Association of non-malignant respiratory disease with predicted lung cancer risk


Session type:

Stephen W. Duffy1, Ghasem Yadegarfar2, David Baldwin3, John Holemans4, Martin Ledson4, Nicholas Screaton5, Robert Rintoul5, Anand Devaraj6, David M. Hansell7, John K. Field2
1Queen Mary, University of London, London, UK, 2University of Liverpool, Liverpool, UK, 3Nottingham University Hospital, Nottingham, UK, 4Liverpool Heart and Chest Hospital, Liverpool, UK, 5Papworth Hospital, Cambridgeshire, UK, 6St George's Hospital, London, UK, 7Royal Brompton Hospital, London, UK


The Liverpool Lung project (LLP) risk model[1] is a validated predictor of lung cancer risk in the next five years, based on age, sex, smoking history, exposure to asbestos, family history of lung cancer, personal history of other malignancy and personal history of pneumonia. The UKLS pilot trial of low-dose CT screening for lung cancer uses the LLP model for selection of high risk subjects (5% or higher 5-year risk of lung cancer)[2].


To date, we have data on 23,575 subjects, of whom 2,818 (12%) had sufficient risk to qualify for recruitment, and of whom 1,402 (6%) are participating in the trial. We estimated the association of five year risk with history of asthma, bronchitis, emphysema, tuberculosis, and COPD, and the extent to which this was driven by the associations of respiratory diseases with smoking. Additionally, we estimated the implications in terms of disease rates in the participating population.


Risk was significantly higher in the presence of each of the five respiratory diseases (p<0.001 in all cases). For example, average predicted 5-year risk in those without COPD was 1.5% (SD 2.6%), and in those with COPD it ws 6.0% (SD 6.2%). This was partly but not entirely driven by associations with smoking, since although attenuated, the effect remained significant after adjustment for smoking, and when restricted to non-smokers (0.5% vs 1.3% 5-year risks). Correspondingly, the proportion of subjects with COPD was significantly higher in participants (12% vs 2%, p<0.001). Similar results were observed for other respiratory diseases.


Associations of non-malignant respiratory diseases with lung cancer risk imply that populations selected on the basis of risk, for screening or other control measures, are likely to have considerable respiratory comorbidity.