BACKGROUND: Chronic obstructive pulmonary disease (COPD) substantially contributes to morbidity and mortality worldwide. We aimed at estimating Global Initiative for Chronic Obstructive Lung Disease (GOLD) spirometric grade-specific mortality in COPD for Germany, using data from a large-scale cohort of patients with COPD. METHODS: Using COSYCONET data, a cohort of 2741 patients diagnosed with COPD was followed over up to 9 years. We estimated mortality rates for GOLD grades 1 to 4 and stratified into age and sex groups. An exponential survival model was used to estimate mortality after checking model assumptions. Additionally, a Cox proportional hazards model was estimated as plausibility check for the exponential model. RESULTS: A total of 345 deaths were observed during the follow-up period. The data fitted well to an exponential survival model when the first year of follow-up was excluded, suggesting a "healthy participant effect". GOLD grade was a strong predictor of mortality, with hazard ratios of 1.6, 3.2, and 8.8 for GOLD 2-4 compared to GOLD 1. Hazard ratios of the Cox model were similar (1.7, 3.4, and 10.0 for grades 2-4 compared to grade 1). At a given grade, mortality strongly increased with age. 1-year mortality ranged from 0.5 % (GOLD 1, <55 years, females) to 54.9 % (GOLD 4, 80+ years, males). Mortality was lower among females by approximately 25 %. CONCLUSION: Based on our findings, mortality in COPD depends on GOLD grade, age, sex and smoking status. Parametric estimation allowed to estimate 1-year mortality for each combination of COPD grade and age group, including uncertainty estimates.
Keywords
