Voice Biomarker Signals Exacerbation Onset

VOICE changes may identify asthma and COPD exacerbations from their first day, supporting future home monitoring.

Researchers found that several acoustic speech features changed at the onset of an exacerbation, before symptoms reached their peak. The findings suggest that a voice biomarker could eventually provide a simple digital method for monitoring people with chronic respiratory disease at home.

Voice Biomarker Tracked Through Mobile Devices

The prospective TACTICAS study included 73 Dutch-speaking participants, comprising 35 people with asthma and 38 with chronic obstructive pulmonary disease (COPD). Their mean age was 62 years.

Participants completed daily voice recordings on their own mobile devices for approximately 3 months. Tasks included sustaining an “a” vowel and reading text. Exacerbation onset, peak symptom burden, and recovery were tracked using a respiratory symptom questionnaire, electronic health records, medication use, and participant-reported symptoms.

Investigators analyzed 39 speech features using multilevel models that accounted for repeated recordings from the same participant. Overall, the study collected 23,799 recordings, including 2,737 during exacerbation periods.

Voice Quality Changed From Exacerbation Onset

Thirty-eight exacerbations occurred among 35 participants. Sixteen were mild, 19 were moderate, and three were severe.

Thirteen voice features changed significantly on the first day of an exacerbation. Sustained vowel duration decreased, while measures reflecting irregular vocal fold vibration and reduced voice quality increased. Changes were also observed in pitch, pulses, periods, jitter, shimmer, and the balance between harmonic sound and noise.

Breathlessness and chest symptoms showed stronger associations with speech features than cough and sputum symptoms. The effects did not differ significantly between asthma and COPD, suggesting that acute airway obstruction may contribute to similar vocal changes across both conditions.

Participants completed at least one speech task on 83% of intended study days, supporting the feasibility of mobile voice monitoring. However, the observed effects were small, and the analysis identified associations rather than demonstrating that voice could predict exacerbations. The study was also limited to Dutch speakers, did not adjust for multiple comparisons, and requires validation in larger, multilingual populations.

The researchers concluded that a voice biomarker may eventually support earlier identification and management of respiratory deterioration, although predictive algorithms and integration into clinical care pathways remain necessary.

Reference
van Bemmel L et al. Voice as biomarker for early exacerbation detection in asthma and COPD: the TACTICAS study. ERJ Open Res. 2026. doi:10.1183/23120541.01737-2025.

Featured Image: andranik123 on Adobe Stock.