Background: Heart failure (HF) is associated with poor prognosis, and the identification of biomarkers of its severity could help in its treatment. In a pilot study, we observed high levels of acetone in the exhaled breath of patients with HF. The present study was designed to evaluate exhaled acetone as a biomarker of HF diagnosis and HF severity.
Methods: Of 235 patients with systolic dysfunction evaluated between May 2009 and September 2010, 89 patients (HF group) fulfilled inclusion criteria and were compared with sex- and age-matched healthy subjects (control group, n = 20). Patients with HF were grouped according to clinical stability (acute decompensated HF [ADHF], n = 59; chronic HF, n = 30) and submitted to exhaled breath collection. Identification of chemical species was done by gas chromatography-mass spectrometry and quantification by spectrophotometry. Patients with diabetes were excluded.
Results: The concentration of exhaled breath acetone (EBA) was higher in the HF group (median, 3.7 μg/L; interquartile range [IQR], 1.69-10.45 μg/L) than in the control group (median, 0.39 μg/L; IQR, 0.30-0.79 μg/L; P < .001) and higher in the ADHF group (median, 7.8 μg/L; IQR, 3.6-15.2 μg/L) than in the chronic HF group (median, 1.22 μg/L; IQR, 0.68-2.19 μg/L; P < .001). The accuracy and sensitivity of this method in the diagnosis of HF and ADHF were about 85%, a value similar to that obtained with B-type natriuretic peptide (BNP). EBA levels differed significantly as a function of severity of HF (New York Heart Association classification, P < .001). There was a positive correlation between EBA and BNP (r = 0.772, P < .001).
Conclusions: EBA not only is a promising noninvasive diagnostic method of HF with an accuracy equivalent to BNP but also a new biomarker of HF severity.