Abstract
Background: Congestive heart failure (CHF) requires continuous monitoring, especially during exacerbation phases. Pulse oximetry, commonly used for critical patient surveillance, has shown diagnostic potential in acute heart failure. In this exploratory pilot study, we investigated oxygen saturation (SpO(2)) data and respiratory sounds from CHF patients to uncover their relationship and to assess the potential of respiratory sound intensity in telehealth or self-monitoring systems, with a view toward future predictive applications. Methods: The relationship between SpO(2) and acoustic intensity was explored by collecting physiological and acoustic data, including infrasound, from 25 CHF patients using an electronic pickup device. These patients had been enrolled in a larger clinical trial that aimed to explore disease exacerbation across various chronic diseases. Four patients experienced exacerbation phases (SpO(2) < 92%), and for each phase, we computed Pearson correlations in two frequency ranges (audible, audible + infrasound). Eight prespecified correlations were assessed, with unadjusted and adjusted p values (Bonferroni, FDR) and effect sizes reported. Results: Significant negative correlations between specific acoustic frequency ranges and SpO(2) variations were found in several patients. In all four patients, inclusion of the infrasound range increased the correlation magnitude compared to the audible range alone, with lower p values in all cases. Adjusted analyses retained significance in Patients 2 and 4 across both frequency ranges. Conclusion: This pilot work identifies consistent moderate-to-strong negative correlations between acoustic intensity and SpO(2) during CHF exacerbations. While not confirmatory, these results support the potential of acoustic intensity as a candidate indicator for early detection, warranting validation in larger studies and predictive modeling frameworks. Trial Registration: EUDAMED Clinical Investigation (CIV) Identification: CIV-NO-21-10-037926.