Renal Histology in ANCA-Associated Vasculitis: Differences Between Diagnostic and Serologic Subgroups

Kidney International, Vol. 61 (2002), pp. 80–89

CELL BIOLOGY – IMMUNOLOGY – PATHOLOGY

Renal histology in ANCA-associated vasculitis: Differences between diagnostic and serologic subgroups

HERBERT A. HAUER,INGEBORG M. BAJEMA,HANS C. VAN HOUWELINGEN, FRANCO FERRARIO,LAURE-HE´ LE` NE NOE¨ L,RU¨ DIGER WALDHERR,DAVID R.W. JAYNE, NIELS RASMUSSEN,JAN A. BRUIJN, and E. CHRISTIAAN HAGEN, on behalf of the EUROPEAN VASCULITIS STUDY GROUP (EUVAS)1

Departments of Pathology and Medical Statistics, Leiden University Medical Center, Leiden, Department of Pathology, Erasmus University Medical Center, Rotterdam, and Department of Internal Medicine, Eemland Hospital, Amersfoort, the Netherlands; Renal Immunopathology Center, Ospedale San Carlo Borromeo, Milan, Italy; INSERM U507, Hoˆpital Necker, Paris, France; Department of Pathology, University of Heidelberg, Heidelberg, Germany; Renal Unit, Addenbrooke’s Hospital, Cambridge, England, United Kingdom; and Department of Otolaryngology, Rigshospitalet, Copenhagen, Denmark

Renal histology in ANCA-associated vasculitis: Differences tion than glomerulonephritis in relation to WG. This difference between diagnostic and serologic subgroups. may be due to a delayed establishment of diagnosis in patients Background. Differences in renal histopathology between with MPA compared to patients with WG. Both active and microscopic polyangiitis (MPA) and Wegener’s granulomato- chronic lesions are more abundantly present in MPO-ANCA– sis (WG), and between anti-neutrophil cytoplasm autoantibody positive patients than in patients with PR3-ANCA–positivity, (ANCA) test results in patients with ANCA-associated vasculi- which suggests that the pathogenesis of renal disease in these tis may provide insight into the differences in pathogenesis ANCA subsets could be different. Our results also suggest and raise the opportunity of classifying the vasculitides more that ANCA test results may be useful in classifying ANCA- accurately. The possible differences in histopathology are in- associated vasculitides. vestigated in this study. Methods. We report an analysis of 173 patients with renal disease in microscopic polyangiitis or Wegener’s granulomato- Rapidly progressive deterioration of renal function sis. A total of 173 renal biopsies, performed at diagnosis, were scored by two observers separately, using a previously stan- is a frequent but clinically unfavorable feature of anti- dardized protocol. Consensus on each biopsy was achieved neutrophil cytoplasm autoantibody (ANCA)-associated during a central review. vasculitis, including microscopic polyangiitis, Wegener’s Results. Normal glomeruli were more common in WG than granulomatosis, and renal limited vasculitis [1]. Usually, Ͻ in MPA (P 0.001). Glomerulosclerosis was more prominent this rapidly progressive deterioration of renal function in MPA than in WG (P ϭ 0.003). Interstitial fibrosis (P Ͻ 0.001), tubular atrophy (P Ͻ 0.001), and tubular casts (P ϭ is histopathologically accompanied by a pauci-immune 0.005) were more frequently present and more severe in MPA crescentic necrotizing glomerulonephritis. However, the than in WG. Presence of glomerulosclerosis was more extensive histopathological features vary among patients from mild in patients with myeloperoxidase (MPO)-ANCA than with focal segmental extracapillary proliferation to diffuse ϭ proteinase 3 (PR3)-ANCA (P 0.022). Interstitial fibrosis crescentic necrotizing glomerulonephritis with granulo- (P ϭ 0.008), tubular necrosis (P ϭ 0.030), tubular atrophy (P ϭ 0.013), and intra-epithelial infiltrates (P ϭ 0.006) were more mas and tubular intra-epithelial infiltrates. In some cases, frequently present and more severe in MPO-ANCA than in extensive glomerulosclerosis is found [2]. PR3-ANCA. The differences between microscopic polyangiitis and Conclusions. Glomerulonephritis in relation to MPA has Wegener’s granulomatosis have been described as fol- more characteristics of chronic injury at the time of presenta- lows [3, 4]. Microscopic polyangiitis is characterized by a non-granulomatous systemic vasculitis that is associ- 1 A complete list of EUVAS participants is in the Acknowledgments. ated, in a majority of cases, with ANCA positivity in a cytoplasmic (C-ANCA, ϳ25%) or perinuclear (P-ANCA, Key words: renal biopsy, glomerulonephritis, microscopic polyangiitis, Wegener’s granulomatosis, renal limited vasculitis, ANCA-associated ϳ60%) staining pattern, as determined by indirect im- vasculitis. munofluorescence [5]. Sera that produce the C-ANCA Received for publication June 8, 2001 pattern nearly always react with proteinase-3 (PR3- and in revised form July 31, 2001 ANCA). The major P-ANCA target is myeloperoxidase Accepted for publication August 1, 2001 (MPO-ANCA). Wegener’s granulomatosis is character-  2002 by the International Society of Nephrology ized by a granulomatous systemic vasculitis with upper

80 EUVAS: Renal histology in vasculitis 81

Table 1. Inclusion criteria for CYCAZAREM Table 2. Inclusion criteria for MEPEX (1, 2 and 3 are required) (1, 2 and 3 are required) 1. A diagnosis of new or previously untreated MPA, RLV, or WG 1. A diagnosis of new or previously untreated MPA, RLV, with active vasculitis, as indicated by the presence of active or WG with or without histological confirmation. necrotizing glomerulonephritis on renal biopsy. 2. Either Renal involvement attributable to active MPA, 2. ANCA-positivity; that is, one of the following possibilities: RLV, or WG with one or more of: 1) C-ANCAϩ 1) Elevated serum creatinine 2) PR3-ANCAϩ 2) Hematuria (Ͼ30 red blood cells per high power 3) MPO-ANCAϩ field) ANCA-negativity is allowed if the disease is confirmed histologically 3) Red cell casts 3. Biopsy-proven necrotizing and/or crescentic glomerulonephritis, in 4) Proteinuria (Ͼ1 gram/24 h) the absence of another defined glomerulopathy, with severe renal And/or Other severe organ involvement or imminent loss of vital impairment; that is, one of the following possibilities: organ function, attributable to active MPA or WG. 1) Oliguria (Ͻ400 mL/24 h) 3. ANCA-positivity; that is, one of the following 2) Intention to commence dialysis within 48 hours of admission possibilities: 3) Creatinine Ն500 ␮mol/L 1) C-ANCAϩ 2) PR3-ANCAϩ 3) MPO-ANCAϩ ANCA-negativity is allowed if the disease is confirmed histologically. conditions enabled us to perform the present study in Abbreviations are in the Appendix. which we determined (1) differences in the occurrence of renal histological lesions between microscopic polyangiitis, Wegener’s granulomatosis, and renal limited vasculitis; and (2) differences in the occurrence of renal histo- airway involvement that is associated, in a majority of logical lesions between patients with ANCA-associated cases, with C-ANCA positivity (ϳ65%) in the presence vasculitis with various ANCA test results. of PR3-ANCA, or P-ANCA positivity (ϳ20%) in the presence of MPO-ANCA [5]. Renal limited vasculitis is METHODS distinguished from microscopic polyangiitis and Wegen- Patients er’s granulomatosis by the absence of extrarenal symptoms of vasculitis. Despite these relatively clear disease Patients in the present study were derived from two of descriptions, a diagnosis for the individual patient may the first-wave trials of the EUVAS [10]: CYCAZAREM, be difficult to establish in clinical practice. Currently, a randomized trial of cyclophosphamide versus azathio- patients are treated similarly, irrespective of their diag- prine during remission of ANCA-positive systemic vascu- nosis. Whether diagnosis-specific modulation of therapy litis [10], and MEPEX, a randomized trial of adjunctive is needed, is unclear. This issue needs to be addressed. therapy for severe glomerulonephritis in ANCA-associ- Knowledge of the differences in renal histopathology ated systemic vasculitis that examined plasma exchange versus intravenous administration of methylpredniso- between these three diseases is relatively limited, though lone [10]. Patients only were included into the present distinct differences may help to establish a more certain study if a renal biopsy, performed at study entry, was avail- diagnosis and may give insight into the pathogenesis of able. Inclusion criteria for CYCAZAREM and MEPEX ANCA-associated systemic vasculitis. are listed in Tables 1 and 2, respectively. Exclusion crite- Clinically, vasculitic syndromes overlap, and diagnoses ria for both trials included age under 18, pregnancy, pre- sometimes may be uncertain. Therefore, differentiation vious malignancy, known HIV positivity, administration between the vasculitic syndromes by means of ANCA of any cytotoxic drug within a year before entry, and hep- test results is sometimes preferred, which may be sup- atitis B antigenemia (only if HepB-ag positive). An addi- ported by the possible pathogenic role of ANCAs in tional exclusion criterion for CYCAZAREM was serum pauci-immune crescentic necrotizing glomerulonephritis creatinine Ͼ500 ␮mol/L. Additional exclusion criteria for in relation to ANCA-associated vasculitis, and by the MEPEX were serum creatinine Ͻ500 ␮mol/L, life-threat- observation that meaningful differences between pa- ening non-renal manifestations of vasculitis, including tients with various ANCA test results exist [6]. Differ- alveolar hemorrhage requiring mechanical ventilation ences in renal histopathology between patients with dif- within 24 hours of admission, on dialysis for more than ferent ANCA test results were investigated but led to two weeks prior to referral, significant baseline renal im- contradictory results (abstract, Bajema IM, J Am Soc pairment (creatinine Ͼ200 ␮mol/L, one year or more be- Nephrol 8:532A, 1997) [7–9]. fore presentation), or a previous episode of biopsy-proven Within the framework of the European Vasculitis Study necrotizing and/or crescentic glomerulonephritis [10]. group (EUVAS; see Acknowledgments), we prospec- tively gathered renal biopsies with clinical information Renal histopathology of a large and homogeneous group of patients with renal Paraffin sections were stained with silver, periodic disease in relation to ANCA-associated vasculitis. These acid-Schiff (PAS), hematoxylin & eosin (H&E), and tri- 82 EUVAS: Renal histology in vasculitis chrome, and were forwarded to two of five participating ANCA, respectively. To determine the different patterns pathologists (IMB, LHN, FF, RW, JAB). Both patholo- of renal lesions in ANCA test subgroups, the ANCA gists scored the biopsies separately, blinded to patient test results also were classified as a combination of the data and the scores of the other observer, and according IIF and ELISA results, as the specificity of ANCA testing to a previously standardized protocol for scoring renal increases by using this combination, compared to IIF or biopsies of patients with ANCA-associated systemic vas- ELISA alone [5]. culitis [11, 12]. In short, each glomerulus had to be scored separately on the presence of fibrinoid necrosis, crescents Statistics (cellular/fibrous and segmental/circumferential), glomer- Differences of quantitative parameters between groups ulosclerosis (local/segmental/global), periglomerular in- were assessed with the t test for independent samples. filtrates, granulomatous reactions, as well as a number Differences of semiquantitative results were tested by of other lesions. The presence of glomerular lesions was using the Mann-Whitney U test. Difference of gender calculated as the percentage of the total number of glo- between groups was analyzed by the chi-square test. meruli in a biopsy. Most interstitial, tubular and vascular lesions were scored as present or absent, except for inter- RESULTS stitial infiltrates (Ϫ/ϩ/ϩϩ/ϩϩϩ), interstitial fibrosis (Ϫ/ϩ/ϩϩ) and tubular atrophy (Ϫ/ϩ/ϩϩ), which were Patients scored semiquantitatively. Together, 39 histological pa- Renal biopsies were performed in 132 of 157 patients rameters were examined. The scores were entered into a that were recruited in CYCAZAREM. As a renal bi- central database (MSAccess) and discrepancies between opsy was an inclusion criterion for MEPEX, a renal the observers were resolved by conference during central biopsy was performed in all 124 patients in this study. reviews, achieving consensus for each biopsy. Of the combined 256 biopsies from these two studies, Immunofluorescence for immunoglobulins and com- 180 were received for re-evaluation. Seven of these 180 plement components, and electron microscopy were per- biopsies were excluded: five biopsies because of the ab- formed locally in all participating centers. For logistical sence of cortical tissue, and two because patients ap- reasons, these procedures were not standardized or re- peared to have antibodies directed against the glomeru- viewed by the EUVAS, but the local conclusions were lar basement membrane. Together, 173 biopsies (98 from reported. CYCAZAREM and 75 from MEPEX) were available for evaluation in the present study. Patient diagnoses Diagnosis (as defined in the Methods section) were microscopic Meeting the inclusion criteria was needed to enter polyangiitis (N ϭ 80), renal limited vasculitis (N ϭ 19), patients in either CYCAZAREM or MEPEX. Diagnos- and Wegener’s granulomatosis (N ϭ 73). The diagnosis tic categories were distinguished (always with respect to of one patient remained uncertain (doubt about whether the inclusion criteria, Tables 1 and 2) by the following extrarenal manifestations were attributable to vasculi- criteria [10] and as such determined by the local physi- tis). ANCA test results are listed in Table 3. The mean cians, without review by the EUVAS: age of all patients was 59 years (range 21 to 83 years). Patients with MPO-ANCA were older than patients with • Microscopic polyangiitis: Systemic vasculitis, pre- PR3-ANCA (63 vs. 58 years, P ϭ 0.033). Patients in- dominantly affecting small vessels, with extrarenal cluded 56% females and 44% males. The gender distri- symptoms, but without airway symptoms compati- bution was comparable in the various groups. Mean se- ble with Wegener’s granulomatosis. rum creatinine and GFR of all patients were 453 ␮mol/L • Renal limited vasculitis: Idiopathic rapidly progres- and 31 mL/min, respectively. Patients with microscopic sive glomerulonephritis without systemic disease polyangiitis and renal limited vasculitis had a signifi- manifestations. cantly lower mean GFR than patients with Wegener’s • Wegener’s granulomatosis: Inflammation of the re- granulomatosis (20 and 22 vs. 44 mL/min, P Ͻ 0.001). spiratory tract together with necrotizing vasculitis Patients with a P-ANCA IIF had a significantly lower affecting small- to medium-sized vessels. mean GFR than patients with a C-ANCA IIF (21 vs. 40 mL/min, P Ͻ 0.001). Patients with MPO-ANCA had ANCA testing a significantly lower mean GFR than patients with Indirect immunofluorescence (IIF) and ELISA were PR3-ANCA (21 vs. 38 mL/min, P ϭ 0.002). performed for local ANCA testing in all participating centers. The staining pattern in the IIF test was scored Histological diagnoses as perinuclear (P-ANCA), cytoplasmic (C-ANCA), atypi- Figure 1 shows the histopathological diagnoses on the cal, or negative. Sera positive for ANCA directed against basis of the renal biopsy findings. Crescentic necrotizing MPO and PR3 were reported as MPO-ANCA and PR3- glomerulonephritis was observed in 112 biopsies (65%); EUVAS: Renal histology in vasculitis 83

Table 3. ANCA test results of biopsies and in approximately one third of those cases Diagnosis N was scored as diffuse (ϩϩ). Tubular intra-epithelial in- IIF ELISA MPA RLV WG Uncertain filtrates were seen in 64% of biopsies. Interstitial vasculitis was present in only 12% of biopsies. Arteriosclerosis P-ANCA MPO-ANCA 35 11 9 0 PR3-ANCA 1 0 0 0 and arteriolosclerosis were present in 70% and 32% of MPO-ANCA & PR3-ANCA 2 0 0 0 biopsies, respectively. Pauci-immunity was concluded in Negative 5 0 0 0 all patients. Electron microscopy was performed in some Not reported 5 2 1 0 C-ANCA MPO-ANCA 0 0 0 0 patients, but histopathological conclusions remained un- PR3-ANCA 17 2 40 0 changed with this procedure. MPO-ANCA & PR3-ANCA 0 0 2 0 Negative 2 0 4 0 Differences between diagnoses Not reported 5 1 6 0 Atypical ANCA MPO-ANCA 1 0 0 0 Table 4 shows the mean percentage of the main glo- PR3-ANCA 0 0 0 0 merular parameters in microscopic polyangiitis, renal MPO-ANCA & PR3-ANCA 0 0 0 0 Negative 0 0 0 0 limited vasculitis, and Wegener’s granulomatosis. In mi- Not reported 0 0 0 0 croscopic polyangiitis, 21% of glomeruli per biopsy were Negative MPO-ANCA 0 0 0 0 normal, whereas in Wegener’s granulomatosis 40% of PR3-ANCA 0 0 0 0 MPO-ANCA & PR3-ANCA 0 0 0 0 glomeruli were normal (Fig. 2A, P Ͻ 0.001). Glomerulo- Negative 1 1 1 0 sclerosis was present in 30% of patients who were diag- Not reported 1 0 2 0 nosed with microscopic polyangiitis versus only 16% of Not reported MPO-ANCA 0 1 1 0 PR3-ANCA 0 0 3 0 patients with Wegener’s granulomatosis (Fig. 2A, P ϭ MPO-ANCA & PR3-ANCA 0 0 1 0 0.002). The percentage of glomeruli with fibrinoid necro- Negative 1 0 0 0 sis or crescents did not distinguish between the diagnoses. Not reported 4 1 3 1 In microscopic polyangiitis, interstitial fibrosis (Fig. 2B) Abbreviations are in the Appendix. and tubular atrophy (Fig. 2C) were more frequently and severely present than in Wegener’s granulomatosis (P Ͻ 0.001 for both differences). Interstitial infiltrates (Fig. crescentic glomerulonephritis without fibrinoid necrosis 2D) and tubular necrosis (Fig. 2E) tended to occur more was present in 40 cases (23%). Two biopsies showed frequently in microscopic polyangiitis than in Wegener’s interstitial vasculitis only, one biopsy was normal, and granulomatosis (P ϭ 0.030 and P ϭ 0.035, respectively). nine biopsies (5%) showed diffuse global glomeruloscle- Tubular casts were more often reported in microscopic rosis. Other biopsies included focal global glomeruloscle- polyangiitis than in Wegener’s granulomatosis (P ϭ 0.005). rosis, interstitial fibrosis, and mesangial matrix increase. However, there were no differences in proteinuria between the diagnostic groups (data not shown). Arterio- Histological parameters in all patients sclerosis was present more often in microscopic polyangi- The histological findings in the entire patient cohort itis than in Wegener’s granulomatosis (P ϭ 0.021). No showed that a mean of 29% of glomeruli per biopsy were difference was observed in the occurrence of interstitial normal, 45% of glomeruli had (predominantly cellular) vasculitis. Biopsies from the 19 patients with renal lim- crescents, and 23% were globally sclerotic. Fibrinoid ne- ited vasculitis were more or less comparable to those crosis of the glomerular tuft was present in 22% of glo- from the patients with microscopic polyangiitis (Table 4 meruli. In five biopsies fibrinoid necrosis was present and Fig. 2). in glomeruli without crescents, although crescents were found in other glomeruli in the same biopsy. Interstitial Differences between MPO-ANCA and PR3-ANCA edema was present in 34% of biopsies. Interstitial infil- The distribution of histological parameters among the trates were present in 92% of biopsies; in half of those ELISA test result groups is presented in Table 5. The scored as mild (ϩ), in the other half as severe (ϩϩ). mean percentage glomerulosclerosis was 15% in patients Mononuclear inflammatory cells were always present, with PR3-ANCA, 25% in patients with MPO-ANCA, and usually, they formed the predominant part of the and 35% in patients with negative ELISA test results infiltrates. Neutrophils were present in 65% of biopsies, (P ϭ 0.022 for the difference between PR3-ANCA and but predominant in only a few cases. Although eosino- MPO-ANCA, P ϭ 0.005 for the difference between neg- phils were seen in 22% of biopsies, they were never ative test results and PR3-ANCA; Fig. 3A). The occur- scored as predominant. Interstitial fibrosis was present rence of fibrinoid necrosis and crescents was similar in in 83% of biopsies and in approximately half of those all groups. In patients with MPO-ANCA positivity, inter- cases it was scored as diffuse (ϩϩ). Tubular casts and stitial fibrosis (Fig. 3B) and tubular atrophy (Fig. 3C) tubular necrosis were present in 87% and 66% of biop- were more frequently and more severely present than sies, respectively. Tubular atrophy was present in 86% in patients with PR3-ANCA positivity (P ϭ 0.008 and 84 EUVAS: Renal histology in vasculitis

Fig. 1. Histopathological diagnoses on the basis of the renal biopsy ﬁndings.

Table 4. Main glomerular lesions in MPA, RLV, and WG

MPA (Nϭ80) RLV (Nϭ19) WG (Nϭ73) P value Lesions Mean SD Mean SD Mean SD MPA vs. RLV RLV vs. WG MPA vs. WG Normal glomeruli 21 22 27 25 40 34 0.291 0.137 Ͻ0.001 Fibrinoid necrosis 21 25 21 26 23 26 0.976 0.656 0.496 Crescentsa 47 28 49 33 42 32 0.790 0.365 0.257 Cellular crescents 45 28 55 27 45 29 0.210 0.205 0.902 Fibrous crescents 7 10 3 6 4 9 0.239 0.715 0.172 Glomerulosclerosis 30 29 23 25 16 23 0.361 0.265 0.002 Periglomerular inﬁltrates 10 16 16 20 9 12 0.121 0.033 0.680 Granulomatous reaction 4 13 2 6 2 8 0.566 0.918 0.379 a The mean percentage of cellular crescents and ﬁbrous crescents together is usually different from the mean percentage of crescents, as the former two lesions were not scored if crescents were absent

P ϭ 0.013, respectively). Tubular necrosis (Fig. 3E), tu- with MPO-ANCA versus patients with PR3-ANCA (data bular intra-epithelial infiltrates (Fig. 3F), and arterioscle- not shown). rosis were more frequently present in patients with MPO-ANCA than in patients with PR3-ANCA (P ϭ Sample size 0.030, P ϭ 0.006, and P ϭ 0.027, respectively). The histo- Reported results are based on analyses without a mini- pathology in the group of patients with negative ELISA mum for the number of glomeruli in the biopsies (biop- test results was more or less similar to that of MPO- sies without cortical tissue were excluded). We tested ANCA positive patients (Table 5 and Fig. 3). It is ques- whether the presence of only a small number of glomer- tionable whether these results are independent of the uli—less than five and less than ten—would influence diagnostic subgroup. To address this issue, we tested our results, but this was not the case (data not shown). whether ANCA specificity influenced renal histology in- dependently of diagnosis. The observed differences in the occurrence of tubular necrosis (P ϭ 0.021) and tubu- DISCUSSION lar intra-epithelial infiltrates (P ϭ 0.005) between MPO- We present the results of a prospective multicenter ANCA and PR3-ANCA positive patients remained sig- study of 173 renal biopsies from a homogeneous group nificant after this correction. No differences between the of patients with newly diagnosed renal disease in relation groups were observed in the occurrence of interstitial to ANCA-associated vasculitis. Our aim was to deter- vasculitis. We also investigated the distribution patterns mine (1) the differences in the occurrence of renal histo- of histological features in patients with P-ANCA versus logical lesions in microscopic polyangiitis, Wegener’s patients with C-ANCA, and of P- and MPO-ANCA– granulomatosis, and renal limited vasculitis; and (2) the positive patients versus C- and PR3-ANCA–positive pa- differences in the occurrence of renal histological lesions tients. Results of these analyses were similar to the above- in patients with ANCA-associated vasculitis with various described histological distribution patterns in patients ANCA test results. EUVAS: Renal histology in vasculitis 85

Fig. 2. Renal histology in patients with microscopic polyangiitis, renal limited vasculitis, or Wegener’s granulomatosis. Differences in presence of glomerular lesions (A), interstitial fibrosis (B), tubular atrophy (C), interstitial infiltrates (D), tubular necrosis (E), and tubular intra-epithelial infiltrates (F), between the diagnostic categories are presented. Symbols in panel A are: (᭿) other; ( ) sclerosis; ( ) crescents; ( ) normal glomeruli. Symbols in panels B and C are: ( ) diffuse; ( ) focal; ( ) absent. Panel D symbols are: (᭿) very dense; ( ) dense; ( ) mild; ( ) absent. Symbols in panels E and F are: ( ) present; ( ) absent. 86 EUVAS: Renal histology in vasculitis

Table 5. Main glomerular lesions in patients with MPO-ANCA, PR3-ANCA, double positivity, and negative ELISA MPO- & Negative MPO- PR3- PR3- ELISA ANCA ANCA ANCA P valueb (Nϭ15) (Nϭ58) (Nϭ63) (Nϭ5) Negative vs. Negative vs. MPO-ANCA vs. Lesions Mean SD Mean SD Mean SD Mean SD MPO-ANCA PR3-ANCA PR3-ANCA Normal glomeruli 20 31 26 23 31 30 29 37 0.402 0.188 0.265 Fibrinoid necrosis 19 22 23 23 24 27 32 41 0.510 0.529 0.925 Crescentsa 45 32 47 26 50 32 70 37 0.766 0.525 0.514 Cellular crescents 46 29 46 24 50 30 70 37 0.907 0.671 0.357 Fibrous crescents 5 9 6 10 6 9 0 0 0.776 0.836 0.880 Glomerulosclerosis 35 34 25 26 15 21 1 2 0.222 0.005 0.022 Periglomerular inﬁltrates 8 11 11 18 8 12 16 18 0.554 0.901 0.219 Granulomatous reaction 1 3 5 15 1 4 0 0 0.401 0.969 0.092 a The mean percentage of cellular crescents and ﬁbrous crescents together is usually different from the mean percentage of crescents as the former two lesions were not scored if crescents were absent b P values of negative, MPO-ANCA, and PR3-ANCA vs. MPO-ANCA positivity are not reported (in general, P values were over 0.100)

Glomerulosclerosis, interstitial fibrosis, tubular atro- ing for the GFR), but no significant differences were phy, tubular casts, and arteriosclerosis occurred more observed (data not shown). often and more pronounced in microscopic polyangiitis In this study, ANCA positivity was present in 149 of and renal limited vasculitis than in Wegener’s granulo- 164 patients in which ANCA test results were reported. matosis. However, none of these lesions were specific for Patients with MPO-ANCA had significantly more glo- one of the diagnostic subgroups, which limits the role of merulosclerosis, interstitial fibrosis, tubular atrophy, and renal histopathology in further sub-categorizing ANCA- arteriosclerosis, but also significantly more tubular ne- associated vasculitis. Except for tubular casts, the above- crosis and tubular intra-epithelial infiltrates. In other mentioned lesions are representative of a chronic disease words, a mixture of both chronic and active lesions was phase, reflecting the result of longer existing inflamma- more abundantly present in patients with MPO-ANCA tion of the kidney, in which irreversible scars develop. than in patients with PR3-ANCA. The observed differ- So-called active lesions (such as tubular intra-epithelial ences between the occurrence of tubular necrosis and infiltrates and glomerular crescents) did not discriminate tubular intra-epithelial infiltrates remained significant between the three diagnostic subgroups. We believe that after correction for diagnosis. One should realize, how- the explanation for these results is that in patients with ever, that the observed differences are small, that none microscopic polyangiitis and renal limited vasculitis, the of the lesions is specific for one of the groups, and that diagnosis is established later than in patients with Weg- active glomerular lesions (that is, fibrinoid necrosis and ener’s granulomatosis. In over 90% of cases, Wegener’s cellular crescents) were present to the same extent. Our granulomatosis presents with upper airway or pulmonary results may be explained by the finding that circulating symptoms, or both [13], which probably reduces the pa- soluble MPO, released by activated polymorphonucle- tient’s and doctor’s delay. Alternatively, although such ars, may bind to resting “innocent bystander polymor- a mechanism is not known in vasculitis, our results could phonuclears,” thereby making them reactive to anti- also indicate that, at the same time point after onset of MPO antibodies, which leads to perpetuating inflamma- the disease, patients with renal limited vasculitis and tion and tissue destruction [14]. In addition, anti-MPO with microscopic polyangiitis may have developed more antibodies may inhibit the inactivation of MPO by ceru- irreversible lesions in the kidney than patients with Weg- loplasmin, its natural inhibitor [15], resulting in persis- ener’s granulomatosis. This may be explained by a faster tent generation of reactive oxygen radicals with en- progression of active lesions to chronic lesions in micro- hanced potential for endothelial cytotoxicity [16]. These scopic polyangiitis and renal limited vasculitis or, alterna- processes may lead to a more active disease, and in tively, by development of chronic lesions without the the long term, to more chronic damage in MPO-ANCA obligatory presence of active lesions. An argument in positive patients than in PR3-ANCA-positive patients. favor of the occurrence of the latter mechanisms would The previously reported tendency to a more frequent be that, after correction for the diagnostic delay, the renal involvement in patients with MPO-ANCA than observed differences in renal histology between the diag- in patients with PR3-ANCA is further evidence for a noses are still present. Therefore, we re-analyzed our stronger renal pathogenic effect of MPO-ANCA com- results by correcting for the diagnostic delay (by correct- pared to PR3-ANCA [12, 17]. EUVAS: Renal histology in vasculitis 87

Fig. 3. Renal histology in patients with PR3-ANCA, MPO-ANCA, PR3- and MPO-ANCA, or negative ELISA. Differences in presence of glomerular lesions (A), interstitial fibrosis (B), tubular atrophy (C), interstitial infiltrates (D), tubular necrosis (E), and tubular intra-epithelial infiltrates (F) between the ELISA test result groups are illustrated. Symbols in panel A are: (᭿) other; ( ) sclerosis; ( ) crescents; ( ) normal glomeruli. Symbols in panels B and C are: ( ) diffuse; ( ) focal; ( ) absent. Panel D symbols are: (᭿) very dense; ( ) dense; ( ) mild; ( ) absent. Symbols in panels E and F are: ( ) present; ( ) absent.

Previous investigations on the occurrence of renal his- in a retrospective analysis of 36 patients with ANCA- topathological differences between patients with MPO- associated vasculitis, that patients with PR3-ANCA posi- ANCA and PR3-ANCA led to contradictory results. In tivity had more fibrinoid necrosis and a higher activity three reports, no differences between these groups were index than MPO-ANCA positive patients, and that pa- found (abstract; ibid) [7, 8]. Franssen et al described tients with MPO-ANCA had more glomerulosclerosis 88 EUVAS: Renal histology in vasculitis and a higher chronicity index [9], which may be explained P. Landais (Hoˆ pital Necker, Paris, France), Ph. Vanhille (Centre Hospi- talier de Valenciennes, Valenciennes, France), P. Bataille (Centre Hospi- by the inhibitory effect by PR3-ANCA binding to PR3 talier General, Bologne sur Mer, France), V. Esnault (CHU Hotel Dieu, on the irreversible inactivation of PR3 by ␣1-antitrypsin Nantes, France), F. van der Woude, R. Waldherr, W. Schmitt, K. An- [18], similar to the mechanism by which MPO-ANCA drassy, O. Hergesell, R. Nowack (Heidelberg University Hospital, Hei- delberg, Germany), K. de Groot, K. Herlyn, W.L. Gross (Rheumaklinik, inhibits inactivation of MPO by ceruloplasmin [19, 20]. Bad Bramstedt, Germany), K.A. Boki, J.N. Boletis (Laikon General In addition, the respiratory burst following activation of Hospital, Athens, Greece), D.S. Emmanouel (University Hospital Crete, tumor necrosis factor-␣ (TNF-␣) primed neutrophils in Heraklion, Greece), C. Feighery (Saint James’s Hospital, Dublin, Ire- land), F. Ferrario, R.A. Sinico (Ospedale San Carlo Borromeo, Milan, vitro by PR3 may be more potent than activation by Italy), G. Gregorini (Spedale Civili, Brescia, Italy), J. Dadoniene´ (Uni- MPO [21], though this difference is not evident in other versity Hospital Vilnius, Vilnius, Lithuania), E.C. Hagen (Eemland reports [22–24]. Hospital, Amersfoort, The Netherlands), C.G.M. Kallenberg, C. Stege- man (University Hospital Groningen, Groningen, The Netherlands), It seems that, although differences exist between J.W. Cohen Tervaert (University Hospital Maastricht, Maastricht, The PR3-ANCA– and MPO-ANCA–associated vasculitis [6], Netherlands), C.A. Verburgh, C.E.H. Siegert, J.A. Bruijn, H.A. Hauer, current knowledge on the mechanisms by which these J. Hermans, J.C. van Houwelingen, C.E. Vergunst (Leiden University Medical Center, Leiden, The Netherlands), E. van Gurp (University ANCAs are involved in the pathogenesis of vasculitis is Medical Center Utrecht, Utrecht, The Netherlands), I.M. Bajema (Dijk- insufficient for the explanation of these differences. Alter- zigt Hospital, Rotterdam, The Netherlands), C. Vasconcelos (Hospital natively, the action of ANCA is just an epiphenomenon. San Antońio, Porto, Portugal), E. Mirapeix, M. Sole´ (Hospital Clinic I Provincial, Barcelona, Spain), E.E. Pettersson, A. Bruchfeld (Hud- In summary, we conclude that glomerulonephritis in dinge University Hospital, Huddinge, Sweden), K.W.A. Westman (Uni- relation to microscopic polyangiitis has more characteris- versity Hospital of Lund, Lund, Sweden), Z. Heigl (Karolinska Hospi- tics of chronic injury at the time of presentation than tal, Stockholm, Sweden), C. Chizzolini (University Hospital Geneva, Geneva, Switzerland), D. MacLahan (University Hospital, Basle, Swit- glomerulonephritis in relation to Wegener’s granuloma- zerland). Pathologists who provided biopsy material: M. Depierreux tosis. We suggest that this difference is due to a delayed (Academic Hospital of the Free University, Bruxelles, Belgium), B. van establishment of diagnosis in patients with microscopic Damme (University Hospital Leuven, Leuven, Belgium), A. Stejska- lova´, Z. Vernerova´ (Charles University, Prague, Czech Republic), T. polyangiitis compared to patients with Wegener’s granu- To¨ rnroth (University of Helsinki, Helsinki, Finland), A.C. Feller (Uni- lomatosis. The finding that both active and chronic lesions versity of Luebeck, Luebeck, Germany), E. Gaffney (Saint James’s are more abundantly present in MPO-ANCA-positive Hospital, Dublin, Ireland), R. Tardanico (Ospedale Civili, Brescia, It- aly), R. Consalonieri (Ospedale Maggiore CA Granda, Milan, Italy), patients than in patients with PR3-ANCA positivity may G. Garibotto (ISUL, Genova, Italy), A.T.M.G. Tiebosch (Academic be an argument in favor of the existence of different Hospital Groningen, Groningen, The Netherlands), C.D. Kooijman routes in the pathogenesis of renal disease in these ANCA (Eemland Hospital, Amersfoort, The Netherlands), M. Sole Arques (Hospital Clinic I Provincial de Barcelona, Barcelona, Spain), F. Algaba subsets, although the mechanism remains unclear. In (Puigvert, Barcelona, Barcelona, Spain), M. Carreras (Hospital de Bell- addition, our results suggest that ANCA test results may vitge, Barcelona, Spain), M. Vaquero Perez (Hospital Universitari Ger- be useful in further classifying ANCA-associated vasculi- mans Trias I Pujol, Badalona, Spain), L. Bernardo (Hospital Dr. Josep Trueta, Girona, Spain), B. Sundelin (Karolinska Hospital, Stockholm, tides. Within the framework of the EUVAS, the pheno- Sweden), P. Alm (University Hospital of Lund, Lund, Sweden), A. type of patients with systemic disease, differentiation of Wernersson (Huddinge University Hospital, Sweden), B. Veress (Uni- the leukocyte infiltration, and the predictive value of the versity Hospital Malmø, Malmø, Sweden), W. Landells (Saint Helier Hospital, London, UK), A.J. Howie (University Hospital Birmingham, renal histopathological parameters in this patient cohort Birmingham, UK), S. Fleming (University Department of Pathology, will be investigated in the near future. Edinburgh, UK), A.P. Griffith (Morriston Hospital, Swansea, UK), P.N. Furness (Leicester Area Histopathology Service, Leicester, UK), H.T. Cook (Hammersmith Hospital, London, UK). ACKNOWLEDGMENTS The EUVAS group was supported by grants from the EU (Contract Reprint requests to H.A. Hauer, M.D., Department of Pathology, nos. BMH1-CT93-1078, CIPDCT94-0307, BMH4-CT97-2328, and Leiden University Medical Center, P.O. Box 9600, Building 1, L1-Q, ERBIC20-CT97-0019). 2300 RC Leiden, The Netherlands. The following investigators participated in the European Vasculitis E-mail: [email protected] Study Group (EUVAS): Co-ordinator and project leader: N. Rasmus- sen (Rigshospitalet, Copenhagen, Denmark). Co-project leader: D.R.W. Jayne (Addenbrooke’s Hospital, Cambridge, UK). EUVAS Members: APPENDIX I. Neumann (Wilhelmina Hospital, Vienna, Austria), D. Abramowicz (Erasmus Hospital, Bruxelles, Belgium), Ph. Madhoun (Edith Cavell Abbreviations used in this study are: ANCA, anti-neutrophil cyto- Medical Institute, Bruxelles, Belgium), J. Sennesael (Academic Hospital plasm autoantibodies; C-ANCA, ANCA with a cytoplasmic staining of the Free University, Bruxelles, Belgium), V. Tesar, I. Rychlik, J. Bartun- as determined by indirect immunofluorescence; CYCAZAREM, ran- kova, J. Luka´s(Charles University Hospital, Prague, Czech Republic), domized trial of CYClophosphamide versus AZAthioprine during B. Ravn Juhl, J. Petersen, C.B. Andersen, and W. Szpirt (Rigshospitalet, REMission of ANCA-positive systemic vasculitis; ELISA, enzyme- Copenhagen, Denmark), A. Wiik (Statens Serum Institut, Copenhagen, linked immunosorbent assay; EUVAS, EUropean VAsculitis Study Denmark), H. Løkkegaard, H. Nielsen (Herlev County Hospital, Den- group; GFR, glomerular filtration rate; IIF, indirect immunofluores- mark), T. Ring (Aalborg Hospital, Aalborg, Denmark), S. Freiesleben cence; MEPEX, randomized trial of adjunctive therapy for severe Sørensen (Bispebjerg Hospital, Bispebjerg, Denmark), P.A. Bacon, A. glomerulonephritis in ANCA-associated systemic vasculitis–intravenous Exley, C.O.S. Savage (University Hospital Birmingham, Birmingham, administration of MEthylprednisolone versus Plasma EXchange; MPA, UK), G. Gaskin, C. Pusey (Hammersmith Hospital, London, UK), C.M. microscopic polyangiitis; MPO, myeloperoxidase; MPO-ANCA, ANCA Lockwood† (Addenbrooke’s Hospital, Cambridge, UK), R. Luqmani directed against myeloperoxidase as determined by ELISA; P-ANCA, (Western General Hospital, Edinburgh, UK), C. Gro¨ nhagen-Riska, A. ANCA with a perinuclear staining as determined by IIF; PR3, protein- Ekstrand (Helsinki University Hospital, Helsinki, Finland), L. Guillevin, ase-3; RLV, renal limited vasculitis; SD, standard deviation; WG, Weg- F. Lhote (Hoˆpital Avicenne, Bobingy, France), Ph. Lesavre, L.H. Noe¨l, ener’s granulomatosis. EUVAS: Renal histology in vasculitis 89

REFERENCES 14. Hess C, Sadallah S, Schifferli JA: Induction of neutrophil re- sponsiveness to myeloperoxidase antibodies by their exposure to 1. Petterson E, Sundelin B, Heigl Z: Incidence and outcome of supernatant of degranulated autologous neutrophils. Blood 96: pauci-immune necrotizing and crescentic glomerulonephritis in 2822–2827, 2000 adults. Clin Nephrol 43:141–149, 1995 15. Segelmark M, Persson B, Hellmark T, Wieslander J: Binding 2. Bajema IM, Hagen EC, van der Woude FJ, Bruijn JA: Wegener’s and inhibition of myeloperoxidase (MPO): A major function of granulomatosis: a meta-analysis of 349 literary case reports. J Lab ceruloplasmin? Clin Exp Immunol 108:167–174, 1997 Clin Invest 129:17–22, 1997 16. Griffin SV, Chapman PT, Lianos EA, Lockwood CM: The inhibi- 3. Jennette JC, Falk RJ, Andrassy K, et al: Nomenclature of sys- tion of myeloperoxidase by ceruloplasmin can be reversed by anti- temic vasculitides. Proposal of an international consensus confer- myeloperoxidase antibodies. Kidney Int 55:917–925, 1999 ence. Arthritis Rheum 37:187–192, 1994 17. Franssen C, Gans R, Kallenberg C, et al: Disease spectrum of 4. Falk RJ, Jennette JC: ANCA small-vessel vasculitis. J Am Soc patients with antineutrophil cytoplasmic autoantibodies of defined Nephrol 8:314–322, 1997 specificity: Distinct differences between patients with anti-protein- 5. Hagen EC, Daha MR, Hermans J, for the EC BCR Project of ase 3 and anti-myeloperoxidase autoantibodies. J Intern Med 244: ANCA Assay Standardization: Diagnostic value of standardized 209–216, 1998 assays for anti-neutrophil cytoplasmic antibodies in idiopathic sys- 18. Dolman KM, Stegeman CA, van de Wiel BA, et al: Relevance temic vasculitis. Kidney Int 53:743–753, 1998 of classic anti-neutrophil cytoplasmic autoantibody (C-ANCA)- 6. Franssen CF, Stegeman CA, Kallenberg CG, et al: Antiprotein- mediated inhibition of proteinase 3-alpha 1-antitrypsin complex- ase 3- and antimyeloperoxidase-associated vasculitis. Kidney Int ation to disease activity in Wegener’s granulomatosis. Clin Exp 57:2195–2206, 2000 Immunol 93:405–410, 1993 7. Jennette JC, Wilkman AS, Falk RJ: Anti-neutrophil cytoplasmic 19. van de Wiel BA, Dolman KM, van der Meer-Gerritsen CH, et autoantibody-associated glomerulonephritis and vasculitis. Am J al: Interference of Wegener’s granulomatosis autoantibodies with Pathol 135:921–930, 1989 neutrophil proteinase 3 activity. Clin Exp Immunol 90:409–414, 1992 8. Fienberg R, Mark EJ, Goodman M, et al: Correlation of antineu- 20. Daouk GH, Palsson R, Arnaout MA: Inhibition of proteinase trophil cytoplasmic antibodies with the extrarenal histopathology 3 by ANCA and its correlation with disease activity in Wegener’s of Wegener’s (pathergic) granulomatosis and related forms of vas- granulomatosis. Kidney Int 47:1528–1536, 1995 culitis. Hum Pathol 24:160–168, 1993 21. Franssen CFM, Huitema MG, Kobold ACM, et al: In vitro neutro- 9. Franssen CF, Gans RO, Arends B, et al: Differences between phil activation by antibodies to proteinase 3 and myeloperoxidase anti-myeloperoxidase- and anti-proteinase 3 associated renal dis- from patients with crescentic glomerulonephritis. J Am Soc Nephrol ease. Kidney Int 47:193–199, 1995 10:1506–1515, 1999 10. Jayne DRW, Rasmussen N: Treatment of antineutrophil cyto- 22. Kettritz R, Schreiber A, Luft FC, Haller H: Role of mitogen- plasm autoantibody-associated systemic vasculitis: Initiatives of activated protein kinases in activation of human neutrophils by anti- the European Community Systemic Vasculitis Clinical Trials Study neutrophil cytoplasmic antibodies. J Am Soc Nephrol 12:37–46, Group. Mayo Clin Proc 72:737–747, 1997 2001 11. Bajema IM, Hagen EC, Hansen BE, et al: The renal histopathol- 23. Falk RJ, Terrell RS, Charles LA, Jennette JC: Anti-neutrophil ogy in systemic vasculitis: An international survey study of inter- cytoplasmic autoantibodies induce neutrophils to degranulate and and intra-observer agreement. Nephrol Dial Transplant 11:1989– produce oxygen radicals in vitro. Proc Natl Acad Sci USA 87:4115– 1995, 1996 4119, 1990 12. Bajema IM, Hagen EC, Hermans J, et al: Kidney biopsy as a 24. Reumaux D, Vossebeld PJ, Roos D, Verhoeven AJ: Effect of predictor for renal outcome in ANCA-associated necrotizing glo- tumor necrosis factor-induced integrin activation on Fc gamma merulonephritis. Kidney Int 56:1751–1758, 1999 receptor II-mediated signal transduction: Relevance for activation 13. Jennette JC, Falk RJ: Small-vessel vasculitis. N Engl J Med 337: of neutrophils by anti-proteinase 3 or anti-myeloperoxidase anti- 1512–1523, 1997 bodies. Blood 86:3189–3195, 1995