Клинопатологические характеристики и результаты у пациентов с диабетической нефропатией сосуществуют с идиопатичной мембранопатией: ретроспективное исследование наблюдений.

Introduction

In recent years, the global incidence of diabetes mellitus (DM) has increased significantly, raising widespread concern [ 1 ]. Diabetic nephropathy (DN), a severe chronic microvascular complication of DM, affects nearly 50% of patients as the disease progresses [ 2 , 3 ]. The clinical course of diabetic nephropathy initially manifests as microalbuminuria; with disease progression, characteristic nodular glomerulosclerosis emerges on renal pathology, accompanied by increasing proteinuria and ultimately leading to end-stage renal disease (ESRD) [ 4 , 5 ]. Membranous nephropathy (MN) is the most common glomerular disease, characterized by subepithelial electron-dense deposits of immunoglobulin along the glomerular basement membrane (GBM) and clinically manifesting as nephrotic syndrome or severe proteinuria [ 6–8 ]. The increasing prevalence of DM has also led to a rise in cases of DM combined with MN (DM-MN) [ 9 ]. However, few clinical studies have examined diabetic nephropathy with membranous nephropathy (DN-MN). Consequently, the clinical characteristics and disease prognosis of DN-MN patients remain unclear, limiting the understanding and management of this condition in clinical practice.

Additionally, differing opinions exist regarding glucocorticoid use in DN-MN patients. Although the Kidney Disease Improving Global Outcomes (KDIGO) 2022 guidelines provide recommendations for the management of DM-associated chronic kidney disease, they do not offer specific guidance for DN combined with MN [ 10 , 11 ]. Some studies suggest that MN patients are primarily treated with glucocorticoids and immunosuppressive therapy [ 12 , 13 ]. However, glucocorticoids can affect blood glucose levels and carry a risk of inducing steroid-induced diabetes, necessitating cautious use in DM patients [ 14 ]. The application of glucocorticoids in the treatment of patients with DN combined with MN remains controversial. Whether hypoglycemic agents combined with glucocorticoids or immunosuppressants can improve the clinical outcomes and prognosis of DN-MN remains to be explored.

This study is a single-center retrospective cohort analysis comparing clinical and pathological differences among the DN-MN group, the DM-MN group, and the MN group. The study aims to evaluate the clinical and pathological features of patients with DN-MN, identify pathological features influencing renal outcomes, and construct a nomogram to predict renal prognosis, providing clinicians with a practical tool for prognosis assessment in DN-MN and DM-MN. Additionally, this study investigates the impact of glucocorticoid use on patient outcomes to offer new strategies for optimizing treatment of this disease.

Materials and methods

Patient selection

This study included patients with DN-MN diagnosed by renal biopsy between February 24, 2015, and October 1, 2023, at the Second Xiangya Hospital of Central South University, in accordance with the KDIGO Guidelines. Patients with DM-MN and patients with isolated MN served as control groups. Exclusion criteria included systemic diseases, malignancy, hepatitis B or C virus infection, tuberculosis, other comorbid conditions, and age under 18 years. The final study cohort comprised 34 patients with DN-MN, 44 patients with DM-MN, and 198 patients with MN for baseline clinical and pathological characteristics analysis. The detailed patient selection flowchart is presented in Figure 1(A) . Verbal informed consent was obtained through both in-person outpatient visits and telephone follow-ups. This study adhered to the Declaration of Helsinki and received approval from the Ethical Review Committee of the Second Xiangya Hospital of Central South University (ethical approval number: 139).

Flowchart of study participants and comparison & correlation of clinicopathological features in DN-MN, DM-MN, and MN patients. (A) Flowchart of study participants. (B) H&E staining (magnification, ×20), Periodic acid-silver methenamine (PASM) staining (magnification, ×400), fluorescence staining for IgG deposition (magnification, ×400) were performed on renal tissue sections from the three groups. In the DM-MN panel, green arrows indicate prominent spike formation. In the DN-MN panel, green arrows indicate spikes and thickening of the glomerular basement membrane, with visible Kimmelstiel–Wilson nodules. (C–D) Electron microscopy, (C) MN, showing subepithelial deposits. (D) DN-MN, showing both subepithelial deposits and diffuse glomerular basement membrane thickening. (E–F) Spearman correlation analyses were performed to examine the relationships between segmental sclerosis (E), IFTA (F), and clinical manifestations in patients with MN. The heatmap represents the mean values of Spearman’s ρ. * p  < 0.05. (the first column). DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; MN: membranous nephropathy; IFTA: interstitial fibrosis and tubular atrophy; SBP: systolic blood pressure; DBP: diastolic blood pressure; LDL: low-density lipoprotein; TP: total protein; TCHO: total cholesterol; 24h Upro: 24-hour urinary protein; eGFR: estimated glomerular filtration rate; BUN: blood urea nitrogen; Cr: serum creatinine; UA: blood uric acid; ACR: albumin-to-creatinine ratio; IgG, immunoglobulin G.

Clinical and pathological data collection

Clinical data and medication records for patients diagnosed by renal biopsy were obtained from the hospital’s electronic medical record system. Basic clinical parameters included age, sex, systolic blood pressure (SBP), diastolic blood pressure (DBP), erythrocyte sedimentation rate (ESR), hemoglobin (Hb), high-density lipoprotein (HDL), low-density lipoprotein (LDL), total protein (TP), albumin (ALB), triglyceride (TG), total cholesterol (TCHOL), blood urea nitrogen serum (BUN), creatinine (Cr), uric acid (UA), 24-h urinary protein (24h Upro), protein-to-creatinine ratio (PCR), albumin-to-creatinine ratio (ACR) and serum anti-PLA2R antibody. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula [ 15 ]. Pathological data for all patients were derived from histopathological reports.

Study outcomes

All patients were followed through the outpatient system or telephone contact from the time of renal biopsy until the occurrence of an endpoint event, death, loss to follow-up, or October 31, 2024, whichever occurred first. Two primary endpoints were evaluated: renal function deterioration and renal complete remission. Renal function deterioration was defined as a composite of end-stage renal disease (ESRD; eGFR <15 mL/min/1.73 m 2 ; renal replacement therapy >3 months) or death. In addition, the complete remission (CR), defined as the achievement of urinary protein <0.3 g/24 h, and stable renal function from the time of renal biopsy [ 16 ], which does not reflect complete remission of DN but based on the KDIGO guidelines’ definition of CR for MN. The median follow-up duration was 26.0 months (Interquartile Range [IQR]: 17.5–40.0 months).

Statistical analyses

The Shapiro-Wilk test was used to assess normality of parameter distributions. Normally distributed continuous variables were expressed as means ± standard deviation (SD), while non-normally distributed variables were expressed as medians with interquartile ranges. Categorical variables were expressed as absolute counts and percentages. The Kruskal-Wallis, ANOVA, Wilcoxon rank-sum test or t-test was used for continuous variable analysis, and the χ 2 test was used for categorical variables. Correlations between pathological features and clinical characteristics were analyzed using Spearman’s correlation and visualized with heatmaps. The Kaplan-Meier method was used to calculate event-free survival for the two composite endpoints (renal survival rate and renal remission rate), with the log-rank test or Gehan-Breslow-Wilcoxon test used to assess differences between survival curves. For pairwise comparisons of survival curves, the Bonferroni correction was applied to the p -values to adjust for multiple comparisons. Multivariate Cox proportional hazard models were employed to identify prognostic factors for renal survival and remission. To select the most relevant prognostic factors for the multivariable Cox proportional hazards model, we employed the Least Absolute Shrinkage and Selection Operator (LASSO) regression method. This technique effectively eliminates multicollinearity issues by applying penalty terms to compress the coefficients of secondary variables to zero, thereby precisely identifying key variables from among candidate predictors. Based on multivariate Cox regression results, statistically significant variables from univariate analysis were selected to construct nomograms for estimating 1-, 3-, and 5-year renal survival probabilities. Receiver Operating Characteristic (ROC) curves and calibration curves were used to assess concordance between actual observations and predicted survival probabilities from the nomogram. Statistical analyses were performed using IBM SPSS for Windows 27.0 and R (x64 4.2.2). All tests were two-sided, with p -values <0.05 considered statistically significant.

Results

Baseline characteristics of patients

Among the 276 patients with biopsy-proven diabetic nephropathy or idiopathic membranous nephropathy, 34 were classified into the DN-MN group, 44 into the DM-MN group, and 198 into the MN group. Baseline eGFR in the study cohort was 68.2 (41.7, 104.8) mL/min/1.73 m 2 , 100.3 (91.4, 112.1) mL/min/1.73 m 2 , and 101.4 (89.2, 113.0) mL/min/1.73 m 2 . Cr, 24-Upro, PCR, and ACR in the DN-MN group were significantly higher than those in the DM-MN and MN groups, indicating impaired baseline renal function in the DN-MN group. ESR and HDL levels also demonstrated significant differences between the DN-MN group and the other groups, whereas no significant differences were observed in the remaining clinical characteristics ( Table 1 ).

Table 1. Baseline clinical features of patients with DN-MN, DM-MN, and MN. Baseline clinical features MN DM-MN DN-MN P-Value n  = 198 n  = 44 n  = 34 Age 49 (40, 58) 54 (47, 61) 53 (45, 63) 0.740 Sex 130/68 23/21 25/9 0.189 SBP (mmHg) 135 (122, 145.3) 137.4 (124, 147.8) 141 (125.3, 156.3) 0.258 DBP (mmHg) 86 (78, 94) 88.6 (81, 96) 87.5 (74.8, 96.3) 0.540 ESR (mm/h) 33.3 (16, 45.3)* 50.6 (19.3, 78) 57.7 (38, 75.3) ‡ <0.001 Hemoglobin (g/L) 133.4 ± 18.5 124.8 ± 18.7 126.5 ± 24.1 0.100 HDL (mmol/L) 1.7 (1.4, 2.0) 1.6 (1.3, 1.9) 1.2 (1.0, 1.5) ‡ <0.001 LDL (mmol/L) 5.1 (3.4, 6.4) 4.9 (3.4, 7.0) 4.1 (3.2, 5.9) 0.758 Total protein (g/L) 47.2 (40.1, 52.9) 44.7 (37.9, 49.8) 44.5 (39.4, 53.4) 0.250 Serum albumin (g/L) 25.0 (19.6, 29.1) 22.0 (17.0, 26.1) 21.6 (17.9, 25.4) ‡ 0.002 Triglycerides (mmol/L) 2.7 (1.4, 3.1) 3.1 (1.8, 3.8) 2.5 (1.6, 3.4) 0.189 Total cholesterol (mmol/L) 7.6 (5.9, 9.0) 7.8 (6.0, 9.7) 5.8 (4.9, 7.3) 0.115 BUN (mmol/L) 6.4 (4.5, 7.4) 6.7 (4.3, 8.8) 6.7 (5.3, 8.7) 0.086 Serum creatinine (μmol/L) 74.6 (61.8, 83.8) 68.0 (51.4, 81.0) † 91.8 (68.6, 153.8) ‡ 0.000 Uric acid (μmol/L) 360.1 (299.1, 428.8) 338.5 (280.2, 400) 352.7 (310.6, 435.3) 0.201 eGFR (ml/min/1.73m 2 ) 101.4 (89.2, 113) 100.3 (91.4, 112.1) † 68.2 (41.7, 104.8) ‡ <0.001 Complement 3 (g/L) 1.0 (0.8, 1.1) 1.0 (0.9, 1.2) 1.0 (0.8, 1.2) 0.296 Complement 4 (g/L) 0.3 (0.2, 0.3) 0.3 (0.2, 0.4) 0.3 (0.2, 0.4) 0.299 24h Upro (mg/24h) 3433.8 (1466.8, 5988.6) 3603.6 (2191.9, 6300.2) † 6332.8 (4083.9, 10898.1) ‡ 0.001 PCR (mg/g) 3414.6 (702.2, 4970.2)* 4953.6 (1788.3, 7069.8) † 6408.0 (3709.7, 9354.8) ‡ 0.000 ACR (mg/g) 1859.2 (273.99, 2737.7) 2522.2 (115.5, 3299.7) † 3140.2 (586.9, 5300.7) ‡ 0.024 PLA2R (RU/ml) 10.65 (0.94, 62.96) 18.58 (2.44, 71.55) 19.22 (9.34, 23.07) 0.496 Continuous variables are presented as mean ± standard deviation for normally distributed data and as median (interquartile range) for non‑normally distributed data. Categorical variables are presented as counts and percentages. Group comparisons for continuous variables were performed using the Kruskal-Wallis, ANOVA, Student’s t -test or Wilcoxon rank‑sum test, and for categorical variables using the chi‑square test. p  < 0.05 between MN and DM‑MN; † p  < 0.05 between DM‑MN and DN‑MN; ‡ p  < 0.05 between MN and DN‑MN. DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; MN: membranous nephropathy; SBP: systolic blood pressure; DBP: diastolic blood pressure; ESR: erythrocyte sedimentation rate; HDL: high-density lipoprotein cholesterol; LDL: low-density lipoprotein cholesterol; BUN: blood urea nitrogen; eGFR: estimated Glomerular Filtration Rate; 24h Upro: 24-h urinary protein; PCR: protein-to-creatinine ratio; ACR: albumin-to-creatinine ratio.

Baseline clinical features of patients with DN-MN, DM-MN, and MN.

Continuous variables are presented as mean ± standard deviation for normally distributed data and as median (interquartile range) for non‑normally distributed data. Categorical variables are presented as counts and percentages. Group comparisons for continuous variables were performed using the Kruskal-Wallis, ANOVA, Student’s t -test or Wilcoxon rank‑sum test, and for categorical variables using the chi‑square test. p  < 0.05 between MN and DM‑MN; † p  < 0.05 between DM‑MN and DN‑MN; ‡ p  < 0.05 between MN and DN‑MN.

DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; MN: membranous nephropathy; SBP: systolic blood pressure; DBP: diastolic blood pressure; ESR: erythrocyte sedimentation rate; HDL: high-density lipoprotein cholesterol; LDL: low-density lipoprotein cholesterol; BUN: blood urea nitrogen; eGFR: estimated Glomerular Filtration Rate; 24h Upro: 24-h urinary protein; PCR: protein-to-creatinine ratio; ACR: albumin-to-creatinine ratio.

Pathological features of patients

Among the 34 patients in the DN-MN group, 20 (58.8%) had segmental sclerosis, 21 (61.8%) demonstrated interstitial fibrosis and tubular atrophy (IFTA), and 25 (73.5%) had moderate to severe segmental mesangial proliferation. In addition, glomerulosclerosis was found in 25 patients (73.5%), glomerular crescents in 8 patients (23.5%), and renal arteriolar hyalinosis in 21 patients (61.8%). These values were statistically different from those in the DM-MN and MN groups ( p  < 0.05), indicating that the renal lesion in the DN-MN group presented more serious renal pathological features compared with the DM-MN and MN groups ( Table 2 ). Representative renal pathological features for patients in the MN, DM-MN and DN-MN groups are shown in Figure 1(B) . In DN-MN patients, notable renal pathological features included Kimmelstiel-Wilson (K-W) nodules, warty dilatation, and extensive IFTA in addition to spike formation. In DM-MN patients, prominent spike formation with chain-like structures was observed. In MN patients, pathological features encompassed thickening of the glomerular basement membrane (GBM), spike formation, and IFTA. Electron microscopy reveals typical subepithelial deposits in the MN patients ( Figure 1(C) ). The DN-MN patient exhibits a combination of these deposits superimposed on a markedly and diffusely thickened glomerular basement membrane ( Figure 1(D) ). Immunofluorescence staining for IgG4 and PLA2R in consecutive glomerular sections from the same patient is provided in Supplementary Figure 1(A) and (B) .

Table 2. Pathological feature scores of patients with DN-MN, DM-MN, and MN. Pathological features Score MN DM-MN DN-MN p -value n  = 198 n  = 44 n  = 34 Glomerulosclerosis 0 70 (35.4) 20 (45.5) 9 (26.5) 0.214   1 128 (64.6) 24 (54.5) 25 (73.5)   IFTA 0 99 (50.0) 25 (56.8) 13 (38.2) 0.044   1 99 (50.0) 19 (43.2) 21 (61.8)   Inflammation 0 174 (87.9) 41 (93.2) 29 (85.3) 0.416   1 24 (12.1) 3 (6.8) 5 (14.7)   Segmental sclerosis 0 120 (60.6) 17 (38.6) 14 (41.2) 0.007 1 78 (39.4) 27 (61.4) 20 (58.8)   Peri-glomerular fibrosis 0 160 (80.8) 34 (77.3) 19 (55.9) 0.041 1 38 (19.2) 9 (20.5) 14 (41.2)   2 0 1 (2.3) 1 (2.9)   Crescent 0 190 (96.0) 42 (95.5) 26 (76.5) 0.018   1 8 (4.0) 2 (4.5) 8 (23.5)   Mesangial proliferation 1 155 (78.3) 28 (63.6) 9 (26.5) 0.001   2 43 (21.7) 16 (36.4) 25 (73.5)   Hyalinosis 0 175 (88.4) 40 (90.9) 13 (38.2) 0.001   1 23 (11.6) 4 (9.1) 21 (61.8)   IgG deposition 0 2 (1.0) 2 (4.5) 0 0.001   1 5 (2.5) 1 (2.3) 7 (20.6)     2 16 (8.1) 3 (6.8) 3 (8.8)     3 122 (61.6) 22 (50.0) 18 (52.9)     4 53 (26.8) 16 (36.4) 6 (17.6)   IgA deposition 0 135 (68.2) 32 (72.7) 17 (50.0) 0.001   1 46 (23.2) 5 (11.4) 12 (35.3)     2 17 (8.6) 4 (9.1) 5 (14.7)     3 0 3 (6.8) 0   IgM deposition 0 93 (47.0) 25 (56.8) 17 (50.0) 0.956   1 52 (26.3) 11 (25.0) 10 (29.4)     2 47 (23.7) 7 (15.9) 6 (17.6)     3 5 (2.5) 1 (2.3) 1 (2.9)     4 1 (0.5) 0 0   C3 deposition 0 55 (27.8) 16 (36.4) 10 (29.4) 0.843   1 47 (23.7) 8 (18.2) 7 (20.6)     2 64 (32.3) 11 (25.0) 10 (29.4)     3 32 (16.2) 9 (20.5) 7 (20.6)   C1q deposition 0 85 (42.9) 20 (45.5) 17 (50.0) 0.776   1 72 (36.4) 14 (31.8) 11 (32.4)     2 34 (17.2) 6 (13.6) 4 (11.8)     3 6 (3.0) 3 (6.8) 2 (5.9)     4 1 (0.5) 1 (2.3) 0   Fib deposition 0 163 (82.3) 39 (88.6) 22 (64.7) 0.002   1 5 (2.5) 2 (4.5) 4 (11.8)     2 22 (11.1) 2 (4.5) 2 (5.9)     3 8 (4.0) 1 (2.3) 6 (17.6)   Kappa chain 0 3 (1.5) 6 (13.6) 3 (8.8) 0.001   1 8 (4.0) 1 (2.3) 6 (17.6)     2 34 (17.2) 4 (9.1) 6 (17.6)     3 135 (68.2) 27 (61.4) 16 (47.1)     4 17 (8.6) 4 (9.1) 2 (5.9)   Lambda chain 0 4 (2.0) 6 (13.6) 3 (8.8) 0.002   1 17 (8.6) 1 (2.3) 7 (20.6)     2 52 (26.3) 6 (13.6) 6 (17.6)     3 109 (55.1) 25 (56.8) 16 (47.1)     4 15 (7.6) 4 (9.1) 1 (2.9)   PLA2R deposition 0 26 (13.1) 13 (29.5) 7 (20.6) 0.007   1 17 (8.6) 3 (6.8) 2 (5.9)     2 48 (24.2) 13 (27.3) 7 (20.6)     3 51 (25.8) 11 (25.0) 3 (8.8)     4 1 (0.5) 1 (2.3) 0   Podocytosis 1 29 (18.6) 14 (46.7) 8 (36.4) 0.002   2 127 (81.4) 16 (53.3) 14 (63.6)   Comparisons were based on the Kruskal-Wallis, ANOVA, Wilcoxon rank-sum test, chi-square test or Fisher’s exact test. Mesangial proliferation score: Grade 1 indicates mild to moderate mesangial proliferation; grade 2 indicates moderate to severe mesangial proliferation. Peri-glomerular fibrosis score: 0 = absence of peri-glomerular fibrosis; 1 ≤ 50%; 2 > 50%. For other binary variables, a score of 0 indicates the absence of the pathological feature, and 1 indicates its presence. Data for some pathological features were not available for all patients due to limitations in archival tissue sections or staining quality. Percentages were calculated based on the number of patients for whom data were available for that specific feature. DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; MN: membranous nephropathy; IFTA: interstitial fibrosis and tubular atrophy; IgG: immunoglobulin G; IgA: immunoglobulin A; IgM: immunoglobulin M; C3: complement 3; C1q: complement component 1q; Fib: fibrinogen.

Pathological feature scores of patients with DN-MN, DM-MN, and MN.

Comparisons were based on the Kruskal-Wallis, ANOVA, Wilcoxon rank-sum test, chi-square test or Fisher’s exact test. Mesangial proliferation score: Grade 1 indicates mild to moderate mesangial proliferation; grade 2 indicates moderate to severe mesangial proliferation. Peri-glomerular fibrosis score: 0 = absence of peri-glomerular fibrosis; 1 ≤ 50%; 2 > 50%. For other binary variables, a score of 0 indicates the absence of the pathological feature, and 1 indicates its presence.

Data for some pathological features were not available for all patients due to limitations in archival tissue sections or staining quality. Percentages were calculated based on the number of patients for whom data were available for that specific feature.

DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; MN: membranous nephropathy; IFTA: interstitial fibrosis and tubular atrophy; IgG: immunoglobulin G; IgA: immunoglobulin A; IgM: immunoglobulin M; C3: complement 3; C1q: complement component 1q; Fib: fibrinogen.

Association between segmental sclerosis, IFTA and clinicopathological features

The correlations between segmental sclerosis, IFTA and clinicopathological features were analyzed by Spearman’s correlation. In the DN-MN group, segmental sclerosis was correlated with Cr levels (Spearman’s ρ  = 0.4203, p  = 0.0133), and IFTA was significantly correlated with Cr levels (Spearman’s ρ  = 0.3856, p  = 0.0243). In addition, segmental sclerosis was correlated with Cr levels (Spearman’s ρ  = 0.3783, p  = 0.0478) in the DM-MN group ( Supplementary Table 1 ). As shown in Figure 1(E–F) , in the MN group, glomerular segmental sclerosis was correlated with Cr levels (Spearman’s ρ  = 0.2432, p  = 0.0006), 24-h proteinuria (Spearman’s ρ  = 0.2184, p  = 0.0020) and eGFR (Spearman’s ρ = −0.1948, p  = 0.0060). IFTA was correlated with Cr levels (Spearman’s ρ  = 0.3186, p  < 0.01), 24-h proteinuria (Spearman’s ρ  = 0.2146, p  = 0.0024), eGFR (Spearman’s ρ = −0.2545, p  = 0.0003), and ACR (Spearman’s ρ  = 0.2313, p  = 0.0010). Other pathological features in the glomeruli of patients in the MN group showed no significant correlation with clinical renal function ( p  > 0.05).

Kaplan–Meier survival analysis

The median follow-up duration for the 276 patients included in the prognostic analysis was 26 months (IQR: 17.5–40.0 months). When renal function deterioration was defined as the endpoint event, 5 patients (14.7%) in the DN-MN group, 2 patients (5.0%) in the DM-MN group, and 4 patients (2.0%) in the MN group reached the endpoint. When renal complete remission was defined as the endpoint event, 7 patients (20.6%) in the DN-MN group, 21 patients (47.7%) in the DM-MN group, and 94 patients (47.5%) in the MN group reached the endpoint. The Kaplan-Meier curves demonstrated that renal function deterioration as the endpoint event, patients in the DN-MN group had significantly worse renal prognosis compared to those in the DM-MN and MN groups ( p  < 0.05) ( Figure 2(A) ) The Kaplan–Meier curves for renal complete remission diverged early in the follow-up period, indicating a difference in early event rates. To formally test this observation, we employed the Gehan-Breslow-Wilcoxon test, which gives more weight to early events, and it confirmed that this early separation was highly statistically significant ( p  < 0.05) ( Figure 2(B and C) ). Additionally, segmental glomerulosclerosis, IFTA, and segmental mesangial proliferation were significantly associated with complete renal remission rates in patients with membranous nephropathy. Furthermore, IFTA was correlated with complete remission rates in patients with DN-MN and DM-MN (all p  < 0.05). ( Supplementary Fig. 2(A–I) ). These findings indicate that patients in the DN-MN group had a poorer renal prognosis compared to patients in the DM-MN and MN groups.

Kaplan–Meier survival analysis in patients with DN-MN, DM-MN, and MN. Kaplan–Meier survival analysis was conducted to compare the time to reach outcome measures among patients with DN-MN, DM-MN, and MN. (A) The outcome measure is ESRD or death. p -values for pairwise comparisons were adjusted using the Bonferroni method. (B) The outcome indicator is renal complete remission (24-hour urinary protein < 0.3 g/d). The curves for the three groups showed significant separation, particularly during the early follow-up period. This early divergence was confirmed to be statistically significant by the Gehan–Breslow–Wilcoxon test ( p  < 0.05). The shaded areas represent the 95% confidence intervals. The numbers at risk at different time points are indicated. (C) The number at risk table is provided to indicate the number of patients remaining in the analysis at specific time intervals. DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; MN: membranous nephropathy.

To identify predictors of complete remission in patients with diabetes-associated MN (DM-MN and DN-MN, n  = 78), we used LASSO regression to select key variables for a multivariable Cox model ( Supplementary Fig. 1(C–F) ). Among them, IFTA, eGFR, and segmental sclerosis were selected by the LASSO model for multivariable Cox regression. The analysis revealed that a greater degree of IFTA was a significant independent negative predictor of remission (HR: 0.3650; 95% CI: 0.1606–0.8297). While baseline eGFR (HR: 1.0214; 95% CI: 1.0074–1.0459) and segmental sclerosis (HR: 0.4549; 95% CI: 0.2026–1.0213) were also selected by LASSO, segmental sclerosis was not statistically significant in the final multivariable model ( Table 3 ). For comparison, a parallel analysis in the primary MN cohort identified a different set of predictors, including eGFR, segmental mesangial proliferation, glomerulosclerosis, male sex, and PCR, with full results detailed in Supplementary Table 3 .

Table 3. Severe IFTA was a significant independent negative predictor of remission in DN-MN and DM-MN patients. Clinicopathological features Multivariate models   HR (95% CI) p -value eGFR 1.0214 (1.0074–1.0459) 0.045 Segmental sclerosis 0.4549 (0.2026–1.0213) 0.076 IFTA 0.3650 (0.1606–0.8297) 0.020 The final multivariable model was constructed using variables selected by the LASSO Cox regression analysis. HR < 1 indicates a lower probability of achieving complete remission. Reference groups for categorical variables were Grade 0. DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; HR: hazard ratio; CI: confidence interval; eGFR: estimated glomerular filtration rate; IFTA: interstitial fibrosis and tubular atrophy.

Severe IFTA was a significant independent negative predictor of remission in DN-MN and DM-MN patients.

The final multivariable model was constructed using variables selected by the LASSO Cox regression analysis. HR < 1 indicates a lower probability of achieving complete remission. Reference groups for categorical variables were Grade 0.

DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; HR: hazard ratio; CI: confidence interval; eGFR: estimated glomerular filtration rate; IFTA: interstitial fibrosis and tubular atrophy.

Nomogram for predicting renal complete remission probability at 1, 3, and 5 years

A nomogram was constructed based on the LASSO regression model results, incorporating predictors of renal complete remission (baseline eGFR, the presence of segmental sclerosis, and the degree of IFTA). The nomogram provided a simplified estimation method for determining 1-, 3-, and 5-year renal complete remission probabilities for patients with DN-MN and DM-MN ( Figure 3(A) ). For example, a DM-MN patient with eGFR = 112.19 mL/min/1.73 m 2 , with segmental sclerosis and IFTA lesions, would have a total score of 150, corresponding to 1-, 3-, and 5-year renal non-remission rates of approximately 99.3%, 82.9%, and 55.6%, respectively. The receiver operating characteristic (ROC) curves and calibration curves demonstrated good agreement between the nomogram-predicted probability of renal complete remission at 1-, 3-, and 5-year and the actual observations ( Figure 3(B and C) ). The 1-, 3-, and 5-year prediction curves for the MN group were equally informative ( Supplementary Fig. 2(J–L) ).

Nomogram for predicting complete remission in DN-MN and DM-MN patients. (A) Prognostic nomogram to predict the probability of non-remission (24-hour urinary protein < 0.3 g/day) at 1-, 3-, and 5-year in patients with DN-MN and DM-MN based on eGFR, degree of segmental sclerosis, and IFTA. (B) The ROC curve and AUC value. (C) Calibration curve. DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy; ROC: receiver operating characteristic; AUC: area under ROC curve; eGFR: estimated Glomerular Filtration Rate; IFTA: interstitial fibrosis and tubular atrophy.

Treatment response in patients

Patients with DM-MN or DN-MN were categorized into non-glucocorticoid and glucocorticoid groups based on whether they received glucocorticoid therapy within two years after diagnosis. There was no significant difference in the number of patients receiving sodium-glucose cotransporter 2 inhibitor (SGLT2i) therapy between the two groups. Angiotensin-converting enzyme inhibitor (ACEI) and angiotensin II receptor blocker (ARB) therapies were administered as clinically indicated based on individual patient requirements. The treatment approaches for the DN-MN and DM-MN groups were statistically analyzed ( Figure 4(A) ). In this study, 13 DM-MN patients were assigned to the non-glucocorticoid group and 31 to the glucocorticoid group, while 18 DN-MN patients were assigned to the non-glucocorticoid group and 16 to the glucocorticoid group. Supplementary Tables 4–5 demonstrate that there were no significant baseline differences between the two groups. To evaluate the effects of different treatments on DM-MN and DN-MN patients, Kaplan-Meier analysis was performed to compare renal complete remission outcomes ( Figure 4(B–C) ). No significant differences were observed in the impact of glucocorticoid use on achieving renal complete remission in DM-MN and DN-MN patients ( p  > 0.05).

Kaplan–Meier survival analysis of DN-MN and DM-MN patients on treatment strategies. (A) Different treatment strategies accounted for the proportion of patients in each group. (B–C) Kaplan–Meier survival analysis comparing the time to reach the outcome indicators in relation to medication in DN-MN and DM-MN patients. (B) DM-MN group. (C) DN-MN group. The outcome indicator was renal complete remission (24-hour urinary protein < 0.3 g/d). The non-glucocorticoid and glucocorticoid therapy were compared to assess the time to reach the renal complete remission outcome for the different treatment strategies within the two groups of patients. The shaded areas represent 95% confidence intervals. DN-MN: diabetic nephropathy combined with membranous nephropathy; DM-MN: diabetes mellitus combined with membranous nephropathy.

Discussion

This cohort study involving 276 biopsy-confirmed patients with DN-MN, DM-MN, and MN found that DN-MN patients exhibit unique clinical and pathological features and have a poorer prognosis. The study also revealed that the degree of IFTA was an independent pathological prognostic factor for unfavorable outcomes in DN-MN patients, whereas the use of glucocorticoids did not affect the complete remission rates in either DN-MN or DM-MN patients.

In diabetic patients with short disease duration who suddenly develop renal insufficiency or active hematuria, this generally indicates concomitant non-diabetic renal disease (NDRD) [ 17 ]. With the increasing prevalence of diabetes, both NDRD and diabetic nephropathy coexisting with NDRD require further investigation. C E Mogensen et al. found that about 10% of biopsied diabetic patients with proteinuria have a component of NDRD [ 18 ]. Singh et al. reported that NDRD accounts for 25.8% (17/66) of type 2 diabetes patients, while diabetic nephropathy combined with NDRD accounts for 19.7% (13/66) [ 19 ]. MN is the most common cause of nephrotic syndrome in adults [ 20 ], with a significant increase in cases concurrent with DM or DN in recent years. Among DM patients with concurrent NDRD, MN is more common [ 21 ]. Currently, the reported prevalence rates of DM-MN and DN-MN vary across the literature. The variability in prevalence rates may be attributed to differences in biopsy indications between centers. In the United States, epidemiologic data indicated that MN accounts for 8.2% (18/220) of diabetic patients with NDRD [ 22 , 23 ], while Chinese data reported a 39.3% (323/822) prevalence of DM-MN patients [ 24 ]. A single-center study in India found the prevalence of concurrent MN in patients with type 2 diabetes to be 8.6% (6/70), with DN-MN accounting for 1.4% (1/70) [ 25 ]. In this retrospective study, DN-MN patients accounted for 8% (34/425) and DM-MN patients for 10.4% (44/425) of cases. When DM patients present with massive proteinuria and a clinical distinction cannot be easily made between DN and NDRD (such as MN), renal biopsy is recommended to clarify the pathological diagnosis and guide treatment.

DM has been identified as an independent risk factor for worsening renal function in patients with membranous nephropathy, though it does not affect remission of proteinuria [ 26 ]. Xie et al. reported that patients with diabetes mellitus combined with membranous nephropathy tend to exhibit more complex clinical manifestations, such as advanced age, reduced glomerular filtration rate, and more severe anemia. Furthermore, DM was associated with a failure to achieve complete remission in MN patients [ 27 ]. However, to date, few reports in the literature have compared the clinical features between DM-MN and DN-MN. In this study, we found that DN-MN patients, compared to other groups, exhibited lower ALB and eGFR, higher Cr and proteinuria, decreased renal survival rates, and poorer complete remission rates. Additionally, the DM-MN group showed elevated erythrocyte sedimentation rate and protein-to-creatinine ratio compared to the membranous nephropathy group, with no significant difference in renal function.

Significantly, after adjusting for key baseline confounders (including eGFR, serum creatinine, and proteinuria) in a multivariable Cox proportional hazards model, our study demonstrated that DN (HR: 0.991; 95% CI: 0.983–0.999), rather than DM, was an independent risk factor for renal remission in patients with MN ( Supplementary Table 2 ). While the hazard ratio in the multivariable model suggested a modest effect size, likely due to adjustment for highly correlated covariates (e.g. massive proteinuria), the independent association remained statistically significant. To further validate this finding and rigorously control for baseline imbalances, we conducted a sensitivity analysis using propensity score matching (PSM) by balancing age, eGFR, and proteinuria. The PSM analysis revealed a much more pronounced impact, confirming that the DN-MN group had a significantly lower likelihood of remission (HR = 0.3, p  = 0.045). Therefore, the unfavorable prognosis observed in the DN-MN group cannot be attributed solely to the higher baseline level of proteinuria. Rather, it is more likely the result of synergistic renal injury caused by diabetes-related pathological changes superimposed on the underlying MN.

Limited studies exist on the pathological characteristics and prognosis of patients with DN-MN. Here, we found that the IFTA pathological feature was associated with the rate of complete remission in the DN-MN group. These data suggest that DN-MN is a disease with a distinct clinicopathological phenotype. Additionally, the research on the impact of the clinicopathological characteristics of DN-MN and DM-MN on disease prognosis was limited. The nomogram model is valuable in clinical medical research, as it integrates multiple independent prognostic factors, such as pathological features, biochemical indicators, and clinical status, displays the weighted contribution of each factor to prognosis and rapidly calculates individual patients’ probability of outcomes at specific time points (e.g. 1-, 3-, and 5-year). We have previously used nomograms to evaluate DN in type 2 diabetic patients [ 28 ]. In this study, we constructed a nomogram model to predict complete remission in DN-MN and DM-MN groups using clinical manifestations and pathological changes. Although we constructed a nomogram to predict complete remission, it requires further clinical testing.

Recent studies have shown that the coexistence of DM and DN presents significant challenges in the diagnosis and treatment of MN patients [ 29 , 30 ]. Currently no standardized treatment guidelines exist for DN-MN patients. While conventional MN treatment approaches are typically applied to diabetic patients with membranous nephropathy alongside standard glycemic control, limited data exist regarding treatment efficacy and prognosis in DN-MN patients. Previous studies have reported positive effects with low-dose glucocorticoids combined with tacrolimus and mycophenolate mofetil for DM-MN management [ 31 ]. Shang et al. found that leflunomide (LEF) combined with low-dose glucocorticoid methylprednisolone was more effective for DN-MN patients than LEF monotherapy [ 32 ]. Here, we found that glucocorticoid use did not significantly influence the prognosis of DN-MN patients, which may suggest that the benefit-risk ratio of conventional therapy needs careful assessment in DN-MN management. The reason may be attributed to study design, sample size, population differences, and other factors including the dosage and duration of treatment with glucocorticoids or other immunosuppressants. In addition, while glucocorticoids may regulate immune dysregulation and inflammatory responses in MN patients, their effects on other etiologies, such as DN conditions, hemodynamic abnormalities and metabolic disturbances, the effect of glucocorticoids or immunosuppressants may be limited [ 33 ]. Future prospective multicenter studies are needed to evaluate the role of glucocorticoids and immunosuppressive agents on clinical efficacy and long-term prognosis of DN-MN.

This study has several limitations. First, due to its retrospective design, we were unable to obtain continuous, longitudinal follow-up data on proteinuria (PCR/ACR), thus precluding the dynamic assessment of proteinuria progression under different treatment regimens. This limitation primarily resulted from the study’s extended time span, inconsistent data recording, and patient attrition. Consequently, our assessment of treatment response relied primarily on clinical endpoints such as complete renal remission. Second, it is worth noting that although baseline proteinuria is a classical risk factor, it was not selected in our final LASSO model despite the inclusion of 26 candidate variables (including ACR). This exclusion is likely attributable to multicollinearity with other potent predictors such as eGFR and IFTA. As LASSO regression favors sparsity to prevent overfitting, it prioritized variables with the strongest independent predictive value in this specific cohort, which warrants further validation in larger samples. Finally, as a single-center retrospective cohort study, potential selection bias may exist, and loss to follow-up with differing observation periods may impact outcome assessment. Future prospective cohort studies should systematically collect dynamic metrics to more precisely characterize treatment response patterns in DN-MN patients.

In summary, this study demonstrates that DN-MN patients present with more severe clinical manifestations, worse renal function, greater renal pathological injury, and poorer prognosis, warranting particular clinical attention. The nomogram model may aid in evaluating the probability of disease remission at 1-, 3-, and 5-year in DN-MN and DM-MN patient populations. Glucocorticoid therapy did not significantly affect the complete renal remission rate in DN-MN and DM-MN patients, which may help to inform future investigations.

Supplementary Material

Supplemental Material

Supplemental Material

Supplemental Material