching treatments to underlying pathophysiological mechanisms.
Several limitations warrant acknowledgment. First, the single-center retrospective design conducted at a tertiary care hospital may limit generalizability to broader diabetic populations. Tertiary care centers typically manage patients with more severe or complicated T2DM, potentially introducing selection bias toward individuals with advanced disease manifestations or multiple comorbidities. Consequently, the applicability of our findings to community-based or primary-care diabetic populations, particularly those with earlier disease stages or less complex clinical profiles, remains uncertain and requires external validation in diverse geographic, ethnic, and healthcare settings. Second, urinary 8-OHdG was measured only at baseline using archived samples stored at -80 °C, representing a single time-point assessment of oxidative stress. Oxidative stress markers are known to fluctuate over time and may be influenced by numerous factors including glycemic control, intercurrent infections, medication changes, dietary patterns, physical activity, and acute stressors during follow-up. This single-time measurement strategy limits our ability to assess temporal variability in 8-OHdG levels and precludes evaluation of whether changes in oxidative stress correlate with trajectory shifts in renal function. Serial 8-OHdG measurements would provide insights into dynamic oxidative stress burden and strengthen causal inference, and future prospective studies should incorporate repeated biomarker assessments. Third, while we adjusted for numerous potential confounders including RAS inhibitor use, several other medications known to influence oxidative stress and renal outcomes were not fully incorporated into multivariable analyses. Specifically, statins, sodium-glucose cotransporter-2 (SGLT-2) inhibitors, and glucagon-like peptide-1 (GLP-1) receptor agonists, all of which possess pleiotropic effects on oxidative stress pathways, were used by subsets of patients but detailed dosing information and treatment duration were incompletely available for comprehensive covariate adjustment. Additionally, key lifestyle factors such as detailed dietary composition, antioxidant supplement intake (including vitamins C and E), physical activity levels, and smoking intensity (pack-years) were not systematically recorded and thus could not be included in final models. We also did not exclude patients based on recent strenuous exercise or acute infections prior to urine collection, which could transiently elevate 8-OHdG levels. The possibility of residual confounding from these unmeasured or incompletely measured variables should be acknowledged when interpreting our findings. Fourth, urinary 8-OHdG was measured using enzyme-linked immunosorbent assay (ELISA), which represents a widely validated and clinically practical approach. However, liquid chromatography-tandem mass spectrometry (LC-MS/MS) is regarded as the gold standard for 8-OHdG quantification due to its superior specificity and reduced potential for cross-reactivity with structurally similar compounds. ELISA-based measurements may occasionally detect related oxidized nucleosides, potentially affecting absolute concentration values, though relative comparisons within the same assay platform remain valid for risk stratification purposes. Fifth, the ROC-derived cutoff value of 15.2 ng/mg creatinine for urinary 8-OHdG demonstrated reasonable discrimination in our cohort, but this threshold was derived from the same dataset used for model development. Without external validation in independent cohorts, the clinical applicability and generalizability of this specific cutoff remain uncertain, and we caution against immediate clinical implementation pending prospective validation studies. Sixth, the 36-month follow-up duration, while adequate for assessing rapid eGFR decline, may be insufficient to capture long-term outcomes such as ESRD development, which typically requires decades in many patients.
In conclusion, elevated urinary 8-OHdG levels are independently associated with rapid eGFR decline in patients with type 2 diabetes mellitus, exhibiting a dose-response relationship and providing incremental prognostic value beyond established clinical risk factors. Incorporation of urinary 8-OHdG into prediction models significantly enhances discriminative performance for identifying high-risk patients. These findings suggest that urinary 8-OHdG represents a promising predictive biomarker for early detection and risk stratification of diabetic kidney disease progression; however, the retrospective observational nature of this study precludes causal inference, and prospective validation in independent cohorts is essential before clinical implementation. Translation of these observations into clinical practice could facilitate personalized medicine approaches by enabling targeted intensive interventions for patients at greatest risk, potentially reducing the burden of diabetic kidney disease and its devastating complications.
Acknowledgements
We sincerely thank our colleagues for their enthusiastic efforts and help, and thank the author for their active assistance!
Author contributions
Ke Chen. submitted ethics approval, collected the data. Juan Wu started the original draft of the manuscript. Juan Wu and Ying Lu was the supervising investigator; he prepared the original elements of the protocol, Ping Sheng supervised the data collection. Ke Chen Reviewed the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding
This work is funded by Hunan Natural Science Foundation (2021JJ70141).
Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethical approval
The study protocol received approval from the Xiangya Hospital of Central South University ethics committee, and all participants provided written informed consent prior to enrollment. The study adhered to the principles outlined in the Declaration of Helsinki and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
Consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.ElSayed NA, Aleppo G, Aroda VR, et al. Chronic kidney disease and risk management: standards of care in diabetes-2025. Diabetes Care. 2025;48(Suppl 1):S239-S255. [DOI] [PMC free article] [PubMed]
- 2.Alicic RZ, Rooney MT, Tuttle KR. Diabetic kidney disease: Challenges, progress, and possibilities. Clin J Am Soc Nephrol. 2017;12(12):2032–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Samsu N. Diabetic nephropathy: challenges in pathogenesis, diagnosis, and treatment. BioMed Res Int. 2021;2021:1497449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Tonelli M, Muntner P, Lloyd A, et al. Risk of coronary events in people with chronic kidney disease compared with those with diabetes: a population-level cohort study. Lancet. 2012;380(9844):807–14. [DOI] [PubMed] [Google Scholar]
- 5.Thomas MC, Cooper ME, Zimmet P. Changing epidemiology of type 2 diabetes mellitus and associated chronic kidney disease. Nat Rev Nephrol. 2016;12(2):73–81. [DOI] [PubMed] [Google Scholar]
- 6.Bennett JE, Stevens GA, Mathers CD, et al. NCD Countdown. 2030: worldwide trends in non-communicable disease mortality and progress towards sustainable development goal target 3.4. Lancet. 2018;392(10152):1072–1088. [DOI] [PubMed]
- 7.Bakris GL, Molitch M. Microalbuminuria as a risk predictor in diabetes: the continuing Saga. Diabetes Care. 2014;37(3):867–75. [DOI] [PubMed] [Google Scholar]
- 8.Alicic RZ, Neumiller JJ, Johnson EJ, Dieter B, Tuttle KR. Sodium-Glucose cotransporter 2 Inhibition and diabetic kidney disease. Diabetes. 2019;68(2):248–57. [DOI] [PubMed] [Google Scholar]
- 9.Fioretto P, Steffes MW, Mauer M. Glomerular structure in nonproteinuric IDDM patients with various levels of albuminuria. Diabetes. 1994;43(11):1358–64. [DOI] [PubMed] [Google Scholar]
- 10.Pugliese G, Penno G, Natali A, et al. Diabetic kidney disease: new clinical and therapeutic issues. Joint position statement of the Italian diabetes society and the Italian society of nephrology on the natural history of diabetic kidney disease and treatment of hyperglycemia in patients with type 2 diabetes and impaired renal function. Nutr Metab Cardiovasc Dis. 2019;29(11):1127–50. [DOI] [PubMed] [Google Scholar]
- 11.Colhoun HM, Marcovecchio ML. Biomarkers of diabetic kidney disease. Diabetologia. 2018;61(5):996–1011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Feakins BG, McFadden EC, Farmer AJ, Stevens RJ. Standard and competing risk analysis of the effect of albuminuria on cardiovascular and cancer mortality in patients with type 2 diabetes mellitus. Diagn Progn Res. 2018;2:13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Wang N, Zhang C. Oxidative stress: A culprit in the progression of diabetic kidney disease. Antioxid (Basel). 2024;13(4):455. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Su S, Ma Z, Wu H, et al. Oxidative stress as a culprit in diabetic kidney disease. Life Sci. 2023;322:121661. [DOI] [PubMed] [Google Scholar]
- 15.Zhong M, Zhu E, Li N, et al. Identification of diagnostic markers related to oxidative stress and inflammatory response in diabetic kidney disease by machine learning algorithms: evidence from human transcriptomic data and mouse experiments. Front Endocrinol (Lausanne). 2023;14:1134325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sakashita M, Tanaka T, Inagi R. Metabolic changes and oxidative stress in diabetic kidney disease. Antioxid (Basel). 2021;10(7):1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hojs NV, Bevc S, Ekart R, et al. Oxidative stress markers in chronic kidney disease with emphasis on diabetic nephropathy. Antioxid (Basel). 2020;9(10):925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Vodošek Hojs N, Bevc S, Ekart R, Hojs R. Oxidative stress markers in chronic kidney disease with emphasis on diabetic nephropathy. Antioxid (Basel). 2020;9(10):925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rajlic S, Treede H, Münzel T, et al. Early detection is the best prevention—characterization of oxidative stress in diabetes mellitus and its consequences on the cardiovascular system. Cells. 2023;12(4):583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zhang P, Li T, Wu X, et al. Oxidative stress and diabetes: antioxidative strategies. Front Med. 2020;14(5):583–600. [DOI] [PubMed] [Google Scholar]
- 21.Zaki M, Abdallah H, Orban H, et al. Assessment of urinary 8-hydroxy-2-deoxyguanosine as oxidative DNA damage for diabetic nephropathy. Egypt J Chem. 2024;67(6):405–8. [Google Scholar]
- 22.Serdar M, Sertoglu E, Uyanik M, et al. Comparison of 8-hydroxy-2’-deoxyguanosine (8-OHdG) levels using mass spectrometer and urine albumin creatinine ratio as a predictor of development of diabetic nephropathy. Free Radic Res. 2012;46(10):1291–5. [DOI] [PubMed] [Google Scholar]
- 23.Spoto B, Politi C, Pizzini P, et al. 8-hydroxy-2’-deoxyguanosine, a biomarker of oxidative DNA injury, in diabetic kidney disease. Nutr Metab Cardiovasc Dis. 2025;35(2):103722. [DOI] [PubMed] [Google Scholar]
- 24.Sanchez M, Roussel R, Hadjadj S, et al. Plasma concentrations of 8-hydroxy-2’-deoxyguanosine and risk of kidney disease and death in individuals with type 1 diabetes. Diabetologia. 2018;61(4):977–84. [DOI] [PubMed] [Google Scholar]
- 25.Goycheva P, Petkova-Parlapanska K, Georgieva E, et al. Biomarkers of oxidative stress in diabetes mellitus with diabetic nephropathy complications. Int J Mol Sci. 2023;24(17):13541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ma X, Ma J, Leng T, et al. Advances in oxidative stress in pathogenesis of diabetic kidney disease and efficacy of TCM intervention. Ren Fail. 2023;45(1):2146512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.From the American Association of Neurological Surgeons (AANS), American Society of Neuroradiology (ASNR), Cardiovascular and Interventional Radiology Society of Europe (CIRSE), Canadian Interventional Radiology Association (CIRA), Congress of Neurological Surgeons (CNS), European Society of Minimally Invasive Neurological Therapy (ESMINT), European Society of Neuroradiology (ESNR), European Stroke Organization (ESO), Society for Cardiovascular Angiography and Interventions (SCAI), Society of Interventional Radiology (SIR), Society of NeuroInterventional Surgery (SNIS), and World Stroke Organization (WSO), Sacks D, Baxter B, et al. Multisociety consensus quality improvement revised consensus statement for endovascular therapy of acute ischemic stroke. Int J Stroke. 2018;13(6):612–632. [DOI] [PubMed]
- 28.Zarezadeh M, Musazadeh V, Foroumandi E, Keramati M, Ostadrahimi A, Mekary RA. The effect of cinnamon supplementation on glycemic control in patients with type 2 diabetes or with polycystic ovary syndrome: an umbrella meta-analysis on interventional meta-analyses. Diabetol Metab Syndr. 2023;15(1):127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.González-Rocha A, Colli VA, Denova-Gutiérrez E. Risk prediction score for chronic kidney disease in healthy adults and adults with type 2 diabetes: systematic review. Prev Chronic Dis. 2023;20:220380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yang L, Wu Y, Wang Y, et al. Prediction factors and models for chronic kidney disease in type 2 diabetes mellitus: A review of the literature. Clin Transl Discov. 2024;4(4):e355. [Google Scholar]
- 31.Swaminathan SM, Rao IR, Shenoy SV, et al. Novel biomarkers for prognosticating diabetic kidney disease progression. Int Urol Nephrol. 2023;55(4):913–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chen Y, Liu X, Shengbu M, et al. Biomarkers: new advances in diabetic nephropathy. Nat Prod Commun. 2025;20(1):1934578X251321758. [Google Scholar]
- 33.Kovačević M, Mališ S, Pavlović D, et al. Plasma activity of the antioxidant enzymes in predicting diabetic nephropathy progression. Int Urol Nephrol. 2022;54(6):1365–72. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.