A75-03 Daily Cough Count Variability and Correlation With PROs in Chronic Cough: Informing Antitussive Trial Endpoints

Rationale

Respiratory clinical trials are starting to evaluate antitussive effects of novel therapies in development. Understanding the natural variability of objective cough frequency and its relationship to patient-reported outcomes (PROs) is essential for optimizing trial design. We conducted a prospective observational study using continuous cough monitoring (CCM) to characterize daily cough variability and examine correlations with PROs in clinically stable refractory and unexplained chronic cough (RCC/UCC) patients.

Methods

Thirty-five RCC/UCC patients were recruited from two cough clinics (Royal Brompton Hospital, Belfast City Hospital). Participants wore the Hyfe CoughMonitor smartwatch during waking hours and continued to monitor by the bedside at night for 30 days, completing daily visual analogue scale (VAS) assessments via a companion smartphone application. Hyfe CoughMonitor wearable is fully privacy-preserving and does not record or store audio or conversations. It passively monitors sound levels, and a validated convolutional neural network runs on-device to timestamp coughs in real-world conditions in real time. The Leicester Cough Questionnaire (LCQ) was administered at enrollment and study completion. Objective cough counts were correlated with daily VAS scores and analyzed for day-to-day variability. Test-retest reliability was assessed for both CCM and LCQ in this stable cohort.

Results

928 person-days comprising 302,868 coughs were recorded. Mean participant age was 65.5 years (66% female). Average hourly cough rates ranged from 2.1 to 51.2 (cohort mean 13.6). Substantial intrasubject day-to-day variability in cough rates was observed (Figure1). The repeated measures correlation coefficient between daily VAS scores and objective cough counts was poor (r = 0.23, 95% CI 0.16 - 0.30), with considerable inter-individual variation. Among participants reporting VAS below 40, 28% had greater than 10 coughs per hour and 14% exceeded 20 coughs per hour. The Pearson correlation coefficient between the first and last LCQ was 0.79. CCM test-retest reliability improved to 0.90 with longer monitoring duration, suggesting that using the average cough rate over consecutive monitoring days provides a more reliable and stable endpoint.

Conclusions

Daily cough rates in RCC/UCC exhibit substantial variability, and subjective VAS assessments correlate poorly with objective cough frequency. Using cough VAS thresholds alone for trial screening may exclude a significant proportion of patients with objectively high cough rates. These findings have direct implications for endpoint selection and enrollment criteria in clinical trials evaluating antitussive effects, including consideration for monitoring duration and the complementary roles of CCM and PROs.

This abstract is funded by: Trevi Therapeutics