Management of ANCA associated vasculitisBMJ 2020; 368 doi: https://doi.org/10.1136/bmj.m421 (Published 18 March 2020) Cite this as: BMJ 2020;368:m421
- Zachary S Wallace, assistant professor of medicine1 2 3,
- Eli M Miloslavsky, assistant professor of medicine2 3
- 1Clinical Epidemiology Program, Division of Rheumatology, Allergy, and Immunology, Mongan Institute, Department of Medicine, Massachusetts General Hospital, Boston, MA, USA
- 2Rheumatology Unit, Division of Rheumatology, Allergy, and Immunology, Department of Medicine, Massachusetts General Hospital, Boston, MA, USA
- 3Harvard Medical School, Boston, MA, USA
- Correspondence to: @zach_wallace_md on Twitter) (or
Anti-neutrophil cytoplasmic antibody (ANCA) associated vasculitis (AAV) is a small to medium vessel vasculitis associated with excess morbidity and mortality. This review explores how management of AAV has evolved over the past two decades with pivotal randomized controlled trials shaping the management of induction and maintenance of remission. Contemporary AAV care is characterized by approaches that minimize the cumulative exposure to cyclophosphamide and glucocorticoids, increasingly use rituximab for remission induction and maintenance, and consider therapies with less toxicity (for example, methotrexate, mycophenolate mofetil) for manifestations of AAV that do not threaten organ function or survival. Simultaneously, improvements in outcomes, such as renal and overall survival, have been observed. Additional trials and observational studies evaluating the comparative effectiveness of agents for AAV in various patient subgroups are needed. Prospective studies are necessary to assess the effect of psychosocial interventions on patient reported outcomes in AAV. Despite the expanding array of treatments for AAV, little guidance on how to personalize AAV care is available to physicians.
Anti-neutrophil cytoplasmic antibody (ANCA) associated vasculitis (AAV) is a necrotizing vasculitis primarily affecting small to medium sized vessels.1 AAV is a heterogeneous condition that includes three clinicopathologic conditions: granulomatosis with polyangiitis (GPA, formerly Wegener’s granulomatosis), microscopic polyangiitis, and eosinophilic granulomatosis with polyangiitis (EGPA).1 It affects men and women from a wide age spectrum and diverse racial backgrounds (fig 1). The most commonly affected organs include the respiratory tract, eyes, kidneys, skin, and nervous system.
Most patients with AAV are ANCA positive at some point in their clinical course, with antibodies targeting either proteinase 3 (PR3-ANCA positive) or myeloperoxidase (MPO-ANCA positive).345 Mounting evidence suggests that phenotyping patients according to ANCA type (PR3-ANCA positive or MPO-ANCA positive), rather than as GPA or microscopic polyangiitis, may better identify homogeneous groups that share similar genetics, pathogenesis, organ involvement, and response to treatment (fig 1).678910
Pivotal trials (table 1 and table 2) over the past several decades have shaped contemporary strategies for induction and maintenance of remission, which have led to improvements in relapse rates, renal outcomes, quality of life, morbidity, and overall survival.2627282930313233 However, AAV remains associated with a higher risk of death than the general population,3435 and patients continue to accrue organ damage from disease activity and treatment.36
Here, we focus on clinical trials that have shaped the contemporary management of induction and maintenance of remission in AAV, specifically GPA and microscopic polyangiitis; EGPA is considered a distinct disease with different approaches to management. We consider how the increasing amount of data from trials and observational studies might inform a personalized approach to AAV care.
Incidence and prevalence of AAV
A population based study conducted in Olmstead County, Minnesota, USA, estimated the annual incidences of GPA and microscopic polyangiitis to be 1.3 (95% confidence interval 0.8 to 1.8) per 100 000 and 1.6 (1.0 to 2.2) per 100 000, respectively. The prevalence of AAV in that study was 42.1 (29.6 to 54.6) per 100 000.37 Similar estimates of annual incidence of GPA were reported in a large US health insurance claims database (1.3 (1.2 to 13.4) per 100 000),38 as well as the UK General Practice Research Database (0.8 (0.8 to 0.9) per 100 000).39
Severity of AAV varies widely, leading many people to differentiate between severe and non-severe disease (fig 2). Severe disease is characterized by manifestations that threaten the function of vital organs, such as diffuse alveolar hemorrhage, glomerulonephritis, mononeuritis multiplex, sensorineural deafness, scleritis, or gangrene.41424344 Non-severe disease, in contrast, is characterized by disease affecting the sinuses, nares, and/or mucocutaneous surfaces, as well as pulmonary nodules, tracheobronchial disease, and arthritis. Despite this dichotomization, manifestations considered severe may be indolent and non-severe manifestations may contribute significantly to disease morbidity.
Sinonasal, pulmonary, and musculoskeletal manifestations are the most common presenting symptoms. A minority of patients present with chronic non-severe disease, often limited to the sinuses and the upper airway. More often, patients with non-severe disease develop severe disease, after months to years of non-severe manifestations.45 This highlights the importance of early recognition and treatment initiation. Severe disease, however, may be fulminant and unheralded. Renal involvement is the most common severe manifestation and can lead to end stage renal disease (ESRD), which is estimated to affect 30% of patients with renal involvement after five years.26
Generally, severe AAV is fatal if left untreated.46 Even with treatment, excess mortality remains. A recent meta-analysis found that the standardized mortality ratio of death in AAV compared with the general population was 2.7 (95% confidence interval 2.3 to 3.2).3547
Pathogenesis of ANCA associated vasculitis
The cause of AAV is unknown, but approximately 20% of disease risk is due to genetic factors, which differ between PR3-ANCA positive and MPO-ANCA positive patients.104849 The role of MPO-ANCA in pathogenesis is well established given observations that the transfer of MPO-ANCA antibodies into experimental mouse models induces vasculitis.5051 Evidence supporting the pathogenic role of PR3-ANCA is less strong but is inferred from certain observations: PR3-ANCA antibodies appear years before clinical presentation; genetic variants in proteinase 3 (the antigenic target of PR3-ANCA) are seen in PR3-ANCA positive AAV; PR3-ANCA titers correlate with disease activity in some patients; and B cell targeted therapy is efficacious in PR3-ANCA positive patients.52 As such, current therapies include broad immunosuppression and B cell targeted treatment.
However, as our understanding of the pathogenesis of AAV expands, so do the therapeutic options. Myeloperoxidase and proteinase 3 are enzymes found in neutrophils, which are the targets of ANCA antibodies and play a key role in pathogenesis.50 Neutrophil extracellular traps (NETs) have been recently recognized as important in the pathogenesis.53 Moreover, although immunoglobulins and complement are infrequently observed in biopsy specimens, the alternative pathway is now considered pathogenically important, and complement inhibitors (such as anti-C5a) may have efficacy in AAV.5455
Sources and selection criteria
We identified relevant studies for inclusion by a search of PubMed, Embase, and Cochrane databases from 2009 to 2019. We searched for the following phrases: “ANCA-associated vasculitis”, “Wegener’s granulomatosis”, “granulomatosis with polyangiitis”, and/or “microscopic polyangiitis”. We reviewed approximately 500 abstracts of systematic reviews, meta-analyses, randomized controlled trials (RCTs), and observational studies that were published in English. We also reviewed published management guidelines after reviewing websites and publications from professional rheumatology (for example, American College of Rheumatology, British Society of Rheumatology), nephrology, and other specialty societies. In addition to studies identified using these approaches, we also included landmark studies. With the exception of high profile RCTs recently reported in abstract form, only peer reviewed studies were eligible for inclusion.2556 We prioritized meta-analyses and RCTs with the exception of the management of specific manifestations (for example, sinus disease) for which data are more limited to case series. We generally excluded studies that enrolled patients with other forms of vasculitis (for example, polyarteritis nodosa) unless results from stratified analyses were available.
Evolution of clinical trial design in ANCA associated vasculitis
Case series from the US National Institutes of Health shaped the initial AAV treatment paradigm. Subsequently, recognition of the importance of multicenter studies in AAV and related conditions led to the formation of the Glomerular Disease Collaborative Network (GDCN), the French Vasculitis Study Group (FVSG), the European Vasculitis Study Group (EUVAS), and the Vasculitis Clinical Research Consortium (VCRC).57 The success of clinical trials organized by these groups confirmed that large, well designed studies in AAV were feasible and transformed AAV treatment. However, review of this experience reveals challenges to designing AAV trials that affect our ability to compare results across studies. Firstly, although recent efforts have been made to standardize trial design, trials historically relied on variable enrollment criteria, primary outcome measures, and definitions of remission and relapse.58 Secondly, the doses and durations of glucocorticoid regimens have varied widely.42 Considering these limitations is important as we review AAV clinical trials from the past 20 years.
Induction of remission
Glucocorticoids in AAV remission induction
Glucocorticoids bind to cytosolic glucocorticoid receptors, causing decreased expression of pro-inflammatory proteins and rapid non-genomic effects.59 Their quick onset of action and robust anti-inflammatory effects have made them a cornerstone of AAV care.60 However, significant toxicity (table 3) and incomplete efficacy as monotherapy necessitate a second immunosuppressive agent.61
The optimal dose, route, and duration of glucocorticoids remains uncertain. Intravenous methylprednisolone, typically in doses of 1000 mg daily for three days, is often used for severe end organ involvement, although the evidence base supporting this practice is weak.62 Ongoing uncertainty about the role of intravenous glucocorticoids is shown by the design of clinical trials for severe AAV, with some studies requiring their use and others using oral glucocorticoids or leaving the decision to the discretion of the investigator.13151617
Similar variation exists in the dose and duration of glucocorticoid therapy, with trials using courses ranging from 5.5 months to more than 24 months.42 Oral glucocorticoids are typically started at prednisone equivalent 1 mg/kg/day and tapered to 30-40 mg/day by one month and 10-20 mg/day by three months.4463
Only recently have studies assessed the optimal duration of oral glucocorticoids. The PEXIVAS trial, which enrolled 704 patients, used a two-by-two factorial design comparing plasma exchange versus no plasma exchange and standard dose versus reduced dose glucocorticoids in a non-blinded manner. The primary outcome was death or ESRD. The reduced glucocorticoid dose arm used 55% of the standard dose regimen over the first six months of therapy,64 achieving the same efficacy as the standard dose arm with fewer serious infections.56 Additional studies are needed to optimize the use of glucocorticoids to balance their efficacy and toxicity. This has been facilitated by the development of a novel tool, the glucocorticoid toxicity index, now being used in clinical trials.65
Refining cyclophosphamide to minimize toxicity
Prospective case series describing the efficacy of oral daily cyclophosphamide in combination with prednisone for GPA heralded the modern era of AAV treatment.454666 Cyclophosphamide is an alkylating agent that causes broad immunosuppression, including B cell depletion.67686970 Despite good response rates, the risk of relapse and toxicity (table 3) became evident with growing experience, leading to multiple studies comparing oral daily with intravenous pulse cyclophosphamide.11467172
Of these, the CYCLOPS trial was the first remission induction trial to use the contemporary regimen in which azathioprine is substituted for cyclophosphamide after three to six months of therapy (see section on remission maintenance).19 CYCLOPS was an unblinded controlled trial that randomized 149 patients with renal disease due to ANCA positive (99%) GPA or microscopic polyangiitis to daily or pulse cyclophosphamide (table 1).15 The primary outcome, time to remission over 18 months, was no different between the daily and pulse groups (hazard ratio 1.1, 95% confidence interval 0.8 to 1.6). The median time to remission was three months in both groups. The daily group received double the amount of cyclophosphamide as the pulse group (16 g v 8 g; P=0.001). No difference was seen in the proportion achieving remission at nine months (88% in both groups).
In a retrospective follow-up study of CYCLOPS, the daily group had a 50% reduction in the risk of relapse (hazard ratio 0.5, 0.3 to 0.9) over a median of 4.3 years compared with the pulse group, independent of ANCA type.73 Leukopenia was more common in the daily group than the pulse group (45% v 26%; P=0.02). In both the 18 month analysis and a long term follow-up study, no difference was seen in risk of malignancy, severe infection, death, or change in renal function between the two treatment arms.
Rituximab versus cyclophosphamide
For patients with complications or for whom cyclophosphamide treatment was not successful, an alternative was needed. Case series reported the efficacy of peripheral CD20+ B cell depletion,747576 motivating the design of the Rituximab in ANCA-Associated Vasculitis (RAVE) trial,16 which transformed the management of AAV (table 1).
RAVE randomized 197 patients with newly diagnosed or relapsing severe ANCA positive GPA or microscopic polyangiitis to a 5.5 month glucocorticoid taper plus either rituximab (four weekly 375 mg/m2 doses) or oral daily cyclophosphamide for three to six months followed by azathioprine. Patients in the rituximab arm received no prescribed treatment beyond glucocorticoids and the first four doses of rituximab. In contrast to other AAV trials, this was a double blind, double dummy, non-inferiority trial which had a primary outcome that required patients to be off glucocorticoids by six months. Most patients were PR3-ANCA positive (68%) and classified as GPA (77%); 48% had no major renal involvement. RAVE excluded patients with diffuse alveolar hemorrhage requiring ventilatory support and/or renal failure with a creatinine concentration greater than 4.0 mg/dL (354 μmol/L).
The primary outcome, complete remission off glucocorticoids at six months, was reached by 64% of patients randomized to rituximab compared with 53% of those randomized to cyclophosphamide (P<0.001 for non-inferiority). Rituximab was not statistically superior to cyclophosphamide (P=0.09) overall but was found to be superior among patients with relapsing disease at baseline (odds ratio 1.40, 95% confidence interval 1.03 to 1.91) and, in a post hoc analysis, among those who were PR3-ANCA positive (odds ratio 2.11, 1.04 to 4.30).7 Rituximab without additional remission maintenance therapy remained non-inferior to cyclophosphamide-azathioprine at 18 months when rates of sustained completion remission were compared (39% v 33%).77 Over 18 months, 74% of patients achieved complete remission off glucocorticoids at any time.
No difference was seen between rituximab and cyclophosphamide with respect to the frequency, severity, or time to relapse over 18 months. Outcomes were similar among patients with major renal involvement at baseline regardless of whether they were randomized to cyclophosphamide or rituximab.1678 No significant difference was seen in the rates of adverse events between the two treatment arms, although cyclophosphamide was associated with more pneumonias (P=0.03).16 This equivalency was surprising but likely reflects the countering effects of glucocorticoid exposure and reduced cyclophosphamide exposure on risk of infection.
A companion EUVAS trial, RITUXVAS, randomized 44 patients (3:1) with severe AAV associated renal vasculitis to rituximab plus two intravenous cyclophosphamide infusions or intravenous cyclophosphamide for three to six months followed by azathioprine.17 Both groups achieved sustained remission at similar rates (76% v 82%; P=0.7) and had similar rates of severe adverse events (42% v 36%; P=0.8).
Given the efficacy and comparable safety profile of rituximab, rituximab is often chosen for induction of remission.79 The optimal dosing regimen has not been studied, but the most common regimens include 375 mg/m2 weekly for four weeks or 1000 mg twice over two weeks. Cyclophosphamide remains a cornerstone of therapy for many providers,80 especially where rituximab is prohibitively expensive.81 The relatively short follow-up in RAVE limited an assessment of differences in long term outcomes, such as malignancy and survival, between the two treatment arms, but recent observational data suggest that rituximab, in contrast to cyclophosphamide, is not associated with an increased risk of malignancy.32
Other approaches to remission induction: methotrexate and mycophenolate mofetil
Methotrexate and mycophenolate mofetil have also been studied for induction of remission. Methotrexate has its effects by increasing extracellular adenosine, whereas mycophenolate mofetil inhibits DNA synthesis.82 The NORAM trial was designed following studies suggesting that methotrexate may have efficacy in non-severe AAV.83848586 NORAM was an open label, non-inferiority EUVAS trial that randomized 100 patients with non-severe GPA or microscopic polyangiitis to either methotrexate or daily cyclophosphamide for 12 months (table 1).12 Patients with organ threatening or life threatening disease were excluded.
Most patients in NORAM were PR3-ANCA positive (74%) and had GPA (94%). The primary outcome, the proportion in remission at six months, was achieved by a similar proportion of patients in the methotrexate and cyclophosphamide arms (90% v 94%), and the median time to remission was also similar (3 v 2 months; P=0.3). However, methotrexate was associated with a longer time to remission among patients with more extensive disease and a shorter time to relapse among all patients than was cyclophosphamide (hazard ratio 1.9, 95% confidence interval 1.1 to 3.3). Over 18 months, methotrexate was associated with more flares than cyclophosphamide (70% v 47%). Compared with cyclophosphamide, methotrexate was associated with fewer episodes of leukopenia (3 v 14; P=0.01) but more episodes of liver dysfunction (7 v 1: P=0.04).
NORAM has two significant limitations. Firstly, it used 12 months of oral cyclophosphamide, so how methotrexate would compare with contemporary three to six month cyclophosphamide regimens is unclear. Secondly, all therapy was tapered and discontinued by 12 months, after which many of the flares occurred. Prolonged methotrexate treatment may have prevented flares and been well tolerated. Methotrexate is commonly used in rheumatology because of its favorable safety profile, as confirmed in a recent large trial evaluating its efficacy for secondary prevention of cardiovascular disease,87 but should be avoided in patients with a glomerular filtration rate below 30 mL/min/1.73 m2.
MYCYC was an open label, non-inferiority, randomized controlled EUVAS trial that compared mycophenolate mofetil with cyclophosphamide for induction of remission (table 1).18 It followed several small prospective studies that found mycophenolate mofetil to be efficacious in MPO-ANCA positive AAV.888990 In MYCYC, 140 patients newly diagnosed as having GPA or microscopic polyangiitis (59% PR3-ANCA positive, 38% MPO-ANCA positive) were randomized to either mycophenolate mofetil (76% on 2 g/day of mycophenolate) or pulse cyclophosphamide; both arms received azathioprine following remission. Patients with life threatening disease and those with rapidly declining renal function were excluded, but most (81%) had renal disease and the severity of renal disease was generally worse than in those enrolled in NORAM.
A similar proportion of patients in the mycophenolate mofetil and cyclophosphamide groups achieved remission at six months (67% v 61%), showing non-inferiority (risk difference 5.7%, 90% confidence interval −7.5% to 19%). However, more patients in the mycophenolate mofetil group relapsed after remission (33% v 19%; incidence rate ratio 1.97, 95% confidence interval 0.96 to 4.23). This difference was strongly driven by differences in response according to ANCA type; 48% of PR3-ANCA positive patients in the mycophenolate mofetil group compared with 24% of PR3-ANCA positive patients in the cyclophosphamide group relapsed. Notably, the rates of serious infection were similar between the mycophenolate mofetil and cyclophosphamide arms (26% v 17%; odds ratio 1.7, 0.68 to 4.19), as were the rates of other adverse events (for example, ESRD and death).
NORAM and MYCYC show that methotrexate and mycophenolate mofetil can induce remission in selected patients but may be associated with a higher risk of relapse. They may be particularly useful for patients in whom conventional therapies have failed or are contraindicated and/or in those at lower risk of flare (such as MPO-ANCA). They may not be ideal for patients in whom a relapse could portend organ failure, significant effect on quality of life, serious toxicity from excess glucocorticoids, or life threatening disease. Of note, mycophenolate mofetil received the lowest grade of endorsement by members of EUVAS in their recommendations for the management of AAV, although these were published before the publication of MYCYC.44
RAVE, NORAM, and MYCYC highlight the relative safety from an infection perspective of contemporary cyclophosphamide regimens. However, these observations are derived from clinical trials in which patient selection and close monitoring may influence infection rates. Observational data may better assess these risks and be used to compare the relative efficacy of different regimens (for example, methotrexate, mycophenolate mofetil) versus rituximab in patients with less severe disease.
Plasma exchange in ANCA associated vasculitis
In cases of diffuse alveolar hemorrhage, rapidly progressive glomerulonephritis, or both, plasma exchange is sometimes used in conjunction with other therapies. Until recently, MEPEX was the only large trial evaluating the efficacy of plasma exchange in AAV patients with biopsy proven glomerulonephritis associated with a creatinine concentration above 5.8 mg/dL (table 1).14 MEPEX was a EUVAS trial that randomized 137 patients (43% PR3-ANCA positive, 52% MPO-ANCA positive) to oral cyclophosphamide and a glucocorticoid taper with either seven plasma exchanges or three days of pulse methylprednisolone. Patients randomized to plasma exchange were more likely to be alive and free of dialysis and to have a creatinine concentration below 5.8 mg/dL at three months than were those randomized to pulse methylprednisolone (69% v 49%; P=0.02). However, four years after randomization, a post hoc analysis found that differences in ESRD and death were not sustained.91 A subsequent meta-analysis concluded that insufficient evidence was available to assess the efficacy of plasma exchange for preventing ESRD and/or death in AAV.92
To assess the ongoing uncertainty about the efficacy of plasma exchange, PEXIVAS randomized AAV patients with glomerulonephritis (with a glomerular filtration rate <50 mL/min/1.73 m2), diffuse alveolar hemorrhage, or both to standard therapy (rituximab or cyclophosphamide and glucocorticoids) plus plasma exchange or standard therapy only.5664 PEXIVAS enrolled 704 patients (41% PR3-ANCA positive, 59% MPO-ANCA positive), 98% of whom had renal disease. In contrast to MEPEX, both arms received pulse methylprednisolone. No difference was seen in the rate of ESRD or death from any cause (co-primary endpoints) between patients randomized to plasma exchange and those randomized to standard therapy (28% v 31%; hazard ratio 0.86, 0.65 to 1.13) at seven years. No difference was seen when the outcomes of ESRD and death were analyzed separately.
Despite differences in inclusion criteria and treatment, both PEXIVAS and MEPEX suggest that plasma exchange does not provide a long term benefit over standard of care in AAV. Post hoc analyses of PEXIVAS and other trials are needed to clarify the potential utility of plasma exchange in subgroups of patients. Patients who are positive for both ANCA and anti-glomerular basement membrane should continue to receive plasma exchange.939495
Summary of remission induction
Induction of remission has evolved considerably over the past few decades (table 1). Today, high dose glucocorticoids plus either rituximab or cyclophosphamide are used more often. Case series using combination rituximab and cyclophosphamide to spare glucocorticoid exposure have been published, but the evidence base supporting this practice remains weak.9697 The main limitation to use of rituximab is its cost,98 although biosimilar rituximab may cost less.99 Toxicity from cyclophosphamide may be minimized by reducing the dose in patients with renal impairment and with advancing age, as well as by limiting duration of treatment. Plasma exchange does not seem to add efficacy for severe disease. In cases of non-severe disease, methotrexate or mycophenolate mofetil may be options, but neither has been compared with rituximab.
Additional studies comparing long term outcomes between treatments, stratified by ANCA type and other disease features, are warranted in addition to cost effectiveness analyses. The cost effectiveness of regimens for induction of remission needs to be evaluated from multiple societal perspectives, considering the impact of biosimilar rituximab costs and incorporating long term observational outcome data as they become available.100
Maintenance of remission
The vast majority of patients with AAV will achieve remission with contemporary regimens, but approximately a third of patients will relapse by 18 months and less than a third will remain in relapse-free remission for more than a decade.77101 A personalized approach to maintenance of remission based on patient specific and disease specific factors would balance the benefits of disease quiescence with the cost and morbidity of prolonged immunosuppression. This is particularly important given that most deaths occurring more than a year after the diagnosis of AAV are due to infection, malignancy, and cardiovascular disease rather than active vasculitis.34 The past 20 years have greatly advanced the approach to maintenance of remission, with several effective agents and treatment strategies now in use.
Cyclophosphamide, azathioprine, methotrexate, and mycophenolate mofetil have been investigated in randomized trials of maintenance therapy (table 2). All studies induced remission with prednisone and cyclophosphamide, but cyclophosphamide dosing strategies (daily v pulse, duration) varied between trials.
CYCAZAREM was an unblinded, controlled EUVAS trial that randomized 144 patients with newly diagnosed severe AAV (57% PR3-ANCA positive, 39% MPO-ANCA positive) to azathioprine or continued cyclophosphamide for maintenance of remission (table 2). Azathioprine is thought to have efficacy because of its effect on DNA synthesis. Over 18 months, the relapse rate (severe and non-severe) was similar between azathioprine and cyclophosphamide (15.5% v 13.7%; P=0.65).19 CYCAZAREM changed the treatment paradigm in AAV, limiting cyclophosphamide treatment to three to six months to avoid serious morbidity associated with prolonged cyclophosphamide exposure (table 3).102103104
Azathioprine was compared with methotrexate in WEGENT (table 2), a randomized, controlled, unblinded FVSG trial of 126 patients with newly diagnosed systemic AAV (60% PR3-ANCA positive, 31% MPO-ANCA positive).20 Methotrexate was hypothesized to have fewer adverse events than azathioprine, but both were found to have similar rates of adverse events (19% v 11%; P=0.21). The surprisingly high adverse event rate in the methotrexate arm may have resulted from a lack of dose adjustment for renal impairment. At 24 months, methotrexate and azathioprine had similar rates of severe and non-severe relapse-free survival (hazard ratio 0.92, 0.52 to 1.65), but the trial was not powered to assess this outcome.20
Azathioprine was found to be superior to mycophenolate mofetil for maintenance of remission in the IMPROVE trial conducted by EUVAS (table 2).21 This was an open label RCT that enrolled 156 patients with newly diagnosed AAV (58% PR3-ANCA positive, 33% MPO-ANCA positive). Over a median of 39 months, participants randomized to mycophenolate mofetil relapsed (severe and non-severe) more often than did those randomized to azathioprine (hazard ratio 1.80, 1.10 to 2.93).21
Given these observations, both azathioprine and methotrexate are considered effective oral agents for maintenance of remission in AAV.44 Mycophenolate mofetil remains a potential alternative agent, particularly given its efficacy in MYCYC, but may be inferior to azathioprine.18 Notable aspects of this evidence base are that most patients had GPA, were PR3-ANCA positive, and had severe disease.
Rituximab’s efficacy in inducing remission,16 combined with the experience of long term B cell depletion in rheumatoid arthritis,105 led to several reports of its use for AAV remission maintenance.106107108 Subsequently, two RCTs compared rituximab with azathioprine for maintenance of remission.
MAINRITSAN 1 (table 2), an unblinded, randomized, controlled FVSG trial of 115 newly diagnosed or relapsing patients with AAV (70% PR3-ANCA positive, 23% MPO-ANCA positive), compared rituximab (500 mg twice over 14 days followed by infusions every six months until month 18) with azathioprine (for 22 months). All patients were treated with pulse cyclophosphamide for induction of remission. Azathioprine was associated with more severe relapses than rituximab (29% v 5%; hazard ratio 6.61, 1.56 to 27.96).22
At five years, 42 months after the last rituximab infusion, rituximab remained superior to azathioprine but relapse-free survival was reduced in both groups (58% v 37%; P=0.01).109 Although few deaths occurred, rituximab was associated with better survival at five years (100% v 93%; P=0.045).109 Moreover, a cost effectiveness analysis from the French societal perspective found rituximab to be cost effective compared with azathioprine.110 A notable criticism of MAINRITSAN 1 is that the dose of azathioprine was decreased from 2 mg/kg to 1.5 mg/kg at 12 months and to 1 mg/kg at 18 months, whereas the rituximab exposure remained unchanged.
The RITAZAREM study randomized 170 patients (72% PR3-ANCA positive, 28% MPO-ANCA positive) with relapsing disease to rituximab 1 g every four months or azathioprine for maintenance of remission after induction with rituximab (table 2). Preliminary data presented in abstract form showed that rituximab was superior to azathioprine for preventing disease relapse (13% v 38%; hazard ratio 0.30, 0.15 to 0.60; P<0.001), further supporting the efficacy of rituximab for maintaining remission.25
Timing of rituximab
MAINRITSAN 1 and RITAZAREM used a fixed interval of rituximab retreatment. However, observations that relapses are less frequent in patients with undetectable peripheral B cells and that rising ANCA titers may predict a flare led to MAINRITSAN 2 (table 2).77111112 This open label RCT conducted by the FVSG compared fixed interval rituximab (as used in MAINRITSAN 1) with retreatment only when peripheral B cells became detectable or a significant rise in the ANCA titer occurred (that is, tailored therapy).113 MAINRITSAN 2 enrolled 162 patients with newly diagnosed or relapsing AAV (47% PR3-ANCA positive, 31% MPO-ANCA positive). Fixed interval and tailored approaches were equivalent with regard to the risk of severe and non-severe relapse over 28 months (10% v 17%; P=0.22). Patients in the tailored arm received significantly fewer rituximab infusions than those in the fixed interval arm (median 3 (interquartile range 2-4) versus 5 (5-5)). Safety was similar in the fixed interval and tailored arms (85% v 91% severe events; P=0.5), but numerically fewer patients had infections in the tailored arm (9 v 16). Although the optimal timing and dose of rituximab retreatment remain uncertain, personalization of retreatment regimens holds promise.
Duration of treatment
The optimal duration of AAV maintenance therapy remains unknown.114 Although some patients can remain in drug-free remission, the prevalence of relapse increases steadily over time.101109 REMAIN, an open label RCT that enrolled 117 patients with severe AAV (52% PR3-ANCA positive, 44% MPO-ANCA positive), found that discontinuing azathioprine and prednisone after 24 months of treatment was associated with a higher risk of non-severe and severe relapse (63% v 22%; odds ratio 6.0, 2.6 to 13.8) and ESRD (7.8% v 0%; P=0.01) compared with continued treatment for 48 months (table 2).23 Similarly, a meta-analysis of 13 studies which included 983 patients with diverse AAV manifestations found that prolonged treatment with glucocorticoids was associated with fewer relapses than shorter durations (hazard ratio 0.4, 0.3 to 0.5).115
Being able to stratify patients according to risk of relapse is necessary to eventually personalize care in a way that optimally balances the risk of relapse and potential toxicity of various regimens. Several risk factors for relapse have been identified (table 4). Multiple studies have shown that PR3-ANCA positive patients have an approximately twofold higher risk of relapse than MPO-ANCA positive patients.16101 More severe renal damage has been associated with a reduced risk of relapse.122129130 Potential serum and urine biomarkers of disease activity have been identified but need further study before clinical use.123127 Risk stratification and development of biomarkers able to predict relapse are critical to personalizing the approach to maintenance therapy.
Treatment challenges associated with specific manifestations
Sinus involvement, tracheobronchial disease, orbital pseudotumor, and interstitial lung disease are particularly challenging manifestations of AAV because they can be the sole manifestation, can be hard to diagnose and treat, and can cause significant, irreversible damage. Studies of these manifestations are limited to case series, providing a limited evidence base for recommendations.
Sinusitis is common in GPA.3645 Evaluation by an experienced otolaryngologist is useful in patients with sinusitis and suspected GPA, because up to 40% of patients with disease limited to the sinuses and respiratory tract may be ANCA negative and an otolaryngologist can help to differentiate active sinus disease from damage.45 Nasal septal perforation and saddle nose deformity are well described manifestations of AAV damage, but bony erosion, persistent nasal congestion, crusting, and epistaxis may persist after appropriate treatment. Distinguishing between damage and active disease is important, as treating damage with immunosuppression may unnecessarily increase the risks of infection and other complications.
No studies have investigated the optimal treatment regimen for sinonasal manifestations, but all agents that have been investigated for induction of remission are thought to be efficacious. In addition, topical treatment with saline and corticosteroids, as recommended for general chronic sinusitis,131 can be helpful.
Subglottic stenosis, tracheobronchitis, and bronchial masses are uncommon manifestations of GPA but cause substantial morbidity.132133134 Tracheobronchial manifestations are often decoupled from disease activity elsewhere and can occur during immunosuppressive therapy, suggesting that these manifestations may not respond to immunomodulation.132133134 In general, bronchial masses are treated with immunosuppression, but the management of subglottic stenosis is controversial. One case series of 21 patients observed that a series of intralesional glucocorticoid and dilations over a mean follow-up of 41 months did not result in any tracheostomies, a marked improvement from previous reports.135 Other case series of up to 43 patients, in contrast, have reported variable success of immunosuppressives (methotrexate, cyclophosphamide, rituximab) in combination with prednisone.132133136
Orbital inflammatory disease
Orbital inflammatory disease is challenging to treat because it often does not fully respond to immunosuppression and relapses are common.137138 One of the largest case series of 59 patients with orbital masses suggested improved outcomes with rituximab compared with cyclophosphamide (91% v 52% response).138 However, the study’s small sample size and retrospective nature limit the generalizability of these findings.
Interstitial lung disease
Interstitial lung disease (ILD) is increasingly observed in association with MPO-ANCA antibodies with or without other manifestations of AAV.139140141 Whereas treatment of ILD in patients with other AAV manifestations follows the typical approaches described earlier, the management of ILD and a positive ANCA test without other AAV manifestations is uncertain. In a small case series that included 36 patients, approximately 20-30% of patients with ANCA positivity and ILD went on to develop AAV, especially those who were MPO-ANCA positive.142 A case series of 12 patients with usual interstitial pneumonia associated with ANCA positivity reported improvement in ILD with immunosuppression,143 but other series (n<20) did not report a benefit.144145 Additional studies of this unique AAV subgroup are needed. The role of nintedanib in fibrotic lung disease associated with ANCA is of interest, given its efficacy in idiopathic pulmonary fibrosis and scleroderma lung disease.146
Supplementary therapies in AAV
Trimethoprim-sulfamethoxazole is an important treatment adjunct for preventing Pneumocystis jirovecii pneumonia (PCP) in patients with AAV (table 3). Although no randomized studies exist to guide PCP prophylaxis in patients with rheumatologic disease, the American Thoracic Society suggests treatment in patients receiving 20 mg or more of prednisone for four weeks or longer, especially if combined with another immunosuppressive agent.147 Uncertainty exists as to whether trimethoprim-sulfamethoxazole has a role in preventing relapse of GPA, based on a randomized study in 81 patients that showed benefit in those with upper airway involvement.148 However, its efficacy is not thought to be sufficient for it to be used as monotherapy.44
In addition to calcium and vitamin D, postmenopausal women and men aged over 50 who are receiving prednisone equivalent of at least 7.5 mg for three months or more or are at high risk for major osteoporotic fracture (based on FRAX tool) should receive treatment with a bisphosphonate or another agent.149 Finally, reviewing and completing appropriate vaccinations is of particular importance in patients with AAV.
Organ transplantation in AAV
Rapidly progressive glomerulonephritis and interstitial lung disease are two manifestations of AAV that can cause irreversible organ failure. Renal transplantation is safe and effective in appropriately selected patients with ESRD due to AAV.150151152 Among AAV patients with ESRD who are waitlisted for a transplant, renal transplantation is associated with a 70% reduction in the risk of all cause death (relative risk 0.30, 95% confidence interval 0.25 to 0.37), largely driven by a reduction in the risk of death due to cardiovascular disease (0.10, 0.06 to 0.16).153 Although recommendations vary, patients with a low glomerular filtration rate (<20 mL/min/1.73 m2) without expected recovery should be evaluated for transplantation candidacy, ideally before dialysis.154 AAV should be in remission for at least 12 months and the ANCA titer should be negative before transplantation.155156 In contrast to renal disease, outcomes in lung transplantation for AAV are poorly described.157
Quality of life in AAV
AAV is associated with depression and anxiety, as well as a worse quality of life compared with the general population.158159160161 These differences often persist in disease remission and highlight the challenges of managing both the physical and psychosocial aspects of living with AAV. When reported, quality of life outcomes in AAV trials have not differed between treatment arms. Differences in quality of life observed in MAINRITSAN 1 are difficult to interpret because completion of the survey was optional and a significant amount of data is missing.162 The recent development of an AAV specific patient reported outcome measure will enable the assessment of AAV specific quality of life (for example, AAV symptoms, side effects, emotional impact, uncertainty) in future studies.163 No studies have evaluated psychosocial interventions in AAV.
The development of novel agents and optimization of the use of existing therapeutics are rapidly evolving research areas in AAV. Given the recognition of the importance of the complement pathway in the pathogenesis of AAV, several agents targeting the complement activation product C5a are being studied, including a phase III trial of avacopan, an oral C5a receptor inhibitor, for induction of remission. Two phase II studies, CLEAR and CLASSIC, showed the potential efficacy of avacopan in inducing remission in combination with rituximab or cyclophosphamide.55164 The CLEAR study suggested that avacopan may have a dramatic steroid sparing effect, which would be a tremendous advance.55 IFX-1, a monoclonal antibody targeting C5a, is being studied in phase II trials.165 In addition to complement pathway targets, a phase III trial is studying abatacept, a cytotoxic T lymphocyte associated protein 4 antagonist, for non-severe GPA.166167
The optimal dose and duration of glucocorticoid treatment remains an area of active investigation in both induction and maintenance of remission.125167168169170 For example, a single center study of 49 patients found that the combination of rituximab and cyclophosphamide with two weeks or less of glucocorticoids led to 96% of patients achieving remission at six months.125
A growing evidence base suggests that personalized approaches to treatment of AAV may be possible. Identifying optimal strategies that balance the adverse events and cost of maintenance therapy against morbidity associated with active disease and treatment is a critical area of future investigation. In particular, remission induction and maintenance studies stratified by ANCA type can further elucidate whether PR3-ANCA positive and MPO-ANCA positive patients respond differently to treatment. Additional studies of non-severe AAV are essential to assess the efficacy of less toxic treatment options.
The European League Against Rheumatism (EULAR) and the British Society for Rheumatology (BSR) have published guidelines on the management of AAV.4481 Notably, both were written before the publication of several recent pivotal trials, so their recommendations should be interpreted cautiously. General agreement exists between the EULAR and BSR guidelines. Both reflect the growing therapeutic options that clinicians have for induction and maintenance of remission in AAV. Both guidelines advocate for prolonged glucocorticoid exposure, which remains controversial. The EULAR and BSR guidelines emphasize the need to routinely assess patients for disease related and treatment related toxicities, including cardiovascular disease, diabetes, and hypogammaglobulinemia, which could not be covered in this review.
Management of AAV has evolved considerably in recent decades, a period characterized by substantial improvements in outcomes, including renal outcomes and overall survival. As our understanding of the pathogenesis of AAV expands, so too does the pipeline of therapeutic options for AAV, especially those with the potential to substantially reduce glucocorticoid exposure. Additional studies are needed to define personalized approaches to the management of AAV.
Should induction and maintenance of remission be managed differently for patients with ANCA associated vasculitis (AAV) who are PR3-ANCA positive and those who are MPO-ANCA positive?
What are the risks and benefits of reducing the cumulative exposure to and/or duration of glucocorticoid therapy used during induction and maintenance of remission?
What is the comparative effectiveness of methotrexate, mycophenolate mofetil, and rituximab for non-severe AAV?
What interventions can effectively improve the quality of life of patients with AAV and meet their psychosocial needs?
Glossary of abbreviations
AAV—ANCA associated vasculitis
ANCA—anti-neutrophil cytoplasmic antibody
BSR—British Society for Rheumatology
EGPA—eosinophilic granulomatosis with polyangiitis
ESRD—end stage renal disease
EULAR—European League Against Rheumatism
EUVAS—European Vasculitis Study Group
FVSG—French Vasculitis Study Group
GDCN—Glomerular Disease Collaborative Network
GPA—granulomatosis with polyangiitis
ILD—interstitial lung disease
NET—neutrophil extracellular trap
PCP—Pneumocystis jirovecii pneumonia
RCT—randomized controlled trial
VCRC—Vasculitis Clinical Research Consortium
Three patients from our practice kindly volunteered to review this manuscript. In response to their feedback, minor edits were made to the manuscript. One was a 40 year old man with ANCA associated vasculitis (AAV) who has cutaneous vasculitis, arthritis, and pulmonary masses. He suggested highlighting the importance of early diagnosis and treatment recommendation and appreciated our proposal to study psychosocial interventions for AAV. The second patient was a 60 year old man with AAV who has sinusitis, pulmonary nodules, and neuropathy. He suggested highlighting the balance of costs and benefits of fixed rituximab maintenance therapy. The third patient was a 67 year old man with AAV causing scleritis, sialadenitis, hearing loss, pancreatitis, sinusitis, and lung nodules. He appreciated the value of this manuscript but did not recommend any specific changes.
Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors
Contributors: ZSW had the idea for the article. Both authors did the literature search and planned and wrote the article. ZSW accepts full responsibility for the finished article, had access to any data, and controlled the decision to publish; he is the guarantor.
Funding: ZSW receives funding from NIH/NIAMS (K23AR073334 and L30 AR070520).
Competing interests: We have read and understood the BMJ Group policy on declaration of interests and declare the follow interests: none.
Provenance and peer review: Commissioned; externally peer reviewed.