From the clinical point of view, WG is characterized by two features: mass formation and ANCA-associated small- to medium-size AAV. The former is usually restricted to the upper and lower respiratory tracts and its characteristic histology features constitute granulomatous and necrotizing inflammation, yet vasculitis may also be found within the granulomatous inflammation 2 . The latter may affect virtually any organ, with a predilection for lung and kidney involvement (necrotizing, predominantly small-vessel pauci-immune vasculitis, that is, manifesting as pulmonary capillaritis and crescentic glomerulonephritis) 3 4 . There is some evidence from in vitro and in vivo studies that ANCAs play a crucial role in the mediation of small-vessel vasculitis 5 6 7 . In WG, ANCAs are mainly directed against Proteinase 3 4 and are implicated in the activation of polymorphonuclear neutrophils by interacting with Proteinase 3 expressed on the plasma membrane of these 5 .

According to a concept by Fienberg 8 , WG starts as 'granulomatous' disease of the upper and lower respiratory tracts and subsequently progresses to generalized ('systemic') disease characterized by clinical vasculitis manifestations, such as pulmonary capillaritis or necrotizing glomerulonephritis. This concept has been incorporated in the European definition of disease stages of AAV introduced in 1995 9 and has been updated several times since then. In the current definitions 10 , the localized disease stage, defined as manifestations restricted to the respiratory tract (that is, rhinitis, sinusitis, pulmonary masses) with no clinical signs of vasculitis (no pulmonary capillaritis), is differentiated from the systemic disease stages (early systemic, generalized and severe disease), all of which are associated with clinical signs of vasculitis. In the systemic disease stages, patients usually present with pronounced constitutional symptoms (weakness, fever, weight loss). Whereas early systemic disease is considered non-life-threatening (that is, occurring with arthritis, episcleritis or purpura), generalized disease is defined as organ- or life-threatening (that is, due to glomerulonephritis or alveolar hemorrhage). In severe disease, organ failure has occurred due to systemic vasculitis manifestations (that is, renal failure; Table 1). The definition of these disease and activity stages 10 has facilitated the generation of evidence-based, stage-adapted treatment regimens. For most disease stages (except for localized disease), evidence from controlled trials is available to guide treatment decisions 11 .

WG usually starts with symptoms due to upper respiratory tract involvement: manifestations affecting the nasal and oral cavity, the sinuses and the trachea are reported to occur in 75 to 93% of patients at diagnosis and in up to 99% during the course of the disease 3 4 12 . Upper respiratory tract involvement not only occurs early in the disease course, but is the most frequent manifestation of WG 4 , and for these two reasons is indicative of WG: patients typically present with rhinosinusitis with bloody discharge, crusting and epistaxis. Nasal crusting ('golden crusts') is a typical finding of endoscopic evaluation.

Granulomatous inflammation and/or mass formation may be found in the nasal cavity, sinuses and within the orbits but also frequently in the lower respiratory tract as pulmonary 'granuloma' 3 4 (Figure 1). Orbital masses, either perforating from the sinuses or developing within the orbit in isolation, are a serious complication of upper respiratory tract disease in 15 to 20% of patients 4 . The nature of granulomatous inflammation is destructive, as a significant proportion of patients develop signs of cartilage and bone erosion during the course of disease (that is, saddle nose deformity or orbital wall destruction), as has recently been shown in a study of localized WG patients 13 . Trachea and bronchi may also be affected by ulcerative or granulomatous inflammation, which may result in subglottic or bronchial stenosis.

The overall incidence of pulmonary involvement is frequent (between 60% and 85%) and constitutes 'granulomatous' as well as 'vasculitic' manifestations 3 4 12 . Pulmonary nodules/masses ('granuloma') have been described on conventional radiography in around 60% of cases 3 . Diffuse alveolar hemorrhage due to vasculitis has been documented in 7 to 45% of patients 3 4 (Figure 2).

MPA may be regarded as 'pure' small- to medium-size vessel vasculitis that is associated with a positive ANCA status in more than 90% of cases, usually with perinuclear ANCAs (P-ANCAs) or myeloperoxidase-specific ANCAs (MPO-ANCAs). MPA has an incidence of three people per million per year and shows a slight male predominance 19 20 with an average onset between 50 and 60 years of age. In contrast to WG, patients with MPA do not present with a classic localized stage of the disease. Ear-nose-throat (ENT) symptoms have been described to occur 20 21 , but granulomatous tumor-like or destructive processes of the upper and lower respiratory tract are lacking. At disease onset, constitutional symptoms are common 19 21 . Patients may present initially with either a non-life-threatening course ('early systemic disease', that is, arthralgia, arthritis, episcleritis), which may progress to 'generalized' disease, or acute organ- or life-threatening disease ('generalized' or 'severe' disease stage, that is, necrotizing glomerulonephritis or DAH) within days or weeks. Similar to WG, there is a predilection for vasculitic manifestations of kidney (glomerulonephritis) and lung (alveolar capillaritis): glomerulonephritis has been reported to occur in 80 to 100% of patients 19 20 21 22 and is even more frequent than in WG; and alveolar hemorrhage due to capillaritis has been reported in 12 to 55% of patients 19 20 21 22 (Figure 2). In biopsy specimens, small- to medium-size necrotizing pauci-immune vasculitis is to be found with no evidence of granuloma formation.

DAH is the classic pulmonary manifestation in MPA and has been described to occur in 12 to 55% of patients 19 20 21 22 . MPA shows the highest frequency of DAH and/or glomerulonephritis among the AAVs, occurring either as isolated organ involvement or together as pulmo-renal syndrome. As in WG, it has a wide spectrum, ranging from asymptomatic disease to a severe life-threatening stage.

Pulmonary fibrosis has been recognized as a pulmonary manifestation of MPA (Figure 3). Only few cases and a larger retrospective case series have been reported in the literature, most of which demonstrated an association of pulmonary fibrosis with MPO-ANCA-positive AAV or MPA 23 24 25 , suggesting a pathogenic role of MPO or MPO antibodies in MPA-associated pulmonary fibrosis. Pulmonary fibrosis may develop after other clinical manifestations of MPA, but has also been reported to occur before disease onset 24 , and it may have a usual interstitial pneumonia (UIP) or noninterstitial pneumonia (NSIP) pattern on HRCT and is associated with lower total lung capacity and increased mortality 26 .

CSS is characterized by peripheral blood hypereosinophilia, tissue eosinophilia, asthma and vasculitis 1 2 . The association with ANCAs is less strong; ANCAs can be detected in around 40% of individuals with CSS, in most cases P-ANCAs or MPO-ANCAs 27 . There is a substantial clinical overlap with the hypereosinophilic syndrome (HES). In fact, in some cases it might be difficult to strictly differentiate between CSS and HES. According to recent classification proposals, CSS may be seen as a subtype of HES 28 . CSS usually develops in three distinct phases: the first phase is indistinguishable from poorly controlled asthma and may last for several years; the second phase is characterized by profound blood eosinophilia; and finally, with additional manifestations of small vessel vasculitis or histological evidence of vasculitis, the diagnosis of CSS can be established 29 . In general CSS may involve any organ system, although the lungs (90 to 100%), the peripheral nervous system (approximately 70%), the skin (50 to 70%) and the heart (approximately 50%) are predominantly affected 30 31 . According to the literature, ENT involvement is seen in about 50% of patients; in our own experience, sinusitis, polyposis or rhinitis are found in more than 90% if examination by an ENT specialist is part of the work-up (unpublished observation). Constitutional symptoms are common.

ANCAs are correlated with distinct clinical manifestations: in ANCA-positive CSS, classical vasculitis manifestations, such as glomerulonephritis, pulmonary vasculitis with DAH or peripheral neuropathy, are more frequent, whereas heart-involvement is seen more often in ANCA-negative CSS 29 . Evidence from genetic studies further supports the view of at least two distinct disease subtypes 32 .

In the absence of surrogate parameters for vasculitis, the evaluation of biopsy specimens is crucial to differentiate between HES and CSS since the latter requires histological evidence of necrotizing vasculitis as long as no clinical surrogate such as glomerulonephritis is present. In the original publication by Churg and Strauss 33 three characteristic features were described: tissue eosinophilia, necrotizing vasculitis and granuloma formation. Tissue infiltration by eosinophils is very common and believed to represent a major pathogenic mechanism; however, it is not specific for CSS and cannot aid delimitation towards HES.

Definitions for disease stages and activity should be used as suggested for AAV (Table 1), although the evidence for those definitions was derived mainly from studies in WG and MPA. The Five Factor Score (FFS) for CSS was found to predict mortality and should be used as a prognostic tool 34 35 .

Respiratory tract involvement is very common in CSS. The vast majority of patients (>90%) suffer from bronchial asthma. In spite of the high frequency of asthma in CSS patients, clinical features of asthmatic manifestations and subtypes of asthma have not yet been described in detail. There are no reliable markers to predict the development of CSS in asthmatics. In the majority of patients asthma is characterized by adult onset and a relatively low prevalence of inhalational allergies. Frequently, a tendency to more severe or 'difficult to treat' asthma is observed. In more than 80% of cases asthma precedes the onset of vasculitis by a median time of 4 years 36 . Lung function testing often shows persistent airflow obstruction in CSS patients with asthma 37 . Some reports link the use of a leukotriene receptor antagonist to CSS, although many experts believe that leukotriene receptor antagonists simply unmask CSS by allowing for reduction of systemic steroids. In a recent study, however, the association between leukotriene receptor antagonist and CSS could not be explained by steroid withdrawal in the majority of cases 38 .

Reported frequencies of nasal polyposis, sinusitis and rhinitis vary and, according to our experience, are underestimated as long as no routine examination by an ENT specialist and/or cranial MRI is included in the work-up procedure. Upper respiratory tract manifestations in CSS can be distinguished from WG due to the lack of mass formation and destruction. Granulomatous lesions are a feature in biopsy specimens, yet the clinical or radio-logical appearance of granulomatous mass formation is usually not seen in CSS.

Pleural effusions can be seen in 10% of cases and might either be a sign of congestive heart failure due to heart involvement or a sequel of eosinophilic pleuritis.

Eosinophilic infiltration of the lung is common and may be detected as a patchy and migratory infiltrate on plain X-ray (Figure 4). Bronchoalveolar lavage is performed in order to prove eosinophilic alveolitis, and tissue infiltration may be seen in trans-bronchial lung biopsies. Other forms of alveolitis and mixed patterns (neutrophilic and lymphocytic alveolitis) also occur 39 . DAH as the clinical equivalent of lung capillaritis can occur but is less frequent than in WG and MPA 30 (Figure 2).

AAV patients should undergo a standardized multidisciplinary evaluation in order to determine disease stage and activity. It is recommended to treat AAV patients at or in cooperation with specialist centers 11 . All AAV patients require a work-up that includes examination of lung (X-ray, lung function testing and HRCT, and fiberoptic bronchoscopy with bronchoalveolar lavage if necessary) and kidneys (ultrasound, assessment of creatinine clearance and microscopic hematuria, red cell casts and proteinuria) and testing for ANCAs. In addition, WG patients should be seen by an otorhinolaryngologist and CSS patients require assessment of pulmonary function and allergy testing. At first presentation, confirmation of diagnosis should be sought by biopsy - for example, nasal mucosa and lung or kidney biopsy. Depending on disease symptoms and suspected organ manifestations, additional evaluations by a neurologist, ophthalmologist and dermatologist and/or additional technical examinations, such as MRI of sinuses and orbits, may be required.

Patients should be seen and evaluated in regular intervals (three- to six-monthly) to assess disease activity and adapt immunosuppressive therapy. Ideally, patients are evaluated by standardized and validated tools to assess disease activity and damage (such as the Birming-ham Vasculitis Activity Score or the Vasculitis Damage Index) 40 41 .

Treatment follows a stage- and disease activity-adapted regimen due to evidence from controlled trials 11 . For remission induction in WG and MPA, methorexate (MTX) should be used in non-organ-threatening early systemic disease and cyclophosphamide in generalized disease in addition to glucocorticoids 42 43 . Rituximab may be an alternative for cyclophosphamide in the future 44 45 , even if not yet listed in the treatment recommendations as data have been published only very recently. For severe disease, namely acute renal failure, additional plasma exchange is recommended 46 . After successful remission induction (of 3 to 6 months), MPA and WG patients should be switched to maintenance therapy with medium-potent immunosuppression (azathioprin, MTX or leflunomide) 47 48 49 . It is recommended that the glucocorticoid dose should not exceed 10 mg prednisone/day during maintenance. It is not yet known how long maintenance therapy is required but this is being investigated in current trials.

Treatment of localized WG and refractory disease represent the current challenges in the treatment of AAV. For localized disease, no controlled trials are available to support therapy decisions. Trimethoprim/sulfomethoxazole may be used for remission induction in upper airway disease in WG 50 51 although it was not sufficient to control disease activity in a significant proportion of patients with persistent localized WG 13 . In clinical practice, patients in the localized disease stage are treated according to disease activity and severity: manifestations such as granulomatous sinusitis or subglottic stenosis may be treated with MTX. In the case of disease progression (that is, granulomatous sinusitis perforating the orbital wall and affecting the optic nerve) patients usually receive highly potent immunosuppression.

In spite of the use of highly potent and potentially toxic immunosuppression, such as cyclophosphamide, around 15 to 20% of patients (localized and generalized) do not respond to therapy ('refractory disease') 4 13 . In this situation, rituximab, infliximab, deoxyspergualin, antithymocyte globulin, mycophenolate mofetil or intra-venous immunoglobulins are recommended 11 , although controlled trials to support these recommendations are still lacking.

Older age, kidney involvement with impaired renal function, pulmonary manifestations at diagnosis and absence of ENT symptoms have been related to an adverse outcome and increased mortality 4 52 53 . Whereas several studies published in the 1990s showed an increased mortality of WG and MPA compared to the general population 54 55 56 57 , a decrease in the standardized mortality ratios was reported in Swedish patients diagnosed before and after 1996 58 ; furthermore, a recent study reported no (more) increased mortality rates in WG patients diagnosed in the 1990s 59 . The evidence from controlled trials and the implementation of stage-and disease activity-adapted regimens have particularly contributed to the improved outcome. Regardless, mortality rates during the first year after diagnoses remain excessively high (50%), especially in those patients who are severely diseased (in the generalized or severe disease stage) and treated with highly potent immunosuppression 60 . Interestingly, patients died rather from infections than from acute vasculitis 60 .

In general, principles of treatment for CSS are the same as for the other AAVs. A stage- and disease activity-adapted approach is recommended. To guide therapy, the FFS as an additional instrument should be used 34 . Glucocorticoids (GCs) represent the mainstay of CSS treatment. In most cases relatively high doses have to be used for longer periods of time. As evidence accumulates that GCs are also the most significant risk factor for severe infections (see below) and long-term use is associated with a high burden of co-morbidity, therapy should aim to keep GC dose as low as possible. Concomitant treatment with high-dose inhaled corticosteroids may help to control asthma in CSS and lower systemic GC doses, although prospective data in this respect are not available. Medium-potent immunosuppressants, such as MTX or azathioprine, may be used for steroid sparing. In non-life- or organ-threatening CSS, for example, MTX might be used for induction of remission but subsequent relapse rates are high 61 . A FFS of ≥1 usually triggers intense immunosuppression with cyclophosphamide 35 . Further indications for the use of cyclophosphamide are severe peripheral nerve involvement or failure of medium-potent drugs to control disease activity. Most experts recommend maintenance therapy after attaining remission for years, although proof of this concept is lacking in CSS.

Despite treatment with GCs and cyclophosphamide, about 10% of patients show refractory disease courses. The reported 5-year mortality rates in recent series are about 80% 62 . Several salvage therapies have been reported in case reports or small case series. Interferon-α is capable of inducing remission but long-term results are disappointing 63 64 . Rituximab, as in WG and MPA, may be a promising approach but data from larger series are still lacking 65 . Some efficiency was reported with tumor necrosis factor-α blocking agents 66 .

A recently published case series reported the steroid sparing potential of mepolizumab, an IL-5 antibody, but failure to induce remission 67 . In another trial the steroid sparing potential was confirmed and the capability of mepolizumab to induce remission in refractory and relapsing CSS was shown 68 . Targeting IL-5 therefore represents the first targeted approach in CSS and warrants further investigations.

Infections in AAVs may be related to immunosuppressive therapy or the severity of the disease or both 81 and have been reported to occur in 6 to 55% of AAV patients 82 83 . In particular, high doses of GCs (often defined as 30 mg prednisolone/day) and cyclophosphamide have been shown to be associated with infections in AAVs 84 . The dose of oral cyclophosphamide at the time point of infection as well as the cumulative oral cyclophosphamide dose were identified as risk factors 84 . Moreover, the rate of infections within the first 3 years of follow-up was related to the cumulative doses of cyclophosphamide and GCs 84 . Biologics are increasingly used for remission induction in AAV, with similar risk of infection compared to conventional therapy: in a randomized controlled trial of oral cyclophosphamide versus rituximab, the rates of infections in both arms were similar (around 7% of subjects had infections of grade 3 or higher) 44 .

In refractory AAV, even more intensive therapy may be administered, as biologics are often used in conjunction with medium- or highly potent immunosuppression. Serious infections were reported in 20% of refractory AAV patients treated with rituximab plus additional medium- or highly potent conventional immunosuppression, most of whom had pneumonias (14%) 85 . Under anti-thymocyte globulin or deoxyspergualin in combination with GCs, even 40% and 78% of patients, respectively, suffered from infections 86 87 .

A study on risk factors for major infection in WG found that half of the major infections occurred within 3 years after WG diagnosis 84 . Furthermore, infections and not active vasculitis represent the main cause of 'early mortality' (mortality within the first year of diagnosis) in AAVs. 'Early mortality' was higher in study populations with more severe disease (in the generalized or severe stage of the disease) who received more intensive immunosuppression (cyclophosphamide or cyclophosphamide plus plasma exchange) 60 . In summary, patients seem to be most vulnerable to infections shortly after diagnosis. Whether this relates to the intensive immunosuppression required for remission induction or whether the disease and disease activity itself also accounts for a suppression of the immune system and an increased risk of death has not been determined.

Most of the controlled trials in AAV give the rate of infection, but do not specify the kind of infection 42 43 . Yet, a large retrospective study on WG patients and major risk factors for infections reported that pneumonia is one of the most frequent infectious complications under immunosuppression in AAVs 84 . Pneumonia accounted for 36% of all major infections, followed by viral infections (17%). Likewise, lower respiratory tract infection has been reported as the most frequent infection in refractory AAV patients treated with rituximab and deoxyspergualin 86 87 . An intensified diagnostic approach including bronchoscopy with bronchoalveolar lavage is recommended in immunosuppressed patients with pneumonia because of the broad spectrum of pathogens and the uncertainties of empirical antimicrobial coverage in this population.

Pneumocystis jirovecii pneumonia (PJP) is a dreaded complication of immunosuppression and has been shown to occur most frequently in patients undergoing intense remission induction therapy 88 . Furthermore, among AAV patients, those with WG seem particularly at risk of developing PJP 81 . Without the use of PJP prophylaxis, the incidence of PJP was reported to be up to 20% 82 . Age, as well as a low lymphocyte count before and during therapy and prolonged GC doses of >15 to 20 mg/day, are risk factors for PJP 88 89 90 91 . Although there are no controlled data on PJP prophylaxis in AAV, it is recommended as infection rates were much higher in trials not using prophylaxis compared to trials encouraging it 82 92 . Mahr and colleagues 93 reported no more occurrence of PJP since the introduction of PJP prophylaxis. PJP prophylaxis with trimethoprim/sulfomethoxazole is encouraged by European guidelines in all patients receiving cyclophosphamide 11 .