Trends in Cause-Specific Mortality in Oxygen-dependent Chronic Obstructive Pulmonary Disease

Rationale: Since the introduction of long-term oxygen treatment (LTOT) in chronic obstructive pulmonary disease (COPD) with chronic hypoxia, the proportion of women and the age of patients starting LTOT have increased markedly. We hypothesize that this might have led to shifts in the causes of death over time. Objectives: To test for time trends in cause-specific mortality in COPD with LTOT. Methods: Patients starting LTOT for COPD in Sweden between January 1, 1987 and December 31, 2004 were included in a national prospective study and monitored until withdrawal of LTOT, death, or December 31, 2004. The primary end point was cause of death obtained from the Swedish Causes of Death Register. Measurements and Main Results: A total of 7,628 patients (53% women) were monitored for a median of 1.7 years (range, 0-18.0 yr). No patient was lost to follow-up and 5,457 patients died during the study. The crude overall mortality increased by 1.6%/year (95% confidence interval [CI], 0.9-2.2%/yr; P < 0.001). The absolute risk of death increased for circulatory disease by 2.8%/year (95% CI, 1.3-4.3%/yr; P < 0.001) and for digestive organ disease by 7.8%/year (95% CI, 1.9-14.0%/ yr; P = 0.009). The absolute risk of death decreased for respiratory disease by 2.7%/year (95% CI, 2.0-3.3%/yr; P < 0.001) and for lung cancer by 3.4%/year (95% CI, 1.1-5.7%/yr; P = 0.004). Conclusions: In oxygen-dependent COPD, mortality has increased over time both overall and of nonrespiratory causes, including cardiovascular disease. This highlights the importance of optimized diagnostics and treatment of comorbidities to decrease morbidity and mortality.