Association of novel adult cough subclasses with clinical characteristics and lung function across six decades of life in a prospective, community-based cohort in Australia: an analysis of the Tasmanian Longitudinal Health Study (TAHS)

Abstract

Cough is a common yet heterogeneous condition. Little is known about the characteristics and course of cough in general populations. We aimed to investigate cough subclasses, their characteristics from childhood across six decades of life, and potential treatable traits in a community-based cohort. For our analysis of the Tasmanian Longitudinal Health Study (TAHS), a prospective, community-based cohort study that began on Feb 23, 1968, and has so far followed up participants in Tasmania, Australia, at intervals of 10 years from a mean age of 7 years to a mean age of 53 years, we used data collected as part of the TAHS to distinguish cough subclasses among current coughers at age 53 years. For this analysis, participants who answered Yes to at least one cough-related question via self-report questionnaire were defined as current coughers and included in a latent class analysis of cough symptoms; participants who answered No to all nine cough-related questions were defined as non-coughers and excluded from this analysis. Two groups of longitudinal features were assessed from age 7 years to age 53 years: previously established longitudinal trajectories of FEV1, forced vital capacity [FVC], FEV1/FVC ratio, asthma, and allergies-identified via group-based trajectory analysis or latent class analysis-and symptoms at different timepoints, including asthma, current productive cough, ever chronic productive cough, current smoking, and second-hand smoking. Of 8583 participants included at baseline in the TAHS, 6128 (71·4%) were traced and invited to participate in a follow-up between Sept 3, 2012, and Nov 8, 2016; 3609 (58·9%) of these 6128 returned the cough questionnaire. The mean age of participants in this analysis was 53 years (SD 1·0). 2213 (61·3%) of 3609 participants were defined as current coughers and 1396 (38·7%) were categorised as non-coughers and excluded from the latent class analysis. 1148 (51·9%) of 2213 participants in this analysis were female and 1065 (48·1%) were male. Six distinct cough subclasses were identified: 206 (9·3%) of 2213 participants had minimal cough, 1189 (53·7%) had cough with colds only, 305 (13·8%) had cough with allergies, 213 (9·6%) had intermittent productive cough, 147 (6·6%) had chronic dry cough, and 153 (6·9%) had chronic productive cough. Compared with people with minimal cough, and in contrast to other cough subclasses, people in the chronic productive cough and intermittent productive cough subclasses had worse lung function trajectories (FEV1 persistent low trajectory 2·9%, 6·4%, and 16·1%; p=0·0011, p<0·0001; FEV1/FVC early low-rapid decline trajectory 2·9%, 12·1%, and 13·0%; p=0·012, p=0·0007) and a higher prevalence of cough (age 53 years 0·0%, 32·4% [26·1-38·7], and 50·3% [42·5-58·2]) and asthma (age 53 years 6·3% [3·7-10·6], 26·9% [21·3-33·3], and 41·7% [24·1-49·7]) from age 7 years to age 53 years. We identified potential treatable traits for six cough subclasses (eg, asthma, allergies, and active and passive smoking for productive cough). The required management of productive cough in primary care (eg, routine spirometry) might differ from that of dry cough if our findings are supported by other studies. Future population-based studies could apply our framework to address the heterogeneity and complexity of cough in the community. The National Health and Medical Research Council of Australia, The University of Melbourne, Clifford Craig Medical Research Trust of Tasmania, Victorian Asthma Foundation, Queensland Asthma Foundation, Tasmanian Asthma Foundation, The Royal Hobart Hospital Research Foundation, the Helen MacPherson Smith Trust, GlaxoSmithKline, and the China Scholarship Council.