SEARCH

SEARCH BY CITATION

Churg-Strauss syndrome (CSS) has traditionally been included within the spectrum of the idiopathic systemic vasculitides. Based on autopsy findings in 13 patients, it was originally described by Churg and Strauss in 1951 as a condition of allergic granulomatosis and angiitis (1). Key elements of the syndrome were (a history of) asthma, extravascular granulomatous inflammation with massive infiltration of eosinophils, and necrotizing vasculitis (also with prominence of eosinophils). These key elements led Churg and Strauss to summarize their observations under the title “Allergic granulomatosis, allergic angiitis, and periarteritis nodosa.”

Subsequently, it was recognized that, in many patients, the disease follows a characteristic 3-stage course (2). Asthma and/or allergic rhinitis with or without nasal polyposis precedes full development of the syndrome, usually by many years. In the next stage, tissue infiltration by eosinophils, with or without granuloma formation, occurs in various organs, particularly in the upper and lower respiratory tract, the gastrointestinal tract, and the myocardium. Finally, systemic vasculitis develops, with necrotizing small-vessel vasculitis clinically apparent in peripheral nerves (mono- or polyneuropathy, mononeuritis multiplex), skin (purpura), and kidneys (necrotizing and crescentic glomerulonephritis). The observation that changing treatment from corticosteroids to leukotriene receptor antagonists in patients with severe asthma can coincide with the presentation of CSS supported the concept of stepwise development of the syndrome (3). Stopping maintenance treatment with corticosteroids in these asthmatic patients would presumably allow the clinical expression of eosinophilic tissue infiltration and vasculitis.

Various definitions and classification criteria for a diagnosis of CSS have been suggested based on the original observations by Churg and Strauss, data from other series, and, particularly, expert opinion. Lanham et al (2) proposed a diagnosis of CSS when the patient shows the triad of asthma, eosinophilia exceeding 1.5 × 109 cells/liter, and vasculitis involving at least 2 organs. In their criteria, they allow a diagnosis to be based on clinical findings alone without histopathologic confirmation.

The American College of Rheumatology (ACR) criteria for CSS are part of an effort to classify the various idiopathic vasculitides based on characteristic clinical signs and symptoms (4). The criteria were established to distinguish individual forms of vasculitis from each other, and not to distinguish patients with vasculitis from those without vasculitis. As such, the ACR criteria are classification criteria and not diagnostic criteria. Nevertheless, the ACR criteria have been used as diagnostic criteria in many studies. In addition, it should be noted that the sensitivity and specificity of the criteria for a particular form of vasculitis were tested on groups of patients with that particular diagnosis based on expert opinion only. Using this approach, the ACR committee came up with 6 criteria for the classification of CSS, including asthma, eosinophilia >10%, neuropathy, pulmonary infiltrate, paranasal sinus abnormality, and biopsy findings showing extravascular eosinophil infiltration. When a patient fulfilled at least 4 of these 6 criteria, a sensitivity of 85% and a specificity of 99.7% were obtained for classifying the patient as having CSS (4). It is clear from these criteria that a patient with asthma, eosinophilia, and sinusitis with extravascular eosinophil infiltration in, for example, a nasal biopsy specimen will be diagnosed as having CSS, despite the absence of vasculitis, when the ACR criteria are used as diagnostic criteria.

Reflecting in part dissatisfaction with the ACR criteria for the idiopathic vasculitides, a group of experts came together in Chapel Hill, NC, in 1993 in order to redefine the various vasculitides. The resulting Chapel Hill Consensus Conference definitions for the idiopathic vasculitides are therefore based on expert opinion. The aim of the Conference was to establish mutually exclusive definitions for the vasculitides and not to develop diagnostic criteria for these disorders (5). Nevertheless, the Chapel Hill definitions have since been used as a diagnostic tool. CSS has been defined by the Chapel Hill Consensus group as eosinophil-rich and granulomatous inflammation involving the respiratory tract in combination with necrotizing vasculitis affecting small-to- medium–size vessels in association with asthma and eosinophilia. This definition requires not only histopathologic evidence of eosinophil infiltration in the context of granulomatous inflammation, but also histopathologic demonstration of necrotizing vasculitis. Although the Chapel Hill criteria have been mentioned in many reports as prerequisites for establishing a diagnosis of a specific form of vasculitis, it is highly questionable whether all patients diagnosed as having that form of vasculitis fulfill all the components of the definitions proposed by the Chapel Hill Consensus group.

From the foregoing discussion, it is clear that uniform diagnostic criteria are lacking for CSS. The limitations of the diagnostic potential of the ACR criteria, when used without the a priori presence of vasculitis, are apparent. The same applies for the Chapel Hill definitions. For CSS, the situation is even more complicated, because the disease shows different phases of development. Is it possible to make a diagnosis of CSS in a patient with asthma, hypereosinophilia, and granulomatous eosinophil-rich inflammation in various organs but without histopathologically demonstrable necrotizing vasculitis? Some investigators have described these cases as “formes frustes” of CSS, but they have nevertheless categorized these cases as CSS (6).

Tissue infiltration by eosinophils: not specific for CSS

  1. Top of page
  2. Tissue infiltration by eosinophils: not specific for CSS
  3. Is CSS part of the antineutrophil cytoplasmic antibody (ANCA)–associated small-vessel vasculitides?
  4. ANCAs in CSS: a clue to pathogenesis?
  5. CSS: just one disease entity?
  6. REFERENCES

As mentioned, the second stage of CSS is characterized by peripheral blood and tissue eosinophilia. Eosinophil infiltration particularly occurs in the upper and lower respiratory tract, the myocardium, and the gastrointestinal tract, and it may be present at other sites as well. Tissue infiltration by eosinophils, however, is observed in a variety of other conditions including drug hypersensitivity, parasitic infections, malignancies, and allergic conditions. In the lungs, the most common condition showing tissue infiltration by eosinophils is chronic eosinophilic pneumonia, which frequently occurs in association with asthma and peripheral blood eosinophilia. As such, the condition may mimic CSS, although the level of eosinophilia is generally higher in CSS, and necrotizing lesions are less apparent.

In this respect, Katzenstein (7) classified the pulmonary pathologic findings in CSS as classic when eosinophilic pneumonia is accompanied by necrotizing vasculitis and granulomatous inflammation; as highly suggestive when accompanied by necrotizing vasculitis; as suggestive when parenchymal necrosis is prominent; and as less likely in the case of eosinophilic pneumonia alone. The same approach can probably be followed when considering pathologic specimens from the upper airways, in that necrotizing inflammation in conjunction with tissue eosinophilia is more likely to reflect CSS than is tissue eosinophilia alone. Vasculitis is frequently not recognized in these biopsy specimens. It should be noted that in Wegener's granulomatosis (WG) as well, nasal biopsy specimens frequently do not show necrotizing vasculitis but rather show only necrotizing inflammation with or without granuloma formation.

Severe blood eosinophilia (>1.5 × 109 cells/liter for >6 months) and tissue infiltration by eosinophils resulting in organ dysfunction are characteristically present in the hypereosinophilic syndrome (HES). A third criterion for this syndrome is the absence of other causes of eosinophilia such as allergy and parasitic infection. Organ dysfunction due to toxic products from activated eosinophils finally results in end-stage fibrosis, which is particularly apparent in myocardial involvement. Diffuse necrosis and vasculitis are generally absent, and (a history of) asthma is rare in HES. Nevertheless, the clinical differentiation of HES from CSS is not always easy, particularly in the absence of apparent small-vessel vasculitis. More recent data on HES suggest that this disorder results from clonal expansions in the eosinophil lineage, myeloproliferation resulting from chromosomal abnormalities, or expansion of eosinophils resulting from T cell abnormalities with massive production of interleukin-5 (8). Although the pathogenesis of HES has not yet been fully elucidated, these recent data point to a common denominator, that is, clonal or T cell–driven polyclonal expansion of eosinophils in HES. The presence of these abnormalities may allow us to better distinguish HES from CSS.

From the foregoing discussion, it is clear that differentiating CSS from other disorders with eosinophilic tissue infiltration can be imprecise. In the absence of histopathologically demonstrable small-vessel vasculitis, the clinical context, in particular the presence of (a history of) asthma, is highly relevant for a diagnosis of CSS. In addition, a critical review of biopsy material is important to characterize the type of inflammation. Necrotizing and granulomatous inflammation with an abundance of eosinophils, also in the absence of vasculitis, point to a diagnosis of CSS.

Is CSS part of the antineutrophil cytoplasmic antibody (ANCA)–associated small-vessel vasculitides?

  1. Top of page
  2. Tissue infiltration by eosinophils: not specific for CSS
  3. Is CSS part of the antineutrophil cytoplasmic antibody (ANCA)–associated small-vessel vasculitides?
  4. ANCAs in CSS: a clue to pathogenesis?
  5. CSS: just one disease entity?
  6. REFERENCES

The idiopathic small-vessel vasculitides can be divided into the immune complex–mediated vasculitides, such as Henoch-Schönlein purpura and cryoglobulinemic vasculitis, and the so-called pauciimmune vasculitides, in which immune deposits are generally absent in affected tissue (5). Within the latter group, ANCAs are frequently present. In WG, in most cases, ANCAs are directed to proteinase 3 (PR3 ANCAs). Up to 90% of patients with WG are positive for PR3 ANCAs. Levels of PR3 ANCAs often fluctuate in relation to disease activity, and the specificity of PR3 ANCAs for small-vessel vasculitis is extremely high (9). The second disorder within the group of pauciimmune small-vessel vasculitides is microscopic polyangiitis (MPA). MPA is also closely associated with the presence of ANCAs, particularly ANCAs directed to myeloperoxidase (MPO ANCAs). MPO ANCAs occur in almost all patients with a renal-limited form of MPA, that is, idiopathic necrotizing and crescentic glomerulonephritis (9). When MPO ANCA–positive patients are compared with PR3 ANCA–positive patients, independently of their diagnosis, respiratory tract granulomas and extrarenal organ manifestations occur more frequently in PR3 ANCA–positive patients (10). Thus, PR3 ANCAs and MPO ANCAs are sensitive and specific markers for the idiopathic pauciimmune small-vessel vasculitides and are each associated with particular clinical and histopathologic findings.

Do ANCAs occur in CSS? While ANCAs are almost uniformly present in patients with WG and MPA, ANCAs have been detected in <50% of CSS patients in most series with large numbers of patients. In a series of 96 patients with CSS from France, 47.6% were ANCA-positive by indirect immunofluorescence, and most exhibited a perinuclear fluorescence pattern suggesting the presence of MPO ANCAs (11). In this issue of Arthritis & Rheumatism, Sinico et al (12) tested 93 patients with CSS for ANCAs and found 35 of them (37.6%) to be positive; most of these ANCA-positive patients (n = 30) were positive for MPO ANCAs, while a minority of them (n = 3) were positive for PR3 ANCAs. In the series of Keogh and Specks (13), who described 91 patients diagnosed as having CSS, 22 of 30 patients (73%) tested before therapy was started were ANCA positive. In that study from a referral center, ANCAs were tested by indirect immunofluorescence alone and appeared to be present exclusively during active disease in the majority of patients. Despite the higher frequency of ANCAs in the latter study, ANCAs seem to occur less frequently in CSS than in the other ANCA-associated vasculitides.

In WG and MPA, renal involvement, histopathologically apparent as necrotizing and crescentic glomerulonephritis, is frequent. Seventy-seven percent of WG patients demonstrate renal involvement during the course of their disease (14), and 90% of patients with MPA show renal involvement, frequently as part of a pulmonary–renal syndrome with alveolar capillaritis in conjunction with necrotizing glomerulonephritis. In CSS, renal involvement is far less frequent. In the series reported by Lanham et al (2) and Guillevin et al (11), 12 of 122 patients (10%) and 8 of 96 patients (8%), respectively, had renal involvement. Sinico et al (12) report that 25 of their 93 patients (26.9%) had renal disease, a higher prevalence possibly due to differences in referral patterns, since various nephrology departments participated in their study. It may be concluded, therefore, that renal involvement is relatively uncommon in CSS.

Interestingly, in the study by Sinico et al (12), ANCAs occurred far more frequently in CSS patients with renal involvement, particularly in those with rapidly progressive glomerulonephritis, than in those without renal involvement. In addition, the presence of ANCAs was associated with purpura, pulmonary hemorrhage, and mononeuritis multiplex, all of which are clinical features of small-vessel vasculitis. Conversely, heart involvement and lung involvement (other than alveolar hemorrhage) were associated with the absence of ANCAs. These data point to a dichotomy in CSS: a subset of patients is ANCA positive and shows clinical and histopathologic features of pauciimmune small-vessel vasculitis, including glomerulonephritis; and a subset is ANCA negative and is clinically characterized by having tissue infiltration by eosinophils resulting in fibrotic organ damage (Table 1). It should be mentioned, however, that this dichotomy was not absolute in the study by Sinico et al, and that the data in their study were retrospectively obtained from chart review. Furthermore, small-vessel vasculitis, which is supposed to underlie purpura, mononeuritis multiplex, pulmonary hemorrhage, and renal involvement, was not consistently confirmed by biopsy in their patients with CSS.

Table 1. Hypothetical spectrum of Churg-Strauss syndrome based on the presence of antineutrophil cytoplasmic antibodies (ANCAs)
 ANCA-associated subsetANCA-negative subset
Clinical associationsNecrotizing glomerulonephritisNasal polyposis
 PurpuraPulmonary infiltrates
 Pulmonary hemorrhageCardiomyopathy
 Mononeuritis multiplexMono-/polyneuropathy
  Eosinophilic gastritis/enteritis
HistopathologySmall-vessel vasculitisTissue infiltration by eosinophils
PathogenesisANCA relatedToxic products from eosinophils

ANCAs in CSS: a clue to pathogenesis?

  1. Top of page
  2. Tissue infiltration by eosinophils: not specific for CSS
  3. Is CSS part of the antineutrophil cytoplasmic antibody (ANCA)–associated small-vessel vasculitides?
  4. ANCAs in CSS: a clue to pathogenesis?
  5. CSS: just one disease entity?
  6. REFERENCES

There is growing evidence that ANCAs are involved in pathogenetic pathways operative in the vasculitides with this serologic pattern. In addition to evidence from clinical associations (9, 10), in vitro studies have shown that ANCAs can activate neutrophils, resulting in endothelial damage (15). Even more convincingly, in vivo experimental data have demonstrated the pathogenic potential of MPO ANCAs. Transfer of splenocytes from MPO-deficient mice immunized with mouse MPO into wild-type mice resulted in the development of (pauciimmune) glomerulonephritis and pulmonary capillaritis (16). These latter experiments, together with previous experimental work (17), strongly suggest a direct pathogenic role for MPO ANCAs. A role of autoreactive B cells is further suggested by the efficacy of rituximab, a B cell–directed anti-CD20 chimeric antibody, in WG patients whose disease was treatment resistant or who could not tolerate cyclophosphamide (18). The need for high-intensity immunosuppressive treatment in ANCA-positive vasculitis patients was also suggested by the more frequent use of cyclophosphamide in ANCA-positive CSS patients than in ANCA-negative CSS patients, as recorded retrospectively in the study by Sinico et al (12). Extrapolating the experimental findings regarding MPO ANCAs to the pathogenesis of CSS, one may consider ANCA-positive CSS as part of the group of ANCA-associated vasculitides. Accordingly, treatment modalities could follow the therapeutic guidelines suggested for the ANCA-associated vasculitides.

Is the pathogenesis of CSS different in patients who are (persistently) negative for ANCAs? The data from the study by Sinico et al (12) suggest that small-vessel vasculitis is generally absent in these ANCA-negative patients and that tissue infiltration by eosinophils is more prominent. The activated eosinophil may be the major pathogenic factor in these patients. Toxic products of eosinophils such as eosinophilic cationic protein and eosinophil-derived neurotoxin can damage tissues, leading to fibrotic changes and functional impairment. As such, eosinophil-induced damage in ANCA-negative CSS patients may follow pathways comparable with those operative in the HES.

CSS: just one disease entity?

  1. Top of page
  2. Tissue infiltration by eosinophils: not specific for CSS
  3. Is CSS part of the antineutrophil cytoplasmic antibody (ANCA)–associated small-vessel vasculitides?
  4. ANCAs in CSS: a clue to pathogenesis?
  5. CSS: just one disease entity?
  6. REFERENCES

As discussed, criteria used in clinical practice for diagnosing CSS are not uniform. In particular, histopathologic proof of necrotizing small-vessel vasculitis is frequently lacking when one reviews case reports or series of patients with CSS as presented in the literature. This may explain differences in the clinical presentation and prevalence of ANCAs between studies. The data presented by Sinico et al (12) suggest that CSS, when presenting as a form of necrotizing small-vessel vasculitis, is an ANCA-associated disease. If so, demonstration of small-vessel vasculitis or, possibly, positivity for MPO ANCAs or PR3 ANCAs should be required in order to make a diagnosis of CSS when this syndrome is considered a form of idiopathic vasculitis. Also, ANCA-associated CSS should be treated in a manner analogous to the treatment of the other ANCA-associated vasculitides. Patients with a tentative diagnosis of CSS who lack vasculitis and are negative for ANCA may have a subset of the HES. Since pathophysiologic pathways and clinical outcomes may differ between ANCA-associated CSS and this latter subset, treatment of ANCA-negative patients might follow therapeutic guidelines suggested for the HES. It is hoped that future studies will define further the role of ANCAs in the pathogenesis of CSS and their significance as a marker to distinguish disease subsets and guide therapy.

REFERENCES

  1. Top of page
  2. Tissue infiltration by eosinophils: not specific for CSS
  3. Is CSS part of the antineutrophil cytoplasmic antibody (ANCA)–associated small-vessel vasculitides?
  4. ANCAs in CSS: a clue to pathogenesis?
  5. CSS: just one disease entity?
  6. REFERENCES