Cluster analysis of volatile organic compounds — A pilot study in patients with severe acute decompensated heart failure

Abstract

Introduction

Acute decompensated heart failure (ADHF) is a global health problem and early detection of high-risk patients for effective treatment is important. Exhaled breath analysis and measurement of volatile organic compounds (VOCs) may be a fast and cost-effective non-invasive diagnostic and screening tool complementing measurement of cardiac biomarkers. Another technique to detect and characterize VOCs is the ion mobility spectrometry (IMS) not requiring vacuum or sample pretreatment.

Methods

This prospective controlled proof-of-concept study prospectively enrolled adult patients with severe ADHF at the University Hospital Heart Centre Brandenburg. Severe ADHF was defined as patients presenting with symptomatic acute decompensation and NTproBNP >7000 pg/dL. Cardiac patients with NT-proBNP 220 pg/dL served as control. A gas chromatography ion mobility spectrometer (GC-IMS) of the type “MultiMarkerMonitor™” from GRAUPNER medical solutions GmbH was used.

Measurement was performed at T0 (within 24 h of admission), T1 (after 3–5 days) and T2 (after 8–10 days).

Results

Forty patients were enrolled in the study, 20 patients with severe ADHF and 20 control patients. In patients with severe ADHF, three clusters with significantly altered maximum peak heights were detected compared to control. There was no change in the peak height of clusters 8, 9 and 206 at the time points T1 and T2 (all p > 0.50).

Also, NT-proBNP was stable over time (p = 0.247). Sixteen control patients (16/20, 80 %) and four with severe ADHF (4/20, 20 %) presented without cluster deviation.

Patients with deviation in at least two clusters had longer hospital stay, 11 days (5.0–15.0) compared to those without deviation, 4 days (2.0–9.5), p = 0.028.

Conclusion

Longer-term follow-up studies are needed to assess the stability and clinical significance of the identified clusters by IMS and their diagnostic and prognostic relevance.