Exploring breath metabolomics as a non-invasive tool for detecting pulmonary vascular disease.

Aims: The aim of this initial study was to explore whether prediction models based on breath metabolome profiles could detect differences between pulmonary vascular disease (PVD) patients and healthy controls. Additionally, we sought to investigate the potential to distinguish between two major subtypes of PVD-pulmonary arterial hypertension (PAH) and chronic thromboembolic pulmonary hypertension (CTEPH)-to support early detection and targeted treatment. Methods and results: We used real-time breath analysis to compare the breath profiles of patients with PVD to healthy controls, and the metabolome of patients with PAH to those with CTEPH. Pathway enrichment analysis was conducted to reveal underlying metabolic pathways. Breath profiles of 75 patients (47 (62.7%) with PAH and 28 (37.3%) with CTEPH) were analysed and compared with those of 115 healthy controls. The prediction models identified PVD with an area under the curve of 0.917 and distinguished PAH from CTEPH with an AUC of 0.764. PVD patients showed significant metabolic alterations, particularly in Conclusion: Breath analysis shows potential as a non-invasive and real-time diagnostic tool by demonstrating detectable differences in breath profiles between PVD patients and healthy controls. Establishing these differences is a critical first step in assessing the feasibility of identifying breath markers for PVD and exploring further differentiation between PAH and CTEPH. Registration: Trial registration number: NCT05458934.