The prevalence of cough in lung cancer: Its characteristics and predictors.

Abstract

162 Background: Cough commonly affects patients with LC but quoted prevalence figures vary and most were obtained before advances in anticancer treatment for LC. This study aimed to determine its prevalence and characteristics (severity, impact and potential predictors) in patients undergoing standard treatment and follow up. Methods: Consecutive LC patients attending clinics at The Christie Hospital, UK during a predefined 5 week period were asked whether they had a cough and invited to participate in its further evaluation according to an ethically approved protocol. Demographic, cancer and cancer treatment data were collected in all. Patients with a cough also completed subjective assessments using the validated cough impact Manchester cough in LC scale (MCLCS) and cough severity visual analogue scale (VAS). Standard statistical tests were used for analysis with significance (p<0.05). Results: Of 223 consecutive LC outpatients approached, the prevalence of cough was 57% (128/223); 202/223 (86%) consented to further data collection (study population) and of these 115 (57%) reported cough. The majority (60/115, 52%) felt that their cough warranted treatment and 23% reported their cough to be painful (26/115). The median VAS score (n=115) was 32mm (25th-75th IQR 20-51, range 0-100, high scores = worse cough severity). The median MCLCS score (n=113) was 22 (25th-75th IQR 16-27, range 1-50, high scores = worse cough impact). There was no significant difference between coughers and non-coughers with respect to age, sex, co-morbidities, smoking, performance status, treatment intent, treatment type, reasons for not receiving treatment, stage and histology. The proportion of non-coughers was lower among patients receiving anticancer therapy (45/115 39% vs 47/87 54%, p=0.04) in univariate analysis. Conclusions: Cough is present in the majority of LC patients, even in those undergoing treatment. Many patients feel that it warrants treatment. Whilst many consider factors such smoking, stage, histology and cancer treatment type to be important determinants of the presence of cough, our data does not reflect this. Improving LC related cough requires more than better cancer therapies.