Bacterial infection-related glomerulonephritis in patients with diabetes

Aim Diabetic patients are prone to infections, thus making them a unique cohort at risk of developing bacterial infection-related glomerulonephritis (IRGN). Methods In total, 1693 adult diabetic patients underwent kidney biopsy between 2005 and 2021 at our tertiary care hospital in South India. Of these, 121 consecutive cases which met criteria of bacterial IRGN were included in this study. Results The mean age of the cohort was 53.1 ± 10.1 years and 83/121 (68.5%) were males. Majority (98.3%) had type 2 diabetes for a median duration of 6 (IQR, 2–12) years. The most common sites of infection were skin (47/121, 38.8%) and urinary tract (15/121, 12.4%). Fifty percent (58/121) of patients had underlying advanced diabetic kidney disease (DKD). Isolated C3 deposits (without immunoglobulin) occurred in 66/121 (54.5%) patients predominantly in advanced DKD patients. IgA-dominant glomerulonephritis occurred in only 9/121 (7.4%) patients. Short-course oral steroid was given to 86/121 (71.1%) patients. Steroid related dysglycemia and immunosuppression related infections occurred in 9/61 (14.8%) and 16/61 (26.2%) patients respectively. Of the 90 patients with follow up details >3 months, 46 (51.1%) progressed to kidney failure over a median period of 0.5 (IQR, 0–7.2) months. Patients diagnosed in the latter half of our study period (2013–2021) were older, less commonly presented with fever, had more pronounced hypocomplementemia and severe renal histology predominantly with a ‘starry sky’ immunofluorescence pattern. Conclusion Superimposed bacterial IRGN on underlying DKD is associated with poor renal outcomes. Use of short course steroid was associated with significant toxicity.


Introduction
Bacterial infection-related glomerulonephritis (IRGN) is an immune complex disease which occurs in response to infections. 1 Diabetic patients are prone to infections making them a unique cohort at risk of IRGN. 2 Superimposed IRGN on underlying diabetic kidney disease (DKD) is associated with poor renal outcomes. 3The role of steroid in diabetic patients with IRGN is controversial. 4In spite of diabetes being the most common risk factor for IRGN, there are no previous studies specifically looking into the clinicopathological characteristics of IRGN in diabetic patients.This retrospective study spanning a period of 17 years reports the demographics, clinical features, profile of infections, histopathological characteristics and outcomes in 121 consecutive diabetic patients with kidney biopsy proven bacterial IRGN at a tertiary care centre in South India.

Study design and population
This was an observational retrospective biopsy registry cohort [Glomerular Research And Clinical Experiments-Infection Related Glomerulonephritis in diabetics (GRACE-IRGN)] study done in the Department of Nephrology at Christian Medical College, Vellore, the single largest not-for-profit tertiary care centre in South India.In total, 1693 adult diabetic patients underwent native kidney biopsies between January 2005 and December 2021 (Figure 1).Of these, 129 consecutive patients fulfilled at least three of the five criteria for diagnosis of bacterial IRGN. 2

1.
Clinical or laboratory evidence of bacterial infection either prior to or at onset of symptoms of glomerulonephritis (GN).
Of the 129 patients, eight were excluded either due to inadequate kidney biopsy tissue (6)  or due to revision of diagnosis at follow-up (2).In total, 121 patients were included in this study (biopsy incidence of 7.1%).This study was approved by the institutional review board and a waiver of consent was given by the ethics committee as the data was deidentified and collected retrospectively.

Baseline assessments and treatment
Data on demographic profile, clinical features, biochemical parameters, histopathology variables, treatment details and follow-up outcomes were retrieved from electronic patient records.Proteinuria was assessed from 24-h timed collection, as is the standard practice at our centre.Estimated glomerular filtration rate (eGFR) was calculated by Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. 5sed on timing of IRGN relative to infection, patients were classified into parainfectious, postinfectious and latent infectious GN.Parainfectious GN was defined as presence of concurrent infection at onset of features of GN. 6 Postinfectious GN was defined as resolution of infection prior to onset of GN. 7 Latent infectious GN was defined as IRGN occurring with no known infective focus.Latent period (in postinfectious GN) was defined as time (in days) between resolution of infection and onset of features of GN. 7 Streptococcal infection was identified by isolation of β hemolytic Streptococcus in culture or by detection of anti-streptolysin or anti-deoxyribonuclease B antibodies in sera.Skin infections were classified as acute when lesions resolved in 2 weeks (e.g., cellulitis, abscess, pyoderma) or chronic when lesions persisted beyond 2 weeks (e.g., chronic non-healing ulcers).Of the 121 patients in our cohort, 49 were diagnosed between 2005 and 2012 ('period 1') and the remaining 72 were diagnosed between 2013 and 2021 ('period 2').
All kidney biopsy specimens were processed for light microscopy and immunofluorescence.Serial 3 μm thick tissue sections were stained with haematoxylin and eosin (H&E), periodic acid schiff reagent, masson's trichrome, and jones methenamine silver for light microscopy.Interstitial fibrosis and tubular atrophy (IFTA) was graded semi-quantitatively based on an estimate of percentage of cortex involved into none (0%), mild (1%-25%), moderate (25%-50%) and severe (>50%).DKD was classified as per Renal Pathology Society classification into four classes. 84-μm cryostat sections were stained with fluorescein-tagged polyclonal rabbit anti-human antisera specific to IgG, IgA, and IgM; complement factors C3 and C1q; and κ and λ light chains (Dako, Carpinteria) for IF studies.Ultrastructural examination (Tecnai T12 Spirit electron microscope) is not routinely done at our centre and was performed only in a minority (6.6%).
The oral steroid protocol at our centre is 2 mg/kg on alternate days for 6-8 weeks followed by a rapid taper over next 6-8 weeks.This was offered to all patients with eGFR <60 mL/min/1.73m 2 with no significant chronicity on kidney biopsy and good glycemic status.Intravenous methyl prednisolone pulse (IVMP) was given to all patients with rapidly progressive glomerulonephritis and/or severe renal histology.

Follow-up procedures and study outcomes
Follow-up clinical and outcome data were collected for each review visit until June 2022.Remission of kidney disease was defined as improvement in eGFR by >25% from kidney biopsy and last visit eGFR >90 mL/min/1.73m 2 or attainment of baseline eGFR.
Stabilisation of kidney disease was defined as change in eGFR not exceeding 25% from kidney biopsy and last visit eGFR >15 mL/min/1.73m 2 .Worsening of kidney disease was defined as decline in eGFR by >25% from kidney biopsy and last visit eGFR >15 mL/min/1.73m 2 .Kidney failure was defined as last visit eGFR ≤15 mL/min/1.73m 2 or initiation of renal replacement therapy.Complete remission (CR) of proteinuria was defined as proteinuria <0.5 g/day and serum albumin >3.5 g/dL at follow-up.Partial remission (PR) was defined as reduction in proteinuria by at least 50% and to 0.5-3.5 g/day.No remission (NR) was defined as those who failed to attain either CR or PR. 3 Results

Classification based on time period of IRGN diagnosis
Patients diagnosed in 'period 2' were older (55.1 ± 10.3 vs. 50.1 ± 9.2, p = .007),less commonly presented with fever (20.8% vs. 34.7%,p = .098)and had more pronounced hypocomplementemia (low C3: 88.4% vs. 65.9%,p = .005)(Table 4).There were no significant differences noted in sites of infection, infectious agents isolated, drug resistant organisms and latency in the two time periods.Histologically, patients diagnosed in 'period In spite of these differences, baseline proteinuria, eGFR and renal outcomes were comparable between the two periods.

Discussion
The burden of IRGN in diabetic population is unknown (biopsy incidence is 7.1% in our cohort).To the best of our knowledge, no previous study has addressed this issue in spite of diabetes being the most common predisposing factor, especially in developed countries [10%-44% vs. 3%-21% (developing countries)].

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts the most common causes of non-diabetic kidney disease in developing countries [8%-21% (developing countries) vs. 0%-4% (developed countries)] 17-26 (Table S7).In addition, in most IRGN cohorts underlying diabetes is a risk factor for kidney failure. 3,27,28jority (88%) of our cohort had underling diabetic kidney disease with almost half of them having advanced DKD.Similar findings were noted in previous studies.In a small series of 18 diabetic patients with IRGN, 16 (88.9%)had underlying DKD. 29In an elderly cohort of IRGN patients, 43 of 53 (81%) diabetic patients had underlying DKD and 38% had advanced DKD. 30 In a large biopsy cohort of type 2 diabetic patients, Mazzucco et al. also observed that 70% cases of IRGN occurred in patients with underlying diabetic kidney disease. 31Patients with advanced DKD had a higher incidence of micro and macrovascular complications in our study.This predisposes them to develop skin ulcerations with superimposed bacterial infections.
Nasr et al. reported that patients with superimposed IRGN on DKD had histologically more mesangial and subendothelial deposits and less subepithelial deposits.They postulated this to be due to mesangial sclerosis causing inadequate mesangial clearing and glomerular capillary wall thickening impeding formation of large subepithelial humps. 11,32This could not be confirmed in our study as electron microscopy was done in only a minority of patients.However, we observed that isolated C3 deposits were more commonly seen in advanced DKD patients.In a biopsy series of 217 diabetic patients where DKD was the only glomerular disease diagnosed, 53.9% had C3 deposits primarily within mesangium.
Patients with isolated C3 deposits in this study had lower C3 level, lower serum albumin, more advanced DKD and poor renal outcomes. 33lf of our cohort had an ongoing infection at onset of GN.Staphylococcal, gram-negative and drug resistant organisms were more commonly isolated from this parainfectious group.This was similar to findings seen in a pooled analysis of 83 cases of staphylococcal infection related glomerulonephritis wherein all patients had an ongoing infection at onset of GN.Two-thirds of this cohort had MRSA infection. 27oury et al. postulated that dysglycemia is associated with more aggressive histological lesions of IRGN based on animal studies. 34However, we did not observe any difference in renal histology nor outcomes in patients with and without dysglycemia at presentation.Nasr et al. reported IgA-dominant IRGN as the predominant histopathological pattern in diabetic patients (50%-75%). 32Patients with diabetic kidney disease have increased serum IgA levels as well as IgA containing immune complexes.This may be due to more frequent subclinical mucosal infections and impaired IgA clearance.However, IgA dominant IRGN occurred in only 7% of our cohort.This may be explained by differences in infectious agents causing IRGN in our cohort (Streptococcus pyogenes and gram-negative infections) as compared to western population (Staphylococcus aureus). 11,32tients diagnosed in the latter half of our study period (2013-2021) were older, less commonly presented with fever, had more pronounced hypocomplementemia and severe renal histology predominantly with a 'starry sky' IF pattern.'Starry sky' pattern (immunoglobulin and C3) is seen in the early phase of IRGN which over a course of time progresses to 'mesangial' pattern pre-dominantly composed of isolated C3 deposits.We observed a decline in 'garland' pattern over the years as reported previously. 3,10,11However, the sites of infection, infectious agents isolated, drug resistant organism, latency and renal outcomes did not differ between the two time periods.Moroni et al. classified cases of IRGN based on the site of infection into typical infections (pharyngitis and skin infections) and atypical infections (other sites of infection).Over the years, the authors observed a significant increase in atypical infections. 10Similarly, a study from Taiwan reported an increase in staphylococcal infections from 50% to 71.4% and decrease in streptococcal infections from 21.4% to 14.3% in IRGN patients diagnosed in time periods 2000-2004 and 2005-2009, respectively. 35fty-one percent of patients progressed to kidney failure in our cohort.The most significant predictor of kidney failure in our cohort was moderate to severe interstitial fibrosis, which was similar to findings in the French Nationwide Study in IRGN patients with IgA deposits. 13Tubulointerstitial fibrosis is mediated by interstitial myofibroblasts which are alpha-smooth muscle actin (α-SMA) positive cells.Oda et al. analysed expression of α-SMA immunohistochemically in kidney biopsy tissue of six patients of post-streptococcal glomerulonephritis at both acute and convalescent phase (3 had complete remission and 3 progressed to CKD).Glomerular α-SMA staining was found to be transient and reversible due to activated mesangial cells, whereas, interstitial α-SMA expression was associated with progression to CKD. 28 As in other glomerular diseases, the progression of IRGN to kidney failure correlates more with interstitial than glomerular lesions.
Majority of our cohort (71%) received treatment with steroid.However, it did not improve renal outcomes.Twenty-six percent of patients treated with steroid developed infections in the follow-up period which was more seen in patients with dysglycemia at kidney biopsy.In a randomised controlled trial comparing efficacy of corticosteroid versus supportive care in 52 patients of adult IRGN, Arivazhagan et al. observed that immunosuppression related adverse events were significantly more in the treatment arm with no difference in renal outcomes at 6 months.There were seven diabetics included in this study (2 received steroid and 5 supportive care).This study however excluded patients with advanced DKD and a high dose steroid protocol was used, unlike in our study. 4 conclusion, bacterial IRGN is an important cause of non-diabetic kidney disease in diabetic patients in developing countries.Over the years, we observe a change in clinical and histopathological presentations of IRGN.Use of steroid is associated with significant toxicity.Renal prognosis is guarded in patients with underlying DKD.The future directions would be incorporation of programs for early diagnosis and aggressive treatment of infections in diabetic patients as an important strategy for primary prevention of chronic kidney disease.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.

Summary at a glance
In this retrospective study spanning a period of 17 years, we report the clinicopathological outcomes in 121 consecutive diabetic patients with biopsy proven bacterial infection-related glomerulonephritis (IRGN).Superimposed bacterial IRGN on underlying diabetic kidney disease is associated with poor renal outcomes.Use of short course steroid was associated with significant toxicity.

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts

Europe PMC Funders Author Manuscripts
Europe PMC Funders Author Manuscripts Predictors of kidney failure by Kaplan-Meier survival analysis.CKD-EPI, chronic kidney disease epidemiology collaboration; DKD, diabetic kidney disease; eGFR, estimated glomerular filtration rate; GN, glomerulonephritis; IFTA, interstitial fibrosis and tubular atrophy.Baseline characteristics of GRACE-IRGN cohort at kidney biopsy.
was presented as mean ± standard deviation (SD) or median [interquartile range (IQR)] or frequency (percentage) according to the types and distribution of variables.Differences among groups of normally distributed variables were analysed by t test or one-way ANOVA.Post hoc comparisons were performed using t test with Bonferroni correction.Differences among groups of non-parametric variables were analysed by Mann-Whitney U test or Kruskal-Wallis test.Categoric variables were compared using Pearson's chi-squared or Fisher's exact test.Survival analysis for time to kidney failure was done by Kaplan and Meier using log rank test for comparison between groups.A Cox proportional hazards model was used to identify predictors of kidney failure.Statistical analyses were performed using Statistical Package for Social Sciences software for Windows, version 21.0 (SPSS Inc., Chicago, IL), and graphs were made using Graph Pad Prism 7.0e (Graph Pad Software Inc., San Diego, CA).A p value of <.05 was taken as significant.

Figure 2 .
Figure 2. Profile of bacterial infection-related glomerulonephritis in diabetics.(A) Indications for kidney biopsy.(B) Classification based on severity of diabetic kidney disease.(C) Common sites of infection in para and post-infectious glomerulonephritis.(D) Classification based on period of IRGN diagnosis.ATI, acute tubular injury; DKD, diabetic kidney disease; eGFR, estimated glomerular filtration rate; IF, immunofluorescence; IFTA, interstitial fibrosis and tubular atrophy; RRT, renal replacement therapy.