Article Text
Abstract
Objective: To evaluate the relevance of monitoring antimyeloperoxidase antibody levels in the management of antimyeloperoxidase-associated vasculitides.
Methods: Thirty-eight patients with antimyeloperoxidase-associated vasculitides were included: microscopic polyangiitis (n = 18), Wegener’s granulomatosis (n = 15) and Churg–Strauss syndrome (n = 5). Baseline characteristics and outcomes were recorded. Serial measurements of antimyeloperoxidase antibody levels were performed (ELISA, positive ⩾20 IU/ml).
Results: All patients achieved vasculitis remission after a mean time of 2.0 months (SD 0.9), with a significant decrease in the mean antimyeloperoxidase antibody level at remission (478 vs 41 IU/ml (SD 598 vs 100); p<0.001). Twenty-eight (74%) patients became antimyeloperoxidase antibody negative. After a mean follow-up of 54 months (SD 38), 12 cases of clinical relapse occurred in 11/38 (29%) patients. Relapses were associated with an increase in antimyeloperoxidase antibody levels in 10/11 (91%) patients (34 vs 199 IU/ml (88 vs 314); p = 0.002). The reappearance of antimyeloperoxidase antibodies after achieving negative levels was significantly associated with relapse (odds ratio 117; 95% CI 9.4 to 1450; p<0.001). Antimyeloperoxidase antibodies showed a positive predictive value of 90% and a negative predictive value of 94% for relapse of vasculitis. Up to 60% of cases of relapse occurred less than 12 months after the reappearance of antimyeloperoxidase antibodies. Relapse-free survival was significantly worse for patients who exhibited a reappearance of antimyeloperoxidase antibodies than in those with persistent negative antimyeloperoxidase antibodies (p<0.001). The antimyeloperoxidase antibodies serum level was strongly correlated with the Birmingham vasculitis activity score and the disease extent index (r = +0.49; p = 0.002).
Conclusion: Through monitoring, antimyeloperoxidase antibodies are a useful marker of disease activity and a good predictor of relapse in antimyeloperoxidase-associated vasculitides.
Statistics from Altmetric.com
The primary systemic vasculitides are a group of rare disorders defined by inflammation of the blood vessel walls, which are classified according to their clinical and histological features and the size of the predominantly affected vessels. Some vasculitides are associated with the presence of antineutrophil cytoplasmic antibodies (ANCA), which are a useful tool for the diagnosis of disease, including Wegener’s granulomatosis (WG), microscopic polyangiitis (MPA) and Churg–Strauss syndrome (CSS). Antiproteinase 3 antibodies are present in more than 90% of patients with WG. Antimyeloperoxidase antibodies are found in 80% of patients with MPA and 40–60% of those with CSS.1 2 3 4 5
More than the diagnostic relevance of ANCA during systemic vasculitides, there is evidence for a pathogenic role of these antibodies:1 6 (1) neonatal vasculitis caused by transplacental transfer of ANCA;7 8 (2) activation of neutrophils and monocytes in vitro to release mediators of acute inflammation;9 10 11 12 13 14 15 16 17 18 19 20 (3) induction of glomerulonephritis and vasculitis by passive administration of antimyeloperoxidase IgG to mice;21 22 23 (4) induction of glomerulonephritis and vasculitis by the transfer of antimyeloperoxidase lymphocytes to mice;23 (5) induction of glomerulonephritis and vasculitis by active immunisation of rats with human myeloperoxidase;24 (6) enhancement of dermal inflammation in mice by the injection of antiproteinase 3;25 (7) stimulation of neutrophils by ANCA and activation of complement via the alternative pathway, initiating severe necrotising inflammation26 and (8) induction of oxidative burst by antimyeloperoxidase antibodies, leading to severe endothelial damage.27
Besides their diagnostic and pathogenic relevance, ANCA may be useful as a surrogate marker of disease activity, as suggested by the fact that disease remission is generally paralleled by the decrease or absence of ANCA, and that the persistent presence or reappearance of ANCA in patients in remission is associated with disease relapse.3 28 29 30 31 32 33 34 In antiproteinase 3-associated vasculitides, predictive values of more than 50% for a relapse within 6 or 12 months for a greater than 75% increase in ANCA level (as determined by ELISA) have been found,35 36 37 and positive antiproteinase 3 antibodies during early follow-up identified patients at increased risk of relapse.38 On the contrary, the risk of relapse in patients who have a persistence of ANCA seems to be very low.39 However, studies are controversial in the setting of antiproteinase 3-associated vasculitides,40 41 42 43 whereas in antimyeloperoxidase-associated vasculitides, data are limited.32 35 44
The aim of this study was to analyse the relevance of serum antimyeloperoxidase antibody level assessment in the management of ANCA-associated vasculitides. We sought to determine whether antimyeloperoxidase antibody levels correlated with disease activity, and whether a decrease in antimyeloperoxidase antibody levels during remission is associated with longer remission and increases are followed by relapse.
Patients and methods
Patients
We included all patients who had received ANCA screening with serial measurements of serum antimyeloperoxidase antibody levels performed in the immunology laboratory of a university hospital between January 1996 and December 2006. Nine patients were excluded because of the absence of follow-up or incomplete data (n = 9). The study population consisted of 38 patients with antimyeloperoxidase-associated vasculitides, including 18 patients with MPA, 15 with WG and five with CSS. Patients’ files were retrospectively reviewed and data collected using a standardised form. At inclusion, all patients were in the active phase of the disease. None of them had received previous treatment for vasculitides. Clinical and biological data (including antimyeloperoxidase antibody levels), treatments and outcomes, were recorded for each patient at the time of the initial evaluation and at each hospitalisation and consultation during follow-up. The diagnoses of MPA, WG and CSS were established by the presence of the classic features and histological findings, and satisfied the criteria defined by the American College of Rheumatology and Chapel Hill Consensus Conference for MPA,45 46 and the American College of Rheumatology for WG46 47 and CSS.48 We also applied the EMEA algorithm for individual vasculitis diagnoses.49
Disease activity scoring
Disease activity was scored using the Birmingham vasculitis activity score (BVAS)50 and the disease extent index (DEI).51 The following inflammatory biological variables were recorded: C-reactive protein levels, erythrocyte sedimentation rate and fibrinogen levels.
Disease state definitions reported from the EULAR/EUVAS consensus group have been used.52 Remission was defined as the absence of clinical signs and symptoms attributable to active vasculitis qualified by the need for ongoing stable maintenance immunosuppressive therapy. Relapse was defined as the recurrence or new onset of clinical signs and symptoms attributable to active vasculitis.
Detection of ANCA by indirect immunofluorescence and measurement of serum antimyeloperoxidase levels by ELISA
ANCA detection by indirect immunofluorescence was performed on ethanol-fixed human neutrophils (INOVA; Menarini, Antony, France). Serum samples were diluted 1 : 20. All patients showed a perinuclear staining pattern. Measurements of serum antimyeloperoxidase antibody levels were performed by ELISA using an antimyeloperoxidase commercial kit (Euroimmun; BioAdvance, Emérainville, France) at diagnosis and during follow-up. Values above 20 IU/ml were considered positive.
Statistical analysis
Data are presented as means (SD) for continuous variables and percentages for qualitative variables. The Fisher’s exact test was used to compare qualitative variables, and the non-parametric Mann–Whitney test was used to compare continuous variables. The Wilcoxon test was used to compare paired quantitative variables. Correlation was evaluated using Spearman’s test. A p value of 0.05 or less was considered significant. Statistical analyses were performed using StatView.
Results
Patient characteristics
The main clinical and biological characteristics of the 38 patients with antimyeloperoxidase-associated vasculitis (14 men and 24 women, mean age 58 years (SD 16, range 16–81)) are given in table 1.
The mean serum antimyeloperoxidase antibody level was higher in patients with kidney involvement than in their negative counterparts (711 vs 245 IU/ml (SD 666 vs 421); p<0.001). The mean antimyeloperoxidase antibodies level was lower in patients with ear, nose and/or throat involvement than in those without (292 vs 734 IU/ml (SD 413 vs 723); p = 0.04). No difference in the antimyeloperoxidase antibody level was found between patients with and without involvement of the lungs, joints, muscles, peripheral nervous system, eyes, skin, heart and central nervous system (data not shown).
Treatment consisted of corticosteroids (prednisone 1 mg/kg per day) in all patients (100%), with pulse methylprednisolone in 20, and associated with pulse cyclophosphamide (0.7 g/m2 per 4 weeks) in 26 patients (68%). Patients treated with corticosteroids alone (n = 12) did not differ from those treated with corticosteroids and cyclophosphamide (n = 26) in terms of age, gender, vasculitis diagnosis, heart, gastrointestinal and central and peripheral nervous system involvement (data not shown), but showed less frequent renal (17% vs 65%, p = 0.01) and lung (25% vs 69%, p = 0.02) involvement, lower BVAS (12.3 vs 20.2 (SE 6.7 vs 5.8); p = 0.002), lower DEI (4.6 vs 7.8 (SD 2.2 vs 2.5); p = 0.002) and antimyeloperoxidase antibody levels (296 vs 562 (SD 519 vs 623); p = 0.03).
The time course of antimyeloperoxidase antibody levels and the clinical outcomes of the patients are summarised in fig 1.
Antimyeloperoxidase antibody levels and sustained remission
All patients (100%) achieved remission after a mean time of 2.0 months (SD 0.9) following induction therapy. All patients showed a decrease in their serum antimyeloperoxidase levels at remission (478 vs 41 IU/ml (SD 598 vs 100); p<0.001; fig 2A). Antimyeloperoxidase remained positive, ie, 20 IU/ml or greater, in 10 (26%) patients and became negative, ie, less than 20 IU/ml in 28 (74%) after 5.7 months (SD 4.6) of treatment. The level of antimyeloperoxidase antibodies at the diagnosis of vasculitis was not correlated with the time course of antimyeloperoxidase antibody levels during treatment (absence of antimyeloperoxidase antibodies vs persistent positive antimyeloperoxidase antibodies; 385 vs 706 IU/ml (SD 526 vs 724); respectively; p = 0.24).
The median duration of remission was 40 months (range 5–104). The duration of remission was not correlated with antimyeloperoxidase antibody levels at diagnosis (>150 vs ⩽150 IU/ml; 43 vs 51 months (SD 29 vs 32), respectively; p = 0.47). Moreover, no correlation was found between the magnitude of the decrease in antimyeloperoxidase antibody levels and the duration of remission (data not shown).
Antimyeloperoxidase antibody levels and relapse
After a mean follow-up of 54 months (SD 38), 12 clinical relapses occurred in 11/38 (29%) patients. Organ involvement at relapse included the kidney (n = 5), lung (n = 5), joints (n = 5), ear, nose and/or throat (n = 5), eyes (n = 6), nerves (n = 2), digestive tract (n = 1) and skin (n = 1). Treatment at relapse consisted of corticosteroids in 11 patients, with pulse methylprednisolone in five, and associated with pulse cyclophosphamide (n = 3), mycophenolate mofetil (n = 2), methotrexate (n = 2), azathioprine (n = 2), intravenous immunoglobulins (n = 1), rituximab (n = 1) and plasmapheresis (n = 1). Patients initially treated with corticosteroids alone did not show more frequent relapses than those treated with corticosteroids and cyclophosphamide (data not shown).
Relapses were associated with an increase in the level of antimyeloperoxidase antibodies compared to that found at the time of remission in 10 out of the 11 cases (91%; 199 vs 34 IU/ml (SD 314 vs 88); p = 0.002; fig 2B). No correlation was found between the magnitude of the increase in antimyeloperoxidase antibody levels and the time of relapse (data not shown).
No differences were observed between relapsers and non-relapsers in terms of gender, age, initial organ involvement and activity of the disease, biological parameters, initial serum antimyeloperoxidase levels and initial treatment. The only significant difference between them was the lower number of MPA (p = 0.03) and higher number of WG diseases (p = 0.01) in the relapser group (table 1).
We next examined whether the time course of antimyeloperoxidase antibody levels was associated with relapses. The reappearance of antimyeloperoxidase antibodies was significantly associated with relapse (odds ratio (OR) 117; 95% CI 9.4 to 1450; p<0.001). The persistence of negative antimyeloperoxidase antibodies was negatively associated with the occurrence of clinical relapse (OR 0.06; 95% CI 0.007 to 0.53; p<0.001). The persistence of positive antimyeloperoxidase antibodies was not associated with the occurrence of clinical relapse (OR 0.2; 95% CI 0.02 to 1.8: p = 0.23).
The time course of antimyeloperoxidase antibody levels during follow-up in the 11 patients who relapsed is indicated in fig 3.
Among the 28 patients who became antimyeloperoxidase antibody negative during induction therapy, the reappearance of antimyeloperoxidase antibodies (n = 10) was associated with clinical relapse in nine patients (positive predictive value 9/10, 90%). The median time between the reappearance of antimyeloperoxidase antibodies and clinical relapse of the vasculitis was 8 months (range 0–89). The relapse occurred in six of nine (67%) patients within 12 months, and during a longer follow-up period for the last three patients (31, 65 and 89 months). Only one patient with the reappearance of antimyeloperoxidase antibodies did not relapse, after a follow-up with positive antimyeloperoxidase antibody levels of 25 months. In patients who remained antimyeloperoxidase antibody negative during treatment (n = 18), only one patient presented with early relapse during treatment with corticosteroids (40 mg/day) and pulse cyclophosphamide (0.7 g/m2 per 4 weeks) 5 months after the diagnosis of vasculitis (negative predictive value 17/18, 94%).
Figure 4 shows the relapse-free survival curve in patients with persistent negative antimyeloperoxidase antibodies and in those who exhibited a reappearance of antimyeloperoxidase antibodies. The relapse-free survival was significantly worse in patients with reappearance of antimyeloperoxidase antibodies than in those with persistent negative antimyeloperoxidase antibodies (p<0.001).
Among the 10 patients who remained antimyeloperoxidase antibody positive under treatment, a relapse occurred in one patient who had an increase in the antimyeloperoxidase antibody level (ie, a >50% increase in the antimyeloperoxidase antibody level) compared with the level at remission (1080 vs 300 IU/ml) and with the patients who did not relapse (fig 5).
Treatments in non-relapser and relapser patients
Treatments at the end of follow-up for non-relapser patients and at the time of clinical relapse for relapser patients were assessed.
At the end of follow-up, non-relapsers were treated with corticosteroids in 21/27 cases (78%) at a mean dose of 9.7 mg/day (SD 4.7), and with immunosuppressant agents (IS) in 12/27 (44%). In non-relapsers without IS at the end of follow-up (n = 15), eight patients had never received IS, and in the seven remaining patients IS had been discontinuated from 31 months (SD 23).
Relapser patients were treated at the time of clinical relapse with corticosteroids in seven of 12 (58%) cases at a mean dose of 10.7 mg/day (SD 13), and with IS in two of 12 (17%). In relapsers without IS at the time of the clinical relapse (n = 10), three patients had never received IS, and in the seven remaining patients IS had been discontinuated from 45 months (SD 19).
We compared treatments in non-relapser and relapser patients. No significant difference was found between the two groups, but relapser patients were receiving corticosteroids and IS less frequently than non-relapser patients (58% vs 78%; p = 0.26; and 17% vs 44%; p = 0.15, respectively).
Antimyeloperoxidase antibody levels and disease activity
Antimyeloperoxidase antibody measurements and evaluation of disease activity were performed at the same time. At the time of clinical relapse of vasculitis for relapser patients and at the time of remission for non-relapser patients, the antimyeloperoxidase antibody serum level was strongly correlated with the BVAS (r = +0.49; p = 0.002) and the DEI (r = +0.49, p = 0.002).
Discussion
Predicting the clinical course and response to therapy of patients with ANCA-associated vasculitides is a major challenge. In contrast to antiproteinase 3-associated vasculitides, which has been widely studied in the literature, the relevance of antimyeloperoxidase antibody level assessment in predicting the relapse of vasculitis has not been well defined. In the present study, we aimed to determine the relevance of antimyeloperoxidase antibody assessment in the prediction of vasculitis relapse. To analyse the course of antimyeloperoxidase antibody levels during the follow-up of vasculitis, we studied patients from a single university centre who had had antimyeloperoxidase antibody assessment performed in the immunology laboratory with the same assay. A major premise of our study was that the reappearance of antimyeloperoxidase antibodies during remission identifies patients at risk of vasculitis relapse. The reappearance of antimyeloperoxidase antibodies was associated with a 117-fold higher rate of relapse in the following months, with a median period of 8 months (range 0–89) between the reappearance of antimyeloperoxidase antibodies and clinical relapse. Two-thirds of the clinical relapses occurred less than 12 months after the reappearance of antimyeloperoxidase antibodies. This may have important implications in the management of ANCA-associated vasculitides. Vasculitis relapses are often unpredictable and cause severe and life-threatening organ damage.53 Treatment of patients before the occurrence of clinical relapse could improve the prognosis of such patients. For instance, patients displaying an increase in ANCA levels during remission may be treated preemptively.54 Along this line, a recent but small prospective open-label trial55 showed that preemptive treatment with rituximab of patients with a rise in ANCA levels during remission could prevent relapse in most cases.
In addition, we found, in accordance with previous data, a higher rate of relapse in patients with WG. However, our study population includes WG patients with antimyeloperoxidase antibodies, which are usually considered to relapse less frequently than patients with antiproteinase 3 antibodies.53 This finding underlines the importance of the disease phenotype in patients with identical antibody profiles, ie, antimyeloperoxidase antibodies, but no conclusion can be drawn because of the small number of patients.
We found that the antimyeloperoxidase antibody level during treatment paralleled the patient’s clinical course. All patients achieved remission, and antimyeloperoxidase antibody levels significantly decreased in all patients between the time of diagnosis of vasculitis and the time of remission. However, no correlation was found between the magnitude of the decrease in antimyeloperoxidase antibody levels and the duration of remission. On the other hand, compared with the level at the time of remission, relapses were associated with an increase in the antimyeloperoxidase antibody level in all patients but one. In our study, relapse occurred exceptionally in patients with persistent negative or persistent positive antimyeloperoxidase antibodies. Indeed, only one out of 10 patients with persistent positive antimyeloperoxidase antibodies relapsed while experiencing a rise in the level of antimyeloperoxidase antibodies. In addition, among the 18 non-relapser patients with negative antimyeloperoxidase antibodies, only one patient exhibited a transient increase in antimyeloperoxidase antibodies levels after the withdrawal of azathioprine, but antimyeloperoxidase antibodies returned to the normal range (<20 IU/ml) within 6 months. For all the remaining relapsers, relapse occurred after the reappearance of antimyeloperoxidase antibodies. Therefore, the positive predictive value of the reappearance of antimyeloperoxidase is 90%, and the negative predictive value of persistent negative antimyeloperoxidase during treatment is 94%. A rise in the level of antimyeloperoxidase antibodies preceded clinical relapse in 91% of cases. In patients with MPA and WG, the positive predictive value of the reappearance of antimyeloperoxidase was 100% and 89%, respectively, whereas the negative predictive value of persistent negative antimyeloperoxidase was 100% for both.
Our findings are consistent with previous studies in which conclusions were limited by the small number of patients.32 35 44 A rise in the level of antimyeloperoxidase antibodies before or at the time of relapse was described in 73% to 100% and was followed by a clinical relapse in 79% to 100% of cases.32 35 44 Taken together, these findings suggest that through monitoring, antimyeloperoxidase antibody levels are a very useful and sensitive tool for the prediction of relapse in the follow-up of antimyeloperoxidase-associated vasculitides. Interestingly, the monitoring of antimyeloperoxidase antibody levels might represent an even more sensitive tool than antiproteinase 3 antibody levels for predicting relapse. Despite apparent rises in antiproteinase 3 antibody levels, there may be no clinical consequences in up to one-third of patients with antiproteinase 3-associated vasculitides.35 36 In addition, the largest study ever published on this topic did not confirm previous findings, as it did not find a correlation between antiproteinase 3 antibodies and disease activity in WG.40 In antimyeloperoxidase-associated vasculitides, such discrepancy does not exceed 9% in the present study and 0% to 21% in previous reports.32 35 44 This improved clinical relevance of antimyeloperoxidase antibody level assessment could be explained by the fact that antiproteinase 3 antibodies and antimyeloperoxidase antibodies exhibit distinct pathogenic roles. Only antimyeloperoxidase antibodies have been shown to be pathogenic in vitro and in vivo in adoptive transfer,7 8 whereas antiproteinase 3 antibodies have never been shown to exhibit any pathogenic role in animal models or humans.
Finally, we found that the level of antimyeloperoxidase antibodies at the time of clinical relapse of vasculitis for relapser patients and at the time of remission for non-relapser patients was correlated with the disease activity score, ie, the BVAS and the DEI. This finding is in accordance with previous reports,27 56 57 and suggests that patients with high levels of antimyeloperoxidase antibodies at relapse present a higher risk of severe organ damage than those with low levels of antimyeloperoxidase antibodies. This strong correlation of antimyeloperoxidase antibodies with BVAS is also consistent with the pathogenic role exhibited by ANCA.1 6
Limitations of the present study are represented by the retrospective character and the absence of systematic measurement of antimyeloperoxidase antibody levels. One cannot exclude the possibility that during the interval periods between each measurement, there may have been fluctuations in antimyeloperoxidase levels, undetected because of the absence of clinical activity of vasculitis. Other limitations of this study are its relatively small size, although it represents the largest study published so far on this topic, and the fact that three diseases (MPA, WG and CSS) are included.
In conclusion, in patients with antimyeloperoxidase-associated vasculitis, antimyeloperoxidase antibody levels during follow-up monitoring are a very useful and relevant surrogate marker of disease activity. In patients with persistent negative antimyeloperoxidase during treatment, the monitoring of antimyeloperoxidase antibody levels has positive and negative predictive values of more than 90%. Prospective studies are needed to confirm our findings and to evaluate the potential benefits of preemptive treatment in patients with clinical remission of vasculitis and a rise in the antimyeloperoxidase antibody levels.
REFERENCES
Footnotes
Competing interests None.
Ethics approval Ethics approval was obtained.