Eosinophilia is associated with worsening asthma severity and decreased lung function, with increased exacerbation frequency. We assessed the safety and efficacy of benralizumab, a monoclonal antibody against interleukin-5 receptor α that depletes eosinophils by antibody-dependent cell-mediated cytotoxicity, for patients with severe, uncontrolled asthma with eosinophilia.We did a randomised, double-blind, parallel-group, placebo-controlled phase 3 study at 374 sites in 17 countries. We recruited patients (aged 12-75 years) with a physician-based diagnosis of asthma for at least 1 year and at least two exacerbations while on high-dosage inhaled corticosteroids and long-acting β2-agonists (ICS plus LABA) in the previous year. Patients were randomly assigned (1:1:1) by an interactive web-based voice response system to benralizumab 30 mg either every 4 weeks (Q4W) or every 8 weeks (Q8W; first three doses every 4 weeks) or placebo Q4W for 48 weeks as add on to their standard treatment. Patients were stratified 2:1 according to blood eosinophil counts of at least 300 cells per μL and less than 300 cells per μL. All patients and investigators involved in patient treatment or clinical assessment were masked to treatment allocation. The primary endpoint was annual exacerbation rate ratio versus placebo, and key secondary endpoints were prebronchodilator forced expiratory volume in 1 s (FEV1) and total asthma symptom score at week 48, for patients with blood eosinophil counts of at least 300 cells per μL. Efficacy analyses were by intention to treat (based on the full analysis set); safety analyses included patients according to study drug received. This study is registered with ClinicalTrials.gov, number NCT01928771.Between Sept 19, 2013, and March 16, 2015, 2681 patients were enrolled, 1205 of whom met the study criteria and were randomly assigned: 407 to placebo, 400 to benralizumab 30 mg Q4W, and 398 to benralizumab 30 mg Q8W. 267 patients in the placebo group, 275 in the benralizumab 30 mg Q4W group, and 267 in the benralizumab 30 mg Q8W group had blood eosinophil counts at least 300 cells per μL and were included in the primary analysis population. Compared with placebo, benralizumab reduced the annual asthma exacerbation rate over 48 weeks when given Q4W (rate ratio 0·55, 95% CI 0·42-0·71; p<0·0001) or Q8W (0·49, 0·37-0·64; p<0·0001). Both benralizumab dosing regimens significantly improved prebronchodilator FEV1 in patients at week 48 compared with placebo (least-squares mean change from baseline: Q4W group 0·106 L, 95% CI 0·016-0·196; Q8W group 0·159 L, 0·068-0·249). Compared with placebo, asthma symptoms were improved by the Q8W regimen (least-squares mean difference -0·25, 95% CI -0·45 to -0·06), but not the Q4W regimen (-0·08, -0·27 to 0·12). The most common adverse events were worsening asthma (105 [13%] of 797 benralizumab-treated patients vs 78 [19%] of 407 placebo-treated patients) and nasopharyngitis (93 [12%] vs 47 [12%]).These results confirm the efficacy and safety of benralizumab for patients with severe asthma and elevated eosinophils, which are uncontrolled by high-dosage ICS plus LABA, and provide support for benralizumab to be an additional option to treat this disease in this patient population.AstraZeneca and Kyowa Hakko Kirin.