Cookie Notice

This site uses cookies. By continuing to browse this site, you are agreeing to our use of cookies. Review our cookies information for more details.

skip to main content

Predicting asthma attacks (exacerbations): performance of current instruments

Published online: February 15, 2018

Exacerbations (asthma attacks) are bothersome events of increased symptoms in patients with asthma, often resulting in urgent health care consultation and absenteeism from activities like work or school. When clinicians are able to estimate which patients are likely to experience an exacerbation in the future, interventions to prevent these events (like a step up in medication) can be applied individually. Several prediction models (instruments that combine risk factors, like increased symptoms and smoking, to estimate future risks) have been developed. Generally, they have not been tested in populations other than those in which they were developed, leaving their actual predictive performance unknown.

In a recent issue of The Journal of Allergy and Clinical Immunology: In Practice, Loymans and collaborators reported on a study aimed at collecting all models that predict exacerbations of asthma and assessing their predictive performance in two different populations: one consisting of patients with mild and one with severe asthma. Prediction models were collected by an extensive search in the medical literature (a systematic review). Performance of prediction was assessed in two ways: 1) discrimination: how well does the model discriminate between patients with and without an exacerbation and 2) calibration: how accurate are the risks calculated by the model?

From 12 reports, 24 models were collected. The models were developed in very distinct populations, the outcome (exacerbation) was not defined uniformly (usually as courses of prednisone, emergency department visits, hospitalizations or combinations thereof), and description of important issues on the development of the model were often missing. Previous exacerbations, increased symptoms, and decreased lung function were elements of most prediction models. Most of the collected models demonstrated similar predictive performance in both distinct populations, although calibration, discrimination or both were usually below the desired level.

Current prediction instruments for asthma attacks need improvement before use in clinical practice. Previous exacerbations, increased symptoms, and decreased lung function are key elements of predicting asthma attacks. This set of elements should be supplemented by novel (bio)markers to increase the predictive performance to a level that merits application in the general asthma population.

The Journal of Allergy and Clinical Immunology: In Practice is an official journal of the AAAAI, focusing on practical information for the practicing clinician.