Exhaled breath acetone during rest and exercise reflects haemodynamic stress and reduced functional capacity across the heart-failure spectrum

Introduction Exhaled breath acetone (EBA) is a volatile organic compound potentially reflecting altered cardiac energetics and haemodynamic stress. Whether EBA measured during exercise relates to congestion and functional limitation across heart-failure (HF) phenotypes remains unclear. Purpose To determine whether EBA measured at rest and during exercise differs between patients with established HF and subjects at risk of HF, and whether EBA tracks markers of cardiac dysfunction, congestion and exercise intolerance assessed by echocardiography, NT-proBNP and cardiopulmonary exercise testing (CPET). Methods We enrolled consecutive outpatients with HF and left ventricular ejection fraction (LVEF) >50% (HFpEF) or <50% (HF with reduced/mildly reduced EF), and subjects with cardiovascular risk factors and/or structural heart disease without established HF (“at risk”). All participants underwent clinical and laboratory assessment, resting transthoracic echocardiography, and combined CPET–exercise stress echocardiography with EBA monitoring at rest (EBArest) and during exercise (EBAex). Mixed exhaled breath samples were collected directly onto sorbent tubes packed with Tenax GR and quantified by thermal desorption–gas chromatography–mass spectrometry (TD-GC-MS). Diabetes was not a significant covariate in the EBA analyses. The cohort included HFpEF (n=62), HF with LVEF <50% (n=41) and at-risk subjects (n=50); HFpEF participants were older and more frequently female. Results EBA was higher in HF compared with at-risk subjects both at rest and during exercise: EBArest 1.5 (IQR 1.0–3.1) vs. 0.9 (0.7–1.2) µg/L and EBAex 2.4 (1.5–4.4) vs. 1.1 (0.9–2.1) µg/L (all p<0.0001). Among HF patients, those in the highest EBArest tertile showed a more severe phenotype, with lower LVEF, greater echocardiographic signs of congestion, higher NT-proBNP, and reduced peak oxygen consumption, consistent with impaired exercise capacity. In multivariable regression, NT-proBNP (p=0.0004) and ventilatory inefficiency (VE/VCO2 slope; p=0.0013) independently predicted EBArest (adjusted R2=0.458). Conclusion EBA is increased in HF compared with at-risk subjects and is associated with congestion biomarkers and ventilatory inefficiency, paralleling reduced functional capacity. These findings support EBA as a promising non-invasive marker of HF severity during rest and exercise, warranting prospective studies to test its incremental value for HF phenotyping and monitoring.