Severe eosinophilic asthma can now be treated with biologic therapies that reduce attacks and, in some cases, allow a corticosteroid dose reductions. However, not all patients respond to these treatments, and it remains difficult to predict in advance who will benefit from them. A study conducted by scientists at the University of Liège and Liège University Hospital proposes using a ‘simple’ breath sample to guide treatment selection even before it begins.
The use of breath for clinical purposes (breathomics) is a rapidly expanding field. From the detection of various cancers to the diagnosis of food intolerances, and the monitoring of chronic conditions, breath analysis has applications in many areas. However, few studies have demonstrated the transferability of identified markers between different patient cohorts or between different clinical applications. A new study shows that asthma markers are transferable between different research centers and can be used for various applications, such as identifying types of asthma or monitoring the effectiveness of treatments.
Biotherapies targeting interleukin-5 (mepolizumab and benralizumab) have transformed the management of severe asthma. However, their effectiveness varies from one patient to another, and the markers currently used to guide their prescription - fractional exhaled nitric oxide (FeNO) and blood eosinophil count – are poor predictors of treatment response. Identifying, at an early stage, which patients are likely to respond represents a genuine clinical need.
“For this study, we analysed the breath of 58 patients with severe asthma, divided into two independent cohorts: thirty-four patients at the University Hospital of Liège and twenty-four at Glenfield Hospital in Leicester,” explains Thibault Massenet, a researcher at the Obiachem Laboratory at the University of Liège. Samples were taken just before the first injection, and the patients were then followed up for six to twelve months.” The breath samples were analysed using two-dimensional gas chromatography coupled with mass spectrometry, a technique that enables the separation and measurement of dozens of volatile compounds. The analysis was based on a signature of seventeen molecules, previously identified in earlier work by the same team as indicative of eosinophilic inflammation.
The two cohorts were analysed separately using distinct statistical methods. In both cases, the breath signature made it possible to distinguish quite clearly between responder and non-responder patients. To put this result into context, doctors currently rely on two tests to guide the prescription of these biologics: FeNO, which measures inflammation based on a gas (nitric oxide) present in exhaled air, and the eosinophil count in a blood sample. In these cohorts, breath analysis predicted the response more accurately than these two gold-standard tests.
The contribution of scientists from ULiège and doctors at Liège University Hospital lies at the heart of this study. “For over ten years, the Faculty of Sciences and the Faculty of Medicine at the University of Liège, together with Liège University Hospital, have been jointly developing expertise in the analysis of breath molecules,” explains Pierre-Hugues Stefanuto , a chemist at the ObiaChem Laboratory. “The main significance of this work lies in the transferability of the markers: compounds identified in a previous study were able to be reused in a new cohort and in another laboratory, with consistent results. This inter-centre reproducibility remains rare in the emerging field of breathomics, the clinical use of breath analysis.”
The authors urge caution in interpreting the findings. Both cohorts are small and contain significantly more responders than non-responders, mirroring real-world prescribing conditions. These results will need to be confirmed by larger prospective trials before any clinical application. Future research will also focus on rapid detection devices for use at the patient’s bedside and on the standardisation of protocols.
“In the longer term, the aim is to offer personalised monitoring based on breath samples, ” concludes Florence Schleich , a pulmonologist at ULiège. “Just as a diabetic patient monitors their blood glucose levels, an asthma patient might one day have their treatment adjusted based on a simple breath sample. The study does not go that far, but it shows that a non-invasive respiratory marker can provide relevant information to guide treatment decisions and brings this technology closer to clinical use.”
American Journal of Respiratory and Critical Care Medicine
., Eosinophil-derived breath biomarkers and response to anti-IL-5/5R biologics in severe asthma
8-Jul-2026