Introduction

The vasculitides are a heterogeneous group of relatively rare conditions that can occur independently or as a secondary feature of an established disease. The most widely accepted classification system is based on vessel size predominantly involved and association with antineutrophil cytoplasmic antibodies (ANCA). Despite the progress in the last years on the field of vasculitides, there are several unmet needs regarding classification, disease activity assessment, predictors of flares and complications, and type of treatment for the different forms.

The aim of this paper is to identify and discuss these unmet needs in vasculitides.

Classification Criteria of Large-Vessel Vasculitis

The classification criteria currently used to define the large-vessel vasculitides (LVV) giant cell arteritis (GCA) [1] (Table 1) and Takayasu arteritis (TAK) [2] (Table 2) were developed in 1990 by the American College of Rheumatology (ACR). These criteria were designed to discriminate between different types of vasculitides but not to differentiate vasculitis from other disorders. In addition, the control group mainly consisted of patients with small-vessel vasculitides, who share only a limited number of features with LVV. Therefore, these criteria should be used to classify, rather than diagnose, patients. However, even for classification purposes, there are still significant shortcomings that limit the use of these criteria. The ACR criteria for GCA strongly focus on patients with cranial manifestations, yet some patients with GCA present predominantly or exclusively with features related to large-vessel involvement. In particular, patients with LVV have less frequently cranial symptoms and a positive temporal artery biopsy [3, 4], hence are less likely to be captured by the ACR criteria. On the other hand, the ACR criteria for TAK lay an emphasis on clinical manifestations of large-vessel disease but do not incorporate the findings of novel vascular imaging techniques that are able to demonstrate early vessel wall changes, which precede the appearance of clinical symptoms and signs [5]. The need to consider imaging findings in diagnosing and assessing patients with LVV has been recognized by the Outcome Measures in Rheumatology (OMERACT) [6] and should inform future classification criteria.

Table 1 1990 ACR criteria for the classification of giant cell (temporal) arteritis (traditional format)
Full size table
Table 2 1990 criteria for the classification of Takayasu arteritis
Full size table

Recently, efforts have been made to overcome the shortcomings of the ACR criteria. With regard to GCA, a diagnostic algorithm has been proposed that includes both biopsy and ultrasonography findings, but this algorithm remains to be yet validated [7]. In addition, a multicentric project, the Diagnostic and Classification in Vasculitis Study (DCVAS) [8], is being conducted with the aim to replace the ACR with new criteria.

Disease Activity Assessment in Large-Vessel Vasculitis

There is evidence supporting the hypothesis that GCA and TAK may represent phenotypes within the spectrum of a single disorder. GCA and TAK present with similar clinical manifestations and similar arterial histopathology revealing granulomatous inflammation [9]. Segmental stenosis, occlusion, dilatation, or aneurysm formation may occur in the vessel wall during the course of the disease. Assessment of the extent of arterial involvement and measurement of current inflammatory status are essential for the optimal management of LVV. However, the lack of an accepted definition of disease activity in LVV presents a major challenge in creating useful and valid outcome tools for the assessment of disease course. One of the major difficulties is the differentiation between activity and damage in LVV [10]. Despite many attempts to adopt standardized approaches to disease activity assessment in LVV, no one measure or set of measures has been accepted as valid and useful for clinical trials [9]. The most commonly adopted measure to evaluate disease activity in TAK is the tool developed in 1994 at the National Institute of Health (NIH) by Kerr et al. and still go by the name of Kerr or NIH criteria [11]. These criteria include four categories: systemic features such as fever and musculoskeletal symptoms; elevated ESR; features of vascular ischaemia or inflammation, such as claudication, diminished or absent pulses, bruits, vascular pain, and asymmetric blood pressure in either upper or lower limbs; and typical angiographic features. The presence of two or more items indicated activity. Subsequent NIH criteria were modified by the addition of CRP to the original version. Advantages of the NIH criteria have been a simple four-category system and the inclusion of angiography. An analysis by the authors, however, revealed that their main drawback was a quite poor sensitivity, with 61% of patients judged as having “inactive” disease incurring progression of angiographic lesions [12]. The Birmingham Vasculitis Activity Score (BVAS) is a validated tool for small-vessel and medium-vessel vasculitis that records the evidence of active vasculitis including multiple manifestations of vasculitis, arranged by organ systems [13], but few studies of LVV have incorporated BVAS [9]. Furthermore, the differences in organ involvement in small-vessel versus LVV are a major concern and common major manifestations of small-vessel disease are rare in LVV. Therefore, the use of BVAS may lead to unnecessary organ evaluation for LVV, whereas cardiovascular findings may be underinvestigated [10]. Furthermore, imaging is not included. The Disease Extent Index-Takayasu (DEI-Tak) was created based on the BVAS with the goal of assessing the extent of the disease, rather than assessing disease activity [14]. Items of the cardiovascular system were given considerable weight in 10 categories, but imaging was not included. The Indian Takayasu Clinical Activity Score (ITAS2010) was an attempt to develop a disease activity score for TAK [15]. This outcome measure evolved from the DEI-Tak and simplified items into six categories, but proportionally the weighting applied to cardiovascular involvement was increased. An activity version (ITAS-A), which included CRP and ESR, is also present. Clinical features must be specific for vasculitis and are scored as absent or present; for clinical features to be considered present, they must be new or have recently worsened. Disease is considered active if at least one organ system scores positive, while inflammatory markers and physician global assessment do not count per se toward assessment of disease activity. The major limitation of ITAS2010 is that the assessment of new vascular signs is by means of physical examination and imaging is not included. The correlation between ITAS2010 and physician global assessment is insufficient, and the tool has not been widely adopted for use in research [9]. In a study of Turkish patients during routine follow-up, ITAS2010 was significantly higher in patients with active disease [16]. However, total agreement between ITAS2010 and physician global assessment was moderate (66.4%) but was better between ITAS2010 and NIH score (82.8%). During follow-up, 14 of 15 patients showing vascular progression with imaging were categorized as having inactive disease according to ITAS2010. No specific tool has been yet designed to assess disease activity of GCA, although the Kerr and ITAS have been empirically used in large-vessel GCA. OMERACT has acknowledged the limitations inherent in the Kerr and ITAS indices and recognized the need to develop finer-tuned assessment tools [6, 9].

Acute-phase response (ESR and CRP) is frequently advocated for disease assessment in LVV, despite being shown to be neither sensitive nor specific enough to monitor disease activity [17, 18]. As in other inflammatory disorders, search for a convenient, reliable, and validated biomarker for LVV is still ongoing.

Imaging procedures are the mainstay to assess large-vessel vasculitis [19]. Color-Doppler sonography (CDS), magnetic resonance (MR) combined with MR angiography (MRA), and contrast-enhanced computerized tomography (CT) combined with CT angiography (CTA) can visualize both the vessel wall and the lumen of large vessels. All these techniques are able to demonstrate early inflammatory signs (vessel wall thickening and mural inflammation) as well as late complications (stenoses and aneurysms). 18F-Fluorodeoxyglucose (FDG) positron emission tomography (PET) is able to detect increased FDG uptake by metabolically active cells, including inflammatory cells infiltrating the vessel wall in vasculitis, while digital subtraction angiography (DSA) is useful to demonstrate luminal changes. Monitoring of luminal changes over time requires morphological imaging. However, the potential role of imaging techniques in monitoring disease activity and response to therapy in LVV patients is still unclear. To this regard, the validity and utility of CT, MR, or 18-FDG-PET to measure the degree of inflammation and disease activity by studying wall enhancement or other parameters remain highly controversial [19].

Predictors of Relapses

Large-Vessel Vasculitis

High-dose glucocorticoids (GCs) are effective in inducing remission in LVV and remain the cornerstone first-line treatment for both GCA and TAK [20]. However, relapses during GC tapering and after GC discontinuation are common, requiring prolonged GC treatment with ensuing GC-related adverse events. Identification of predictors of flares can aid in optimizing therapeutic strategies, minimizing disease flares, and reducing the cumulative GC dose, which is an important risk factor for GC-related side effects.

In observational studies, clinical relapses have been reported in 34 to 74.5% of patients with GCA [21,22,23,24,25,26,27]. These differences are likely related to the lack of an accepted definition of disease flares, with different definition of relapse used in different studies. When a strict definition was used, considering relapse as reappearance of GCA-related clinical manifestations accompanied by elevated ESR and/or CRP, relapses were reported in 34 to 41% of patients [21, 23, 25,26,27]. Most relapses occur within the first 5 years from the diagnosis, almost half of cases within the first 2 years of treatment, when prednisone dose is reduced to less than 10 mg/day [27]. As expected, patients with relapses have longer duration of therapy and higher cumulative prednisone dose and are at higher risk of GC-related side effects [24]. The most consistent predictor of relapse at GCA diagnosis is the presence of a strong initial inflammatory response, characterized by fever, anemia, and severe inflammation at temporal artery biopsy [21, 22, 27]. Other reported predictors of relapses at GCA diagnosis are the involvement of extracranial large vessel, a rise in anticardiolipin antibody levels, female sex, and the presence of hypertension and diabetes [4, 23, 25]. In one study, patients treated with initial prednisone dose higher than 40 mg/day achieved earlier GC discontinuation [25]. Relapses during the course of TAK are common but reported data on relapse frequencies are lacking. In retrospective studies, 42–46% of patients experienced at least one relapse during the first 5 years, and 20% occurred in the first year [28, 29]. In a recent series of Korean patients in remission, 22% had a relapse during a follow-up of 37 months, which was mainly associated with type V angiographic classification (i.e., involvement of the entire aorta and its branches) [30]. In another study, the cutoff level > 1.2 mg/month in the mean dose reduction rate of prednisolone was the only significant independent predictor for disease relapse [31]. In a recent multicenter retrospective French study, male sex, elevated CRP level, and carotidynia were independently associated with a twofold higher risk of relapse [28]. Patients older than 40 years at diagnosis seem at lower risk of disease relapse [32].

ANCA-Associated Vasculitis

Relapses are common in ANCA-associated vasculitis (AAV) with several reports indicating that they occur in about 38–54% of patients after a follow-up between 44 and 62 months [33]. Within the 5-year follow-up of the European Vasculitis Study Group (EUVAS) cohort, 38% of patients had at least one relapse [34]. PR3-ANCA and cardiovascular involvement at diagnosis were independently associated with a higher relapse risk, while renal function was inversely related (renal insufficiency was associated with a lower risk for relapse). Others studies reported an increased risk of relapses in patients with granulomatosis with polyangiitis (GPA) compared with those with microscopic polyangiitis (MPA) [35, 36]. In some series, patients with initial involvement of the lung and/or ENT were at higher risk of relapses [37]. The role of ANCA level is still under evaluation. Recently, a single-center study of 166 AAV patients found that an increase in ANCA level (PR3-ANCA or MPO-ANCA) was predictive of relapses, particularly in patients who had presented with renal involvement and in those with nonrenal severe disease [38]. In a recent study, data from the Rituximab versus Cyclophosphamide for ANCA-Associated Vasculitis (RAVE) trial were used to evaluate the association of an increase in PR3-ANCA level with subsequent relapse [39]. In this study, the association of an increase in PR3-ANCA level with the risk of subsequent relapse was affected by the disease phenotype and remission induction treatment. An increase in PR3-ANCA level during complete remission was associated with an increased risk of relapse among patients with renal involvement or alveolar hemorrhage and those treated with rituximab. Finally, the analysis of the databases of two EUVAS trial (CYCLOPS and IMPROVE) showed that patients who remained ANCA positive (both PR3-ANCA and MPO-ANCA) at the time of switch from induction to remission maintenance therapy had a higher relapse rate compared with those who had no detectable ANCA at switch [40].

Results from the 5-year follow-up of EUVAS patients have shown that cyclophosphamide-sparing strategies by using either pulsed intravenous cyclophosphamide or methotrexate compared with daily oral cyclophosphamide as induction therapy were associated with a higher relapse rate of vasculitis [41, 42].

After cyclophosphamide/prednisolone-based induction, longer remission maintenance therapy with azathioprine/prednisolone (48 months from diagnosis) was associated with fewer relapses compared to shorter remission maintenance therapy (24 months from diagnosis) [43]. In a recent trial enrolling patients with AAV in complete remission after a cyclophosphamide/glucocorticoid induction regimen, maintenance therapy with azathioprine was associated with an increased risk of relapses compared to rituximab [44]. The role of glucocorticoid therapy for relapse prevention is still unclear, but a meta-analysis found that studies with longer courses of glucocorticoids were associated with fewer relapses [45].

Predictors of Complications

Large-Vessel Vasculitis

In GCA, disease severity is mainly related to morbidity, while mortality is generally not increased. Visual loss and cerebrovascular accidents (CVAs) are the most feared cranial ischemic complications of the disease, occurring in up to 20 and 5% of newly diagnosed GCA patients, respectively. Consistent predictors of GCA-related ischemic complications are lacking. Reported predictors include previous ischemic events, marked intimal hyperplasia, and calcification on temporal artery biopsy, moderately (but not very high) serum inflammatory markers at diagnosis, older age, hypertension, smoking, previous ischaemic heart disease, atherosclerosis, and absence of systemic manifestations [46]. Since GCA-related ischemic complications occur early on in the disease course and GCs prevent further ischemic complications, therapy should be started as soon as the diagnosis of GCA is suspected.

Patients with GCA have 6.6–17.3-fold increased risk of developing thoracic aortic aneurysm compared with the general population [47, 48]. The incidence of aortic aneurysm/dissection increases 5 years after GCA diagnosis and continues to increase thereafter [49]. Patients with GCA who develop aortic aneurysm and/or aortic dissection have an increased mortality compared with the general population [49]. Consistent predictors of aortic aneurysm in GCA are lacking. Reported predictors of aortic aneurysm/dilatation are male gender, younger age at diagnosis of GCA, increasing time since diagnosis of GCA, earlier cessation of prednisolone, hypertension, hyperlipidemia, coronary artery disease, increased aortic FDG uptake by PET at GCA diagnosis, aortic regurgitant murmur at GCA diagnosis, extracranial large-vessel involvement, a combination of polymyalgic symptoms and elevated laboratory markers of inflammation at GCA diagnosis, and lower erythrocyte sedimentation rate and higher hemoglobin concentration at the time of aneurysm screening [3, 4, 50].

At a population level, life expectancy of patients with GCA is comparable to that of the general population and the mortality ratio does not differ according to gender [51]. However, patients diagnosed and treated in the hospital setting may be at higher mortality risk compared to GCA patients in the population as a whole. Reported predictors for increased mortality in GCA patients include aortic aneurysm/dissection, comorbid disease, visual loss, and higher maintenance GC doses [49, 51].

Adverse events related to GC are common in GCA, occurring in up to 86% of patients. In a population-based study, predictors of GC-related adverse events were older age and cumulative dose of GC [24].

Although earlier diagnosis and initiation of treatment has improved the outcome [31], both morbidity and mortality are increased in patients with TAK [52], with a high rate of new, severe manifestation during follow-up [53]. Vascular complications (new arterial occlusion, myocardial infarction, and/or hear failure, aortic regurgitation, new-onset or worsening arterial aneurysm, occurrence of stroke/transitory ischemic attack, end-stage renal failure) are common during the course of the disease. Reported predictors of vascular complications are progressive clinical course at diagnosis, thoracic aorta involvement, and retinopathy [28].

Assessment of damage due to disease or treatments such as corticosteroids is becoming the cornerstone of long-term follow-up of vasculitis patients. Age, resistant disease course, disease duration, and cumulative GC dose were independent predictors of damage assessed by the vasculitis damage index (VDI) in a cohort of TAK patients from Turkey [54].

Most pregnancies in patients with TAK are successful; however, women with TAK are predisposed to complications, particularly during the peripartum period. Severe hypertension and preeclampsia are the most frequent complications of pregnancy in women with TA [55]. In a recent systematic review of more than 200 pregnancies in women with TAK, up to 20% of pregnancies were complicated by either intrauterine growth restriction (IUGR) or low birth weight. Patients with renal artery and abdominal aorta involvement experienced more frequent complications of preeclampsia and IUGR [56].

Reported mortality in TAK ranges between 3 and 21% [57]. Disease phenotype and severity of disease expression due to ethnicity, differences in medical therapy, and variations in access to surgical therapy may give rise to different mortality rates. Predictors of poor outcome are progressive disease, retinopathy, severe hypertension, aortic valve insufficiency, aneurysms, severe functional disability, and cardiac involvement [58,59,60]. Overall survival rates of 82.9% at 15 years and of 87.2% at 10 years have been reported from Japan and from Korea, respectively [58, 59]. In a recent US series, the overall survival was 97% at 10 years and 86% at 15 years. Mortality was increased compared with the general population (standardized mortality ratio, 3.0; 95% CI, 1.0–8.9) [29]. Mortality directly related to TA usually occurs from congestive cardiac failure, cerebrovascular events, myocardial infarction, aneurysm rupture, hemorrhage, and renal failure [58,59,60].

ANCA-Associated Vasculitis

Patients with AAV have a very poor prognosis if not diagnosed, evaluated, and treated properly. The introduction of treatment with immunosuppressive therapy has dramatically improved patient survival but with considerable side effects. It is therefore important to recognize and characterize the risk factor that might predict the manifestations associated with poor outcome and prognosis in order to select and stratify patients and treatment strategies.

Several reports have been published on survival of AAV patients. Patient survival has been reported to be ~ 70% at 5 years of follow-up in cohorts comprising GPA and MPA, while in cohorts with exclusively GPA, it is ~ 79% [33]. Many studies have documented a worse outcome for elderly patients and those with renal insufficiency at time of diagnosis of AAV. The patient survival at 1, 2, and 5 years within the 5-year follow-up of the EUVAS cohort was 88, 85, and 78%, respectively [61]. AAV patients had a 2.6-fold increased risk of death compared with a matched general population. Significant predictors of mortality were advanced age, a severely decreased glomerular filtration rate (eGFR < 15 mL/min), and a high Birmingham Vasculitis Score (BVAS) at entry. The main causes of death within the first year of follow-up were infection and active vasculitis, while after the first year cardiovascular (CV) events, malignancy and infection were the leading cause of death [61]. In 1996, the French Vasculitis Study Group validated the five-factor score (FFS) [62], revised in 2011 [63]. The role of the FFS was to evaluate necrotizing vasculitis prognosis and identify clinical manifestations associated with death. The current version includes four negative factors associated with poor prognosis: age (older than 65 years), GI involvement, cardiac involvement and renal involvement, and one positive factor associated with better outcome: ENT involvement [63].

There are indications that patient survival has improved during the last decades, and in GPA, it has been reported to be as high as 95% at 47 months of follow-up [64], possibly reflecting earlier diagnosis, more individually tailored therapy, and more accurate follow-up of patients.

Renal involvement is common in patients with AAV. Patients with MPO-ANCA are at higher risk of end-stage renal failure compared to those with PR3-ANCA [65]. Furthermore, patients presenting with renal insufficiency or dialysis dependency at the time of diagnosis have a worse outcome [66, 67]. Patients with end-stage renal failure treated by kidney transplantation generally have a good outcome, with a low relapse rate [68]. The degree of glomerular sclerosis at renal biopsy seems to be predictive of renal outcome, with a higher degree of sclerosis correlating with a worse renal prognosis [69].

Patients with AAV are at high risk of infection. At 5-year follow-up, 30% of patients from EUVAS trials required hospital treatment for infections and infections were the leading cause of death [61]. A Canadian retrospective study identified lower initial GFR, longer duration of corticosteroid treatment, and presence of lymphocytopenia as risk factors for infections [70].

Historically, the treatment of AAV was associated with an increased risk of malignancy, in particular hematological malignancies, bladder cancer, and non-melanoma skin cancer [71, 72]. In a Danish cohort of patients with GPA diagnosed between 1973 and 1999, the risk of cancer was increased in patients who received more than a cumulative dose of 36 g of cyclophosphamide but not for those who had never received cyclophosphamide or less than 36 g cumulative dose [72]. During the 5-year follow-up of the EUVAS cohort, there was an increased risk for non-melanoma skin cancer but not for other types of cancer [73]. A recent German study also did not find an increased incidence of cancer in patients with AAV. This may reflect a reduced exposure to cyclophosphamide in more recent years [64].

Damage refers to irreversible scarring which does not respond to immune-suppressive therapy. Long-term data from EUVAS trials showed that after a mean of 7.3 year follow-up, only 7.9% of patients had no item of damage whereas 34.4% had accumulated at least five items [74]. Predictors of high level of damage were older age, worse renal function, higher disease activity at presentation, and the number of relapses and duration of glucocorticoid therapy during follow-up [75].

Can Glucocorticoid-Free Treatment Regimens Be Envisaged for Large-Vessel Vasculitis?

GCs remain the mainstay of treatment for LVV but are fraught with a plethora of potential side effects. In a population-based study of 120 patients with GCA diagnosed between 1950 and 1991 in the Olmsted County in the USA, 86% patients developed adverse events thought to be related to GC [24]. Side effects included bone fractures in 46 patients, avascular necrosis of the hip in 3, diabetes mellitus in 11, infections in 37, generalized infection in 2, pneumonia in 18, gastrointestinal hemorrhage in 5, hypertension in 26, and posterior subcapsular cataract in 49. A high cumulative GC dose was a significant predictor of adverse events. Similar findings were revealed by a 15-year (1978–1992) survey from Israel, which showed that 58% of patients with GCA developed serious GC-related adverse events [76]. These data caution against a too liberal use of GC, and current recommendations suggest that the dose and duration of GC therapy should be minimized in order to avoid incurring in unwanted complications [77]. In addition, because effective non-steroidal agents are now available to treat LVV [20], the question arises as to whether GC-free regimens might be conceivable, similarly to what has been proposed for systemic lupus erythematosus [78]. To date, no controlled studies have been performed to establish the efficacy and safety of GC-free therapies in LVV, but a few cases have been published documenting treatment of LVV without GC.

The proinflammatory cytokine TNF-α has been shown to be expressed in the inflamed temporal arteries from patients with GCA [79]. Therefore, TNF-α blocking agents have been considered as treatment for GCA, although monotherapy with anti-TNF agents has rarely been reported. In 2003, Andonopoulos et al. described two patients who received the monoclonal anti-TNF-α antibody infliximab as monotherapy for GCA [80]. Both patients had biopsy-proven GCA and presented with cranial symptoms; they were treated with infliximab 3 mg/kg at onset and after 2 and 4 weeks. The first patient did well for 3 months but subsequently relapsed and required treatment with methylprednisolone, to which he had a good response. The second patient presented with a recurrence of symptoms and a rise in the erythrocyte sedimentation rate 6 weeks after the third infusion and was thus switched to methylprednisolone, again with a satisfactory response. Uthman et al. described another patient with biopsy-proven GCA and cranial symptoms treated with infliximab (two infusions of 5 mg/kg 1 month apart) as monotherapy; she responded well but was later given an additional infusion because of a rise in the erythrocyte sedimentation rate, after which no further treatment was required [81]. Because these findings are very limited and the efficacy of infliximab unclear, anti-TNF-α agents do not appear to qualify as a potential therapeutic alternative to GC in GCA. In line with these findings, a subsequent randomized controlled trial showed that infliximab provided no benefit over placebo in reducing the risk of flares in patients with GCA upon tapering of GC following successful GC-induced remission [82].

Interleukin-6 (IL-6) as emerged as key cytokine in the pathogenesis of GCA [83] and TAK [84], and serum IL-6 levels have been demonstrated to correlate with disease activity in both disorders [83, 84]. The relevance of IL-6 to the clinical expression of LVV has recently been borne out by a randomized controlled trial conducted in GCA that demonstrated that the IL-6 inhibitor tocilizumab plus a prednisone tapering course was superior to placebo and prednisone for the achievement of sustained remission at 52 weeks [85]. These data may suggest a potential role for tocilizumab as an alternative therapy to GC in LVV. So far, the literature contains six patients with LVV treated with tocilizumab monotherapy. Pazzola et al. described two patients with LVV (one with idiopathic aortitis and one with GCA) treated with monthly iv tocilizumab 8 mg/kg for 6 months [86]. Both patients had a quick (4–8 weeks) clinical and laboratory response to tocilizumab, while FDG PET showed resolution or marked improvement of FDG vascular uptake after 6 months. Clinical remission was documented until the last visits.

Salvarani et al. reported two patients with TAK who received monthly iv tocilizumab 8 mg/kg for 6 months [87]. Again, both patients had a quick clinical and laboratory response paralleled by a significant decrease in FDG vascular uptake; remission was subsequently maintained by instituting methotrexate therapy.

In a Swiss series of seven LVV patients, two (both with GCA) received monthly iv tocilizumab (8 mg/kg) infusions without concomitant GC [88]. Both patients had constitutional manifestations and one had jaw claudication; none of them had visual symptoms. Clinical manifestations resolved and laboratory markers normalized without the need of adding GC.

Although the data of tocilizumab monotherapy in LVV are scanty, the efficacy of tocilizumab in this setting may suggest that it could be a suitable candidate for a trial testing its efficacy and tolerability in LVV. Because IL-6 has angiogenic properties potentially able to counteract ischemic complications [89], IL-6 inhibition with tocilizumab might theoretically precipitate ischemic events (although this effect has not materialized in studies where tocilizumab was used with GC [85]). Therefore, it would be prudent to exclude patients at risk for ocular ischemic manifestations from clinical trials on tocilizumab monotherapy. In addition, close monitoring of vascular inflammation would be required, since persistent signs of vascular inflammation at imaging have been reported despite clinical remission in LVV patients treated with tocilizumab [90].

Would Rituximab Be an Option in ANCA-Negative GPA and MPA Patients?

The AAV are a group of diseases characterized by a predominantly small-vessel vasculitis and the presence of ANCA in most but not all cases [91]. In fact, 20–50% of patients with GPA with limited disease and 10% with systemic/severe disease are ANCA-negative [92, 93].

The limited form of GPA is mainly characterized by necrotizing granulomatous lesions with chronic sinusitis, pulmonary nodules, subglottic stenosis, orbital pseudotumor, and pachimeningiitis. Limited and/or early systemic disease in GPA and MPA has been successfully treated with methotrexate (MTX), which in a randomized control trial (RCT) showed the same efficacy as cyclophosphamide (CYC) for the induction of remission [94]. In contrast, in case of more extensive disease, MTX was less effective than CYC. Despite optimal standard therapy, MTX may be not sufficient to prevent relapse or to achieve a sustained remission.

Rituximab (RTX) was introduced for the treatment of ANCA-associated systemic vasculitides (AAV) with the assumption that it would target the B cells and inhibit the production of pathogenic antibodies [95]; the RITUXVAS and RAVE trials enrolled only ANCA-positive GPA and MPA patients, while ANCA-negative patients were excluded [96, 97]. In this regard, it would be interest to understand if these results can be extended to ANCA-negative patients.

The role of B cells in the pathogenesis of the disease is complex and not only ANCA related.

Rituximab is recommended for the management of severe relapsing GPA or MPA refractory to traditional immunosuppressants, especially to treat vasculitic manifestations, while its use for treatment of granulomatous disease manifestations remains controversial [98]. In fact, in some case series [99, 100], RTX was less effective for the treatment of granulomatous manifestations, especially orbital granulomata and pachymeningitis, while other case series and case reports suggest that RTX may represent an option even in this subset of patients, independently from ANCA status [98, 101,102,103]. Even in case series with ANCA-positive and ANCA-negative patients with severe disease, RTX was able to induce remission [104, 105], with poorer outcomes in patients with orbital masses.

The very limited number of cases does not allow to draw definitive conclusions, but the response to RTX seems more related to the characteristics of the disease (granulomatous vs vasculitic lesions) rather than to the ANCA status.

Rituximab Use in Eosinophilic Granulomatosis with Polyangiitis

The treatment of eosinophilic granulomatosis with polyangiitis (EGPA) is based on GCs sometimes combined with traditional immunosuppressants, depending on the severity of the disease. CYC is traditionally considered the first choice in patients with serious organ involvement, while methotrexate (MTX) or azathioprine (AZA) can be used for remission induction in less severe disease and are often employed for maintenance therapy [106].

Nevertheless, the considerable rate of side effects related to GC and/or CYC, the high rates of relapses (41%) [107], and the fact that some EGPA patients do not respond to CYC therapy raises the need for alternative therapies.

Two RCT showed that RTX, a monoclonal antibody against CD20 protein, was as effective as CYC in inducing remission of ANCA-associated vasculitis [96] and appeared superior in relapsing cases [97]. However, these trials enrolled only ANCA-positive GPA and MPA patients, while EGPA was excluded.

There is growing evidence supporting a role of B cells in EGPA. As a matter of fact, about 40% of EGPA patients are MPO-ANCA positive [108], and albeit their pathogenicity in EGPA is still debated, their strong association with the vasculitic manifestations makes this hypothesis likely. Furthermore, in active EGPA patients, high serum IgE and IgG4 levels are detected, independently of ANCA status: all these findings seem to confirm the role of humoral immunity and B cells in EGPA [109, 110]. Another possible mechanism of action of RTX in EGPA may be related to the B cell-dependent activation of T lymphocytes with IL-5 production, the main survival factor for eosinophils [110].

Table 3 displays case series and case reports on the use of RTX difficult cases of EGPA [111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131], in which it seems effective and safe. Remission rate seems higher in ANCA-positive patients [111].

Table 3 Characteristics of the published cases of EGPA treated with RTX
Full size table

To our knowledge, there are at least 80 EGPA refractory or difficult patients in which RTX was able to induce at least partial remission in 74 cases, even though in the majority of cases the response criteria were not specified and in particular neither the assessment of clinical disease activity with a validated clinical tool nor the GC dosage before and after therapy was uniformly reported.

However, RTX showed efficacy in the majority of cases and in a wide spectrum of manifestations (especially pulmonary infiltrates, asthma, glomerulonephritis, ENT manifestations, cutaneous vasculitis, heart involvement, peripheral neuropathy).

Although not licensed for use in EGPA, we agree with the task force recommendations for evaluation and management of EGPA [132] that suggest to use RTX especially in ANCA-positive patients with renal involvement or severe refractory disease despite conventional therapy or when cytotoxic agents are contraindicated or undesirable (e.g., to prevent CYC exposition to young patients and/or toxicity in those who have already received high CYC doses). However, prospective randomized control studies are needed to properly assess the efficacy and safety of RTX in EGPA.

Biologic Therapy in PAN

Polyarteritis nodosa (PAN) is a vasculitis characterized by the involvement of medium and small vessels [133]. The treatment of PAN is usually based on the combination of GCs and immunosuppressants such as CYC, AZA, MTX, or mycophenolate mofetil (MMF), depending on the severity of the disease [134,135,136].

The use of biological agents could be an option for the treatment of refractory PAN patients without hepatitis B virus (HBV) infection, in case of disease resistant to cyclophosphamide and at least one synthetic immunosuppressive agent. RTX has not formally been evaluated in PAN patients, but its use is supported by the efficacy in patients with ANCA-associated vasculitis [96, 97] and few case reports [122, 137,138,139,140,141,142,143]. Other case reports suggest the use of anti-TNF alfa, mainly etanercept (ETA), and infliximab (IFX) in refractory cases [143,144,145,146,147,148,149,150,151,152,153,154,155,156,157].

The very limited number of cases does not allow to draw definitive conclusions, but RTX or anti-TNF alfa may represent a possible alternative therapy in case of refractory or difficult to treat PAN patients.

Unmet Needs in Behçet’s Disease

Behçet’s disease (BD) is a systemic inflammatory disorder whose clinical hallmark is recurrent oral and genital ulcers variably associated with skin and organ involvement. Given its protean clinical manifestations, BD has been classified within different disease frames over the years. It is currently mostly considered a systemic vasculitis, although histology is actually consistent more often with periphlebitis rather than vasculitis proper. Indeed, in BD lesions, the inflammatory infiltrate does not invade nor destroy the vessel wall but tends to surround blood vessels [158].

Additionally, due to some recent insights into its pathogenic mechanisms, there is a current tendency to consider BD a polygenic autoinflammatory disease, at least for some disease phenotypes resembling the “on/off” and self-regulating mechanisms of autoinflammation [159].

BD has a worldwide distribution and both genders can be affected; it can affect all ages as well, but onset is more common in the third decade of life [160], with a more severe course observed in young males [161].

The highest prevalence of the disease is found in the countries along the ancient Silk Route, namely in Turkey, Middle East region, Iran, Saudi Arabia, China, Korea, and (11.9–370 per 100,000 population) [162,163,164,165]. However, the migration of the Middle and Far East populations to the Mediterranean basin over the centuries and an increased disease recognition since the definition of BD as nosological entity at the end of 1930 may partially justify the increased prevalence recorded in some countries, in recent years [162,163,164,165].

Different clinical phenotypes and prognosis of BD patients according to their ethnic background have been reported. A change in disease expression has been attributed to an increased awareness of the disease and a greater accessibility to hospitals, as well as improvements in both hygienic conditions and therapeutic strategies [158].

Despite the new insights into the genetics, new data on epidemiology, disease expression, and distribution, there are still several unmet needs in BD. Diagnosis remains exquisitely “clinical” although plenty of sets of diagnostic/classification criteria have been proposed since the description of the disease in 1937 [166]. Such a criteria showed many limitations in different populations and may have different relevance according to the phenotypes of disease.

About the therapeutic approach, the armamentarium in BD appears still limited as compared to other rheumatologic disease. Hence, treatment of refractory cases, with different organ involvement, may represent a real challenge. We focused our attention on some “unmet needs” in BD regarding (1) current available diagnosis and classification criteria and (2) the therapeutic approach in the cases of refractory clinical manifestations and new therapeutic options.

The “Saga” of Diagnostic and Classification Criteria in Behçet’s Disease

The diagnosis of BD is based on a combination of clinical symptoms and signs, since there is no specific histologic, laboratory, or radiologic finding.

The diagnostic difficulties in correctly defining a patient as being affected with BD—given the wide differential diagnosis required as well as the “vague” clinical signs and symptoms in some cases at the onset—have incited authors in dealing with experts consensus and complex statistical exercises, over the years. Nevertheless, the “saga” of diagnostic and classification criteria in BD has not ended yet [167].

In 1990, the International Study Group on Behçet’s Disease (ISGB) composed of experts from seven countries (France, Iran, Japan, Tunisia, Turkey, UK, and USA) presented the ISGB criteria [168], which consider oral ulcers as a pivotal prerequisite to classify a patient as having BD (Table 4).

Table 4 International Study Group for Behçet’s Disease (ISGB)—1990 criteria [168]
Full size table

ISGB criteria have a very good specificity—at the expense of sensitivity and accuracy—thus, they represent a useful tool to enroll patients in clinical trials, as required for good classification criteria [169]. Some concerns have been raised for diagnosing individual patients.

It is worth noticing that 3% of the entire BD population considered by the ISGB group was excluded, on the basic assumption that oral ulcer is almost always present in BD. Such criteria are by definition inapplicable to the odd BD patients with no oral ulcers.

In a longitudinal study from our group, 87% of patients with BD fulfilled the ISGB criteria at 10 years from the diagnosis, while only 23% of patients with early BD met the criteria (unpublished data).

A posteriori considerations of ISGB diagnostic limitations also include the correct differentiation between gastrointestinal involvement in the course of BD and idiopathic inflammatory bowel disease (IBD) [170], since this distinction remains quite challenging nowadays, in cases with suspected BD with prominent gastrointestinal manifestations.

In 2006, the new International Criteria for Behçet’s Disease (ICBD) were presented; these criteria have subsequently validated in some countries and recently revised [171] (Table 5).

Table 5 International Criteria for Behçet’s Disease (ICBD) [171]
Full size table

Differently from the ISGB criteria, the ICBD criteria included as items both vascular and neurological manifestations. In comparison to ISGB, ICBD have shown better sensitivity (97 vs 77%), lesser specificity (97 vs 99%), and a better accuracy (97 vs 87%).

Nevertheless, the performance of the ICBD criteria needs to be validated in further different populations; therefore, additional validation studies are required.

Treatment of Refractory Clinical Manifestations and New Therapeutic Options: When and Why?

At the EULAR 2016 meeting, new recommendations for the management of BD were developed by a multidisciplinary committee [172].

However, some points remain unclear, and the same authors reiterate the absence of strong evidence for many of them. For instance, no evidence-based evidence is available for the treatment of gastrointestinal BD. Corticosteroids, sulfasalazine, azathioprine, thalidomide, or TNF antagonists should be tried before surgery, apart from emergency conditions [172].

Similarly, there is no firm evidence on how managing major vessel disease in BD. For the management of both pulmonary and peripheral arterial aneurysms, high-dose corticosteroids and cyclophosphamide are recommended; still, in deep vein thrombosis or other arterial lesions, the use of anticoagulants, antiplatelet, or antifibrinolytic agents remains debated.

Given the relatively small therapeutic armamentarium available for BD, treatment of refractory clinical manifestations may be a real challenge.

The major differences from the 2008 EULAR recommendations for BD and the 2016 updated ones were the recommended use of TNF antagonists in refractory cases with all types of involvement and as a first-line therapy in severe neuro-BD [172,173,174]. The most studied drug with the strongest evidence remains infliximab, followed by adalimumab; more recently, the use of other TNF antagonists has been reported [173]. Among them, certolizumab showed favorable results in a retrospective survey of 13 refractory BD patients [175].

New anti-TNF antibody are going to be tested, such as DLX105, a smaller molecule whose tissue penetration is claimed better that larger ones [176].

Preliminary results on the management of six BD patients have been presented in a 2-week open-label study presented during the 2016 17th International Conference on BD. However, it was tested only on muco-cutaneous lesions that tend to have a relapsing-remitting course [177].

Apremilast is an oral phosphodiesterase-4 inhibitor with immunomodulating properties. It has been proposed as a therapeutic option for active oral ulcers in BD, following the promising results of a phase II trial [178]. We are eager to know the results on the recently ended phase III randomized double-blind placebo-controlled trial on the same issue.

Recent insights into the critical role of Th1 and Th17 in the pathogenesis of BD and the correlation of these cytokines with the disease activity led some authors to conduct a small prospective study on the efficacy and safety of ustekinumab on refractory muco-cutaneous BD, with good results [179,180,181].

Evidence on the efficacy and safety of IL-1 inhibitors (anakinra, canakinumab) in refractory BD still comes from small case series and no randomized controlled trial. The latter are needed to establish as a pivotal aim the disease phenotype that can mostly benefit from these drugs [182,183,184].

Similarly, a recent systematic review of intravenous immunoglobulin (IVIG) in BD patients reported six patients successfully treated with this agent, suggesting IVIG as a promising agent especially in patients unsuitable for immunosuppressive therapy [185]. Plasma cytokine evaluation in a BD patient during IVIG therapy showed decrease in CXCL8 levels with a correlation with BD clinical activity [186]. Treatment with tocilizumab is anecdotal as well, with conflicting results. The first report on the efficacy in neuro-BD dated back to 2012 [187]. Our group reported a single-center experience on three patients with refractory neuro-BD with non-convincing results in the long term [188].

Conclusions

Our understanding of vasculitides has advanced, but we still need to clarify methods of diagnosis, evaluation of disease activity and damage, and optimum management. Biomarker and outcome investigations might identify risk factors for an aggressive course leading to treatment tailored to disease severity. International collaborative efforts are ongoing for the development of standardized classification and diagnostic criteria in order to define homogeneous patient groups for randomized clinical trials and to differentiate primary systemic vasculitis from their mimickers.

Different multicenter, randomized, placebo-controlled trials assessing the efficacy and safety of biological agents in the treatment of vasculitides have been recently completed, and others are ongoing. Therefore, in the near future, we will probably face a wider use of biological agents in patients with a more severe prognosis and in patients with relapsing disease. However, the indications and timing for biological treatment in patients with vasculitides should be better defined.

Finally, international concerted efforts could be instrumental in the development of repositories of biological specimens that are essential for translational studies. In turn, a better understanding of the molecular mechanisms associated with the pathogenesis of vasculitides should provide opportunities to improve therapy.