Kidney involvement in myelodysplastic syndromes

ABSTRACT Introduction The objective of this study was to describe kidney involvement in patients with myelodysplastic syndromes (MDS), their treatments, and outcomes. Methods We conducted a multicenter retrospective study in seven centers, identifying MDS patients with acute kidney injury (AKI), chronic kidney disease (CKD), and urine abnormalities. Results Fifteen patients developed a kidney disease 3 months after MDS diagnosis. Median urine protein-to-creatinine ratio was 1.9 g/g, and median serum creatinine was 3.2 mg/dL. Ten patients had AKI at presentation, and 12 had extra-renal symptoms. The renal diagnoses included anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV), ANCA negative vasculitis, C3 glomerulonephritis, immune complex-mediated glomerulonephritis, polyarteritis nodosa, and IgA vasculitis. All patients but one received a specific treatment for the MDS-associated kidney injury. The effect of MDS treatment on kidney injury could be assessed in six patients treated with azacitidine, and renal function evolution was heterogenous. After a median follow-up of 14 months, four patients had CKD stage 3, five had CKD stage 4, and three had end stage kidney disease. On the other hand, three evolved to an acute myeloid leukemia and three died. Compared to 84 MDS controls, patients who had kidney involvement were younger, had a higher number of dysplasia lineages, and were more eligible to receive hypomethylating agents, but no survival difference was seen between the two groups. Compared to 265 AAV without MDS, the ten with MDS-associated pauci-immune vasculitis were older, ANCA serology was more frequently negative, and more cutaneous lesions were seen. Conclusion The spectrum of kidney injuries associated with MDS is mostly represented by vasculitis with glomerular involvement, and especially AAV.


INTRODUCTION
Hematologic malignancies can be associated with renal complications.In the last few years, the concept of monoclonal gammopathy of renal significance ( MGRS) has allowed a better characterization of kidney diseases related to monoclonal gammopathies [1 ].Nevertheless, renal complications associated with myeloid neoplasms are less described.
The objective of this work was to study the clinical and histological presentation of patients with renal impairment during MDS, as well as their therapeutic management.Furthermore, two analyses were performed: ( i) comparison of the hematological phenotype of patients with a kidney injury occurring during MDS with MDS control patients, and ( ii) comparison of the renal phenotype of AAV associated with MDS with AAV without MDS.

Patients' selection
We included patients from three different sources ( Supplementary Fig. 1) .The first group of patients were recruited at the Brigham and Women's Hospital, and the Massachusetts General Hospital, at Boston, where we screened-in all consecutive patients with a diagnosis of MDS with a kidney biopsy in the last 20 years.The second one included patients with a renal injury related with a diagnosis of vasculitis associated with MDS, from a French multicenter retrospective study by Roupie et al. who enrolled patients from different databases conducted by the "Groupe Francophone des Myélodysplasies" ( GFM) , and "French Network of Dysimmune Disorders Associated with Hemopathies" ( MINHEMON) groups [16 ].Finally, we reached out to various French scientific societies ( "Société Nationale Française de Médecine Interne," and "Société Francophone de Néphrologie Dialyse et Transplantation") to do a retrospective collection of patients cases that fitted our inclusion criteria.
Controls with MDS without any kidney injury were extracted from GFM prospective nationwide database.Controls with AAV without any MDS diagnosis come from a French study who included patients with a kidney biopsy analyzed at the Paris Cité Nephropathology unit at Necker hospital [17 ].

MDS definitions
Diagnosis criteria for MDS were based on the WHO classification revised in 2016: MDS with single lineage dysplasia ( MDS-SLD) , MDS with multilineage dysplasia ( MDS-MLD) , MDS with ring sideroblasts ( MDS-RS) , MDS-RS with single lineage dysplasia ( MDS-RS-SLD) , MDS-RS with multilineage dysplasia ( MDS-RS-MLD) , MDS with isolated del( 5q) , MDS with excess of blasts ( MDS-EB) , and unclassifiable MDS ( MDS-U) [2 ].We excluded patients with MDS secondary to an immunosuppressive treatment or radiotherapy ( therapy-related MDS) , and MDS secondary to infections, and toxic causes.Cytogenetic analysis and molecular findings were collected when available.

Definition of kidney diseases
We included adult patients ( > 18 years old) diagnosed with MDS and a kidney disease.We excluded patients with kidney disease diagnosed before MDS, and remaining stable after the myeloid neoplasm diagnosis.
According to the National Kidney Foundation's Kidney Disease Outcome and Quality Initiative ( KDOQI) guidelines, CKD ( chronic kidney disease) is defined as either a decreased GFR ( glomerular filtration rate < 60 ml/min/1.73m 2 ) , or markers of kidney damage: albuminuria with an albumin creatinine ratio ≥30 mg/g, urine sediment abnormalities, electrolyte and other abnormalities due to tubular disorders and abnormalities detected by histology among other things [18 ]. eGFR ( estimated glomerular filtration rate) was calculated according to the CKD-EPI formula [19 ].Acute kidney injury ( AKI) is defined according to the Kidney Disease: Improving Global Outcomes as follows: increase in serum creatinine by ≥0.3 mg/dl ( ≥26.5 μmol/l) within 48 hours, or increase in serum creatinine to ≥1.5 times baseline, or urine volume < 0.5 ml/kg/hour for 6 hours.Kidney biopsies were locally reviewed in each pathology center.

Clinical data
Clinical and biological data were retrospectively collected from the patients' charts.Clinical data included were gender, age at the time of the hematological and renal diagnosis, history of hypertension, and diabetes mellitus.Extra-renal organ involvement was recorded if it was directly attributable to systemic manifestations of the MDS.
Biological data, at the time of hematological and renal diagnosis, included blood cell counts, serum creatinine, eGFR, urinary protein-to-creatinine ratio, and urine sediment.
Regarding underlying MDS, data collection included bone marrow cytology and cytogenetic analyses, revised-International Prognosis Scoring System-Revised ( IPSS-R) scales and its components, treatments, and the evolution to acute myeloid leukemia ( AML) .The presence and the type of somatic mutations by next-generation sequencing were recorded when available.
At the time of the renal diagnosis, we collected: antinuclear antibodies and titer, ANCA testing by IF and/or enzyme-linked immunosorbent assay, serum albumin, C-reactive protein ( CRP) , C3/C4 levels, and the type of treatment.

Follow-up
During follow-up, events such as dialysis, kidney transplantation, occurrence of AML, or patient death were collected.Evaluation of renal function with proteinuria and serum creatinine was performed at the diagnosis of MDS, at the time of kidney disease, and at last follow-up.

Statistics
The descriptive data of the patients were expressed in number ( frequency) for the binary variables and in median ( interquartile range, IQR) for the continuous variables.Owing to non-normal distribution of the values, we used nonparametric tests.Comparisons of continuous variables between two groups were made using the Mann-Whitney test, and those of binary variables were made using the Chi-square test.A Kaplan-Meier method was used to estimate overall survival.The significance level was P < .05.

Ethics
This study was conducted according to the Declaration of Helsinki and fulfilled the recommendations of the French "Commission Nationale Informatique & Libertés" CNIL.According to its policy, this study was approved by the Institutional Review Board of the Brigham and Women's Hospital ( protocol number 2019P003787) , who validated a waiver of consent.

Characteristics of the patients
Overall, 15 patients [male gender, n = 9; median age 76 years 71-80) ] with a kidney disease associated with MDS were included in this study.

Kidney and systemic presentations
Kidney disease was diagnosed following MDS in eight cases [median time 3 months ( 0-22) ], and concomitantly in six cases.Patients mainly presented with a rapidly progressive glomerulonephritis ( n = 10, 71.4%) , a nephrotic syndrome ( n = 2, 14.3%) , isolated microscopic hematuria in one patient, acute kidney injury associated with proteinuria in one patient, and isolated proteinuria in one patient.Median urine protein-to-creatinine ratio was 1.88 g/g ( 0.6-3.1)( Table 1 ) .Median serum albumin was 3.2 g/dl ( 2.7-3.6) .Eleven patients presented with a microscopic hematuria.Median serum creatinine was 3.2 mg/dl ( 1.15-4.56) .Ten ( 66.7%) patients presented with AKI, and three ( 17.6%) required hemodialysis at diagnosis.Three patients had a previous diagnosis of CKD of unspecified cause with an eGFR between 43 and 51 ml/min/1.73m 2 .

Renal diagnoses and histological features of MDS patients
Kidney biopsy was performed in nine ( 60%) out of the 15 patients.For the other patients, the diagnosis was made according to the clinical and biological data, or based on another organ biopsy, because of kidney biopsy's contraindications ( Supplementary Table S1) .
For the patients who did not have a kidney biopsy, two had a diagnosis of microscopic polyangiitis based on the clinical and biological data confirming the ANCA positivity with a MPO specificity, one had an ANCA negative vasculitis with a diagnosis based on an association between AKI, pulmonary involvement, and purpura lesions, one had a granulomatosis with polyangiitis with an ANCA positivity specific to PR3, and one a polyarteritis nodosa based on the imaging ( Supplementary Table S1) .
Regarding the three patients with a C3 glomerulonephritis, and an immune complex-mediated glomerulonephritis, serum complement C3 and C4 levels were normal.The only patient with a type II cryoglobulinemia had a granulomatosis with polyangiitis.

Outcomes
At admission, one out of the nine patients with a previous diagnosis of MDS was started on a hematological disease-modifying therapy ( azacitidine) 2 months before the renal diagnosis was made.Following the identification of MDS-associated kidney disease, treatment was started or modified in five patients ( 33.3%) , and consisted in azacitidine ( n = 5) , a switch for one patient from lenalidomide to azacitidine, and an allogenic stem cell transplantation ( SCT) for another patient ( Table 3 ) .Among the six patients with a diagnosis of MDS concomitant or less than two months after renal involvement, three received azacitidine, one underwent an allogenic SCT and one had a withdrawal of azacitidine because of severe infections.After a median follow-up of 14 months ( 6.6-74.3), three evolved to AML.Hematologic malignancy was considered as stable in seven ( 46.7%) patients.Three patients died at last follow-up ( survival rate 80%) .
All patients but one received a specific treatment for the MDS-associated kidney injury including steroids.The immunosuppressive regimen details are listed in Table 3 .Three patients received plasma exchanges, nine had an induction treatment with rituximab, whereas three had cyclophosphamide.After treatment, data on renal function was available for 13 patients.Renal function improved in eight patients ( 53%) , and worsened in three ( 20%) .Among the three patients who underwent renal replacement therapy ( hemodialysis) at diagnosis, none recovered any renal function.

Comparison to AAV without MDS
MDS-associated AAV ( n = 10) were compared with a cohort of AAV without MDS ( n = 265) ( Table 4 ) .Patients with MDSassociated AAV were older than patients with AAV without MDS ( 76 vs 63 years old; P = .025) .The serum creatinine at diagnosis was similar between the two groups ( 2.5 mg/dl in AAV patients without MDS, and 3.6 mg/dl with MDS; P = .781) .According to the Berden classification, there was one sclerotic type in the MDS-associated AAV group compared to the AAV without MDS group ( n = 107, 40.4%) .There was no difference in the renal risk score between the two groups ( P = .523) .There is a tendency of a higher frequency of ANCA negative AAV in the MDS-associated AAV group ( n = 2 vs 11, P = .075) .Regarding organ involvement, skin was the most prevalent systemic manifestations in patients with MDS-associated AAV, compared with controls AAV ( 40% vs 9.1%; P = .012) .In terms of induction regimen, MDS-associated AAV patients were more likely to be treated with rituximab than cyclophosphamide ( 60% vs 22.3%; P = .013and 30% vs 74.7%; P = .005,respectively) .It must be noted that the median year at AAV diagnosis was 2018 ( 2011-2020) for MDS-associated AAV and 2012 ( 2009-2016) for AAV.The median follow-up for

Comparison between MDS with and without kidney injuries
There were 84 MDS patients without a kidney injury.Compared to these controls, patients who had a kidney involvement were younger ( 78 vs 75 years, respectively, P = .048)( Table 5 ) .Even if there was no difference in prognosis scores ( IPSS-R > 3.5 14.3% and 14.3%, P = 1) , MDS patients with kidney injury had a higher number of dysplastic lineages ( two dysplasia, P = .044;three dysplasia, P = .005) .MDS patients with kidney injury were more eligible to receive a specific hematologic treatment with hypomethylating agents ( 60% and 26.2% in the control MDS group, P = .003) .The median follow-up from MDS diagnosis was 23 months  in patients without a kidney injury, and 8 months  in patients with a kidney injury ( P = .277) .Patients with MDS-associated kidney injury tended to progress more frequently to AML ( 8.3% and 30%, respectively, P = .071) .The median survival was 66, and 225 months in MDS with and without kidney injury, respectively [ P = 0.7309, odd ratio: 0.71 ( 95% CI 0.24-2.11)] ( Fig. 2 b) .

DISCUSSION
Even if the knowledge about kidney injuries associated with hematological malignancies has increased recently with the concept of MGRS, data about renal complications associated with myeloid neoplasms are scarce, especially in patients with MDS.
In our study, the development of kidney disease tended to be early after MDS diagnosis ( 3 months) , and was concomitant in 40%.This was earlier than in previous large cohort [15 ].We identified vasculitis as the most frequent kidney disease associated with MDS, with a high prevalence of AAV.Systemic vasculitis associated with MDS were already described in the literature.However, a recent study reported different patterns of kidney lesions: Schwotzer et al. documented the spectrum of kidney diseases from MDS patients with kidney biopsies, and reported that acute tubulointerstitial nephritis was the predominant kidney injury [15 ].Yet, tubulointerstitial nephritis is mainly reported in chronic myelomonocytic leukemia, a hematological disorder at the frontier between MDS and myeloproliferative disorders.The fact that we have mostly identified glomerular diseases and systemic vasculitis can be explained by the recruitment of our patients.Indeed, some of our patients were included from a cohort of vasculitis associated with MDS/CMML [16 ].Furthermore, in a large cohort studying 123 patients with systemic and auto-immune manifestations associated with MDS/CMML, 32% of the manifestations were systemic vasculitis [6 ].Our analysis comparing AAV with and without MDS highlights the fact that we should look for MDS among elderly AAV patients, or with a cutaneous manifestation at AAV diagnosis.In such cases, a bone marrow biopsy should be performed if cytopenia is present.
Even if most of the cases published to date are from small series and case reports, it has also been described that extra-renal vasculitis can be associated with MDS, arguing for the association MDS-crescent glomerulonephritis that we observe here [16 ].Whether the mechanisms of lesion development are shared    between MDS or classical crescent glomerulonephritis should be explored in a dedicated study.Hence, auto-immunity seems to be a cornerstone in the setting of kidney injuries related to MDS.Extra-renal manifestations were almost systematically present when the renal disease occurred ( 80%) .Skin ( 50%) and lungs ( 42%) were the most prevalent organs involved, with alveolar hemorrhage, purpura, and skin eruption, which was concordant with the Schwotzer et al. study in which half of the patients presented extra-renal manifestations [15 ].
In our study, auto-antibodies were detected in 11 patients, with nine patients having ANCA.Even if antinuclear antibodies were present in 23% of the patients with systemic and auto-immune manifestations related with MDS/CMML in the Mekinian et al. study, there was no statistically significant difference in the frequencies of various auto-antibodies in MDS/CMML patients with and without associated auto-immune disorders, except for anti-DNA antibodies [6 , 20 ].Contrary to CMML patients, MDS patients with auto-immune disorders did not have higher ANCA positivity [20 ].
In our study, the mortality rate was 20%, with three transformations toward AML, which was similar to another large cohort [15 ].Among patients treated with azacitidine, two improved their renal function, and three worsened it, among whom two stayed on long-term hemodialysis.Overall, for the 13 patients with data after the renal treatment, renal function improved in eight patients ( 53%) and worsened in three ( 20%) .Among the three patients who underwent renal replacement therapy ( hemodialysis) at diagnosis, none recovered any renal function.Our results were similar to those of Schwotzer et al., with 10 patients evolving to CKD 3 or 4, and 5 to end stage kidney disease at last follow-up.Thus, which patients with renal complications of MDS would benefit from hypomethylat-ing agents remains to be determined.A prospective phase 2 study sponsored by the GFM assessed azacitidine efficacy and tolerance in patients with MDS and steroids dependent or refractory systemic inflammatory and auto-immune manifestations ( SAID) [21 ].In this prospective trial, 66% of MDS/CMML patients with SAID obtained SAID complete or partial response after six cycles of azacitidine.Thus, azacitidine may represent an interesting treatment for MDS/CMML patients with associated SAID, acting on both hematological disease and systemic inflammation.
The most represented subtypes of MDS were MDS-MLD and MDS-EB, two subgroups at high risk of transformation toward AML.Nevertheless, our analysis comparing MDS patients with and without kidney injury did not reveal a statistically significant difference in terms of prognosis and overall survival, this may be due to the difference of follow-up between the two groups.However, we noted that patients with a kidney injury tented to be more treated with a hypomethylating agent.Mekinian et al. showed in their study enrolling 123 patients, that MDS/CMML associated with systemic inflammatory and auto-immune disorders had worse baseline prognostic factors than control MDS/CMML, but with a similar overall survival [6 ].
Limitations of this work are mainly due to its retrospective design and the small size of the cohort, even though this is the second largest published study to date.That is why it is not possible to completely rule out the hypothesis of a fortuitous association of renal diseases with MDS.To push the analysis further, the next objectives will be to collect kidney biopsies samples to characterize the pathological and molecular features of the kidney injuries associated with MDS.
The present study clearly indicates that the kidney, along with other organs, is target of auto-immunity in the setting of MDS.MDS are associated with systemic vasculitis, especially with a glomerular involvement such as AAV.A diagnosis of AAV in a patient with cytopenias, other than inflammatory anemia, should lead to research into a possible MDS.

Figure 1 :
Figure 1: Kidney biopsy of a 28-year-old man with a history of mono-MAC syndrome ( monocytopenia and mycobacterial infection syndrome) and MDS referred to renal clinic for evaluation of a nephrotic syndrome with hematuria.Focal proliferative and crescentic glomerulonephritis, pauci-immune cellular crescents and fibrocellular or fibrous crescents with segmental scarring ( hematoxylin and eosin staining, magnification ×177) .Courtesy of Helmut G Rennke.

Figure 2 :
Figure 2: ( a) Kaplan-Meier curve for the overall survival according to AAV with and without MDS.AAV with MDS group ( n = 10) .AAV without MDS group ( n = 265) Last follow-up at 5 years.P = .0190,log rank test.( b) Kaplan-Meier curve for the overall survival according to MDS with and without kidney injury.MDS with kidney injury ( n = 15) .MDS without kidney injury ( n = 84) .Last follow-up at 5 years.P = .7309,log rank test.

Table 4 : Characteristics of ANCA-associated vasculitis according to the presence/absence of MDS.
a Missing data.