Predicting responders to reslizumab after 16 weeks of treatment using an algorithm derived from clinical studies of patients with severe eosinophilic asthma

Rationale: Reslizumab is a humanized anti–IL-5 monoclonal antibody used as add-on maintenance treatment for patients with uncontrolled eosinophilic asthma. Objectives: To predict response and nonresponse to intravenous reslizumab at 52 weeks with an algorithm we developed based on clinical indicators from pivotal clinical trials. Methods: Patients aged 18 years and older who met Global Initiative for Asthma 4 or 5 criteria and received intravenous reslizumab (n = 321) in two trials (www.clinicaltrials.gov identifiers, NCT01287039 and NCT01285323) were selected as the data source. A mathematical model was constructed that was based on change from baseline to 16 weeks in Asthma Control Questionnaire and Asthma Quality of Life Questionnaire scores and FEV 1 , and number of clinical asthma exacerbations during the year before enrollment and in the first 16 weeks of treatment, and these measures were evaluated for their ability to predict the outcome at 52 weeks: responder, nonresponder, or indeterminate. Measurements and Main Results: The algorithm predicted that 276 patients would be classified as responders; in 248 (89.9%), the prediction was correct. In comparison, 26 patients were predicted to be nonresponders; 50.0% of these predictions were correct. Nineteen patients were classified as indeterminate. The algorithm had 95.4–95.5% sensitivity and 40.6–54.1% specificity. Jackknife and cross-study validation confirmed the robustness of the algorithm. Conclusions: Our algorithm enabled prediction at 16 weeks of treatment of the response to intravenous reslizumab treatment at 52 weeks, but it was not suitable for predicting nonresponse. A positive score at 16 weeks should encourage continued treatment, and a negative score should prompt close monitoring to determine whether discontinuation is warranted.