TY - JOUR T1 - Predictors of mortality in a well-characterised group of patients with COPD JF - European Respiratory Journal JO - Eur Respir J VL - 40 IS - Suppl 56 SP - P3455 AU - Saher B. Shaker AU - Asger Dirksen Y1 - 2012/09/01 UR - http://erj.ersjournals.com/content/40/Suppl_56/P3455.abstract N2 - Background: COPD is a leading cause of mortality worldwide with an increasing incidence.Objective: to identify clinical, physiological and radiological predictors of mortality in a well-characterised group of patients with COPD.Material & methods: patients who participated in 3 clinical trials were included in this study. In these trials, patients were evaluated with clinical data, lung function tests and quantitative computed tomography (relative area of emphysema < -910 HU [RA-910]). The date and cause of mortality were reported during 10 years, and the data of those who were still living by February 1st 2012 were censored at this time point. Data were analysed using Cox proportional hazard regression model first in a univariate model then in a stepwise multiple regression model.Results: Data from 208 patients with moderate to severe COPD were available. A total of 104 patients died. The median survival time was 10.4 yrs (95% C.I. 9.4-∞). Age, packyears, FEV1, DLCO and RA-910 were significant predictors of mortality in a univariate model. In a multivariate model with a stepwise selection, age (p=0.005), packyears (p=0.02) and RA-910 (p=0.017) emerged as the significant predictors of mortality, whereas FEV1 (p=0.05) and DLCO (p=0.14) did not reach statistical significance.Conclusion: In this population of smokers with moderate to severe COPD, age, packyears and the degree of emphysema on CT as RA-910 were the most important predictors of mortality. ER -