Churg-Strauss syndrome

Study objectives: We previously reported eight patients who developed Churg-Strauss syndrome in association with zafirlukast treatment for asthma and postulated that the syndrome resulted from unmasking of a previously existing condition due to corticosteroid withdrawal and not from a direct drug effect. The availability of montelukast, a new leukotriene receptor antagonist with a different molecular structure, permitted us to test this hypothesis. Our goals were to ascertain whether the Churg-Strauss syndrome developed in patients taking montelukast and other novel asthma medications, and to describe potential mechanisms for the syndrome.

Design: Case series.

Setting: Outpatient and hospital practices of pulmonologists in the United States and Belgium.

Patients: Four adults (one man, three women) who received montelukast as treatment for asthma; two women who received salmeterol/fluticasone therapy, but not montelukast.

Results: Churg-Strauss syndrome developed in the four asthmatic patients who received montelukast. In each case, there was a long history of difficult-to-control asthma characterized by multiple exacerbations that had required frequent courses of oral systemic corticosteroids or high doses of inhaled corticosteroids for control. Two other asthmatics who received fluticasone and salmeterol but not montelukast therapy developed the same syndrome with tapering doses of oral or high doses of inhaled corticosteroids.

Conclusions: The occurrence of Churg-Strauss syndrome in asthmatic patients receiving leukotriene modifiers appears to be related to unmasking of an underlying vasculitic syndrome that is initially clinically recognized as moderate to severe asthma and treated with corticosteroids. Montelukast does not appear to directly cause the syndrome in these patients.

(CHEST 2000; 117:708-713)

Key words: asthma; Churg-Strauss syndrome; corticosteroid; leukotriene; montelukast; vasculitis

Abbreviations: ACR = American College of Rheumatology; CSS = Churg-Strauss syndrome; EF = ejection fraction; EMG = electromyography; ESR = erythrocyte sedimentation rate; GPRD = General Practice Research Database

We previously reported eight cases of Churg-Strauss syndrome (CSS) that were associated with withdrawal of corticosteroid treatment in patients with asthma treated with the cysteinyl leukotriene receptor antagonist zafirlukast.[1] In that report, we postulated that the induction of CSS was not due to direct effects of zafirlukast, but rather to an underlying systemic eosinophilic disorder that was chronically masked by systemic or high-dose inhaled corticosteroid treatment of what was perceived to be severe asthma. Since that time, there have been three other published case reports of CSS in relationship to zafirlukast use. In two of these cases, zafirlukast appeared to provide protection against the airway obstruction due to leukotriene excess and thus provided the opportunity for withdrawal of corticosteroid treatment, which in turn unmasked the underlying and previously unrecognized CSS.[2,3] However, another recently published report described the occurrence of CSS in association with zafirlukast use in two patients who had not received systemic steroid treatment.[4] As of November 15, 1998, the estimated exposure of zafirlukast was 1.3 million patients; in association with this leukotriene-modifying medication, 33 cases of vasculitis (as determined by biopsy or the presence of surrogate clinical markers of vasculitis) meeting the American College of Rheumatology (ACR) diagnostic criteria for CSS[5] have been reported to Zeneca Pharmaceuticals and the US Food and Drug Administration. In almost all of these cases, the systemic eosinophilic disorder (ie, CSS) became manifest after corticosteroid withdrawal (Catherine M. Bonuccelli, MD; personal communication; December 16, 1998).

In early 1998, another cysteinyl leukotriene receptor antagonist, montelukast (Singulair; Merck; Whitehouse Station, NJ), became available for use as a treatment for asthma. Because the mechanism of action of montelukast is similar to that of zafirlukast, even though their chemical structures are different, we reasoned that newly incident cases of CSS in patients receiving montelukast for asthma would provide data to support our hypothesis that antileukotriene treatment in certain patients perceived to have severe persistent asthma could allow withdrawal of systemic corticosteroid treatment and subsequent unmasking of signs and symptoms consistent with CSS. In the safety trials leading to the approval of montelukast, approximately 2,600 subjects received the drug; in this cohort, the incidence of side effects was comparable to placebo. Neither systemic eosinophilic disorders nor cases of CSS were reported. Since the release of montelukast, there have been an estimated 350,000 patient-years of exposure to this agent.[6] Merck has received several reports of cases of CSS in association with montelukast use, with an estimated case rate of 60 cases per million patient-years.[7] We present four cases of CSS in association with montelukast use that became manifest after corticosteroid withdrawal and suggest a mechanism for this syndrome in patients taking this class of medications (see Table 1).

Table 1--Characteristics of Asthmatic Patients Who Developed CSS in Association With Systemic Corticosteroid-Sparing Agents(*)

(*) F = female; M = male; N/A = not available; + = present; - = absent.

([dagger]) As noted in Green and Vayonis.[4]

CASE REPORT 1

A 62-year-old white woman with chronic asthma, rhinitis, sinusitis, and eosinophilia and multiple courses of systemic corticosteroids was started on montelukast in March 1998 for asthma control. Her asthma improved such that she was able to significantly reduce the dosing of her inhaled corticosteroids. Three months later, she developed musculoskeletal pains, weakness, and sensory loss in both lower and upper extremities. She was admitted to hospital with polyneuropathy and was noted to have an erythematous rash on her right hand. Her total eosinophil count was 9.97 x [10.sup.9]/L (49% of her total WBC count). Electromyography (EMG) was compatible with mononeuritis multiplex. Skin and muscle biopsies revealed eosinophilic vasculitis compatible with CSS. Chest radiography showed no pulmonary infiltrates. Montelukast was discontinued, and treatment with systemic corticosteroids and cyclophosphamide resulted in complete resolution of her rash and eosinophilia within a few days, and partial resolution of her polyneuropathy.

CASE REPORT 2

A 25-year-old man with a history of childhood wheezing and exertional dyspnea, but no history of sinusitis, received fluticasone and albuterol for treatment of mild persistent asthma. Since he had increased asthma symptoms despite treatment with high-dose fluticasone (1,720 [micro]g/d) and salmeterol (200 [micro]g/d), therapy with montelukast was initiated in March 1998. Symptoms progressed over subsequent months until he required three courses of systemic corticosteroids, the last of which ended in August 1998. In October 1998, he developed diarrhea, fatigue, and shortness of breath, as well as a vasculitic rash on the dorsum of his hand. On admission to hospital, bibasilar rales were present and stool was positive for occult blood. He was hypoxemic, with a [PO.sub.2] of 63 mm Hg. His WBC count was 13.3 x [10.sup.9]/L with 26% eosinophils, and his erythrocyte sedimentation rate (ESR) was 100 mm/h. Chest radiography revealed diffuse interstitial infiltrates suggestive of congestive heart failure, and echocardiography revealed global left ventricular hypokinesis with an ejection fraction (EF) of 25%. EMG was notable for right median and right peroneal neuropathy consistent with mononeuritis multiplex. Montelukast was discontinued, and after treatment with high-dose IV corticosteroids, both dyspnea and diarrhea abated, and eosinophilia resolved within 48 h. None of the symptoms recurred over 6 weeks of follow-up; at that time, echocardiography revealed an EF of 45%. A mild neuropathy persists.

CASE REPORT 3

A 38-year-old woman with a 10-year history of sinusitis and asthma and documented history of eosinophilia was maintained on treatment with inhaled steroids and [Beta]-agonist. She had been treated with systemic corticosteroids three to four times per year for asthma symptoms and had a 4-month course of prednisone for treatment of sinusitis through early August 1998. She was started on montelukast (but not systemic corticosteroids) in mid-August 1998 in the hope of avoiding further systemic steroid treatment, and presented in October 1998 with increased cough and shortness of breath. Chest radiograph revealed patchy infiltrates, and CBC count revealed a WBC count of 18.7 x [10.sup.9]/L with 32% eosinophils. Neurologic examination was normal, but perinuclear antineutrophil cytoplasmic antibody was positive (titer 1:20) and ESR was 38 mm/h. She denied fevers, myalgias, GI symptoms, and rash; both echocardiography and urinalysis were normal. Bronchoscopy with lavage revealed 87% eosinophils, and all cultures were negative. After montelukast therapy was discontinued, initial treatment with low doses of prednisone yielded no relief; however, all of her symptoms, chest radiograph abnormalities, and eosinophilia resolved within a few days of treatment with prednisone, 80 mg/d.

CASE REPORT 4

A 63-year-old woman with chronic sinusitis and corticosteroid-dependent asthma presented in September 1998 with 1 month of worsening neuropathic pain and weakness in both lower extremities and in her right hand. Oral prednisone had been discontinued 6 months prior to presentation when she began to take fluticasone, 880 [micro]g/d, and montelukast daily. Physical examination was notable for bilateral wheezes and marked weakness in upper and lower extremities. Chest radiograph was without infiltrates, but her CBC was notable for a leukocytosis of 40 x [10.sup.9]/L with 55% eosinophils. EMG confirmed right median sensory and bilateral lower extremity sensorimotor involvement. Montelukast was discontinued, and the patient received high-dose IV methylprednisolone; eosinophilia and neurologic symptoms improved dramatically over the course of the next several days.

In addition to these patients, we recently encountered two women with sinusitis and severe allergic asthma that required frequent courses with systemic corticosteroids. After tapering their doses of oral corticosteroids, they developed CSS during treatment with salmeterol and high doses of the inhaled corticosteroid fluticasone (see Table 1). While neither subject had received leukotriene-modifying therapy for asthma, each developed eosinophilia, pulmonary infiltrates, and end-organ vasculitis (one patient had a rash and the other had neuropathy) after withdrawal of systemic corticosteroids. Each one improved with a combination of cyclophosphamide and reinstitution of systemic corticosteroid therapy.

DISCUSSION

First described by Churg and Strauss in 1951, CSS is an uncommon vasculitis of unknown etiology that is also known as allergic angiitis and granulomatosis.[8] In 1990, the ACR developed diagnostic criteria for the syndrome; it is a disease characterized by at least four of the following six features: (1) moderate to severe asthma, (2) peripheral blood eosinophilia ([is greater than] 10%), (3) mononeuropathy or polyneuropathy, (4) pulmonary infiltrates, (5) paranasal sinus abnormality and (6) extravascular eosinophils.[4]

In the cases of CSS that occurred with zafirlukast use, which we previously reported,[1] two patients had evidence of CSS prior to initiation of zafirlukast and all patients had been on high-dose corticosteroids at or near the time of syndrome onset. We thus proposed that zafirlukast provided such a salutary therapeutic effect that the physicians caring for these patients were able to withdraw systemic corticosteroid therapy, thereby unmasking the underlying CSS. This phenomenon has been referred to as the forme fruste of CSS.[9] A similar mechanism is postulated for the one case of CSS reported in the literature in association with yet another leukotriene receptor antagonist, pranlukast.[10]

In the above-reported cases of CSS in patients treated with montelukast for asthma, the syndrome was again manifest in association with a leukotriene-modifying drug, albeit one with a different biochemical structure. Each of the patients had a history of multiple asthma exacerbations that often required systemic corticosteroids for control. These patients had all received corticosteroids (either oral systemic or high-dose inhaled) that were being tapered at the time the syndrome became manifest, likely because of superior asthma control resulting from the addition of montelukast to their treatment regimen or because of the waxing and waning nature of the disorder. We propose that prior to their treatment with leukotriene modifiers, the high-dose inhaled corticosteroids used in these patients in conjunction with intermittent systemic corticosteroid treatment had suppressed manifestations of systemic eosinophilia. It is well established that inhaled corticosteroids have sufficient systemic absorption to account for such suppressive effects[11,12] (although these may not be able to suppress the manifestations as the disease progresses), and the forme fruste of CSS has been described in association with tapering of inhaled steroids.[8] Furthermore, each of these cases fell within the spectrum of the natural course of CSS, with "indolent allergic disease that evolves into asthma with multiple exacerbations that may progress to eosinophilia and finally, multiorgan eosinophilic vasculitis."[13] Hence, the findings in these patients, as well as those of another recent case report,[14] are compatible with the hypothesis that while a leukotriene-modifying agent could have directly caused CSS, we believe it is more probable that the drugs either unmasked, or were coincident with the progression of an underlying systemic eosinophilic disorder.

After 350,000 patient-years of exposure to montelukast, a case rate of ~60 cases of CSS that meet ACR criteria per million patient-years have been reported to Merck in association with this drug.[7] This incidence is strikingly similar to the incidence observed with zafirlukast (also ~60 cases/million asthmatics/yr) (personal communication; Catherine M. Bonuccelli, MD; December 16, 1998). As the incidence of CSS in the general population has previously been reported to be 2.4 to 3.3 cases/million/ yr,[15,16] it appears that these newly observed rates in association with leukotriene modifiers represent a large increase in the rate of CSS. However, a recent analysis of CSS prevalence by Martin et al[17] suggests that these background rates are not necessarily representative of CSS in the asthmatic population that this syndrome afflicts: CSS was observed in 64.4 patients/million patient-years in a cohort of subjects receiving nonleukotriene modifier asthma drugs; the prevalence of CSS was only 1.8/million patient-years of observation in the general population. This suggests that the incidence of CSS in an asthmatic population is both considerably higher than that previously reported for the general nonasthmatic population, and also very similar to the rates observed with these drugs.

In order to ascertain the background incidence of this syndrome in an asthma population prior to the release of leukotriene modifying drugs, we have preliminarily reviewed the information in the General Practice Research Database (GPRD), a large computerized medical care resource in the United Kingdom, in which physicians have enrolled [is greater than] 4 million UK residents since 1987. We applied the 1990 ACR CSS criteria to the computer records of [is greater than] 300,000 asthmatic subjects in the GPRD and identified an upper limit of approximately 43 possible cases of CSS over an 8-year period. From this review, we have derived a preliminary crude estimate of an incidence in this population of between 6 and 18 cases/million asthmatics/yr (Table 2); these results will be reported separately after a detailed review of available clinical records. While lower that that of the study by Martin et al,[17] population samples and study methodology differed, and there may have been underdiagnosis of CSS due to the lack of availability of novel asthma therapies that could have delayed institution of suppressive corticosteroids that may have masked CSS. Nevertheless, a higher-than-previously reported CSS incidence remains. Compared with these numbers, the apparent increase in case rates with leukotriene modifiers that we obtained implies a causative role of these medications in CSS.

(*) From personal communication with Catherine M. Bonuccelli, MD, December 16, 1998.

Indeed, an allergic response to therapy with these new drugs may be responsible for CSS in some of these asthmatic patients who are inherently predisposed to allergic diathesis by virtue of their eosinophilia at baseline. However, the facts that similar case rates were observed with two medications with distinct molecular structures, that similar rates were observed in the study by Martin et al,[17] and that no cases have been observed in steroid-naive individuals suggest an allergic reaction is unlikely to be the primary pathophysiology. For instance, given the fact that the relationship between leukotriene biology and eosinophils is still not well understood, one possible explanation is that blockade of the cysteinyl leukotriene-1 receptor by these agents may lead to vasculitis by allowing the presence of other agents. For example, the eosinophil chemoattractant, leukotriene B4, which is not affected by these leukotriene receptor antagonists, but which otherwise would have been inhibited by corticosteroids used to treat the bronchoconstriction resulting from activation of the cysteinyl leukotriene-1 receptor, could be responsible for causing vasculitis. However, the fact that cases of CSS have been reported to the US Food and Drug Administration in association with the 5-lipoxygeanse inhibitor, zileuton,[18] which also blocks leukotriene B4, makes this possibility less likely. Another hypothesis to account for the increased incidence of CSS is that there may just be increased reporting of cases of CSS in relationship to a new drug or due to increased awareness of the disease since our initial report with zafirlukast. Alternatively, previous reports may have underestimated the incidence due to underreporting--a possibility if cases of forme fruste CSS were diagnosed as severe asthma rather than CSS. Before the introduction of antileukotriene therapy, patients who developed progressive disease heralded by airway obstruction were perceived to have severe asthma and treated with higher doses of inhaled or oral corticosteroids. While eosinophils may produce bronchoconstrictor substances such as leukotrienes, with the availability of leukotriene modifiers, patients do not necessarily develop asthmatic symptoms that require corticosteroids because their airway obstruction is prevented by the leukotriene receptor antagonists. Thus, corticosteroid treatment is not begun, and other systemic eosinophilic manifestations of the disease (that would otherwise have been quelled by corticosteroids) may become manifest. In these cases it appears that blockade of leukotriene receptors modulates the asthmatic component of hypereosinophilia, but not the systemic effects that require the previously tapered corticosteroids for control.

Indeed, the unmasking of CSS is not exclusively observed with treatment by a leukotriene modifier; other systemic steroid-sparing medications that modulate airway obstruction, including beclomethasone, flunisolide, and cromolyn,[19] carry warnings that vasculitis or pulmonary infiltrates with eosinophilia may occur. Furthermore, the occurrence of multiple cases of CSS in association with fluticasone recently prompted its manufacturers to change its labeling.[20] Similarly, we now report two subjects with asthma who had previously required multiple corticosteroid courses for control; after tapering of systemic corticosteroid therapy to a combination of high-dose inhaled corticosteroids (fluticasone or budesonide) and the long-acting [Beta]-agonist salmeterol,[8] they developed a similar syndrome. While neither subject received a leukotriene modifier, each of the two subsequently developed pulmonary infiltrates and eosinophilia with end-organ vasculitic involvement (neuropathy or rash) compatible with CSS. In these cases as well, it appears that the use of novel corticosteroid-sparing agents to quell asthmatic symptoms has allowed for the unmasking or progression of a systemic eosinophilic syndrome that would otherwise be masked by systemic corticosteroids. While some patients who develop the syndrome may be on high doses of suppressive doses of inhaled corticosteroids, the natural course of the disease is that in some patients, despite even systemic corticosteroids, it may progress to the point that the patient requires cytotoxic agents.

For many patients, the leukotriene modifiers remain safe and effective medications in the armamentarium against asthma.[21] Although this type of analysis cannot establish or dismiss cause and effect, our data suggest that neither montelukast nor zafirlukast is likely to have a direct causative role in CSS pathogenesis in these patients, but rather either unmasked or coincided with the natural course of a progressive preexisting condition. While CSS is potentially life threatening, it remains a rare entity, both with and without these medications. While physicians must be especially vigilant for it in patients whose corticosteroids are tapered, the syndrome does not appear to be a direct result of antileukotriene treatment, and choice of asthma therapy need not be influenced based on the prevalence of this syndrome. The unmasking of a vasculitic syndrome previously recognized as asthma is possible in any patient receiving a medication that suppresses asthmatic airway obstruction by interfering with the pathways leading to that obstruction. As new medications with this potential are introduced into asthma treatment regimens, it is likely that additional cases of CSS will be recognized.

ACKNOWLEDGMENT: We would like to thank Michael R. Jordan, MD, of New England Medical Center, Boston, MA, and Mohammed Sand Anwar MD, MPH, of St. Francis Hospital and Medical Center, Hartford, CT, for their help in preparation of this article. We would like to thank Hershel lick and Susan Jick of the Boston Collaborative Drug Group, Lexington MA, for their help in analyzing the GPRD.

REFERENCES

[1] Wechsler ME, Garpestad E, Kocher O, et al. Pulmonary infiltrates, eosinophilia and cardiomyopathy in patients with asthma receiving zafirlukast, JAMA 1998; 279:455-457

[2] Katz RS, Papernik M. Zafirlukast, and Churg-Strauss syndrome [letter]. JAMA 1998; 279:1949

[3] Knoell DL, Lucas J, Allen JN. Churg-Strauss syndrome associated with zafirlukast. Chest 1998; 114:332-334

[4] Green RL, Vayonis AG. Churg-Strauss syndrome after zafirlukast in two patients not receiving systemic steroid treatment. Lancet 1999; 353:725-726

[5] Masi AT, Hunder GG, Lie JT, et al. The American College of Rheumatology 1990 criteria for the classification of the Churg-Strauss syndrome (allergic granulomatosis and angiitis). Arthritis Rheum 1990; 33:1094-1100

[6] NPAplus (National Prescription Audit) data provided by IMS Health, Westport, CT, October 1998

[7] Post marketing research data. Rahway, NJ: Merck, 1999

[8] Churg J, Strauss L. Allergic angiitis and periarteritis nodosa. Am J Pathol 1951; 27:277

[9] Churg A, Brallas M, Cronin SR, et al. Formes frustes of Churg-Strauss syndrome. Chest 1995; 108:320-323

[10] Kinoshita M, Shiraishi T, Koga T, et al. Churg-Strauss syndrome after corticosteroid withdrawal in an asthmatic patient treated with pranlukast. J Allergy Clin Immunol. 1999; 103:534-535

[11] Toogood JH. Complications of topical steroid therapy for asthma. Am Rev Respir Dis 1990; 141:S89-S96

[12] Barnes PJ, Pedersen S, Busse WW. Efficacy and safety of inhaled corticosteroids: new developments. Am J Respir Crit Care Med 1998; 157:S1-S53

[13] Lanham JG, Elkon KB, Pusey CD, et al. Systemic vasculitis with asthma and eosinophilia: a clinical approach to the Churg-Strauss syndrome. Medicine 1983; 63:65-81

[14] Franco J, Artes MJ. Pulmonary eosinophilia associated with montelukast. Thorax 1999; 54:558-560

[15] Watts RA, Carruthers DM, Scott DG. Epidemiology of systemic vasculitis: changing incidence or definition? Semin Arthritis Rheum 1995; 25:28-34

[16] Reid AJ, Harrison BD, Watts RA, et al. Churg-Strauss syndrome in a district hospital. Q J Med 1998; 91:219-229

[17] Martin RM, Wilton LV, Mann RD. Prevalence of Churg-Strauss Syndrome, vasculitis, eosinophilia and associated conditions: retrospective analysis of 58 prescription-event monitoring cohort studies. Pharmacoepidemiol Drug Safety 1999; 8:179-189

[18] Food and Drug Administration, Freedom of Information Act, data on file, 1999

[19] Physicians' Desk Reference. 52nd ed. Montvale, NJ: Medical Economics Company, 1998

[20] Glaxo Wellcome. Letter to health care professionals, January 1999

[21] Suissa S, Dennis R, Ernst P, et al. Effectiveness of the leukotriene receptor antagonist zafirlukast for mild to moderate asthma. Ann Intern Med 1997; 126:177-183

(*) From the Departments of Medicine, Brigham & Women's Hospital and Harvard Medical School (Drs. Wechsler, Westlake, and Drazen), New England Medical Center (Drs. Finn and Gunawardena), Boston, MA; Oregon Health Sciences University (Dr. Barker), St. Francis Hospital and Medical Center (Dr. Haranath), Hartford, CT; and University Hospital (Drs. Pauwels and Kips), Ghent, Belgium.

Dr. Drazen has been a consultant (within the past 5 years) for the following pharmaceutical companies: Abbott, Bayer, Eli Lilly, Forest Laboratories, Genetics Institute, Genome Therapeutics, GlaxoWellcome, Hoechst-Marion-Rousseau, Merck, Pfizer, Roche Bioscience, Schering Sepracor, and Zeneca. Dr. Drazen's laboratory receives support for asthma trials from Abbott, Astra, Genetics Institute, Immunologics, Merck, Millenium, Schering, Wyeth-Ayerst and Zeneca. Dr. Drazen holds no equity position in any of the entities above. Dr. Drazen serves on the scientific advisory board of Lifemasters Health Systems, Inspire Pharmaceuticals, and Aradigm Medical Products; in each of these, he holds equity positions. Dr. Pauwels has had financial affiliations and has served as a lecturer and as an occasional consultant with the following: Astra, Allmiral, Bayer, Boehringer-Ingelheim, Byk Gulden, Glaxo Wellcome, Hoechst Marion Roussel, Menarini, Merck Sharp & Dohme, Mitsubishi, Novartis, Pfizer, Pierre Fabre, Roche, Rhone Poulenc Rorer, SmithKline Beecham, Schering Plough, Zeneca, and 3M Riker. Dr. Kips has served as a lecturer for Astra, Boehringer Ingelheim, Glaxo Wellcome, Hoechst Marion Roussel, and Merck, Sharp & Dohme.

Manuscript received June 10,1999; revision accepted October 26,1999.

Correspondence to: Jeffrey M. Drazen, MD, Pulmonary and Critical Care Division, Department of Medicine, Brigham & Women's Hospital, 75 Francis St, Boston, MA 02115; e-mail: jdrazen@rics.bwh.harvard.edu

COPYRIGHT 2000 American College of Chest Physicians
COPYRIGHT 2000 Gale Group