Classification criteria have an important role and practical use in everyday rheumatology. Improvement in therapy and our understanding of the aetiopathogenesis of vasculitis have driven the need to have better descriptors and groupings of diseases. This in turn will allow newer therapy and further understanding to have a greater impact.
The American College of Rheumatology (ACR) classification criteria have been an important advance but have limitations. There remains confusion between classification and diagnostic criteria and definitions. We hope to resolve this using evidence-based improvements in classification and diagnostic criteria. Further understanding of the underlying causative mechanisms could lead to diagnostic testing, eliminating the need for classification criteria.
Introduction
The vasculitides are an uncommon heterogeneous group of diseases that share the pathological hallmark of blood vessel-wall inflammation. The objective of classification criteria is to classify a specific patient into a standardised category for research with an aim to improve patient management . Given the wide variability in the aetiology, pathological features, clinical expression, prevalence and prognosis of these diseases the classification of vasculitis has proved problematic.
Zeek in 1952 was the first to propose a classification for vasculitis . She adopted the term of ‘necrotising angiitis’ and described five distinct vasculitides from a review of the literature. These forms included hypersensitivity angiitis, granulomatous allergic angiitis (recognisable as eosinophilic granulomatosis with polyangiitis (EGPA)), rheumatic arteritis, periarteritis nodosa (polyarteritis nodosa (PAN)) and temporal arteritis (GCA). Recognition of various other types of vasculitides, including Takayasu’s arteritis (TAK) and granulomatosis with polyangiitis (GPA, Wegner’s granulomatosis), in subsequent years led to development of various other classification systems as outlined in Table 1 . Through the years, these classification systems slowly built upon Zeek’s criteria and separated the vasculitides by distinct parameters including size of vessel involvement, histological features and primary or secondary forms.
| Author(s)/date | Key points |
|---|---|
| Zeek, 1954 | First classification system, and has served as the basis for all subsequent classification systems |
| Alarçon-Segovia & Brown, 1964 | Similar scheme as Zeek, but GPA and IgAV added |
| De Shazo, 1975 | Similar to Zeek, with the addition of few subgroups |
| Gilliam & Smiley, 1976 | Better appreciation of the degree of overlap in size of vessels involved between vasculitis subgroups |
| Fauci et al., 1978 | Kawasaki’s disease added to the vasculitis subtypes |
| Lie, 1994 | Subdivisions into primary or secondary forms, as well as predominant vessel size involvement |
The most commonly used classification criteria within clinical trial research are adapted from the American College of Rheumatology (ACR)-proposed criteria. The Chapel Hill Consensus Conference (CHCC) nomenclature is also frequently used in studies. These are a set of definitions rather than classification or diagnostic criteria but are often mistakenly used as such. These two systems are not without their limitations and both are used incorrectly as diagnostic criteria, which was not their primary design. There are no widely used diagnostic tools for vasculitis.
Within the scope of this review we will argue the need to update the classification criteria, particularly in light of our increased understanding of the underlying pathogenesis and improved diagnostic techniques. We will also outline the current international collaborative efforts underway which are attempting to do this.
Throughout this review we have used the new definitions of vasculitis as outlined in the updated 2012 CHCC . Eponyms have been replaced with suitable non-eponymous terminology. This terminology is summarised in Table 2 . In particular CHCC 2012 adopted the recommendations of the ACR, American Society of Nephrology (ASN) and European League Against Rheumatism (EULAR) to replace ‘Wegener’s granulomatosis’ with ‘granulomatosis with polyangiitis’(GPA) . Churg–Strauss syndrome was replaced with the term ‘eosinophilic granulomatosis with polyangiitis’ to provide continuity within the anti-neutrophil cytoplasm antibody (ANCA)-associated vasculitis group. Henoch–SchÖnlein purpura was replaced by the name ‘IgA vasculitis’ (IgAV) based on the established acceptance and evidence that the defining pathological feature is of abnormal IgA deposition in vessel walls.
| Abbreviation | Definition | Previous terminology |
|---|---|---|
| GPA | Granulomatosis with polyangiitis | Wegener’s granulomatosis |
| EGPA | Eosinophilic granulomatosis with polyangiitis | Churg–Strauss |
| MPA | Microscopic polyangiitis | |
| IgA vasculitis | IgA vasculitis | Henoch–Schonlein purpura |
| Other new terminology and minor modifications | Previous terminology |
|---|---|
| Anti-GBM disease | Goodpastures disease |
| Cutaneous arteritis | Cutaneous PAN |
| Cutaneous leukocytoclastic angiitis | |
| Hypocomplementemic urticarial vasculitis |
Historical perspective
Why do we need classification criteria? Classification criteria aim to differentiate patients known to have a form of vasculitis in order to create homogenous cohorts for clinical research. This prevents heterogeneous patients being assessed for the same outcome and leading to erroneous conclusions. One example of this was when Leib et al., in 1979 reported outcomes of 64 cases of PAN . Patients were treated with steroid and immunosuppressant, steroid alone or no treatment. They reported that survival was significantly increased in the combination group. Nine patients had renal biopsies. The histology included patients with focal proliferative glomerulonephritis suggesting that most patients actually had microscopic polyangiitis (MPA) or GPA. Forty-eight percent of patients also had eosinophilia and almost certainly some patients will have had EGPA, especially those accepted on clinical grounds alone. Treatment strategies for ANCA-associated vasculitis are now different from PAN. In the latter, corticosteroids and immunosuppressive agents can enhance viral replication if related to hepatitis B infection. The treatment of this group would include a short course of corticosteroids, plasma exchange and antiretroviral therapy. Guillevin et al. attempted to define clinical, radiological and immunological characteristics of MPA and to separate them from PAN and EGPA . Patients presenting with microaneurysms or multiple-vessel stenosis did not have an ANCA. Skin involvement, glomerulonephritis and the presence of ANCA were found to be significantly more frequent in patients with normal versus abnormal angiograms. Conversely, hypertension, renal vasculitis and hepatitis B infection were significantly more common in patients with abnormal angiograms. This highlighted the potential value of ANCA as a positive predictor of MPA and a negative value predictive of PAN.
Historical perspective
Why do we need classification criteria? Classification criteria aim to differentiate patients known to have a form of vasculitis in order to create homogenous cohorts for clinical research. This prevents heterogeneous patients being assessed for the same outcome and leading to erroneous conclusions. One example of this was when Leib et al., in 1979 reported outcomes of 64 cases of PAN . Patients were treated with steroid and immunosuppressant, steroid alone or no treatment. They reported that survival was significantly increased in the combination group. Nine patients had renal biopsies. The histology included patients with focal proliferative glomerulonephritis suggesting that most patients actually had microscopic polyangiitis (MPA) or GPA. Forty-eight percent of patients also had eosinophilia and almost certainly some patients will have had EGPA, especially those accepted on clinical grounds alone. Treatment strategies for ANCA-associated vasculitis are now different from PAN. In the latter, corticosteroids and immunosuppressive agents can enhance viral replication if related to hepatitis B infection. The treatment of this group would include a short course of corticosteroids, plasma exchange and antiretroviral therapy. Guillevin et al. attempted to define clinical, radiological and immunological characteristics of MPA and to separate them from PAN and EGPA . Patients presenting with microaneurysms or multiple-vessel stenosis did not have an ANCA. Skin involvement, glomerulonephritis and the presence of ANCA were found to be significantly more frequent in patients with normal versus abnormal angiograms. Conversely, hypertension, renal vasculitis and hepatitis B infection were significantly more common in patients with abnormal angiograms. This highlighted the potential value of ANCA as a positive predictor of MPA and a negative value predictive of PAN.
Current classification criteria
1990 American College of Rheumatology
In 1990, the ACR published a series of classification criteria for seven types of systemic vasculitis: GCA, TAK, GPA, EGPA, PAN, IgAV and hypersensitivity vasculitis (HSV) . Rheumatologists from 48 centres submitted a total of 1020 cases. Patients with Kawasaki disease, insufficient evidence of vasculitis and vasculitis secondary to connective-tissue disease were excluded. Patients within each of the vasculitis subgroups were compared to the rest of the group. A total of 807 patients with vasculitis were analysed looking for criteria to distinguish them from each other, with diagnosis made on the basis of expert opinion. No prior definitions of vasculitis had been given to the experts. The criteria help identify a homogeneous group for epidemiology studies and facilitate comparison between different therapeutic strategies. The goals of the ACR committee were to provide a standard way to evaluate patients in therapeutic and epidemiologic studies. The criteria were not intended to be used as diagnostic tools, but are often used as such. The sensitivity for the criteria ranged from 71.0% to 95.3% and specificity 78.7%–99.7%, with the most sensitive and specific criteria in EGPA, GCA and TAK. The positive predictive value of fulfilling any of the ACR criteria for diagnosis of vasculitis was low (17–29%) . The ACR criteria were widely accepted and beneficial for advancing the field of vasculitis research. However, they are not without their limitations ( Table 3 ).
| Type of Vasculitis | Sensitivity | Specificity | Limitations |
|---|---|---|---|
| GCA | 93.5% | 91.2% | Temporal artery biopsy is an important diagnostic tool but is not an obligatory criterion. |
| TAK | 90.5% | 97.8% | Newer imaging modalities, such as CT PET maybe useful, but are not included |
| GPA | 88.2% | 92% | No clear discrimination between GPA and MPA, or other mimics of GPA. Does not incorporate ANCA test. |
| EGPA | 85% | 99.7% | No inclusion of common features such as cardiac manifestations and rash. Does not incorporate ANCA test. |
| PAN | 82.2% | 86.6% | No absolute requirement for arteriography, or biopsy findings. No clear discrimination between PAN and MPA. |
| IgAV | 87.1% | 87.7% | Do not distinguish between IgAV from allergic reactions, or infectious related purpura. Common features; arthritis and nephritis are excluded. Age set as important criteria, but almost 30% of patients were above the age of 20. |
| Hypersensitivity vasculitis | 71% | 83.9% | Difficult to distinguish from IgAV |
| Microscopic polyangiitis | Not recognized by ACR |
The ACR criteria do not include MPA or other primary vasculitides including cryoglobulinaemia, Behҫet’s, relapsing polychondritis, primary central nervous system (CNS) vasculitis and Cogan’s syndrome. They were developed by determining clinical features that distinguished one form of vasculitis over another. They therefore cannot be used to distinguish between vasculitis and other diseases. There is considerable overlap between the criteria, meaning that patients can fall into more than one category . There is considerable heterogeneity within current groupings. They do not include information about biochemical, pathological or radiological investigations.
The ACR classification was developed before the widespread use of ANCA testing. The different forms of ANCA may help to differentiate between GPA, MPA and EGPA.
The sensitivity of ANCA varies according to the assay used and study design, including differences in disease severity and treatment, with figures varying from 34% to 92%. There is also geographical variation in sensitivity. Optimal results for ANCA testing are from combined use of indirect immunofluorescence and enzyme linked immunosorbent assay (ELISA). A meta-analysis of seven studies using these techniques gave a weighted pooled sensitivity of 85.5% for myeloperoxidase antibodies on ethanol-fixed neutrophils (MPO-pANCA) and 98.6% for c-ANCA antigen specific for proteinase 3 (PR3-cANCA) . Specificity seems to be more consistent between studies. The absence of ANCA is also useful diagnostically, particularly in differentiating PAN from MPA.
Radiological techniques have also advanced significantly in the time since the ACR criteria were developed. CT angiography (CTA) and magnetic resonance angiography (MRA) are less invasive than conventional angiography and have similar sensitivity (95% and 100%, respectively) and specificity levels (100%) when assessing for TAK . Ultrasonography (US) may have a role in detecting mural changes, particularly in early disease where CTA does not perform as well. US may also have an important role in GCA, where the abnormal ‘halo’ sign around arteries increases the likelihood of disease and will therefore affect the pre-test probability of a temporal artery biopsy. Compared with biopsy, it has a sensitivity of 69% and specificity of 82% . The role of US is being further assessed in a multicentre study, aiming to recruit 400 patients (TABUL (The Temporal Artery Biopsy vs. Ultrasound in diagnosis of Giant Cell Arteritis) study). Positron emission tomography scanning (PET) may also prove useful in the diagnosis of large-vessel vasculitis, although its radiation dose is large, particularly when combined with CT.
Further examples of where imaging may have a role in diagnosis include PAN, where CTA and MRA may detect medium-vessel aneurysms. Imaging can also be useful to guide biopsies and therefore increase diagnostic yield, such as CT and MRI in GPA to visualise sinus inflammation and destruction. However, there is no consensus regarding the role of radiology in the diagnosis and management of vasculitis and no widely acknowledged gold standard .
Paediatric vasculitis
In 2006, EULAR and the Paediatric Rheumatology European Society (PReS) produced consensus criteria for the classification of childhood vasculitis . Prior to the publication of these criteria, there were no widely accepted paediatric vasculitis criteria, and paediatricians used adult guidelines. The authors of the criteria acknowledged that there were forms of vasculitis such as Kawasaki that occurred exclusively in childhood, other forms such as GCA that did not occur at all and others such as PAN and GPA that had very different clinical features in children compared with adults. The guidelines were intended to be used to aid research and the aim was that they would be easily applied and widely accepted. The criteria are intended to classify those children in whom the diagnosis of primary vasculitis had already been made and not to differentiate vasculitis from mimicking conditions.
The EULAR/PReS guidelines use vessel size to categorise vasculitis based on the CHCC definitions ( Table 4 ). The existing ACR criteria for different forms of adult vasculitis were then modified and are outlined in Table 5 .
| Vessel size | Vasculitis subtypes |
|---|---|
| Large | TAK |
| Medium | Childhood PAN |
| Cutaneous Polyarteritis | |
| Kawasaki disease | |
| Small | Granulomatous |
| GPA | |
| EGPA | |
| Non-granulomatous | |
| MPA | |
| IgA vasculitis | |
| Isolated cutaneous leucocytoclastic vasculitis | |
| Hypocomplementaemic urticarial vasculitis | |
| Other | Behcet’s disease |
| Secondary vasculitis | |
| Vasculitis associated with connective tissue diseases | |
| Isolated vasculitis of the central nervous system | |
| Cogan syndrome | |
| Unclassified |
| Class of vasculitis | Classification criteria |
|---|---|
| IgAV | Palpable purpura mandatory plus 1 of:
|
| Kawasaki disease | Fever for five days mandatory plus 4 of:
|
| Childhood PAN | Systemic illness with a biopsy showing small and medium-vessel necrotising vasculitis or angiographic evidence of aneurysms or occlusions mandatory plus 2 of:
|
| Cutaneous polyarteritis |
|
| GPA | 3 from:
|
| TAK | Angiographic abnormalities of aorta or its main branches mandatory plus 1 of:
|
Related posts:
Update on Vasculitis
Infections in vasculitis
Cardiovascular disease due to accelerated atherosclerosis in systemic vasculitides
Fertility and pregnancy in vasculitis
Imaging in vasculitis
Corrigendum to “Impact of comorbidities on measuring indirect utility by the Medical Outcomes Study Short Form 6D in lower-limb osteoarthritis” [Best Pract & Res Clin Rheumatol 26 (2012) 627–635]
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
