Churg-Strauss Syndrome
Peripheral neuropathies are seen in many systemic vasculitic diseases such as polyarteritis nodosa, Churg-Strauss syndrome, Wegener's granulomatosis, cryoglobulinemias, and hypereosinophilic syndrome. In our patient who presented with a rapidly ascending neuropathy, the diagnosis of Guillain-Barre syndrome was initially entertained, and the patient received five days of intravenous imunoglobulin after developing a fever during the plasma exchange. However, her CSF studies never demonstrated albumino-cytologic dissociation, even when repeated a few weeks after the initial study. Furthermore, Guillain-Barre syndrome (acute inflammatory demyelinating polyradiculoneuropathy) typically exhibits more evidence of demyelination on electrodiagnostic studies, whereas our patient had a predominantly axonal picture. Variants of the Guillain-Barre syndrome, including acute motor axonal neuropathy (AMAN) and acute motor-sensory axonal neuropathy (AMSAN) can occur, but are much less common than the typical Guillain-Barre syndrome. All of these syndromes exhibit a monophasic temporal progression, with deterioration over a few days to several weeks, followed by recovery. However, our patient was transferred for re-evaluation when she apparently relapsed during her inpatient rehabilitation. At this time, the history of fever (to 103 F), rash, and eosinophilia suggested high concern for a systemic vasculitis mimicking the Guillain-Barre syndrome. The patient's asthma and "allergic rhinitis" had been diagnosed only three years previously, and since then she had experienced multiple exacerbations. The patient did have transient eosinophilia (24%) on her initial admission, but a consultant initially attributed this to her asthma, and the same day, the patient received methylprednisolone as part of the treatment for her neuropathic pain, with disappearance of her eosinophilia the following day. In retrospect, we conclude that steroid treatment, which likely partially treated the patient's inflammatory syndrome, probably masked the eosinophilia. During her re-evaluation, muscle and skin biopsies demonstrated tissue and vascular infiltration with eosinophils, which together with her history of recent-onset asthma-like symptoms and her clinical picture of relapsing neuropathy, suggested the diagnosis of Churg-Strauss syndrome.
Churg and Strauss first described a syndrome in 1951 of asthma, allergic rhinitis, pulmonary and systemic small vessel vasculitis, and extravascular granulomas, which they termed "Allergic Granulomatous and Angiitis" (Churg and Strauss, 1951). These original cases were classified by autopsy results. Historically, there has been debate as to whether Churg-Strauss syndrome (CSS) is a separate entity from polyarteritis nodosa. The American College of Rheumatology published six formal classification criteria in 1990:
The presence of four or more of these six criteria yielded a sensitivity of 85% and specificity of 99.7% for a clinical diagnosis of Churg-Strauss syndrome (Masi et al, 1990). This patient's case fits this clinical definition, with four of the six criteria (asthma, eosinophilia, polyneuropathy, and biopsy containing extravascular eosinophils).
The differential diagnosis of CSS includes polyarteritis nodosa (PAN) , Wegener's granulomatosis, Loeffler's syndrome, and the hypereosinophilic syndrome. Probably the most confusion arises from trying to distinguish PAN and CSS. Both PAN and CSS can be associated with asthma, but the incidence of asthma in PAN ranges from 4-54% (Wilson and Alexander, 1945), whereas it is an essential element in CSS. The types of involved blood vessels appear to be different; CSS typically affects small arteries and veins, whereas PAN affects both small and medium-sized arteries (Vertzman, 1980). The dominant cellular infiltrate in these two disorders is also different; in PAN there is a neutrophilic predominance, and in CSS the eosinophilic infiltrate is more striking. Necrotizing extravascular granulomas are not usually observed in PAN. Severe renal involvement in CSS is rare, with renal failure occurring in one patient of 30 in one study (Chumbley et al., 1977) and one patient of 16 in another study (Lanham et al., 1984); early renal involvement is much more common in PAN. (Note that untreated vasculitic syndromes of nearly any variety may confer a high risk of renal involvement). PAN is often associated with evidence of hepatitis B infection (Vertman, 1980), unlike CSS.
Like CSS, Wegener's granulomatosis can present with necrotizing granulomas, vasculitis, and lung involvement. However, in patients with Wegener's granulomatosis, asthma appears to occur at the same frequency as in the general population (Churg, 1963). Initially, pulmonary findings in CSS relate to the consequences of asthma, but in Wegener's granulomatosis, ulceration and necrosis are typical (Fauci and Wolff, 1973). CSS usually affects the lower airways, but a diagnosis of Wegener's granulomatosis typically requires evidence of granulomatous inflammation in the upper and lower respiratory tracts. Eosinophilia, while it can occur in Wegener's granulomatosis, is not as common or as prominent as in CSS. Histologically, CSS has more fibrinoid necrotizing epithelioid and eosinophilic granulomas, whereas Wegener's granulomatosis is typified by coagulative or liquefactive necrosis in epithelioid granulomas. In Wegener's granulomatosis, focal necrotizing glomerulonephritis is common and is a frequent cause of death, whereas renal disease (as stated above) is uncommon in CSS. Although high titers of serum anti-neutrophil cytoplasmic antibodies (ANCA) are often considered to support a diagnosis of Wegener's granulomatosis, elevations of serum ANCA titers have been reported in other inflammatory conditions (see below).
Loeffler's syndrome is characterized by transient shifting pulmonary infiltrates associated with peripheral eosinophilia and commonly with asthma. However systemic vasculitis is not a usual clinical feature of this condition, and therefore neuropathy is uncommon.
Hypereosinophilic syndrome (HES) encompasses a heterogeneous group of conditions defined with eosinophilia (more than 1.5 x 109/L) for longer than 6 months, in the absence of a known cause of eosinophilia (Chusid et al., 1975). Manifestations can range from isolated skin rash to eosinophilic leukemia. The heart can have endomyocardial involvement by eosinophilic infiltrates, and thromboembolic complications are frequent (Parillo et al., 1979). Both HES and CSS can produce the clinical image of Loeffler syndrome or eosinophilic gastroenteritis. Pathologically, HES shows tissue infiltration by masses of eosinophils, but actual angiitis and granuloma formation are uncommon (Lanham et al., 1984).
The three classic histologic criteria described by Churg and Strauss of necrotizing vasculitis, tissue eosinophilic infiltration, and extravascular granulomas were found together at autopsy in only 24% of patients with CSS, and in only 13% of tissue biopsies in a review of 45 autopsy cases and 37 living patients (Lanham et al.,1984). In the same review, only 38% of the tissue biopsies had granulomas and only 40% of the autopsy samples (Lanham et al., 1984). Moreover, the clinical features of the patients with granulomas and without granulomas did not show any important clinical differences. Because of this observation, these authors proposed clinical criteria requiring only asthma, peak peripheral blood eosinophilia > 1.5 X 109/L, and systemic vasculitis involving two or more extrapulmonary organs (Lanham et al., 1984).
Three phases of CSS have been proposed to occur, beginning with allergic disease (rhinitis, sinusitis, and/or asthma), followed by a second phase of blood and tissue eosinophilia (including chronic eosinophilic infiltration of the lung and occasionally eosinophilic gastroenteritis). A third clinical stage of CSS is defined by symptoms and signs of systemic vasculitis (Lanham et al., 1984). Allergic rhinitis may be seen in about 3/4 of patients (Olsen et al., 1995). The mean delay between onset of asthma and development of symptoms of vasculitis has been reported to be approximately three years (Lanham et al., 1984), but intervals as long as 8 years have been reported (Chumbley et al., 1977). The mean delay between onset of asthma and onset of vasculitis was shortened in the patients who died; suggesting that a short time from onset of asthma to development of vasculitis is an unfavorable prognostic sign (Chumbley et al., 1977). The mean age of onset of asthma is 40-45 years, but varied widely with a range of 7-74 years in one study (Lhote and Guillevin, 1998); an older mean age of onset of asthma (50 years) has also been reported (Masi et al., 1990). The male to female ratio in CSS varies from 1:1 to 3:1 depending on the series (Lhote and Guillevin, 1998). The annual incidence is estimated at 2.4 per million (Watts, 1995).
Vasculitis in CSS resembles that seen in polyarteritis nodosa, but severe renal involvement is much less common. When an asthmatic patient begins to have myalgia, arthralgia, fatigue, and/or weight loss, suspicion for vasculitis should increase. The causes of death in CSS are heart failure (48%), renal failure (18%), cerebral hemorrhage (16%), gastrointestinal perforation or hemorrhage (8%), and respiratory failure (2%) (Lanham et al., 1984). Cardiovascular disease occurs in roughly half of CSS patients (Chumbley et al., 1977; Lanham et al., 1984), but can vary greatly, from asymptomatic electrocardiographic abnormalities to sudden death (Lie and Bayardo, 1989; Hunsaker et al., 1992). Cutaneous lesions are present in about two-thirds of patients (Lhote and Guillevin, 1998). The most common cutaneous lesions are palpable purpuric lesions of the lower extremities, but livedo reticularis, maculopapular erythema, and urticaria may be seen (Davis et al., 1997). Skin biopsy may show leukocytoclastic vasculitis, granulomas, or eosinophilic invasion of the dermis.
Neurologic involvement is present in more than 60% of patients with CSS [Lanham et al. (1984) report 75%; Chumbley et al. (1977) report 63%; Lhote et al. (1998) and Sehgal et al. (1995) report 62%]. Mononeuritis (mononeuropathy) multiplex is the most common manifestation in all of these reported series. In a review of 47 patients with CSS evaluated at the Mayo Clinic, Seghal et al. (1995) documented 17 patients with multiple mononeuropathies, seven with distal symmetric polyneuropathy, and one with asymmetric polyneuropathy. In addition, there were three patients with cerebral infarction. Less commonly noted peripheral manifestations included radiculopathies, ischemic optic neuropathy, and bilateral trigeminal neuralgia (see also Weinstein et al., 1983; Lanham et al., 1984; Guillevin et al., 1988). In a more recent review of 28 patients with CSS, mononeuritis multiplex was again the most common neurologic manifestation, with the highest involvement occurring in the common peroneal nerve, followed by tibial, sural, ulnar, and median nerves. The interval from onset of neuropathy to maximal impairment was rapid (2-4 weeks). In all cases in this series, peripheral neuropathy preceded skin and visceral organ involvement (Hattori et al., 1999). It is suggested that ischemic nerve injury, resulting from inflammation of small blood vessels, results in predominantly axonal degeneration, with a degree of segmental demyelination co-existing in some cases (Chalk et al., 1993). Central nervous system involvement including seizures, confusion, coma , and cerebral infarction have been reported in CSS (Churg and Strauss, 1951; Fauci et al., 1978; Cupps and Fauci, 1981; Lanham et al., 1984; Cavallaro et al., 1988).
As stated previously, peripheral hypereosinophilia is one of the criteria for diagnosis of CSS. In the acute stage, an inflammatory pattern may be seen with leukocytosis, elevated erythrocyte sedimentation rate, and elevated serum IgE. Complement levels are usually normal. CSS is strongly associated with perinuclear antineutrophil cytoplasmic antibodies (P-ANCA); in one series 67% of patients (Sehgal et al., 1995) and in another 60% of patients (Guillevin et al., 1993) exhibited positive P-ANCA titers. In a recent review, 86% of CSS patients had a positive rheumatoid factor assay (Hattori et al., 1999). Antinuclear antibodies, anti-dsDNA antibodies, and hepatitis B surface antigen assays were negative in these. It is suggested that the eosinophilia can be used to monitor activity of disease (Chumbley et al., 1977).
Sural nerve biopsy is the diagnostic procedure most often reported in CSS patients with peripheral neuropathy. In our patient, the decision to do a muscle biopsy reflected the severity of the rash on her lower extremities, which raised a concern for infection if a sural nerve biopsy was undertaken. Sural nerve biopsies in the 28 patients reviewed by Hattori et al. (1999) showed the following characteristics: myelinated fiber density was severely but variably diminished, and unmyelinated fiber loss correlated with myelinated fiber loss. In teased fiber preparations, the frequency of fibers with evidence of active axonal degeneration ranged from 30-70%. Necrotizing vasculitis of "medium" (100-150 micron diameter) to "small" (30-50 micron diameter) vessels was seen in the epineurial space in 54% of cases. Vessel walls usually showed necrotizing and hyaline degeneration with the inner elastic lamina destroyed in most involved arteries. Occlusion and recanalization of vessels were frequently seen. Granuloma formation with vessel necrosis was only seen in four cases, while moderate to severe eosinophilic tissue infiltrates were also seen in only four cases. Vasculitic infiltrates of CD45+ and CD3+ T lymphocytes in most cases show equal proportions of CD8+ and CD4+ cells. Macrophages were present in moderate to high numbers in regions of active vasculitis.
As in other vasculitides, the mortality rate for CSS if untreated is high. In Churg and Strauss' original report all but two of their 13 patients died of the illness. The five year survival rate for untreated CSS has been reported as 25%. Mortality has decreased with the use of glucocorticoids, but the treated five year survival rate is still reported as slightly more than 50% in commonly used medical texts. More recent studies suggest that more than 75% of patients survive after treatment (Guillevin, 1996). Prednisone is usually started at the dose of 1 mg/kg/day, and tapered progressively after one month over a 9-12 month period (Cottin and Cordier, 1999). After the inflammatory disease is controlled, an alternate-day regimen is preferred. There may be difficulty in stopping the corticosteroids after the vasculitis is controlled because of residual asthma. Early reports suggested that steroids alone could be used for CSS (Chumbley et al., 1977; Lanham et al., 1984); however, some patients require combined therapy with cytotoxic agents such as cyclophosphamide and azathioprine as well as glucocorticoids. Most of the trials studying CSS have also included patients with polyarteritis nodosa because of the rarity of CSS. If it is assumed that the vasculitides of PAN and CSS are similar enough to use identical therapeutic strategies (which may or may not be the case), then several trials suggest efficacy of both glucocorticoids and glucocorticoids plus cyclophosphamide (Guillevin et al., 1988; 1991). Monthly intravenous pulsed cyclophosphamide therapy (0.6 g/m2) for one year, adjusted to renal function, is preferred to oral cyclophosphamide as it allows a lesser cumulative dose, is less toxic and appears equally effective in controlling the disease activity (Lhote and Guillevin, 1998; Gayraud et al., 1997). Treatment with high dose methylprednisolone at initiation may be helpful (Lhote and Guillevin, 1998). Plasma exchange with corticosteroids alone or with corticosteroids and cyclophosphamide did not improve outcome at 10 years (Guillevin et al., 1991). Guillevin et al. (1996), in a prospective study, identified five conditions associated with a poorer prognosis: proteinuria greater than 1 gm/day, renal insufficiency (serum creatinine greater than 140 umol/L), cardiomyopathy, and gastrointestinal or central nervous system involvement.
Our patient received three days of high dose intravenous methylprednisolone, as well as a total dose of 2.5 grams of intravenous cyclophosphamide divided over five days, and was continued on oral prednisone tapered on a monthly basis. She has also been continued on monthly cyclophosphamide booster infusions. She is doing well and has had good return of her strength six months after her presentation.
There has recently been considerable discussion of the relation between the use of leukotriene receptor inhibitors and the development of Churg-Strauss syndrome. The most commonly held view is that, at least in most patients, the reduction of steroid dosage frequently achieved with the use of these agents allows unmasking of previously undiagnosed Churg-Strauss syndrome (e.g., Wechsler et al., Drug Safety 21:241-251, 1999). In this patient, such an explanation was difficult to adopt, although we cannot rule out a possible effect of the brief steroid doses administered during the exacerbation of her asthmatic illness prior to the onset of neuropathy. Whether there is an absolute increase in the number of cases of CSS associated with the use of leukotriene receptor inhibitors for the management of asthma remains to be determined. Wechsler et al. (1999) prudently suggest that with the use of leukotriene receptor inhibitors, "physicians must be wary for the signs and symptoms of the Churg-Strauss syndrome, particularly in patients with moderate to severe asthma in whom corticosteroids are tapered."
-- Dennis R. Mosier, M.D., Ph.D.
Email comments: