Progress in treatment of ANCA-associated vasculitis

Autoantibodies to neutrophil cytoplasmic antigen-associated vasculitis (AAV) is characterised by inflammation of blood vessels. The introduction of immunosuppressive therapy with glucocorticoids and cyclophosphamide transformed AAV from a fatal condition to a largely treatable condition. Over the past 30 years, considerable progress has been made refining immunosuppressive regimens with a focus on minimising toxicity. There is, however, a high unmet need in the treatment of AAV. A proportion of patients are refractory to current therapies; 50% experience a relapse within 5 years and treatment toxicity contributes to mortality and chronic disability. As knowledge of the pathogenesis of vasculitis grows, it is mirrored by the availability of biological agents, which herald a revolution in the treatment of vasculitis. Lymphocyte-targeted and cytokine-targeted agents have been evaluated for the treatment of AAV and are entering the routine therapeutic arena with the potential to improve patient outcomes. As rare diseases, treatment advances in vasculitis depend on international collaborative research networks both to establish an evidence base for newer agents and to develop recommendations for patient management.

Rates of cardiovascular disease and malignancy are increased and it is unclear to what extent these refl ect the underlying disease process or its treatment [10]. Early treatment-related leucopaenia and infection, and later cardiovascular disease, are now the predominant causes of death, rather than uncontrolled vasculitis [11].
Treatment-related toxicity, together with the 10 to 30% of patients who do not respond to traditional immunosuppressive agents and pursue a refractory course and the 50% of patients who relapse within 5 years despite continued immunosuppression, has shifted the focus of treatment for AAV [12]. From learning how to best use traditional agents to minimise toxicity without losing effi cacy, to the advent of new targeted biological agents, a second revolution in the treatment of AAV is on the horizon.

Pathogenesis of AAV
Both genetic susceptibility and environmental exposures contribute to the aetiology of AAV. ANCA have been demonstrated pathogenic in animal models but their contribution to human disease remains unclear. Transfer of murine myeloperoxidase-ANCA IgG to mice without functioning B cells or T cells results in a pauci immune, necrotising crescentic glomerulonephritis, similar to that seen in AAV in humans [3]. Human anti-PR3 antibodies from individuals with AAV were recently demonstrated to cause comparable renal and pulmonary lesions in mice with a humanised immune system [13]. Two lines of investigation have linked infections to ANCA formation through molecular mimicry [14,15]. Fimbriated bacteria can induce novel ANCA, antibodies to human lysosome membrane protein-2, which can induce a crescentic glomerulo nephritis in animal models [15]. Microbial superantigens are responsible for T-cell dysregulation in Kawasaki disease [16], and infection with Staphylococcus aureus is associated with relapse of GPA [17,18]. In sensitive assays, the PR3-ANCA binding level is predictive of outcome and rises in ANCA precede relapse. Th e clinical correlation with ANCA is closest for alveolar and glomerular capillaritis, both lesions capable of being induced by ANCA in experimental models. Patients who are consistently ANCA-negative, however, may fi t the clinical phenotype of AAV, and the effi cacy of B-cell depletion with rituximab is not associated with ANCA status. Despite the pauci immune nature of the histology in ANCA vasculitis, there is evidence of a role for immune complexes and complement activation in renal vasculitis.
Th ere is activation of circulating T lymphocytes and B lymphocytes and infi ltration of plasmoblasts into aff ected tissues [19]. Autoreactive B cells are necessary for the development of autoantibody-producing cells, but may play a more important role in supporting autoreactive Tcell activity through antigen presentation, co-stimulation and the direct production of proinfl ammatory cytokines, such as IL-6 and TNFα. In view of their role as precursors of ANCA-secreting plasma cells, B cells are a therapeutic target in AAV. T cells also play an important role in the pathogenesis of AAV [20]. Th e autoantibodies are class switched (that is, mainly IgG), which means the autoreactive B cell has received cognate T-cell help. Also, in biopsy specimens T cells are seen to infi ltrate tissues, causing damage via direct cytotoxicity and by the recruitment and activation of macrophages [21]. A CD8 T-cell gene expression signature has been associated with relapse frequency in AAV [22], and T-cell targeted therapies including cyclosporin [23] and alemtuzumab have been eff ective [24].

Treatment
After making a diagnosis of systemic vasculitis, treatment is divided into remission induction, then 3 to 6 months of intensive therapy, followed by a longer period of remission maintenance therapy. Disease state defi nitions have been agreed for active disease, remission, low disease activity state and refractory disease [25]. Further key management concepts are the subgrouping of patients by severity, the minimisation of adverse event risk and proactive management of increased cardiovascular and malignancy risks.
GPA has been subdivided into limited or nonsevere versus generalised or severe -the most important determi nant being the presence or absence of nephritis. An analogous system was developed by the European Vasculitis Study Group to include both GPA and MPA (Table 2) [25] in order to defi ne eligibility for a sequence of randomised controlled trials aimed at defi ning the optimal standard of care.
no randomised controlled trials to support their use. Evidence is also lacking to guide dosage. Current regimens start oral prednisone/prednisolone at 1 mg/kg/ day, reducing to 10 to 20 mg daily by 12 weeks, with up to 3 g intravenous (i.v.) methylprednisolone for generalised or severe presentations. Although they contribute to disease control, there is increasing evidence that highdose steroids contribute to morbidity [26]. Th e PEXIVAS trial aims to address the effi cacy of a rapidly reducing glucocorticoid regimen, as well as the place of plasma exchange in severe AAV (ClinicalTrials.gov: NCT00987389).

Cyclophosphamide
Cyclophosphamide, together with high-dose gluco corticoids, was introduced empirically for induction treatment of AAV over 40 years ago, as a lymphocyte-depleting therapy following its use in lymphoproliferative diseases. Together with glucocorticoids, cyclophosphamide remains the standard of care in consensus guidelines for the treatment of generalised disease [27]. Th e use of cyclophosphamide has been refi ned, however, employing strategies to reduce exposure and cumulative toxicity.
Th e effi cacy of a sequential design whereby cyclophosphamide was replaced by an alternative at the time of remission was explored in the CYCAZAREM trial [28]. In this study, 155 patients were randomised to receive either 1 year of oral cyclophosphamide or 3 months (maximum 6 months) of oral cyclophosphamide followed by azathioprine. Both groups followed a similar steroid taper. Th ere was no diff erence in relapse rates between the two groups at 18 months, demonstrating that the duration of exposure to cyclophosphamide may be safely reduced.
An i.v. pulsed administration has potential advantages over daily oral cyclophosphamide of reduced exposure, bladder protection and improved compliance. A metaanalysis of three studies of i.v. cyclophosphamide, including 143 patients, demonstrated that pulsed regimens reduced cumulative cyclophosphamide exposure by 50%, and were at least as eff ective at inducing remission, with fewer infective and myelosuppressive side eff ects but possibly with an increased risk of relapse [29].
Th e CYCLOPS study randomised 149 patients with generalised AAV to receive either i.v. cyclophosphamide, 15 mg/kg at 2-week intervals for the fi rst three doses and every 3 weeks thereafter, or daily oral cyclophosphamide 2 mg/kg/day [30]. Induction therapy was continued for 3 months after remission was achieved in both groups prior to a switch to azathioprine for maintenance therapy until the end of the 18-month trial. Th ere was no diff erence in time to achieve remission or the proportion of patients in remission at 9 months between the groups, despite the i.v. cyclophosphamide group receiving 8.2 g as compared with 15.9 g cyclophosphamide for the daily oral route. Th e reduced dose resulted in fewer leucopaenic episodes, but this did not translate to fewer infections. Long-term follow-up (median 4.3 years), however, has shown that pulsed i.v. cyclophosphamide is asso ciated with a higher relapse rate, but this does not seem to impact survival or long-term morbidity [31]. Th e CYCLOPS protocol also standardised cyclophosphamide dose reductions for age over 60 years and for renal impairment, which improved the safety of these regimens.

Methotrexate
Following nonrandomised studies of methotrexate in limited GPA, 100 patients with a new diagnosis of early systemic AAV, without critical organ manifestations and with creatinine <150 μmol/l were randomised to receive either methotrexate targeting 25 mg weekly or oral cyclophosphamide in the NORAM study [32]. Both groups followed the same glucocorticoid regimen, and treatment was withdrawn at 12 months. Methotrexate was not inferior at inducing remission, but remission was slower than with cyclophosphamide in those patients with more extensive disease or pulmonary involvement. Th ere was less leucopaenia with methotrexate, but more liver dysfunction and a higher relapse rate.

Adjunctive therapies for severe disease
For those presenting with severe vasculitis when vital organ function is already compromised, adjunctive therapies have been considered that include, plasma exchange, i.v. methylprednisolone, intravenous immunoglobulin (IVIg) and TNFα blockade. Despite an absence of randomised evidence there is widespread use of i.v. methylprednisolone in doses between 1,000 and 3,000 mg for generalised and severe presentations. Advanced renal failure at presentation correlates with an increased risk of end-stage renal failure and death, and the aims of therapy are both to control disease and recover organ function [6]. Several small trials have produced confl icting results regarding the merits of plasma exchange, although improvements in recovery rates for renal vasculitis presenting with advanced renal failure were found [33]. Th e MEPEX trial compared adjunctive therapy with plasma exchange or i.v. methylprednisolone in 137 patients with AAV and serum creatinine >500 μmol/l (5.8 mg/dl) at presentation [34]. Both groups received oral cyclophosphamide and steroids. Plasma exchange decreased the risk of progression to end-stage renal failure by 24% at 12 months, but had no eff ect on longer-term renal function or survival [35].
Th e mode of action of plasma exchange is unclear, but removal of complement and coagulation factors, ANCA and other circulating factors is likely to be important. A smaller trial found benefi t in terms of renal outcome for patients presenting with less severe nephritis [36]. Th ere remains uncertainty regarding the long-term benefi ts of plasma exchange, and a recent meta-analysis failed to fi nd convincing evidence that there was an infl uence on the composite endpoint of death or end-stage renal disease after 1 year [37].
Severe alveolar haemorrhage is the most common vasculitic cause of early death, and, in view of similarities in pathogenesis with renal vasculitis, plasma exchange has been recommended. Single-centre reviews have suggested a role for plasma exchange, but one study of ventilator-dependent haemorrhage still found a 50% mortality despite this intervention [38,39].

Newer induction agents B-cell depletion therapies (rituximab)
Following the successful use of rituximab in rheumatoid arthritis and increasing appreciation of a role for B cells in the pathogenesis of AAV, several studies have reported responses to ritxumab for patients with relapsing or refractory disease [40]. Th e largest multicentre series of 65 patients with refractory disease reported complete remission in 75%, partial remission in 23%, and only one nonresponder. However, 57% of those who achieved remission relapsed after a median of 12 months [41].
Two randomised control studies -RITUXVAS [42], a randomised trial of a rituximab-based regimen versus cyclophosphamide in ANCA-associated renal vasculitis; and RAVE [43], rituximab for ANCA-associated vasculitis -have demonstrated that rituximab has similar effi cacy to cyclophosphamide for the induction of remission in AAV. Table 3 summarises the key features of these studies.
Neither study reported a reduction in severe adverse events with rituximab, which suggested that high-dose glucocorticoids and the disease itself are the main contributors to these events, and that current cyclophosphamide regimens are, in expert centres, reasonably safe. Although two doses of cyclophos phamide were administered in addition to rituximab in the RITUXVAS study, the RAVE study used no immunosuppressive in the rituximab group. Response rates were comparable, so it is now justifi ed to withdraw immunosuppression at the time of rituximab treatment. Th e response rates in both the rituximab and cyclophosphamide limbs of the RAVE study were lower in those patients who had glucocorticoids withdrawn by 6 months, highlighting their importance in remission induction strategies. However, this observation must be balanced against their longer-term toxicity, and the optimal time for withdrawal is yet to be established. For the subgroup presenting with relapsing disease as compared with new-onset disease, rituximab was more eff ective in obtaining remission than cyclophosphamide and can be recommended for this indication. Th e role of rituximab for new-onset disease is more controversial; both trials had a short duration of 12 and 6 months, respectively, and longer-term results are awaited.
Meanwhile rituximab can be recommended as an alternative to cyclophosphamide in generalised, severe or refractory AAV when the latter is contraindicated. However, controlled data for the use of rituximab in localised and early systemic disease are lacking at present.

Remission maintenance therapy
AAV is a relapsing condition, with disease and treatment factors infl uencing relapse risk. In a retrospective cohort study of 246 patients with AAV and renal involvement, the relapse rate was 34% with a median time to relapse of 13 months after induction of remission [12]. Aggregated data from four European Vasculitis Study Group trials, encompassing the spectrum of disease severity, reported a 50% relapse rate at 5 years (unpublished data), with a higher rate for GPA (70%) than for MPA (30%). Relapse risk is increased by upper respiratory tract involvement, nasal carriage of S. aureus infection and the absence of renal disease. Persistent ANCA positivity at the time of remission and a subsequent rise in ANCA level can be predictive of relapse [48][49][50]. Th e utility of ANCA monitor ing varies with disease type, however, and ongoing treatment dissociates changes in ANCA from changes in disease activity, which accounts for some negative reports [51]. Withdrawal of immunosuppression or glucocorticoids is associated with a higher relapse rate [52]. Th e cyclophosphamide group had a relapse rate of 46.5% at 18 months in the NORAM trial [32], when therapy was stopped at 12 months, compared with 13.7% at the same time point in the cyclophosphamide group in the CYCAZAREM trial [28], when azathioprine was sub stituted for cyclophosphamide at 12 months and continued until the end of the study. Consensus guidelines advocate continuation of maintenance immunosuppression for at least 18 to 24 months [27,53]. Th e REMAIN study is comparing the relapse rate between 2 years and 4 years of maintenance therapy [54].

Standard remission maintenance therapy Azathioprine
Cyclophosphamide is eff ective at inducing remission in AAV, and has been continued for prolonged periods to maintain remission. However, the dangers of high cumulative toxicity, particularly bladder cancer and myeloproliferative disease, have led to arguments for safer therapies [9]. Th e CYCAZAREM trial compared continued cyclophospha mide 1.5 mg/kg/day after remission induction to a switch to azathioprine 2 mg/kg/day [28]. No diff erence in subse quent relapse rate or other endpoints was seen at 18 months, and this study has validated a sequential approach to treatment with cyclophosphamide with drawal at the time of remission.

Methotrexate
Th e WEGENT study randomised 126 patients to receive either methotrexate or azathioprine as maintenance therapy after successful induction of remission with cyclo phosphamide [55]. Th ere were no diff erences between groups in either relapse or toxicity. Use of methotrexate is usually restricted to those with creatinine <150 μmol/l (1.8 mg/dl).

Glucocorticoids
Th ere is variance in current practice regarding the use of prednisolone for remission maintenance. A review of published randomised trials in which the glucocorticoid protocol was defi ned found by multivariate analysis that glucocorticoid withdrawal was the strongest predictor for relapse [52].

Alternative remission maintenance therapies
A study involving 64 patients with GPA, comparing metho trexate (20 mg weekly) with lefl unomide (30 mg daily) as a maintenance therapy, was terminated early due to a higher major relapse rate in the methotrexate group (46%) compared with the lefl unomide group (23%) [56]. However, this was balanced by an increased rate of severe adverse events in the lefl unomide group. Th e position of lefl unomide as a remission maintenance agent remains uncertain.
Long-term treatment with sulfamethoxazole/trimethoprim is based on the assumption that controlling bacterial infection will reduce relapse risk [18]. When dosed at sulfamethoxazole/trimethoprim 960 mg twice a day in GPA there were reduced respiratory tract infec tions and respiratory tract relapses in a double-blind trial [57]. It is unclear whether this antibiotic has intrinsic antivasculitis activity separate from its antibiotic role, but the study supports the need for careful microbiological evaluation and treatment in GPA, especially in patients with respiratory tract involvement pursuing a relapsing course.

Mycophenolate mofetil
Following its success in the fi eld of transplantation, MMF has been considered as an alternative to azathioprine for remission maintenance therapy in AAV. Results from a small pilot study in 11 patients were promising, with only one relapse occurring after 14 months of follow-up [58]. Th e IMPROVE randomised study that compared MMF 2 g/day and azathioprine 2 mg/kg/day as maintenance agents after cyclophosphamide induction in 174 patients, however, found an increased hazard ratio of 1.7 for relapse in the MMF group, and a shorter time to relapse [59]. MMF cannot be recommended as a fi rst-line remission maintenance agent in AAV but it may have a role for patients intolerant of azathioprine for whom methotrexate is contraindicated by renal failure.

Refractory disease
Patients with refractory disease encompass those who have disease progression or inadequate disease control despite induction therapy; those who relapse on main tenance immunosuppression, and those who are intolerant of standard therapy, most commonly cyclophosphamide. Refractory disease has been defi ned in a European League Against Rheumatism statement to facilitate further clinical investigation [25]. In such challenging patients, it is important to identify underlying drives such as infection or malignancy, and then to consider whether further immunosuppressive agents or alternative agents are indicated.
Rituximab is now the preferred agent for refractory disease [41]. Rituximab was more eff ective than cyclophosphamide in the relapsing subgroup of the RAVE trial [40], although a retrospective series suggests it is less eff ective for granulomatous lesions, especially orbital masses than for vasculitic manifestations [60].

Additional agents Intravenous immunoglobulin
Small studies have demonstrated a benefi t of IVIg in persistent disease [61,62]. Postulated mechanisms for action of IVIg include the interference of binding of ANCA to their antigens through idiotypic mechanisms and the inhibition of ANCA-induced neutrophil activation [63,64]. Th irty-four patients with persistent disease activity were randomised to receive either IVIg (a single course at 2 g/kg) or placebo in a double-blind trial. Fourteen of 17 patients in the IVIg group as compared with 6/17 patients who received placebo had a reduction in disease activity, but this eff ect was not sustained beyond 3 months [65]. IVIg has also been used to maintain remission in patients with refractory disease [66]. Cost and limited availability have been barriers to the use of IVIg; however, it remains an alternative to consider in sepsis or pregnancy in order to minimise or avoid the use of standard agents.

TNF blockade
TNF blockade ameliorates disease in experimental models of AAV [67]. However, a benefi cial eff ect of TNF blockade in human disease has been more diffi cult to demonstrate. Etanercept did not improve the rates of stable remission in GPA when added to glucocorticoids and cyclophosphamide or methotrexate [68]. Rates of serious adverse events exceeded 50% in both groups and a possible increase in solid malignancies was recorded in the etanercept group [69]. All cases of malignancy occurred in patients with prior cyclophosphamide exposure, however, and TNF blockade should therefore be avoided in these patients [70]. Both infl iximab and adalimumab have been proposed as adjunctive agents in renal vasculitis and a potential for steroid sparing has been demonstrated [71][72][73]. Prolonged infl iximab for refractory AAV has been complicated by infection and relapse.

Alternative strategies T-cell-targeted agents
T-cell depletion with anti-thymocyte globulin led to remissions in refractory AAV in the SOLUTION trial but was poorly tolerated [74]. Alemtuzumab, an anti-CD52 humanised monoclonal antibody, causes similar profound lymphocyte depletion. A retrospective cohort study found remission of vasculitis in 60/71 (85%) refractory AAV patients after 134 mg alemtuzumab. A minority had sustained treatment-free remissions for many years, but 43/60 (72%) patients relapsed after a median of 9 months [24]. Retreatment was eff ective. Infective morbidity was high in those over 50 years old and with fulminant disease or vital organ failure. Th ese results have confi rmed a pathogenic role for T cells and the potential for T-cell depletion to produce sustained remissions. A trial of abatacept that blocks co-stimulation via CD28 in GPA is ongoing (ClinicalTrials.gov: NCT00468208).

Gusperimus
Gusperimus (previously termed 15-deoxyspergualin) is a synthetic analogue of the antibiotic spergualin, and suppresses lymphocyte and macrophage function and impairs neutrophil production. Forty-two of 44 (95%) patients with relapsing GPA responded to six cycles, each lasting 21 days with a 7-day washout period, of 0.5 mg/kg gusperimus, followed by azathioprine for 6 months [75]. Twenty of 44 (45%) patients achieved complete remission, although 43% of patients relapsed a median of 6 months after achieving remission. Severe adverse events occurred in 53% of patients and were predominantly leucopaenias, although they rarely led to treatment discontinuation. Extended courses of gusperimus as a maintenance agent have been shown to be eff ective and relatively safe [76,77].

Autologous stem cell transplantation
Th ere is limited experience of stem cell transplantation in vasculitis. In the 15 cases recorded in the European Blood and Marrow Transplantation database there was a response rate of over 90%, although relapses still occurred [78]. Th is approach is yet to be evaluated prospectively, and appropriate patient selection -avoiding those with high prior cyclophosphamide exposure and irreversible end organ damage -is crucial in view of the high risk of complications.

Churg-Strauss syndrome
Although a form of AAV, the treatment of CSS has a number of distinct features. Glucocorticoid therapy alone may be suffi cient, and patients are stratifi ed on the basis of the Five Factor Score. Cardiac, gastrointestinal and central nervous system involvement are adverse prognostic features, as is creatinine >140 μmol/l and proteinuria >1 g in 24 hours [79]. A total 93% of 72 patients with a Five Factor Score of 0 achieved remission with glucocorticoids alone, although 35% relapsed on prednisolone taper, and many required long-term maintenance glucocorticoids in order to control asthma [80].
Cyclophosphamide is the fi rst-line therapy for those with any adverse prognostic features. Th is treatment is highly eff ective when administered at a dose of 600 mg/m 2 every 2 weeks for a month and 4-weekly thereafter, with 87.5% of 48 patients achieving remission. Cumulative cyclophosphamide exposure infl uences relapse risk, however, with 85.7% of patients relapsing after six doses compared with 61.9% after 12 doses [81].
Regardless of disease severity, relapses are common, and current maintenance strategies for CSS are as for GPA and MPA. Rituximab has been shown to be effi cacious in refractory CSS. Th is observation is based on only 20 reported cases, however, and there were two reports of severe bronchospasm during rituximab infusion, which have raised concerns over possible hypersensitivity reactions triggered by rituximab in CSS [82]. In both situations, antihistamines but not steroids were administered at the time of rituximab infusion, and it may be that co-administration of intravenous steroids may reduce this risk.
IL-5 is an important survival factor for eosinophils, and mepolizumab (a monoclonal antibody directed against IL-5) is a promising novel therapeutic agent. Two pilot studies, using monthly mepolizumab infusions but involving only 17 patients in total, have demonstrated the effi cacy and steroid sparing potential of mepolizumab in CSS [83,84]. Relapses were common on discontinuation of the drug, however, and larger, longer-term studies of this approach are required.

Future therapies
With the demonstration of the effi cacy of rituximab in AAV, other B-cell-targeted therapies hold potential. Th e humanised anti-CD20 and anti-CD22 antibodies ocrelizu mab, ofatumumab and epratuzumab have diff erences in pharmacodynamics and frequency of antiglobulin responses compared with rituximab [85]. Th e Blymphocyte-stimulating protein (BLyS; also known as BAFF), a B-cell survival factor, is a further target. BLyS exerts its modulatory eff ect on B cells via three receptors; the transmembrane activator and calcium-modulator and cyclophilin inhibitor; B-cell maturation protein; and the BAFF receptor. Patients with active untreated GPA have signifi cantly higher serum BLyS levels than healthy controls [86]. Also, following B-cell depletion with rituximab in SLE, BLyS levels signifi cantly increase and then normalise as B-cell populations recover [87]. B-cell maturation protein is found on the surface of plasma cells and BLyS seems to protect lymphoid resident B cells from depletion, so BLyS inhibition may reduce plasma cell numbers [88] and enhance the susceptibility of tissue-residing B cells to depletion [89]. Belimumab is a humanised monoclonal antibody that inhibits BLyS. Effi cacy and tolerability when added to standard therapy, but without a steroid-sparing eff ect, have recently been demonstrated in two phase III clinical trials in patients with SLE, without signifi cant renal involvement.
Belimumab may yet prove to be very useful in AAV.
Co-stimulatory blockade, which aims to prevent autoreactive T-cell activation by B cells or other antigenpresenting cells, should work in AAV. Abatacept (a CTLA4-Ig fusion protein) is under evaluation in GPA. Further biological agents targeting downstream cytokines are in trials in other autoimmune diseases, such as tocilizimab (an anti-IL-6 agent) in rheumatoid arthritis and IL-17 blocking agents in Crohn's disease. Historically, many treatments for AAV have been adopted from other autoimmune diseases, and at least some of these newer biological agents used presently for other indications will probably be tried in AAV.
Other therapeutic mechanisms of relevance to vasculitis include inhibitors of downstream infl ammatory signalling pathways. After phosphorylation, spleen tyrosine kinase initiates several signalling cascades involving phosphatidylinositol-3-kinase that lead to aberrant neutrophil responses -including release of superoxide, which in turn leads to tissue damage. Inhibition of spleen tyrosine kinase in both mouse models and in human rheumatoid arthritis has exerted a positive clinical eff ect [90,91]. Inhibition of phosphatidylinositol-3-kinase suppresses infl ammation in mouse models of rheumatoid arthritis and SLE [92]. Although the pathology of AAV contains few immune deposits, complement depletion has abrogated experimental myeloperoxidase-ANCA renal vasculitis and is a potential target in human disease.

Discussion
Th e introduction of combined glucocorticoid and immuno suppressive therapy transformed the survival of patients with systemic vasculitis. Although there has been little change in the actual agents used to treat AAV, there has been considerable progress in optimising treatment regimens to minimise toxicity [93]. Despite these eff orts, however, cyclophosphamide and high-dose glucocorticoids remain toxic and 10% of patients are intolerant of these therapies. In addition, relapsing disease is common, with over 50% of patients experi encing a relapse within 5 years despite continued immuno suppression. Furthermore, refractory disease aff ects 10 to 20% of patients and poses a signifi cant management challenge. Th ere is therefore a clear need for more effi cacious and safer therapies.
Rituximab is the best studied biological agent in AAV. Th e recently published randomised controlled trials RITUXVAS and RAVE have shown that rituximab is similar to cyclophosphamide, in terms of both effi cacy and safety, for induction of remission in AAV in the short term. Subgroup analysis in the RAVE trial found the effi cacy of rituximab to be superior to at 6 months for patients with relapsing disease.
Potential adverse eff ects of any new therapy must be carefully evaluated prior to recommendation of use. Infection, particularly in the context of hypogammaglobulinaemia, is a concern following administration of rituximab. In a retrospective cohort of 105 patients with primary systemic vasculitis or SLE, 39% developed a mild infection, most commonly of the urinary tract, chest or upper respiratory tract. A total 28.5% of patients developed a severe infection, which is higher than in studies of rituximab in rheumatoid arthritis but lower than in vasculitis trials using alemtuzumab or deoxy spergualin. Patients with low IgG and IgM levels have been shown to be at signifi cantly increased risk of infection, and immunoglobulin levels, as well as white cell counts, should be routinely monitored following rituximab administration [94]. In a retrospective series, late-onset neutropaenia occurred in 3/13 patients with GPA, all of whom developed infection needing intravenous antibiotics [95]. Further study is required into longer-term outcomes after rituximab treatment, including cardiovascular disease, infertility and malignancy.
Concern has been expressed regarding the development of progressive multifocal leucoencephalopathy, an opportunistic infection caused by the JC virus. At least 57 cases have been reported following rituximab therapy, four in patients with autoimmune disease (two SLE patients, one rheumatoid arthritis patient and one immune thrombocytopaenia patient) [96]. Th is association is confounded by the use of immunosuppression prior to the administration of rituximab, and the observation that progressive multi focal leucoencephalopathy is seen in patients with autoimmune disease not treated with rituximab. Unfortunately, these reports highlight the possibility of rare, severe, adverse eff ects that may not be detected in clinical trials. It is important that patients who receive new biological agents enter registry followup, to document late outcomes and side eff ects. Relapses are common after rituximab, and maintenance strategies are required that keep patients well yet avoid complications of prolonged B-cell depletion.
Improved patient subgrouping and tailoring of therapy will reduce unnecessary exposure to toxic therapies. Older people pose particular challenges. Th e peak incidence of AAV occurs in the age group of 65 to 70 years. One-year mortality is 23% in those aged over 60 years, compared with just 5% in the under-60 age group. Much of the excess mortality is due to infection. More attention is being focused on older people. Th e CORTAGE study, coordinated by the French Vasculitis Study Group, is comparing low-dose glucocorticoids with standard-dose glucocorticoids, with standard immuno suppression in patients over 65 years old. Newer biomarkers are required to predict treatment response, relapse risk and drug selection. Urinary monocyte chemoattractant protein-1 correlates with the degree of active vasculitis on renal biopsy and may become useful in renal vasculitis [97]. With advances in understanding of the genetic basis of disease and early results of transcriptome analysis, there is a real possibility that other clinically useful tests will become available.
Advances in classifi cation and evidence-based therapy of vasculitis have resulted from collaborative networks both within and between countries. To date, there has been little pharmaceutical investment in these disorders. With increasing physician collaboration and industry interest in rare diseases, progress is likely to accelerate. Consensus recommendation statements have been published that will lead to harmonisation of therapy. With the establishment of registries and quality indicators there is hope that the variable quality of care that patients have received in the past will be improved.

Conclusion
Considerable progress has been made over recent years in refi ning the use of traditional immunosuppressive agents, namely glucocorticoids and cyclophosphamide, in systemic vasculitis, with the objective of minimising exposure and hence toxicity. Progress in classifying disease severity has provided the rationale to tailor the potency of immunosuppressive regimes to the extent of disease. Th is has facilitated the stratifi cation of patients into a number of randomised controlled trials looking at newer immunosuppressive agents (Table 2).
However, the pace of change is accelerating with greater understanding of the pathogenesis of vasculitis, and a number of targeted biological agents are entering clinical use. Th ere is an ongoing need for improved therapies for patients who are intolerant of current immunosuppression, and for those who pursue a relapsing or refractory course. A range of therapeutic agents is likely to be required for a condition as heterogeneous as systemic vasculitis. Th e effi cacy, safety and tolerability of each agent must be thoroughly assessed, however, and the international collaborative clinical networks established in the vasculitis community will make it possible to develop this evidence base.
As the number of therapeutic options grows, the treatment of vasculitis will probably become more complex for clinicians. With the introduction of biomarkers to more accurately assess disease activity and relapse risk, however, we should aim to simplify treatment for patients by individually tailoring their immunosuppressive regimen. As disease control improves, we should focus more on longer term causes of morbidity such as cardiovascular disease and on important patient-centred outcomes, including quality of life. Strauss syndrome; GPA, granulomatosis with polyangiitis; IL, interleukin; i.v., intravenous; IVIg, intravenous immunoglobulin; MMF, mycophenolate mofetil; MPA, microscopic polyangiitis; SLE, systemic lupus erythematosus; TNF, tumour necrosis factor.