Antibiotics confound breath-based respiratory disease detection in calves

Abstract

Breath analysis offers a promising, non-invasive approach for early disease detection in livestock, particularly for bovine respiratory disease (BRD). However, the widespread use of antibiotics in veterinary practice raises concerns about their potential to confound breath-based diagnostics. In this study, we analysed breath samples from 65 calves to evaluate the diagnostic performance of volatile organic compounds (VOCs) as biomarkers of BRD. We identified several candidate ions, including formaldehyde and acetone/propanal, that showed significant elevation at the onset of disease and a modest detection capability (AUC = 0.53-0.77). However, the strongest discrimination between diseased and controls occurred after antibiotic administration, suggesting a confounding pharmacological influence. To isolate this effect, a secondary trial was conducted in which healthy calves were treated with oxytetracycline dihydrate (Alamycin® LA 300). This revealed rapid and transient changes in the breath volatilome, with significant increases in dimethyl sulphide and (C₅H₁₀O)H + within 1 hour, and other VOCs, including formaldehyde and acetone/propanal within 24 hours of treatment. These findings demonstrate that antibiotics can substantially alter breath VOC profiles, potentially mimicking or masking disease signals. We conclude that breath-based diagnosis of BRD holds promise but must account for treatment history to avoid misclassification. Moreover, the reproducible and time-resolved nature of the VOC response suggests that breath analysis could also be developed as a tool for monitoring antimicrobial exposure and optimising therapeutic dosing in livestock.