Efficacy of benralizumab for patients with severe eosinophilic asthma: a retrospective, real-life study

From July 2018 to December 2019, 24 Japanese patients with SEA received benralizumab injections (30 mg every 4 weeks for the first three injections and every 8 weeks for the subsequent injections) at Jikei University Hospital, Tokyo, Japan. All asthma patients were diagnosed by respiratory physicians based on Japanese guidelines [13] or the Global Initiative of Asthma (GINA) guidelines [4]. Severe asthma was defined as requiring a high dose of inhaled corticosteroids (ICSs) plus at least one of the following additional control measures: long-acting β-2 agonists (LABAs), long-acting muscarinic antagonists (LAMAs), leukotriene receptor antagonists (LTRAs), a xanthine derivative and a daily OCS [3, 4, 13]. Xanthine derivatives are routinely prescribed in Japan because the Japanese guideline for the treatment of adult asthma recommends the use of xanthine derivatives in the setting of insufficient control with conventional control measures. SEA was defined as severe asthma with eosinophilic airway inflammation, which was defined as a peripheral blood eosinophil count ≥300/μl.

The present study was approved by the Ethics Committee of Jikei University [30–319 (9340)]. Based on the ethical guidelines of Jikei University, informed consent was not necessary for this retrospective study, and we performed opt-out consent on the website of our hospital. The benralizumab prescription rule was based on the Pharmaceuticals and Medical Devices Agency in Japan. The inclusion criteria for initiating benralizumab treatment in patients were as follows: 1. the patient had at least two exacerbations requiring OCS treatment in the previous year or before the introduction of mepolizumab; 2. if the patient did not receive mepolizumab treatment, the patient’s blood eosinophil count was at least 150/μl at baseline or ≥ 300/μl in the previous year; or 3. the patient received OCS maintenance therapy or another biologic (omalizumab or mepolizumab) regardless of the number of exacerbations or the blood eosinophil count.

We retrospectively examined the following characteristics: sex, age, comorbidities of eosinophilic diseases, smoking status, body mass index (BMI), baseline treatments including biologics, and the duration of asthma. We examined and evaluated the following parameters at baseline and at the final follow-up after 4 months: peripheral blood eosinophil and basophil counts, serum IgE, fractional exhaled nitric oxide (FeNO), the Asthma Control Test (ACT) score, pulmonary function test results [the forced vital capacity (FVC), forced expiratory volume in 1 s (FEV₁), FEV₁/FVC, %FEV₁ and % peak expiratory flow (%PEF)], and daily OCS maintenance doses as prednisone equivalents (mg). A change of 200 ml or greater was adopted as a significant change in the FEV₁ [12, 14‐16]. The FeNO level was measured using a NIOX VERO™ device (Aerocrine AB, Stockholm, Sweden) with a 50 ml/s flow rate according to the American Thoracic Society/European Respiratory Society recommendations [17]. The number of annual exacerbations of asthma symptoms requiring systemic CS was defined as the total number of exacerbations × 12/the total duration of the observation period (months). Furthermore, we evaluated changes from baseline in these parameters. To evaluate clinical efficacy, we utilized the ACT score and the Global Evaluation of Treatment Effectiveness (GETE) score; the ACT score is clinically useful as a simple scoring system, and scores of 20–25 are classified as well-controlled asthma [4]. The minimal clinically important difference (MCID) was an ACT score of three points [18]. The GETE score was evaluated based on symptom severity, medication use and pulmonary function tests at the final follow-up after at least 4 months of benralizumab treatment [19]. The GETE score has five classifications: excellent, good, moderate, poor and worsening. A responder is defined as a patient with a good/excellent response when treated with benralizumab. The GETE score after mepolizumab treatment was used if the overall evaluation did not change for the patients switched from mepolizumab.

The primary endpoint was the GETE score. We performed a subgroup analysis based on previous mepolizumab treatment with eosinophilic chronic rhinosinusitis (ECRS) or chronic rhinosinusitis with nasal polyps (CRSwNP) and with aspirin-exacerbated respiratory disease (AERD) because these factors affect the evaluation of clinical efficacy. The secondary endpoints included each parameter, each biomarker, changes in the number of asthma exacerbations and in daily CS doses, and the proportion of patients with a change in FEV₁ from baseline (≥ 200 ml). To examine sinusitis symptoms and findings, we utilized nasal discharge, nasal congestion and olfactory loss reported in the medical records.

All statistical analyses were performed using StatView version 5 (SAS Institute, Inc., Cary, NC, USA), and a post-hoc power analysis was performed with EZR (version 1.37, Saitama Medical Center, Jichi Medical University, Saitama, Japan) [20], which is a graphical user interface for R (version 3.4.1, The R Foundation for Statistical Computing, Vienna, Austria). All values are expressed as the mean ± standard deviation (SD). A p-value < 0.05 was considered statistically significant. To determine the characteristics of the responders, we statistically analyzed univariate models of patient characteristics, such as sex, age, BMI, peripheral blood eosinophil and basophil counts, previous mepolizumab treatment, and comorbidities. The factors associated with patient characteristics were analyzed using the Mann-Whitney U test, Fisher’s exact test, or the Wilcoxon signed-rank test (univariate model). Because the number of patients in the present study was small, we re-evaluated the clinical parameters, the percentage change in the number of annual exacerbations and maintenance CS doses with a post hoc power analysis (α-error < 0.05, cut-off 0.80). Furthermore, logistic regression analysis was performed to evaluate the identified characteristics of the responders (multivariate model), including the peripheral blood eosinophil count (≥ 300 /μl), ECRS or CRSwNP [11] as a comorbidity and other variables that achieved p < 0.20 in the univariate model.

Nine men and 15 women received benralizumab treatment with eight doses (median). The characteristics of the 24 patients are shown in Table 1. Among them, 11 had directly switched to benralizumab from mepolizumab treatment based on asthma symptoms (n = 2), the interval between hospital visits (n = 6) and discussions with physicians (n = 3) (Additional file 1). No significant differences in patient characteristics were found between the patients with and without previous mepolizumab treatment except for eosinophilic otitis media, which was frequently present in patients with previous mepolizumab treatment. In addition, we assessed the patients’ characteristics in the following subgroups (Additional file 2): males, females, AERD (−) and AERD (+). We showed that the %FEV₁, %PEF and ACT score were significantly lower in males than in females using post-hoc power analysis. In addition, the AERD (+) group showed significantly lower ACT scores and tended to have a lower %FEV₁ at baseline than the AERD (−) group (Additional file 2).

The changes in clinical parameters and biomarkers are shown in Table 2. The peripheral blood eosinophil and basophil counts significantly decreased. However, no significant differences in the changes in pulmonary function, FeNO, the ACT score, the number of annual exacerbations or maintenance CS doses from baseline were found between all patients or in the presence or absence of previous mepolizumab treatment. The %FVC, %FEV₁ and FEV₁ were reevaluated using post-hoc power analysis, and no significant difference was found before and after benralizumab treatment regardless of previous mepolizumab treatment. The ACT score tended to increase in all patients (n = 24) and in the group without previous mepolizumab (n = 13) treatment without significant differences (Table 2). Furthermore, the total number of patients with a final ACT score = 25 (well-controlled asthma) or an ACT score increase ≥3 (significant change) was 11 (85%) among the patients without previous mepolizumab treatment (n = 13) (data not shown). We determined the %FEV₁ and ACT score before and after benralizumab treatment in the AERD (−) and (+) groups (Additional file 3). Although the %FEV₁ tended to improve in the AERD (−) group, no significant difference in the %FEV₁ was observed before and after benralizumab treatment in the AERD (+) group.

We show the GETE scores of all patients, the previous mepolizumab treatment (−) / (+) groups and the AERD (−) / (+) groups in Fig. 1. The total responder rate to benralizumab treatment was 58% (14 patients), including good and excellent responses. Regardless of previous mepolizumab treatment, the response rate was approximately 60%. The response rate in the AERD (−) group was higher than that in the AERD (+) group [71% (12/17) vs 29% (2/7), p = 0.085].

We compared baseline clinical parameters and comorbidities between the responders and nonresponders to benralizumab treatment (Table 3). A significant difference in the %FVC at baseline was found between these two groups. No significant differences in clinical parameters at baseline were observed between the responders and nonresponders in the previous mepolizumab treatment (−) and (+) groups. To further examine the involvement of baseline clinical parameters in the efficacy of benralizumab, we analyzed changes in the FEV₁ of ≥200 ml and < 200 ml (Table 3). The change in the FEV₁ from baseline was 200 ml or greater in eight patients (33%), including three patients who were switched from mepolizumab treatment (Table 3). The %FVC and %FEV₁ at baseline were significantly lower in patients with a change in the FEV₁ ≥ 200 ml than in those with a change in the FEV₁ < 200 ml.

We performed subgroup analyses using univariate logistic regression of the GETE score as follows: 1) patient characteristics and parameters, 2) with or without previous mepolizumab treatment, 3) with or without AERD as a comorbidity, and 4) with or without ECRS as a comorbidity (Table 4). Then, we selected the peripheral blood eosinophil count (≥ 300 /μl), ECRS or CRSwNP, AERD, FeNO (≥ 50 parts-per-billion) and BMI (≥ 25 kg/m²) for multivariate logistic regression analysis of the GETE score. We found that the number of patients with AERD was significantly lower in the GETE responder group than in the non-responder group [odds ratio (OR) 0.035, 95% confidence interval (CI) (0.002–0.72), p = 0.03] (Table 4).

Four patients discontinued benralizumab treatment for the short term for the following reasons: adverse effects [headache (n = 1) and injection site pain (n = 1)], poor effectiveness for asthma (n = 1) and other (n = 1) (data not shown).

We evaluated sinusitis symptoms in 18 of the 20 patients with comorbid SEA and ECRS. Based on patient symptoms in the medical records, nine patients (50%) reported the efficacy of benralizumab treatment for sinusitis (data not shown).