Clinicopathologic features and outcomes of patients with DNAJB9-positive fibrillary glomerulonephritis Authors: Ozge Hurdogan, Safak Mirioglu, Egemen Cebeci, Ahmet Burak Dirim, Sidar Bagbudar, Cagla Safak Karaoglan, Ayse Serra Artan, Ozgur Akin Oto, Isin Kilicaslan, Savas Ozturk, Halil Yazici, Yasemin Ozluk Published: 2026 ABSTRACT ABSTRACTBackgroundFibrillary glomerulonephritis (FGN) is a very rare glomerular disease characterized by non-branching fibril deposition and DNAJB9 positivity. We aimed to analyze our cohort to better elucidate clinicopathologic features and outcomes of patients with FGN.MethodsAdult patients diagnosed with FGN between 2007 and 2025 and showing DNAJB9 positivity were included. Primary composite outcome was defined as doubling of serum creatinine from baseline, undergoing dialysis or transplantation, development of stage 5 chronic kidney disease or death.ResultsFifty native kidney biopsies of 44 patients were examined; 17 belonged to males (34%). Most common patterns of injury were mesangial (24, 48%) and membranoproliferative (16, 32%). Mean fibril diameter was 13.7 ± 2.5 nm. Eight biopsies (16%) showed atypical variants such as congophilia, light-chain restriction on frozen immunoflorescence and immunoglobulin G (IgG) negativity. Out of 44 patients, 25 (56.8%) had adequate follow-up data; the mean age was 48.4 ± 12.7 years. Over a median of 27 (7.5–89.5) months, 13 patients (52%) reached primary composite outcome, 10 (40%) underwent kidney replacement therapies and 3 (12%) died. Multivariate logistic regression models showed hemoglobin predicted the primary outcome whereas histopathological features or immunosuppressive use did not. Baseline serum C3 was associated with primary outcome (area under the curve 0.759, 0.525–0.916) with 134 mg/dL as cut-off value. The kidney survival rate was 30.8% in patients with serum C3 ≤134 mg/dL and 87.5% in serum C3 >134 mg/dL. However, multivariate regression models failed to show a clear association between serum C3 levels and primary outcome.ConclusionsAtypical histopathological features are not uncommon in FGN, complicating the differential diagnosis. Prognosis is still quite dismal despite immunosuppressive use. Lower baseline hemoglobin might indicate poorer outcomes.Keywords: C3, DNAJB9, electron microscopy, fibrillary glomerulonephritis, immunosuppression What was known: Fibrillary glomerulonephritis (FGN) is a very rare glomerular disease with poor prognosis, characterized by non-branching fibrils and DNAJB9 positivity. Fibrillary glomerulonephritis (FGN) is a very rare glomerular disease with poor prognosis, characterized by non-branching fibrils and DNAJB9 positivity. Atypical histopathological features complicate the differential diagnosis. Atypical histopathological features complicate the differential diagnosis. Immunosuppressive therapies have shown inconsistent benefits in slowing disease progression. Immunosuppressive therapies have shown inconsistent benefits in slowing disease progression. This study adds: Atypical histopathological features such as congophilia, light-chain restriction and immunoglobulin G negativity were present in 16% of DNAJB9-positive FGN cases. Atypical histopathological features such as congophilia, light-chain restriction and immunoglobulin G negativity were present in 16% of DNAJB9-positive FGN cases. Baseline hemoglobin level independently predicted adverse outcomes whereas histopathological features and immunosuppressive use did not. Baseline hemoglobin level independently predicted adverse outcomes whereas histopathological features and immunosuppressive use did not. Patients with lower baseline serum C3 levels presented with more aggressive disease and suffered from poorer kidney survival; however, multivariate regression models failed to show a clear association. Patients with lower baseline serum C3 levels presented with more aggressive disease and suffered from poorer kidney survival; however, multivariate regression models failed to show a clear association. Potential impact: This study highlights the diagnostic heterogeneity of FGN. This study highlights the diagnostic heterogeneity of FGN. Baseline hemoglobin might be a prognostic marker to improve risk stratification, and serum C3 levels at diagnosis may be associated with more severe disease. Baseline hemoglobin might be a prognostic marker to improve risk stratification, and serum C3 levels at diagnosis may be associated with more severe disease. Efficacy of current immunosuppressive strategies is limited, underscoring the need for novel therapeutic approaches. Efficacy of current immunosuppressive strategies is limited, underscoring the need for novel therapeutic approaches. INTRODUCTION Fibrillary glomerulonephritis (FGN) is first described as a cause of nephrotic syndrome with amyloid-like, non-branching, randomly oriented fibrils [1]. On light microscopic examination, glomeruli show various patterns including mesangial proliferative, membranoproliferative, membranous and diffuse sclerosing, with mesangial pattern being the most prevalent. Immunofluorescence examination reveals smudgy deposits for immunoglobulin G (IgG), C3, kappa and lambda. IgG deposition is frequently dominated by IgG1 and IgG4 subclasses. Kappa and lambda light chains are predominantly polytypic, with light-chain restriction observed in about 5% of cases [2]. Ultrastructural examination has been essential for the diagnosis of FGN for decades. The deposits consist of non-branching, randomly oriented fibrils measuring 15–25 nm in diameter (mean ∼20 nm) on electron microscopy. They are typically located within the mesangium but may also extend into the glomerular basement membrane and, less commonly, exhibit subendothelial or subepithelial distribution [2]. Congophilic variants of FGN and IgG-negative FGN have also been described [3, 4]. A major advancement in the understanding of FGN was achieved in 2018, when DnaJ homolog subfamily B member 9 (DNAJB9) was identified as a highly enriched protein within FGN deposits via laser capture microdissection of glomeruli followed by tandem mass spectrometry. DNAJB9 was also demonstrated to co-localize with IgG and components of the classical complement pathway in FGN, while being absent in amyloidosis, other glomerular diseases and normal controls [5, 6]. DNAJB9 immunostaining has emerged as a highly reliable diagnostic marker for FGN, demonstrating exceptionally high sensitivity and specificity in the largest cohorts studied and is currently used in routine diagnosis of FGN. Whether it is an autoantigen or a protein that binds to misfolded immunoglobulins is not known [7]. Patients with FGN can present with nephrotic syndrome or nephrotic-range proteinuria, hematuria and kidney injury [2, 8]. FGN can be associated with autoimmune diseases, diabetes mellitus and malignancies [2, 8]. The prognosis is quite dismal, progressing to kidney failure in approximately half of the patients in 5 years [8]. Recurrent disease after transplantation is generally indolent, most probably associated with the background immunosuppressive treatment regimen [9]. FGN is a very rare entity, and several variants—such as Congo-positive, monotypic and immunoglobulin-negative cases—have been described, further complicating its diagnosis. Due to its rare nature, there have been few studies originating from large cohorts [8, 10–12]. In this study, we aimed to examine the clinicopathologic characteristics of FGN within our own cohort, including cases exhibiting these atypical features. Additionally, we aimed to evaluate the correlation between histopathologic features and treatment approaches, clinical outcomes, disease progression and patient survival. MATERIALS AND METHODS Case selection The digital database of our pathology department was searched for FGN cases diagnosed between 2007 and 2025, including both in-house material and outside consultations. Cases diagnosed by electron microscopy, prior to the routine use of DNAJB9 immunostaining, were retrospectively subjected to immunohistochemical examination. Only DNAJB9-positive cases were included in the study. Fifty native kidney biopsies of 44 adult patients were identified, and 25 patients from two centers had adequate follow-up data. Histopathological evaluation The following parameters were documented from the original pathology reports: total number of glomeruli; number of globally/segmentally sclerotic glomeruli; light microscopic pattern; presence of glomerular basement membrane duplication; degree of interstitial fibrosis and tubular atrophy; presence, number and type of crescents; and presence and degree of vascular lesions. Light microscopic patterns were reclassified as mesangial, membranoproliferative, membranous, diffuse proliferative and diffuse sclerosing [2]. The presence of congophilia under polarized light was also assessed and recorded. Immunofluorescence findings were reviewed for the presence and staining intensity of IgG, IgM, IgA, C3, C1q, kappa and lambda light chains. The presence of light-chain restriction was determined based on immunofluorescence findings. Cases with light-chain restriction on frozen immunofluorescence were subjected to paraffin immunofluorescence if adequate tissue was present. The formalin-fixed paraffin-embedded blocks from biopsies lacking DNAJB9 immunohistochemical staining were retrieved from the archives. For each biopsy, the tissue block containing renal cortex with higher number of glomeruli was selected, and 3-µm-thick sections were prepared for staining. Following deparaffinization and rehydration, antigen retrieval was performed with EDTA for 60 min at 37°C, using anti-DNAJB9–human specific rabbit antibody (polyclonal), Sigma-Aldrich®. The slides were processed in the Ventana Medical System-Benchmark XT IHC/ISH Staining module. The slides were evaluated as positive or negative by a nephropathologist. Ultrastructural findings were obtained from the reports of electron microscopic biopsies that had previously been examined using a JEOL® 1011 transmission electron microscope. When additional evaluation was required, new 80-nm-thick grids were prepared with a Reichert® Om U3 ultramicrotome and stained with uranyl acetate and lead citrate. The localization of fibrillary deposits and the diameters of the fibrils were documented. Measurements of fibril diameter were performed at magnifications of 50 000–60 000×. The mean fibril diameter was calculated for each case. Clinical data and outcomes Hypertension was defined as systolic blood pressure (BP) ≥140 mmHg or diastolic BP ≥90 mmHg or using antihypertensive agents. Urinary protein-to-creatinine ratio in the first morning specimens was used to estimate proteinuria throughout the follow-up, and 24-h urine collection was selectively used to solve any discrepancies arising from the spot urine assessment. Nephrotic-range proteinuria was defined as proteinuria level of ≥3 g/day in the absence of nephrotic syndrome. Estimated glomerular filtration rates (eGFRs) were calculated by using the Chronic Kidney Disease Epidemiology Collaboration 2009 formula [13]. Decision to use angiotensin-converting enzyme inhibitors, angiotensin-receptor blockers, sodium-glucose cotransporter-2 inhibitors and immunosuppressives, and selection of the agents were at the discretion of the treating nephrologist at two centers. Primary composite outcome was defined as doubling of serum creatinine from baseline, undergoing dialysis or transplantation, development of stage 5 chronic kidney disease (CKD) (eGFR <15 mL/min/1.73 m2) or death. A composite outcome was used to increase event numbers and statistical power in this rare disease cohort, consistent with prior studies evaluating kidney outcomes [14, 15]. Secondary outcome was complete or partial remission (CR or PR). CR was determined to be achieved when proteinuria decreased to ≤0.3 g/day with normal serum albumin and creatinine concentrations, while PR was considered as a proteinuria reduction of ≥50% (and a proteinuria value of <3 g/day in patients with nephrotic syndrome or nephrotic-range proteinuria at baseline) and stabilization or improvement in kidney function [16]. Associations of clinical, laboratory and histopathological features with primary composite outcome were also evaluated. Statistical analyses Results were expressed as mean ± standard deviation when normally distributed or as median [interquartile range (IQR)] otherwise. Categorical variables are shown as frequencies (%). Comparisons of continuous variables between the groups were evaluated with t-tests or the Mann–Whitney U test where appropriate. Differences in the proportions of the groups were compared using the chi-squared test. Relationships were determined by Pearson correlation coefficient or Spearman’s rho. Variables with P-values <.05 in univariate analyses were selected for the multivariate logistic regression models with enter method. Sex, age at diagnosis and variables which had been known to affect the outcomes were also included. Results of the regression models are presented as odds ratios (ORs) and 95% confidence intervals (CIs). The receiver operating characteristic (ROC) curve was used and the area under the curve (AUC) value was calculated to assess the performance of variables in discriminating the primary composite outcome. The optimal cut-off value in ROC curve was calculated with Youden’s index. Periods of primary outcome-free kidney survival were analyzed using Kaplan–Meier curves by using log-rank test, and this period for each patient was computed from baseline evaluation to the last follow-up or primary composite outcome. Statistical analyses were performed with SPSS for Windows (SPSS version 25.0, IBM Corp., Armonk, NY, USA) and MedCalc for Windows (MedCalc version 19.0, MedCalc Software, Ostend, Belgium). All analyses were two-sided and a P-value of <.05 was considered as statistically significant. Ethical issues This study was approved by the Istanbul University Istanbul Faculty of Medicine Ethical Committee (2025/1321) and complied with the Declaration of Helsinki and its later amendments. All patients provided informed consent to extract their data to the research database. RESULTS Histopathological characteristics Fifty native kidney biopsies of 44 adult patients were examined. Seventeen belonged to male patients (34%). The median number of glomeruli per biopsy was 19 (IQR 13–25.5). The median percentage of global sclerotic glomeruli was 25% (11.32–42.73). No segmentally sclerotic glomerulus was identified. The light microscopic injury patterns, grades of tubular atrophy, interstitial fibrosis, vascular lesions, immunofluorescence staining patterns, and the intensity of immunoglobulin and complement staining are presented in Table 1. Typical histopathologic features of a patient are depicted in Fig. 1. Table 1:Light and immunofluorescence microscopic features of all kidney biopsies.Characteristics, n (%)Kidney biopsies with FGN diagnosis (n = 50)Light microscopic patternMesangial24 (48)Membranoproliferative16 (32)Membranous4 (8)Diffuse proliferative2 (4)Diffuse sclerosing4 (8)Presence of GBM duplication13 (26)Presence of crescents8 (16)Presence of interstitial fibrosis39 (78)Presence of tubular atrophy43 (86)Presence of arteriosclerosis27 (54) aIF staining patternMesangial7 (14)GBM3 (6)Mesangial and GBM37 (74) aIgG staining ≥2+ on IF40 (85.1) aIgA staining ≥2+ on IF1 (2.1) aIgM staining ≥2+ on IF1 (2.1) aC3 staining ≥2+ on IF38 (80.9) aC1q staining ≥2+ on IF3 (6.4) bKappa staining ≥2+ on IF12 (26.7) bLambda staining ≥2+ on IF18 (38.3)aImmunofluorescence staining was available for 47 biopsies (94%).bKappa and lambda staining was available for 45 biopsies (90%). GBM, glomerular basement membrane; IF, immunofluorescence. Light and immunofluorescence microscopic features of all kidney biopsies. Immunofluorescence staining was available for 47 biopsies (94%). bKappa and lambda staining was available for 45 biopsies (90%). GBM, glomerular basement membrane; IF, immunofluorescence. (a) A glomerulus showing expansion of the mesangial area by fibrillary material (periodic acid–Schiff, ×400). (b) Immunofluorescence staining reveals smudgy IgG deposits in the mesangial area and along the glomerular basement membranes (anti-IgG FITC, ×400). (c) Immunofluorescence staining demonstrates smudgy C3 deposition in the mesangium and along the glomerular basement membranes (anti-C3 FITC, ×400). (d) Positive immunohistochemical staining in both mesangial areas and glomerular basement membranes for anti-DNAJB9 antibody (×400). (e) Electron microscopy shows electron dense deposits on mesangial area and glomerular basement membranes (lead citrate-uranyl acetate, ×3000). (f) Randomly arranged, non-branching fibrils measuring approximately 15–20 nm in diameter (lead citrate-uranyl acetate, ×20 000). Five cases (10%) demonstrated congophilic deposits, and six cases (12%) showed light-chain restriction on frozen immunofluorescence. In congophilic cases, Congo-red positivity was focal and less intense than typically observed in amyloidosis, with only subtle birefringence under polarized light rather than the strong and diffuse pattern characteristic of amyloidosis. One biopsy was negative for IgG by frozen immunofluorescence, as well as showing light-chain restriction in frozen and paraffin immunofluorescence and congophilia. Detailed clinicopathologic features of these atypical cases are summarized in Table 2. The histopathological features of an atypical FGN case with lambda restriction and congophilia are shown in Fig. 2. Table 2:Clinicopathological characteristics of cases with atypical histopathological features.Age (years), sexLight microscopic patternGlobal sclerosis/total glomeruliInterstitial fibrosisTubular atrophyAtypical histopathological featureFrozen IF staining characteristicsMean fibril diameter (nm) and fibril localization on EMClinical presentationTreatment and outcomes43, maleMembranoproliferative5/39NoneMildCongophiliaIgG: +++, IgA: –, IgM: –, C3: ++, C1q: +15; membranous, mesangialNANA53, maleaMembranoproliferative8/20MildMildCongophiliaIgG: +++, IgA: –, IgM: –, C3: ++, C1q: –12; membranous, mesangial, subendothelialSerum creatinine: 2.32 mg/dL, proteinuria: 0.5 g/day, hematuriaTreated with glucocorticoids. Underwent hemodialysis after 6 months of follow-up41, maleMembranous0/9NoneNoneLambda restriction on frozen IFIgG: ++, IgA: –, IgM: –, C3: +, C1q: –NA; membranous, subendothelialSerum creatinine: 1.05 mg/dL, proteinuria: 1.11 g/day, no hematuria. No evidence of monoclonal disorders. History of spondyloarthritisReceived conservative treatment. Last serum creatinine was 1.02 mg/dL and proteinuria was 2.68 g/day after 26 months of follow-up47, femaleDiffuse proliferative5/13MildModerateLambda restriction on frozen IFIgG: +++, IgA: –, IgM: +, C3: +++, C1q: –NASerum creatinine: 5.7 mg/dL, hematuria, no proteinuria. No evidence of monoclonal disordersNo treatment data. Underwent hemodialysis after 48 months of follow-up51, femaleMesangial4/28NoneMildKappa restriction on frozen IFIgG: +++, IgA: –, IgM: –, C3: ++, C1q: +14; membranous, mesangial, subendothelialSerum creatinine: 0.9 mg/dL, proteinuria: 2.4 g/day, hematuria. No evidence of monoclonal disorders. History of type 2 diabetes and autoimmune thyroiditisReceived conservative treatment. Underwent hemodialysis after 128 months of follow-up56, femaleMesangial9/17MildMildKappa restriction on frozen IFIgG: +, IgA: –, IgM: –, C3: –, C1q: –NA; membranous, mesangialDiagnosed with multiple myelomaReceived chemotherapy. Underwent hemodialysis after 77 months of follow-up45, femalebMesangial10/25MildMildLambda restriction of frozen IF, congophiliaIgG: +++, IgA: –, IgM: –, C3: ++, C1q: –NASerum creatinine: 1.12 mg/dL, proteinuria: 1.3 g/day, hematuria. No evidence of monoclonal disorders. History of Crohn’s diseaseNA48, femalecMesangial3/26MildMildIgG negativity, congophilia, lambda restriction on frozen and paraffin IFIgG: –, IgA: –, IgM: –, C3: –, C1q: –10; membranous, subendothelialSerum creatinine: 0.56 mg/dL, proteinuria: 2.5 g/day, no hematuria.IgG kappa was found in the serum and evaluated as an unrelated MGUS after a bone marrow biopsy. History of type 2 diabetes and rheumatoid arthritisReceived conservative treatment. Last serum creatinine was 0.74 mg/dL and proteinuria was 1 g/day after 10 months of follow-upaThis patient has a previous biopsy, referred as a consultation biopsy, showing membranoproliferative pattern and congophilia, lacking immunofluorescence examination.bParaffin immunofluorescence showed no restriction for kappa and lambda.cParaffin immunofluorescence showed IgG negativity and lambda restriction. EM, electron microscopy; IF, immunofluorescence; MGUS, monoclonal gammopathy of undetermined significance; NA, not available. Clinicopathological characteristics of cases with atypical histopathological features. This patient has a previous biopsy, referred as a consultation biopsy, showing membranoproliferative pattern and congophilia, lacking immunofluorescence examination. Paraffin immunofluorescence showed no restriction for kappa and lambda. Paraffin immunofluorescence showed IgG negativity and lambda restriction. EM, electron microscopy; IF, immunofluorescence; MGUS, monoclonal gammopathy of undetermined significance; NA, not available. (a) A glomerulus with expansion of the mesangial area by fibrillary material (hematoxylin and eosin stain, ×400). (b) Immunofluorescence reveals smudgy IgG deposition in the mesangium and glomerular basement membranes (anti-IgG FITC, ×400). (c) Absence of staining for kappa antibody on immunofluorescence (anti-kappa FITC, ×400). (d) Immunofluorescence staining shows monotypic lambda light-chain deposits in the mesangium and glomerular basement membranes (anti-lambda FITC, ×400). (e) Positive immunohistochemical staining in both mesangial areas and glomerular basement membranes for anti-DNAJB9 antibody (×400). (f) Congo red staining reveals apple-green birefringence in the mesangial area (polarized light, ×400). All cases demonstrated positive immunohistochemical staining for DNAJB9, which was a decisive finding supporting FGN diagnosis. Electron microscopic evaluation was available in 37 cases (74%), including all biopsies performed prior to the routine implementation of DNAJB9 staining. Mean fibril diameter on electron microscopy was 13.7 ± 2.5 nm. Localization of fibrils was mesangial and membranous and/or subendothelial in 20 (55.6%), membranous and subendothelial in 10 (27.7%), intramembranous in 3 (8.3%) and subendothelial in 3 (8.3%). Clinical features and outcomes Out of 44 patients, 25 (56.8%) had adequate follow-up data and 9 patients (36%) were male. Mean age was 48.4 ± 12.7 years at diagnosis. Diabetes mellitus was present in 6 patients (24%). Autoimmune or autoinflammatory diseases were also found in 6 (24%). One patient (4%) was diagnosed with multiple myeloma. Median serum creatinine and eGFR levels were 1.2 (0.76–2) mg/dL and 63.1±36 mL/min/1.73 m2, respectively. Median proteinuria was 2.4 (0.92–4.8) g/day and 14 patients (56%) suffered from hematuria (Table 3). Baseline characteristics of patients with (n = 25) and without (n = 19) adequate follow-up data are provided in Supplementary data, Table S1. Patients without follow-up data had higher proteinuria (P = .049) and lower serum albumin (P = .001) levels. Table 3:Baseline clinical features of patients with adequate follow-up data.FeaturesPatients with FGN (n = 25)Male sex, n (%)9 (36)Age, mean ± SD48.4 ± 12.7Diabetes mellitus, n (%)6 (24)Hypertension, n (%)15 (60)Systolic BP, mean ± SD131.9 ± 25.1Diastolic BP, mean ± SD84.1 ± 13.3Autoimmune and inflammatory diseases, n (%)6 (24)Malignancies, n (%)1 (4)Hemoglobin (g/dL), mean ± SD12.1 ± 1.9Serum creatinine (mg/dL), median (IQR)1.2 (0.76–2)eGFR (mL/min/1.73 m2), mean ± SD63.1 ± 36Serum uric acid (mg/dL), mean ± SD6.9 ± 2.12Serum albumin (g/dL), mean ± SD3.8 ± 0.7Proteinuria (g/day), median (IQR)2.4 (0.92–4.8)Nephrotic syndrome or nephrotic-range proteinuria11 (44)Hematuria, n (%)14 (56)Serum C3 (mg/dL) a, mean ± SD131.1 ± 30.9Serum C4 (mg/dL) a, mean ± SD36.7 ± 10aReference values for serum C3 and C4 were 90–180 and 10-40 mg/dL, respectively. SD, standard deviation. Baseline clinical features of patients with adequate follow-up data. Reference values for serum C3 and C4 were 90–180 and 10-40 mg/dL, respectively. SD, standard deviation. Patients were followed for a median duration of 27 (7.5–89.5) months. Treatment data were available for 22 patients and angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers were used in 19 (86.4%). Immunosuppressive agents were prescribed to 13 patients (59.1%), and glucocorticoids and rituximab were the most common choices. Thirteen patients (52%) reached primary composite outcome and 10 (40%) underwent kidney replacement therapies. First replacement modality was hemodialysis in eight, peritoneal dialysis in one and kidney transplantation in one. Three patients (12%) died. Cause of death was alveolar hemorrhage and hemorrhagic shock after vaginal delivery in two patients and unknown in one. Remission was seen in only three patients (two PR and one CR) (Table 4). Table 4:Treatment and outcomes of patients with adequate follow-up data.CharacteristicsPatients with FGN (n = 25)Treatmenta, n (%)ACEi or ARB19 (86.4)SGLT2i4 (18.2)Immunosuppression13 (59.1)Glucocorticoids9 (40.9)Rituximab9 (40.9)Mycophenolic acid analogs6 (27.3)Azathioprine1 (4.5)Intravenous cyclophosphamide1 (4.5)Outcomes, n (%)Primary composite outcome13 (52)Kidney replacement therapies10 (40)Doubling of serum creatinine2 (8)Death3 (12)bPR or CR3 (12)aTreatment data were available for 22 patients.bOne patient died after experiencing doubling of serum creatinine and one patient passed away while treated with hemodialysis. ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; SGLT2i, sodium-glucose cotransporter-2 inhibitor. Treatment and outcomes of patients with adequate follow-up data. Treatment data were available for 22 patients. One patient died after experiencing doubling of serum creatinine and one patient passed away while treated with hemodialysis. ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; SGLT2i, sodium-glucose cotransporter-2 inhibitor. Predictors of primary composite outcome Univariate logistic regression analyses demonstrated that hemoglobin at diagnosis (OR 0.465, 95% CI 0.235–0.919, P = .028), use of immunosuppression (OR 7.875, 95% CI 1.105–56.123, P = .039) and glucocorticoids (OR 7.875, 95% CI 1.105–56.123, P = .039) might be the predictors (Supplementary data, Table S2). Four multivariate logistic regression models were constructed. Models 1 and 2 showed that hemoglobin predicted the primary composite outcome (OR 0.340, 95% CI 0.123–0.938, P = .037 and OR 0.238, 95% CI 0.069–0.825, P = .024) (Table 5). Table 5:Multivariate logistic regression analysis for the primary composite outcome.VariablesOR (95% CIs) P Model 1Male sex0.239 (0.014–4.152).326Age at diagnosis0.926 (0.834–1.027).144Hemoglobin at diagnosis0.340 (0.123–0.938) .037 Use of immunosuppression3.796 (0.354–40.717).270Model 2Male sex6.148 (0.397–95.283).194Age at diagnosis0.906 (0.809–1.014).085Hemoglobin at diagnosis0.238 (0.069–0.825) .024 eGFR at diagnosis1.014 (0.978–1.051).456Model 3Male sex3.937 (0.122–127.258).440Age at diagnosis0.880 (0.730–1.062).182Hemoglobin at diagnosis0.095 (0.008–1.091).059Serum C3 at diagnosis0.998 (0.932–1.068).946Model 4Male sex0.305 (0.028–3.326).330Age at diagnosis0.977 (0.890–1.072).619eGFR at diagnosis1.005 (0.967–1.044).808Serum C3 at diagnosis0.951 (0.895–1.011).105Statistically significant P-values <.05 are shown in bold. Multivariate logistic regression analysis for the primary composite outcome. Statistically significant P-values <.05 are shown in bold. Univariate analyses also revealed that serum C3 might be associated with the primary composite outcome (OR 0.961, 95% CI 0.920–1.004, P = .073). Therefore, a ROC analysis was performed. Serum C3 levels had an AUC value of 0.759 (0.525–0.916) for the discrimination of the primary composite outcome (P = .018). A serum C3 level of ≤134 mg/dL (reference: 90–180 mg/dL) discriminated the outcome with a 90% sensitivity and 63.6% specificity (Fig. 3a). Kaplan–Meier analysis showed that primary outcome-free kidney survival rate was 30.8% in the patients with serum C3 ≤134 mg/dL and 87.5% in patients with serum C3 >134 mg/dL (P = .230) (Fig. 3b). Patients with serum C3 ≤134 mg/dL had a higher percentage of global glomerulosclerosis (P = .012), higher levels of serum creatinine at diagnosis (P = .046), a higher rate of glomerular basement membrane duplication (P = .023) and a lower rate of mesangial pattern of injury (P = .027). Patients with serum C3 ≤134 mg/dL were also characterized by higher use of immunosuppressive agents (P = .001) and poorer primary composite outcome (P = .011) (Supplementary data, Table S3). Serum C3 levels were negatively correlated with the intensity of C3 deposition in the tissue (r= –0.604, P = .005). However, multivariate regression models failed to show a clear association between serum C3 levels at diagnosis with the primary composite outcome (Table 5). (a) A serum C3 level of ≤134 mg/dL discriminated the primary composite outcome with a 90% sensitivity and 63.6% specificity. (b) Kaplan–Meier analysis demonstrated that primary composite outcome-free kidney survival rate was 30.8% in the patients with serum C3 ≤134 mg/dL and 87.5% in patients with serum C3 >134 mg/dL (P = .230 with log-rank test). DISCUSSION This study comprehensively characterizes the histopathological spectrum of 50 kidney biopsies diagnosed as DNAJB9-positive FGN including atypical histopathological variants, and correlates these findings with available clinical features. DNAJB9 is a relatively recent biomarker introduced into routine practice, and earlier case series were primarily defined based on the presence of fibrils identified by electron microscopy. Electron microscopy was available in most of our biopsies, and cases without ultrastructural evaluation were included only when the diagnosis was supported by the characteristic morphologic and immunofluorescence findings together with strong DNAJB9 immunoreactivity. In more recent studies, however, DNAJB9 often has not been systematically evaluated in all cases and tested only in a subset [8, 17, 18], except for one study in which all cases were selected based on DNAJB9 positivity [19] similar to our cohort. In this study, the mesangial pattern was the most common histopathological pattern on light microscopy consistent with most published series [8, 10, 18–21]. There are other studies in literature reporting a higher prevalence of membranoproliferative pattern, albeit less frequently [22, 23]. Given the rarity of this disease and its dismal prognosis, the relationship between histopathological features—particularly the pattern of injury on light microscopy—and clinical course has been a major subject of investigation in many studies. Although it is generally accepted that diffuse sclerosing pattern portends the worst prognosis, whereas mesangial and membranous patterns are associated with more favorable outcomes, most series failed to show such a correlation [7]. In one cohort, a diffuse proliferative pattern was a significant predictor of progression to kidney failure on multivariate analysis [8], whereas in another study, the degree of glomerulosclerosis was a strong predictor of the outcome [10]. Interestingly, in our cohort, neither the light microscopic injury pattern and the percentage of global glomerulosclerosis, nor other chronicity parameters such as tubular atrophy and interstitial fibrosis, showed a significant association with kidney survival. The limited number of cases in some histologic subgroups may have reduced the power of these analyses, and the retrospective study design and potential treatment-related bias should be acknowledged. Taken together, our findings are in line with many prior series and suggest that kidney survival cannot be reliably predicted by a single histopathological parameter. In our cohort, we report cases with atypical histopathological characteristics such as congophilia, light-chain restriction on frozen immunofluorescence and IgG negativity, comprising 16% of all cases in total. Congo-red positivity was identified in three cases (6%), consistent with the wide range (0%–27%) reported in the literature [17–19]. The largest series of Congo red–positive FGN showed no major clinical or pathological differences compared with Congo red–negative FGNs, except for a higher rate of circulating monoclonal proteins, and mass spectrometry analyses confirmed the absence of a full amyloid protein signature, supporting its distinction from true amyloidosis [3]. Consistent with these findings, we did not identify an association between Congo positivity and clinical features or outcomes. Light chain restriction was identified in 10% of our cases, and this frequency also varies across published series [17–19]. However, light-chain restriction on frozen immunofluorescence may not persist on paraffin immunofluorescence and a diagnosis of monoclonal FGN requires confirmation by paraffin immunofluorescence and IgG subclass restriction [7, 24–26]. We could not perform paraffin immunofluorescence in all cases with light-chain restriction due to limited tissue availability. Consistent with prior studies showing that monotypic FGN is usually not associated with dysproteinemia or overt hematolymphoid neoplasia, we identified no evidence of clinical monoclonal disorders in our series, except for a single patient with multiple myeloma. Our cohort also includes one case of IgG negative, congophilic FGN with lambda restriction. Immunoglobulin-negative FGN cases have been described in a limited number of publications [4, 27, 28]; however, no FGN cases have been reported in which IgG negativity is accompanied by concurrent congophilia and light-chain restriction to the best of our knowledge. Progression to kidney failure is quite common in patients with FGN, which has been reported between 25% and 53% in various studies [8, 10, 12, 17, 19]. Accordingly, kidney failure rate was 40% in our cohort. Immunosuppressive treatment was used in 59% of our patients without success. Notably, six of nine patients who were treated with rituximab experienced the primary composite outcome. Retrospective studies suggested a potential benefit of rituximab in FGN [20, 29], and a pilot trial showed that rituximab was associated with stabilization of kidney function but did not reduce proteinuria [30]. Very recently, a combination therapy consisting of rituximab, cyclophosphamide and glucocorticoids resulted in better outcomes in a small cohort [31]. One might argue that B-cell depletion achieved by rituximab may not sufficiently suppress disease activity in FGN, but further studies are needed to test this hypothesis. Several variables have been suggested as predictors of dismal outcomes: age, sex, proteinuria, serum creatinine at baseline [8, 10, 12, 19] and various histopathological features mentioned before. We found that only hemoglobin level at diagnosis was associated with the primary composite outcome. Oxygen delivery to kidneys is reduced in the presence of anemia, and hypoxia may increase fibrosis of the kidneys aggravating hypoxia itself, culminating in a vicious cycle that paves the way to kidney failure [32]. Anemia is a known complication of CKD, but clinical evidence suggests that it may also be a risk factor for disease progression [15, 33]. In addition, anemia has been reported as a prognostic factor in patients with IgA nephropathy and C3 glomerulopathy [15, 34]. Interestingly, we showed that serum C3 levels may have an association with the outcomes of FGN. Patients with a level of ≤134 mg/dL presented with more aggressive disease, and were mostly treated with immunosuppressive agents but suffered from poorer outcomes. Notably, serum C3 levels have been demonstrated as a significant predictor in various glomerular diseases [35–37]. Even though most of our patients presented with 2–3+ C3 deposition in kidneys, we found that lower serum C3 levels were correlated with higher intensity of C3 deposition. Activation of the complement system might be seen as a secondary phenomenon in FGN, possibly triggered by IgG and C3 deposition in the kidneys. To the best of our knowledge, this is the first study suggesting a possible association of serum C3 levels with outcomes in patients with FGN. Nevertheless, this finding should be interpreted with caution since multivariate regression models failed to show a clear association. External validation is also needed. Among the limitations of our study, one is its retrospective design, precluding cause–effect relationships. Second, the number of patients with adequate follow-up data was limited although overall cohort size is sufficiently large. Patients without follow-up data presented with more severe nephrotic features, raising the possibility of survivor bias. In addition, IgG subclassification and paraffin immunofluorescence was not performed in all patients with light-chain restriction, therefore, a definitive diagnosis of monoclonal FGN could not be established in these cases. On the other hand, our study also has important strengths. Diagnosis of FGN was confirmed by DNAJB staining in all biopsies and majority of the cases (74%) were subjected to electron microscopic examination. We also report a notable subset of cases with atypical histopathological features, complicating FGN diagnosis. In conclusion, this study provides a comprehensive clinicopathologic evaluation of a large DNAJB9-positive FGN cohort and highlights the considerable histopathological heterogeneity of the disease, including a notable proportion of atypical features. Our findings demonstrated that histopathological features failed to predict clinical outcomes. Baseline hemoglobin emerged as a possible biomarker discriminating poor prognosis. Serum C3 levels at diagnosis might be associated with more severe disease, but external validation of this finding is required. Notably, immunosuppressive use was not associated with better outcomes. These results underscore the persistent unmet need for reliable prognostic tools and effective therapies for the management of FGN. Supplementary Material ACKNOWLEDGEMENTS This study was approved by the Istanbul University Istanbul Faculty of Medicine Ethical Committee (2025/1321). Contributor Information Ozge Hurdogan, Department of Pathology, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Safak Mirioglu, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey; Department of Immunology, Aziz Sancar Institute of Experimental Medicine, Istanbul University, Istanbul, Turkey. Egemen Cebeci, Division of Nephrology, Istanbul Haseki Training and Research Hospital, Istanbul, Turkey. Ahmet Burak Dirim, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Sidar Bagbudar, Department of Pathology, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Cagla Safak Karaoglan, Department of Pathology, Basaksehir Cam and Sakura City Hospital, Istanbul, Turkey. Ayse Serra Artan, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Ozgur Akin Oto, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Isin Kilicaslan, Department of Pathology, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Savas Ozturk, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Halil Yazici, Division of Nephrology, Department of Internal Medicine, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. Yasemin Ozluk, Department of Pathology, Istanbul Faculty of Medicine, Istanbul University, Istanbul, Turkey. AUTHORS’ CONTRIBUTIONS O.H., S.M. and Y.O. designed the study. All authors participated in data collection. O.H. and S.M. performed the statistical analyses and drafted the manuscript. All authors significantly contributed to the content and reviewed the article. CONFLICT OF INTEREST STATEMENT Outside the submitted work, S.M. received support for meeting registration and travel from Amgen. The remaining authors have no disclosures. FUNDING No funding. DATA AVAILABILITY STATEMENT Deidentified data are available upon reasonable request from the corresponding author. References Associated Data Supplementary Materials Data Availability Statement Deidentified data are available upon reasonable request from the corresponding author.