Cough phenotypes in children

Abstract

Introduction: Distinction of phenotypes among children with cough can help understand underlying causes. We used a data-driven approach to identify cough phenotypes in children and assess their clinical validity.

Methods: We included 531 children aged 5-16 years with cough (i.e. without cold, at night, more than others or more than 4 weeks) but no wheeze from the Swiss Paediatric Airway Cohort. We applied latent class analysis to identify phenotypes using cough characteristics and symptoms from parental questionnaires and selected the best fitting model and optimum number of classes using Akaike Information Criteria. We tested the validity of the phenotypes by comparing characteristics, objective tests, doctors’ diagnosis, and one-year prognosis.

Results: We identified four cough phenotypes: allergic (61%), infectious (4%), mixed (9%) and nonspecific (25%). Children of the allergic phenotype were older and had more frequently a history of atopy. FeNO differed across groups (p<0.001) being highest in the allergic (17.9 ppb, 95% CI 15.6-19.5) and lowest in the infectious phenotype (7.0 ppb, 95% CI 5.0-17.7). Similarly, positive allergy tests were common in the allergic (70%) and rare in the nonspecific phenotype (31%). 80% of children in the allergic phenotype were diagnosed with asthma compared to 44-55% in other phenotypes. A diagnosis of chronic cough was more frequent in the mixed (44%) and infectious phenotypes (39%). At one-year follow up, children of the allergic phenotype more often developed wheeze and cough persisted less often.

Conclusion: The four identified phenotypes of cough without wheeze are valuable to understand underlying causes and predict short-term prognosis which is essential for clinical management.

Funding: SNSF 320030_182628