Predictors of non-diabetic kidney disease in diabetics: A Saudi Arabian perspective

Abstract

BACKGROUND

Diabetic patients with atypical presentation are often challenging in terms of diagnosis and management. Kidney biopsy is not routinely done in diabetics, and clinicians are always in a dilemma in such a scenario to decide whether to do a biopsy or not. Since non-diabetic kidney diseases (NDKD) are common, and some patients may have NDKD superimposed on diabetic kidney diseases (DKD), therefore, kidney biopsy may be warranted to rule out NDKD.

AIM

To determine the prevalence of NDKD, DKD, or mixed lesions, identify predictors of NDKD, and investigate renal and patient survival, as well as factors associated with these outcomes.

METHODS

This retrospective observational study was conducted on patients with biopsy-proven NDKD, DKD, and mixed lesions (having both NDKD and DKD). Binary logistic regression models were constructed to identify predictors of NDKD. Kaplan-Meier survival analysis was performed to compare time to kidney failure and patient survival across the three histological groups. Multivariable Cox proportional hazards regression was used to identify clinical and pathological factors associated with kidney failure and all-cause mortality.

RESULTS

A total of 103 biopsies were analyzed. Sixty-four (62.1%) had NDKD alone or mixed lesions. The most common NDKD pathologies were interstitial nephritis in 12 (29.2%), focal segmental glomerulosclerosis in 10 (24.4%), and immune complex–mediated glomerulonephritis in five (12.2%) patients. Compared to DKD, NDKD was associated with significantly lower odds of proteinuria > 3.5 g/day [odds ratio (OR), 0.02; P = 0.0015], retinopathy (OR = 0.04; P = 0.0067), and diabetes duration ≥ 10 years (OR = 0.01; P = 0.0002). However, NDKD had higher odds of anemia (Hemoglobin < 12 g/dL; OR = 9.56; P = 0.0107) and creatinine levels > 180 μmol/L (OR = 18.68; P = 0.0063). Kaplan-Meier analysis showed significant differences in renal survival (log-rank P = 0.0033). Patients with NDKD have the best outcomes, while those with DKD have the worst. In a multivariable Cox regression analysis, increasing age, creatinine, arteriosclerosis, and severe interstitial fibrosis and tubular atrophy were independently associated with kidney failure. At the same time, the use of renin angiotensin system blockers was protective (hazard ratio = 0.43, P = 0.02). Kaplan-Meier curves for patient survival also differed significantly (log-rank P = 0.018); patients in the mixed group showed the highest mortality, while those with NDKD showed the lowest. Mortality was independently associated with older age, hypoalbuminemia, diabetic retinopathy, arteriosclerosis, and higher creatinine.

CONCLUSION

NDKD and mixed lesions are frequent in diabetic patients. These histological lesions carry distinct prognostic implications. Clinical features such as a shorter diabetes duration, absence of retinopathy, anemia, and elevated creatinine levels suggest NDKD and warrant biopsy. NDKD had better renal and patient survival rates, while mixed lesions had the worst outcomes. Older age, hypoalbuminemia, retinopathy, arteriosclerosis, and elevated creatinine were key predictors of mortality.

Key Words: Diabetes; Kidney biopsy; Non-diabetic kidney disease; Predictors; Renal survival; Patient survival

Core Tip: Diabetic patients have non-diabetic kidney diseases (NDKD) alone or with diabetic kidney diseases (DKD) in 62.1%. Compared to DKD, NDKD was associated with significantly lower odds of heavy proteinuria, retinopathy, and long-standing diabetes. However, NDKD had higher odds of anemia. NDKD patients experienced better renal and patient survival outcomes. Increasing age, creatinine, arteriosclerosis, interstitial fibrosis, and tubular atrophy were independently associated with kidney failure. Renin angiotensin system blockers were found to be protective in preventing kidney failure. Mortality was independently associated with older age, hypoalbuminemia, diabetic retinopathy, arteriosclerosis, and higher creatinine.

INTRODUCTION

Diabetes mellitus is a global health challenge across the world. It is estimated that the number of adults with diabetes will increase from 630 million in 1990 to 828 million in 2028. This represents a rise in prevalence from 7% to 14% between 1990 and 2022[1]. Around 40% of type 2 diabetes mellitus patients will develop chronic kidney disease (CKD)[2]. The prevalence of type 2 diabetes mellitus is alarmingly high in Saudi Arabia, rising from 8.5% in 1992 to 39.5% in 2022[3], and it’s the leading cause of CKD in Saudi Arabia[4]. Given this growing burden of diabetes and its renal complications, an important question arises about how best to evaluate kidney dysfunction in these patients. The decision to do a kidney biopsy in patients with diabetes is often tricky. Generally, nephrologists may be inclined to attribute kidney dysfunction to diabetic kidney disease without doing a biopsy. However, it is essential to mention that having non-diabetic kidney disease (NDKD) is not uncommon in diabetics[5,6]. Clinicians may opt for a kidney biopsy in diabetic patients without retinopathy, a rapid decline in kidney function, a rapid increase in proteinuria, the sudden appearance of nephrotic syndrome, active sediment, or the emergence of signs and symptoms of systemic diseases[7]. Considering the dilemma between diabetic and NDKD, obtaining a histological diagnosis is crucial. A kidney biopsy can accurately identify a variety of lesions, ranging from classic diabetic nephropathy to mixed lesions and entirely NDKD. Our study aimed to explore the histological spectrum in diabetic patients undergoing biopsy and to determine predictors of NDKD. We also investigated renal and patient survival, as well as factors associated with these outcomes.

MATERIALS AND METHODS

This retrospective study was conducted at the Center of Renal Diseases and Transplantation at King Fahad Armed Forces Hospital, Jeddah. Following approval by the ethics review committee, data were collected between 2000 and 2024 for all diabetic patients who underwent a kidney biopsy. The study included patients with both type 1 diabetes mellitus and type 2 diabetes mellitus who were 14 years of age or older and underwent a native kidney biopsy at our center between 2000 and 2024.

The prevalence of NDKD, mixed lesions, and isolated diabetic kidney diseases (DKD) was calculated using the total number of diabetic patients who underwent native kidney biopsy at our center between 2000 and 2024 (n = 103) as the population at risk. In addition, we also evaluated long-term outcomes, including kidney failure and all-cause mortality. We also studied the predictors of kidney failure and all-cause mortality using survival analysis. The predictors of kidney failure studied included clinical factors such as proteinuria > 3.5 g/day, hemoglobin < 12 g/dL, presence of retinopathy, duration of diabetes ≥ 10 years, age, and serum creatinine > 180 μmol/L, as well as histopathological features such as moderate to severe interstitial fibrosis and tubular atrophy (IFTA), global glomerulosclerosis ratio, and podocyte effacement. A multivariate Cox proportional hazards model was used to identify independent predictors of all-cause mortality, including age, serum creatinine, hemoglobin < 12 g/dL, serum albumin < 30 g/L, retinopathy, and arteriosclerosis.

Kidney failure was defined as a composite endpoint comprising an estimated glomerular filtration rate (eGFR) of < 20 mL/minute/1.73 m², initiation of dialysis, or kidney transplantation. The eGFR threshold of < 20 mL/minute/1.73 m² was chosen due to its clinical relevance, as this is the level at which, under current practice guidelines and recommendations, patients are typically referred for preemptive kidney transplant evaluation and vascular access planning[8].

Biopsy was performed in patients with atypical presentation, such as unexplained kidney dysfunction, sudden onset of proteinuria, hematuria, rapidly declining renal function, shorter duration of diabetes, or absence of diabetic retinopathy. Patients with adequate biopsy tissue were included in the analysis. We excluded patients with previously known NDKD before the diagnosis of diabetes. Patients with a kidney transplant, or biopsy done for mass lesions, or those under 14 years of age were also excluded. One core each for light microscopy and immunofluorescence studies was analyzed. The biopsy tissues were examined under light microscopy using hematoxylin and eosin, Masson’s trichrome, and periodic acid methenamine silver stains to assess glomerular, tubular, interstitial, and vascular pathological changes. Immunofluorescence studies were performed using antisera against human immunoglobulin G, immunoglobulin A (IgA), immunoglobulin M, C3, C1q, kappa, and lambda light chains, as well as fibrinogen, to assess immune complex deposition. Electron microscopy was performed in selected cases to analyze details of basement membrane changes and other ultrastructural alterations. Two experienced renal pathologists interpreted all findings. The presence of nodular or diffuse mesangial expansion, hyalinosis of afferent and efferent arterioles, thickening of the glomerular basement membrane, and the presence of exudative lesions such as fibrin cap and capsular drop were labeled as diabetic nephropathy[7].

The pathological diagnosis and classification of DKD were based on the 2010 classifications of the Renal Pathology Society, which included four groups[9]. Class I included glomeruli that appeared normal on light microscopy but exhibited thickening of the glomerular basement membrane. Class II was defined by mesangial expansion without nodular sclerosis. Class III was characterized by the presence of at least one Kimmelstiel-Wilson nodular lesion, with less than 50% of glomeruli showing global sclerosis. Class IV represented advanced diabetic glomerulosclerosis, with more than 50% of glomeruli globally sclerosed. IFTA was classified as absent (< 10%), mild (10%-24%), moderate (25%-50%), and severe (> 50%) of the tissues[10]. Arteriosclerosis was reported as none (0% luminal narrowing), mild (≤ 25% luminal narrowing), moderate (26%-50% luminal narrowing), and severe (> 50% luminal narrowing or reported as severe)[11].

The following clinical and laboratory parameters were collected at the time of kidney biopsy, where available: Age, gender, type and duration of diabetes, history of hypertension, smoking, dyslipidemia, diabetic retinopathy, neuropathy, and family history of diabetes or CKD. Cardiovascular comorbidities such as coronary artery disease, peripheral arterial disease, and stroke were noted. Past history and duration of proteinuria, as well as deranged renal function, were also reported. The presence of hematuria and proteinuria was also noted. Laboratory investigations, including serum albumin, total protein, serum creatinine, eGFR, hemoglobin, calcium, phosphate, intact parathyroid hormone (PTH), and glycated hemoglobin, were also collected. Data for medications such as insulin, oral hypoglycemic agents, sodium-glucose cotransporter-2 inhibitors (empagliflozin or dapagliflozin), antihypertensive therapy, use of angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, and statins were also collected. Immunological investigations, including complement levels (C3, C4), antinuclear antibody, anti-double-stranded DNA (anti-dsDNA), C-antineutrophil cytoplasmic antibodies (ANCA), and P-ANCA, as well as serology for hepatitis B surface antigen, anti-hepatitis C antibodies, and human immunodeficiency virus, were also recorded. Due to the retrospective nature of the study, some data were not available for all patients. Therefore, only the available information was analyzed for each variable using a complete-case approach. To minimize bias, we ensured careful selection of variables with sufficient data and conducted sensitivity analyses where possible to assess the robustness of our findings. As a result, the reduced sample size for some analyses related to NDKD or DKD may have limited the statistical power and generalizability of these results. This has been recognized as a limitation of the study.

Based on kidney biopsy findings, lesions were classified as pure DKD, NDKD, and mixed lesions (NDKD + DKD). For analysis to identify predictors of NDKD, patients were grouped into two groups. Group I included those with pure DKD, while Group II included those with NDKD. This grouping was used to compare clinical and laboratory features between patients with isolated diabetic nephropathy and those with additional or alternative renal pathologies. Kaplan-Meier curves for DKD, mixed lesions, and NDKD were used to assess time to kidney failure using eGFR at the time of biopsy as the starting point and kidney failure as the endpoint. The Kaplan-Meier survival plot was also used to compare time to death among three histologically defined groups: DKD, mixed, and NDKD, using eGFR at the time of biopsy, the time-to-event variable, and mortality as the event. Risk factors for kidney failure and mortality for the whole cohort using Cox regression analysis.

Statistical analysis

Continuous variables were summarized using median and interquartile range. Categorical variables were reported as frequencies and percentages. Continuous variables included age, duration of diabetes, hemoglobin A1c, creatinine levels (at biopsy, 1 year, and 2 years), eGFR (at 1 and 2 years), hemoglobin, PTH, calcium, phosphate, total protein, albumin, proteinuria (g/day), total glomeruli count, global and segmental sclerosis counts and ratios, IFTA severity, arteriosclerosis severity, and hyalinosis severity. Categorical variables included gender, diabetes type, presence of retinopathy, hypertension, dyslipidemia, neuropathy, cardiovascular and cerebrovascular comorbidities, medication use (including angiotensin enzyme inhibitors, angiotensin receptor blockers, statins, insulin, oral hypoglycemics, and sodium-glucose cotransporter-2 inhibitors, indication for biopsy, presence of hematuria, crescent formation, IFTA (≥ 25%), podocyte effacement, antinuclear antibody, anti-dsDNA, complement levels (C3, C4), ANCA (C-ANCA, P-ANCA), kidney failure, and mortality.

Group comparisons were made using Student’s t-test, Mann-Whitney U test, χ², or Fisher’s exact test, as appropriate. Multivariable logistic regression analysis was performed to identify independent predictors of DKD vs NDKD and mixed lesions. Survival analysis was carried out using Cox proportional hazards regression to identify predictors of kidney failure and all-cause mortality. Kaplan-Meier curves with log-rank tests were used to compare time to kidney failure and mortality across biopsy groups. A P-value < 0.05 was considered statistically significant.

RESULTS

Kidney biopsy findings

Three hundred and fifty-nine biopsies were performed in the study period. Out of 359 biopsies, 106 were performed on individuals with diabetes. After excluding three kidney biopsies for mass lesions, a total of 103 out of 359, representing 28.69% of the total, were analyzed. Sixty-four (62.1%) patients had NDKD alone or mixed lesions (NDKD with DKD). Of these, 41 (39.8%) were diagnosed with NDKD, 23 (22.3%) with mixed lesions, and 39 (39.8%) with pure DKD. The most frequent pathology in the NDKD group was interstitial nephritis in 12 (29.2%), followed by focal segmental glomerulosclerosis (FSGS) in 10 (24.4%) and immune complex-mediated glomerulonephritis in five (12.2%) patients. Among mixed lesions, immune complex-mediated glomerulonephritis was present in six (26.1%) and IgA nephropathy in four (17.4%). Other less frequent findings included membranous nephropathy, lupus nephritis, chronic pyelonephritis, and rare entities such as C1q nephropathy, C3 glomerulopathy, fibrillary GN, and pauci-immune vasculitis. The detailed distribution of biopsy findings is presented in Table 1. Figure 1 shows the prevalence of NDKD, DKD, and mixed lesions. Figure 2 shows light microscopy and electron microscopy showing a mixed lesion of FSGS and diabetic nephropathy. Among patients with DKD, the most common histological classification was Class 4, seen in 16 patients (41.0%). This was followed by Class 3 in 15 patients (38.5%) and Class 2 in seven patients (17.9%). Only one patient (2.6%) had early-stage disease classified as Class 1. Overall, the majority of patients had moderate to advanced DKD lesions. Table 2 shows the stages of histologically proven DKD.

Figure 1 Illustrates the prevalence of pure diabetic nephropathy and non-diabetic kidney diseases, as well as non-diabetic kidney diseases + diabetic nephropathy. DKD: Diabetic kidney diseases; NDKD: Non-diabetic kidney diseases.

Figure 2 Shows focal segmental sclerosis with early diffuse diabetic nephropathy changes. A: Segmental glomerular sclerosis showing obliteration of glomerular capillary lumina (arrow) and adhesion of the sclerotic tuft to Bowman’s capsule (star) (Periodic acid–Schiff stain, × 40); B: Ultrastructurally, the mesangium is expanded by increased mesangial matrix (star), and the glomerular basement membrane is thickened (arrow) (Transmission electron microscopy, uranyl acetate and lead citrate stain); C: Electron micrograph demonstrating loss of normal glomerular basement membrane ultrastructural detail (arrow) and subendothelial deposition of dense material consistent with hyalinosis (star) (Transmission electron microscopy, uranyl acetate and lead citrate stain).

Table 1 Histological spectrum of non-diabetic kidney diseases vs mixed lesions, n (%).

Biopsy category	NDKD	NDKD + DKD	Total
Interstitial nephritis	12 (29.2)	-	12 (18.2)
FSGS	10 (24.4)	3 (13.0)	13 (20.3)
Immune complex glomerulonephritis	5 (12.2)	6 (26.1)	11 (17.2)
IgA	2 (4.9)	4 (17.4)	6 (9.4)
Membranous	4 (9.8)	2 (8.7)	6 (9.4)
Lupus nephritis	4 (9.8)	-	4 (6.2)
Chronic pyelonephritis	-	2 (8.7)	2 (3.1)
Minimal change disease	3 (7.3)	-	3 (4.7)
Multiple myeloma	1 (2.4)	1 (4.3)	2 (3.1)
C1Q nephropathy	-	1 (4.3)	1 (1.6)
C3 glomerulopathy	-	1 (4.3)	1 (1.6)
Fibrillary glomerulonephritis	-	1 (4.3)	1 (1.6)
Pauci-immune vasculitis	-	1 (4.3)	1 (1.6)
Post-infectious glomerulonephritis	-	1 (4.3)	1 (1.6)
Total	41 (100.0)	23 (100.0)	64 (100.0)

DKD: Diabetic kidney diseases; NDKD: Non-diabetic kidney diseases; FSGS: Focal segmental glomerulosclerosis; IgA: Immunoglobulin A.

Table 2 Biopsy classification in the diabetic kidney disease group (n = 39).

DKD classification	n (%)
Class 4	16 (41.0)
Class 3	15 (38.5)
Class 2	7 (17.9)
Class 1	1 (2.6)

DKD: Diabetic kidney diseases.

Baseline demographic and clinical characteristics

The median age in our cohort was significantly higher in the DKD + NDKD group (69.0 years) compared to the DKD (53.0 years) and NDKD (50.0 years) groups (P = 0.0003). There were no significant differences in gender distribution across groups (P = 0.3518). We observed that type 1 diabetes mellitus was more common in the DKD group, present in 10 patients (n = 10, 26.3%) as compared to DKD + NDKD (n = 1, 4.3%) and NDKD (n = 1, 2.4%) groups (P = 0.0020). The duration of diabetes varied significantly among the groups. We observed the longest duration in DKD (median 168 months), followed by DKD + NDKD (108 months), and the shortest in NDKD (60 months) (P < 0.0001). Hemoglobin A1c levels were similar between the groups (P = 0.2877). Diabetic retinopathy was significantly more common in the DKD group (n = 23, 59.0%) than in the DKD + NDKD (n = 9, 39.1%) and NDKD (n = 10, 24.4%%) groups (P = 0.0070).

While examining comorbidities, we found no significant differences in the prevalence of hypertension (P = 0.3005) or dyslipidemia (P = 0.9336) across all groups. However, we found that neuropathy was more frequent in the DKD group (n = 22, 61.1%) compared to the DKD + NDKD group (n = 11, 55%) and the NDKD group (n = 2, 9.1%) (P = 0.0003). Coronary artery disease was more common in the DKD group (n = 23, 65.7%) than in DKD + NDKD (n = 8, 42.1%) and NDKD (n = 1, 14.2%) (P = 0.0000). Similarly, peripheral artery disease and stroke were more prevalent in DKD patients (n = 30, 85.7% and n = 28, 77.8% respectively) and were less frequent in the other two groups (both P < 0.0001).

We analyzed the indication for the biopsy. We found that sudden renal function decline was the most common indication in the NDKD group (58.5%) and DKD + NDKD group (52.2%). However, it was a less common indication in DKD (30.8%). Conversely, a rise in proteinuria was the predominant indication for biopsy in DKD patients (51.3%) but was less common in DKD + NDKD (30.4%) and rare in NDKD (17.1%). Hematuria as an indication for biopsy was not significantly different across all groups (P = 0.2809). Analysis of glomerular sclerosis revealed important features in our study. We found the lowest total glomeruli count in the NDKD group (median 10.0, interquartile range 8.0-17.0) compared to the DKD and mixed groups (P = 0.0154). Global glomerulosclerosis was significantly more common in the DKD group, with a median of 5.0 globally sclerosed glomeruli. This was in contrast to 4.0 in the mixed group and 2.0 in the NDKD group (P = 0.0007). Similarly, the global sclerosis ratio was highest in the DKD group (median 0.3). From this, we can infer that nearly one-third of the glomeruli were globally sclerosed. This was higher when compared to 0.2 in the mixed group and 0.1 in the NDKD group (P = 0.0212). We also found higher segmental sclerosis count and ratio in DKD. However, this achieved statistical significance only for the count (P = 0.0399), and not for the ratio (P = 0.6571). From these findings, we can deduce that DKD is associated with more extensive chronic glomerular damage. On the other side, NDKD may present with relatively preserved glomerular architecture.

We observed more crescents in DKD + NDKD (21.7%) and NDKD (12.2%) compared to DKD (0.0%) (P = 0.0567). This indicates more active inflammatory injury in non-diabetic or mixed lesions. IFTA ≥ 25% were most prevalent in the DKD + NDKD group (77.3%), followed by NDKD (69,7%) and DKD (41.0%) (P = 0.0073). The severity of IFTA differed significantly across the three biopsy-based groups (P = 0.0019). It was interesting to note that severe IFTA (> 50%) was observed exclusively in the DKD group (Group I), affecting 25.6% of patients. At the same time, it was absent in both the mixed (DKD + NDKD) and NDKD groups. Moderate IFTA (25%-50%) was also more common in the DKD group (33.3%), compared to 22.7% in the mixed group and 30.3% in the NDKD group. The absence of IFTA (< 10%) was infrequent across all groups. We observed no significant differences in the presence or severity of arteriosclerosis, nor hyalinosis scores, across the groups. The severity of podocyte effacement, as assessed by electron microscopy, differed significantly among the three groups (P = 0.0042) on univariate analysis. Diffuse effacement (> 50%) was most frequent in the NDKD group (52.0%), compared to 30.0% in the mixed group and 26.3% in the DKD group. Focal effacement (≤ 50%) was also more commonly reported in the NDKD and mixed groups (36.0% and 20.0%, respectively) than in the DKD group (7.9%). We also looked into renal and patient outcomes. Kidney failure, defined as a composite of sustained eGFR < 20 mL/minute/1.73 m², initiation of dialysis, or kidney transplantation, was significantly more common in the DKD group (97.4%) compared to the DKD + NDKD (65.2%) and NDKD (61.0%) groups (P = 0.0002). This reflects poorer renal survival in patients with pure diabetic nephropathy.

We found comparable creatinine levels at the time of the biopsy; however, we observed significant divergence at 1 year, where the median creatinine was highest in the DKD group (483.5 μmol/L), compared to the DKD + NDKD group (352.0 μmol/L) and the NDKD group (124.0 μmol/L) (P = 0.0000). A similar trend was observed at 2 years (P = 0.0002). We found significantly lower eGFR at 1 year (P = 0.056) and 2 years (P = 0.0372). This indicates a more rapid decline in kidney function in patients with DKD. It was interesting to note that anemia was also more prevalent in DKD patients. DKD patients had the lowest median hemoglobin levels (10.0 g/dL), followed by DKD + NDKD (11.1 g/dL) and NDKD (12.1 g/dL) (P = 0.0013). We observed significantly higher protein-to-creatinine ratios in the DKD group. The DKD group had a protein-to-creatinine ratio (7.04 g/mmol) compared to the DKD + NDKD (1.80) and NDKD (2.90) groups (P = 0.0187). We found no significant differences in albumin, calcium, phosphate, and PTH levels.

Kidney failure (defined as a composite of sustained eGFR < 20 mL/minute/1.73 m², dialysis initiation, or kidney transplantation) was noted most frequently in the DKD group (97.4%), followed by the mixed group (65.2%) and NDKD group (61.0%). The difference across groups was statistically significant (P = 0.0002). This suggests a markedly higher risk of progression to kidney failure among patients with pure diabetic kidney disease. We observed the highest mortality in the mixed group (30.4%), followed by the DKD group (20.5%) and the NDKD group (12.2%). However, these differences between the three groups were not statistically significant (P = 0.2038). Cardiac causes and malignancy were less common. Table 3 describes baseline demographic and clinical characteristics.

Table 3 Baseline demographic and clinical characteristics, n (%).

Variable	DKD (n = 39)	DKD + NDKD (n = 23)	NDKD (n = 41)	P value
Gender				0.3518
Male	20 (51.3)	16 (69.6)	25 (61.0)
Female	19 (48.7)	7 (30.4)	16 (39.0)
Age, median (IQR)	53.0 (42.0-66.0)	69.0 (61.5-74.0)	50.0 (46.0-61.0)	0.0003
Diabetes diagnosis				1.0000
Yes	39 (100.0)	23 (100.0)	41 (100.0)
Diabetes type				0.0020
Type 1	10 (26.3)	1 (4.3)	1 (2.4)
Type 2	28 (73.7)	22 (95.7)	40 (97.6)
Duration of diabetes (months)	168.0 (120.0-240.0)	180.0 (36.0-216.0)	60.0 (12.0-108.0)	< 0.0001
Haemoglobin A1c level	7.1 (5.9-8.3)	7.2 (6.3-9.2)	6.8 (6.3-9.2)	0.5502
Retinopathy				0.0070
Yes	23 (59.0)	9 (39.1)	10 (24.4)
No	16 (41.0)	14 (60.9)	31 (75.6)
Hypertension				0.3005
Yes	34 (91.9)	20 (100.0)	35 (97.2)
No	3 (8.1)	0 (0.0)	1 (2.8)
Dyslipidaemia				0.9336
Yes	23 (60.5)	13 (59.1)	22 (56.4)
No	15 (39.5)	9 (40.9)	17 (43.6)
Neuropathy				0.0003
Yes	22 (61.1)	11 (55.0)	2 (9.1)
No	14 (38.9)	9 (45.0)	20 (90.9)
Coronary artery disease				0.0000
Yes	23 (65.7)	8 (42.1)	1 (4.2)
No	12 (34.3)	11 (57.9)	23 (95.8)
Peripheral artery disease				< 0.0001
Yes	30 (85.7)	12 (63.2)	1 (4.0)
No	5 (14.3)	7 (36.8)	24 (96.0)
Stroke				0.0001
Yes	28 (77.8)	12 (63.2)	5 (20.8)
No	8 (22.2)	7 (36.8)	19 (79.2)
Blood pressure-lowering medications				0.5112
≥ 1 medication	36 (92.3)	19 (82.6)	36 (87.8)
0 medication	3 (7.7)	4 (17.4)	5 (12.2)
ACE inhibitor use				0.1394
Yes	11 (30.6)	3 (13.0)	6 (14.6)
No	25 (69.4)	20 (87.0)	35 (85.4)
ARBs use				0.6285
Yes	21 (56.8)	11 (47.8)	19 (46.3)
No	16 (43.2)	12 (52.2)	22 (53.7)
Statin use				0.6736
Yes	29 (78.4)	19 (82.6)	30 (73.2)
No	8 (21.6)	4 (17.4)	11 (26.8)
Insulin use				0.000
Yes	29 (76.3)	15 (65.2)	11 (26.8)
No	9 (23.7)	8 (34.8)	30 (73.2)
Oral hypoglycemics				0.2538
Yes	17 (45.9)	8 (34.8)	23 (56.1)
No	20 (54.1)	15 (65.2)	18 (43.9)
SGLT2 inhibitor use				0.8871
Yes	2 (5.7)	1 (4.3)	3 (7.3)
No	33 (94.3)	22 (95.7)	38 (92.7)
Creatinine at biopsy (µmol/L)	295.0 (166.0-596.5)	323.0 (193.5-547.5)	210.0 (105.0-580.0)	0.3004
Creatinine at 1 year	483.5 (211.0-694.8)	352.5 (157.0-551.2)	124.0 (89.5-182.5)	0.0000
Creatinine at 2 years	491.0 (243.0-891.0)	302.0 (134.0-561.0)	122.0 (92.0-169.0)	0.0002
eGFR at 1 year	16.0 (11.0-25.5)	34.0 (19.5-44.5)	45.5 (22.8-77.2)	0.0056
eGFR at 2 years	19.0 (15.0-22.0)	30.5 (13.0-47.8)	55.0 (32.0-77.0)	0.0372
Hemoglobin (g/dL)	10.0 (9.1-11.1)	11.1 (9.8-14.5)	12.1 (10.4-78.3)	0.0013
Yes (Hb < 12)	30 (76.9)	14 (60.9)	18 (43.9)
No (Hb ≥ 12)	9 (23.1)	9 (39.1)	23 (56.1)
PTH (pmol/L)	20.4 (15.0-33.0)	11.1 (7.5-25.9)	20.4 (11.3-28.2)	0.0634
Calcium (mmol/L)	2.06 (1.90-2.15)	2.03 (1.90-2.31)	2.11 (1.96-2.22)	0.3066
Phosphate (mmol/L)	1.41 (1.28-1.64)	1.50 (1.23-1.96)	1.47 (1.26-1.73)	0.9176
Total protein (g/L)	58.0 (51.0-67.0)	58.0 (51.2-66.0)	64.0 (56.0-75.0)	0.1400
Albumin (g/L)	30.0 (26.0-33.0)	27.5 (23.8-32.0)	31.0 (27.0-37.0)	0.1404
Proteinuria = yes > 3.5 g/day	21.0 (47.7)	10.0 (22.7)	13.0 (29.5)
Proteinuria = no < 3.5 g/day	16.0 (30.2)	12.0 (22.6)	25.0 (47.2)
Proteinuria, median (IQR)	4.6 (2.2-7.8)	3.3 (0.8-7.0)	2.1 (0.9-5.0)	0.1859
Indication for biopsy				0.0243
Sudden renal decline	12 (30.8)	12 (52.2)	24 (58.5)
Proteinuria rise	20 (51.3)	7 (30.4)	7 (17.1)
Systemic markers	2 (5.1)	1 (4.3)	0 (0.0)
Hematuria (Indication)	5 (12.8)	3 (13.0)	10 (24.4)
Hematuria presence				0.2809
Present	27 (69.2)	17 (73.9)	23 (56.1)
Absent	12 (30.8)	6 (26.1)	18 (43.9)
Total glomeruli count	19.0 (12.0-28.0)	18.0 (10.5-32.5)	10.0 (8.0-17.0)	0.0154
Global sclerosis (count)	5.0 (2.0-11.0)	4.0 (2.2-5.8)	2.0 (1.0-3.2)	0.0007
Global sclerosis ratio	0.3 (0.1-0.6)	0.2 (0.1-0.4)	0.1 (0.1-0.4)	0.0212
Segmental sclerosis (count)	3.0 (0.5-6.0)	1.0 (0.0-5.0)	1.0 (0.0-2.5)	0.0399
Segmental sclerosis ratio	0.1 (0.0-0.2)	0.0 (0.0-0.2)	0.0 (0.0-0.2)	0.6571
Crescent presence				0.0567
Present	0 (0,0)	5 (21.7)	5 (12.2)
Absent	39 (100)	18 (78.3)	36 (87.8)
IFTA presence (≥ 25)				0.0073
Present	16 (41.0)	17 (77.3)	23 (69.7)
Not present	23 (59.0)	5 (22.7)	10 (30.3)
IFTA severity score				0.0019
Absent < 10	5 (12.8)	2 (9.1)	3 (9.1)
Mild 10-24	11 (28.2)	15 (68.2)	20 (60.6)
Moderate 25-50	13 (33.3)	5 (22.7)	10 (30.3)
Severe > 50	10 (25.6)	0 (0.0)	0 (0.0)
Median (IQR)	2.0 (1.0-2.5)	1.0 (1.0-1.0)	1.0 (1.0-2.0)
Arteriosclerosis presence				0.2367
Present	19 (50.0)	9 (42.9)	10 (71.4)
Absent	19 (50.0)	12 (57.1)	4 (28.6)
Arteriosclerosis severity (0-3)				0.2508
None	19 (50.0)	12 (57.1)	4 (28.6)
Mild 25	6 (15.8)	4 (19.0)	7 (50.0)
Moderate 26-50	9 (23.7)	3 (14.3)	2 (14.3)
Severe > 50	4 (10.5)	2 (9.5)	1 (7.1)
Arteriosclerosis median (IQR)	0.5 (0.0-2.0)	0.0 (0.0-1.0)	1.0 (0.2-1.0)
Hyalinosis severity (0-3)				0.3221
None	8 (20.5)	6 (28.6)	4 (22.2)
Mild	15 (38.5)	10 (47.6)	10 (55.6)
Moderate	15 (38.5)	3 (14.3)	4 (22.2)
Severe	1 (2.6)	2 (9.5)	0 (0.0)
Hyalinosis median (IQR)	1.0 (1.0-2.0)	1.0 (0.0-1.0)	1.0 (1.0-1.0)
Podocyte effacement				0.0042
Diffuse > 50	10 (26.3)	6 (30.0)	13 (52.0)
Focal ≤ 50	3 (7.9)	4 (20.0)	9 (36.0)
None	2 (5.3)	1 (5.0)	0 (0.0)
Not reported	23 (60.5)	9 (45.0)	3 (12.0)
Podocyte effacement median (IQR)	4.0 (1.2-4.0)	2.5 (1.0-4.0)	1.0 (1.0-2.0)
ANA (antinuclear antibodies)				0.5140
Positive	5 (13.5)	4 (20.0)	8 (24.2)
Negative	32 (86.5)	16 (80.0)	25 (75.8)
Anti-dsDNA Antibodies				0.3852
Positive	0 (0.0)	1 (12.5)	1 (4.2)
Negative	18 (94.7)	7 (87.5)	23 (95.8)
C3 complement (low vs normal)				0.1198
Low	0 (0.0)	4 (33.3)	7 (19.4)
Normal	11 (100.0)	8 (66.7)	29 (80.6)
C4 complement (low vs normal)				0.6700
Low	1 (4.2)	0 (0.0)	2 (5.6)
Normal	23 (95.8)	14 (100.0)	34 (94.4)
C-ANCA (cytoplasmic ANCA)				0.6378
Positive	0 (0.0)	1 (0.0)	0 (0.0)
Negative	35 (100)	16 (100.0)	15 (100.0)
P-ANCA (perinuclear ANCA)				1.0000
Positive	0 (0.0)	1(6.25)	(0.0)
Negative	35 (100.0)	15 (93.75)	14 (100.0)
Kidney failure (as per study definition)	38 (97.4)	15(65.2)	25 (61.0)	0.0002
Mortality				0.2038
Yes	8 (20.5)	7 (30.4)	5 (12.2)
No	31 (79.5)	16 (69.6)	36 (87.8)

DKD: Diabetic kidney diseases; NDKD: Non-diabetic kidney diseases; IQR: Interquartile range; ACE: Angiotensin converting enzyme; ARBs: Angiotensin II type 1 receptor blockers; SGLT2: Sodium-glucose cotransporter 2; eGFR: Estimated glomerular filtration rate; PTH: Parathyroid hormone; IFTA: Interstitial fibrosis and tubular atrophy; ANA: Antinuclear antibody; ANCA: Antineutrophil cytoplasmic antibodies.

Clinical and histological predictors of NDKD vs DKD

A clinical and laboratory-based logistic regression model was used to differentiate NDKD from DKD. We identified several clinical predictors that were independently associated with NDKD vs DKD. Nephrotic-range proteinuria (> 3.5 g/day) was significantly less common in NDKD [OR = 0.02, 95 % confidence interval (CI): 0.00-0.22, P = 0.0015]. Anemia (hemoglobin < 12 g/dL) was markedly more common in NDKD (OR = 9.56, 95%CI: 1.68-54.30, P = 0.0107). The presence of diabetic retinopathy was strongly associated with DKD and was significantly less common in NDKD (OR = 0.04, 95%CI: 0.00-0.38, P = 0.0067). Similarly, a diabetes duration of ≥ 10 years was rare among NDKD cases (OR = 0.01, 95%CI: 0.00-0.12, P = 0.0002). Furthermore, serum creatinine levels greater than 180 μmol/L were more frequently observed in NDKD (OR = 18.68, 95%CI: 2.26-154.80, P = 0.0063). Although podocyte effacement was more common with NDKD in univariate analysis, logistic regression comparing NDKD to DKD revealed that podocyte effacement was significantly less likely in NDKD (OR = 0.06, 95%CI: 0.01-0.39, P = 0.0033). This suggests that, after adjustment for other factors, podocyte injury remains more characteristic of DKD. The global sclerosis ratio did not differ significantly between groups (OR = 0.13, 95%CI: 0.01-3.21, P = 0.2108). Moderate and severe IFTA showed a borderline association with NDKD (OR = 0.22, 95%CI: 0.04-1.26, P = 0.0893).

When comparing mixed lesions (DKD + NDKD) to pure DKD, most clinical predictors of NDKD were not found to be significantly associated with the presence of mixed pathology. We found that only diabetes duration ≥ 10 years was significantly related to mixed lesions (OR = 0.07, 95%CI: 0.01-0.71, P = 0.0249). This suggests that long-standing diabetes may be prone to developing dual pathology. Moreover, we found that age was a significant predictor (OR = 1.13, 95%CI: 1.04-1.23, P = 0.0055) for the mixed lesions. This implies that older individuals may be more likely to have superimposed non-diabetic changes. Table 4 shows the predictors of NDKD vs DKD, as determined through logistic regression analysis.

Table 4 Multinomial logistic regression (Non-diabetic kidney diseases vs diabetic kidney diseases and mixed vs diabetic kidney diseases).

Group	Predictor	OR (95%CI)	P value	Significance
A clinical + lab-based model
NDKD vs DKD	Proteinuria > 3.5 g/day	0.02 (0.00-0.22)	0.0015	Significant
	Haemoglobin < 12 g/dL	9.56 (1.68-54.30)	0.0107	Significant
	Retinopathy present	0.04 (0.00-0.38)	0.0067	Significant
	DM duration ≥ 10 years	0.01 (0.00-0.12)	0.0002	Significant
	Age	1.01 (0.94-1.08)	0.8127	Not significant
	Creatinine > 180 μmol/L	18.68 (2.26-154.80)	0.0063	Significant
Mixed vs DKD	Proteinuria > 3.5 g/day	0.66 (0.08-5.70)	0.7030	Not significant
	Haemoglobin < 12 g/dL	1.90 (0.31-11.51)	0.4850	Not significant
	Retinopathy present	0.35 (0.05-2.58)	0.3049	Not significant
	DM duration ≥ 10 years	0.07 (0.01-0.71)	0.0249	Significant
	Age	1.13 (1.04-1.23)	0.0055	Significant
	Creatinine > 180 μmol/L	1.48 (0.13-16.50)	0.7483	Not significant
A histopathological model
NDKD vs DKD	Global sclerosis ratio	0.13 (0.01-3.21)	0.2108	Not significant
	Podocyte effacement	0.06 (0.01-0.39)	0.0033	Significant
	IFTA moderate to severe	0.22 (0.04-1.26)	0.0893	Borderline
Mixed vs DKD	Global sclerosis ratio	0.38 (0.04-3.73)	0.4048	Not significant
	Podocyte effacement	0.30 (0.08-1.19)	0.0870	Borderline
	IFTA moderate to severe	0.19 (0.05-0.77)	0.02	Significant

OR: Odds ratios; CI: Confidence interval; DKD: Diabetic kidney diseases; NDKD: Non-diabetic kidney diseases; DM: Diabetes mellitus; IFTA: Interstitial fibrosis and tubular atrophy.

Renal survival analysis

Kaplan-Meier analysis identified significant differences in renal survival across the three groups (log-rank P = 0.0033). We compared Kaplan-Meier curves for DKD, mixed lesions, and NDKD, assessing time to kidney failure using eGFR at the time of biopsy as the starting point and kidney failure as the endpoint. Patients with NDKD showed the most favorable renal outcomes, with slower progression to kidney failure. The curve for DKD demonstrated the steepest decline in kidney function over time, suggesting that DKD carries the highest risk of progression. The mixed group (DKD + NDKD) exhibited intermediate survival, possibly indicating that treatment of superimposed pathology may slightly modify the natural course of diabetic kidney disease. Figure 3A shows Kaplan-Meier survival curves comparing time to kidney failure across the three groups.

Figure 3 Kaplan-Meier curves. A: Kaplan-Meier curves compare the time to kidney failure across three biopsy-based groups. The log-rank test P-value = 0.0033, indicating a statistically significant difference in kidney failure risk between the groups; B: Shows Kaplan-Meier survival curves comparing time to death between patients with diabetic kidney diseases, non-diabetic kidney diseases, and mixed lesions. Shaded areas show the 95% confidence intervals, representing uncertainty in survival estimates. The log-rank test P-value = 0.018, indicating a statistically significant difference in mortality across the groups. DKD: Diabetic kidney diseases; NDKD: Non-diabetic kidney diseases.

We performed Cox regression analysis to identify predictors of kidney failure. Kidney failure was defined as a composite outcome of sustained eGFR < 20 mL/minute/1.73 m², initiation of dialysis, or receipt of a kidney transplant. A Cox proportional hazards regression model was developed to identify predictors of progression to kidney failure among 72 patients with complete data. The model performed well (C-index = 0.93) and achieved strong statistical significance (log-rank χ² = 106.9, P < 0.001). We observed that increasing age [hazard ratio (HR) 1.04, P = 0.01], higher serum creatinine at biopsy (HR = 1.01, P < 0.005), arteriosclerosis (HR = 1.56, P = 0.01), and moderate-to-severe IFTA (HR = 1.85, P < 0.005) were significantly associated with the development of kidney failure. Use of RAAS blockade was significantly protective (HR = 0.43, P = 0.02). Table 5 shows a multivariable Cox proportional hazards model for time to kidney failure and mortality.

Table 5 The multivariable Cox proportional hazards model (kidney failure and mortality).

Variable	HR	95%CI	P value	Significance
Multivariable Cox proportional hazards model (Kidney failure)
Age	1.04	1.01-1.07	0.01	Significant
Creatinine	1.01	1.01-1.01	< 0.005	Significant
RAS blocker (ACEi/ARBs)	0.43	0.21-0.89	0.02	Protective factor
Arteriosclerosis	1.56	1.11-2.19	0.01	Significant
IFTA	1.85	1.26-2.72	< 0.005	Significant
Retinopathy	1.60	0.79-3.25	0.19	Not significant
Multivariable Cox proportional hazards model (Mortality)
Age	1.05	1.00-1.09	0.04	Significant
Creatinine	1.01	1.00-1.01	< 0.005	Significant
Hemoglobin < 12 g/dL	2.25	0.66-7.72	0.20	Not significant
Albumin < 30 g/L	4.26	1.09-16.61	0.04	Significant
Retinopathy	3.66	1.05-12.72	0.04	Significant
Arteriosclerosis	2.47	1.36-4.50	< 0.005	Significant

HR: Hazard ratio; CI: Confidence interval; RAS: Renin-angiotensin system; ACEi: Angiotensin converting enzyme inhibitors; ARBs: Angiotensin II type 1 receptor blockers; IFTA: Interstitial fibrosis and tubular atrophy.

Mortality analysis and patient survival

The Kaplan-Meier survival plot compared the time to death among three histologically defined groups: DKD, mixed, and NDKD, using eGFR at the time of biopsy as the time-to-event variable and mortality as the event. The log-rank test yielded a P-value of 0.018, indicating a statistically significant difference in mortality across the groups. Patients in the NDKD group exhibited the highest survival probability, indicating a lower risk of mortality over time. On the other side, the mixed group showed the worst survival outcomes, suggesting a higher mortality risk. The DKD group had an intermediate survival pattern, falling between the other two groups. These results highlight the prognostic value of biopsy-based classification. The presence of both diabetic and non-diabetic lesions may exhibit worse outcomes than either condition alone. Figure 3B shows the Kaplan-Meier survival plot comparing the time to death among three histologically defined groups using eGFR at the time of biopsy as the time-to-event variable and mortality as the event.

A multivariable Cox regression model was used to determine predictors of mortality. We found that older age (HR = 1.05, 95%CI: 1.00-1.09, P = 0.04), hypoalbuminemia (albumin < 30 g/L; HR = 4.26, 95%CI: 1.09-16.61, P = 0.04), diabetic retinopathy (HR = 3.66, 95%CI: 1.05-12.72, P = 0.04), and arteriosclerosis (HR = 2.47, 95%CI: 1.36-4.50, P < 0.005) were independently associated with increased mortality risk. Higher serum creatinine also significantly predicted mortality (HR = 1.01, 95%CI: 1.00-1.01, P < 0.005). Table 5 shows predictors of mortality in patients with diabetes who underwent a kidney biopsy.

DISCUSSION

We report the first detailed study based on kidney biopsy from Saudi Arabia. In this retrospective analysis, we found several pertinent findings. These findings add to the understanding of the different spectrum of kidney disease in diabetic patients. NDKD, either alone or mixed with diabetic lesions, was more common than DKD alone. Distinct clinical features helped to differentiate NDKD from DKD. These features included less proteinuria, shorter diabetes duration, and absence of retinopathy in cases of NDKD. NDKD was also associated with better renal and patient survival. Conversely, DKD and mixed pathologies were linked to worse outcomes, with histological and clinical factors such as arteriosclerosis and elevated creatinine independently predicting kidney failure and mortality.

Prior to our study, limited research had been conducted in Saudi Arabia. These studies were primarily descriptive, focusing solely on histopathologies with a small sample size[12,13]. The first study analyzed 16 biopsies and found NDKD in 50% of diabetic patients, with membranous glomerulonephritis and interstitial nephritis being most common[12]. Another study reported 10 patients with mixed lesions, identifying IgA nephropathy in five (50%), interstitial nephritis in two (20%), membranoproliferative glomerulonephritis in two (20%), and membranous nephropathy in one (10%)[13]. These studies provided important descriptive data but were limited by small sample sizes and purely descriptive designs. None explored predictors, outcomes, or comparative analysis between DKD and NDKD. Our study addresses these gaps by integrating clinical, pathological, and outcome data to provide a more comprehensive understanding. We searched for predictors of DKD that could differentiate it from NDKD and evaluated long-term outcomes using Cox regression and Kaplan-Meier survival analysis, approaches not previously addressed.

We found that NDKD alone or in combination with DKD was more common than isolated DKD. NDKD was seen in 39.8%, NDKD and DKD together in 22.3%, and DKD in 37.9%. FSGS was the predominant pathology (20.3%), followed by interstitial nephritis (18.2%), emphasizing the importance of kidney biopsy for accurate diagnosis. Clinicians should have a low threshold to biopsy diabetic patients with atypical presentations. Identifying the correct etiology of NDKD and providing appropriate therapy, along with optimal diabetes control, may improve outcomes. A narrative review reported that the prevalence of isolated NDKD ranges from 0 to 68.6% (average 40.6%) and diabetic nephropathy plus NDKD ranges from 0 to 45.5% (average 18.1%)[14]. There is also variability in predominant histopathological diagnoses. We found FSGS most common (20.3%), followed by interstitial nephritis (18.2%). Other studies similarly reported FSGS[15-17], while some found membranous[5] or acute tubular interstitial nephritis[18] as predominant. Such variability likely reflects differences in patient selection, biopsy indications, epidemiology, and diagnostic practices.

We identified several independent predictors distinguishing NDKD from DKD. Shorter diabetes duration (< 10 years) and absence of diabetic retinopathy were significantly associated with NDKD. Our findings align with prior literature showing that diabetic retinopathy indicates DKD, while its absence suggests NDKD[19-22]. However, biopsy-proven NDKD can still have diabetic retinopathy; in our cohort, 24.4% of NDKD had it. Conversely, only 59% of DKD and 39.1% of mixed lesions had retinopathy. Thus, 41% of DKD, 60.9% of mixed, and 75.6% of NDKD lacked retinopathy. The prevalence of retinopathy is higher in DKD, though 23.6% of biopsy-proven DKD may not have it[23]. Hence, retinopathy is predictive but not exclusive of DKD. Additionally, diabetes duration ≥ 10 years was uncommon in NDKD, suggesting that shorter duration indicates NDKD. Similar findings were reported in prior studies[18,22,24,25].

Nephrotic-range proteinuria was less common in NDKD compared to DKD, and this association remained robust in logistic regression. Prior studies also found higher proteinuria in DKD[26-28]. Liu et al[26] reported nephrotic-range proteinuria in 63.2% of DKD vs 42.3% in NDKD, while Lee et al[27] found 62.8% in DKD vs 24% in NDKD. In our study, median proteinuria was highest in DKD (4.6 g/day), followed by mixed (3.3 g/day), and lowest in NDKD (2.1 g/day). In practice, clinicians often biopsy diabetic patients with worsening renal function or high proteinuria to rule out NDKD. Although high proteinuria suggests NDKD, our findings indicate that marked proteinuria may still represent DKD. In such cases, additional features such as shorter diabetes duration, absence of retinopathy, or atypical findings (e.g., active urinary sediment or sudden nephrotic syndrome) should prompt consideration of NDKD. A thorough evaluation of these factors can guide diagnostic and management decisions.

Diabetic patients are more prone to anemia due to chronic inflammation, reduced erythropoietin production, and impaired oxygen sensing[29]. In univariate analysis, anemia was more severe in DKD, but multivariable analysis showed hemoglobin < 12 g/dL was independently associated with higher odds of NDKD. This discrepancy may result from confounding factors and differences in analyzing hemoglobin as continuous vs binary. Previous studies reported mixed findings; one found anemia, reduced RBC count, and low hemoglobin associated with NDKD on univariate analysis, but only RBC count remained significant on multivariate analysis[30], while another identified anemia as a DKD predictor[31]. These mixed results highlight the complex relationship between anemia and kidney pathology in diabetes and should be interpreted cautiously.

In multivariable regression, serum creatinine > 180 μmol/L was independently associated with higher odds of NDKD compared to DKD. This contrasts with another study identifying creatinine > 97 μmol/L as predictive of DKD[32]. The difference may reflect variations in patient populations and geography. Higher baseline creatinine in our NDKD group may relate to the diverse pathologies - interstitial nephritis, immune complex glomerulonephritis, FSGS, and lupus nephritis - often causing acute or rapidly progressive injury. Also, biopsies were more often performed in the NDKD group due to sudden renal decline (58.5% vs 30.8%), explaining higher creatinine levels. These findings suggest that, especially in populations with higher NDKD prevalence, elevated creatinine should not preclude biopsy, emphasizing individualized assessment over fixed thresholds.

Kaplan-Meier survival analysis for kidney failure revealed significant differences (log-rank P = 0.0033). DKD patients had the poorest renal survival and fastest progression, NDKD the best, and mixed lesions intermediate. Similar findings were reported in the CKD-Research of Outcomes in Treatment and Epidemiology cohort, where DKD patients had a 50% higher risk of eGFR decline (HR = 2.30) and kidney replacement therapy (HR = 1.64) compared to NDKD, with faster progression[33]. This indicates that diabetic nephropathy follows a more aggressive course than many non-diabetic renal pathologies, highlighting the prognostic significance of biopsy-based classification.

Predictors of kidney failure in our cohort included older age and higher serum creatinine, both independently associated with faster progression. Use of renin-angiotensin system (RAS) blockers was significantly protective. Histologically, arteriosclerosis and moderate-to-severe IFTA were strong predictors. These associations are biologically plausible; aging reduces nephron number and reserve, while comorbidities and altered immune responses increase vulnerability[34]. Higher creatinine may reflect advanced disease or concurrent NDKD. RAS blockers protect by lowering glomerular pressure and limiting fibrosis, with well-established efficacy in reducing proteinuria[35], slowing CKD progression, and reducing mortality[36,37].

Our cohort also showed significant mortality differences on Kaplan-Meier analysis. NDKD patients had the highest survival, mixed lesions the worst, and DKD intermediate. These results suggest histopathological diagnosis impacts both renal and patient survival. A Korean study similarly reported best survival in NDKD, worst in mixed, and intermediate in DKD[38]. Jin Kim Y et al[39] also found better survival in NDKD and intermediate outcomes in mixed pathology. Cox regression showed that increasing age, hypoalbuminemia, diabetic retinopathy, arteriosclerosis, and higher serum creatinine independently predicted mortality. Mortality in our cohort was multifactorial, reflecting poor physiological reserve with aging, malnutrition, inflammation, vascular damage, and impaired kidney function. A Spanish study found older age, peripheral vascular disease, higher creatinine, and DKD as mortality risk factors[40]. These findings underscore the importance of comprehensive clinical and pathological assessments to identify high-risk patients and plan interventions.

Our study has several strengths. It represents the first comprehensive biopsy-based analysis of diabetic patients with kidney disease from Saudi Arabia. Earlier local studies had small sample sizes and lacked clinical correlation. Our study integrated histopathological findings with clinical predictors and long-term outcomes using robust statistical models, including logistic regression and Cox analysis, enabling differentiation between DKD and NDKD and identification of predictors of kidney failure and mortality.

We acknowledge limitations. Being a single-center retrospective analysis, findings cannot be generalized and may be prone to referral bias. Since biopsies were performed based on clinical suspicion, selection bias toward atypical presentations is possible. Despite these limitations, our cohort provides valuable insights into the clinical-pathological presentation of DKD and NDKD in this underrepresented population. We recommend that future studies aim to validate our findings in larger, multicenter cohorts to improve generalizability. Future studies should also explore the utility of non-invasive biomarkers in differentiating DKD from NDKD, potentially reducing the need for biopsy. Finally, studies evaluating the impact of histology-guided treatment strategies on long-term renal and patient survival are needed.

CONCLUSION

NDKD and mixed lesions are common in patients with diabetes and often coexist with DKD. A shorter diabetes duration, absence of diabetic retinopathy, anemia, and elevated serum creatinine should prompt the clinician to consider NDKD and the need for a kidney biopsy. NDKD is associated with better renal survival and a lower risk of progression to kidney failure compared to DKD. Factors such as older age, higher serum creatinine at biopsy, arteriosclerosis, and moderate-to-severe interstitial fibrotic arteriopathy are independently associated with an increased risk of kidney failure. In contrast, RAS blockade has a protective effect. In terms of mortality, patients with NDKD had the best survival, while those with mixed lesions had the worst outcomes. Independent predictors of mortality included increasing age, hypoalbuminemia, diabetic retinopathy, arteriosclerosis, and elevated serum creatinine.