The potential biomarker value of soluble CD36 in the treatment of diabetic kidney disease: evidence from GLP-1 and insulin interventions

This study was designed as a real-world, prospective, observational cohort study conducted at the Affiliated Hospital of Qingdao University. Patients with type 2 diabetes mellitus (T2DM) who were hospitalized in the Department of Endocrinology and Metabolic Diseases between 2024 and 2025 were screened for eligibility. All participants met the predefined inclusion and exclusion criteria and provided written informed consent prior to enrollment. The study protocol was approved by the Medical Ethics Committee of the Affiliated Hospital of Qingdao University (Ethical Approval No.QYFYWZLL29898).To minimize selection bias, an stratified assignment strategy was employed. Eligible patients were stratified and assigned in a 1:1:1 ratio to the control group, the GLP-1RA group, or the insulin group using a computer-generated block randomization system, managed by a research assistant who was not involved in clinical treatment or data collection. The allocation sequence was blinded to the primary investigators, thereby enhancing objectivity and comparability between groups. Although this was an observational study and not a randomized controlled trial (RCT), the use of structured allocation procedures strengthened the internal validity of the group comparisons. The study was conducted in strict accordance with the principles of Good Clinical Practice (GCP), the Declaration of Helsinki, and relevant Chinese regulations on clinical research ethics. All participants provided written informed consent. Baseline demographic data (e.g., height, weight, age) were collected by trained healthcare professionals and kept strictly confidential. All patient-related information was used solely for research purposes and was not disclosed or repurposed without authorization. Sodium-glucose co-transporter-2 (SGLT2) inhibitors were excluded from this study to avoid confounding effects on renal outcomes, given their known nephroprotective properties and potential influence on inflammation and oxidative stress. This also allowed a more focused comparison between GLP-1 receptor agonists and insulin with respect to their impact on sCD36 and UACR.

Based on a priori power analysis using G*Power (effect size: Cohen's d = 0.8; α = 0.05; power = 80%), a minimum of 70 participants per group was required to achieve statistical significance. A total of 191 patients were ultimately enrolled in this study (Figure 1), meeting the criteria for valid statistical analysis. Participants were screened according to predefined inclusion and exclusion criteria and assigned to one of three treatment groups based on their therapeutic regimen: the control group, the GLP-1RA group, and the insulin group. All participants received a 12-week course of treatment. Metformin (500 mg, twice daily) was administered as background therapy in all groups. The control group received no additional intervention. The GLP-1RA group received subcutaneous injections of a GLP-1 receptor agonist (0.5 mg weekly) in addition to metformin. The insulin group received daily subcutaneous insulin injections, with dosages adjusted individually based on glycemic status.

Data Collection: At baseline and after 12 weeks of treatment, fasting venous blood samples were collected to measure the following clinical indicators: triglycerides (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), free fatty acids (FFA), glycated hemoglobin (HbA1c), fasting insulin, fasting C-peptide, sCD36, urinary albumin-to-creatinine ratio (UACR), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and serum uric acid. Blood was collected into procoagulant separation tubes (≥1 mL), stored at 4°C, and centrifuged within 2 hours at 3000 rpm for 10 minutes. The supernatant was aliquoted and stored at –80°C for subsequent analysis. Serum sCD36 concentrations were measured using a commercially available enzyme-linked immunosorbent assay (ELISA) kit, following the manufacturer's protocol. All samples were analyzed in the same experimental batch. Laboratory staff were blinded to group assignments, and the inter-assay variability was controlled within 10%.

Statistical Analysis: Changes in key clinical indicators before and after treatment were compared within and between groups to evaluate the efficacy of different therapeutic regimens in glycemic control, lipid regulation, renal protection, and modulation of sCD36 levels. In addition, intergroup comparisons were performed, and adverse events were recorded to assess the overall safety and clinical utility of each intervention.

Inclusion Criteria: Participants were eligible for inclusion if they met all of the following criteria:1. Age between 15 and 75 years;2. Body mass index (BMI) ranging from 18.5 to 40 kg/m²;3. Poor glycemic control, defined as glycated hemoglobin A1c (HbA1c) levels between 7.5% and 10% despite adherence to diet and exercise interventions;4. Urinary albumin-to-creatinine ratio (UACR) between 30 and 300 mg/g, indicating microalbuminuria;5. No prior use of GLP-1RAs or insulin for glycemic management.

Exclusion Criteria: Participants were excluded from the study if they met any of the following criteria:1. History of acute diabetic complications such as diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic syndrome (HHS);2. Presence of treatment-resistant hypertension requiring four or more antihypertensive agents for control;3. Recent acute cardiovascular or cerebrovascular events, active infections, ketoacidosis, stress-related conditions, hepatic or renal dysfunction, anemia, other endocrine disorders, primary acute or chronic glomerulonephritis, urinary tract infections, urinary obstruction, or other urinary system diseases;4. Use of glucocorticoids or other medications that affect blood glucose levels, or use of diuretics and other drugs that may interfere with urinary albumin excretion rate (UAER) measurements;5. Presence of cardiac pacemakers, neurostimulators, or metallic prosthetic heart valves;6. Comorbid hepatic-pancreatic-renal conditions, including liver cirrhosis, active hepatitis, or impaired renal function (estimated glomerular filtration rate, eGFR < 60 mL/min/1.73 m²);7. History of malignancy or other serious illnesses within the past five years at the time of screening.

All statistical analyses were performed using IBM SPSS Statistics version 29.0 (IBM Corp., Armonk, NY, USA). Continuous variables were expressed as mean ± standard deviation (SD). For normally distributed data, paired t-tests were used to compare pre- and post-treatment values. For non-normally distributed data, the Wilcoxon signed-rank test was applied. Comparisons among the three groups were conducted using one-way analysis of variance (ANOVA), ANCOVA was employed to adjust for significant baseline differences in HbA1c and other metabolic parameters. These variables were entered as covariates to reduce potential bias in the estimation of treatment effects. When statistical significance was observed, the least significant difference (LSD) test was used for post hoc pairwise comparisons. For non-normally distributed variables, the Kruskal–Wallis test was employed, followed by Dunn–Bonferroni post hoc correction for multiple comparisons. Categorical variables were expressed as frequencies and percentages and compared using the chi-square test or Fisher's exact test as appropriate.

Baseline characteristics among the three groups (control, GLP-1RA, and insulin groups) were compared using one-way analysis of variance (ANOVA) for normally distributed continuous variables, the Kruskal–Wallis test for non-normally distributed variables, and the chi-square (χ²) test for categorical variables. Within-group changes in metabolic parameters before and after treatment were evaluated using paired t-tests for normally distributed variables or the Wilcoxon signed-rank test for non-normally distributed variables. Key metabolic parameters analyzed included triglycerides (TG), total cholesterol (TC), low-density lipoprotein (LDL), high-density lipoprotein (HDL), free fatty acids (FFA), HbA1c, fasting insulin, fasting C-peptide, sCD36, urinary albumin-to-creatinine ratio (UACR), alanine aminotransferase (ALT), aspartate aminotransferase (AST), and serum uric acid. Between-group comparisons of changes in these metabolic and biochemical parameters after treatment were conducted using ANOVA or the Kruskal–Wallis test, as appropriate. To control for potential pre-treatment differences, analysis of covariance (ANCOVA) was applied, followed by post hoc testing using the least significant difference (LSD) method to identify pairwise group differences.

sCD36, the circulating form of CD36, is typically released into the bloodstream via proteolytic cleavage from tissue-expressed CD36. Measurement of serum sCD36 levels provides a non-invasive approach to indirectly assess CD36 activity in renal tissue. The following analyses were performed: Group Comparisons: Changes in sCD36 levels before and after treatment were compared among the control, GLP-1RA, and insulin groups using analysis of variance (ANOVA) or the Kruskal–Wallis test, depending on data distribution. Post hoc comparisons between groups were conducted using Bonferroni correction to identify statistically significant pairwise differences.GLP-1RA–Specific Analysis: In patients treated with GLP-1RAs, the change in sCD36 levels before and after treatment (ΔsCD36) was calculated. The relationship between ΔsCD36 and improvements in renal function, as reflected by changes in UACR, was further examined. Correlation and Regression: Correlations between sCD36 levels and renal function indicators, such as UACR and estimated glomerular filtration rate (eGFR), were evaluated using either Spearman or Pearson correlation analysis, depending on variable distribution. To assess whether sCD36 was independently associated with renal function, multivariable linear or logistic regression models were employed, adjusting for potential confounders including age, sex, body mass index (BMI), and duration of diabetes. Subgroup Analysis and Comparative Evaluation: Subgroup comparisons were performed to assess the magnitude of change in sCD36, UACR, and other core indicators. Differences in therapeutic effects between the GLP-1RA and insulin groups were also analyzed to evaluate the relative impact of each intervention on sCD36 modulation and renal function outcomes.

Statistical significance was defined as p < 0.05 (*), and highly significant results were indicated as p < 0.01 (**). To enhance the reliability of the findings, all tests were two-tailed, with a significance threshold set at 5%. For variables demonstrating significant correlations, partial correlation analyses were further conducted to account for the influence of potential confounding factors.

A total of 191 participants completed the study and were categorized into three groups according to their treatment regimen: control group (n = 63), insulin group (n = 71), and GLP-1RA group (n = 57). Baseline characteristics for all groups are summarized in Table 1. No statistically significant differences were observed among the groups at baseline with respect to body weight, duration of diabetes, lipid parameters (TG, TC, LDL-C, HDL-C, FFA), liver function indicators (AST, ALT), urinary albumin-to-creatinine ratio (UACR), or sCD36 levels (all p > 0.05).

Post-treatment comparisons of clinical parameters among the three groups are presented in Table 2A. Significant between-group differences were observed in several metabolic and renal function markers, including fasting blood glucose (FBG), glycated hemoglobin (HbA1c), fasting insulin, fasting C-peptide, homeostasis model assessment of insulin resistance (HOMA-IR), β-cell function index (HOMA-β), urinary albumin-to-creatinine ratio (UACR), and sCD36 levels (all p < 0.05; see Figure 2).Notably, sCD36 levels were significantly lower in the GLP-1RA group [median: 195.20 ng/mL, interquartile range (IQR): 160.45–314.75], followed by the insulin group [364.60 ng/mL, IQR: 279.10–394.10], with the highest levels observed in the control group [386.10 ng/mL, IQR: 323.60–471.30]. The difference among groups was statistically significant (F = 5.098, p < 0.05).In contrast, no significant differences were found among groups for lipid parameters (TG, TC, LDL-C, HDL-C, FFA) or liver function markers (AST, ALT) (p > 0.05 for all). Combined with baseline findings, where no between-group differences were noted in UACR, sCD36, or lipid/liver function parameters, these results suggest that the observed post-treatment changes in UACR and sCD36 may be attributed to the different therapeutic regimens, whereas changes in lipid and hepatic markers were not statistically significant.

To further evaluate the effects of different therapeutic strategies, within-group comparisons were performed for key clinical parameters before and after treatment (Table 2B). Significant improvements were observed in glycemic control and lipid metabolism across all three groups following treatment (p < 0.05).Regarding liver function, both AST and ALT levels significantly decreased in the GLP-1RA group (AST: p < 0.001; ALT: p < 0.05), while AST levels also decreased significantly in the insulin group (p < 0.05). No significant changes were observed in the control group. For pancreatic function markers, fasting insulin levels increased significantly in the control and insulin groups (p < 0.05), but not in the GLP-1RA group. Fasting C-peptide levels significantly increased in both the insulin (p < 0.05) and GLP-1RA (p < 0.001) groups, indicating potential enhancement of β-cell function. The homeostasis model assessment of insulin resistance (HOMA-IR) was significantly reduced in the insulin group (p < 0.05), but remained unchanged in the GLP-1RA and control groups. In contrast, the β-cell function index (HOMA-β) increased significantly across all three groups (p < 0.001), suggesting a general improvement in islet β-cell function following treatment. For DKD-related markers, both the insulin and GLP-1RA groups showed significant reductions in the urinary albumin-to-creatinine ratio (UACR) after treatment (p < 0.001 for both), whereas no significant change was observed in the control group. In addition, both UACR and sCD36 levels were significantly decreased in the insulin and GLP-1RA groups (p < 0.001; Figure 3), with the greatest reduction observed in the GLP-1RA group, suggesting a potentially superior renoprotective effect of GLP-1RA therapy.

This study further analyzed between-group differences in clinical indicators before and after treatment among the control, insulin, and GLP-1RA groups (Table 3A). Significant differences were observed among the groups in fasting blood glucose (FBG), glycated hemoglobin (HbA1c), fasting insulin, fasting C-peptide, homeostasis model assessment of insulin resistance (HOMA-IR), β-cell function index (HOMA-β), urinary albumin-to-creatinine ratio (UACR), and sCD36, with all p < 0.05.In parallel, within-group comparisons confirmed that all three groups experienced some degree of clinical improvement after treatment (Table 2B). Of particular interest, baseline levels of UACR and sCD36 did not differ significantly between the groups (p > 0.05); however, significant between-group differences were observed post-treatment (Table 3A), suggesting differential effects of the interventions on renal injury markers. Pairwise post hoc comparisons revealed that the GLP-1RA group exhibited significantly greater reductions in both UACR and sCD36 levels compared to the insulin and control groups (all p < 0.05), highlighting the superior renoprotective potential of GLP-1RA therapy in the management of DKD. Regarding metabolic indicators such as FBG, HbA1c, fasting insulin, C-peptide, HOMA-IR, and HOMA-β, significant baseline differences already existed among groups. These differences persisted in the post-treatment data and remained significant even in the change-from-baseline values (Table 2B), raising concerns that baseline imbalances may confound the interpretation of treatment effects. Given the statistically significant baseline differences in HbA1c, HOMA-IR, and fasting insulin levels across groups (p < 0.05), we performed ANCOVA to adjust for these imbalances when comparing post-treatment outcomes. This statistical approach allowed us to mitigate the influence of confounding variables and isolate the treatment effects more accurately. The ANCOVA results (Table 3B) showed that significant overall between-group differences remained for fasting insulin (F = 8.35, p < 0.001), HOMA-IR (F = 12.56, p < 0.001), and HOMA-β (F = 5.20, p = 0.006). However, subsequent pairwise comparisons using the least significant difference (LSD) method failed to detect statistically significant differences between individual treatment groups (all p > 0.05), indicating that although overall trends were present, definitive pairwise treatment effects could not be confirmed. In summary, the findings suggest that the different therapeutic regimens exert significantly different effects on DKD-related biomarkers, particularly UACR and sCD36, with GLP-1RA therapy demonstrating a potentially superior clinical benefit. For other metabolic indicators, such as fasting insulin, HOMA-IR, and HOMA-β, although overall group differences were observed, further studies with larger sample sizes are needed to confirm the specific intergroup effects. After adjusting for baseline HbA1c levels using ANCOVA, the between-group differences in post-treatment outcomes remained significant for sCD36 and UACR (p < 0.05), supporting the robustness of our findings.

Correlation analyses revealed that HbA1c, sCD36, and uric acid were significantly associated with UACR both before and after treatment (p < 0.05). To further investigate the strength of these associations, a generalized linear model (GLM) with a gamma distribution and log link function was employed to analyze baseline, post-treatment, and change-from-baseline values for HbA1c, uric acid, and sCD36. At baseline, only sCD36 was significantly associated with UACR (Table 4, regression coefficient = 0.007, Exp(B) = 1.007, 95% CI: 1.006–1.007, Wald χ² = 921.179, p < 0.001), indicating that elevated baseline sCD36 was strongly associated with increased risk of DKD progression. In contrast, baseline HbA1c (regression coefficient = 0.021, Exp(B) = 1.021, 95% CI: 0.994–1.049, p = 0.129) and uric acid (regression coefficient ≈ 0, Exp(B) = 1.000, 95% CI: 1.000–1.001, p = 0.156) were not significantly associated with UACR. After treatment, both sCD36 and uric acid remained significantly associated with UACR. Specifically, sCD36 had a regression coefficient of 0.006 (Exp(B) = 1.006, 95% CI: 1.005–1.006, Wald χ² = 728.521, p < 0.001), while uric acid had a coefficient of 0.001 (Exp(B) = 1.001, 95% CI: 1.000–1.001, Wald χ² = 8.009, p < 0.05). Post-treatment HbA1c was not significantly associated with UACR (regression coefficient = 0.036, Exp(B) = 1.037, 95% CI: 0.986–1.091, p = 0.159).Analysis of change values (post-treatment minus baseline) revealed that the change in HbA1c was significantly and negatively associated with UACR (regression coefficient = –0.410, Exp(B) = 0.664, 95% CI: 0.518–0.850, Wald χ² = 10.511, p < 0.05), suggesting that HbA1c reduction was closely related to improved DKD status. Additionally, the change in sCD36 was positively associated with the change in UACR (Exp(B) = 1.000, 95% CI: 0.997–1.002, p<0.001), indicating that a reduction in sCD36 was significantly correlated with DKD improvement. However, changes in uric acid were not significantly associated with UACR change (regression coefficient = 0.011, Exp(B) = 1.011, 95% CI: 1.005–1.017, Wald χ² = 12.440, p>0.05).