Mortality by cause of death and spirometric pattern in a population-based study

Authors:

Abstract
Background: In Sweden, cardiovascular disease (CVD) and cancer are the most common causes of death. In population-based studies mortality is higher among subjects with than without COPD, but data on causes of death among subjects with COPD are scarce. Aim: To study major causes of death in a population-based COPD-cohort and non-obstructive referents divided into normal lung function (NLF) and restrictive spirometric pattern (RSP). Methods: Subjects with airway obstruction (FEV1/VC<0.70), were, together with non-obstructive referents, identified from the OLIN studies examinations in 2002-04. Spirometric groups: NLF (FEV1/VC≥0.70&VC≥80% predicted), RSP (FEV1/VC≥0.70&VC<80% predicted), and COPD (post- bronchodilator FEV1/VC<0.70). Causes of death were collected from the Swedish National Board of Health and Welfare until December 2015. Results: The study-population included n=742 NLF, n=251 RSP, and n=736 COPD. The cumulative mortality was lower in NLF than in RSP and COPD; 25.1 vs 41.4 and 40.6% (p<0.001vs NLF for both). In NLF, RSP, and COPD, respectively, the cumulative mortality by cause of death was 1.1, 3.6, and 6.0% respiratory; 9.8, 16.3, and 16.4% CVD; and 8.1, 11.2, and 11.3% cancer. In a multivariate Cox regression model there was an increased risk (HR;95%CI) for respiratory mortality associated with RSP (2.6;1.0-6.7) and COPD (3.5;1.6-7.7) when compared with NLF, and corresponding estimates for CVD death were 1.3;0.9-1.9 and 1.5;1.1-2.0, while cancer death yielded 1.1;0.7-1.8 and 1.2;0.8-1.7, respectively. Conclusion: Subjects with RSP and COPD have a similar and increased mortality compared to those with NLF, and it was mainly associated with respiratory and CVD but not cancer cause of death.