Pre- and post-bronchodilator lung function as predictors of mortality in the Lung Health Study

Background: Chronic obstructive pulmonary disease (COPD) is supposed to be classified on the basis of post-bronchodilator lung function. Most longitudinal studies of COPD, though, do not have post-bronchodilator lung function available. We used pre- and post bronchodilator lung function data from the Lung Health Study to determine whether these measures differ in their ability to predict mortality. Methods: We limited our analysis to subjects who were of black or white race, on whom we had complete data, and who participated at either the 1 year or the 5 year follow-up visit. We classified subjects based on their baseline lung function, according to COPD Classification criteria using both pre- and post-bronchodilator lung function. We conducted a survival analysis and logistic regression predicting death and controlling for age, sex, race, treatment group, smoking status, and measures of lung function (either pre- or post-bronchodilator. We calculated hazard ratios (HR) with 95% confidence intervals (CI) and also calculated area under the curve for the logistic regression models. Results: By year 15 of the study, 721 of the original 5,887 study subjects had died. In the year 1 sample survival models, a higher FEV1 % predicted lower mortality in both the pre-bronchodilator (HR 0.87, 95% CI 0.81, 0.94 per 10% increase) and post-bronchodilator (HR 0.84, 95% CI 0.77, 0.90) models. The area under the curve for the respective models was 69.2% and 69.4%. Similarly, using categories, when compared to people with normal lung function, subjects with Stage 3 or 4 disease had similar mortality in both the pre-(HR 1.51, 95% CI 0.75, 3.03) and post-bronchodilator (HR 1.45, 95% CI 0.41, 5.15) models. In the year 5 sample, when a larger proportion of subjects had Stage 3 or 4 disease (6.4% in the pre-bronchodilator group), mortality was significantly increased in both the pre-(HR 2.68, 95% CI 1.51, 4.75) and post-bronchodilator (HR 2.46, 95% CI 1.63, 3.73) models. Conclusions: Both pre- and post-bronchodilator lung function predicted mortality in this analysis with a similar degree of accuracy. Post-bronchodilator lung function may not be needed in population studies that predict long-term outcomes.