Vocal biomarkers in COPD: capturing disease severity using voice

Abstract

<b>Title:</b> Vocal biomarkers in COPD: capturing disease severity using voice <b>Introduction:</b> Voice characteristics (e.g. number of syllables per breath) in COPD are different from those of healthy subjects. These characteristics also change during an acute exacerbation of COPD (AECOPD) making voice a potential vocal biomarker for COPD patients. We hypothesized that such voice characteristics during an AECOPD would be more significantly affected in patients with more severe disease (i.e. higher GOLD stage). <b>Methods:</b> Speech recordings from COPD patients during an exacerbation and after recovery were analyzed using software package ‘Praat’. The primary outcomes of interest were: sustained vowel production and syllables per breath from ‘the story reading fragment’.&nbsp;Secondary analyses were performed for other vocal biomarkers (shimmer, HNR ratio, degree of voice breaks). Differences in voice characteristics were analyzed based on severity of COPD (Gold stage 2 vs. 3-4) and disease state (stable vs exacerbation). <b>Results:</b> The results of 10 COPD patients (5 COPD Gold 2 and 5 Gold 3-4) were compared respectively in the stable and exacerbated state. There was a trend towards a lower number of syllables per breath in a exacerbated state in both COPD GOLD severity groups. This difference was significantly different in the GOLD 2 group (p=0.016). The remaining outcomes of interest showed trends towards significance, in which the exacerbated state showed worse outcomes compared to the stable disease state <b>Conclusions:</b> GOPD GOLD stage influences specific acoustic characteristics in the voices of COPD patients during a stable and an exacerbated state. Voice could thus potentially be used to detect AECOPDs in patients with COPD.