General population-based lung function trajectories over the life course: an accelerated cohort study

Abstract

Background: Lung function is a key determinant of health, but current knowledge on lung function growth and decline over the life course is based on fragmented, potentially biased data. We aimed to empirically derive general population-based life course lung function trajectories, and to identify breakpoints and plateaus.

Methods: We created an accelerated cohort by pooling data from eight general population-based child and adult cohort studies from Europe and Australia. We included all participants with information on lung function, smoking status, BMI, and asthma diagnosis status from at least two visits. We used cross-classified three-level linear mixed models to derive sex-specific life course trajectories of FEV₁, forced vital capacity (FVC), and FEV₁/FVC ratio based on observations at ages 4-80 years, and Bayesian time-series decomposition to identify breakpoints and plateaus. We repeated sex-specific analyses with separate stratification for asthma status (never had asthma vs persistent asthma, where persistent was defined as the risk factor being present at all participant visits) and smoking status (never smoker vs persistent smoker).

Findings: The accelerated cohort included 30 438 participants born between 1901 and 2006 (15 703 [51·6%] female and 14 735 [48·4%] male; mean age 26 [SD 16] years), who provided a total of 87 666 observations (range 2-8 observations per participant). In female participants, FEV₁ increased non-linearly in two phases, at a mean of 234 (95% CI 223 to 245) mL/year until age 13 (95% credible interval [CrI] 12 to 15) years, then at 99 (76 to 122) mL/year until a peak at age 20 (18 to 22) years, and subsequently decreased throughout the rest of adulthood (-26 [-27 to -25] mL/year). In male participants, the pattern was similar, with an increase in FEV₁ of 271 (263 to 280) mL/year until age 16 (14 to 18) years, which slowed to 108 (93 to 124) mL/year until reaching a maximum at age 23 (21 to 25) years, decreasing thereafter (-38 [-39 to -37] mL/year), representing a later peak than in female participants. In female participants, FVC increased non-linearly in two phases, at 232 (95% CI 222 to 243) mL/year until age 14 (95% CrI 12 to 15) years, then at 77 (59 to 94) mL/year until peaking at age 20 (19 to 22) years, after which it decreased throughout the rest of adulthood (-26 [-27 to -25] mL/year). In male participants, FVC also increased in two phases, at 326 (315 to 337) mL/year until age 15 (13 to 17) years, then at 156 (144 to 168) mL/year until a peak at 23 (19 to 30) years, and subsequently declined in two phases (-22 [-29 to -14] mL/year until age 42 [38 to 50] years, then -36 [-38 to -34] mL/year thereafter). No plateau after the peak was observed for either lung function parameter in both sexes. FEV₁/FVC ratio decreased throughout life from the starting age of 4 years in both sexes with some distinct patterns. Stratified analysis showed that persistent asthma (vs never had asthma) was related to an earlier FEV₁ peak, lower FEV₁ throughout adulthood, and lower FEV₁/FVC ratio across the life course in both sexes. Persistent smoking (vs never smoking) was related to an accelerated decrease in FEV₁ and FEV₁/FVC ratio during adulthood in both sexes. No statistically significant plateau was observed in any lung function parameter across the strata of asthma or smoking status.

Interpretation: In both sexes, FEV₁ and FVC increased in two phases, with a fast increase until around age 13-16 years, and then a slower increase until a peak. Neither parameter showed a plateau phase after the peak, and decreases started earlier than previously described. FEV₁/FVC ratio decreased throughout life. These observations provide an essential, but previously unavailable, framework to assess and monitor lung health over the life course.

Funding: EU Horizon 2020, Wellcome, European Respiratory Society, AstraZeneca, Chiesi, GSK, Menarini Group, and Sanofi.

Figure 1

Summary of participant visits, ages, and follow-up times* according to source cohort study, and contribution of each cohort study to the accelerated cohort (absolute observation numbers) by 5-year age groups In the bubble graph and histogram, the unit of analysis is the observation (ie, person-visit). Min=minimum. Max=maximum. *This information might not match the source study characteristics, as this table reports data for participants included in the current analysis only. See the Methods section and appendix (pp 2–5) for details.

Figure 2

Frequency of current asthma, active smoking, and obesity by year of age in the accelerated cohort* Frequency (%) is represented by the red solid line, with 95% CIs shown in pink. Current asthma, active smoking, and obesity trends were smoothed with use of locally estimated scatterplot smoothing with a span of 0·4, applied separately to the proportions and their 95% CIs. Frequency values by year of age are provided in the appendix (pp 17–18). *The unit of analysis is the observation (ie, person-visit). Each participant contributed in at least 2 years and up to 8 years, depending on the number of visits (observations).

Figure 3

Sex-specific FEV₁* trajectories during the life course (A) FEV₁ trajectories in each cohort study and in the accelerated cohort, with 95% CIs for the accelerated cohort. (B) Mean age (with 95% CrIs) at lung function breakpoints and mean lung function change (with 95% CIs) between breakpoints. FEV₁ is expressed in litres (L) and change in FEV₁ is expressed in mL/year. Vertical dashed lines indicate the breakpoints, blue shading indicates 95% CrIs, and arrows indicate the mean FEV₁ change between breakpoints or the mean age at breakpoints. (C) Probability of positive change (light blue), plateau (dark blue line), and negative change (blue) at each year of age of the trajectory. 95% CrI=95% credible interval. *Height-standardised and study-standardised (appendix p 15).

Figure 4

Sex-specific FVC* trajectories during the life course (A) FVC trajectories in each cohort study and in the accelerated cohort, with 95% CIs for the accelerated cohort. (B) Mean age (with 95% CrIs) at lung function breakpoints and mean lung function change (with 95% CIs) between breakpoints. FVC is expressed in litres (L) and change in FVC is expressed in mL/year. Vertical dashed lines indicate the breakpoints, blue shading indicates 95% CrIs, and arrows indicate the mean FVC change between breakpoints or the mean age at breakpoints. (C) Probability of positive change (light blue), plateau (dark blue line), and negative change (blue) at each year of age of the trajectory. 95% CrI=95% credible interval. FVC=forced vital capacity. *Height-standardised and study-standardised (appendix p 15).