Hypereosinophilic syndrome: approach to treatment in the era of precision medicine

A 50-year-old woman with a history of urticaria, sinusitis with nasal polyps, and eosinophilia presented with dyspnea and palpitations. Absolute eosinophil count (AEC) was 23 400/μL. Echocardiogram showed severe mitral regurgitation, marked bilateral atrial enlargement, and severe pulmonary hypertension. A bone marrow biopsy performed several years earlier was normocellular with increased eosinophils, but there was no evidence of myeloid neoplasia.

Since the first description of eosinophilic endomyocardial fibrosis by Loeffler in 1934,¹  the definition and classification of hypereosinophilic syndromes (HESs) have evolved considerably. In 1975, Chusid et al²  described 14 cases of “idiopathic” HES defined by the presence of an AEC ≥1500/μL for 6 months (or death before that time), the absence of a known secondary cause, and evidence of eosinophilic end organ manifestations. The patients in that study included individuals with “Churg’s eosinophilic granulomatosis” (or eosinophilic granulomatosis with polyangiitis [EGPA]), idiopathic eosinophilia with gastrointestinal involvement (eosinophilic gastrointestinal disease), and eosinophilic leukemia, clearly showing the heterogeneity of HES. In the 40 years since their report, more effective therapies have become available, making the 6-month requirement to make a diagnosis increasingly problematic. Moreover, the identification of recurrent molecular abnormalities and secondary causes driving the eosinophilic manifestations in some patients with HES, as defined by Chusid et al,²  has led to controversy regarding the definition and classification of this rare syndrome.

For the purposes of this review, HES will be defined according to a consensus definition developed by a multispecialty group of experts as (1) AEC ≥ 1500/μL and clinical manifestations attributable to eosinophilia or (2) tissue hypereosinophilia with blood eosinophilia (AEC above the upper limit of normal for the reference laboratory).^3-5  In contrast to the World Health Organization definition of HES,⁶  which excludes myeloid neoplasms associated with recurrent molecular abnormalities, the consensus definition includes PDGFRA-associated myeloid neoplasms, eosinophilic endomyocardial fibrosis in the setting of helminth infection (Loeffler’s endocarditis), and other HESs of known cause, because the clinical manifestations of these syndromes can be indistinguishable from those of idiopathic HES.

The same consensus group proposed dividing this heterogeneous group of patients with HES into clinically defined categories in an attempt to capture differences that might reflect the underlying etiology of the eosinophilia and/or predict treatment responses.⁴  The 6 proposed clinical variants are as follows: (1) myeloid hypereosinophilic syndrome (MHES)—HES with presumed or proven clonal eosinophilia, (2) lymphocytic variant hypereosinophilic syndrome (LHES)—HES with an aberrant or clonal lymphocyte population driving the eosinophilia, (3) overlap HES—eosinophilic disease restricted to a single organ system or defined eosinophilic syndromes that overlap in presentation with idiopathic HES (eg, EGPA), (4) associated HES—HES in the setting of a known secondary cause that requires treatment directed at the cause rather than the eosinophilia itself (eg, helminth infection, drug hypersensitivity, and immunodeficiency) (Table 1), (5) familial HES, and (6) idiopathic HES—HES that does not fall into any of the other categories. Although these categories continue to be refined as technical advances in diagnostic testing and the availability of novel targeted therapies augment our understanding of the underlying pathogenesis of HES, they have proven useful in predicting response to conventional and novel targeted therapies.^7-9 

The patient underwent emergency mitral valve replacement with a bovine bioprosthesis and tricuspid repair. The excised valve showed evidence of eosinophilic infiltration and fibrosis. Prednisone (1 mg/kg) was initiated for presumed idiopathic HES.

Both the acuity and nature of the clinical presentation are key in guiding the choice of initial therapy in a patient presenting with an AEC ≥ 1500/μL (Figure 1). Because systemic glucocorticoids (GCs) are the first-line agent for the treatment of most forms of HES, the identification of patients who require alternative therapies is of paramount importance in the initial evaluation. These include patients with (1) HES due to a secondary cause for which specific therapy is available (associated HES) (Table 1); (2) myeloid neoplasms associated with imatinib-sensitive mutations, including FIP1L1-PDGFRA, or known to be GC resistant; and (3) overlap syndromes that can be managed with topical agents, such as eosinophilic esophagitis. Equally important is the decision to treat or observe asymptomatic patients with AEC ≥ 1500/μL and no features of MHES or evidence of end organ manifestations (ie, hypereosinophilia of unknown significance), because biomarkers that predict disease progression are lacking.^10,11 

High-dose GCs (eg, prednisone at 1 mg/kg per day) remain the initial therapy of choice for most patients who present with life-threatening or potentially disabling manifestations of HES. In patients with potential exposure to Strongyloides at any time during their life, empiric ivermectin therapy (200 μg/kg orally daily for 2 days) should be given concomitantly to prevent potentially fatal GC-associated hyperinfection syndrome.¹²  Intravenous administration of GC should be considered for patients who are acutely ill and/or have gastrointestinal involvement that may impair absorption. Cyclophosphamide should be added if EGPA is suspected based on clinical features, including but not restricted to asthma, sinus disease, and pulmonary infiltrates, and criteria for poor prognosis are met.¹³ 

Patients with known imatinib-sensitive mutations, including FIP1L1-PDGFRA and translocations involving PDGFRB, should receive imatinib mesylate (100-400 mg daily) as first-line therapy. These patients are commonly GC resistant and have a poor prognosis without effective therapy. The response to imatinib therapy in PDGFR-associated myeloid neoplasms is typically rapid (within days) and complete. Because molecular testing can take several days, not all PDGFR mutations are detected using currently available testing, and false negative results can occur, imatinib should also be considered as first-line therapy for patients who present with clinical features of this diagnosis (eg, male gender, splenomegaly, elevated serum B12 and/or tryptase levels, and a hypercellular marrow with dysplastic eosinophils and mast cells) and unknown or negative PDGFR status.⁷  High-dose GCs should be added for the first few days if there is evidence of cardiac involvement (elevated serum troponin or echocardiographic abnormalities) to prevent myocardial necrosis, a rare complication of imatinib in this patient population.¹⁴ 

Although every attempt should be made to perform a limited diagnostic evaluation before the administration of GC, therapy should not be delayed for this purpose in the face of worsening signs and symptoms. More stable patients with presumed HES should undergo comprehensive evaluation before initiation of treatment. Initial diagnostic testing should focus on identification of the etiology of the eosinophilia, classification by clinical variant, and delineation of end organ involvement (Table 2). In an acutely ill patient, tests that are affected by GC therapy should be prioritized.

The patient’s eosinophilia continued to rise, peaking at 33 700/μL, despite high-dose prednisone therapy (200 mg by mouth daily), and she was hospitalized for congestive heart failure. Echocardiogram revealed stenosis of the bioprosthetic mitral valve. Hydroxyurea (1 g daily) and ultimately, cyclophosphamide (750 mg/m² monthly) were added. She improved after the addition of cyclophosphamide, with normalization of her eosinophil count. Hydroxyurea was discontinued, and prednisone was tapered. After her second dose of cyclophosphamide, she presented with atrial fibrillation and a left atrial clot and underwent mitral valve annuloplasty <4 months after her initial cardiac surgery. Pathology of the excised valve tissue again showed eosinophilic infiltration. After the second surgery, she was discharged on interferon-α (1 mU subcutaneously daily). One month later, her AEC rose above 1000/μL, and cyclophosphamide was restarted at 70 mg/m² monthly. Despite this, her AEC remained uncontrolled (1000-6000/μL).

The most common second-line agents for the treatment of GC-resistant, PDGFR mutation–negative HES are imatinib, hydroxyurea, interferon-α, methotrexate, and cyclosporine.^15,16  Although the recent Food and Drug Administration approval of 2 anti–interleukin-5 (anti–IL-5) antibodies (mepolizumab and reslizumab) and 1 antibody to IL-5 receptor-α (benralizumab) for eosinophilic asthma is beginning to change this paradigm, only mepolizumab is approved for a variant of HES (EGPA), and cost remains an important factor.

Which second-line agent to use in an individual patient is a complex decision that depends on a number of factors, including anticipated efficacy, side effect profile, and patient and physician preference (Table 3). A brief overview of the most commonly used second-line agents follows. Mepolizumab will be discussed in the next section.

Although clearly the drug of choice for PDGFR-associated myeloid neoplasms, the utility of imatinib in patients without a known imatinib-sensitive mutation remains controversial, with efficacy rates ranging from 9% to 60% depending on the series.^17-19  This is likely due in part to the heterogeneity of the patients included in these studies. In a recent prospective study of imatinib therapy in FIP1L1-PDGFRA–negative patients with GC-refractory HES, 50% of those with MHES (as defined by ≥4 features suggestive of a primary myeloid disorder) responded to imatinib in contrast to 0% of those without MHES.⁷  Some PDGFR-negative patients seem to require higher doses of imatinib and respond more slowly.¹⁸  Consequently, imatinib (400 mg daily for at least 4 weeks) is recommended. Patients experiencing a suboptimal or partial response should undergo repeat bone marrow examination, because unmasking of pre–B-cell acute lymphocytic leukemia has been reported. Although second generation tyrosine kinase inhibitors, including nilotinib,²⁰  dasatinib, and sorafenib, have shown efficacy in small numbers of patients with imatinib-resistant or -intolerant PDGFRA-associated disease, their effects in mutation-negative HES are unknown.

Hydroxyurea has been used for the treatment of GC-resistant HES since 1978²¹  and is effective in reducing eosinophil counts in a majority of patients, although durable and complete responses are relatively uncommon with monotherapy. Dose-related cytopenias are a major limiting factor.¹⁵  Low-dose hydroxyurea (500 mg daily) has been reported to potentiate the effects of interferon-α without increasing toxicity.²² 

First used for the treatment of HES resistant to GC and hydroxyurea, interferon-α has been used as a second-line agent for the treatment of HES for decades.¹⁶  In addition to its direct effect on eosinophils, interferon-α downregulates Th2 cytokine production, providing a theoretical advantage for the treatment of LHES. Cytogenetic remissions have also been reported in MHES. Unfortunately, HES is a chronic disease, and drug discontinuation due to the debilitating side effects of interferon-α is common.¹⁵  Pegylated interferon, at doses similar to those used for treatment of viral hepatitis, seems to be equally effective and slightly better tolerated.²³ 

Cyclosporine has effects on T-cell activation and cytokine production, and case reports have documented clinical response in patients with LHES, despite persistence of the aberrant clonal T-cell population. In a retrospective series of 188 patients, 11 (6%) were treated with cyclosporine (150-500 mg daily). Although 3 of 5 subjects who received cyclosporine monotherapy had a partial or complete response, the majority of patients (82%) discontinued cyclosporine due to toxicity.

Although there are little published data directly supporting the use of methotrexate as a GC-sparing agent in HES, it is commonly used in EGPA at relatively low doses (7.5-20 mg weekly) and seems to be as effective as cyclophosphamide in maintaining remission with markedly less toxicity.²⁴  Consequently, methotrexate has been considered an alternative second-line therapy in HES, particularly in patients with pulmonary symptoms or findings suggestive of EGPA. In view of the toxicity profile, which includes black box warnings for rare but potentially fatal toxicities and secondary lymphomas, it is likely that methotrexate use will soon be replaced by mepolizumab (and other targeted therapies) with efficacy in this subgroup of HES patients.

Her eosinophil count remained significantly elevated despite the addition of interferon-α. Approval was obtained for compassionate use mepolizumab (antibody to IL-5; 750 mg intravenously monthly). She resumed monthly cyclophosphamide with only partial suppression of her hypereosinophilia and debilitating side effects, including fatigue, malaise, nausea, and vomiting lasting for 3 to 5 days after each dose. She was subsequently enrolled on a placebo-controlled, double-blind clinical trial of benralizumab (afucosylated antibody to IL-5 receptor-α; 30 mg subcutaneously monthly). At the time of enrollment, her AEC was 6650/μL, despite monthly cyclophosphamide infusions. She showed a dramatic response with AEC 0/μL and no progression of cardiac disease on benralizumab monotherapy for >3 years.

Of the novel biologics recently approved for asthma, the humanized anti–IL-5 antibody mepolizumab (Nucala; GlaxoSmithKline) is the best studied in HES. A double-blind, placebo-controlled phase 2 trial in 85 PDGFRA-negative patients showed that monthly mepolizumab (750 mg intravenously) was safe and effective as a steroid-sparing agent in HES, including LHES.^25,26  Long-term safety and efficacy of this dose were confirmed in a second study.²⁷  More recently, mepolizumab (300 mg subcutaneously monthly) was approved for the treatment of EGPA after a double-blind, placebo-controlled phase 3 trial showed a significant reduction in flares with mepolizumab therapy.²⁸  A phase 3 study of mepolizumab (300 mg subcutaneously monthly) to prevent flares in HES is ongoing (NCT02836496). It should be noted that all of the above studies enrolled GC-responsive subjects and used a dose of mepolizumab that is higher than the dose approved for eosinophilic asthma (100 mg subcutaneously monthly). A recent study examining mepolizumab response in a cohort of patients with HES who received the drug as part of a multicenter compassionate use trial suggests that patients with treatment-resistant HES may be less likely to respond to mepolizumab than GC-responsive patients (43% vs 83% complete response, respectively).⁹  In the same analysis, clinical variant was a useful predictor of response to mepolizumab, with the highest and lowest response rates in patients with overlap HES and MHES, respectively. Although some trial participants were able to reduce mepolizumab dosing in this retrospective trial, 3 of 11 who attempted dose reduction were unsuccessful. Although the monoclonal anti–IL-5 antibody, reslizumab (Cinqair; Teva Pharmaceuticals), is likely to have a similar safety and efficacy profile as mepolizumab, data in HES are limited to a small pilot study in 4 patients.²⁹  Of note, reslizumab is dosed by weight and may provide a theoretical advantage in obese patients.

Benralizumab is an afucosylated antibody to IL-5 receptor-α that depletes eosinophils in the blood, bone marrow, and tissue through enhanced antibody-dependent cell-mediated cytotoxicity. Approved for use in eosinophilic asthma, a recently completed phase 2 trial in HES showed promising results,³⁰  and several additional trials in HES clinical variants are ongoing or planned (NCT03473977 and NCT03010436).

Other agents in clinical development include dexpramipexole (Knopp Inc.) and AK002 (Allakos Inc.). Dexpramipexole is an oral agent with an excellent safety profile that was developed for the treatment of amyotrophic lateral sclerosis and found serendipitously to lower peripheral eosinophil counts.³¹  In a recent pilot study, dexpramipexole (150 mg orally twice daily) was found to deplete eosinophils in the blood, bone marrow, and tissues of 4 of 10 patients with HES through induction of lineage-specific maturational arrest.³²  The drug was well tolerated, and additional studies are planned. AK002 is an afucosylated monoclonal antibody to Siglec-8, an inhibitory receptor expressed on eosinophils, basophils, and mast cells, that is expected to deplete eosinophils in a manner similar to benralizumab. A clinical trial of AK002 for eosinophilic gastritis is currently underway (NCT03496571). Other targeted agents, including JAK inhibitors³³  and alemtuzumab,³⁴  have been used in small numbers of patients with LHES with some efficacy, although side effects can be significant and long-term outcomes are lacking. A clinical trial of ruxolitinib for the treatment of GC-refractory HES is ongoing (NCT00044304).

HES is a heterogeneous group of disorders with varied etiologies, clinical manifestations, and prognoses. Although the recent explosion in targeted therapies with improved toxicity profiles has dramatically improved outcomes in some patients with HES, with the exception of imatinib-sensitive mutations, the factors predicting response to a given therapy are still largely unknown. Given the high cost of these new therapies and the challenges in designing and implementing the clinical trials necessary for regulatory approval (and insurance reimbursement) in rare diseases, selection of the right drug for an individual patient is becoming increasingly important. Recent data suggest that clinical variant may be useful in this regard.

This work was funded by the Division of Intramural Research, National Institute of Allergy and Infectious Diseases, National Institutes of Health.

Amy Klion, Human Eosinophil Section, Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, 4 Memorial Dr, Bethesda, MD 20892; e-mail: aklion@niaid.nih.gov.