Treatment-refractory idiopathic hypereosinophilic syndrome: Pitfalls and progress with use of novel drugs


  • Conflict of interest: Nothing to report

  • A physician or group of physicians considers presentation and evolution of a real clinical case, reacting to clinical information and data (boldface type). This is followed by a discussion/commentary.

A 48-year-old Caucasian female presented with throat pain, malaise, and arthralgias, for which she received a course of amoxicillin. Her myalgias worsened and a complete blood count revealed a white blood cell count of 33 × 109/L with 51% eosinophils, 43% neutrophils, 5% lymphocytes, and 1% monocytes. The hemoglobin level was 14.2 g/dL and the platelet count was 190 × 109/L. A working diagnosis of drug (amoxicillin)-related eosinophilia was made, and she was started on prednisone 40 mg daily.

Eosinophilia (absolute eosinophil count [AEC] > 0.5 × 109/L) and severe eosinophilia (arbitrarily defined as AEC > 1.5 × 109/L) are associated with a variety of disorders. Initial evaluation of eosinophilia should focus on distinguishing secondary from primary (i.e., clonal or idiopathic) causes, and a detailed medical history is indispensable in this regard [1]. Common secondary causes of eosinophilia include parasitic infections, atopic disorders, drug reactions, granulomatous disorders, and malignancy (Hodgkin lymphoma, peripheral T-cell lymphoma, acute lymphoblastic lymphoma/leukemia, etc.) [2].

The patient's malaise and myalgias persisted, and her white blood cell count increased to 80 × 109/L with 72% eosinophils. She had no significant travel history or exposure to animals or hazardous chemicals. Her physical exam was remarkable only for mild tenderness upon palpation of muscles in all extremities.

A useful diagnostic algorithm for assessing a patient with primary eosinophilia has been described by Tefferi, et al. [1] The first step is to screen for presence of FIP1L1-PDGFRA and BCR-ABL1 mutations in peripheral blood (PB) using either fluorescence in situ hybridization (FISH) or reverse-transcriptase polymerase chain reaction (RT-PCR). Presence of BCR-ABL1, which signifies a diagnosis of chronic myeloid leukemia (CML), should always be excluded in this setting. If FIP1L1-PDGFRA testing is positive, a working diagnosis of the World Health Organization (WHO)-defined entity “Myeloid and lymphoid neoplasms with eosinophilia and abnormalities of PDGFRA, PDGFRB or FGFR1” is reached, even without a bone marrow (BM) examination [3]. If FIP1L1-PDGFRA testing is negative, a BM examination should be performed; cytogenetic studies will identify presence of rare rearrangements such as 5q33 (PDGFRB), 4q12 (PDGFRA rearrangement not involving FIP1L1), or 8p11.2 (FGFR 1). A BCR-ABL1-negative myeloproliferative neoplasm (MPN) with prominent eosinophilia and a PDGFRA or PDGFRB rearrangement represents an imatinib-sensitive disease, whereas FGFR1 translocations are primarily found in aggressive myeloid or lymphoid hematological malignancies characterized by refractoriness to conventional drug therapy. BM morphological examination is important to exclude presence of a well-defined myeloid malignancy with associated eosinophilia such as systemic mast cell disease, chronic myelomonocytic lymphoma, myelodysplastic syndrome, or acute myeloid leukemia. In the setting of primary eosinophilia and absence of the aforementioned molecular abnormalities or distinct myeloid neoplasm, the presence of >2 or 5% blasts in PB or BM, respectively, or presence of another clonal cytogenetic or molecular abnormality satisfies the diagnosis of “chronic eosinophilic leukemia-not otherwise specified (CEL-NOS).” Finally, T-cell receptor gene rearrangement and T-cell immunophenotyping studies should be performed to assess for an occult, cytokine-producing, clonal T-cell population, consistent with a diagnosis of “lymphocytic variant” hypereosinophilia. Only if all testing for a clonal process is negative, the eosinophilia is persistent for at least 6 months, and no cause for a secondary eosinophilia is discovered, can a patient be diagnosed as having idiopathic hypereosinophilia (IH). When such a patient is documented to have developed end-organ damage as a direct result of hypereosinophilia, a diagnosis of idiopathic hypereosinophilic syndrome (IHES) is appropriate. Screening tests for end-organ damage, including chest X-ray, electrocardiogram (ECG), echocardiogram, and serum troponin level are indicated in all patients with IH, with further organ-specific evaluation guided by features of each patient's presentation [1].

A BM aspirate and biopsy revealed a hypercellular (90%) marrow with markedly increased mature eosinophils and no abnormal mast cell population or other myeloid neoplasm. Lymphocyte immunophenotyping and T-cell receptor gene rearrangement studies were unremarkable and the BM karyotype was normal. Testing for KITD816V, BCR-ABL1 and FIP1L1-PDGFRA was normal. Tryptase was within normal limits at 4.49 ng mL−1 (<11.5 ng mL−1). Serial stool ova and parasite evaluations and serologies for a number of potential infectious etiologies were negative. Further evaluation included a negative chest X-ray, normal creatine kinase (24 U/L), an ECG which revealed non-specific lateral T-wave inversions and an elevated troponin T (0.25 ng/mL). Echocardiogram revealed normal left ventricular size and ejection fraction (60%) and a mildly thickened, mildly regurgitant mitral valve. Treatment was started with 40 mg dexamethasone daily, which led to an improvement in symptoms and a decrease in AEC from 46 × 109/L to 9.6 × 109/L after 3 days of therapy.

Diagnostic criteria for IHES were initially proposed by Chusid et al. in 1975. [4] Several modifications to these criteria have been proposed, including loosening the requirements related to duration and degree of eosinophilia in the setting of documented end organ disease [5]. Given the evidence of cardiac involvement, our patient met these revised criteria for IHES and higher dose steroids were initiated while the evaluation continued.

Steroids are commonly used as initial therapy for IHES given their ability to rapidly decrease the eosinophil count and stabilize organ function, possibly through suppression of inflammatory cytokines [6]. An unusual feature of this case was the partial, transient control of the eosinophil count despite use of high-dose corticosteroids.

Given her lack of symptoms consistent with acute coronary syndrome and elevated troponin T, eosinophilic myocarditis was suspected. The patient underwent cardiac magnetic resonance imaging (CMRI), which demonstrated diffusely increased left ventricular wall thickness and increased T2 signal within the subendocardial aspect of the myocardium, suggestive of myocarditis in the acute necrotic stage. A thin rim of nonenhancing epicardial/subepicardial tissue was noted in the left ventricle consistent with mural thrombus or fibrosis. The posterior mitral valve appeared to be tethered by the infiltrative material, leading to mitral regurgitation. Repeat CBC demonstrated an increase in the AEC to 24.8 × 109/L. Given these findings, the patient was hospitalized for observation and further management.

Upon admission, anticoagulation with intravenous heparin was initiated. Because of the worsening eosinophilia, the dose of dexamethasone was increased and she was given 100 mcg of pegylated interferon alfa-2a (Peg-IFNα) subcutaneously (SQ). Three days following admission, she developed acute onset weakness of the right hand and right sided neglect. Subsequent imaging revealed extensive acute to subacute infarcts scattered throughout bilateral cerebral and cerebellar hemispheres. Her AEC had decreased to 8.8 × 109/L, but given the stroke, treatment was escalated to 1 g methylprednisolone daily and alemtuzumab 30 mg SQ thrice weekly. Antimicrobial prophylaxis with valacyclovir and pentamidine was started (sulfa allergy was reported).

Cardiac involvement in IHES represents a potentially life-threatening complication [6, 7]. Echocardiography is often unremarkable in the acute phase of eosinophilic myocarditis [8], and CMRI can be useful in detecting the disease at an early stage [9]. Endomyocardial biopsy is often useful in confirming a diagnosis of CM and was considered in this case, but was deferred given the high associated risk in the setting of recent cardioembolic strokes and the consequent ongoing need for anticoagulation. Intracardiac thrombosis is a frequent finding in eosinophilic myocarditis and is a significant concern, as up to 25% of patients suffer an embolic event [8, 10]. The primary goal of treatment in this setting is control of the underlying eosinophilic process, with anticoagulation likely playing an adjunctive role. Data regarding anticoagulation is scant and its use should be limited to those with documented intracardiac thrombosis or thromboembolic event, as its efficacy in preventing such complications is unknown [10]. Arrhythmia, often life threatening, is another potential complication [11] and prophylactic measures to prevent sudden cardiac death can be considered.

Given the rapidly progressive disease with uncontrolled eosinophilia, a more aggressive treatment strategy was deemed necessary and thus the patient was started on alemtuzumab with high dose corticosteroids. Alemtuzumab is a humanized IgG1k anti-CD52 monoclonal antibody and eosinophils have been shown to express CD52 [12]. The data supporting alemtuzumab use in IHES are limited. In one study [13], 10 of 11 patients with refractory IHES treated with alemtuzumab achieved a complete hematologic response (CHR), defined as normalization of AEC; the median time to CHR was 2 weeks (range 0.5–5 weeks). Consequently, alemtuzumab is an attractive option in cases such as ours when rapid decrease in the AEC is indicated.

The patient's AEC decreased to within the normal range within 3 days of starting alemtuzumab, and the remainder of the hospital course was uncomplicated. She was dismissed with an external defibrillator for 1 month given the risk of sudden cardiac death.

The patient received 12 doses of alemtuzumab over 4 weeks without complication; steroids were also tapered off over this period. Circulating eosinophils were undetectable by the end of this course, and given this robust response, monthly “maintenance” alemtuzumab was planned. Prior to receiving the first such treatment, she developed fevers, fatigue and a nonproductive cough and was subsequently diagnosed with Pneumocytis jiroveci pneumonia (PcP). Alemtuzumab was discontinued and steroids reinitiated as part of PcP therapy. She completed a 21 day course of atovaquone followed by desensitization and initiation of prophylaxis with trimethoprim-sulfamethoxazole. The AEC slowly rose to 0.3 × 109/L over the 3 months following alemtuzumab discontinuation.

This case demonstrates one of the complications of profound immunosuppression related to alemtuzumab therapy; other potential complications include CMV reactivation and iatrogenic immunodeficiency-associated lymphoproliferative disorders [14, 15]. The risk:benefit ratio of alemtuzumab retreatment in this setting is unclear given the small number of published cases; in one series [13], only 1 of 10 patients attaining CHR with alemtuzumab maintained this response, and retreatment with alemtuzumab after relapse was attempted in a number of cases with varying results. Given the severity of our patient's IHES manifestations and prior efficacy of alemtuzumab, we opted to retreat with alemtuzumab despite the infectious complication.

Prior to restarting alemtuzumab, CMRI revealed complete regression of the previously reported subendocardial process [16]. She received 2 monthly doses of alemtuzumab, which led to stabilization of the AEC within the normal range. Given the improvement in the CMRI findings as well as ongoing concerns regarding risks of long-term alemtuzumab therapy, alemtuzumab was discontinued and treatment with Peg-IFNα at a dose of 180 mcg SQ weekly was started.

After a thorough discussion with the patient and other members of the health care team, a consensus decision to change treatment to Peg-IFN-α was reached. IFN-α is a commonly utilized therapy for IHES, with a subset of patients achieving objective hematologic and end-organ responses. [17] Pegylated forms of IFN-α have been utilized successfully [18], but whether such formulations are more efficacious or better tolerated for treatment of primary eosinophilia than nonpegylated forms remains to be established. Long-term use of IFN-α is challenging given the many potential toxicities, which have been well described [19]. One strategy we have used is addition of low-dose prednisone (30 mg for the first month, 15 mg for the second month, then discontinue) to IFN-α at the time of starting treatment as a means of improving tolerability and possibly efficacy.

The patient initially tolerated Peg IFN-α well, but soon required several step-wise dose reductions due to toxicities including myalgias, transaminitis, hypothyroidism, and a worsening of underlying depressive symptoms. Over the course of 6 months, the dose was decreased to 45 mcg SQ weekly. Over this period, the AEC slowly increased to a peak of 0.85 × 109/L, at which point prednisone was reintroduced, which normalized the AEC. Despite the modestly increased AEC, there was no evidence of recurrent cardiac involvement or other end organ damage, and surveillance with monthly troponin T values and CMRI every 6 months was pursued without event. Given the significant toxicity profile, Peg IFN-α was discontinued and mepolizumab (750 mg intravenous once monthly) was started. The pretreatment interleukin-5 level was normal (<7.8 pg/mL).

Our patient's experience with Peg IFN-α is not atypical; despite its efficacy, treatment could not be continued at a dose intensity that maintained the AEC within the normal range, which was deemed an important therapeutic end point. Fortunately, the AEC never rose above 1 × 109/L and there was no evidence of recurrent cardiac involvement, ascertained through use of several surveillance methods. This illustrates the difficulty in identifying a discrete, predictable AEC threshold above which end-organ damage is evident; anecdotally, it appears this threshold varies from patient-to-patient.

Mepolizumab is a fully humanized antihuman interleukin-5 (IL-5) monoclonal antibody which inhibits interaction between IL-5 and IL-5 receptors on the surface of eosinophils, and early studies using this agent have shown promise in treating IHES patients [20]. The largest such trial enrolled 85 patients with FIP1L1-PDGFRA-negative IHES in a randomized, double-blind, placebo-controlled fashion [21]. These patients had lead-in treatment with corticosteroids to normalize the AEC and were subsequently randomized to either mepolizumzb (750 mg) or placebo every 4 weeks for a total of 36 weeks, with concurrent tapering of corticosteroids based on the AEC. The primary endpoint was a reduction in prednisone dose to <10 mg daily for a period of at least 8 weeks. Patients receiving mepolizumab were significantly more likely to reach this endpoint (83% vs. 43%; hazard ratio 2.90; P < 0.001) without increased risk of adverse events compared to placebo. Mepolizumab is currently not approved by the FDA for HES treatment and our patient availed of it via a compassionate use program supported by the manufacturer [22].

The patient has tolerated mepolizumab therapy extremely well and the Peg IFN-α-related toxicities improved within a month of its discontinuation. Her AEC has consistently been <0.2 × 109/L in the 6 months since starting mepolizumab. The lymphopenia associated with both alemtuzumab and Peg IFN-α has resolved and all infectious prophylaxis has been discontinued without complication. She is currently 21 months from her diagnosis with excellent control of her disease and without any obvious mepolizumab-related toxicities.


IHES is a clinical syndrome rather than a distinct, molecularly defined disease entity. Given its rarity, heterogeneity, lack of understanding regarding pathogenetic mechanisms, and toxicity of available therapies, it remains a difficult condition to treat. There is no established standard of care for IHES treatment, but as this case illustrates, there is an emerging role for novel therapies such as alemtuzumab and mepolizumab. In general, in cases of confirmed IHES, the decision to initiate treatment hinges more on the presence of symptoms and/or end-organ damage than the AEC. If a decision is made to follow an asymptomatic patient with significant eosinophilia off of treatment, one must ensure a comprehensive plan of periodic evaluations to assess for development of end organ damage (troponin T, echocardiogram, etc.), in which case therapeutic intervention may be warranted.

In patients who do require treatment due to end organ damage or a markedly elevated eosinophil count, steroids remain a cornerstone of therapy, with steroid sparing therapy indicated when >10 mg of prednisone daily is required to maintain control of IHES-related symptoms. Clonal eosinophilias with evidence of PDFGRA or PDGFRB rearrangements are almost universally responsive to imatinib therapy; in FIP1L1-PDGFRA mutated cases, treatment is generally initiated at a low dose (100 mg daily) [23]. Imatinib has also been used in IHES treatment [24], typically at higher doses (400–800 mg daily); here, in general, the depth and quality of responses are inferior as compared to those seen in patients harboring discrete imatinib-sensitive molecular abnormalities. In addition to the aforementioned therapies, drugs such as cyclosporine, hydroxyurea, and cladribine also have a role in IHES treatment, often in combination with steroids; we refer the reader to appropriate reviews for a comprehensive discussion of this subject matter [1, 6, 25]. As knowledge regarding the molecular pathogenesis of hypereosinophilia continues to increase, new therapeutic targets are likely to emerge, allowing for individualized therapy based on the underlying genetic alterations.