Disease spectrum of patients with antineutrophil cytoplasmic autoantibodies of defined specificity: distinct differences between patients with anti-proteinase 3 and anti-myeloperoxidase autoantibodies


Franssen Department of Internal Medicine, Division of Nephrology, University Hospital Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands (fax: +31 50 3619310; e-mail: m.j.s.graler@int.azg nl).


Franssen C, Gans R, Kallenberg C, Hageluken C, Hoorntje S (University Hospital, Groningen; Free University Hospital, Amsterdam; and Catharina Hospital, Eindhoven; the Netherlands). Disease spectrum of patients with antineutrophil cytoplasmic autoantibodies of defined specificity: distinct differences between patients with antiproteinase 3 and antimyeloperoxidase autoantibodies. J Intern Med 1998; 244: 209–16.


To compare the disease spectrum of consecutive patients with antineutrophil cytoplasmic autoantibodies directed against proteinase 3 (anti-PR3) or myeloperoxidase (anti-MPO).


Retrospective analysis.


Three teaching hospitals in the Netherlands.

Main outcome measures

Clinical features at presentation, histopathological characteristics and outcome.


All consecutive patients who tested positive for anti-PR3 (n= 46) or anti-MPO (n= 46) over an 8-year-period.


At diagnosis, patients with anti-PR3 had a higher vasculitis activity index than patients with anti-MPO (P < 0.001). The mean (SD) number of affected organs in the anti-PR3 group exceeded that of the anti-MPO group (3.9 (1.4) and 2.2 (1.1), respectively; P < 0.01). The combination of renal and respiratory tract involvement was present in as many as 78.3% of patients with anti-PR3 and in only 23.9% of patients with anti-MPO (P < 0.01). Renal-limited disease exclusively occurred in patients with anti-MPO. Granulomas were found in 41.3% of anti-PR3- but in only 4.3% of anti-MPO-positive patients (P < 0.01). All anti-PR3-positive patients had Wegener's granulomatosis or microscopic polyangiitis. By contrast, diagnoses in the anti-MPO group were more diverse: idiopathic necrotizing crescentic glomerulonephritis (26.1%), microscopic polyangiitis (26.1%), Churg–Strauss syndrome (4.3%), Wegener's granulomatosis (2.2%), giant cell arteritis (2.2%), clinically suspected vasculitis (19.6%), as well as miscellaneous nonvasculitic disorders (19.6%). During follow-up, 10 anti-PR3-positive patients had 11 relapses whereas only 3 patients with anti-MPO relapsed (P= 0.04).


A large divergence was seen in the disease spectrum between patients with anti-PR3 and those with anti-MPO. In particular, extra-renal disease manifestations, granuloma formation and relapses were more prominent in anti-PR3- than in anti-MPO-positive patients.


Antineutrophil cytoplasmic autoantibodies (ANCA) represent a group of antibodies directed against myeloid lysosomal enzymes and are useful serological markers for vasculitis [ 1, 2]. Two major categories of ANCA can be recognized by indirect immunofluorescence microscopy (IIF) on ethanol-fixed neutrophils: a cytoplasmic (C-ANCA) and a perinuclear (P-ANCA) staining pattern [ [3][4]3–5]. Sera that produce the C-ANCA pattern nearly always react with proteinase 3 (PR3) [ 6, 7]. The major P-ANCA target is myeloperoxidase (MPO), but antibodies against other neutrophil antigens, such as elastase, lactoferrin and cathepsin G, and even antinuclear antibodies can produce a similar staining pattern [ [8][9][10][11]8–12]. In some P-ANCA sera the antigenic specificity remains elusive [ 5]. It follows that sera producing a P-ANCA pattern by IIF must be further characterized by antigen-specific assays [ 2, 5, 12, 13].

Antibodies against PR3 (anti-PR3) are predominantly found in patients with active Wegener's granulomatosis and microscopic polyangiitis [ 2, [14][15] 14–16]. Antibodies against MPO (anti-MPO) are also found in patients with microscopic polyangiitis as well as in patients with Churg–Strauss syndrome, idiopathic necrotizing crescentic glomerulonephritis, and drug-induced vasculitis [ 2, [17][18][19][20] 17–21]. Both anti-PR3 and anti-MPO have also been reported in extra-renal and localized forms of vasculitis [ [22][23][24]22–25]. Studies that have compared the clinical and histopathological associations between patients with anti-PR3 and anti-MPO are scarce. In most studies, patient selection criteria have included the presence of a renal biopsy showing necrotizing crescentic glomerulonephritis [ 1, 22, 25]. This patient selection may lead to under-representation of patients who lack renal involvement as well as patients who are not biopsied since they had only mild renal involvement and presumably followed a more favourable renal course. To better describe and compare the disease spectrum of anti-PR3- and anti-MPO-associated disease we analysed the clinical and histopathological features of consecutive anti-PR3- and anti-MPO-positive patients who were not selected for the presence of any particular organ manifestation. To this objective, we analysed all newly diagnosed patients who tested positive for anti-PR3 and anti-MPO over an 8-year-period in three hospitals in the Netherlands.

Materials and methods


The study was performed in three teaching hospitals in the Netherlands (Catharina Hospital, Eindhoven; Free University Hospital, Amsterdam; St Joseph Hospital, Veldhoven). Since January 1985, clinicians in these hospitals have been testing patients suspected of having vasculitis and/or glomerulonephritis for ANCA by IIF. As of January 1988, IIF-positive sera have been assayed at the Central Laboratory of The Netherlands Red Cross Blood Transfusion Service for the presence of antibodies against PR3, MPO and elastase. From the data files of this laboratory all patients from the above hospitals who had tested positive for anti-PR3 or anti-MPO between January 1988 and January 1993 were identified. In addition, we included all patients who had tested ANCA-IIF positive between January 1985 and January 1988, and whose stored sera revealed anti-PR3 or anti-MPO (n= 6). The case records of all patients were reviewed by two of the authors (CFMF and SJH). Patients with a documented episode of vasculitis or glomerulonephritis prior to January 1985 were excluded.

Serological methods

Serum was assayed in the standard ANCA-IIF test in accordance with the first international workshop on ANCA [ 4, 26]. The result of the test was considered positive when a granular cytoplasmic staining with central accentuation (C-ANCA) or a (peri)nuclear staining (P-ANCA) was present in the majority of the neutrophils at a serum dilution of 1:16 or more. Antigen-capture ELISAs with mouse monoclonal antibodies to PR3, MPO and elastase were performed as described elsewhere [ 6, 19, 27]. All sera were tested for the presence of antibodies against PR3, MPO and elastase.

Parameters and definitions

Clinical and histological data were collected from the patient charts using a standard protocol. Extra-renal organ involvement was categorized as follows [ 28]: Upper respiratory tract: nasal mucosal ulceration, serous otitis media, sinusitis, tracheal stenosis; Lower respiratory tract: pulmonary infiltrates, coin lesions, alveolar haemorrhage; Skin: palpable purpura, ulcers or nodules; Nervous system: mononeuritis multiplex, peripheral neuropathy; Musculoskeletal tract: arthralgia, arthritis, polymyalgia; Eyes: (epi)scleritis, keratitis, uveitis, retinal vasculitis; Heart: pericarditis, myocardial infarction, cardiomyopathy; Gastrointestinal tract: bowel perforation, aneurysms by abdominal angiography. Renal involvement was defined as microscopic haematuria (> 5 red blood cells per high-power field) and/or proteinuria (biuret method or albustix testing; lower detection limits 0.2 g L−1 and 0.3 g L−1, respectively) occurring on two separate occasions. Progressive renal function loss was defined as doubling of serum creatinine within the 3 months before the start of immunosuppressive treatment. All organ manifestations that occurred in a period of 6 months before a diagnosis of vasculitis was made were cumulatively scored. Organ manifestations were considered only if they could not be accounted for by infectious or atherosclerotic disease, or drug toxicity. For each patient the total number of affected organs was counted. The upper respiratory tract, lower respiratory tract and kidneys (organs typically affected in Wegener's granulomatosis; maximum count 3) and skin, nervous system, eyes, heart, musculoskeletal tract and gastrointestinal tract (other organ manifestations; maximum count 6) were counted separately. For each patient, a vasculitis activity score at the time of diagnosis was computed [ 29]. The results of the score ranged from 0 to 63. A relapse of vasculitis was defined as described elsewhere [ 30].

Diagnostic classification

Patients with histopathological proof of necrotizing vasculitis or necrotizing crescentic glomerulonephritis were classified according to the Chapel Hill Consensus Conference definitions [ 31]. Patients were subclassified as having limited Wegener's granulomatosis when they fulfilled the definition of Wegener's granulomatosis but renal involvement was lacking. Idiopathic necrotizing crescentic glomerulonephritis was defined as focal segmental or diffuse necrotizing glomerulonephritis with crescents but negative immunofluorescence studies and without extra-renal organ involvement [ 32]; patients with arthralgia, but no arthritis, as the only extra-renal manifestation were also included in this group.

Statistical analysis

Differences in the prevalence of organ involvement and the number of patients who had a relapse of vasculitis were tested with the chi-square test. Differences in the patients' ages, the number of affected organs and the vasculitis activity scores were assessed by means of the Wilcoxon rank-sum test. The patient survival curves were calculated using Kaplan–Meier estimates for survival distribution. The starting date of the survival curves was the date on which immunosuppressive treatment was instituted. If no specific treatment was given, the date of diagnosis served as the starting date. Differences between curves were tested with the log-rank test. A stepwise procedure, based on the log-rank test, was followed for testing the influence of the following parameters on the survival curves: age, sex and type of antibody and organ involvement. If needed, survival curves were corrected for influential parameters using a proportional hazard model. The level of significance used was 0.05.



Between January 1985 and January 1993, sera from 1780 patients were tested in the participating hospitals for the presence of ANCA by IIF. A total of 220 patients tested positive. Antigen-specific assays for anti-PR3, anti-MPO and anti-elastase were not performed in 48 patients (C-ANCA: 24; P-ANCA: 24) and were negative in 62 others (C-ANCA: 16; P-ANCA: 46). Those 110 patients were excluded from further analysis. The clinical diagnoses of those patients have been reported elsewhere [ 33]. In the remaining 110 patients ANCA with specificity for anti-PR3, anti-MPO or anti-elastase were present. Six of these patients had antibodies against more than one target antigen and were excluded (four had anti-PR3 and anti-MPO, one had anti-PR3 and anti-elastase, one had anti-MPO and anti-elastase). Twelve other patients were excluded because they had been diagnosed with vasculitis or glomerulonephritis before January 1985. The remaining 92 patients were included: 46 patients had anti-PR3 and 46 patients had anti-MPO. None had anti-elastase as the only antibody specificity.

Patients' characteristics and organ involvement

Patients with anti-PR3 were younger than patients with anti-MPO (median age 56 and 63 years, respectively; interquartile range 42–63 and 51–74 years, respectively; P < 0.05). There was no difference in the male/female ratio between the anti-PR3 (26/20) and the anti-MPO group (25/21). The organ involvement is shown in Fig. 1. The upper respiratory tract, the nervous system and the eyes were more often affected in the anti-PR3 than in the anti-MPO group. The total number of affected organs in the anti-PR3 group exceeded that of the anti-MPO group (median 4 and 2, respectively; interquartile range 3–5 and 1–3, respectively; P < 0.01). This was true both for organs typically affected in Wegener's granulomatosis (median 2 and 1, respectively; interquartile range 2–3 and 1–2, respectively; P < 0.01) as well as for the other organs (median 2 and 1, respectively; interquartile range 1–2 and 0–1.75, respectively; P < 0.01). Accordingly, the vasculitis activity score was higher in the anti-PR3 than in the anti-MPO group (median score 24 and 15, respectively; interquartile range 21–27 and 11–16, respectively; P < 0.001). As shown in Table 1, laboratory parameters at diagnosis did not differ between both antibody groups.

Figure 1.

Organ involvement according to antibody group. The solid bars denote the patients with anti-PR3; the hatched bars denote the patients with anti-MPO. Abbreviations: NS, not significant; RT, respiratory tract.

Table 1.  Laboratory findings at the time of diagnosis Thumbnail image of

The combination of renal and upper and/or lower respiratory tract involvement was found in as many as 36 patients (78.3%) with anti-PR3 but in only 11 patients (23.9%) with anti-MPO (P < 0.01). Renal-limited disease (with or without arthralgia) occurred in 20 patients (43.5%) with anti-MPO and in none of the patients with anti-PR3 (P < 0.01).

Histopathological findings and diagnostic classification

Table 2 shows the diagnostic classification of patients with histopathological proof of vasculitis and/or necrotizing crescentic glomerulonephritis according to the Chapel Hill Consensus Conference definitions. A diagnosis of Wegener's granulomatosis was predominantly found in the anti-PR3 group. Only one patient with anti-MPO met the histopathological criteria for limited Wegener's granulomatosis, i.e. presence of granuloma formation. Idiopathic necrotizing crescentic glomerulonephritis was exclusively seen in anti-MPO-positive patients.

Table 2.  Diagnostic classification Thumbnail image of

In 8 anti-PR3- and in 18 anti-MPO-positive patients there was no histopathological proof of vasculitis or necrotizing crescentic glomerulonephritis as their biopsies had revealed findings nondiagnostic for vasculitis (anti-PR3: 7 patients; anti-MPO: 10 patients) or because no biopsies were performed (anti-PR3: 1 patient; anti-MPO: 8 patients). The clinical features of these patients are shown in Table 3. All anti-PR3-positive patients had clinically suspected vasculitis. By contrast, diagnoses in the anti-MPO-positive patients included clinically suspected systemic (Patient numbers (nos) 1–4; 8.7%) or renal-limited vasculitis (nos 5–9; 10.9%) as well as nonvasculitic diseases (19.6%) such as postinfectious glomerulonephritis (nos 10, 11), rheumatic diseases (nos 12–15) and other nonvasculitic diseases (nos 16–18).

Table 3.  Clinical data of patients without histopathological proof of vasculitis or necrotizing crescentic glomerulonephritis Thumbnail image of


Except for one patient who was free of symptoms after local treatment of a subglottic mass, all anti-PR3-positive patients (98%) received immunosuppressive treatment: 39 patients (85%) had cyclophosphamide and corticosteroids and 6 patients (13%) had corticosteroids only. In the anti-MPO group, 33 patients (72%) were treated with immunosuppressive drugs: 23 (50%) had cyclophosphamide and corticosteroids, and 10 patients (22%) had corticosteroids only. Thirteen anti-MPO-positive patients (11%) were not treated with immunosuppressive drugs: one patient with renal-limited disease showing extensive sclerosis and fibrosis on renal biopsy; a 67-year-old female with renal involvement and transient vasculitis of the skin showing stable renal function at follow-up; two patients with renal-limited disease and slowly progressive renal function loss; two patients with postinfectious glomerulonephritis; a patient with limited Wegener's granulomatosis who responded to sulfamethoxazole-trimethoprim; a patient with subglottic stenosis who was treated with dilation of the subglottic stenosis; a patient with vasculitis confined to the skin successfully treated with 4,4'-diaminodiphenyl-sulfone; a patient with rheumatoid arthritis; a patient with sarcoidosis, and a patient with a bacterial pulmonary infection.


Patient survival did not differ significantly between patients with anti-PR3 and anti-MPO (P= 0.28). Survival at 2 years was 86% in the anti-PR3 and 78% in the anti-MPO group. The results were not affected when the survival analysis was corrected for the age difference between anti-PR3- and anti-MPO-positive patients or when the survival analysis was restricted to those patients who had been treated with immunosuppressive drugs or to those patients with biopsy-proven vasculitis or necrotizing crescentic glomerulonephritis (data not shown). When patients with anti-PR3 and anti-MPO were analysed together (n= 92) patients without renal involvement (n= 23) tended to have a better survival than patients with renal involvement (n= 69): survival at 2 years was 95% for patients without and 78% for patients with renal involvement (P= 0.06).

During follow-up, 10 patients with anti-PR3 had a total of 11 relapses whereas only 3 patients with anti-MPO had a relapse of vasculitis (P= 0.04). Of the variables tested the only significant variable which influenced patient survival was age. This parameter was inversely proportional to the probability of survival (P < 0.01).


Since its first description in 1982 [ 34], many studies on the diagnostic potential of ANCA have appeared, soon followed by studies on the diagnostic value of antibodies directed against the major ANCA target antigens, PR3 and MPO. Studies that have specifically compared clinical and histopathological features and outcome between anti-PR3- and anti-MPO-positive patients, however, are relatively scarce [ 1, 25, 35, 36]. In the present study patients with anti-PR3 had more widespread organ involvement and a higher vasculitis activity score at presentation than patients with anti-MPO. In particular, upper respiratory tract and peripheral nervous system involvement were more prominent in anti-PR3- than in anti-MPO-positive patients. Granulomatous inflammation was almost exclusively seen in anti-PR3-positive patients.

Different diagnoses prevailed in each antibody group. All anti-PR3-positive patients with biopsy-proven vasculitis had either Wegener's granulomatosis or microscopic polyangiitis. By contrast, diagnoses in anti-MPO-positive patients with biopsy-proven vasculitis were heterogeneous and included microscopic polyangiitis, idiopathic crescentic glomerulonephritis, Wegener's granulomatosis, Churg–Strauss syndrome and giant cell arteritis. We used a restrictive definition of Wegener's granulomatosis that required histopathological proof of extra-renal granulomatous inflammation. It should be noted, however, that Wegener's granulomatosis and microscopic polyangiitis are closely related and share the vasculitic involvement of the small-sized vessels [ 1, 2, 31]. The most important distinction between both conditions is granuloma formation. Histopathological data, however, are often inconclusive and granulomas may be missed in a biopsy due to sampling errors or incomplete work-up [ 37, 38]. Thus the classification of these patients depends largely on the policy of performing biopsies of affected organs.

All anti-PR3-positive patients had either biopsy-proven or clinically suspected vasculitis. By contrast, as many as one out of every five patients with anti-MPO had nonvasculitic diseases. These comprised postinfectious glomerulonephritis, rheumatoid arthritis, systemic lupus erythematosus, sarcoidosis, idiopathic colitis and bacterial pneumonia. Notably, a finding of anti-MPO is not equivalent to the presence of vasculitis and treatment should therefore not be instituted because of the finding of anti-MPO alone. Since immunosuppressive treatment is associated with significant morbidity and mortality, histopathological proof of vasculitis is desirable in both anti-PR3- and anti-MPO-positive patients.

In line with earlier observations, we found a comparable patient survival between patients with anti-PR3 and anti-MPO [ 1, 35]. Interestingly, the relapse rate was higher in the anti-PR3 than in the anti-MPO group. Other authors also suggested a higher relapse rate in patients with anti-PR3 than in patients with anti-MPO [ 35]. This finding may have implications for the duration of the maintenance phase of immunosuppressive treatment in anti-MPO- vs. anti-PR3-positive patients.

Why do patients with anti-PR3- and anti-MPO-associated vasculitis have different clinical and histopathological characteristics? If we assume that ANCA have a pathogenic role [ 2, 15, 39], one may speculate that anti-PR3 and anti-MPO interact in a different way with granulocytes and/or endothelial cells, influencing the site and extent of the resulting vascular injury. In addition, a difference in T-cell-mediated immunity of patients with anti-PR3- and MPO-associated vasculitis may play a role in the observed higher frequency of granuloma formation in anti-PR3- than in anti-MPO-positive patients. Interestingly, PR3-reactive T cells were observed more frequently in patients with anti-PR3 than in controls, whereas MPO-reactive T cells were observed at the same frequency in anti-MPO patients and controls [ 40].

In conclusion, there is a large divergence in the disease spectrum between patients with anti-PR3 and anti-MPO. Extensive extra-renal disease, granuloma formation and relapses are more prominent in anti-PR3- than in anti-MPO-positive patients. These findings may have consequences for follow-up and treatment of ANCA-positive patients.


The authors gratefully acknowledge the financial support from the Scientific Committee of the Catharina Hospital. They also thank K. M. Dolman (Central Laboratory of The Netherlands Red Cross Blood Transfusion Service, Amsterdam) for providing the data concerning the ANCA target antigens.