COVID-19 Vaccination and New Onset Glomerular Disease: Results from the IRocGN2 International Registry

Abstract

Key Points 

• IgAN and MCD are the most common de novo glomerular diseases reported after COVID-19 vaccination, particularly after mRNA vaccination.
• Membranous nephropathy, pauci-immune GN, and collapsing GN have also been attributed to COVID-19 vaccination, some with dual histologies.
• Recovery of kidney function and proteinuria remission is more likely in IgAN and MCD by 4–6 months compared with the other glomerular diseases.

Background 

Patients with de novo glomerular disease (GD) with various renal histologies have been reported after vaccination against SARS-CoV-2. Causality has not been established, and the long-term outcomes are not known. To better characterize the GDs and clinical courses/outcomes, we created the International Registry of COVID-19 vaccination and Glomerulonephritis to study in aggregate patients with de novo GN suspected after COVID-19 vaccine exposure.

Methods 

A REDCap survey was used for anonymized data collection. Detailed information on vaccination type and timing and GD histology were recorded in the registry. We collected serial information on laboratory values (before and after vaccination and during follow-up), treatments, and kidney-related outcomes.

Results 

Ninety-eight patients with GD were entered into the registry over 11 months from 44 centers throughout the world. Median follow-up was 89 days after diagnosis. IgA nephropathy (IgAN) and minimal change disease (MCD) were the most common kidney diseases reported. Recovery of kidney function and remission of proteinuria were more likely in IgAN and MCD at 4–6 months than with pauci-immune GN/vasculitis and membranous nephropathy.

Conclusions 

The development of GD after vaccination against SARS-CoV-2 may be a very rare adverse event. Temporal association is present for IgAN and MCD, but causality is not firmly established. Kidney outcomes for IgAN and MCD are favorable. No changes in vaccination risk-benefit assessment are recommended based on these findings.

Introduction

The swift development, clinical trial testing, manufacturing, and global distribution of COVID-19 vaccines less than a year since the World Health Organization (WHO) declaration of a pandemic was an unprecedented achievement. According to the COVID-19 vaccine tracker, as of September 2022, more than five billion people received at least one vaccine dose.¹ Various COVID-19 vaccine platforms are currently available, but the most widely distributed vaccines are the two mRNA-based vaccines, Pfizer-BioNTech and Moderna mRNA-1273, and viral vector vaccines, Oxford-AstraZeneca and Johnson & Johnson/Janssen. These vaccines have shown to be highly effective in preventing severe COVID-19 with excellent safety profiles.

The mRNA vaccines received emergency use authorization in December 2020, followed by authorization of the vector-based vaccines. Reports of de novo and relapses of glomerular disease (GD) temporally associated with COVID-19 vaccination (termed CVAGD for COVID-19 vaccine–associated GD) began to emerge by mid-2021 with an increasing number of reports published.^2–27 To date, a cause-effect relationship has not been established. Given the relative rarity of these events and limited follow-up, we created an International Registry of COVID-19 vaccination and Glomerulonephritis to collect consistent data on potential cases of CVAGD to analyze in aggregate. The aim of the study was to describe the range of kidney histologies, vaccine types and timing of administration relative to the onset of kidney disease manifestations, clinical features at presentation, and kidney outcomes.

Methods

Study Design and Participants

This retrospective cohort study included subjects of any age with de novo GN presenting within 3 months after vaccination against SARS-CoV-2. Inclusion criteria were (1) new-onset biopsy-proven GN and (2) received at least one vaccine dose within 3 months before biopsy. Exception to biopsy requirement was a positive serum test for phospholipase A2 receptor (PLA2R) antibody consistent with membranous nephropathy (MN). Exclusion criteria include active COVID-19 or other infection at GN diagnosis or preexisitng GN. Owing to pandemic-related biopsy delays, we included patients outside the 3-month window if compatible kidney-related symptoms (i.e., edema, hematuria) were evident within 2 weeks of vaccination. Data were deidentified, and a computer-generated code was assigned to allow for editing as data became available.

Data Collection

The registry was launched on June 8, 2021. Information regarding the registry with a link to the secure public survey (https://redcapsurvey.niddk.nih.gov/surveys/?s=LCDAMFD9JA) was posted on various social media platforms, newsletters, networks, and international professional societies for nephrologists and pathologists. Practicing providers were able to voluntarily report on patients with new-onset GN suspected after COVID-19 vaccination based on timing of symptoms and clinical presentation. REDCap (Research Electronic Data Capture),^28,29 a secure web-based electronic data capture tool hosted by National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health (NIH), was used to manage data. A separate comprehensive survey was completed for each patient using the same web address, but a unique code was generated for each case submission. This code allowed the reporter to return to the survey at any time to modify and add follow-up data related to changes in patient clinical status, new laboratory values, boosters, outcomes, etc. The survey allowed for both retrospective and prospective data collection. Emails generated from the REDCAP system were periodically sent to reporters to prompt them to update records after the initial survey was submitted. Before closing the registry for analysis, additional email notifications were sent out prompting all reporters to add the most recent clinical information relevant to outcomes.

Registry data elements include sociodemographics, comorbidities, COVID-19 vaccine(s), kidney and nonkidney symptoms, histology, treatment, and outcomes. Complete remission was defined as proteinuria <0.3 g/g (and negative serologies if relevant), and partial remission as <3.5 g/g and 50% proteinuria reduction. Other vaccines administered simultaneously were recorded in addition to other potential triggers. Laboratory parameters collected were serum creatinine (SCr), serum albumin, proteinuria, serologies (i.e., ANCA, antiglomerular basement membrane [anti GBM], PLA2R). eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration 2021³⁰ equation. Parameters were collected at (1) baseline prevaccination (most proximate before vaccination), (2) biopsy (after vaccination, at the time of kidney biopsy±2 weeks), and (3) follow-up. Data were checked for quality by MW and NS.

Ethical Considerations

The project was reviewed by the NIH Office of Institutional Review Board Operations and determined that storage and analysis of anonymous, deidentified data collected by the registry did not qualify as human subjects research defined under federal regulations (45 Code of Federal Regulations 46.102) and did not require institutional review board approval.

Statistical Analysis

Data are described using frequencies (percentages) and simple descriptive statistics (mean±SD or median [interquartile range]), as appropriate. Data were exported from REDCap, and variables were created for analysis. Data were assessed for distributional assumptions, and appropriate statistical tests were performed. Repeated measures analyses were performed using mixed models, with compound symmetry correlations between intervals. In situations where models would not converge, proc Generalized Linear Models-Mixed (GLIMMIX) was used. Post hoc comparisons among histologies or time intervals were corrected for multiplicity using the Bonferroni method, and corrected P-values are reported. Repeated measures analyses could not be computed for less common histologies because of limited follow-up and/or sample sizes. Categorical data were compared using the Fisher exact test. Statistical evidence was based on effect sizes, data variability, and P-values. Data were analyzed using SAS v9.4 (SAS Institute, Inc, Cary, NC).

Results

Demographics

Data from a total of 104 patients were entered into the International Registry of COVID-19 vaccination and Glomerulonephritis from June 8, 2021, to April 29, 2022. Six were excluded due to relapse of primary GN or insufficient data. In total, 98 patients with de novo GN attributed to vaccines were included in the analysis from 44 centers in 13 countries.

Most of the patients were White; 53% were female; and the mean (±SD) age at diagnosis was 48.3±18.4 years (range: 15–79) (Supplemental Table 1 and Table 1). Most (64%) were from the United States. Hypertension was present in 35%, and 16% had preexisting chronic kidney disease, which was attributed to causes unrelated to GD (e.g., kidney donation, nephrectomy for renal cell cancer, medication toxicity for treatment of other conditions, long-standing hypertension). Most (73%) received mRNA vaccines, with fewer patients (12%) receiving the vector vaccines or inactivated whole-virus vaccines (10%). Eleven (11%) received additional vaccine(s) (i.e., influenza, zoster) before the onset of GN. Nonsteroidal anti-inflammatory drug (NSAID) use was reported in 4% before or around the time of vaccination.

Table 1 - Study population baseline characteristics

Characteristic	n	%
No. of patients	98	100.0
Age, in yr
Mean±SD	48.3±18.4
Median (IQR)	49.5 (33.0–66.0)
Range	15–79
Female	52	53.1
Race
White	48	49.0
Asian	26	26.5
Black	8	8.2
American Indian/Alaska Native	1	1.0
Other/unknown	15	15.3
Ethnicity
Not Hispanic/Latino	63	64.3
Hispanic/Latino	12	12.2
Unknown	23	23.5
Countries
Australia	3	3.1
Finland	1	1.0
Greece	3	3.0
India	13	13.3
Israel	1	1.0
Mexico	5	5.1
Morocco	1	1.0
New Zealand	1	1.0
Singapore	3	3.1
Spain	2	2.0
Turkey	1	1.0
United Kingdom	1	1.0
United States	63	64.3
Vaccination name a
Pfizer-BioNTech BNT162b2	53	54.1
Moderna mRNA-1273	19	19.4
Johnson & Johnson/Jansen	3	3.1
Oxford-AstraZeneca ChAdOx1 nCoV-19 (AZD1222)	9	9.2
Sinovac/Coronavac	3	3.1
Bharat Biotech BBV152 Covaxin	7	7.1
Other b	6	6.1
GN after vaccine dose
Primary vaccination	89	90.8
First dose	36	40.5
Second dose	53	59.6
Booster c	9	9.2
First booster	7	87.5
Second booster	1	12.5
GD diagnosis
MCD	28	28.6
IgAN	24	24.5
Pauci-immune crescentic GN/Vasculitis	16	16.3
MN d	10	10.2
CG	3	3.1
FSGS	1	1.0
Anti-GBM disease	5	5.1
Lupus nephritis (class V) e	2	2.0
Immune complex GN NOS f	2	2.0
Light chain (AL) amyloidosis	1	1.0
Global glomerulosclerosis	1	1.0
Mixed glomerular histologies g	5	5.1
Pauci immune GN+anti-GBM disease	1	1.0
IgAN+FSGS	1	1.0
IgAN+CG	1	1.0
IgAN+MCD	1	1.0
MN+CG	1	1.0
Elapsed time from vaccine to kidney-related symptoms
Within hours	2	2.0
Within 1–2 d	15	15.3
3–4 d	6	6.1
5–7 d	26	26.5
2 wk	28	28.6
3 wk	4	4.1
4 wk	9	9.2
5 wk	2	2.0
6 wk	1	1.0
7 wk	1	1.0
8 wk	2	2.0
3 mo	2	2.0
Kidney-related symptoms
New-onset swelling (edema)	53	54.1
Macroscopic hematuria	29	33.7
Microscopic hematuria	36	36.7
Hypertension (new onset or worsening)	40	40.8
Foamy urine	33	33.6
Non–kidney-related symptoms
Fatigue	37	37.8
Fever	15	15.3
Rash	11	11.2
Muscle aches	27	27.6
Weakness	14	14.3
Joint pain	8	8.2
Neuropathy	2	2.0
Shortness of breath	10	10.2
History of COVID-19 infection before vaccine	5	5.1
Comorbidities
Hypertension	34	34.7
Chronic kidney disease h	16	16.3
Diabetes	13	13.3
Obesity	13	13.3
Asthma, chronic obstructive pulmonary disease	8	8.2
Cancer	6	6.1
Hepatitis C	2	2.0
HIV	1	1.0
Rheumatoid arthritis	3	3.1
Psoriatic arthritis	1	1.0
Deep venous thrombosis	2	2.0
Stroke	1	1.0
Guillain-Barre syndrome	1	1.0
Lupus (without kidney disease)	2	2.0
Neurofibromatososis 1	1	1.0
Nephrectomy	1	1.0
Smoker (current)	3	3.1
Total follow-up for cohort, in d	(n=86)
Mean±SD	115±101
Median (IQR)	89 (27–177)
Range	5–448

Data are presented in frequency and percentage. Total percentages may not equal to 100% due to rounding. IQR, interquartile range (25th–75th percentiles); GD, glomerular disease; NOS, not otherwise specified; MCD, minimal change disease; IgAN, IgA nephropathy; MN, membranous nephropathy; CG, collapsing glomerulopathy; PLA2R, phospholipase A2 receptor.

^aThe number of total vaccines is greater than the number of patients because of heterologous administration of vaccines in two patients (Moderna+J&J, Pfizer+AstraZeneca).

^bOther vaccine: Indicated mRNA but did not specify which one.

^cFirst or second booster specified in eight of nine subjects and serves as the denominator.

^dOne MN diagnosis was based exclusively on positive anti-PLA2R antibody.

^eOccurred in patients with known SLE without prior renal manifestations/nephritis.

^fMorphologic patterns include (1) a membranoproliferative pattern with subendothelial and subepithelial glomerular deposits and IgG dominance and (2) mesangial proliferation and mesangial deposits and negative IgA staining on immunofluorescence.

^gFor all subsequent analyses, the five cases of mixed histologies were classified under one subgroup; therefore, the higher “n” in each subgroup beyond Table 1 reflects this reclassification: pauci-immune GN+anti-GBM was classified as pauci-immune crescentic GN; IgAN+MCD and IgAN+FSGS were classified as IgAN; and IgAN+CG and MN+CG were classified as collapsing GN.

^hUnrelated to GN, attributed to various causes including kidney donation, nephrectromy for renal cell cancer, medication toxicity related to treatment of other conditions, and hypertension.

Biopsy Findings

Diagnoses included minimal change disease (MCD, n=28), IgA nephropathy (IgAN, n=24), pauci-immune crescentic GN/vasculitis (n=16), MN (n=10), collapsing glomerulopathy (CG, n=3), FSGS (n=1), anti-GBM disease (n=5), lupus nephritis (n=2), immune complex GN (n=2), AL amyloidosis (n=1), and global glomerulosclerosis (n=1) (Table 1). One patient with MN was diagnosed based on a positive anti-PLA2R antibody test (ELISA) and did not undergo kidney biopsy. MCD (29%) and IgAN (27%) were the most common histologies, followed by pauci-immune GN (17%) and MN (10%). Five patients showed dual glomerulopathy: pauci-immune GN with anti-GBM; IgAN mixed with CG, MCD, and FSGS; and MN mixed with CG (Table 1). For subsequent analyses, the classification of pauci-immune GN+anti-GBM was “pauci-immune,” IgAN+MCD and IgAN+FSGS were “IgAN,” and CG+IgAN and CG+MN were “CG.”

Onset of Disease and Clinical Presentation

The onset of GN was suspected after the second vaccine dose in more than half of patients (Table 1), but there was significant heterogeneity based on vaccine type and GN (Table 2). Minimal change disease (MCD) was more frequently reported after the first vaccine dose, whereas the second dose was more frequently implicated in IgAN (P=0.033) and other histologies. Kidney-related symptoms developed within 2 weeks of the suspected vaccine dose in 75% of patients (Tables 1 and 2), but onset varied between GN and vaccines, with a trend toward earlier symptom onset for MCD, IgAN, and MN and with the Pfizer vaccine (Table 2). Most common kidney-related symptoms were edema (54%), hypertension (41%), foamy urine (34%), and macroscopic hematuria (34%). Macroscopic hematuria occurred most frequently in IgAN (54%) but was also reported in anti-GBM (60%), pauci-immune (47%), and MCD (7%).

Table 2 - Timing of onset of kidney-related symptoms based on vaccine type, dose, and glomerular disease

Patient Cohort Based on Glomerular Disease	n (%)	Vaccine	Timing from Vaccine to Kidney-Related Symptoms	Culprit Vaccine Dose	Timing from Culprit Vaccine to Biopsy, Days a
All patients	98 (100%)	Pfizer-BioNTech (54.1%)	Within 5–7 d (34.5%)	Dose 2: 54 (55.7%)	(n=97)
		Moderna (19.4%)	2 wk (31.0%)	Dose 1: 36 (37.1%)	35±35
		AstraZeneca (9.2%)	Within 1–2 d (6.9%)	Booster: 7 (7.2%)	21 (13–44)
			4 wk (6.9%)		2–177
MCD	28 (28.6%)	Pfizer-BioNTech (67.9%)	Within 5–7 d (35.7%)	Dose 1: 15 (53.6%)	27±23
		AstraZeneca (14.3%)	2 wk (28.6%)	Dose 2: 12 (42.9%)	20 (13–38)
		Moderna (7.1%)	Within 1–2 d (7.1%)	Booster: 1 (3.6%)	2–88
		Covaxin (7.1%)	4 wk (7.1%)
IgAN	26 (26.5%)	Pfizer-BioNTech (57.7%)	Within 1–2 d (38.5%)	(n=25)	39±46
		Moderna (15.4%)	2 wk (23.1%)	Dose 2: 17 (68.0%)	22 (5–47)
		Other (11.5%)	Within 5–7 d (15.4%)	Dose 1: 6 (24.0%)	2–177
				Booster: 2 (8.0%)
Pauci-immune crescentic GN/Vasculitis	17 (17.3%)	Pfizer-BioNTech (52.9%)	2 wk (35.3%)	Dose 2: 12 (70.6%)	36±34
		Moderna (41.2%)	Within 5–7 d (29.4%)	Dose 1: 3 (17.7%)	19 (7–58)
		AstraZeneca (5.9%)	4 wk (11.8%)	Booster: 2 (11.8%)	4–105
MN	10 (10.2%)	Pfizer-BioNTech (50.0%)	Within 5–7 d (30.0%)	Dose 2: 6 (60.0%)	33±23
		Moderna (30.0%)	4 wk (30.0%)	Dose 1: 3 (30.0%)	33 (13–38)
		J&J (10.0%)	Within 3–4 d (20.0%)	Booster: 1 (10.0%)	5–75
		AstraZeneca (10.0%)
Anti-GBM disease	5 (5.1%)	Pfizer-BioNTech (40.0%)	2 wk (60.0%)	Dose 2: 3 (60.0%)	38±29
		Covaxin (40.0%)	3 wk (20.0%)	Dose 1: 2 (40.0%)	28 (19–42)
		Sinovac (20.0%)	3 mo (20.0%)		15–86
CG	5 (5.1%)	Moderna (60.0%)	Within 5–7 d (60.0%)	Dose 2: 3 (60.0%)	42±61
		Pfizer-BioNTech (20.0%)	2 wk (20.0%)	Dose 1: 2 (40.0%)	18 (14–20)
		AstraZeneca (20.0%)	6 wk (20.0%)		7–150

Data are presented in frequency and percentage, unless otherwise specified. Histologies with n<3 are not included. Columns 3–5 represent the top three common responses; thus, the total percentages will not equal to 100%. Available n is provided if data were missing. MCD, minimal change disease; IgAN, IgA nephropathy; MN, membranous nephropathy; CG, collapsing glomerulopathy; IQR, interquartile range (25th–75th percentiles).

^aTiming data are mean±SD, median (IQR), and range.

Clinical presentation included nephrotic syndrome in 55%, subnephrotic range proteinuria in 35%, microscopic hematuria in 37%, and isolated hematuria in 2.4%. Thromboses (renal vein or inferior vena cava) were diagnosed in three. Other contemporaneous diagnoses included autoimmune hepatitis (n=1) and hemophagocytic lymphohistiocytosis (n=1). Among pauci-immune GN, concurrent complications included pulmonary hemorrhage (n=3), pericarditis (n=1), and leukocytoclastic vasculitis rash (n=1).

Laboratory Parameters and Histopathology

At biopsy, mean (±SD) SCr was highest in anti-GBM (8.4±6.2 mg/dl) and lowest in MN (1.4±1.1 mg/dl) (Supplemental Table 2). Proteinuria was highest in MCD and CG.

Crescents were present in 29% of patients, including all pauci-immune and anti-GBM, and in 23% of patients with IgAN (Table 3). Endocapillary proliferation was predominant in IgAN (26%). Tubulointerstitial nephritis (TIN) was a coexisting finding in 25% of patients, as was acute tubular injury.

Table 3 - Selected histologic features on kidney biopsies among patients with glomerular disease

Feature	Frequency of Finding (All Patients)	MCD	IgAN	Pauci-Immune Crescentic GN/Vasculitis	MN a	Anti-GBM Disease	CG
	(n=98)	(n=28)	(n=26)	(n=17)	(n=10)	(n=5)	(n=5)
Collapsing features	6 (6.1%)	0	0	1 (5.9%)	0	0	5 (100.0%)
Crescents	28 (28.6%)	0	6 (23.1%)	17 (100%)	0	5 (100%)	0
TRI	1 (1.0%)	0	0	0	0	0	1 (20.0%)
Focal acute tubular injury	9 (9.2%)	2 (7.1%)	3 (11.5%)	1 (5.9%)	1 (10.0%)	1 (20.0%)	1 (20.0%)
Diffuse acute tubular injury	24 (24.5%)	4 (13.3%)	6 (23.1%)	6 (35.3%)	2 (20.0%)	3 (60.0%)	2 (40.0%)
Acute tubular necrosis	7 (7.1%)	2 (7.1%)	1 (3.9%)	2 (11.8%)	0	0	2 (40.0%)
TIN/inflammatory infiltrate	25 (25.5%)	6 (21.4%)	6 (23.1%)	7 (41.2%)	1 (10.0%)	3 (60.0%)	1 (20.0%)
Endocapillary proliferation	11 (11.2%)	0	7 (26.9%)	2 (11.8%)	0	0	0
Electron microscopy deposits
Subepithelial	10 (10.2%)	b 1 (3.6%)	0	0	c 6 (60.0%)	0	1 (20.0%)
Subepithelial humps	2 (2.0%)	0	1 (3.9%)	0	1 (10.0%)	0	0
Subendothelial	3 (3.1%)	0	1 (3.9%)	1 (5.9%)	0	0	0
Mesangial	33 (33.7%)	2 (7.1%)	22 (84.6%)	3 (17.7%)	1 (10.0%)	1 (20.0%)	1 (20.0%)
Global sclerosis	1 (1.0%)	0	0	0	0	0	0
Other					PLA2R+: 4/9 (44%)
Immunostaining d					THSD7A: 3/3: neg
					NELL1: 3/3: neg
					EXT1/2: 3/3: neg

Data are presented in frequency and percentage. Total percentages may not equal to 100% due to rounding. Histologies with n<3 are not included. MCD, minimal change disease; IgAN, IgA nephropathy; MN, membranous nephropathy; CG, collapsing glomerulopathy; TRI, tubuloreticular inclusions; TIN, tubulointerstitial nephritis; PLA2R, phospholipase A2 receptor; THSD7A, thrombospondin type 1 domain-containing 7A; NELL1, neural epidermal growth factor-like 1; EXT1/2, exostosin.

^aOne patient did not undergo kidney biopsy but was diagnosed with MN based on a positive serum anti-PLA2R antibody test.

^bRare deposit insufficient to explain diffuse process effacement, thus classified as MCD rather than MN by the pathologist.

^cNot all patients examined by electron microscopy.

^dMN antigen immunostaining not performed in all patients; denominator represents the number of patients tested.

Five (50%) patients with MN were PLA2R-positive by tissue staining (Table 3) and/or anti-PLA2R serology (Supplemental Table 3). The others were of an unknown antigen type. Among pauci-immune patients, 76% were ANCA-positive, without a consistent autoantibody pattern. Rather, there was a mixture of C- and P-ANCAs, double seropositive (n=1) and copositivity with anti-GBM (n=1). ANCAs were rare in IgAN. Antinuclear antibodies (ANA) were present in various GNs, but anti-double stranded DNA (anti-dsDNA) and hypocomplementemia were uncommon in nonlupus patients (Supplemental Table 3). Among three CG with available apolipoprotein L1 (APOL1) genotyping, two Black patients had two APOL1 risk alleles.

Management and Outcomes

Renin-angiotensin-aldosterone system inhibitors were initiated in 30% (Table 4); 60% received immunosuppressive therapy, with corticosteroids alone or in combination with anti CD20 therapy, cyclophosphamide, mycophenolate, or calcineurin inhibitors. The kidney disease was self-limiting in 11%. Complete or partial remission (spontaneous or with immunosuppression) was achieved in 60% of patients overall, in 86% of patients with MCD and 61% of patients with IgAN, but fewer remissions among other subgroups (Table 5); 14% did not achieve remission and 9% required dialysis, the majority with pauci-immune or anti-GBM disease. Five patients died, one of whom required hemodialysis. The histologies associated with these patients are summarized in Table 5.

Table 4 - Management of the common glomerular diseases reported after vaccination

Treatment	All Patients	MCD	IgAN	Pauci-Immune GN	MN	Anti-GBM Disease	CG
	(n=98)	(n=28)	(n=26)	(n=17)	(n=10)	(n=5)	(n=5)
Observation only	14 (14.3%)	1 (3.6%)	6 (23.1%)	1 (5.9%)	4 (40.0%)	0	0
RAAS inhibition	29 (29.6%)	5 (17.9%)	14 (53.9%)	3 (17.7%)	6 (60.0%)	0	1 (20.0%)
Immunosuppression	70 (71.4%)	26 (92.9%)	10 (38.5%)	16 (94.1%)	5 (50.0%)	0	4 (80.0%)
Corticosteroids	59 (60.2%)	26 (92.9%)	8 (30.8%)	13 (76.5%)	2 (2.0%)	3 (60.0%)	3 (60.0%)
Rituximab	14 (14.3%)	2 (7.1%)	0	6 (35.3%)	3 (30.0%)	2 (40.0%)	0
Cyclophosphamide	13 (13.3%)	0	2 (7.7%) a	8 (47.1%)	0	3 (60.0%)	0
Plasma exchange	10 (10.2%)	0	0	5 (29.4%)	0	5 (100.0%)	0
Mycophenolate mofetil	8 (8.2%)	0	2 (7.7%)	0	0	0	2 (40.0%)
Tacrolimus	4 (4.1%)	1 (3.6%)	1 (3.9%)	0	1 (10.0%)	0	0
Cyclosporine	1 (1.0%)	1 (3.6%)	0	0	0	0	0
Azathioprine	1 (1.0%)	0	0	1 (5.9%)	0	0	0
Other	2 (2.0%)	0	0	1 (5.9%)	0	0	0

Data are presented as frequency and percentage. Total percentage may not equal to 100% due to rounding.

Histologies with n<3 are not included. MCD, minimal change disease; IgAN, IgA nephropathy; MN, membranous nephropathy; CG, collapsing glomerulopathy; RAAS, renin-angiotensin-aldosterone system.

^aCyclophosphamide used in these patients because of acute kidney injury in the setting of crescents on kidney biopsy.

Table 5 - Outcomes based on selected glomerular disease subgroup

Outcome	All Patients	MCD	IgAN	Pauci-Immune GN	MN	Anti-GBM Disease	CG
	(n=98)	(n=28)	(n=26)	(n=17)	(n=10)	(n=5)	(n=5)
Spontaneous remission	4 (4.1%)	1 (3.6%)	3 (11.5%)	0	0	0	0
PR without treatment	7 (7.1%)	1 (3.6%)	4 (15.4%)	0	2 (2.0%)	0	0
PR after treatment	23 (23.5%)	7 (25.0%)	7 (26.9%)	4 (23.5%)	1 (10.0%)	1 (20.0%)	2 (40.0%)
CR after treatment	24 (24.5%)	15 (53.6%)	2 (7.7%)	4 (23.5%)	0	1 (20.0%)	0
No response to treatment/no remission	14 (14.3%)	1 (3.6%)	3 (11.5%)	3 (17.7%)	3 (30.0%)	1 (20.0%)	2 (40.0%)
Progression of disease with decline in GFR	4 (4.1%)	0	1 (3.9%)	0	1 (10.0%)	0	1 (20.0%)
Persistent proteinuria	16 (16.3%)	1 (3.6%)	5 (19.2%)	2 (11.8%)	5 (50.0%)	0	2 (40.0%)
Relapsed after achieving remission	2 (2.0%)	2 (7.1%)	0	0	0	0	0
Relapsed after vaccine rechallenge a	6	3 b	3 c
Required RRT	9 (9.2%)	0	2 (7.7%)	4 (23.5%)	0	2 (40.0%)	1 (20.0%)
Died d	5 (5.1%)	1 (3.6%)	0	2 (11.7%)	0	1 (20.0%)	1 (20.0%)

Data are presented as frequency and percentage. Total percentages may not equal to 100% due to rounding. Histologies with n<3 are not included. MCD, minimal change disease; IgAN, IgA nephropathy; MN, membranous nephropathy; CG, collapsing glomerulopathy; PR, partial remission; CR, complete remission.

^aLimited data on vaccine rechallenges available, thus denominator not known.

^bOne patient with MCD developed nephrotic syndrome within 6 days of the first vaccine dose, spontaneously remitted, and then relapsed after the second vaccine dose requiring immunosuppression. Two patients with MCD treated with immunosuppression after diagnosis (achieved remission) relapsed after additional doses. In one of these patients, two additional vaccine doses were tolerated well, but another relapse occurred after the fourth dose.

^cTwo patients who developed gross hematuria within days of the first vaccine dose had recurrence of hematuria within hours of the second vaccine dose leading to diagnostic renal biopsy of IgAN. One patient with onset of IgAN during pregnancy after the second vaccine dose had recurrence of gross hematuria after the third dose that resolved spontaneously.

^dOne patient with CG required RRT. Comorbidities in deceased included chronic kidney disease (3), cancer (2), viral hepatitis (1), rheumatoid arthritis (2), and chronic obstructive lung disease (1). The cause of death was unknown in two. Withdrawal of care due to cancer progression, cardiogenic shock, and respiratory failure/pulmonary hemorrhage contributed to death in three patients.

Among 36 patients with GN onset attributed to the first vaccine dose, 44% did not receive subsequent vaccine doses and 25% received the second dose. Details of vaccine rechallenges and relapses are limited. A different vaccine platform was used in two of the vaccine rechallenges (changing from the mRNA platform to the viral vector) (Table 5). Two patients with MCD experienced disease relapses despite no additional vaccine doses.

Longitudinal Outcomes

Median duration of follow-up was 89 (27–177) days. Overall, when histologies were combined, there was a substantial increase in median SCr (mg/dl) from baseline (0.9 [0.8–1.1]) to biopsy (1.7 [1.0–3.7], P<0.001) with improvement at follow-up (1.2 [0.9–1.7] P = 0.043) (Figure 1A). eGFR (ml/min per 1.73 m²) decreased from 90 (67–109) (mean 88±26) at baseline to 42 (15–81) (mean 51±39) at biopsy (P<0.001) with some improvement at follow-up (70 [43–92], mean 67±32; P = 0.052). Although there was no statistical difference in SCr between baseline and follow-up (P = 0.38), overall eGFR remained significantly lower at follow-up compared with baseline (P = 0.001), suggesting incomplete or lack of renal recovery (Figure 1C).

Figure 1: Longitudinal changes in serum creatinine and eGFR show decline in renal function, and subsequent recovery varies based on glomerular disease histology after vaccination. All histologies combined (A, C) and based on histology (B, D), at baseline (prevaccination), and at/around the time of kidney biopsy and follow-up (approximately 4–6 months). Data are presented as mean ±95% confidence interval. Data were assessed for statistical assumptions, and longitudinal data were analyzed using mixed models. Post hoc comparisons were corrected for multiplicity using the Bonferroni method, and corrected P-values are reported along the brackets. The timing of the culprit vaccine dose relative to the timing of renal biopsy is represented by the vertical red lines (solid=median; dash=interquartile range; dot=range). Combined histologies include those most commonly observed (n>3).

In MCD and IgAN, kidney function tended to decline from baseline to biopsy (Figure 1, B and D, Supplemental Table 2), but there were no statistical differences in eGFR between baseline and follow-up (P = 1.0 and P = 0.12, respectively), suggesting recovery. By contrast, in pauci-immune GN and MN, eGFR at follow-up was lower compared with that at baseline (P = 0.009 and P = 0.035, respectively).

Most did not have proteinuria at baseline (Supplemental Table 2). At biopsy, overall median proteinuria (g/g) was 4 (2–9), which decreased to 1 (0–4), (P<0.001) by follow-up (Figure 2A). Persistent proteinuria was reported in 16% of patients (Table 5) after a mean follow-up of 120±109 days. Changes in proteinuria based on histology are presented in Figure 2B and Supplemental Table 2. Substantial improvements in proteinuria were seen over time with MCD and IgAN, for whom the majority achieved remission (Table 5). Specifically, in MCD, median proteinuria (g/g) decreased from 10 (7–15) at biopsy to 0 (0–3) at follow-up (P<0.001). For IgAN, proteinuria decreased from 2 (0.3–2) at biopsy to 1 (0–2) at follow-up, although statistical evidence was not strong (P = 0.081). By contrast, there were no statistical differences in proteinuria from biopsy to follow-up for pauci-immune GN (2 [1–4] versus 3 [1–4]; P = 0.18) and MN (6 [4–11] versus 10 [3–13]; P = 1.0).

Figure 2: Longitudinal changes in proteinuria show significant proteinuria at time of kidney biopsy, and remission at follow-up varied based on glomerular disease histology after vaccination. All histologies combined (A) and based on histology (B), at baseline (prevaccination), and at/around the time of kidney biopsy and follow-up.

Overall, median serum albumin (g/dl) decreased from 4.2 at baseline (4.0–4.5) to 2.7 (2.1–3.4) (P<0.001) at biopsy with recovery to 3.8 (3.1–4.3) at follow-up (P<0.001) but lower compared with baseline (P = 0.047) (Figure 3, A and B and Supplemental Table 2). There were no substantial changes in albumin in patients with IgAN (P = 0.17). In MCD, albumin decreased expectedly from baseline to biopsy (P<0.001) with a trend toward recovery at follow-up (P = 0.048), whereas in MN, albumin was lower at follow-up versus baseline (P<0.001).

Figure 3: Longitudinal changes in serum albumin. Degree of hypoalbuminemia at time of kidney biopsy and potential for recovery at follow-up varies based on glomerular disease histology after vaccination. All histologies combined (A), and based on histology (B) at baseline (prevaccination), and at/around the time of kidney biopsy and follow-up.

Discussion

To the best of our knowledge, this represents the largest series to date of de novo GN temporally related to COVID-19 vaccination or any vaccine, in general. IgAN and MCD were the dominant kidney diseases and more common after mRNA-based vaccines. The dominance of these GNs is consistent with published reports and an analysis of the WHO's pharmacovigilance database, Vigibase.³¹ Of 143 patients with nephrotic syndrome in Vigibase after COVID-19 vaccination, 72% received Pfizer. Minimal lesion GN was predominantly associated with the Pfizer vaccine, whereas IgAN was modestly associated with Moderna. In our cohort, the onset of symptoms was more frequent after the second dose, with the exception of MCD, which was slightly more common after the first dose.

The age distribution for the GN subtypes is similar to primary GN. Patients with pauci-immune GN, anti-GBM, and MN tended to be older than those with MCD, IgAN, or CG. There were few pediatric patients in our registry (n=4), which may reflect the timing of vaccine approval in this population or lower vaccine doses. Published pediatric patients with GN are predominantly IgAN and MCD.^26,32–35 Racial distribution was noteworthy, in that AA represented a minority of patients in our registry and 38% developed CG. This distribution should be interpreted with caution in light of vaccine access issues and hesitancy among underrepresented minorities, particularly early in the vaccination rollout. Nevertheless, this contrasts with the AA predominance observed with COVID-19–associated nephropathy.

When all histologies were combined, eGFR was significantly lower at follow-up compared with prevaccination baseline. Among the different histologies, patients with IgAN and MCD tended to have more favorable kidney prognosis, with improvement in GFR and proteinuria, and were more likely to show remission by 4–6 months. By contrast, MN and pauci-immune GN were less likely to achieve remission or renal recovery during the short-term follow-up. However, the kinetics of proteinuria reduction in MN is often delayed, lagging behind immunologic responses. The less favorable clinical course of pauci-immune GN is likely influenced by the nature of the disease, lower GFR before vaccination and at presentation, older age, and comorbidities. Longer term follow-up is needed to assess the potential for kidney recovery among all the histologies.

Among histological subgroups in which serology is the basis for classification, autoantibody (AB) profiles were mixed. For the MN subgroup, 50% were PLA2R-positive, but the AB profile was undefined in PLA2R-negative patients. There are rare reports of neural epidermal growth factor-like 1³⁶ and thrombospondin type 1 domain-containing 7A-positive MN after COVID-19 vaccination.³⁷ Our limited data set precludes meaningful interpretation of whether suspected postvaccine MN is “secondary” to the vaccine (possibly with unique AB profiles) or “unmasked” subclinical MN. Indeed, PLA2R AB may develop years before MN diagnosis.³⁸ Similarly, in pauci-immune patients, there was no consistent AB pattern. Thus, the serologies do not provide insight into potential distinct mechanisms in vaccine-associated GN.

Co-occurrence of histolopathologies, mainly podocytopathy mixed with IgAN or MN, was observed in a subset of cases. Whether these findings are fortuitous, directly pathogenically linked to vaccination or whether one glomerular lesion enhances susceptibility after vaccination is uncertain. This may be a potential unmasking mechanism. In addition, TIN was superimposed on glomerular pathologies in 25% of patients. Patients with isolated TIN after COVID-19 vaccination have been reported^20,39–42 and may be another unmasking mechanism by exacerbating an unrecognized GN.

According to WHO guidelines, adverse events (AE) following immunization can be defined as any untoward, unfavorable, or unintended medical occurrence after vaccination, not necessarily exhibiting a consistent, causal relationship with the vaccine product. WHO recommends a four-step algorithm for assessing causality of AE after vaccination,⁴³ but Bradford Hill criteria may also be used⁴⁴ and considers criteria of temporality, consistency, strength of association, specificity, and biological plausibility. Using these criteria to assess causality for GN is not straightforward: (1) Temporality and consistency: Kidney-related symptoms (with exception of gross hematuria) tend to be subtle and nonspecific making it challenging to define a temporal relationship between vaccination and GN. However, among GNs, temporality is more consistent for IgAN and MCD. Although temporality does not prove causation, lack of temporality does not disprove causation because autoimmune responses may evolve slowly and clinical manifestations may lag. (2) Strength of association: Relapses of MCD and IgAN shortly after vaccine rechallenge adds strength to this association. (3) Specificity: There is lack of specificity to one COVID-19 vaccine. Non–COVID-19 vaccines are also implicated in de novo GN. While the proportion of events related to mRNA vaccines is higher than other platforms, this may reflect more widespread global distribution and a much larger denominator. (4) Biologic plausibility: A unique biological mechanism by which the COVID-19 vaccine(s) potentially triggers GN is yet to be clarified. The mRNA technology promotes more potent immune responses than inactivated viral vaccines or adjuvanted protein vaccines, but given the heterogeneity and spectrum of GN observed, multiple mechanisms as part of a multihit process in a susceptible host is likely and hypothesized elsewhere.^11,45–47

It is possible that some GN cases are coincidental and unrelated to vaccination. In this regard, we identified comorbidities and medications that may be alternative contributors to GN in some patients, i.e., HIV and psoriatic arthritis in IgAN, NSAID use and Guillian-Barre syndrome in MCD, hepatitis C/colon cancer/NSAIDs in PLA2R-negative MN, lymphoma in CG, and hydralazine use in pauci-immune GN. In other cases, histology patterns such as global glomerulosclerosis and amyloid are more suggestive of long-standing disease that preceded vaccination. Three patients with IgAN had documented hematuria (single-episode gross hematuria or intermittent microscopic) before vaccination. This hematuria had been attributed to other reasons (cystitis, nephrolithiasis, urinary tract infection) but may have reflected preexisting GD. Similarly, it is also possible that patients with preexisting CKD (attributed to other causes) had underlying GD. However, they did not have documented hematuria or significant proteinuria at baseline. Contributions from coadministered vaccines or prior COVID-19 infection cannot be excluded. Finally, increased health care utilization after vaccination compared with the height of the pandemic may be a nonimmunologic factor unmasking preexisting GN.

Optimal management is not defined, but standard immunosuppressive treatments have led to responses. It is unknown whether the molecular and immunologic mechanisms of CVAGD are similar to non-CVAGD. An improved understanding of immune responses and kinetics in CVAGD (and between GNs) and potential for spontaneous recovery versus relapse is needed to inform management.

The issue and safety of rechallenging with future boosters (using the same or different vaccine platform) needs clarification because some subjects relapsed after additional doses. Such decisions will need to be individualized, factoring in histologic diagnosis, previous recovery, comorbidities, and consideration that patients with GN are at higher risk of severe COVID-19.^48,49

Our study has several limitations. The statistical power of analyses was limited, especially for the less frequent outcomes. The nature of the data set does not allow firm causal conclusions or estimate GN frequency. As with all passive reporting systems, our registry is subject to under/incomplete reporting. The registry relies on patient identification based on clinical judgment and voluntary submission, leading to possible selection bias toward more overt, earlier AEs, which may affect result generalizability. We purposefully did not apply a case-level clinical review to assess causality (or exclude cases). This allowed us to report an unbiased range of GN suspected after vaccination. Thus, misattribution is possible. We did not include control spontaneous GN patients for comparison. However, a study from Switzerland using data from all Swiss pathology institutes did not find any differences in the incidence of the four common GNs before or after initiation of the COVID-19 vaccination campaign.⁵⁰ Finally, the risk of GN may be underestimated if the real risk interval was longer than the time window defined in inclusion criteria. However, there is lack of precise references on plausible time frames for appearance of AEs after vaccination, particularly for autoimmune diseases. Despite these limitations, the strengths of this study are international representation, comparisons between histologic subgroups and vaccines, and longitudinal data.

In conclusion, the exchange of information provided in this report, although preliminary, is important for transparency, awareness, and avoidance of misinformation. Any concerns regarding the safety of SARS-CoV-2 vaccines need to be taken seriously and warrant thorough investigation. Importantly, this report highlights the many challenges and complexity of assessing causality between AEs and vaccination. The true incidence of GN conferred by COVID-19 vaccines cannot be calculated, but risk appears low, likely close to the background incidence of these conditions in large populations⁵⁰ and represents a very small percentage of vaccinated individuals. Importantly, most patients with MCD and IgAN improved with conventional therapies. These outcomes are reassuring, particularly considering the prevalence of acute and chronic kidney injury and GN with less favorable outcomes linked to COVID-19. Given the well-established benefits of vaccination in prevention of severe COVID-19, no change in the current vaccination guidelines is suggested.

Collaborative research efforts, centralized repositories, and longitudinal data are needed to understand the potential role of COVID-19 vaccines in the development of GN. This is relevant in light of authorization of the bivalent COVID-19 boosters based on mRNA platforms, which will likely continue to be adapted in the future to target new SARS-CoV-2 variants.