Idiopathic nodular glomerulosclerosis and diabetic nephropathy: a histological and prognostic comparison highlighting the impact of smoking

Abstract

Introduction

Idiopathic nodular glomerulosclerosis (ING) and diabetic nephropathy (DN) display (almost) identical kidney biopsy findings. Our aim was to investigate potential disparities in histology and clinical outcomes, stratified by smoking status, in a Belgian cohort.

Methods

This retrospective case-control study included 64 patients with DN and 11 with ING. Eleven DN patients were matched with 11 ING patients for histological analysis based on sex, age, and histological lesions. Clinical and histological characteristics were compared using parametric and nonparametric methods. Estimated glomerular filtration rate (eGFR) decline and overall survival were assessed using linear mixed modelling with random intercepts and Kaplan-Meier survival analysis, respectively, stratified by smoking.

Results

Light microscopy revealed a trend toward more ischemic collapse in ING versus DN (36.4% vs. 0%, p = 0.090). Aberrant vessels were distributed differently between groups, being present in all ING cases but absent in 27.3% of DN cases (p = 0.045). Overall survival of smoking ING patients was lower compared with smoking DN patients (p = 0.010). The rate of eGFR decline was comparable between DN and ING (-5.41 versus − 6.27 ml/min/1.73 m² per year, p = 0.322). However, active smokers overall had a significantly faster decline in eGFR (p < 0.001).

Conclusions

Histology analysis confirmed that ING and DN share similar features, however subtle differences exist, with distribution of aberrant vessels being different in ING. Clinically, active smokers had worse outcomes in ING, despite no difference in GFR decline between DN and ING groups.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12882-026-04889-y.

Introduction

Idiopathic nodular glomerulosclerosis (ING) is a rare glomerular disease characterized by mesangial matrix expansion and frequently by nodular glomerulosclerosis with Kimmelstiel–Wilson–like nodules. Histologically, ING is indistinguishable from diabetic nephropathy (DN), making clinicopathological correlation essential for diagnosis [1, 2]. In addition to diabetes, other causes of nodular glomerulosclerosis—such as monoclonal immunoglobulin deposition disease (MIDD) and amyloidosis—must be excluded [1–4].

Only limited case reports and small case–control studies have described ING. Reported histological features include diffuse mesangial expansion, nodular glomerulosclerosis, interstitial fibrosis and tubular atrophy (IFTA), and arteriolar hyalinosis or sclerosis, all of which overlap extensively with DN [1–9]. Several mechanisms have been proposed to explain ING, including metabolic disturbances [3, 6, 10, 11], and smoking-related injury [12]; however, these hypotheses remain largely speculative.

Diabetic nephropathy is a common complication of diabetes, affecting approximately 20–30% of patients, and is classified histologically into four classes based on glomerular basement membrane thickening and mesangial and sclerotic lesions [13, 14]. Despite their histological overlap, whether ING and DN represent distinct clinicopathological entities with different prognostic trajectories remains uncertain.

Given the rarity of ING and the limited comparative data, we conducted a retrospective case–control study comparing patients with ING and DN. The primary objective was to identify detectable histological differences between ING and DN at the time of kidney biopsy. The secondary objective was to assess whether clinical outcomes, including kidney function decline and survival, differed between the two entities during follow-up. Based on prior observations suggesting a role for smoking in ING, all analyses were stratified by smoking status, which was examined as a potential effect modifier of both histological findings and clinical outcomes.

Methods

Patient population

This retrospective case-control study was approved by the Ethics Committee of Ghent University Hospital (UZG) under reference numbers ONZ-2022-0305 and THE-2023-0197 and was performed in full accordance with the Declaration of Helsinki. All native kidney biopsy specimens (n = 4553) registered in the Flemish FCGG-NBVN registry from January 2017 to March 2023 were screened for the presence of the primary diagnoses “idiopathic nodular glomerulosclerosis” and “diabetic nephropathy”. Kidney biopsy specimens available in UZG from October 2022 to March 2023 were included when a primary diagnosis of ING or DN was reported. The FCGG registry provides comprehensive population data on kidney disease in Flanders, covering a region of 6.5 million inhabitants [17, 18].

The diagnoses of ING and DN were confirmed by two experienced (nephro)pathologists (AD and PK). Cases with less pronounced nodular glomerulosclerosis or with predominant diffuse mesangial matrix increase were retained in the ING group provided that other diagnoses or patterns such as focal segmental or global glomerulosclerosis, amyloidosis, MIDD, glomerulonephritis and non-specific arterionephrosclerosis patterns were clearly excluded. Native biopsies containing fewer than 10 glomeruli were considered non-representative [19, 20] and excluded. Two patients initially classified as ING were reclassified as DN after chart review. The pattern of involvement - diffuse or nodular - in both ING and DN was definedaccording to the 2010 Renal Pathology Society (RPS) classification of DN [14, 21]. Class II was defined as diffuse mesangial expansion, whereas class III was defined as nodular glomerulosclerosis. For clarity, these groups are referred to as follows: ING nodular (nodular glomerulosclerosis in ING patients), ING diffuse (diffuse mesangial expansion in ING patients), DN nodular (nodular glomerulosclerosis in DN patients) and DN diffuse (diffuse mesangial expansion in DN patients).

DN patients classified as severity class I were excluded, as ING cannot be diagnosed on LM when only GBM thickening is present. Biopsy reports were screened for immunofluorescence results including IgG, IgA, IgM, C3c, C1q, kappa or lambda staining. Details of the selection process for the different cohorts are provided in Supplementary Fig. S1.

For clinical analysis, patients were included after applying the aforementioned exclusion criteria. (Supplementary Fig. S1) For histological analyses, each ING patient was matched to a DN patient based on sex, age (± 5 years), and type of mesangial expansion (diffuse versus nodular) (Supplementary Fig. S1), using the case-control matching algorithm in IBM SPSS Statistics for Windows, version 29 (IBM Corp., Armonk, N.Y., USA). In total, 11 DN patients could be matched with 11 ING patients. Eventually, the matched cohort included 8 patients with nodular sclerosis and 3 with diffuse mesangial expansion were retained in each group.

In the matched DN group, the median age was 72.1 years, 81.8% were male and 72.7% had nodular lesions. In the matched ING group, median age was 75.8 years, 81.8% were male, and 45.5% of patients had nodular lesions (P = 0.387).

Clinical parameters

Clinical data, including demographics, clinical characteristics and laboratory values were obtained from the FCGG-NBVN registry and from the electronic health records (EHR) of UZ Ghent patients. Sex was recorded as documented in the EHR, which generally corresponds to sex assigned at birth. Follow-up data were collected until five years after biopsy. For recording smoker status, active smokers were defined as those reporting current smoking at the time of biopsy, while non-active smokers included former and never smokers. Detailed information on smoking intensity or duration was not consistently available.

Histological characteristics

Biopsy slides from the matched ING and DN subset were scanned with a digital slide scanner (Hamamatsu Nanozoomer 2.ORS) at 40x magnification. All biopsies were systematically reviewed by both the researcher (KB) and an experienced nephropathologist (AD) using a scoring form for assessment of histologic characteristics (Supplementary Table S1). The scoring was performed blinded to the clinical diagnosis. Any discrepancies between KB and AD were resolved through joint review and discussion. Well-defined definitions for these characteristics were used, based on consensus definitions in meeting reports [14, 21–24] or definitions in case reports for uncommon histological features [25–31].

Light microscopy

For LM, Periodic Acid-Schiff (PAS) staining was used for assessment of DN and ING kidney biopsies. Interstitial hemorrhage, glomerular capillary thrombosis, GBM and mesangiolysis were additionally evaluated with silver staining. Hematoxylin and eosin (HE) staining was used to evaluate the nature of interstitial inflammation when present.

The definitions for histological evaluation are summarized in Supplementary Table S1.

Diffuse mesangial expansion was defined as an increase of mesangial extracellular material with an interspace width exceeding that of two mesangial cell nuclei, in one or more peripheral mesangial areas [14, 22]. Mesangial hypercellularity was assessed in the most severely affected mesangial area [22, 25]. Other histological features such as glomerular focal segmental glomerulosclerosis (FSGS), acute tubular necrosis (ATN), collapsing lesions, etc. were reported as categorical variables according to their presence or absence. The presence of periglomerular fibrosis and/or fragmentation of Bowman’s capsule was assessed together as “abnormalities at Bowman’s capsule”. All characteristics were evaluated exclusively in non-globally sclerosed glomeruli. IFTA was systematically scored using a visual analogue scale [32]. Aberrant vessels outside the Bowman’s capsule were reported according to their localization: at the glomerular vascular pole, periglomerular, or at both sites [27].

Statistical analyses

Data were reported as median with interquartile range (IQR) or mean ± standard deviation (SD), depending on data distribution. Statistical analyses were performed using IBM SPSS Statistics for Windows, version 29 (IBM Corp., Armonk, N.Y., USA) and graphs were generated with R Studio, version 4.3 [33]. Comparisons between ING and DN groups were made using two-tailed Fisher’s exact test for categorical variables. For continuous variables, a non-parametric Mann-Whitney U test, Students t-test, or analysis of variance (ANOVA) was applied as appropriate. Post-hoc pairwise comparisons were performed using Tukey’s analysis following a significant ANOVA. The same analyses were applied to the matched subset. Kaplan-Meier survival analysis was used to assess the composite outcome (need for dialysis, kidney transplantation, death and/or doubling of serum creatinine). The log-rank test was applied for between-group comparisons. For longitudinal data, a linear mixed model (LMM) with random intercepts was used. A two-sided p-value of < 0.05 was considered statistically significant.

Results

Incidence of ING and DN

Of the 4553 patients who underwent a kidney biopsy between January 2017 and March 2023 included in the FCGG-NBVN registry, 293 had a primary diagnosis of DN and 13 patients were diagnosed with ING. After detailed chart review, only 11 patients of the ING cases had no evidence of diabetes at the time of biopsy or during subsequent follow-up. The overall incidence of DN in the FCGG-NBVN biopsy cohort was 10.3 per 1000 person-years, whereas the incidence of ING was 0.4 per 1000 person-years.

Demographics and clinical differences

DN vs. ING

In the unmatched cohort, ING patients were older than DN patients, with a median age of 75.8 years (IQR 18.3) compared to 66.9 years (IQR 20.4) in the DN cohort (p = 0.151). Both groups were predominantly male (ING 81.8, DN 70.3%, p = 0.717). A history of smoking at the time of biopsy was reported in 75.0% of the ING patients compared to 56.0% of DN patients (p = 0.449). Obesity was present in 42.6% of DN patients versus 12.5% of ING patients (p = 0.136) (Table 1).

Table 1.

Demographics and clinical features compared between ING and DN subjects

	ING (n = 11)	DN (n = 64)	p-value
Gender (M: F)	(9:2)	(45:19)	0.717
Median age, in years (IQR)	75.8 (20.0) (n = 11)	66.9 (20.5) (n = 64)	0.151
Median BMI, in kg/m² (IQR)	25.9 (3.6) (n = 8)	28.4 (8.0) (n = 47)	0.433
Obesity (> 30 kg/m²)	12.5% (n = 1/8)	42.6% (n = 20/47)	0.136
History of smoking • Active smoker	75.0% (n = 6/8) • 50.0% (n = 4/8)	56.0% (n = 28/50) • 24.0% (n = 12/50)	0.449 • 0.198
Intake lipid lowering drugs at time of biopsy	75.0% (n = 6/8)	76.0% (n = 38/50)	1.000
Intake antihypertensive drugs at time of biopsy	100.0% (n = 8/8)	94.1% (n = 48/51)	1.000
Intake RAAS blockers at time of biopsy	62.5% (n = 5/8)	70.0% (n = 35/50)	0.694
Intake VEGF agents at time of biopsy	0.0% (n = 0/8)	2.0% (n = 1/49)	1.000
Intake oral antidiabetic agents at time of biopsy	0.0% (n = 0/8)	53.1% (n = 26/49)	0.006
Insulin therapy at time of biopsy	0.0% (n = 0/8)	51.0% (n = 25/49)	0.007
Active malignancy at time of biopsy	12.5% (n = 1/8)	2.0% (n = 1/49)	0.272
Renal insufficiency at time of biopsy • Acute: chronic	81.8% (n = 9/11) • 5:4	96.8% (n = 61/63) • 21:40	0.103 • 0.194
Median creatinine at time of biopsy, mg/dl (IQR)	1.8 (1.1) (n = 7)	2.2 (1.1) (n = 51)	0.347
Median eGFR at time of biopsy, ml/min/1.73 m² (IQR)	28.1 (41.2) (n = 7)	29.1 (21.6) (n = 51)	0.761
Hematuria at time of biopsy	36.4% (n = 4/11)	28.1% (n = 18/64)	0.721
Proteinuria at time of biopsy • <= 1.0 g/g • 1.0–3.5 g/g • >= 3.5 g/g	100.0% (n = 11/11) • 18.2% (n = 2/11) • 18.2% (n = 2/11) • 63.6% (n = 7/11)	98.4% (n = 63/64) • 15.6% (n = 10/63) • 31.3% (n = 20/63) • 51.6% (n = 33/63)	1.000 • 0.716
Mean serum albumin at time of biopsy, g/L ± SD	32.9 ± 8.2 (n = 11)	34.1 ± 8.2 (n = 64)	0.834
Nephrotic syndrome at time of biopsy	36.4% (n = 4/11)	29.7% (n = 19/64)	0.728
De novo nephrotic syndrome after biopsy, during FU	14.3% (n = 1/7)	8.7% (n = 4/46)	0.522
Dialysis during FU	42.9% (n = 3/7)	36.7% (n = 18/49)	1.000
Transplantation during FU	0.0% (n = 0/7)	4.1% (n = 2/49)	1.000
Death during FU	33.3% (n = 3/9)	16.4% (n = 9/57)	0.347
Vascular events during FU	0.0% (n = 0/6)	14.3% (n = 7/49)	1.000
Hospitalization during FU	42.9% (n = 3/7)	49.0% (n = 24/49)	1.000

Comparison of demographic and clinical features between subjects with primary diagnosis of ING or DN. Median + Interquartile Range (IQR) is reported for continuous non-parametric variables. Mean ± standard deviation (SD) is reported for continuous parametric variables. Conversion factors for creatinine in mg/dl to µmol/l, ×88.4

Abbreviations: BMI = body mass index; eGFR = estimated glomerular filtration rate; FU = follow-up; RAAS = renin-angiotensin-aldosterone system; VEGF = vascular endothelial growth factor

For categorical variables, a two-tailed Fisher’s exact test was used, while for the continuous variables such as age, BMI, creatinine and eGFR a Mann-Whitney U test was used (except for serum albumin, for which an unpaired t-test was used)

There was no significant difference in serum creatinine or eGFR at the time of biopsy (ING: creatinine 1.8 mg/dl [IQR 1.1], eGFR 28.1 ml/min/1.73 m²; DN: creatinine 2.2 mg/dl [IQR 1.2], eGFR 29.1 ml/min/1.73 m²). Proteinuria was present in nearly all cases (100% in ING vs. 98.4% in DN; p = 1.000) and hematuria was observed in 36.4% of the ING cases. Mean serum albumin did not differ between DN (34.1 ± 8.2 g/l) and ING patients (32.9 ± 8.2 g/l; p = 0.834). Additional clinical features at the time of biopsy and during follow-up are summarized in Table 1.

Among DN patients, 12.5% (n = 8) had type 1 diabetes mellitus (DM) and 87.5% (n = 56) had type 2 DM. Two patients initially classified as ING were reclassified as DN: one was receiving oral antidiabetic treatment at the time of biopsy and the other was diagnosed with diabetes mellitus 4.5 months after biopsy. Additionally, one patient had a pre-existing diagnosis of cystic fibrosis.

Diffuse vs. nodular

Clinical features were compared across four patient subgroups in the non-matched cohort based on histology (ING nodular, ING diffuse, DN nodular, DN diffuse). No significant difference in age was observed between the four groups (p = 0.059).

Serum albumin was significantly different among the four histological subgroups (p < 0.001). Post-hoc tests revealed a significant difference between DN nodular (28.7 ± 5.8 g/l) and DN diffuse (38.6 ± 7.4 g/l) with a mean difference of 9.88 g/l (p < 0.001).

Additionally, there was a significant difference in the prevalence of nephrotic syndrome across subgroups (p = 0.040). Notably, 60.0% of DN patients with nodular lesions had nephrotic syndrome, compared to 20.0% of DN patients with diffuse lesions.

Histological features

Matched ING and DN patients were compared histologically using light microscopy (LM) and results are summarized in Table 2. Diffuse mesangial expansion and nodular sclerosis were present in both entities. Focal segmental glomerulosclerosis (FSGS) and mesangiolysis were observed more frequently in ING patients (FSGS 72.7%, mesangiolysis 45.5%) than in DN patients (FSGS 54.5%, mesangiolysis 9.1% respectively), but these differences were not statistically significant (p = 0.659 and p = 0.149respectively). ING subjects also showed a trend toward more ischemic-type collapse compared to DN patients (36.4% vs. 0%, p = 0.090) (Table 2). Podocyte prominence did not differ between the groups.

Table 2.

Histological features compared between ING and DN subjects

	ING (n = 11)	DN (n = 11)	p-value
Median number of glomeruli (IQR)	30.0 (21.0)	14.0 (7.0)	0.028
Mean proportion global sclerotic glomeruli in % ± SD	7.2 ± 6.4	6.3 ± 4.1	0.079
Mean proportion glomerular obsolescence in % ± SD	7.5 ± 6.7	6.6 ± 4.1	0.132
Diffuse glomerulosclerosis, % • Mild • Moderate • Severe	100.0% (n = 11) • 0.0% (n = 0) • 45.5% (n = 5) • 54.5% (n = 6)	100.0% (n = 11) • 18.2% (n = 2) • 36.4% (n = 4) • 45.5% (n = 5)	/ • 0.670
Nodular sclerosis	72.7% (n = 8)	45.5% (n = 5)	0.387
Distribution NS • None • Focal + segmental • Global	• 54.5.3% (n = 6) • 36.4% (n = 4) • 9.1% (n = 1)	• 27.3% (n = 3) • 54.5% (n = 6) • 18.2% (n = 2)	0.505
Score NS • Absent • 1–50% • > 50%	• 54.5% (n = 6) • 36.4% (n = 4) • 9.1% (n = 1)	• 27.3% (n = 3) • 36.4% (n = 4) • 36.4% (n = 4)	0.317
Mesangial hypercellularity	100.0% (n = 11)	90.9% (n = 10)	1.000
Score hypercellularity • Mild (4–5 cells/area) • Moderate (6–7 cells/area) • Severe ( > = 8 cells/area)	• 18.2% (n = 2) • 63.6% (n = 7) • 18.2% (n = 2)	• 45.5% (n = 5) • 36.4% (n = 4) • 9.1% (n = 1)	0.407
FSGS	72.7% (n = 8)	54.5% (n = 6)	0.659
Mesangiolysis	45.5% (n = 5)	9.1% (n = 1)	0.149
Glomerulonecrosis	0.0% (n = 0)	0.0% (n = 0)	/
Capillary thrombosis	0.0% (n = 0)	0.0% (n = 0)	/
Ischemic-type collapsing lesions	36.4% (n = 4)	0.0% (n = 0)	0.090
Abnormality Bowman’s capsule	100.0% (n = 11)	72.7% (n = 8)	0.214
Crescents	0.0% (n = 0)	0.0% (n = 0)	/
Tuft adhesions	81.8% (n = 9)	45.5% (n = 5)	0.183
Insudative changes in Bowman’s capsule	9.1% (n = 1)	9.1% (n = 1)	1.000
Capsular hyaline drops	18.2% (n = 2)	9.1% (n = 1)	1.000
Hyaline caps	18.2% (n = 2)	0.0% (n = 0)	0.476
Tip lesions	0.0% (n = 0)	0.0% (n = 0)	/
Membranous changes	0.0% (n = 0)	0.0% (n = 0)	/
Duplication of GBM	9.1% (n = 1)	0.0% (n = 0)	1.000
Detachment of podocytes	18.2% (n = 2)	18.2% (n = 2)	1.000
Prominence of podocytes	63.6% (n = 7)	63.6% (n = 7)	1.000
Glomerular deposits	0.0% (n = 0)	0.0% (n = 0)	/
IFTA • < 25% • 25–50% • > 50%	100.0% (n = 11) • 18.2% (n = 2) • 45.5% (n = 5) • 36.4% (n = 4)	100.0% (n = 11) • 27.3% (n = 3) • 45.5% (n = 5) • 27.3% (n = 3)	/ • 1.000
Interstitial inflammation, % • Absent • In relation to IFTA • In areas without IFTA	90.9% (n = 10) • 9.1% (n = 1) • 63.6% (n = 7) • 27.3% (n = 3)	90.9% (n = 10) • 9.1% (n = 1) • 72.7% (n = 8) • 18.2% (n = 2)	1.000 • 1.000
ATN	18.2% (n = 2)	18.2% (n = 2)	1.000
Interstitial hemorrhage	0.0% (n = 0)	0.0% (n = 0)	/
Median number of large vessels (IQR)	4.0 (3.0)	2.0 (2.0)	0.088
Arteriosclerosis • No thickening • Intima < media • Intima > media	90.0% (n = 9/10) • 10.0% (n = 1) • 20.0% (n = 2) • 70.0% (n = 8)	100% (n = 10/10) • 0.0% (n = 0) • 10.0% (n = 1) • 90.0% (n = 9)	1.000 • 0.582
Hyalinosis arterioles • Absent • At least one area • More than one area	90.9% (n = 10) • 9.1% (n = 1) • 45.5% (n = 5) • 45.5% (n = 5)	100.0% (n = 11) • 0.0% (n = 0) • 27.3% (n = 3) • 72.7% (n = 8)	1.000 • 0.387
Luminal narrowing • None • Mild (10%) • Moderate (10–20%) • Severe (> 20%)	• 9.1% (n = 1) • 27.3% (n = 3) • 27.3% (n = 3) • 36.4% (n = 4)	• 0.0% (n = 0) • 9.1% (n = 1) • 54.5% (n = 6) • 36.4% (n = 4)	0.390
Hyalinosis vascular pole	54.5% (n = 6)	9.1% (n = 1)	0.063
Aberrant vessels • Absent • Glomerular vascular pole only • Periglomerularly • In both places	100.0% (n = 11/11) • 0.0% (n = 0/11) • 54.5% (n = 6) • 0.0% (n = 0) • 45.5% (n = 5)	72.7% (n = 8/11) • 27.3% (n = 3/11) • 18.2% (n = 2) • 27.3% (n = 3) • 27.3% (n = 3)	0.214 • 0.045

Comparison of histological features between subjects with primary diagnosis of ING or DN. Mean ± standard deviation (SD) is reported for continuous parametric variables

For categorical variables, A Fisher’s Exact test was used, while for non-parametric continuous variables (reported as median) a Mann-Whitney U test was used. For parametric continuous variables (reported as mean) an unpaired t-test was used.

No significant glomerular deposits were detected on immunofluorescence in either group.

Insudative changes, alterations in the GBM or podocytes, IFTA, concomitant ATN, and interstitial inflammation did not differ significantly between ING and DN patients. Regarding the arteries, arterioles, and capillaries, no differences were observed in arteriosclerosis and arteriolar hyalinosis (p = 1.000) (Table 2). Aberrant vessels, as defined in Supplementary table S1, showed a tendency towards being more often present in ING patients (100.0% versus 72.7%; p = 0.214), and showed a significantly different distribution between groups (p = 0.045). In ING, all patients (100.0%) had aberrant vessels at the vascular pole (compared with only 45.5% of DN patients) and 45.5% of ING patients had the additional finding of aberrant vessels in the periglomerular distribution (compared with only 9.2% in DN). (Table 2).

Prognosis

Survival analyses

Clinical events following the biopsy of ING and DN patients were analyzed in the unmatched ING and DN cohorts.

The primary outcome was a composite of dialysis initiation, kidney transplantation, death and/or doubling of serum creatinine. Differences in cumulative survival from biopsy to these events are illustrated in Fig. 1A. Overall, there was no significant difference in mean time to the composite outcome between ING and DN patients (ING: 516 days, 95% CI [339.8-692.2] versus DN (970.9 days, 95% CI [724.8-1217.1]; p = 0.094).

Fig. 1.

Kaplan–Meier survival analysis of the composite outcome (dialysis, kidney transplantation, death, or doubling of serum creatinine) comparing patients with ING and DN. (A) Non-stratified analysis comparing ING versus DN (log-rank test, p = 0.09). (B) Analysis stratified by smoking (active versus non-active smokers). Among active smokers, survival differed significantly between ING and DN (log-rank test, p = 0.01), whereas no significant difference was observed among non-active smokers (log-rank test, p = 0.49)

When stratified for smoking, there is however a significant difference among active smokers(p = 0.010) (Fig. 1B). Active smokers in the ING cohort had a mean time-to-event of 398 days (95%CI [198–598]) compared to 1356 days (95%CI [852 to 1860]) for active smokers in the ING cohort. Among non-active smokers, the composite outcome did not differ significantly between groups (p = 0.49) (Fig. 1B).

A similar trend was observed when comparing the time-to-composite outcome for the four histological diagnoses. The overall survival did not differ significantly (p = 0.346) (Supplementary Fig. S2).

When stratified by smoking status, a significant difference in survival was seen among active smokers (p = 0.005) (Supplementary Fig. S3). The mean survival time was lower in ING patients with 239 days (95%CI [53–425]) in the ING nodular cohort and 557 days (95%CI [453–661]) in the ING diffuse cohort, compared to DN patients, with 1269 days (95%CI [497–2041]) in the DN nodular cohort and 1404 days (95%CI [817–1991]) in the DN diffuse cohort (Supplementary Fig. S3). Among non-smokers, again the composite outcome did not differ significantly (Supplementary Fig. S4).

Longitudinal analysis of eGFR

ING vs. DN

Linear mixed model analyses of eGFR values showed no significant interaction between diagnosis (ING versus DN) and time.

The estimated annual eGFR decline in ING patients was − 6.27 ml/min/1.73m² per year (95%CI [-7.83 to -4.70], which was similar to DN patients, who had a decline of -5.41 ml/min/1.73m² per year (95%CI [-8.67 to -2.16] (p = 0.322 for diagnosis x time interaction) (Fig. 2).

Fig. 2.

Predicted values based on the random intercept model of eGFR (ml/min/1.73 m²) in function of the time from biopsy to eGFR measurement (in years), compared between DN (-5.41 ml/min/1.73 m² per year 95%CI [-8.67 to -2.16] and ING (-6.27 ml/min/1.73 m² per year 95%CI [-7.83 to -4.70]) (p = 0.322 for diagnosis-time interaction and p < 0.001 for the time slope)

Smokers vs. non-smokers

When comparing active smokers to non-smokers (ex-smokers + never-smokers) at the time of biopsy, LMM analysis demonstrated a significant interaction between smoking status and time (p < 0.001) indicating that eGFR decline differed between the groups. The estimated slope for active-smokers was − 8.62 ml/min/1.73 m² per year (95%CI [-10.10 to -7.16]), whereas for non-smokers it was − 4.96 ml/min/1.73 m² per year (95%CI [-8.01, to -1.90]) (p < 0.001) (Fig. 3).

Fig. 3.

Predicted values based on the random intercept model of eGFR (ml/min/1.73 m²) in function of the time from biopsy to eGFR measurement compared between active smokers (-8.62 ml/min/1.73 m² per year with 95%CI [-10.10 to -7.16]) and non-smokers (ex-smokers and never smokers) (-4.96 ml/min/1.73 m² per year with 95%CI [-8.01 to -1.90]) (p < 0.001 for the smoker-time interaction and p < 0.001 for the time slope)

Discussion

In this study, we found that DN and ING patients share many clinical characteristics. They have no significantly different kidney and overall survival, although a trend towards worse clinical outcome was observed in the ING group. DN and ING share also many histological characteristics, with a difference for distribution of aberrant arteriolar vessels in our small exploratory study.

Cardiovascular risk factors, including hypertension, hyperlipidemia, obesity, and smoking, were highly prevalent in the ING group, supporting a potential pathogenic role of vascular injury in this disease [4–7, 34]. Although the proportion of patients with a smoking history was numerically higher in ING than in DN, this difference did not reach statistical significance, likely reflecting the limited sample size. Nevertheless, smoking emerged as a key modifier of clinical outcomes. Though our study design and sample size do not support strong conclusions on prognosis, active smokers exhibited a significantly steeper decline in eGFR compared to non-active smokers, irrespective of diagnosis. This suggests a detrimental effect of smoking on kidney disease progression. Moreover, after stratification by smoking status, patients with ING who were active smokers reached the composite outcome earlier than their DN counterparts, with proportionally more deaths observed in the ING cohort. Similar patterns were noted across histological subgroups, particularly among patients with nodular lesions, which already seem to have a worse prognosis than patients with diffuse lesions [35, 36]. These findings align with prior reports indicating an increased risk of progression to end-stage kidney disease among ING patients who continue smoking [5], and underscore smoking as an important, potentially modifiable risk factor in both ING and DN. Mechanistically, smoking-related endothelial dysfunction and ischemia may contribute to disease progression and histological overlap between the two entities, although this hypothesis requires confirmation in larger prospective studies.

The higher age within the ING group is likely to have influenced the composite outcome, potentially contributing to the decline in kidney function in this cohort. However, we should refrain from drawing firm conclusions due to the inherent risk of selection bias, resulting from baseline disparities between the two patient populations and the relatively small sample size. A prospective registration of outcomes in a larger population would therefore be valuable to validate these findings.

Aberrant vessels were observed in both DN and ING, with differing distributions in our exploratory cohort. Neovascularization pathways may contribute to their formation in both conditions [1, 4, 15, 37], potentially reflecting altered hemodynamic responses or VEGF expression [4]. In ING, aberrant vessels predominantly localized to the vascular pole compared with DN, consistent with prior findings [4], although methodological differences exist and we present only a small cohort. We hypothesize that ischemic changes in ING, related to risk factors such as hypertension and smoking, promote neovascularization, potentially mediated by nicotine-induced endothelin-1 activation and downstream hypoxia-related pathways [9, 38–41]. Ischemic-type collapsing lesions were more frequent in ING, though not statistically significant (p = 0.090). Aberrant vessels at the vascular pole have also been widely reported in DN [16, 27, 37], supporting the hypothesis that glomerular tuft herniation drives capillary outgrowth and tuft adhesions [27], which may extend into periglomerular areas. Further studies are required to clarify whether distinct pathogenic mechanisms underlie aberrant vessel formation in DN versus ING [42].

Limitations of this study include its retrospective design and the relatively small sample size, which may have reduced statistical power and increased the risk of type II error. In addition, the limited number of events precluded multivariable adjustment, thereby preventing causal inference. However, the limited number of patients is largely unavoidable given the rarity of ING. There was variability in longitudinal eGFR measurements in the DN cohort compared to the ING cohort. Specifically, the absence of fixed time points resulted in a declining number of measurements over time, which may have influenced the interpretation of the LMM analysis. Although matching for sex, age and histological pattern was performed, potentially other unmeasured confounders may still have influenced the outcomes. Validation in larger, preferably prospective cohorts with standardized follow-up and more granular registration of comorbidities is warranted.

Second, the risk of representation bias in these studies could be high, especially in the DN group. DN is frequently diagnosed based on clinical features such as proteinuria and laboratory values. Only patients with severe or atypical clinical presentations of the disease are typically biopsied. This selection bias is common in biopsy-based research.

Third, more granular data on the degree of exposure to risk factors such as smoking (e.g., cumulative pack-years) or hypertension (e.g., graded severity) were lacking. However, reliance on pack-year calculations alone may introduce misclassification and underestimation of smoking exposure, since reporting in health records can be incomplete and self-reported history is subject to recall bias [43]. Similarly, the duration and severity of hypertension could not be reliably assessed due to the retrospective design of the study. Future prospective studies should include standardized blood pressure measurements. Nevertheless, given the comparable distribution of hypertension between groups, it is unlikely that this factor significantly biased either the histological or outcome data.

A major strength of this study lies in its comprehensive and structured evaluation of a broad spectrum of histological features. It provides a detailed overview of histological differences and similarities between the two entities, albeit in a small cohort, and suggests the ischemia-driven pathogenesis hypothesis in ING should be further examined. Inclusion of immunohistochemical markers, such as CD31 for endothelium, could be an added value in future studies.

Overall, this study provides an exploratory comparison of long-term clinical outcomes, including kidney function, between patients with idiopathic nodular glomerulosclerosis and diabetic nephropathy. While histological features largely overlap, subtle differences—such as a different distribution of aberrant vessels in ING—were observed. Clinically, our findings suggest that active smokers may show less favorable outcomes in patients with ING, although these observations should be interpreted cautiously and require confirmation in prospective studies in larger study populations.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A. D., S. V. L.: Conceptualization; K. B., P. K., A. D., S. V. L.: Data curation; K. B., I. V. G., A. D., S. V. L.: Formal analysis; K. B., A. D., S. V. L.: Investigation; K. B., A. D., S. V. L.: Methodology; K. B., A. D., S. V. L.: Project administration; K. B., A. D., S. V. L.: Resources; A. D., S. V. L.: Supervision; K. B.: Visualization; K. B., A. D., S. V. L.: Roles/Writing - original draft; K. B., I. V. G., P. K., A. D., S. V. L.: Writing - review & editing.

Funding

This study was not supported by any sponsor or funder.

Data availability

All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.

Declrations

Ethical approval

This study protocol was reviewed and approved by the ethical committee of Ghent University Hospital, reference number INZ-2022-0305 with exemption from requiring a written informed consent.

Competing interests

The authors declare no competing interests.