Introduction and objectives Few studies have investigated cough frequency in neurological patient groups, in which cough may be impaired or increased in the presence of aspiration. This study aimed to (1) validate the Leicester Cough Monitor (LCM) on a stroke unit, where background coughs might contaminate one patient’s cough recordings; and (2) observe cough frequency longitudinally in a convenience sample of acute stroke survivors.

Methods To validate the LCM, 15-minute recordings were made from 5 patients on a stroke unit. LCM results were compared with real-time cough counts by a researcher present in the room (visual and auditory). To observe cough frequency longitudinally, 21 stroke survivors underwent 24-hour LCM recordings at baseline (<2 weeks post stroke), week 1 and 4. Participants (14 men, mean (SD) age 60 (15) years) had moderate stroke impairment (median (IQR) NIHSS score 8 (5, 11)) with cortical (n = 9), subcortical (n = 9), brainstem (n = 2) and cerebellar (n = 1) strokes. Five randomly selected recordings were analysed by a second researcher, blinded to subject characteristics and not present during the recordings.

Results In the validation study, the real-time observer counted 67 subject coughs plus 81 background coughs in total. The LCM returned a subject cough count of 68, not significantly different to the observer’s count (p = 0.99) with excellent agreement (ICC 0.996, 95% CI: 0.967, >0.999). Inter-rater reliability for LCM hourly cough counts was good (ICC 0.973, 95% CI: 0.789, 0.997). In the longitudinal cohort, average cough frequency was higher at baseline and reduced over time, with wide individual variability (Table 1) and higher cough frequency during day-time. There were no significant associations between cough frequency and sex, age, stroke site, stroke severity, swallowing safety, smoking status or ACE-inhibitor use.

Conclusions This study is limited due to the small sample size and should be regarded as exploratory. It was possible to validate the LCM for application on an acute stroke unit. The findings might serve hypothesis-generation: For example, is cough frequency after stroke increased, indicating sub-clinical levels of swallowing impairment and aspiration threat, which trigger frequent protective coughs?