Lung Function Trajectories and Associated Mortality among Adults with and without Airway Obstruction

Abstract

Rationale: Spirometry is essential for diagnosis and assessment of prognosis in patients with chronic obstructive pulmonary disease (COPD). Objectives: To identify FEV₁ trajectories and their determinants on the basis of annual spirometry measurements among individuals with and without airway obstruction (AO) and to assess mortality in relation to trajectories. Methods: From 2002 through 2004, individuals with AO (FEV₁/VC < 0.70, n = 993) and age- and sex-matched nonobstructive (NO) referents were recruited from population-based cohorts. Annual spirometry until 2014 was used in joint-survival latent-class mixed models to identify lung function trajectories. Mortality data were collected during 15 years of follow-up. Measurements and Main Results: Three trajectories were identified among the subjects with AO and two among the NO referents. Trajectory membership was driven by baseline FEV₁% predicted (FEV₁%pred) in both groups and also by pack-years in subjects with AO and current smoking in NO referents. Longitudinal FEV₁%pred depended on baseline FEV₁%pred, pack-years, and obesity. The trajectories were distributed as follows: among individuals with AO, 79.6% in AO trajectory 1 (FEV₁ high with normal decline), 12.8% in AO trajectory 2 (FEV₁ high with rapid decline), and 7.7% in AO trajectory 3 (FEV₁ low with normal decline) (mean, 27, 72, and 26 ml/yr, respectively) and, among NO referents, 96.7% in NO trajectory 1 (FEV₁ high with normal decline) and 3.3% in NO trajectory 2 (FEV₁ high with rapid decline) (mean, 34 and 173 ml/yr, respectively). Hazard for death was increased for AO trajectories 2 (hazard ratio [HR], 1.56) and 3 (HR, 3.45) versus AO trajectory 1 and for NO trajectory 2 (HR, 2.99) versus NO trajectory 1. Conclusions: Three different FEV₁ trajectories were identified among subjects with AO and two among NO referents, with different outcomes in terms of FEV₁ decline and mortality. The FEV₁ trajectories among subjects with AO and the relationship between low FVC and trajectory outcome are of particular clinical interest.

Keywords: FEV1; chronic obstructive pulmonary disease; natural history; prognosis.

Figure 1.

FEV₁% predicted (FEV₁%pred) trajectories estimated using joint-survival latent class mixed models separately among 993 individuals with airway obstruction (AO) (FEV₁/VC < 0.70) and 993 nonobstructive (NO) referents without AO (FEV₁/VC ⩾ 0.70) at baseline. The solid lines represent the observed means of FEV₁%pred, and the shaded areas represent 95% confidence intervals of the means. For subjects with AO who participated in bronchodilation testing, the highest FEV₁%pred values from pre- or postbronchodilation testing were used as outcomes. For the NO referents, pre-BD FEV₁%pred values were used as outcomes. BD = bronchodilatory; T1 = trajectory 1 (FEV₁ high with normal decline); T2 = trajectory 2 (FEV₁ high with rapid decline); T3 = trajectory 3 (FEV₁ low with normal decline).

Figure 2.

Histograms of individual prebronchodilatory FEV₁ decline (ml/yr) estimates within each of the trajectories, separately among individuals with airway obstruction (FEV₁/VC < 0.70) and referents without airway obstruction (FEV₁/VC ⩾ 0.70) at baseline. T1 = trajectory 1; T2 = trajectory 2; T3 = trajectory 3.

Figure 3.

Survival functions for each trajectory and HRs with 95% CIs (T1 as reference), estimated using Cox regression models including sex, age, pack-years of smoking, and overweight and obesity as covariates, separately among 993 subjects with AO and 993 age- and sex-matched NO referents. AO = airway obstruction; CI = confidence interval; HR = hazard ratio; NO = nonobstructive; T1 = trajectory 1 (FEV₁ high with normal decline); T2 = trajectory 2 (FEV₁ high with rapid decline); T3 = trajectory 3 (FEV₁ low with normal decline).