High rate of renal relapse in 71 patients with Wegener's granulomatosis under maintenance of remission with low-dose methotrexate




To examine the long-term efficacy of low-dose intravenous methotrexate (MTX) with and without concomitant glucocorticoids (GC) for remission maintenance in patients with generalized Wegener's granulomatosis (WG) in an open-label, prospective, standardized trial.


After induction of remission by cyclophosphamide and GC, 71 patients (41 males, 30 female) with initially generalized WG received low-dose methotrexate at 0.3 mg/kg body weight once weekly. At study-start 55 of 71 (77.5%) patients were on low-dose GC (mean 5.9 mg/day) which was tapered during the study. All patients underwent interdisciplinary staging at 3-month (and later at 6-month) intervals to assess disease activity and extent as well as side effects. End points were the first relapse or the end of study (January 2001).


Within a mean followup period of 25.2 months, 26 patients (36.6%) experienced a relapse after a mean of 19.4 months. Seventeen (65.4%) of these 26 patients had terminated GC therapy at the time of relapse. There was no difference in relapse rates among patients with and without concomitant GC at study start. Relapses occurred mainly in the initially involved organ systems, preferentially in the ear, nose and throat tract in 18 of 26 patients and the kidney in 16 of 26 patients. One renal relapse presented as rapid, progressive glomerulonephritis with lethal outcome. Further, 14 relapses were accompanied by a significant rise in creatinine values. In 15/26 patients the relapse was paralleled or preceded by a significant rise of antineutrophil cytoplasmic antibody titer. Two patients ceased MTX prematurely because of persistent leukopenia.


Weekly MTX is a well tolerated therapy for long-term maintenance of remission. However, one-third of the patients relapsed during ongoing MTX treatment, irrespective of whether they were still receiving GC. Because more than half of the relapses affected the kidney, close monitoring is indispensable.


Daily oral cyclophosphamide (CYC) with concomitant glucocorticosteroids (GC) is still indispensible for induction of remission in Wegener's granulomatosis (WG), especially in patients with life-threatening disease or severely impaired renal function. In view of the well recognized toxicity from this regimen (1, 2), the length of treatment with CYC for induction of remission has generally been reduced from originally over 2 years to a maximal length of 6–12 months (2, 3). Furthermore, even under prolonged CYC therapy, recurrence of the disease in up to 50% of the patients within the first 5 years after diagnosis could not be prevented (2, 4, 5). These circumstances have changed today's treatment of WG, because there is a strong requirement for a rather potent and low-toxicity maintenance of remission medication on which a patient can possibly stay for several years. Weekly low-dose methotrexate (MTX) has proven to be an effective drug for this situation (6–8). Its major limitation is its accumulation in patients with moderately to severely impaired renal function with subsequent bone marrow toxicity, because it is mainly eliminated by the kidneys. As opposed to other patient cohorts in which MTX was successfully used for induction of remission even in patients with glomerulonephritis in the course of their vasculitis with either no or mild renal insufficiency (6, 9), we observed 5 of 17 patients who developed a de novo glomerulonephritis under MTX for induction of remission (10). In order to assess the long- term outcome and safety with MTX we prospectively followed the outcome of 71 patients who were treated with MTX for maintenance of remission for a mean period of 25.2 months after induction of remission with CYC and GC.



A total of 71 patients, 41 male and 30 female with generalized WG were included in the study. The mean ± SD age was 49 ± 13.3 years (range 12–77). All patients fulfilled the American College of Rheumatology 1990 classification criteria (11) and the Chapel Hill Consensus Conference on the Nomenclature of Systemic Vasculitis 1992 definition (12) for WG. Fifty-eight of the patients (82%) had biopsy-confirmed WG with necrotizing vasculitis, granuloma, or both in 1 or more organ systems. In the patients in whom WG could not be confirmed histologically, the diagnosis was made on the basis of the typical history, characteristic clinical findings, and a positive classic antineutrophil cytoplasmic antibody (cANCA) on indirect immunofluorescence (IFT) with antigen specificity for proteinase-3 (PR-3) (Table 1). Sixty-nine of the 71 patients (97%) were cANCA positive, and all of them showed antigen specificity for proteinase-3. The remaining 2 patients were ANCA negative. Patients were recruited between January 1995 and June 1999. All consecutive patients without contraindications regarding MTX were enrolled, after induction of remission and informed consent. Fifty-three patients (75%) were enrolled during their initial episode of WG, and the remaining 18 patients had experienced a relapse of their disease.

Table 1. Baseline characteristics of 71 patients with WG treated with low-dose MTX with and without concomitant GC for maintenance of remission*
Parametern (SD)% (range)
  • *

    WG = Wegener's granulomatosis; MTX = methotrexate; GC = glucocorticoids; cANCA = classic antineutrophil cytoplasmic antibody; PR3 = proteinase-3; ACR = American College of Rheumatology (11); CHC = Chapel Hill Consensus Conference (12); CYC = oral cyclophosphamide; DEI = Disease Extent Index.

  • This calculation refers to the 56 patients who were still receiving GC at study start.

+cANCA at diagnosis69/7197
+PR3-ANCA at diagnosis69/69100
Biopsy proven WG58/7182
ACR 1990 criteria71100
CHC 1992 definition71100
CYC therapy before71100
CYC duration (months), mean, median13.8 (11.2)10 (1–66)
GC therapy for remission induction70/7198.6
GC therapy at study start55/7177.5
Daily GC dose (mg) at study start, mean (SD)5.9 (4.9)(1–25)
DEI at diagnosis, mean (SD)10.1 (2.2)(3–15)
DEI at start of study, mean (SD)2.1 (1.7)(0–6)
+cANCA at study start56/6986
cANCA titer at study start 1:; mean256(0–2048)
Followup under MTX, months, mean25.2, 14.5(1–71)

End points of the study were defined as the following: first relapse, scheduled termination of the maintenance treatment (see treatment protocol below), or end of the observation period in January 2001.

Exclusion criteria for treatment with MTX were chronic liver disease or regular alcohol intake that could not be ceased, renal insufficiency irrespective of its etiology with a serum creatinine > 1.5 mg/dl at study start, bone marrow insufficiency (leukopenia < 4,000/μl, hemoglobin < 10 gm/dl, or thrombocytopenia < 100,000/μl), or inadequate contraception.

Ethical approval for the study was obtained in 1992 from the ethics committee of the University of Lübeck.

Treatment protocol

Induction of remission

All 71 patients received induction of remission therapy consisting of oral continuous CYC, and 70 of 71 patients received concomitant GC. The initial oral CYC treatment was 2 mg/kg/day in all patients, given as a single morning dose that was always combined with an equivalent dose of mesna. The remainder of the mesna was divided into 2–3 doses daily (2). The CYC dose was adjusted according to the leukocyte count and tapered when leukocyte counts were lower than 4,000/nl. Leukocyte counts were performed during the treatment with daily CYC at least twice a week. The mean duration of daily oral CYC therapy was 13.8 months (median 10 months, range 3–66). The concomitant treatment with GC was started with 1 mg/kg/day for 2 weeks, which was subsequently tapered weekly by 10 mg until a dosage of 20 mg/day was reached, then finally tapered by 2.5 mg every 3 weeks until a dosage of 5 mg/day was reached. In sustained remission the dosage was reduced by 1 mg monthly. In most patients a GC dosage lower than the Cushing threshold was reached after 3–6 months.

Maintenance of remission

Patients were switched to maintenance medication with MTX after reaching complete or stable partial remission (unchanged smouldering disease for at least 3 months). To be eligible to switch to the MTX therapy patients had to fulfill criteria for remission over a period of 3 months and be taking a daily GC dosage ≤7.5 mg (i.e., in patients with longstanding remission but high GC dose, CYC therapy was maintained) (2). Seven patients were taking doses >7.5 mg GC (5 patients on 10 mg/day, 1 patient on 15 mg/day, 1 patient on 25 mg/day), but were still switched to MTX because of persistent low leukocyte counts on CYC therapy. The MTX therapy was started with 1 weekly intravenous dosage of 7.5 mg and increased (provided there were no contraindications to dose escalation such as leukopenia) until a dosage of 0.3 mg/kg of body weight once a week was reached. An equivalent dosage of folic acid was given 1 day after MTX application. The mean MTX dosage during the observation period was 22.5 mg/week. If a patient was not taking GC and was in complete remission for 6 months under the maintenance medication, the latter was tapered by 2.5 mg MTX per month and finally ceased. The mean Disease Extent Index (DEI) was 10 (median 10, range 3–15) at the time of diagnosis, and 2 (median 2, range 0–6) at time of switching to MTX (Table 1). Fifty-five patients (77.5%) received GC at a mean dose of 5.9 mg per day (median 5 mg) at study start. If possible the dosage was gradually tapered and discontinued as described above.

Assessment of disease activity and extent

Complete remission was defined as the absence of pathologic findings in clinical, radiologic, and seroimmunologic data, irrespective of the ANCA titer. Partial remission was defined as partial stable improvement of the disease over a period of at least 3 months. The reemergence of clinical symptoms attributable to active WG after a phase of complete or partial remission over at least 3 months was considered a relapse (1, 7). Major relapse means life- or organ-threatening disease activity with manifestation of acute renal failure, acute pulmonary hemorrhage, or involvement of heart or central nervous system requiring reinstitution of CYC. Minor relapse denotes a non–life-threatening flare of the vasculitis, usually treated by transient increase of the MTX and GC dose.

At 3-month (and later at 6-month) intervals all patients underwent a set of interdisciplinary clinical and seroimmunologic examinations for the presence of active disease. The same team of specialist clinicians in internal medicine, otorhinolaryngology, ophthalmology, neurology, and radiology performed the examinations each time (2, 7). The extent of WG was assessed using the recently validated DEI (13).

ANCA titers were determined serially at approximately 3-month intervals by indirect IFT as described earlier (14). In patients who experienced a relapse, the 3 sera obtained prior to the relapse were later retested for ANCA titer by IFT and direct PR3-enzyme-linked immunosorbent assay (ELISA) in order to assess to what extent a relapse is preceded or accompanied by a rise in ANCA titer. A significant rise in ANCA titer was considered if there was an increase of ≥2 titer steps for the IFT and ≥5 units/ml in the direct PR3-ANCA ELISA.

Statistical analysis

The data were analyzed with the SPSS statistical package (SPSS, Chicago, IL). Descriptive statistics (means, standard deviations, medians, and ranges) were calculated for each variable. ANCA titers were transformed by logarithm for the calculation of means. All tests of significance (t-test, chi-square test) were 2-tailed, and P values less than 0.05 were considered significant. For the presentation of the cumulative probability of relapse Kaplan-Meier survival curves were used.

In order to compare our population with 31 patients that were analyzed by Langford et al (8) we entered age and DEI into our database and performed t-tests for unrelated groups.

Because of the sample size and confirmation by the Kolmogorov-Smirnov test, normal distributions for all variables except CYC duration were assumed. Therefore we present means and standard deviations in the results section and tables. CYC duration was not normally distributed and therefore was presented as median, mean, and range. The higher mean indicates outliers in the upper range of CYC duration.


The epidemiologic data of the 71 patients with WG included in this study are presented in Table 1. Within a mean observation period of 25.2 months (median 24 months, range 1–71), 26 (36.6%) of 71 patients with WG receiving maintenance treatment of weekly low-dose MTX with or without concomitant corticosteroids experienced a relapse after a mean of 19.4 months (median 19 months; range 1–49). Figure 1 shows the appearance of the relapses within the study period of 77 months. In the patients who relapsed, the mean DEI increased from 2.2 (median 2, range 0–6) at study start to 4.7 (median 4, range 2–9) at the time of relapse. At that time, all patients were still receiving MTX a mean dosage of 18 mg/week (median 15 mg/week; range 7.5–27.5). Seventeen (65.4%) of the 26 patients who relapsed had already discontinued their GC treatment (6 patients before the start of the study, and 11 during maintenance treatment with MTX). The remaining 9 patients were given a mean GC dose of 5.1 mg (median 4.5 mg, range 1–12 mg) at the time of relapse. Approximately one-third of the 55 patients who were still taking GC and one-third of the 16 patients who were not taking GC at the start of the study subsequently relapsed (36.4% versus 37.5%).

Figure 1.

Cumulative probability of relapse in patients with Wegener's granulomatosis receiving treatment with intravenous methotrexate.

Eighteen relapses were classified as major, 16 of which required a switch to CYC treatment. One patient was treated with local laser therapy and local GC injections for a severe subglottic stenosis. Another patient received an increased dose of MTX plus additional cyclosporin A. When this failed to control the disease he also received CYC pulses, instead of daily CYC, because of a CYC-induced cystitis due to previous CYC therapy. This patient subsequently died of uncontrollable disease (for details see below). All 18 patients with a major relapse were additionally given an increased dose of GC (0.5–1 mg/kg). Eight relapses were determined to be minor, because they could be controlled by an increase of the GC dose and/or the MTX dose alone. The mean followup time to major relapse was 16 months, whereas the mean time to a minor relapse was 27.1 months (P = 0.019).

Most relapses occurred in the ear, nose, and throat tract (18/26) and, surprisingly, in the kidney (16/26). Relapses occurred to a lesser extent in the lung (5/26), skin (2/26), the peripheral nervous system (2/26), and the eyes (1/26). Constitutional and rheumatic symptoms (i.e., arthralgias, arthritides, myalgias) reemerged in 11 of 26 patients, respectively (Table 2). Most relapses affected organ systems that were already affected by the initial manifestation of the disease. However, a case of glomerulonephritis and a case of cutaneous vasculitis were each observed in 2 patients for the first time in their disease course at relapse during MTX therapy. Also 3 patients displayed arthralgias and arthritides for the first time at relapse (Table 2).

Table 2. Organ manifestations in 26 patients at the time of relapse during maintenance of remission with low-dose MTX*
Organ system according to DEIPatients, n (%)New organ manifestation at relapse receiving MTX, n
  • *

    MTX = methotrexate; DEI = Disease Extent Index (13).

  • A rise in serum creatinine was experienced by 15/16 patients. One presented with a rapidly progressive glomerulonephritis, and 1/16 had a new nephritic sediment only.

Upper respiratory tract18 (69.2)0
Kidney16 (61.5)1
Rheumatic symptoms11 (42.3)3
Constitutional symptoms11 (42.3)0
Lung6 (23.1)0
Skin2 (7.7)1
Peripheral nervous system2 (7.7)0
Inflammatory eye involvement (no granulomatosus changes)1 (3.8)0
Gastrointestinal tract0 (0)0
Central nervous system0 (0)0
Heart0 (0)0

Of note is the high number of renal relapses that occurred during MTX therapy. Of these 16 patients, 14 exhibited a rise in serum creatinine from normal values to 1.5–2 mg/dl plus a new nephritic sediment. One additional patient even developed a rapidly progressive glomerulonephritis (RPGN) combined with a severe pulmonary infiltration and pulmonary hemorrhage while receiving MTX and concomitant GC. He subsequently died of progressive disease that could not be controlled by an intensification of immunosuppressive therapy. One patient had a nephritic sediment only.

Concerning predictive factors for relapse under maintenance of remission treatment with MTX, neither the DEI at diagnosis of the WG nor the DEI at the start of the study differed significantly between patients who relapsed and those that did not (mean values 10.3 versus 10.0, and 2.2 versus 2.0) (Table 3). Figure 2 illustrates that the cumulative probability to experience a relapse was lower for patients who were in complete remission (DEI = 0), compared with patients in a state of partial remission at study start, although this difference in probability failed to reach statistical significance. Table 3 also shows that there were no differences in pretreatment, mean GC dose at study start, and number of patients not taking GC at study start between patients who relapsed and those who did not. The median duration of pretreatment with CYC was 10 months in both groups. The only significant difference between the 2 groups was found in the mean ANCA titer at the start of the study, which was significantly higher in patients who subsequently relapsed compared with those who did not relapse (1/128 versus 1/32). Apart from a slightly more frequent vasculitic eye involvement at study start in patients with subsequent relapse (3/26 versus 0/45, P = 0.045, Fisher's exact test) there were no statistically significant quantitative differences in organ manifestation patterns either at diagnosis or at study start between patients who relapsed and those who did not. Seven of twenty-six patients (27%) who relapsed and 11 of 45 (24%) who did not relapse during therapy had exhibited previous relapses.

Table 3. Patients who relapsed compared with patients who did not relapse during maintenance of remission with low-dose MTX*
 Relapses (n = 26)Nonrelapses (n = 45)t-test
  • *

    MTX = methotrexate; CYC = oral cyclophosphamide; GC = glucocorticoids; DEI = Disease Extent Index; cANCA = classic antineutrophil cytoplasmic antibody.

  • Represents the mean MTX dose of the 23/45 patients who did not relapse, and who were still taking MTX at study end. At that time 22/45 were not taking MTX and in complete remission.

  • SD is presented in titer steps.

Male/female, n17/924/21Chi-square, NS
Age at diagnosis, mean ± SD (range) years47.4 ± 12.6 (12–69)49.9 ± 13.8 (25–77)NS
CYC-duration, mean ± SD months16.3 ± 14.312.4 ± 8.9NS
CYC-duration, median (range) months10 (5–66)10 (3–50)
GC for remission induction, n (%)25 (96.2)45 (100)NS
DEI at diagnosis, mean ± SD (range)10.3 ± 2.1 (7–13)10.0 ± 2.3 (7–15)NS
DEI at start of study, mean ± SD (range)2.2 ± 2.0 (0–5)2.0 ± 1.6 (0–6)NS
DEI = 0 at study start, n (%)9 (35)13 (20)NS
GC-dose at study start, mean ± SD (range) mg/day4.7 ± 4.9 (0–20)4.5 ± 4.2 (0–25)NS
Patients off GC at study start, n (%)6 (23)10 (22.2)NS
Followup (= time to end-point), mean ± SD (range) months19.4 ± 11.4 (1–49)28.5 ± 15.2 (9–77)P = 0.01
Weekly MTX dose at end-point, mean ± SD (range) mg18.04 ± 6.5 (7.5–30)17.6 ± 6.3 (7.5–30)NS
Daily GD dose at end-point, mean ± SD (range) mg1.7 ± 3.2 (0–12)1.02 ± 1.9 (0–5)NS
Off GC at end point, n (%)17 (65)34 (75)NS
cANCA titer at study start, mean ± SD (range)128 ± 2.7 (0–1024)32 ± 3.7 (0–1024)P = 0.024
cANCA negative at study start, n (%)2/26 (8)8/43 (19)NS
Figure 2.

Cumulative relapse-probability for patients with Wegener's granulomatosis in complete remission (Disease Extent Index = 0) versus patients with Disease Extent Index > 0.

In 15 of 26 patients who experienced a relapse while receiving MTX, the relapse was preceded or paralleled by a significant increase in ANCA titer by IFT (n = 6) or direct PR3-ELISA (n = 9). Only 3 patients showed a parallel increase in ANCA titer in both tests, including the patient who developed RPGN in the course of his fatal relapse.

In the group of patients who did not relapse, 34 of 45 could taper their GC dose to zero during the study period, and 22 of the patients could also cease their MTX medication because of complete remission. Complete remission was exhibited by 13 patients at study start and by 25 patients at study end, indicating that MTX has the potential to further suppress the disease activity. Among those who did not relapse, 12 of 45 (27%) experienced an increase of ≥2 titer steps in their cANCA titer by IFT during the study period (versus 6/26 [23%] of those who relapsed, see above), 13 of 45 (29%) exhibited a decrease of the ANCA titer to the same extent, and 7 of the latter group eventually became ANCA negative.

The following side effects were observed during therapy with MTX: leukopenia in 9 of 71 patients, leading to temporary dose reduction of MTX in 7 and withdrawal from MTX in 2 patients, and one infectious episode in 7 of 71 patients. Five of the infections were bacterial, specifically: 2 urinary tract infections caused by Escherichia coli, 1 upper airway infection caused by Staphylococcus aureus, and upper airway infections caused by Pseudomonas aeruginosa that occurred twice in 1 patient with preexisting bronchiectases. All bacterial infections required antibiotic treatment. In two other patients segmental herpes zoster occurred, requiring antiviral therapy. In no case did the patient need to be hospitalized because of the infection or its treatment. None of the observed infections were serious. Pneumocystis carinii pneumonia, cytomegalovirus infections, mucositis, or MTX-induced pneumonitis did not occur. Apart from the 2 withdrawals mentioned above, there were no further cessations of MTX therapy due to adverse effects.


We have followed 71 patients with initially generalized WG, after induction of remission with CYC and GC, under low-dose MTX therapy for maintenance of remission. This is the largest reported cohort of patients with WG treated with MTX. Just over one-third of the patients experienced a relapse, mostly comprising several organ systems (during the maintenance phase after a mean time of 19.4 months from study start). Alarmingly, in almost two-thirds of the relapses the kidneys were involved, including a rise in creatinine in almost all patients.

Today's challenge in treating WG lies in the prevention of relapses rather than in the achievement of the first remission. In the literature, relapse rates for patients with WG are variable but generally unsatisfactorily high: 50% in 8 years (1), 71% after a followup of more than 5 years (2), 44% after 18 months from diagnosis of the disease (15), and 46% after a median of 27 months in a cohort consisting of patients with WG and microscopic polyangiitis (5). In our cohort of 71 patients with WG, we observed a slightly lower relapse rate (36.6% after a mean period of 19.4 months from start of the maintenance regimen) that corresponds to a mean of 35.7 months from diagnosis. In accordance with the literature (5), all our relapses occurred while the patients where still taking MTX, although two-thirds of them had already discontinued taking GC.

In the attempt to minimize toxicity from long-term treatment with CYC (1), low-dose MTX has successfully been introduced into the therapy of WG for induction (1, 6, 9, 10) and maintenance of remission (7, 8). All reports including the present showed correspondingly that MTX was a relatively safe drug in terms of adverse effects. In addition, the occurrence of opportunistic infections appeared to be related to the dose of concomitant corticosteroids rather than attributable to MTX alone. This can be seen by comparing 2 cohorts of WG patients treated with similar regimens of low-dose MTX but different doses of GC, one low (7) and one higher (6).

However, as opposed to a group from the NIH who used MTX successfully in WG patients with active glomerulonephritis for induction and maintenance of remission (8, 9), we have seen an alarming number of renal relapses (16/26 = 61.5%) under maintenance of remission with MTX. It must be emphasized that all but one of these patients had a rise in serum creatinine from normal values to 1.5–2 mg/dl; one patient even experienced a RPGN and subsequently died. This happened despite a regular, thorough, and meticulous disease assessment program for activity and extent in our center every 3–6 months. Furthermore, this was combined with patient education in order to train them to detect warning signs of an imminent relapse as early as possible.

If one compares our patient cohort with Langford's (8) whose study has a similar set-up (induction with CYC, maintenance with MTX), one can state that the patients in Langford's cohort are significantly younger (37 versus 49 years, P = 0.000), and have a lower DEI at initiation of induction of remission treatment with CYC and GC (mean DEI 5.4 versus 10.1, P = 0.000). The percentage of renal involvement prior to induction of remission was insignificantly lower in Langford's cohort, compared with ours (17/31 patients, [54.8%] versus 51/71 patients [71.8%]; P = 0.444). Whether the advanced patient age and the higher DEI in our cohort could account for a higher renal relapse rate compared with that in Langford's cohort remains speculative. So far there are no published data showing that these parameters would be associated with a higher relapse rate. However, as Langford stated, the severity of the disease and the presence of glomerulonephritis were not associated with a higher frequence of relapse in their cohort consisting of a total of 31 patients (8), which is probably too small of a number of representative statistic calculations. In our significantly larger cohort there was no difference between patients who did and did not relapse in terms of DEI and age at diagnosis, nor was the incidence of renal involvement different in the 2 groups, either at diagnosis, or at start of the study.

Although Langford et al did not specify the mean GC dose accompanying MTX, the dose seems comparable with that used in our study. In the Langford et al study supplemental folinic acid was given in doses of 5–10 mg (8) whereas our patients received folic acid in the same dose as MTX. Studies have confirmed a positive effect of folic or folinic acid in preventing side effects of MTX (16–19). Van Ede et al showed that patients who received folate supplementation needed slightly higher doses of MTX to obtain the same clinical effect as patients who did not receive folic acid supplementation (17). Ortiz et al did not confirm differences in folic acid dose in their meta-analysis, however similar conclusions were drawn by Endresen and Husby (18) and by Shiroky (19). It therefore seems unlikely that different doses of folic or folinic acid were responsible for the different results between our study and that of Langford et al. In addition, MTX was administered intravenously in all patients in our study, leading to a higher bioavailability compared with an oral application of MTX in the NIH study.

In our own preceding study (7) with a much smaller number of patients (who were not enrolled in this study) the relapse rate of patients receiving MTX as maintenance of remission medication was lower than the one presented here. However, the mean followup time was also shorter than in the cohort presented here, and it seems likely that the incidence of relapse is a function of time. This would also be supported by the preliminary results of a prospective, randomized, controlled European Vasculitis Study Group (EUVAS) trial CYCAZAREM, where the relapse rates under CYC and azathioprine were 17% and 16% respectively, but followup was only up to month 18 from diagnosis (20).

The question remains: what is the value of the continued use of a maintenance of remission medication and for how long should it be given? In the attempt to use CYC for as short a time period of time as possible, the maintenance of remission medication aims at consolidating the achieved remission, including conversion of a partial remission into a complete remission. This happened in 9 responders in our cohort, who were in partial remission at study start and finished the study in complete remission. Whether immunosuppression beyond the function of consolidation of the first remission carries more benefit than risks (infections, bone marrow suppression, permanent damage) awaits further assessment. This is the aim of the randomized, prospective, controlled EUVAS trial REMAIN. It tests the hypothesis that prolonged maintenance therapy with low-dose prednisolone and azathioprine reduces long-term morbidity in systemic vasculitis by reducing the frequency of relapse, when compared with cessation of therapy in the second year (3). Conversely, prolonged immunosuppression, even with drugs that are less aggressive than CYC may enhance the risk of myelodysplastic syndrome and lymphoma, which in turn renders potent immunosuppression in case of a relapse impossible (21).

Unfortunately, there are still no reliable predictors available that identify patients who have an increased risk for relapse, which would facilitate the decision of prolonged immunosuppression. In our present cohort, apart from a significantly higher median ANCA titer at study start in the cohort of patients who relapsed, there were no differences concerning age, sex, and disease extent at diagnosis, study start, or pretreatment between the 2 groups, that would indicate that either was at higher risk for relapse. Patients with a DEI = 0 at study start experienced fewer relapses than patients with a DEI > 0, but this difference failed to reach statistical significance. Even serial ANCA testing, at least in our hands, has not proven to be a reliable tool to identify patients at risk for imminent relapse. The relapse was preceded or paralleled by an increase in ANCA titer in only 57.7% of the patients, and this figure resulted from the addition of the results of 2 ANCA tests (the IFT and direct PR3-ANCA ELISA). Therefore, in order to reach that percentage one would have to do at least 2 tests serially, which is expensive and time-consuming. In summary, serial ANCA testing with a single-test system as used in daily practice was not sufficient to predict relapses.

In conclusion, MTX with and without concomitant GC is a safe medication for maintenance of remission in WG. However, 36.6% of the 71 study patients experienced a relapse at the end of the second year of treatment. In almost two-thirds of the patients, the relapse was with renal involvement, among them 1 patient with RPGN. Patients that are not in complete remission and have a high ANCA titer at the time of switch from CYC to MTX may be at higher risk for a future relapse. We still recommend the use of MTX as maintenance of remission medication for patients with generalized WG following remission induction with CYC. However, we emphasize that close monitoring, including urinary sediment, is mandatory in order to treat a relapse as early as possible, because the most frequently involved organ in relapse in our cohort was the kidney. This is a major problem, because this situation usually requires a switch back to CYC and is associated with a poorer prognosis. Since the use of MTX is impossible in patients with moderately to severely impaired renal function, a promising alternative drug for maintenance of remission may be leflunomide, because it does not accumulate with decreased renal function. In a preliminary pilot study for maintenance of remission in 20 WG patients after a median followup of 24 months, 9 minor relapses but only 1 major relapse was observed (22). Recently, a prospective, randomized, controlled, multicenter German study to compare MTX with leflunomide for maintenance of remission in WG has been started. The identification of risk factors that predict a future relapse is one of the major issues for future studies in order to design more individual maintenance of remission strategies.