Introduction

The antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitides (AAV) comprise granulomatosis with polyangiitis (GPA, formerly Wegener’s granulomatosis), microscopic polyangiitis (MPA) including its renal limited form (idiopathic pauci-immune necrotizing crescentic glomerulonephritis, and Churg–Strauss syndrome [1]). These diseases, in particular GPA and MPA, have a very poor prognosis if left untreated. Mortality in GPA was described as 63 % (35 out of 56 patients) at 6 months after diagnosis [2]. The introduction of (oral) cyclophosphamide (CYC) by the pioneering work of Fauci et al. at the National Institute of Health (NIH), USA, has dramatically improved survival [3]. Oral CYC was used in GPA both for induction and maintenance treatment. A regimen of (in most cases) oral CYC with glucocorticoids at the NIH resulted in marked improvement of 91 % of 158 GPA patients with 75 % of patients achieving complete remission [4]. However, 50 % of patients achieving remission relapsed, 13 % of patients died, 86 % of patients had serious irreversible morbidity resulting from the disease, and 42 % of patients suffered from irreversible side effects of treatment [4]. Long-term oral CYC was especially associated with hemorrhagic cystitis and bladder cancer (with a cumulative risk of 5 % after 10 years follow-up increasing to 16 % after 15 years [5]). In addition, opportunistic infections, particularly during leukopenia resulting from CYC, and infertility were frequently observed. In the NIH series, 46 % of patients (73 out of 158) experienced 140 episodes of serious infection during CYC treatment [4]. Thus, although prognosis was strongly improved by the introduction of oral CYC, the serious side effects of CYC in conjunction with the far from optimal efficacy prompted the search for alternative treatment modalities. In this review, I will discuss these various modalities and suggest that we might abandon CYC for the treatment of AAV in the near future.

Do We Need Cyclophosphamide for Induction of Remission?

As mentioned, the introduction of oral CYC has dramatically improved survival in AAV. Nevertheless, side effects of oral CYC are considerable, significant morbidity and damage still occur and relapsing disease is frequent. Therefore, reduction in dosage or abandonment in general of CYC for induction of remission in AAV, if possible, would be welcome.

Oral or Pulse Cyclophosphamide?

The first step to reduce the dosage of CYC consisted of replacing oral CYC by pulse intravenous CYC for induction of remission. In the so-called CYCLOPS study, 149 patients with AAV and non-life-threatening renal involvement were randomized to either daily oral CYC (2 mg/kg) or pulse i.v. CYC (15 mg/kg, three times every 2 weeks, thereafter every 3 weeks until remission), both arms with prednisone [6]. Azathioprine (AZA) was given after remission was obtained in both arms. No difference was observed in time to remission and proportion of patients reaching remission (87.7 % vs. 88.1 %). Thirteen out of 76 patients relapsed after induction of remission with i.v. CYC compared to 6 out of 73 patients in the oral group (NS). The cumulative dose of CYC was significantly lower in the i.v. group compared to the oral group (8.2 g vs. 15.9 g, p < 0.001), and less leukopenia occurred in the i.v. group. Long-term follow-up [7] disclosed significantly more relapses in the i.v. group, but outcome in terms of endstage renal failure, serum creatinine, duration of immunosuppression or adverse events was not different between both groups. This study shows that i.v. CYC can replace oral CYC for induction of remission in generalized AAV.

Can CYC be Replaced by Mycophenolate Mofetil?

A randomized controlled trial comparing mycophenolate mofetil (MMF) with intravenous pulse CYC for induction of remission (acronym MYCYC, coordinated by the European Vasculitis Study Group (EUVAS)) is still underway. Small open series, however, have shown that MMF in a target dose of 2 g daily is effective in inducing remission of AAV with moderate renal involvement [8, 9]. Stassen et al [10] used MMF (2 g daily with corticosteroids) for induction of remission in 32 consecutive patients with AAV who relapsed and could not be treated with CYC. Twenty-five (78 %) patients reached complete remission and in all but one of the remaining patients, partial remission after 12 months. MMF was tapered by 500 mg every 3 months in patients in remission. This resulted in the frequent occurrence of relapses (relapse-free survival at 1, 3, and 5 years was 63, 38, and 27 %, respectively). In a small controlled study [11] MMF was as effective as i.v. CYC in Chinese MPA patients with renal involvement. Thus, awaiting the results of the MYCYC trial, MMF may be used for induction of remission in patients with non-life-threatening AAV who cannot tolerate CYC.

Methotrexate for Induction of Remission

In milder cases of AAV, methotrexate (MTX) has been used for induction of remission. In a controlled trial, 100 patients with active AAV but without critical organ manifestations and a serum creatinine <150 μmol/l, were randomized either to oral CYC (2 mg/kg) or MTX (20–25 mg/week) in combination with prednisone in both arms [12]. Drugs were tapered after induction of remission and withdrawn at 12 months. The percentage of patients reaching remission at 6 months was not different between both arms (89.8 % for MTX, 93.5 % for CYC). Relapses at 18 months were, however, more frequent in the MTX-arm (69.5 %) than in the CYC-arm (46.5 %) with a shorter time to remission in the MTX group. The high relapse rate in both arms is undoubtedly due to the lack of long-term maintenance treatment in this study. With respect to adverse events, leukopenia was more frequent in the CYC group and liver function disturbances in the MTX group. Thus, in milder cases of AAV, MTX can be used but relapses are frequent.

Rituximab, a New Player?

More recently, rituximab (RTX) has been used for induction of remission in AAV [13, 14]. In the so-called RAVE study, 197 patients with severe GPA or MPA, either at first presentation or at the time of relapse, were included. Patients got 1–3 g i.v. methylprednisolone and were randomized either to RTX (375 mg/m2 i.v. weekly for 4 weeks) or oral CYC (2 mg/kg/day), both in combination with prednisone. Steroids were tapered and the primary endpoint was complete remission off prednisone at 6 months. Sixty three out of 98 (64 %) patients on RTX reached the primary endpoint versus 52 out of 99 (53 %) patients on CYC. This led to the conclusion that RTX is not inferior to CYC for induction of remission. In relapsing patients, RTX was even superior to CYC (66.7 % reaching the primary endpoint on RTX versus 42.0 % on CYC, p = 0.01). There were no significant differences between both arms in adverse events. Thus, RTX is at least as effective for induction of remission of severe GPA and MPA as oral CYC. This has led to registration of RTX for induction of remission of AAV by the FDA. Interestingly in this study, no maintenance treatment was applied in the RTX group whereas the CYC group received maintenance treatment with AZA. At 18 months after study entry, no differences in relapse rate were observed between both arms [15]. As expected, more relapses occurred in PR3-ANCA-positive patients than in MPO-ANCA patients. More relapses were also seen in patients included with a relapse than in patients included with a new diagnosis of AAV.

Another controlled study has also shown the potential of RTX to induce remission in severe cases of AAV [14]. In this study, 44 patients with newly diagnosed AAV and (severe) renal involvement were randomized, in a 3:1 ratio, either to RTX (375 mg/m2 i.v. weekly for 4 weeks), with two i.v. pulses of CYC, or to pulse i.v. CYC (as in the CYCLOPS study), with corticosteroids in both arms. The primary endpoint was sustained remission at 12 months. There was no difference in the number of patients reaching the primary endpoint between both arms (76 % in the RTX group and 82 % in the CYC group had sustained remission at 12 months). Also, no differences were observed in mortality, severe adverse events, and renal function at 12 months.

The above-mentioned studies show that RTX can be used for induction of remission in patients with severe AAV, even in those with impaired renal function at presentation. In addition, many open series have demonstrated that RTX could be effective in cases of AAV refractory to conventional treatment [16, 17]. This has led to recommendations for the use of RTX in AAV [18, 19]. The general conclusion from these overviews is that RTX is an effective therapy in refractory AAV and that the drug is as effective as CYC in newly diagnosed cases. However, in view of the relatively short observations on RTX, long-term follow-up is needed to get a better insight into efficacy and safety in the long run.

Adjunctive Treatment in Life-Threatening Disease

Do we need more intensive treatment in patients with AAV who present with life-threatening disease, in particular renal failure and respiratory insufficiency? It is common use to start with i.v. methylprednisolone, 1 g on 1–3 consecutive days, in patients with severe AAV at presentation. Also, plasma exchange has been suggested as an additive way to remove pathogenic substances including ANCA in life-threatening disease. In a randomized controlled study, the adjunctive efficacy of seven plasma exchanges was compared with that of methylprednisolone (1 g i.v. on 3 consecutive days) [20]. Patients with newly diagnosed AAV and a serum creatinine of >500 μmmol/l were included (n = 137). At 3 months, 69 % of patients on plasma exchange compared to 49 % on methylprednisolone (p = 0.02) were alive and dialysis independent (primary outcome). Plasma exchange reduced the risk of endstage renal disease from 43 to 19 % at 12 months as compared with methylprednisolone.

Currently, the adjunctive effect of plasma exchange is being evaluated in patients with newly diagnosed or relapsing AAV with severe vasculitis but not just renal or respiratory insufficiency. This study, called PEXIVAS, will also evaluate different corticosteroid regimens.

Current recommendations for induction of remission in AAV as based on EULAR and British Society for Rheumatology (BSR) recommendations [19, 21] are given in Table 1.

Table 1 EULAR and BSR recommendations for induction of remission in ANCA-associated vasculitis (21,19)
Full size table

Maintenance of Remission

As mentioned before, the cumulative dose of CYC is a critical factor for the development of side effects such as bladder cancer, myelosuppression, and infertility. In the original NIH approach, oral CYC was continued for at least 1 year after induction of complete remission and then tapered by 25 mg decrements every 2 to 3 months until discontinuation but doses was increased when symptoms reappeared [4]. As remission was reached within 6 months in most cases, accumulation of CYC usage particularly occurred during maintenance treatment. So, the question was whether maintenance treatment with CYC could be replaced by maintenance azathioprine treatment.

Azathioprine Instead of Cyclophosphamide for Maintenance

The EUVAS compared CYC with AZA for maintenance treatment of patients with AAV [22]. Patients with active AAV with renal involvement or other threatened loss of function of a vital organ were included and remission was induced with oral CYC (2 mg/kg/day) and prednisolone (1 mg/kg/day with tapering). Patients received at least 3 months of induction treatment with oral CYC and prednisolone. At remission, patients were randomized either for oral CYC (1.5 mg/kg/day) or AZA (2 mg/kg/day), both with 10 mg of prednisolone. After 12 months of maintenance treatment all patients continued with AZA (1.5 mg/kg/day) and 7.5 mg of prednisolone. Primary end point was the occurrence of relapses within 18 months from study entry. This study showed, firstly, the efficacy of the induction regimen with oral CYC as 93 % (144 out of 155) of patients reached remission, although seven patients (5 %) died during the first 3 months. Secondly, no differences were seen in the occurrence of relapses between both arms: 10 relapses occurred in the CYC group (13.7 %) and 11 relapses in the AZA group (15.5 %), p = 0.65. No differences were seen in severe adverse events between both arms during the study period. Also, renal function at the end of the study and disease activity scores were not different between both arms. This so-called CYCAZAREM study shows that, during a period of 12 months, AZA is as effective as CYC for maintenance of remission without differences in side effects. However, a few remarks have to be made. First, long-term follow-up is necessary to demonstrate that AZA is, indeed, as efficacious for persisting remission and restoration of renal function as oral CYC. Secondly, long-term follow-up is also required for assessment of late side effects of treatment such as bladder cancer, myelodysplasia, and other malignancies, as well as assessment of cumulative damage. These data are eagerly awaited. Nevertheless, the results of this study have led to the replacement of CYC by AZA for maintenance treatment of AAV.

Is Methotrexate Effective for Maintenance of Remission?

As AZA still may have undesired adverse events in some patients, the question arises whether AZA can be replaced by MTX as maintenance treatment of AAV. The French Vasculitis Study Group addressed this question [23]. Patients with active AAV, that is GPA or MPA, were included. Induction treatment consisted of intravenous pulses of CYC in combination with corticosteroids (three times 1 g of methylprednisolone followed by oral prednisone, 1 mg/kg/day with tapering after 3 weeks). After remission had been achieved, patients were randomized to either AZA (2 mg/kg/day) or MTX (0.3 mg/kg/week increasing to 25 mg/week) for a period of 12 months after which drugs were discontinued. Primary end point was an adverse event requiring discontinuation of the drug or death; secondary endpoints were severe adverse events and relapses. In this study, 79 % (126 out of 159) of patients reached remission. After 12 months of maintenance treatment, there were no differences between both arms in primary endpoint (7 patients had to stop AZA versus 11 patients, including one death, in the MTX group, p = 0.21). Also, no differences were observed in the secondary end points (severe adverse events in 5 out of 63 patients in the AZA group versus 11 out of 63 patients in the MTX group; 23 relapses in the AZA group versus 21 relapses in the MTX group, most relapses occurring after discontinuation of the drugs). So, again, during short-term follow-up, MTX appears as effective as AZA for maintenance of remission whereas a slight tendency seems present for more adverse events during MTX treatment. We may conclude from this study that MTX can be used for maintenance treatment of AAV in patients who cannot tolerate AZA. Nevertheless, long-term follow-up is needed for a definite conclusion as to the potency of MTX versus AZA for keeping patients in remission and for evaluation of side effects of drugs and damage. In this respect, an open series study in which MTX was used for maintenance treatment demonstrated a relatively high relapse rate of 36.6 % after 20 months with more than half of the relapses affecting the kidney [24].

Is Mycophenolate Mofetil an Alternative for Azathioprine?

MMF has been suggested, at least in one study [25], to be more efficacious than AZA for maintenance treatment in lupus nephritis. So the EUVAS group started a study to compare MMF with AZA for maintenance treatment in AAV [26]. Induction treatment in this study consisted of CYC (either orally or as intravenous pulses) in combination with corticosteroids. At remission, patients were randomized to either AZA (2 mg/kg/day for 12 months, 1.5 mg/kg/day for another 6 months, thereafter 1 mg/kg/day for another 24 months) or MMF (2 g/day for 12 months, 1.5 g/day for another 6 months, thereafter 1 g/day for another 24 months). Surprisingly, relapses were more common in the MMF-arm (42 out of 76 patients) than in the AZA-arm (30 out of 80 patients, p = 0.03). Other outcome measures such as renal function and cumulative damage did not differ. This study demonstrates that AZA is the preferred drug for maintenance treatment in AAV but its efficacy is still not optimal as a substantial number of patients still relapse.

Rituximab for Maintenance of Remission?

In view of the limited efficacy of AZA to maintain remission in AAV, different approaches are being considered. Rituximab, a chimeric monoclonal antibody directed against the CD20 molecule on B-lymphocytes, has been used for induction of remission in AAV, as discussed before, and proven at least as effective as oral CYC in a large randomized controlled trial [13]. In this trial, patients who achieved remission on oral CYC were switched to AZA for maintenance of remission whereas patients achieving remission on rituximab did not receive any maintenance treatment. Interestingly, no differences in relapse-free remission were observed at 18 months after inclusion between both arms (preliminary data, [15]). However, also in this study, percentages of relapsing patients were still high in both arms. This prompted the use of rituximab as maintenance treatment in AAV. In a retrospective analysis, Rhee et al [27] reported their experience with rituximab maintenance treatment (1 g i.v. every 4 months) in 39 patients with AAV. They showed that this approach allowed discontinuation of immunosuppressives and corticosteroids in a substantial number of patients without the occurrence of relapses. These data are in line with the preliminary data from the Mayo Clinic [Specks U, personal communication] demonstrating the potential of intermittent rituximab administration to prevent the occurrence of relapses in patients with AAV, particularly in those at risk for relapsing disease (see later). However, prospective controlled trials with long-term follow-up are needed to show not only the efficacy but also the safety of intermittent rituximab administration for maintaining patients with (relapsing) AAV in remission. Currently, an international randomized controlled trial (RITAZAREM) has been started to compare rituximab with AZA for maintaining remission in AAV.

In conclusion, we can abandon CYC as a drug for maintaining remission in AAV (Table 2). AZA appears as the drug of choice with, if not tolerated, the possibility to replace it with MTX. However, many patients still relapse, particularly when these drugs are tapered or discontinued. Future trials should demonstrate if rituximab is more effective than AZA or MTX for keeping patients into remission. Long-term safety of rituximab may, however, be an issue. Otherwise, more insight into the factors that underlie the occurrence of relapses is necessary. This insight will have therapeutic consequences such as the selection of those patients who need long-term maintenance treatment (personalized medicine). A discussion on risk factors for relapse follows in the next paragraph.

Table 2 EULAR and BSR recommendations for maintenance of remission in ANCA-associated vasculitis [19, 21]
Full size table

Risk Factors for Relapse in AAV

As mentioned, relapses are frequent in AAV but not every patient has relapsing disease. Knowing which patients are at high risk of relapse would certainly influence decisions on duration and level of maintenance treatment. Relapses occur more frequently in patients with PR3-ANCA compared to those with MPO-ANCA which corresponds with a diagnosis of GPA versus a diagnosis of MPA [28]. Furthermore, relapsing disease is associated with chronic nasal carriage of Staphylococcus aureus and with persistence of ANCA after induction of remission [29] (Fig. 1). Also, upper-airway, lung-, and heart involvement are associated with relapse [28, 30] whereas a serum creatinine level >200 μmol/l at the time of diagnosis is associated with a reduced risk of relapse [31]. The association of relapsing disease with PR3-ANCA-positive GPA with involvement of the respiratory tract suggests that factors triggering relapse are residing in the (upper) airways. In this respect, the role of S. aureus is intriguing [32]. Here, the potential of maintenance treatment with co-trimoxazole to strongly reduce the occurrence of relapses is highly relevant [33] (Fig. 2).

Fig. 1
figure 1

Disease-free interval and carrier status. Disease-free interval of 57 patients with GPA grouped according to Staphylococcus aureus carrier status. The time of the disease-free interval was counted from the beginning of the most recent period of disease activity (either initial diagnosis or relapse; p < 0.001). From Ref. [29], with permission

Full size image
Fig. 2
figure 2

Disease-free interval from the start of co-trimoxazole or placebo treatment to relapse in patients with GPA. The difference in the disease-free interval between the co-trimoxazole group and the placebo group was statistically significant (by the log-rank test) at 24 months (relative risk of relapse, 040; 95 % confidence interval, 017 to 0.98). From Ref. [33], with permission

Full size image

In conclusion, based on the data presented, there is a rationale to individualize maintenance treatment in AAV. Patients with GPA with persisting or recurrent upper-airway involvement and chronic nasal carriage of S. aureus can be selected for maintenance treatment with co-trimoxazole. Ongoing maintenance treatment (with AZA or RTX) should be considered for PR3-ANCA patients with recurrent relapses.