Development and validation of a novel risk score for asthma exacerbations: The risk score for exacerbations

BACKGROUND: Identifying patients at risk of future severe asthma exacerbations, those whose asthma might be less treatment responsive, or both might guide treatment selection. OBJECTIVE: We sought to investigate predictors for failure to achieve Global Initiative for Asthma (GINA)-defined good current asthma control and severe exacerbations on treatment and to develop a simple risk score for exacerbations (RSE) for clinical use. METHODS: A large data set from 3 studies comparing budesonide/formoterol maintenance and reliever therapy with fixed-dose inhaled corticosteroid/long-acting β2-agonist therapy was analyzed. Baseline patient characteristics were investigated to determine dominant predictors for uncontrolled asthma at 3 months and for severe asthma exacerbations within 12 months of commencing treatment. The RSE, right censored at 6 months to include all 3 studies, was based on the dominant predictors for exacerbations in two thirds of the data set and validated in one third. RESULTS: Patients (n = 7446) whose symptoms were not controlled on GINA treatment steps 3 and 4 and with 1 or more exacerbations (as judged by a clinician based on patient records, history, or both) in the previous year were included. On multivariate analysis, GINA step, reliever use, postbronchodilator FEV1, and 5-item Asthma Control Questionnaire score were dominant (all P < .001) predictors for both the risk of uncontrolled asthma and severe exacerbations. Additional dominant predictors for uncontrolled asthma were smoking status and asthma symptom scores and an additional predictor for severe exacerbation was body mass index. An exponential increase in risk was observed with increments in RSE based on 5 selected predictors for exacerbations. CONCLUSION: Risk of uncontrolled asthma at 3 months and a severe exacerbation within 12 months can be estimated from simple clinical assessments. Prospective validation of these predictive factors and the RSE is required. Use of these models might guide the management of asthmatic patients.