Videostroboscopic findings in acute heart failure: results from the AHF-Voice strobo substudy
M Bauser, F Kraus, V Cejka, F Sahiti, C Morbach, R Pryss, K Rak, S Frantz, J Hoxha, S Stoerk, F KerwagenAbstract
Background
Several studies have demonstrated voice alterations in patients with acute heart failure (AHF), with vocal biomarkers emerging as a promising tool for the non-invasive monitoring of congestion in heart failure. Vocal fold dysfunction, potentially caused by congested vocal fold edema, is hypothesized to underlie these voice changes, though objective stroboscopic evidence supporting this hypothesis is still lacking. Videostroboscopy is recommended by the European Laryngological Society for assessing vocal fold function.
Method
In the AHF–Voice Strobo substudy of the prospective, monocentric AHF-Voice cohort, we investigated pathophysiological vocal changes during acute heart failure using videostroboscopy and body composition analysis at admission, discharge, 6 weeks, and 6 months. Stroboscopic recordings were independently evaluated by a blinded phoniatric expert according to ELS criteria, including vibratory amplitude, mucosal wave, symmetry, regularity, glottic closure, and supraglottic activity. Laryngeal secretion severity was assessed using the Secretion Severity Rating Scale (0-3, four-point scale). Pharyngeal residue in the vallecula and pyriform sinus was graded on an adapted five-point Yale Residue Scale (1-5). Velopharyngeal closure was rated on an ordinal four-point scale (0-3). Spearman correlations were used to assess cross-sectional associations at Admission. Cumulative link mixed models with subject-specific random intercepts were applied to evaluate longitudinal effects, accounting for repeated measurements and missing data.
Results
Between Apr 2023 and Nov 2025, 50 patients hospitalized for AHF were included in the Strobo substudy. In this group, the mean age was 73±11 years, 22% were women, and 26% had de novo HF. The median NT-proBNP level at admission was 6573 [quartiles 2472; 14535] pg/ml; 84% were in NYHA class III/IV; and the mean LVEF was 44±17%. Body weight correlated moderately with ELS parameters, including mucosal wave (ρ= −0.37, p=0.018), glottic closure (ρ= −0.34, p=0.029), and amplitude (ρ= +0.32, p=0.036). Higher log NT-proBNP levels were also associated with greater glottic abnormalities (ρ= +0.32, p=0.042), Figure. Both the Secretion Severity Rating Scale and velopharyngeal disorder showed concordant trends with overhydration (ρ= +0.29, p=0.060; ρ= +0.28, p=0.066). None of these associations remained statistically significant after correction for multiple testing. Longitudinal mixed-effects models demonstrated a significant overall time effect for velopharyngeal disorder (χ²=8.49, df=3, p=0.037), although no adjusted pairwise post hoc comparisons reached statistical significance.
Conclusion
In AHF, videostroboscopic findings showed only weak associations with disease severity and remained largely stable over time. More sophisticated methods including high-speed videolaryngoscopy might be needed to detect potential minor changes of vocal fold function.For image description, please refer to the figure legend and surrounding text.