Predictors of mortality in patients with emphysema and severe airflow obstruction

Abstract

Purpose: Limited data exist describing risk factors for mortality in patients having predominantly emphysema.

Subjects and methods: A total of 609 patients with severe emphysema (ages 40-83 yr; 64.2% male) randomized to the medical therapy arm of the National Emphysema Treatment Trial formed the study group. Cox proportional hazards regression analysis was used to investigate risk factors for all-cause mortality. Risk factors examined included demographics, body mass index, physiologic data, quality of life, dyspnea, oxygen utilization, hemoglobin, smoking history, quantitative emphysema markers on computed tomography, and a modification of a recently described multifunctional index (modified BODE).

Results: Overall, high mortality was seen in this cohort (12.7 deaths per 100 person-years; 292 total deaths). In multivariate analyses, increasing age (p=0.001), oxygen utilization (p=0.04), lower total lung capacity % predicted (p=0.05), higher residual volume % predicted (p=0.04), lower maximal cardiopulmonary exercise testing workload (p=0.002), greater proportion of emphysema in the lower lung zone versus the upper lung zone (p=0.005), and lower upper-to-lower-lung perfusion ratio (p=0.007), and modified BODE (p=0.02) were predictive of mortality. FEV1 was a significant predictor of mortality in univariate analysis (p=0.005), but not in multivariate analysis (p=0.21).

Conclusion: Although patients with advanced emphysema experience significant mortality, subgroups based on age, oxygen utilization, physiologic measures, exercise capacity, and emphysema distribution identify those at increased risk of death.

Figure 1.

Kaplan-Meier estimates of the probability of death as a function of number of years after randomization for medically treated patients segregated by age. The p value was derived from the log rank test for the comparison between subgroups over a median follow-up period of 3.9 yr.

Figure 2.

Kaplan-Meier estimates of the probability of death as a function of number of years after randomization for medically treated patients segregated by modified BODE index. The p value was derived from the log rank test for the comparison between subgroups over a median follow-up period of 3.9 yr.