Background: No test for asthma has high enough accuracy to be used in isolation. Office-based tests typically have poor sensitivity due to the variable nature of the disease. Exhaled volatile organic compounds (VOCs) reflect host metabolism and may offer biomarkers reflecting fundamental pathophysiological alterations not captured by other tests.

Aim: To investigate whether exhaled VOCs differentiate between untreated symptomatic adults with and without physician confirmed asthma.

Methods: Patients were referred from primary care with suspected asthma. The diagnosis was made by a panel of asthma specialists based on clinical history, examination and the results of all tests recommended in major asthma guidelines. VOCs were sampled (ReCIVA, Owlstone Medical) and analysed using thermal desorption-gas chromatography-mass spectrometry. Univariate and multivariate analysis were performed to filter, extract, and model diagnostic accuracy of VOC features; data from a randomly selected 25% of samples were reserved for independent testing of the model.

Results: Breath data from 106 adults were included in this analysis, 66 with confirmed asthma and 40 without. 39/639 VOC features were significantly different between groups. These were subsequently used to train a supervised model which resulted in a diagnostic sensitivity of 94% and specificity of 40% on the independent test set.

Conclusion: In this, the first study to test VOCs at the point of asthma diagnosis in untreated patients, a select group of VOC features classified patients with asthma from those without. Perhaps uniquely, the VOC model had a clinically useful sensitivity in the test set, potentially providing a rule-out test if validated in a prospective cohort.