Controversies in Small Vessel Vasculitis - Comparing the Rheumatology and Nephrology Views

Ronald J. Falk; Gary S. Hoffman


Curr Opin Rheumatol. 2007;19(1):1-9. 

In This Article

Antineutrophil Cytoplasmic Antibodies and Nomenclature: What's in a Name?

I have not embraced the notion of a spectrum of disease within the universe of vasculitis that should be classified as 'antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV), ANCA vasculitis or ANCA disease'. My colleague, Dr Falk, is among those who have proposed that AAV is a group of like diseases that includes Wegener's granulomatosis (WG), microscopic polyangiitis (MPA), renal-limited pauci-immune glomerulonephritis (RLPIGN) and Churg-Strauss syndrome (CSS). He and many other authorities have proposed that AAV are phenotypic variants of a single entity. This assertion places a great deal of weight on the presence of an antibody, ANCA. I believe the term AAV does not recognize the importance of immunologic and clinical differences among this proposed aggregate of diseases in which heterogeneity influences presentation, complications, selection of treatment strategies and outcomes.

WG, MPA, RLPIGN and CSS have unique clinical differences. A patient with WG may present with ear, nose, throat and tracheal features, none of which are a part of MPA or RLPIGN. The strategies to treat such complications extend beyond deciding which cytotoxic agent to add to glucocorticosteroid therapy. In WG, apart from being possible signs of active disease, one must consider whether chronic otitis media and conductive hearing loss may require placement of tympanotomy tubes, whether nasal complications require debridement of crusts or measures that improve moisturization and lubrication, and whether subglottic or large bronchus stenosis requires surgical interventions. Unlike MPA, CSS or renal-limited disease, active WG can also present with inflammatory mass lesions that may appear in any organ and must be differentiated from a comorbid malignancy or abscess. Although asthma and peripheral eosinophilia are infrequently found in WG, MPA and RLPIGN, they are characteristic manifestations of CSS. Noting such differences does not deny the fact that in some patients it may be difficult, at the time of presentation, to distinguish WG from MPA or to know whether RLPIGN will evolve into these diseases. In my experience, however, such cases constitute a small minority.

Striking histopathologic differences provide further testimony to the distinct nature of these diseases. Whereas the renal histopathologic characteristics of all four diseases are similar, lesions in other organs may be quite distinct. Nonspecific inflammation and granulomatous lesions are common in WG and are often present in the absence of vasculitis. Granulomatous changes, however, are not part of the pathologic characteristics in MPA or RLPIGN. Biopsies from patients with CSS typically show an abundance of eosinophils, which play an important role in tissue injury, but are not characteristic of the other disease entities.

The term 'AVV' implies that ANCA is critical to the pathogenesis of these diseases. If the presence of ANCA is essential to disease pathogenesis, it would be expected to be present in all patients in whom disease is active and titers would increase prior to or concurrent with disease exacerbations. The literature that has addressed these issues has become vast, and is remarkable for differences between studies with regard to spectrum of diseases included, distribution of organ system involvement, and methods by which disease activity and ANCA are measured. Not withstanding these limitations, a brief review of recent large controlled studies in which ANCA was obtained in all patients prior to or at the time of enrollment may serve to inform our discussion about the controversies surrounding AVV. The WG Etanercept Trial (WGET)[1] was a collaboration primarily among rheumatologists. WGET included 180 patients in whom it was required that WG be unequivocally active for study entry. The range of disease severity and distribution of organs involved was broad. Milder presentations of disease, for sake of convenience, were labeled 'limited' WG. We agree with most authors who are critical of the term 'limited' and its lack of precision, and that is why we have been careful to define it with regard to disease severity and organ distribution. Individuals with milder or limited WG did not have immediate critical organ- or life-threatening involvement. Fifty-two patients (29%) in WGET were considered to have 'limited' disease; 128 had more severe disease. About 54% of patients had glomerulonephritis, but only 21% of all patients had abnormal serum creatinine values. The mean serum creatinine value at enrollment for all patients was 1.74 mg/dl (0.97 in limited WG and 2.06 mg/dl in the more severe WG group). Patients could qualify for randomization within the 'limited WG' group if they had renal disease, but those features had to be minimal, which led to creatinine values being elevated in less than 4%, compared to 28% of patients in the more severe group. This spectrum, including absent or mild renal disease, is not likely to occur in the practice of nephrologists who provide care for patients with WG.

Among WGET patients with severe disease, a past history of ever having been ANCA-positive was about 90% by either immunofluorescent or enzyme immunoassay techniques. In patients with milder ('limited') disease, however, positive results were noted in about 80% by immunofluorescence and 70% by enzyme immunoassay (P = 0.028). The European Vasculitis Study Group (EUVAS) has also carried out some of the first randomized controlled studies of treatment for SVV (WG, MPA and RLPIGN) that required all patients to have active disease at study entry. EUVAS includes many nephrologists, which accounts for a greater percentage of patients in their randomized trial of maintenance therapy for AAV having glomerulonephritis (94%), impaired renal function and ANCA-positivity (above 90%).[2] These and other published reports suggest that more severe illness is associated with a higher frequency of ANCA; however, there still remains a minority of patients within the 'AAV' population who are ANCA-negative, or if they do produce undetectable quantities of ANCA, detection is beyond the limits of current tests. If we assume the former, it would be difficult to accept that ANCA is essential for disease pathogenesis, i.e. disease would not occur if ANCA was not present.

Birck et al.[3] attempted to perform a systematic review of studies that examined the relationship of ANCA titer change to determine its value as a measure of disease activity. They concluded that while the value of ANCA as an adjunct to diagnosis is well established, the usefulness of serial determinations to assess disease activity or likelihood of relapse could not be determined from the existing literature.

A more recent derivative study for the WGET Research Group[4] has found that increases in antiproteinase 3 antibody titers did not predict disease relapses. These observations lead to the important clinical recommendation that ANCA titers not be used as a guide to adjusting immunosuppressive therapy. The lack of correlation of increasing ANCA and disease activity is an additional argument for ANCA not being essential to the pathogenesis of at least WG. In following this argument it should not be construed that ANCA lacks importance or does not play some role in pathogenesis.

Some authors have cited a particular case report as strong proof of ANCA being causal in MPA. This case involved a mother with severe pre-eclampsia and a past history of MPA who passed anti-myeloperoxidase antibody by the transplacental route to her fetus. The premature child (33 weeks gestation, 4lb 5oz) became ill, appearing to have respiratory distress syndrome that was treated with ventilatory assistance, antibiotics, surfactant lavage, exchange transfusions and corticosteroids. Biopsy proof of vasculitis or glomerulonephritis was not available. This event was the subject of two publications.[5,6] In considering all the data available from both papers, it seems certain that antibody to myeloperoxidase was transmitted from the mother and may have played a role in the infant's subsequent illness. However, it is also possible that prematurity, respiratory distress syndrome and its complications might have explained all of the newborn's findings. Until other examples of transplacental passage of ANCA are reported to cause disease in an otherwise uncomplicated pregnancy, it may be prudent to withhold judgment on how this case informs the argument of ANCA and disease causation.

Let us consider the experience with CSS. CSS has been said by some authors to be one of the AAV. Two recent studies have again raised questions about whether such nomenclature might be misleading, while also asking a related interesting question: if ANCA is present, might it influence disease expression? This conclusion might be presumed from the foregoing observations in WG and MPA in which patients with more severe disease, including renal, are more likely to be ANCA-positive. Sablé-Fourtassou et al.[7] studied 112 patients with active CSS among whom only 38% had circulating ANCA. ANCA-positive patients had a higher frequency of glomerulonephritis (35 vs. 4%), while ANCA-negative patients had a higher frequency of heart disease (49 vs. 12%). Sinico et al.[8] also demonstrated ANCA-positivity in only 38% of 93 patients, but subsets of patients with ANCA were again more likely to have glomerulonephritis (51 vs. 12%) and those lacking ANCA were more likely to have heart disease (22.4 vs. 5.7%). Other phenotypic differences were noted in ANCA-positive and -negative patient subsets, but were discordant between studies and therefore may be of more questionable significance. On the basis of these findings, these authors hypothesize that ANCA may modify the risk of vasculitis in certain organs.

These observations support the argument that ANCA has important roles in some forms of SVV: (1) as a diagnostic tool, (2) influencing pathogenesis for certain subsets of patients and (3) influencing disease phenotypes. As ANCA has not been shown to be essential to the pathogenesis of WG, MPA and CSS, however, to place it at the center of the pathogenesis universe for all of these diseases would seem premature. The danger of fixing our thoughts to the notion and nomenclature of 'AAV' risks biasing clinicians towards placing excessive weight on serologic findings and influencing investigators away from exploring other important pathways (e.g. roles of T, dendritic, macrophage, natural killer, regulatory T and other cells) that contribute to these complex diseases. The same may not be said for RLPIGN, where the entity itself is partly defined by the presence of the serologic abnormality.

In 1954, Godman and Churg concluded that WG, CSS and MPA (but not polyarteritis nodosa) were morphologically and probably pathogenetically related.[9] I share the opinion that SVV is the clinical equivalent of a genus that encompasses the different species - MPA, WG and renal-limited SVV. In my opinion, differences in disease phenotype are engendered by the individual host responses to pathogenic pressures more than WG, MPA and kidney- or lung-limited disease being completely separate disease entities. We are well accustomed to recognizing protean clinical manifestation even within disease entities related to single pathogenic factors; syphilis, tuberculosis and diabetes mellitus, to name but a few. For example, the patient with diabetes mellitus may have minimal signs and symptoms of disease or severe manifestations that include kidney failure or cardiomyopathy. These patients all have diabetes, not different diseases. The Chapel Hill nomenclature, widely accepted as useful in most of the globe, provided disease definitions for vasculitis.[10] While noting the common basis of SVV, differences between WG, CSS and MPA were described based on specific pathologic features. ANCA was not part of the definition. While these definitions are relatively easy to distinguish on paper, in clinical practice this separation can be difficult. Are pulmonary infiltrates a consequence of a capillaritis or a granulomatous lesion? With what degree of certainty can one distinguish MPA from WG on clinical grounds alone? Some practitioners suggest that all ear, nose and throat disease is a consequence of WG. There is no more basis for this assertion than the claim that the Earth was flat! Many patients with MPA have capillaritis in their upper respiratory tract with no evidence of granulomatous disease. In modern practice, searching for a granulomatous lesion in the lung or ear, nose and throat is rarely performed, and, even if sought, not always found. Thus, to a certain extent, stamping the label of WG or MPA on a patient is frequently in the mind of the beholder. This imperfection in disease diagnosis is mitigated by the fact that induction and most maintenance therapy do not require pathological confirmation of a granulomatous disease process, but in my opinion do require a biopsy diagnosis of SVV.

The concept of 'limited' disease is troubling because clinical acumen is not sensitive enough to detect subtle vasculitis in the spleen, lymph nodes or even in the respiratory tract at times. Moreover, limited disease may imply 'mild' disease, which might not in fact be the case. For example, in the WGET trial, limited disease included patients who had glomerulonephritis.[11] I would argue that glomerulonephritis often leading to end-stage kidney disease in this population can never be considered 'mild'. The entry serum creatinine in the overall WGET trial was approximately 1.7 mg/dl. From a nephrologist's perspective, this suggests that the mean serum creatinine in the entire patient group had moderate to severe kidney disease corresponding to a 30-50% loss of kidney function. Our experience with patients who had 'limited' or 'milder' glomerulonephritis is evident in prospective observations of patients who were treatment-resistant.[12] These patients typically had past histories of repetitive bouts of hematuria and advancing proteinuria as a consequence of relapsing and remitting focal and necrotizing glomerulonephritis that resulted in glomerular scarring. In our population, 22% of patients had resistant disease that resulted in end-stage glomerulonephritis and another 11% of patients who initially responded to treatment progressed to end-stage kidney disease without ever experiencing a disease relapse. In reality, the serum creatinine may not reach the 'abnormal range' until 50% of kidney function is lost. Certainly, this would not be described as 'limited' or mild from the perspective of the kidney, or for the patient who may end up on dialysis. Glomerular hematuria is a better marker for glomerular disease than is the serum creatinine; however, it is generally under-recognized or under-detected and is frequently dismissed as due to benign causes such as urinary tract infection or menses. Emerging studies into gene expression and lymphocyte alteration suggest that significant autoimmune and/or inflammatory responses can occur in relatively asymptomatic patients.

Why concentrate on ANCA? There are emerging and almost incontrovertible data that ANCA are pathogenic, especially myeloperoxidase ANCA. There are a wealth of in-vitro data supporting the observation that ANCA are capable of activating leukocytes resulting in neutrophil and monocyte endothelial interaction that results in endothelial damage using a number of outcome measures.[13,14,15] The advent of two different animal models in which antimyeloperoxidase antibodies in mice and in rats are able to engender a necrotizing and crescentic glomerulonephritis with widespread vasculitis in the lung, spleen, lymph node and other organ systems is evidence that antimyeloperoxidase antibodies are capable of causing disease by themselves.[16,17] It is interesting that the phenotype of this disorder may be different depending upon the host animal rather than on the very nature of the antimyeloperoxidase antibody. Some animals develop more pulmonary disease; some animals develop more kidney disease. The disease is more severe if the animals are boosted with lipopolysaccharide and is made less severe if neutrophils are lacking.[16,17,18] There is clinical evidence that ANCA are pathogenic, including an example in a fetus receiving myeloperoxidase ANCA from the mother and that subsequently developed an ANCA SVV which dissipated once the antibody disappeared.[6] The detractors from the pathogenic ANCA theory invariably invoke two important concerns. ANCA titers are not sufficiently correlative with disease activity to be used for the management of immunosuppression in all patients and there is a percentage of patients with SVV who mimic ANCA-positive patients, but who do not have these circulating autoantibodies. These are valid concerns, but teeter on the reality that 94% of patients with MPA, WG or pauci-immune necrotizing glomerulosclerosis are ANCA-positive. While changes in ANCA titers (which are as yet to be standardized) do not correlate with our (insensitive) measure of disease activity, there is mounting and convincing evidence that they do correlate with subsequent level of activity or relapse. For example, persistence of a positive proteinase 3 test after immunosuppressive therapy was shown to be associated with an increased risk of relapse in two independent studies.[19,20] I would suggest that the ANCA pathogenesis model is a two-hit model that requires multiple factors, including ANCA, to be correctly aligned and that its cognate antigen is available on the surface of the leukocyte. It is also true that the remaining 10-15% of patients who are ANCA-negative always remain ANCA-negative. Whether this minority of patients has a different pathogenetic mechanism or the antibody has not been recognized remains unclear. A strong argument stems from the experience in CSS, where there is more agreement than disagreement. Recent studies of ANCA in CSS suggest a varying range of patients with ANCA positivity.[7,21] The majority of patients with CSS who are ANCA-positive also have glomerulonephritis and those lacking ANCA are more likely to have cardiac disease. These distinctions raise interesting questions about the clinical nature of CSS, which may eventually help guide therapy.

From a practical perspective, the majority of grade A studies pertaining to MPA, WG and renal-limited disease have used ANCA as an entry into their studies. These include all of the EUVAS studies, including CYCAZAREM,[2] NORAM,[22] CYCLOPS[23] and MEPEX.[24] The EUVAS community includes nephrologists, rheumatologists, and otolaryngologists who have found common approaches to these diseases. Our approach is based on histology, pathogenesis and prognosis. Prompt diagnosis and induction therapy are more important than waiting for a diagnosis that allows for arbitrary separation of patients into one specific disease entity or another. In our own practice, patients with biopsy-proven SVV who have a possible myeloperoxidase or proteinase 3 ANCA are diagnosed as having ANCA SVV. With this multidisciplinary approach to early diagnosis and treatment, prompt therapy can begin; thus, earlier disease remission can be attained rather than waiting for granulomatous lesions to be detected on biopsy in order to identify a disease as 'WG' or 'MPA'. I cannot recall the last time I relied upon lung or ENT biopsy to provide a specific phenotypic diagnosis, unless those were the only organ systems involved with the disease.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.