Main

Hypereosinophilic syndrome (HES) is defined by the presence of hypereosinophilia and evidence of eosinophil-mediated clinical manifestations1. Clinical presentation is highly heterogeneous, with patients experiencing symptoms of varying severity, potentially affecting multiple organ systems2. These features of HES, alongside rarity, often lead to substantial delays in diagnosis and treatment2.

Myeloid-HES, also known as primary neoplastic HES, is characterized by clonal eosinophil proliferation1. Myeloid-HES is associated with a wide variety of molecular abnormalities, such as gene rearrangements involving tyrosine kinase receptors (typically the FIP1L1::PDGFRA fusion gene)1. In patients with rearrangements of the FIP1L1::PDGFRA fusion gene, the tyrosine kinase inhibitor imatinib has shown remarkable efficacy3. However, treatment options for patients with imatinib-insensitive HES, including those with lymphocytic HES (an indolent T cell lymphoproliferative disease) or idiopathic HES (mechanistic cause not yet identified), are limited4.

Treatment for patients with nonmyeloid HES typically involves the use of corticosteroids and/or immunosuppressive or cytotoxic therapies, which can reduce inflammation and control symptoms but often lead to adverse effects and may still result in suboptimal disease control5. Biologic therapies have expanded treatment options in HES4,6,7. Results from two clinical trials demonstrated that mepolizumab, a monoclonal antibody that inhibits eosinophil activation and differentiation by binding interleukin (IL)-58, reduced disease flares compared to placebo and had corticosteroid-sparing effects in patients with uncontrolled FIP1L1::PDGFRA-negative HES9,10. On the basis of these findings, mepolizumab (300 mg subcutaneous, every 4 weeks) was approved for the treatment of HES. While the approval of mepolizumab for HES has improved clinical outcomes9, treatment responses vary. Some patients do not achieve a clinical response, or respond only partially to mepolizumab, highlighting the need for alternative therapies4,6,7. A retrospective analysis of the off-label use of biologics in HES suggests that failure to respond to one biologic does not preclude a response to a different biologic11.

The central role of eosinophilic inflammation and eosinophil-mediated end organ damage in the pathophysiology of HES, as well as the efficacy of IL-5 inhibition with mepolizumab, suggests that a more direct eosinophil-depleting approach may also be beneficial2. Benralizumab is a humanized, afucosylated monoclonal antibody that binds to the α subunit of the IL-5 receptor expressed on eosinophils12. Benralizumab causes rapid, near-complete depletion of eosinophils in peripheral blood, bone marrow, airway sputum, skin, esophagus and gastrointestinal tissues13,14,15,16,17 via antibody-dependent cell-mediated cytotoxicity12. It has demonstrated clinical benefit and corticosteroid-sparing effects in severe eosinophilic asthma18,19,20,21 and eosinophilic granulomatosis with polyangiitis (EGPA)22 and is approved as an add-on maintenance treatment for patients aged 6 years and older with severe eosinophilic asthma, and for adults with EGPA. In a phase 2 study of patients with symptomatic, treatment-refractory FIP1L1::PDGFRA-negative HES, 90% of patients achieved a ≥50% reduction in absolute eosinophil counts (AECs) from baseline to week 12, and reduced tissue eosinophilia was observed in patients with severe disease14. Long-term follow-up from this study suggests that benralizumab remains effective and well tolerated23.

Here, we describe the results of a phase 3 trial that assessed the efficacy and safety of benralizumab compared to placebo in patients with FIP1L1::PDGFRA-negative HES.

Results

Patient disposition

The trial enrolled patients between 20 July 2020 and 13 November 2024, with the final patient visit for the double-blind period on 7 May 2025. Overall, 134 patients underwent randomization and 133 received ≥1 dose of treatment; 67 received benralizumab and 66 received placebo. In total, 97% (65/67) of benralizumab-treated patients and 92.5% (62/66) of placebo-treated patients completed the 24-week double-blind period (Fig. 1).

Fig. 1: Trial profile.
Fig. 1: Trial profile.
Full size image

The flow of patients through the study, from enrollment through to completion of the double-blind period. Q4W, every 4 weeks.

Baseline demographics and disease characteristics

The demographic and clinical characteristics of the patients at baseline were generally balanced between the treatment groups and representative of the targeted patient population with HES (Table 1 and Supplementary Table 2). Overall, 62% (82/133) of patients were female, 38% (51/133) were male and the median (range) age was 51.0 (14.0–87.0) years. Four adolescents were enrolled in the study (three received benralizumab and one received placebo). Most patients (75.2%, 100/133) had idiopathic HES and 12.0% (16/133) of patients had lymphocytic HES. The median (range) time since diagnosis was 1.9 (0.1–32.3) years, and patients had experienced a median (range) of 2 (0–12) HES flares in the 12 months before enrollment. Before oral corticosteroid (OCS)-responsiveness assessment and randomization, the median (range) AEC was 1,600 (1,000–25,150) cells µl−1. Around three-quarters of patients (76.7%, 102/133) were receiving background systemic OCS at enrollment, at a median (range) prednisone-equivalent dose of 5.0 (0.0–30.0) mg day−1.

Table 1 Baseline characteristics
Full size table

Primary endpoint

HES flare was observed in 13 patients (19.4%) receiving benralizumab and 28 patients (42.4%) receiving placebo. There was a delay in time to first HES flare in those receiving benralizumab versus placebo, with a statistically significant 65% reduction in risk (hazard ratio (HR) 0.35, 95% CI 0.18 to 0.69, P = 0.0024) (Fig. 2). Similar results were observed in a sensitivity analysis assessing whether changes in background therapy during the double-blind period impacted the time to first flare. Nine patients (five benralizumab, four placebo) with potentially relevant changes were censored at the point of medication change (Supplementary Fig. 2). Across predefined subgroups, the treatment effect of benralizumab versus placebo was consistent and favored benralizumab (Supplementary Fig. 3a).

Fig. 2: Time to first HES flare (primary endpoint).
Fig. 2: Time to first HES flare (primary endpoint).
Full size image

The time to first HES flare was analyzed using a stratified log-rank test (two-sided significance level of 0.05) adjusted for region. HR and 95% CIs were estimated using a Cox proportional hazards model with treatment group and region as covariates. For patients who did not experience a HES worsening/flare, the time to first HES worsening/flare was right censored at the end of the double-blind period corresponding to the earliest date of: the first benralizumab open-label dose, study day 183, the date of last contact and the data cutoff date. An HR <1 favors benralizumab. Flare was defined as HES clinical manifestations or laboratory abnormalities that resulted in an increase/burst of OCS ≥10 mg day−1 prednisolone equivalent for at least 2 days or an increase or addition of new cytotoxic and/or immunosuppressive therapy or hospitalization.

Key secondary endpoints

All endpoints met the prespecified hierarchical testing strategy (P < 0.05) and achieved statistical significance at the 1% level (Table 2). The proportion of patients who experienced a HES flare or withdrew from the study during the double-blind period was significantly lower in the benralizumab group compared to the placebo group: 22.4% (15/67) versus 45.5% (30/66), respectively (odds ratio (OR) 0.31, 95% CI 0.14 to 0.69, P = 0.0033) (Fig. 3a). This represented a 52% relative reduction in the proportion of patients experiencing a flare (rate ratio (RR) 0.48, 95% CI 0.29 to 0.80, P = 0.003). Significantly fewer HES flares occurred in those treated with benralizumab versus placebo (0.41 versus 1.23 flares per year, respectively), with a 66% reduction in the annualized rate of flares (RR 0.34, 95% CI 0.18 to 0.63, P = 0.0008) (Fig. 3b). A higher proportion of patients on benralizumab experienced no HES flare events during the 24-week double-blind period compared to placebo (80.6% (54/67) versus 57.6% (38/66), respectively). Fewer patients on benralizumab 13/67 (19.4%) experienced one flare, while 20/66 (30.3%) patients on placebo experienced one flare. No patients in the benralizumab group experienced more than one flare, while 8/66 (12.1%) in the placebo group had two flares.

Fig. 3: Key secondary endpoints.
Fig. 3: Key secondary endpoints.
Full size image

a, The proportion of patients with HES flares was analyzed using a logistic regression model with treatment group and region as covariates to estimate the OR and 95% CIs. This analysis included patients who withdrew from the study without having flares as an event; these patients were censored at the time of study withdrawal in the analysis of time to first flare. b, The number of HES flares (annualized rate) was analyzed using a negative binomial model with treatment group and region as covariates. The logarithm of follow-up time was used as an offset variable. Flare rates were estimated for each treatment group (error bars represent 95% CIs) and the RR and 95% CIs were calculated for benralizumab versus placebo. c, The time to first hematologic relapse (AEC ≥1,000 cells µl−1) was analyzed using a stratified log-rank test, adjusted for region. HR and 95% CIs were estimated using a Cox proportional hazards model with treatment group and region as covariates. Patients who did not experience hematologic relapse were right censored at the end of the double-blind period corresponding to the earliest date of: the first benralizumab open-label dose, study day 183, the date of last contact and the data cutoff date. d, The LS mean change from baseline to week 24 in PROMIS Fatigue scores was analyzed using a mixed model for repeated measures, with treatment group, baseline score, visit, region and treatment visit interaction as covariates. All available observations were included with no imputation for missing data, assuming missing data were missing at random. The standardized T-score range for Short Form 7a is from 29.4 to 83.2; higher scores indicate more severe fatigue. P values from week 4 to 20 are nominal. Hierarchical fixed sequence testing of key secondary endpoints was conducted in the order presented (a–d) at a two-sided significance level of 0.05. Error bars represent 95% CIs. Values of HR <1, OR <1, RR <1 and LS mean difference <0 favor benralizumab.

Table 2 Summary of efficacy outcomes
Full size table

There was a significant delay in the time to first hematologic relapse: benralizumab patients were 92% less likely to relapse at any given point during the double-blind period versus placebo (HR 0.08, 95% CI 0.03 to 0.20, P < 0.0001). A clear separation between the treatment groups was demonstrated as early as week 4 (Fig. 3c). Fatigue was significantly improved in the benralizumab group versus the placebo group. The least squares (LS) mean difference in Patient-Reported Outcomes Measurement Information System (PROMIS) Fatigue scores between groups was –4.72 (95% CI –7.64 to –1.80, P = 0.0017) at week 24. The difference between groups in fatigue was already evident by the first timepoint measured (week 4) (Fig. 3d).

Additional secondary endpoints

Following a protocol-specified OCS-responsiveness assessment, blood eosinophils were reduced at randomization in both placebo and benralizumab groups. After treatment, blood eosinophils reduced to near-complete depletion in the benralizumab group; however, in the placebo group, blood eosinophils increased and remained higher than the benralizumab group throughout the double-blind period (Supplementary Fig. 4).

Very few patients receiving benralizumab experienced hematologic relapse or withdrew from the study compared to those receiving placebo: 9.0% (6/67) versus 63.6% (42/66), respectively (OR 0.05, 95% CI 0.02 to 0.13, nominal P < 0.0001) (Supplementary Fig. 5).

Most patients on benralizumab sustained AEC <500 cells µl−1 for 24 weeks: 91.0% (61/67) compared to 12.1% (8/66) in the placebo group (OR 87.87, 95% CI 26.09 to 295.97, nominal P < 0.0001) (Supplementary Fig. 6).

Over the double-blind period, 25.4% (17/67) of patients receiving benralizumab versus 48.5% (32/66) receiving placebo required an increase in corticosteroid dose (OR 0.35, 95% CI 0.16 to 0.73, nominal P = 0.005) (Supplementary Fig. 7a). Fewer patients in the benralizumab versus placebo group required at least one corticosteroid increase during the double-blind period (Supplementary Fig. 7b). The LS mean cumulative OCS use during the 24-week double-blind period was lower in those on benralizumab than those on placebo (nominal P = 0.0213) (Supplementary Fig. 7c).

Improvements were observed in patients treated with benralizumab in both the physical and mental component summary scores of the Short Form-36 version 2 (SF-36v2) as early as week 12 (nominal P = 0.0104 and P = 0.0026, respectively) and were sustained to week 24 (nominal P = 0.0008 and P = 0.0852, respectively) (Supplementary Fig. 8).

Patient Global Impression of Severity (PGI-S) and Patient Global Impression of Change (PGI-C) data showed perceived improvements from baseline to week 24 among patients receiving benralizumab (Supplementary Figs. 9 and 10).

PK and immunogenicity

Median benralizumab serum concentrations were generally similar between week 16 and 24 (Supplementary Fig. 11). Although the number of anti-drug antibody (ADA)-positive patients was too low to draw firm conclusions, no consistent impact of ADA on efficacy, safety or pharmacokinetics (PK) was observed (Supplementary Table 3).

Safety

The proportion of patients experiencing any adverse event (AE) during the double-blind period was similar between treatment groups (benralizumab n = 43 (64.2%), placebo, n = 44 (66.7%)) (Table 3). The most common AEs were headache (11/67 (16.4%) benralizumab-treated patients and 5/66 (7.6%) placebo-treated patients), upper respiratory tract infection (5/67 (7.5%) and 5/66 (7.6%)) and coronavirus disease 2019 (4/67 (6.0%) and 4/66 (6.1%)). A full table of AEs is provided in Supplementary Table 4. Five (7.5%) benralizumab- and five (7.6%) placebo-treated patients experienced a serious AE, none of which were considered to be related to treatment. There was one death in the benralizumab group due to sepsis, which was not considered related to treatment by investigator assessment; please refer to the Supplementary Information for further details. Safety was consistent with the known safety profile of benralizumab.

Table 3 AEs during the double-blind treatment period
Full size table

Post hoc analyses

As multiple organ involvement is present in many patients with HES, additional analyses were performed to assess the efficacy of benralizumab among patients with involvement of each major organ system, primary (Supplementary Fig. 3a) or other (Supplementary Fig. 3b). Benralizumab reduced the risk of the first flare versus placebo, regardless of the organ system involved.

Discussion

In the phase 3 NATRON study, treatment with benralizumab delayed the time to first HES flare and resulted in a statistically significant 65% reduction in the risk of first HES flare compared to placebo. Fewer benralizumab-treated patients flared, and benralizumab therapy was also associated with a lower annualized flare rate and delayed time to first hematologic relapse compared to placebo. Benralizumab significantly improved fatigue, a key symptom of HES that negatively impacts health-related quality of life24,25, with improvements observed as early as 4 weeks and maintained to 24 weeks. Consistent with the known mechanism of action of benralizumab, as well as previous observations in small trials or cohorts of patients with HES11,14,26, durable near-complete depletion of blood eosinophils was observed in patients treated with benralizumab. Safety and tolerability results were consistent with the known safety profile for benralizumab, including in adolescents. Overall, the results of the NATRON study support the use of benralizumab as a therapy to manage patients with FIP1L1::PDGFRA-negative HES.

Strengths of the NATRON trial include the high retention rate as well as the inclusion of patients with varied HES disease subtypes and clinical manifestations (most commonly pulmonary, dermatologic and gastrointestinal involvement), which generally reflect disease presentation in the real-world patient population27. The results of the key secondary endpoints supported the primary observations. Importantly, the HES flare definition was related to clinical features without relying solely upon blood eosinophil counts, and included physical examination and assessment of symptoms, ensuring that the efficacy of benralizumab was not overestimated owing to its eosinophil-depleting mechanism of action, allowing for a more balanced comparison with placebo. The tapering of background therapy was not permitted during the double-blind period of the study, even when HES disease was well controlled. This enabled a clearer assessment of disease-related symptoms without confounding effects related to OCS withdrawal. The OCS-responsiveness assessment at screening served as a safety measure to ensure that flares could be managed clinically without the need for investigators to monitor eosinophil counts. Furthermore, it was evident that high-dose OCS bursts are not an effective therapeutic approach for HES, as eosinophil counts returned to elevated levels in the placebo group by week 4.

The NATRON efficacy results and trial population are similar to those observed in the phase 3 registrational trial for mepolizumab9. Although a standardized definition of severe HES is not yet available, the elevated eosinophil counts, frequency of flares in the prior 12 months and the number of organs involved at baseline collectively indicate that the characteristics of the patient population enrolled was suitable for evaluating the efficacy of a biologic in this trial. While there are no head-to-head studies in HES, switching to a different anti-IL-5/receptor biologic when a patient fails to respond to another can result in clinical improvement11; therefore, having both biologic options for HES will benefit some patients.

The OCS-sparing potential of benralizumab has been demonstrated in severe eosinophilic asthma and EGPA21,22. In the head-to-head phase 3 study comparing benralizumab and mepolizumab in EGPA, more patients receiving benralizumab during the double-blind period achieved complete discontinuation of OCS22. This observation was further supported by the results in the first year of the open-label extension (OLE), when the proportion of patients who discontinued OCS increased after switching from mepolizumab to benralizumab28.

The OCS-sparing ability of benralizumab has also been observed in phase 2 and small cohort studies in patients with HES9,11,14,26. Fewer benralizumab-treated patients in this study required increases in OCS doses; however, as patients were not permitted to taper their background OCS dose, no conclusions on the OCS-sparing ability of benralizumab can be drawn. The ongoing NATRON OLE will provide data on the long-term durability of efficacy of benralizumab in patients with HES, and will also provide insights on the ability of benralizumab treatment to reduce exposure to corticosteroids and cytotoxic/immunosuppressive therapies.

While the NATRON results provide important insights, there are some limitations to the study that should be acknowledged. First, the sample size was small owing to the rarity of the disease. Unfortunately, this limited the ability to draw conclusions on efficacy on the basis of subgroups, particularly in those with the L-HES subtype (in which no patients in the benralizumab group experienced a HES flare during the double-blind period). Although the NATRON trial population was broadly similar to that seen in the registrational mepolizumab trial9, the duration of HES since diagnosis (1.9 years) was shorter than that observed in mepolizumab trials (approximately 5.5 years). NATRON excluded patients with life-threatening HES and, therefore, the benefit of benralizumab in acute life-threatening situations requires further study. Given that the double-blind period of the study was limited to 24 weeks, long-term efficacy and safety outcomes will be assessed in the ongoing OLE.

In conclusion, these data reinforce the central role of eosinophils in the pathophysiology of HES and support targeting eosinophils as a therapeutic strategy. Benralizumab demonstrated efficacy and safety in the treatment of HES, with evidence of clinical benefit apparent at the earliest visits in the trial, thus offering perspectives of rapid disease control in patients with active disease manifestations. These results may therefore expand the potential therapeutic options for patients with this disease.

Methods

Study design

The NATRON study, a double-blind, 24-week, phase 3, randomized, placebo-controlled trial with an ongoing OLE, assessed the efficacy and safety of benralizumab in patients with HES (Supplementary Fig. 1). This study was conducted at 40 sites across 15 countries (Argentina, Austria, Belgium, China, Denmark, France, Germany, India, Israel, Japan, the Netherlands, Poland, South Korea, the UK and the USA).

All patients remained on stable background HES therapy during screening and the double-blind period; however, modifications were allowed if required for a HES flare (treatment intensification) or an AE (treatment de-escalation) thought to be due to background therapy. Stable background therapy included oral, topical, nasal or inhaled corticosteroids, immunosuppressive or cytotoxic agents, interferon-alpha (IFN-α), and other medications used to control HES and/or manage HES symptoms.

Patients who completed the 24-week double-blind period were eligible to enter an OLE where all patients received benralizumab. The end-of-study definition is described in the Supplementary Information.

The trial was conducted in accordance with the ethical principles of the Declaration of Helsinki and is consistent with the International Council for Harmonisation Good Clinical Practice guidelines, the applicable regulatory requirements and the AstraZeneca policy on bioethics. The protocol, protocol amendments and any other relevant documents were reviewed and approved by the Independent Ethics Committees/Institutional Review Boards listed in the Supplementary Information. All patients provided written informed consent. International Council for Harmonisation E6 Good Clinical Practice guidelines were followed. Patients enrolled in the study were not compensated for their participation; however, reasonable reimbursement of expenses incurred by the patients (for example, travel and parking) was provided if allowed by local regulations. This was stated clearly in the informed consent form.

An independent safety monitoring board, comprising two clinicians and a statistician, monitored overall patient safety. To inform study design and conduct, 60 patients with HES were surveyed online with support from the American Partnership for Eosinophilic Disorders patient advocacy group, Invitae and Covance Patient Engagement to provide insights on study participation, the placebo-controlled design, extended dosing, site visits and support strategies for long-term study engagement. The full protocol is publicly available at https://www.astrazenecaclinicaltrials.com/study/D3254C00001/. This trial is registered with ClinicalTrials.gov (NCT04191304) and the OLE is ongoing (closed to recruitment).

Patients were ≥12 years of age, had a diagnosis of HES (defined as history of persistent eosinophilia (>1,500 cells μl−1) without secondary cause on two examinations (≥1 month apart) and evidence of end-organ manifestations attributable to eosinophilia), were receiving documented stable HES therapy for ≥4 weeks before screening and were experiencing an HES flare at screening or had a history of ≥2 HES flares that required escalation in therapy within 12 months before screening.

Exclusion criteria were the presence of an FIP1L1::PDGFRA fusion tyrosine kinase gene rearrangement or other known imatinib-sensitive mutation; a confirmed diagnosis of EGPA or systemic mastocytosis; the presence of life-threatening HES complications, as judged by the investigator; a history of thrombotic complications, stroke or cardiac damage; a disease severity that made the patient inappropriate for inclusion; hypereosinophilia of unknown significance; a known, preexisting, clinically significant endocrine, autoimmune, metabolic, neurological, renal, gastrointestinal, hepatic, hematologic, respiratory or other systemic disorders not associated with HES that were uncontrolled with standard treatment and, in the opinion of the investigator, could have increased risk of patient safety, interfered with study outcomes or impaired completion of the study; a documented history of clinically significant cardiac damage; a known active liver disease at the time of study; a malignancy or history of malignancy within 5 years before screening; chronic or active infections requiring systemic treatment or clinically significant viral, bacterial or fungal infection within 4 weeks before visit 1; an untreated or inadequately treated helminth parasitic infection within 24 weeks before visit 1 without documented resolution; a known immunodeficiency disorder other than those attributable to OCS or HES-related therapy; a positive HIV test; any clinically significant abnormal findings on physical examination, vital signs, hematology or clinical chemistry during screening that, in the investigator’s opinion, could have posed a safety risk, influenced study results or impaired study completion; evidence of prior benralizumab treatment failure; treatment with injectable corticosteroids within 4 weeks before randomization; receipt of any investigational product within 30 days or five half-lives (whichever was longer) before visit 1 or concurrent participation in another interventional clinical study, excluding noninterventional registry or cohort studies; receipt of any marketed or investigational biologic within 4 months or five half-lives before informed consent, unless on stable background biologic therapy unlikely to interfere with safety or efficacy assessments; receipt of live attenuated vaccines within 30 days before visit 1; a history of hypersensitivity to any biologic therapy, corticosteroids or components of the investigational product; receipt of immunoglobulin or blood products within 30 days before visit 1; a known or suspected alcohol or substance abuse that could have interfered with protocol compliance; and pregnancy, breastfeeding or lactation at the time of the study. The full protocol can be accessed at https://www.astrazenecaclinicaltrials.com/study/D3254C00001/.

Following enrollment, eligible patients entered a 3-day screening period. To proceed to randomization, two criteria had to be met: first, an AEC ≥1,000 cells μl−1 at local laboratory testing on the date of enrollment, and second, a demonstration of corticosteroid responsiveness defined as an AEC <1,000 cells μl−1 after 2 days of OCS administration (1 mg kg−1 day−1 prednisone/prednisolone equivalent) given in addition to the patient’s background therapy for HES before randomization. This OCS-responsiveness assessment served as a safety measure to ensure that flares could be managed clinically without the need for investigators to monitor eosinophil counts, as eosinophil counts were blinded during the study. The OCS dose equivalency is shown in Supplementary Table 1.

Patients who met eligibility criteria at the end of the screening period were stratified by geographic region (North America, Europe, Asia and rest of the world) and HES flare status (active flare or historic flares at study entry). The inclusion of HES flare status as a stratification factor was introduced in June 2022; before this amendment, patients were required to be actively flaring at entry.

Randomization and masking

All patients were centrally assigned to a randomized study treatment using Interactive Web Response Systems (IWRS)/Interactive Voice Response Systems (IVRS). As patients became eligible for randomization, unique randomization codes were assigned sequentially in each stratum from a randomization list prepared by a computerized system provided on behalf of AstraZeneca. The randomization sequence was computer-generated centrally using a permuted block design with a fixed block size of 4 and stratified by geographic region (North America, Europe, Asia and rest of the world) and HES flare status at screening. Patients, sponsor, site staff and investigators were blinded to treatment allocation and to patients’ blood and biopsy leukocyte counts during the double-blind treatment period and up to week 4 of the OLE. All packaging and labeling ensured blinding for all sponsor and investigational site staff. The following personnel had access to the randomization list during the study: those generating the randomization list, personnel at the IWRS/IVRS company, AstraZeneca’s supply chain department, drug safety services representatives (data entry site case handlers), the bioanalytical laboratory performing the PK sample analysis, the independent statistical data analysis center, the unblinded programmer and the unblinded medical monitor. The IWRS/IVRS provided the investigators with the kit identification number to be allocated to the patient at the dispensing visit.

Procedures

Patients were randomly assigned 1:1 into the 24-week double-blind period where they received benralizumab 30 mg (accessorized prefilled syringe) subcutaneously every 4 weeks or a matching placebo, in addition to the patient’s background therapy for HES. The rationale for benralizumab dosing is described in the Supplementary Information.

In addition to screening and scheduled study visits every 4 weeks, patients were advised to contact the study site each time they thought their symptoms were worsening and attend the site for a flare visit assessment. If study treatment was discontinued early, patients attended a discontinuation visit 4 ± 1 weeks after the last dose.

HES flares were assessed by the investigator at all scheduled or flare visits. Vital signs and blood samples were collected at screening and all scheduled and flare site visits; for scheduled dosing visits, samples were taken before the administration of study treatment.

Patient-reported outcome assessments were completed by patients using an electronic device at study site visits before other study procedures. Patient-reported outcome assessments included: (1) PROMIS Fatigue short form 7a, measured at screening and scheduled study visits (that is, every 4 weeks); (2) SF-36v2 (acute recall), measured at screening, visit 6 (week 12) and visit 9 (week 24); (3) PGI-S, measured at screening and scheduled study visits (every 4 weeks); and (4) PGI-C, measured at visit 4 (week 4) and scheduled study visits (every 4 weeks).

PK and immunogenicity assessments were conducted at screening and pre-dose at scheduled visits 4 (week 4), 5 (week 8), 7 (week 16) and 9 (week 24). AEs included events reported between screening (visit 1) and last contact with the patient. Serious AEs included events recorded from written informed consent throughout the duration of the study.

Outcomes

The primary endpoint was time to first HES flare during the 24-week, double-blind treatment period. A flare was defined as HES clinical manifestation or laboratory abnormality resulting in an increase of OCS ≥10 mg day−1 prednisone equivalent for ≥2 days, or an increase or addition of a new cytotoxic and/or immunosuppressive therapy, or hospitalization. Flares were assessed by the investigator through complete or brief physical examinations, an investigator-led HES symptom interview, laboratory assessments and other routine safety assessments; please refer to the Supplementary Information for further details. If patients were unable to attend the study site for flare assessment, medical records were collected and an investigator-led HES symptoms interview was recommended. Time to first HES flare was calculated as the number of days from the date of randomization to the start date of the first flare event, plus 1 day. The start date of HES flare was defined as the first day of increased dose/burst of OCS, first day of any increase or addition of new cytotoxic and/or immunosuppressive therapy, or date of hospital admission, whichever occurred first.

Secondary endpoints that were multiplicity protected within the prespecified statistical testing hierarchy were defined as ‘key’. Key secondary endpoints were: (1) the proportion of patients with HES flares, with those who withdrew from the study without having experienced a flare considered as having had a flare event; (2) the annualized rate of HES flares, assessed over a maximum follow-up period of 24 weeks or, for patients lost to follow-up, the follow-up time was defined as the duration from randomization to the last timepoint at which flare status could be evaluated, with distinct flares defined as those with onset occurring ≥14 days after the resolution of the previous flare; (3) the time to first hematologic relapse (AEC ≥1,000 cells μl−1), calculated as the number of days from the date of randomization to the start date of first hematologic relapse plus 1 day; and (4) the change from baseline to week 24 in PROMIS Fatigue, with a standardized total score calculated for each visit over the double-blind period.

Other secondary endpoints included the proportion of patients with hematologic relapse (including those who withdrew from the study) during the double-blind period, the proportion of patients with AEC <500 cells µl−1 for 24 weeks, the proportion of patients requiring an increase in corticosteroid dose at any time during the double-blind period and other patient-reported outcomes (SF-36v2, PGI-S and PGI-C).

PK was assessed through benralizumab serum concentrations and immunogenicity assessed through ADA assays and neutralizing antibody testing, as previously described29. Safety was assessed through reporting of AEs, serious AEs, vital signs and clinical laboratory variables.

Statistical analysis

All efficacy endpoints, demographics and baseline characteristics were analyzed using the full analysis set, which included all randomized patients who received ≥1 dose of study treatment according to the intention-to-treat principle, irrespective of adherence to the protocol and continued trial participation.

The primary analysis data cutoff was to occur after 38 patients had a first HES flare event and all randomized patients had completed the 24-week double-blind period. It was estimated that approximately 38 first HES flare events during the double-blind period were required to detect a statistically significant difference between treatment groups at the two-sided 5% significance level with approximately 80% power if the true treatment effect is an HR of 0.389 (equivalent to 30% of patients receiving benralizumab experiencing an event by the end of the double-blind period versus 60% of patients receiving placebo). On the basis of these assumptions, a sample size of approximately 120 was expected, although recruitment could continue beyond this to provide confidence that sufficient events would be observed once complete follow-up was achieved. A sensitivity analysis was conducted to assess the impact of patients changing systemic background therapy before HES flares, by censoring any patients with systemic OCS or immunosuppressive therapy changes that were considered to have a potential to impact the chance of the patient flaring. Please refer to the Supplementary Information for further details on the sensitivity analysis. Subgroups were analyzed for the primary endpoint, and included sex, age, region (a stratification factor), race, HES subtype (based on investigator assignment), baseline blood eosinophil counts, baseline OCS dose, HES flare status (a stratification factor), primary organ involvement and time since HES diagnosis.

To account for multiplicity, if the primary endpoint was found to be statistically significant, key secondary endpoints were tested using a hierarchical fixed sequence approach at two-sided 0.05 in the following order: (1) the proportion of patients who experience a HES flare during the double-blind period; (2) the number of HES flares (annualized rate) during the double-blind period; (3) the time to first hematologic relapse during the double-blind period; and (4) the change from baseline in PROMIS Fatigue score at week 24.

Time-to-event endpoints were analyzed using a stratified log-rank test, adjusted for region. HRs and 95% CIs were estimated using a Cox proportional hazards model with treatment group and region as covariates. For patients who had not experienced the event, the time to event was censored at the end of the double-blind period corresponding to the date of the first of benralizumab open-label dose, study day 183, the date of last contact or the data cutoff date, whichever occurred first.

The proportion of patients with a HES flare was analyzed using a logistic regression model with treatment group and region as covariates to obtain the OR and 95% CIs (withdrawals were included as flare events in the primary analysis). The annualized rate of flares during the double-blind period was analyzed using a negative binomial model. The logarithm of the follow-up time was used as an offset variable in the model. The model included covariates of treatment group and region to obtain estimates of the flare rates in each treatment group and the rate ratio for benralizumab versus placebo.

For continuous change from baseline endpoints, including changes in PROMIS Fatigue standardized T-scores, data were analyzed using a mixed model for repeated measures, with treatment group, baseline score, visit, region and treatment visit interaction as covariates. No imputations were performed for missing data; all available observations were included in the analyses, which assumed missing data were missing at random. Safety was analyzed descriptively on the basis of the safety analysis set, which included all patients who received ≥1 dose of study treatment. The PK analysis set included all patients who received ≥1 dose of benralizumab with ≥1 quantifiable serum PK observation post-first dose. All data analyses were performed with SAS System (SAS Institute Inc.) software. Please refer to the Supplementary Information for additional details on statistical analyses.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.