University of Groningen. Anca associated vasculitis Boomsma, Maarten Michiel

Transcription

1 University of Groningen Anca associated vasculitis Boomsma, Maarten Michiel IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2001 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Boomsma, M. M. (2001). Anca associated vasculitis: occurrence, prediction, prevention, and outcome of relapses Groningen: s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 Prediction of relapses in Wegener s granulomatosis by measurement of antineutrophil cytoplasmic antibody levels: a prospective study Maarten M. Boomsma 1, Coen A. Stegeman 2, Marcel J. van der Leij 1, Wia Oost 1, Jo Hermans 3, Cees G.M. Kallenberg 1, Pieter C. Limburg 4,1, Jan Willem Cohen Tervaert 1,2 Department of Internal Medicine, Division of 1 Clinical Immunology, 2 Nephrology and 4 Rheumatology, University Hospital Groningen and 3 Division of Medical Statistics, Leiden University Medical Center, The Netherlands Arthritis and Rheumatism 2000;43:

3 Abstract Prediction of relapses in Wegener s granulomatosis (WG) by measuring levels of antineutrophil cytoplasmic antibodies (ANCA) directed against proteinase 3 (PR3) or myeloperoxidase (MPO) remains a controversial issue. To assess the value of serial quantification of ANCA by indirect immunofluorescence (IIF) and antigen-specific enzymelinked immunosorbent assay (ELISA) for monitoring disease activity in patients with WG, a prospective observational study was conducted in patients with WG attending an outpatient clinic in the Netherlands. One hundred patients with WG (85 with PR3-ANCA, 15 with MPO- ANCA) were studied prospectively from 1996 to Serum samples were obtained and analyzed every 2 months for ANCA levels. Disease activity was prospectively assessed without knowledge of the ANCA levels. Relapses occurred in 37 of 100 patients (37%). Thirtyfour (92%) of the 37 patients showed a rise in the level of ANCA preceding their relapse, as detected by ELISA or IIF. The predictive value of an increase in ANCA titers for relapse was 57% (17 of 30) for cytoplasmic/classic ANCA (canca; by IIF), 71% (27 of 38) for PR3- ANCA (by ELISA), and 100% (3 of 3) for MPO-ANCA (by ELISA). The predictive value of a rise in ANCA as measured by ELISA or IIF did not substantially improve following concomitant measurement of the IgG3 subclass of PR3-ANCA. Forty-three percent of patients who showed a rise in canca (by IIF) and 29% with a rise in PR3-ANCA (by ELISA) did not subsequently experience a relapse. Serial measurement of ANCA levels is valuable for the early prediction of relapses in patients with WG. Introduction Wegener s granulomatosis (WG) is a systemic disease characterized by necrotizing granulomatous inflammation of the upper and lower respiratory tract, glomerulonephritis, and vasculitis (Godman et al., 1954). Prednisolone and cyclophosphamide treatment is very effective in inducing disease remission in WG (Hoffman et al., 1992). Once remission is achieved, medication is tapered. The majority of patients, however, experience relapses during tapering or when treatment has been stopped, and more than 80% of patients experience 1 or more relapses within 10 years (Hoffman et al., 1992). Since relapses are associated with morbidity and mortality, early detection and prediction of relapses is of great importance (Exley et al., 1997). In patients with active WG, antineutrophil cytoplasmic antibodies (ANCA) directed against either proteinase 3 (PR3) or myeloperoxidase (MPO) can be detected in most cases. We have previously reported that rises in the titer of cytoplasmic / classic ANCA (C-ANCA) often precede disease activity, suggesting that serial measurement of ANCA titers is useful in the followup of patients with WG (Cohen Tervaert et al., 1989). Furthermore, we have found that treatment based on rises in C-ANCA could prevent the occurrence of relapses (Cohen Tervaert et al., 1990A). 76

4 Prediction of relapses in WG The clinical utility of serial measurements of ANCA has been studied by several other groups and the value in monitoring disease activity remains controversial (Egner et al., 1990; Chan et al., 1993; Kerr et al., 1993; Davenport et al., 1995; Cohen Tervaert et al., 1996; Lesavre et al., 1996; Hoffman et al., 1995). In a pooled analysis of published data, only 48% of rises in ANCA titers as measured by indirect immunofluorescence (IIF) were followed by a relapse, and only 51% of relapses were preceded by rising titers (Cohen Tervaert et al., 1996). Different definitions of disease activity, study designs, and methods of ANCA testing may underlie the large variability in results. Moreover, the IIF technique has a disadvantage in that levels of ANCA are difficult to quantify (Kerr et al., 1993; Hagen et al., 1993). Because solidphase assays that utilize purified antigens are better suited to quantify specific antibody levels, this technique may improve our understanding of the relationship between ANCA levels and disease activity. Furthermore, in a small retrospective study, we have shown that rises in the IgG3 subclass of ANCA were a more specific predictor of relapses than were rises in C- ANCA as measured by IIF (Cohen Tervaert et al., 1994). For the present report, we performed a prospective study in 100 patients with WG. Their ANCA levels were measured at least every 2 months by IIF and antigen-specific enzymelinked immunosorbent assay (ELISA). In addition, the IgG3 subclass of ANCA was measured. Changes in the levels of ANCA were related to clinical parameters of disease activity. Patients and methods Patients Patients with WG who had been positive for PR3- or MPO-ANCA (Cohen Tervaert et al., 1990B) during an active phase of the disease were eligible for this study. The diagnosis of WG was established by the presence of classic symptoms and histologic findings, and all patients fulfilled the criteria for WG as defined by the American College of Rheumatology (Jenette et al., 1994; Leavitt et al., 1990). At inclusion, all patients were in complete remission. Study design Inclusion started in March 1996 and ended in March Most patients (n = 77) were included in the study between March and May 1996, while the remaining patients were included after May Followup was completed through March Patients were closely monitored at the vasculitis outpatient clinic of the University Hospital Groningen. Patients were seen at least every 3 months and ANCA were determined at least every 2 months. At each visit, the same physician (CAS or JWCT) performed the clinical examination. The physician was blinded with respect to the results of ANCA measurements. At each visit, standard laboratory evaluation included microscopic analysis of the urine 77

5 sediment, a collection of 24-hour urine for protein and creatinine determination, and measurement of serum creatinine, C-reactive protein concentration, and erythrocyte sedimentation rate. In the study, we focused on the relapse-free period after inclusion. We calculated the risk of a first relapse following a rise in ANCA detected between March 1996 and March Some patients underwent successive relapses during the period studied. These patients remained under medical control, but successive relapses have not been reported herein. Treatment Patients were treated at initial presentation and at the moment of relapse with prednisolone and cyclophosphamide, with or without plasma exchange, according to our standard protocols (Franssen et al., 1998). In some patients, cyclophosphamide treatment was stopped 3 6 months after induction of remission and replaced by azathioprine (Jayne et al., 1997). In addition, several patients received co-trimoxazole maintenance therapy (Stegeman et al., 1996). Disease activity scoring At each visit, disease activity was scored using the Birmingham Vasculitis Activity Score (BVAS) (Luqmani et al., 1994). Complete remission was defined as the complete absence of symptoms or signs attributable to active vasculitis (BVAS = 0), in combination with a normal serum C-reactive protein concentration. A relapse was defined as described previously (Cohen Tervaert et al., 1989). Patients had either the recurrence of biopsy-proven granulomatous inflammation of the respiratory tract, glomerulonephritis, or recurrent arthralgias in combination with other signs of vasculitis with or without rising C-reactive protein levels. Signs of vasculitis included cavitory pulmonary infiltrates, decreased renal function in combination with (microscopic) hematuria and proteinuria, necrotizing scleritis, episcleritis, vasculitis of the skin, paresis with loss of sensory function, and serous otitis media. If infection proved to be the cause of symptoms, the patient was not considered to have an exacerbation of disease. The moment of a clinical relapse was defined as the time at which immunosuppressive treatment was started or intensified. The study was carried out in accordance with the 1997 World Medical Association Declaration of Helsinki (World Medical Association, 1997). Detection of ANCA by IIF ANCA detection by IIF was performed on ethanol-fixed granulocytes as described previously (Cohen Tervaert et al., 1993). Serum samples were diluted in phosphate buffered saline (PBS) and tested at 2-fold serial dilutions starting at 1:20. Slides for ANCA testing were read independently by 2 observers who were not informed of the clinical state of the patient. 78

6 Prediction of relapses in WG Isolation of antigens Isolation of antigens was performed as described previously by Leid et al (Leid et al., 1993) with minor modifications. Briefly, enriched leukocyte populations were collected by centrifugation, and cells were washed and subsequently disrupted using nitrogen cavitation. Alpha granules were isolated by centrifugation on a discontinuous Percoll gradient (Pharmacia Fine Chemicals, Uppsala, Sweden). Antigens were extracted from the alpha granules using PBS containing 2% Triton X-100 (Sigma, St. Louis, MO) and were subsequently subjected to centrifugation. Supernatants were dialyzed and applied to a Bio-rex column (Bio-Rad, Richmond, CA). Samples from the flow-through containing PR3 were purified using a Superdex-75 column (Pharmacia). MPO was eluted from the Bio-rex column using 1 M NaCl, dialyzed to sodium acetate buffer, absorbed to a concanavalin A Sepharose gel (Pharmacia), and eluted with 1 M α-methyl-dmannoside (Sigma). Samples containing MPO were purified using a Superdex-100 column (Pharmacia). The purity of PR3 and MPO was checked by electrophoresis. Specific ELISAs were used to exclude contamination of PR3 with MPO, elastase, and lactoferrin, and contamination of MPO with PR3, elastase, and lactoferrin (Brouwer et al., 1993). Detection of ANCA by solid-phase ELISA Detection of PR3-ANCA was performed as described previously (Popa et al., 1999) with minor modifications. In short, Nunc Maxisorp 96-well microtiter plates were coated at 37 C for 1.5 hours with PR3 (1µg/ml) in coating buffer (0.1 M sodium carbonate, ph 9.6) which was inactivated by 5 m M phenylmethylsulfonyl fluoride. Blocking buffer (0.1 M Tris HCl, 0.3 M NaCl, 1% bovine serum albumin [BSA]) was subsequently added and left for 1 hour. Sera were diluted using incubation buffer (0.1 M Tris HCl, 0.3 M NaCl, 1% BSA, 0.05% Tween 20 [Sigma]) and subsequently applied at a dilution of 1:100 and 1:300. A standard curve was made using a reference serum sample that was included in each test. Values were expressed as arbitrary units/ml. Values of 6 units/ml (mean + 2 SD of 65 normal controls) were considered to be positive. The intra- and interassay variations were <10%. For detection of the IgG3 subclass of PR3-ANCA, sera were incubated at a dilution of 1:100 for 1 hour at room temperature. Monoclonal mouse anti-human IgG3 antibodies (clone HP 6050; Sigma) were added for 1 hour at a dilution of 1:5.000, followed by sheep anti-mouse IgG conjugated with alkaline phosphatase (Sigma) at a dilution of 1:500. Values of 6 units/ml (mean + 2 SD of 32 normal controls) were considered to be positive. The intraassay variation was 14.3% and the interassay variation was 22.4%. Detection of MPO-ANCA was performed as described previously (Franssen et al., 1998). Sera were diluted 1:60 and 1:180 using incubation buffer and incubated for 1 hour at room 79

7 temperature. Values of 8 units/ml (mean + 2 SD of 65 normal controls) were considered to be positive. The intra- and interassay variations were <10%. For detection of the IgG3 subclass of MPO-ANCA, serum samples were incubated as described for the IgG3 subclass of PR3-ANCA on plates coated with MPO. Definition of a rise in ANCA For the detection of rises in ANCA, the titer in each sample was compared with that in the previously obtained sample, using the same assay. Only titer increases that occurred between March 1996 and March 1998 were taken into consideration. In addition, rises in ANCA had to occur within a period of 2 months. A rise in ANCA titer as measured by IIF was defined as an increase of at least 2 titer steps (4- fold) compared with that in the previous sample. With reference to a sample that was negative for ANCA, a rise was subsequently noted to occur when the next serum sample had a titer of at least 1:40 (Cohen Tervaert et al., 1990A). A receiver operating characteristics (ROC) curve was calculated to determine, with optimal sensitivity and specificity, a rise in ANCA as measured by ELISA. In an ROC curve, sensitivity versus specificity is displayed for various cutoff values of rises in ANCA. For preparation of this curve, only rises in ANCA that amounted to at least 10 arbitrary units were taken into account. Each point on the curve represents the relationship between sensitivity and specificity for a single cutoff value. The numbers along the curve refer to the cutoff value for each point. An increase in the IgG3 subclass of ANCA as measured by ELISA was considered to be a concomitant rise when a sample was found positive at the moment of a rise in ANCA by IIF or ELISA (as determined against a negative sample obtained 2 months prior to the rise). Statistical methods The main statistical analysis was oriented to the relapse-free interval, or equivalently the relapse risk, during the study period (March 1996 to March 1999). The prognostic value of a rise in ANCA as measured by IIF and ELISA, observed from the time that a patient was included in the study until March 1998, for predicting a relapse was assessed using a Cox regression with time-dependent covariates (Matthews et al., 1996). The prognostic value of previous relapses, renal function, and tapering of immunosuppressive treatment at inclusion in relation to the occurrence of a relapse was examined using the log-rank test and Cox regression analysis (Matthews et al., 1996). Significance testing in cross-tables was done using Fisher s exact test. A 2-sided P value less than or equal to 0.05 was considered significant. 80

8 Prediction of relapses in WG Table 1 Patient characteristics at study entry Characteristics PR3 MPO (n = 85) (n = 15) Mean age (range), years 58 (24-86) 54 (22 75) Sex, no. male/no. female 45 / 40 5 / 10 Included March-May 1996 / after May 1996, no. of patients 66 / / 4 Mean no. of previous relapses (range) 1.2 (0-12) 0.2 (0-1.0) Immunosuppression at inclusion, no. of patients Yes 47 8 No 38 7 PR3 = proteinase 3; MPO = myeloperoxidase. Results Patient characteristics during followup One hundred patients with WG were included in this study. Patient characteristics are listed in Table 1. Eighty-five patients had tested positive for PR3-ANCA and 15 patients for MPO- ANCA during an active phase of their disease. At inclusion, 55 patients were receiving decreasing doses of immunosuppressive drugs (31 being treated after their initial presentation, 24 after a relapse) and 45 patients were free of immunosuppressive medication (25 after initial treatment, 20 after treatment of a relapse). Patients were followed up for a median of 1,054 days (range 178 1,095 days) until March 1999 or until death. During the followup period (between 178 and 877 days following inclusion into the study), 3 patients died in relation to active disease and 4 patients died with no apparent relation to active disease. One or more relapses occurred during followup in 37 (37%) of 100 patients. At the moment of relapse during the study period, ANCA were detectable by both IIF and ELISA in all but 1 patient. No significant differences were observed between patients with or without a relapse with respect to sex, age, ANCA specificity, or organ involvement at diagnosis. Nineteen (51%) of 37 patients with a relapse had no history of previous relapses. The remaining 18 patients (49%) had a history of 1 12 relapses (median 2.0). Twenty-two (59%) of 37 patients had relapses that involved more than 1 organ system, 9 (24%) had relapses that involved an organ system in combination with general symptoms such as fever, malaise, arthralgias, and/or myalgias, and 6 (16%) had involvement of 1 organ system (3 renal, 1 ear, nose, and throat, and 2 nervous system) alone (Table 2). Twenty-one patients (57%) had renal involvement during the relapse, whereas the remaining 16 patients (43%) had nonrenal relapses that were characterized by granulomatous inflammation of the respiratory tract or arthralgias in combination with other symptoms and signs of vasculitis (Table 2). The median disease activity score was 7 (range 3 18) and the median C-reactive protein level was 49 mg/l (range 2 288) at the moment of clinical diagnosis of relapse. 81

9 Table 2 Clinical characteristics of patients who underwent a relapse Patient/ age/sex Organ involvement during a relapse ANCA specificity Days hospitalized ESR (mm/hour) CRP (mg/liter) 1/68/F Renal,ENT,Syst,Chest PR3 50 N.T /58/M Renal,ENT,Muc,Chest PR /70/M Renal,ENT,Syst,Nerv PR /70/F Renal,ENT,Syst,Nerv PR /33/M Renal,ENT,Syst,Cut,Muc,Chest PR /44/M Renal,ENT,Syst PR /51/M Renal,Syst PR /57/F Renal PR /73/F Renal,Syst,Abd, PR /48/M Renal,Syst,Cut PR /24/F Renal,ENT PR /86/M Renal,Syst,Chest,Nerv, PR3 14 N.T /69/M Renal,Chest PR /72/M Renal,Syst PR /47/M Renal,ENT,Syst PR /58/M Renal,ENT,Syst,Muc,Chest PR /55/M Renal,ENT,Syst,Cut,Chest PR /67/M Renal,ENT,Syst,Chest,Nerv PR /38/M Renal MPO /24/F Renal MPO /75/F Renal,Syst MPO /50/M ENT,Syst,Muc PR /53/M ENT,Syst PR /70/F ENT PR /38/F ENT,Syst,Cut PR /68/M ENT,Syst,Nerv PR /58/M ENT,Syst PR /62/M ENT,Chest,Nerv PR /49/F ENT,Syst,Muc,Chest,Abd PR /80/F ENT,Syst,Muc PR /53/F ENT,Syst,Muc PR /30/F ENT,Syst PR /39/F ENT,Syst MPO /49/F Syst,Muc PR /76/F Nerv PR /76/M Nerv PR /57/F Syst,Nerv, PR BVAS ANCA = antineutrophil cytoplasmic antibodies; ESR = erythrocyte sedimentation rate at relapse; CRP = C- reactive protein at relapse; BVAS = Birmingham Vasculitis Activity Score at relapse; Renal = positive findings on biopsy and/or decreased creatinine clearance in combination with proteinuria and erythrocyturia; ENT = ear, nose, and throat; Syst = systemic (e.g., fever, arthralgia, myalgia, malaise); PR3 = proteinase 3; ND = not done; Muc = genital/mouth ulcers or red eyes/proptosis; Nerv = nervous system; Cut = cutaneous; Abd = abdominal; MPO = myeloperoxidase. Died during relapse. 82

10 Prediction of relapses in WG Disease-free (%) Time (months after titer rise) ELISA IIF Figure 1 Percentage of patients with Wegener s granulomatosis who did not experience disease relapses in the indicated time period after a rise in antineutrophil cytoplasmic antibodies as measured by either indirect immunofluorescence (IIF; n = 30) or antigen-specific enzyme-linked immunosorbent assay (ELISA; n = 38). The numbers above the horizontal axis indicate the number of patients who were still at risk for a relapse at 6, 12, 18, 24, and 30 months after the rise in antibody levels as detected by ELISA (upper numbers) or IIF (lower numbers). Table 3 Relapse risk percentages after inclusion and relative risk values for prognostic factors with respect to relapse in all patients at study entry Relapse Log-rank Relative risk n risk after inclusion, % P 95% CI Previous Relapse No Yes ( ) CrCl >60 ml/min No Yes ( ) Immunosuppression No 45 27% 1.00 Yes 55 45% ( ) CrCl = creatinine clearance; 95% CI = 95% confidence interval. Levels of canca and PR3-ANCA during followup A total of 1,191 serum samples were prospectively analyzed for the presence of ANCA between March 1996 and March For each patient, a median of 14 samples (range 3 35) were analyzed. Thirty (35%) of 85 PR3-ANCA positive patients demonstrated a rise in C- ANCA as measured by IIF, occurring within 2 months of the previous measurement. Seventeen of these 30 patients had a relapse following this increase (positive predictive value 57%) (Table 4). A relapse occurred in 11 patients within 6 months, in 4 patients within 12 months, and in 2 patients during a longer followup period (ranging up to 616 days after the 83

11 rise in C-ANCA) (Figure 1). In 55 patients, no rise in C-ANCA was detected by IIF. In 16 of these patients (29%), a relapse occurred during followup Figure 2 Receiver operating characteristics curve indicating direct proteinase 3 (PR3) enzyme-linked Sensitivity (%) immunosorbent assay detection of increases in anti- PR3 antibodies. Each point represents the relationship between sensitivity and specificity of a single cutoff value (shown as a number beside each point) for the prediction of relapses Specificity (%) 25 0 Based on an ROC curve (Figure 2), an increase in titer of 75% or more was chosen to indicate a rise in PR3-ANCA by ELISA. On the basis of this cutoff value, 38 (45%) of 85 patients showed a rise in PR3-ANCA by ELISA, and 27 of these 38 patients experienced a relapse following this increase (positive predictive value 71%) (Table 4). A relapse occurred in 15 patients within 6 months, in 7 patients within 12 months, and in 5 patients during longer followup (ranging up to 616 days after the rise in PR3-ANCA) (Figure 1). In 47 patients, no rise in PR3-ANCA was detected by ELISA. In 6 (13%) of these 47 patients, a relapse occurred during followup (Table 4). Table 4 Rises in ANCA in relation to the occurrence of relapses in PR3-positive patients a Method ANCA increase Relapse (n = 33) Relapse-free (n = 52) IIF No Yes ELISA No 6 41 Yes a Values are the number of patients. A rise by indirect immunofluorescence (IIF) was defined as an increase of at least 2 titer steps (4-fold) compared with that in the previous sample, or a titer of at least 1:40 in a sample previously negative for proteinase 3 (PR3) antineutrophil cytoplasmic antibodies (ANCA) (Cohen Tervaert et al., 1990A) (P = 0.02 by Fisher s exact test for IIF contingency table). A rise by enzyme-linked immunosorbent assay (ELISA) was defined as an increase in titer of at least 75% compared with that in the previous sample and amounting to at least 10 arbitrary units (P < by Fisher s exact test for ELISA contingency table). 84

12 Prediction of relapses in WG Combining the results of IIF and ELISA, a rise in ANCA by either IIF or ELISA occurred in 49 (58%) of 85 PR3-ANCA positive patients (Table 5). Only 2 (6%) of 33 patients who had a relapse showed no prior rise in ANCA by either technique. These 2 patients had persistently high ANCA levels. A rise in ANCA as detected by both IIF and ELISA occurred in 19 (22%) of 85 PR3-ANCA positive patients (Table 5). In 13 of these patients (68%), the rise in ANCA was followed by a relapse, whereas in 6 patients, a rise in ANCA as detected by both techniques was recorded without an ensuing relapse. When the Cox regression model was calculated with rises in ANCA as the time-dependent covariates, a rise in ANCA as measured by ELISA was a stronger predictor of an ensuing relapse within 1 year of observation than a rise in ANCA as determined by IIF (ELISA rise RR 18.6, 95% CI ; IIF rise RR 4.59, 95% CI ) (Table 5). A rise in either C-ANCA or PR3-ANCA was the strongest predictor of an ensuing relapse within 1 year (RR 42.5, 95% CI ). Table 5 Analysis of the prognostic value for relapse of a rise in ANCA following inclusion in 85 patients with PR3-ANCA associated Wegener s granulomatosis ANCA status n Relative risk 95% CI Absence of rise by IIF Rise by IIF Absence of rise by ELISA Rise by ELISA No rise both by IIF and ELISA Rise by IIF and ELISA Absence of rise by IIF and ELISA Rise by IIF or ELISA % CI = 95% confidence interval (See Table 4 for other definitions). Specificity of a rise in ANCA A rise in C-ANCA as detected by IIF was not followed by a relapse in 13 (43%) of 30 patients, and in 11 (29%) of 38 patients, no relapse occurred after a rise in PR3-ANCA as detected by ELISA. The median length of followup after the detection of a rise was 480 days for C-ANCA (range 418 1,046 days) and 859 days for PR3-ANCA (range 418 1,082 days). The rise in ANCA did not occur during switching from cyclophosphamide to azathioprine in any of these patients. We studied the course of ANCA levels by IIF and ELISA following a rise in ANCA in patients without a relapse and compared these with the levels in patients with a relapse after a rise in ANCA. No differences in the levels of ANCA or in the rate of increase between these 2 groups of patients were seen (data not shown). 85

13 IgG3 subclass of PR3-ANCA A relapse occurred in all 5 patients (positive predictive value 100%) who had a concomitant rise in C-ANCA (by IIF) and IgG3-ANCA, whereas relapses occurred in 12 (48%) of 25 patients with a rise in C-ANCA but no rise in IgG3-ANCA (P = 0.05). A rise in both PR3- ANCA (by ELISA) and IgG3-ANCA was followed by a relapse in 10 of 11 patients (positive predictive value 91%), whereas a relapse was diagnosed in 16 (62%) of 26 patients with a rise in PR3-ANCA but no rise in IgG3-ANCA (P = 0.12). Levels of perinuclear ANCA (panca) and MPO-ANCA during followup Fifteen patients had tested positive for MPO-ANCA during an active phase of their disease and were followed up at our vasculitis clinic. No rise in P-ANCA was detected with the use of IIF between March 1996 and March In accordance with the ROC for the PR3 ELISA, a rise in the titer of MPO-ANCA of 75% or more was also arbitrarily chosen as the cutoff value for the MPO ELISA. Three patients had a rise in MPO-ANCA (by ELISA) and all 3 developed a relapse after a median time of 215 days (range days). In addition, we observed 1 relapse that occurred without an increase in MPO-ANCA. This patient tested positive for MPO-ANCA at diagnosis, but the ANCA levels declined during followup and no ANCA were detectable at the time of relapse. A rise in MPO-ANCA was not accompanied by a rise in the IgG3 subclass of MPO-ANCA in any of these patients. Discussion In our prospective study of 85 patients with PR3-ANCA associated WG, we found that in 31 (94%) of 33 patients who experienced a relapse, the relapse was preceded by a rise in ANCA as measured by either IIF or ELISA. The 2 remaining patients with a relapse but no rise in ANCA had persistently elevated ANCA levels. These data suggest that quantification of ANCA is a sensitive tool for predicting disease activity in patients with WG. By using the IIF technique for quantification of ANCA, we found that in 52% of the PR3- ANCA positive patients who experienced a relapse, the relapse was preceded by a rise in C- ANCA. This value is lower than that found in our previous study (4), in which all relapses were preceded by a rise in C-ANCA (as detected by IIF) when serum samples were obtained at least every 4 weeks, as opposed to every 8 weeks in this study. The results of the present study are, however, in agreement with those of other studies in which sampling was less frequently performed (6 10). Using the ELISA technique to measure ANCA levels, we found that 82% of the PR3-ANCA positive patients had a relapse that was preceded by a rise in PR3-ANCA. We therefore conclude from this study that ELISA is more sensitive than IIF in predicting disease activity. 86

14 Prediction of relapses in WG In 15 patients with WG who tested positive for MPO-ANCA, we found that 3 (75%) of 4 relapses were preceded by a rise in ANCA as measured by ELISA. No rises in MPO-ANCA occurred in patients without a relapse. In 1 patient with a relapse but without a rise in ANCA, ANCA could not be detected. No rises in P-ANCA were detected in our patients. Although these data are promising with regard to the predictive value of a rise in MPO-ANCA, the small number of patients and relapses make any conclusion inappropriate. Combining the results from patients with MPO-ANCA and PR3-ANCA associated WG, relapses occurred in 37 (37%) of the 100 patients studied. Thirty-four (92%) of these 37 patients showed a rise in ANCA preceding their relapse as detected by ELISA or IIF. Serial determination of ANCA by ELISA with purified antigens has been reported in 2 other small studies (Jayne et al., 1995; Kyndt et al., 1999). In both studies, patients with different forms of ANCA-associated vasculitis were studied. Jayne et al found rises in ANCA prior to or at the moment of a relapse in 74% of patients. ANCA levels were detected by the ELISA technique only, and patient followup was restricted to the first year after diagnosis (Jayne et al., 1995). In the second study, Kyndt et al found that 73% of MPO-ANCA positive patients had a rise in MPO-ANCA (by ELISA) prior to or at the moment of relapse, whereas only 33% of PR3-ANCA positive patients had a rise in ANCA as detected by either IIF or ELISA (Kyndt et al., 1999). Comparing the results of the latter study with those of our study is, however, difficult since no precise definitions for rises in ANCA as measured by ELISA were given and serum samples were not tested simultaneously with the previous sample in the same test. Finally, the patient samples in that study were tested in intervals as long as 3 months, which might explain why many rises in ANCA were found at the time of the relapse instead of prior to the relapse. On the basis of the present and previous studies, it is clear that there are patients who do not show a rise in ANCA prior to their relapse. Detection of the levels of antibodies to the proenzymes of either PR3 or MPO might be valuable in these patients (Sun et al., 1998). Predictive values for rises in ANCA varied in previous studies from 23% to 77% (Cohen Tervaert et al., 1996). In the present study, we found that 17 (57%) of 30 PR3-ANCA positive patients with a rise in C-ANCA as measured by IIF experienced a relapse. The predictive value of rises in ANCA as measured by ELISA was higher, since 30 (73%) of 41 PR3-ANCA and MPO-ANCA positive patients with a rise in PR3- or MPO-ANCA had a relapse. Finally, in the present study, we partially confirmed previous data (Cohen Tervaert et al., 1994) on the increase in predictive value by the additional measurement of the IgG3 subclass of PR3-ANCA. The positive predictive value for a relapse of a rise in C-ANCA by IIF increased from 57% to 100%, and that of a rise in PR3-ANCA by ELISA from 71% to 91% when a rise in ANCA was accompanied by a rise in the IgG3 subclass of PR3-ANCA. However, the sensitivity for the prediction of a relapse decreased enormously, making the additional measurement of the IgG3 subclass of PR3-ANCA of little clinical utility. 87

15 Since not all increases in ANCA were followed by a relapse, we wondered whether rises in ANCA not followed by a relapse occurred after switching from cyclophosphamide therapy to azathioprine therapy, as has been reported by Jayne et al (Jayne et al., 1995). In our study, we did not observe this phenomenon. Patients with a rise in ANCA not followed by a relapse, however, may have had subtle signs of disease activity, since C-reactive protein levels and erythrocyte sedimentation rates were slightly higher in these patients during the first 3 months after the rise in ANCA than in patients with WG without a rise in ANCA (data not shown). We postulate that these patients may have had low-grade disease activity. Consistent with this observation, we reported previously that patients with necrotizing crescentic glomerulonephritis associated with persistently elevated levels of MPO-ANCA experienced a poor renal outcome without overt signs of a relapse of disease (Franssen et al., 1998). Since rises in ANCA occur in most patients prior to a relapse, one may question whether treatment based on ANCA levels might be justified. However, starting escalating immunosuppressive treatment in these patients would expose a substantial number of the patients to the unnecessary risk of drug-related morbidity, since in 29% of the patients, such a rise was not followed by a relapse during an extended period of followup. Another possible drawback of preventive treatment is the finding that there is great variability in time lag between the detection of a rise in ANCA and a relapse. Although during extended followup, a relapse occurred in 71% of patients after a rise in PR3-ANCA, only 39% of the patients with such a rise had a relapse of their disease within 6 months. In other words, a rise in PR3- ANCA is not usually an indicator of an imminent relapse. Furthermore, it is questionable whether the rise in ANCA is of pathophysiologic relevance. As indicated in studies involving experimental models of vasculitis (Heeringa et al., 1998), it is clear that ANCA in itself are not sufficient for disease induction. In a rat model, we demonstrated that immunization with MPO results in the induction of MPO-ANCA without the development of lesions and that, subsequently, immune-complex deposition is necessary before vasculitis and glomerulonephritis are induced (Heeringa et al., 1998). It has been suggested that during infections, these immune complexes may be formed after release of PR3 and MPO from activated neutrophils and/or monocytes, and subsequently bound to endothelial cells (Muller-Kobold et al., 1999). In conclusion, this prospective study on a large cohort of patients with WG demonstrates that rises in ANCA are found in the majority of patients prior to a relapse. Furthermore, we showed that measuring ANCA by ELISA is superior to measuring ANCA by IIF for the prediction of an ensuing relapse. The additional measurement of rises in the IgG3 subclass of PR3-ANCA increases the positive predictive value of a rise in ANCA but, given its low sensitivity, is of little clinical utility. 88

16 Prediction of relapses in WG Literature Brouwer E, Huitema MG, Klok PA, de Weerd H, Cohen Tervaert JW, Weening JJ, et al. (1993) Antimyeloperoxidase-associated prolifera-tive glomerulonephritis: an animal model. J Exp Med 177: Chan TM, Frampton G, Jayne DR, Perry GJ, Lockwood CM, Cameron JS (1993) Clinical significance of antiendothelial cell antibodies in systemic vasculitis: a longitudinal study comparing anti-endothelial cell antibodies and anti-neutrophil cytoplasm antibodies. Am J Kidney Dis 22: Cohen Tervaert JW, van der Woude FJ, Fauci AS, Ambrus JL, Velosa J, Keane WF, et al. (1989) Association between active Wegener s granulomatosis and anticytoplasmic antibodies. Arch Intern Med 149: Cohen Tervaert JW, Huitema MG, Hené RJ, Sluiter WJ, The TH, van der Hem GK, et al. (1990A) Prevention of relapses in Wegener s granulomatosis by treatment based on antineutrophil cytoplasmic antibody titre. Lancet 336: Cohen Tervaert JW, Goldschmeding R, Elema JD, van der Giessen M, Huitema MG, van der Hem GK, et al. (1990B) Autoantibodies against myeloid lysosomal enzymes in crescentic glomerulonephritis. Kidney Int 37: Cohen Tervaert JW, Mulder L, Stegeman CA, Elema J, Huitema M, The TH, et al. (1993) Occurrence of autoantibodies to human leucocyte elastase in Wegener s granulomatosis and other inflammatory disorders. Ann Rheum Dis 52: Cohen Tervaert JW, Mulder AHL, Kallenberg CGM, Stegeman CA (1994) Measurement of IgG3 of antiproteinase 3 antibodies to neutrophil cytoplasm is useful in monitoring disease activity in Wegener s granulomatosis [abstract]. J Am Soc Nephrol 5: 828. Cohen Tervaert JW, Stegeman CA, Kallenberg CG (1996) Serial ANCA testing is useful in monitoring disease activity in patients with ANCA associated vasculitides. Sarcoidosis Vasc Diffuse Lung Dis 13: Davenport A, Lock RJ, Wallington T (1995) Clinical significance of the serial measurement of autoantibodies to neutrophil cytoplasm using a standard indirect immunofluorescence test. Am J Nephrol 15: Egner W, Chapel HM (1990) Titration of antibodies against neutrophil cytoplasmic antigens is useful in monitoring disease activity in systemic vasculitides. Clin Exp Immunol; 82: Exley AR, Carruthers DM, Luqmani RA, Kitas GD, Gordon C, Janssen BA, et al. (1997) Damage occurs early in systemic vasculitis and is an index of outcome. QJM 90: Franssen CF, Stegeman CA, Oost-Kort WW, Kallenberg CG, Limburg PC, Tiebosch A, et al. (1998) Determinants of renal outcome in anti-myeloperoxidase-associated necrotizing crescentic glomerulonephritis. J Am Soc Nephrol 9: Godman GC, Churg J (1954) Wegener s granulomatosis: pathology and review of the literature. Am Med Arch Pathol 58: Hagen EC, Andrassy K, Chernok E, Daha MR, Gaskin G, Gross W, et al. (1993) The value of indirect immunofluorescence and solid phase techniques for ANCA detection: a report on the first phase of an international cooperative study on the standardization of ANCA assays. EEC/BCR Group for ANCA Assay Standardization. J Immunol Methods 159: Heeringa P, Brouwer E, Cohen Tervaert JW, Weening JJ, Kallenberg CG (1998) Animal models of antineutrophil cytoplasmic antibody associated vasculitis. Kidney Int 53: Hoffman GS, Kerr GS, Leavitt RY, Hallahan CW, Lebovics RS, Travis WD, et al. (1992) Wegener granulomatosis: an analysis of 158 patients. Ann Intern Med 116: Hoffman GS, Kerr GS (1995) Rise in ANCA titer: to treat or not to treat. Am J Med 98: Jayne DR, Gaskin G, Pusey CD, Lockwood CM (1995) ANCA and predicting relapse in systemic vasculitis. QJM 88: Jayne DR, Rasmussen N (1997) Treatment of antineutrophil cytoplasm autoantibody-associated systemic vasculitis: initiatives of the European Community Systemic Vasculitis Clinical Trials Study Group. Mayo Clin Proc 72: Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross WL, et al. (1994) Nomenclature of systemic vasculitides: proposal of an international consensus conference. Arthritis Rheum 37: Kerr GS, Fleisher TA, Hallahan CW, Leavitt RY, Fauci AS, Hoffman GS (1993) Limited prognostic value of changes in antineutrophil cytoplasmic antibody titer in patients with Wegener s granulomatosis. Arthritis Rheum 36: Kyndt X, Reumaux D, Bridoux F, Tribout B, Bataille P, Hachulla E, et al. (1999) Serial measurements of antineutrophil cytoplasmic autoantibodies in patients with systemic vasculitis. Am J Med 106: Leavitt RY, Fauci AS, Bloch DA, Michel BA, Hunder GG, Arend WP, et al. (1990) The American College of Rheumatology 1990 criteria for the classification of Wegener s granulomatosis. Arthritis Rheum 33: Leid WR, Ballieux BE, van der Heijden I, Kleyburg-van der Keur C, Hagen EC, van Es LA, et al. (1993) Cleavage and inactivation of human C1 inhibitor by the human leukocyte proteinase, proteinase 3. Eur J Immunol 23:

17 Lesavre P, Kyndt X, Vanhille P, Reumaux D, Guillevin L, Noel LH (1996) Serial ANCA testing has limited value during the follow-up of disease activity in patients with ANCA-associated vasculitis. Sarcoidosis Vasc Diffuse Lung Dis 13: Luqmani RA, Bacon PA, Moots RJ, Janssen BA, Pall A, Emery P, et al. (1994) Birmingham Vasculitis Activity Score (BVAS) in systemic necrotizing vasculitis. QJM 87: Matthews DE, Farewell VT (1996) Using and understanding medical statistics. 3rd ed. Basel: Karger. Muller Kobold AC, van der Geld YM, Limburg PC, Cohen Tervaert JW, Kallenberg CG (1999) Pathophysiology of ANCA-associated glomerulonephritis. Nephrol Dial Transplant 14: Popa ER, Stegeman CA, Bos NA, Kallenberg CG, Cohen Tervaert JW (1999) Differential B- and T-cell activation in Wegener s granulomatosis. J Allergy Clin Immunol 103: Stegeman CA, Cohen Tervaert JW, de Jong PE, Kallenberg CG (1996) Trimethoprim-sulfamethoxazole (cotrimoxazole) for the prevention of relapses of Wegener s granulomatosis. N Engl J Med 335: Sun J, Fass DN, Viss MA, Hummel AM, Tang H, Homburger HA, et al. (1998) A proportion of proteinase 3 (PR3)-specific anti-neutrophil cytoplasmic antibodies (ANCA) only react with PR3 after cleavage of its N- terminal activation dipeptide. Clin Exp Immunol 114: World Medical Association (1997) World Medical Association Declaration of Helsinki: recommendations guiding physicians in biomedical research involving human subjects. JAMA 277: