Clinical relevance of lung function trajectory clusters in middle-aged and older adults

Xander Bertels, James C. Ross, Rosa Faner, Michael H. Cho, M. Arfan Ikram, Guy G. Brusselle, Lies Lahousse
ERJ Open Research 2024 10: 00793-2023; DOI: 10.1183/23120541.00793-2023

Abstract

Background The determinants and health outcomes of lung function trajectories in adults among the general population are poorly understood. We aimed to identify and characterise clusters of lung function trajectories in adults aged ≥45 years.

Methods Gaussian finite-mixture modelling was applied to baseline and annualised change of forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC) and FEV1/FVC ratio z-scores in participants of the Rotterdam Study, a prospective population-based cohort study, with repeated spirometry (n=3884; mean±sd age 64.7±8.9 years). Longitudinal outcomes were all-cause mortality, respiratory outcomes (symptoms, COPD (FEV1/FVC <0.7 in absence of asthma), preserved ratio impaired spirometry (PRISm; FEV1/FVC ≥0.7 and FEV1 or FVC <80%)), smoking cessation and weight changes. Independent risk factors, including genetics, were identified by multiple logistic regression.

Results We identified eight trajectory clusters, with the reference group having persistently normal spirometry (prevalence 42.8%). Three clusters showed higher mortality, adjusted for confounders: 1) the persistently low FEV1 cluster (prevalence 6.8%, hazard ratio (HR) 1.71, 95% CI 1.37–2.13); 2) rapid FEV1 decliners (prevalence 4.6%, HR 1.48, 95% CI 1.10–1.99); and 3) FVC decliners (prevalence 3.7%, HR 1.49, 95% CI 1.09–2.03). In contrast, FVC improvers (prevalence 6.7%, HR 0.61, 95% CI 0.41–0.90) and persistently high FEV1 (prevalence 29.2%, HR 0.82, 95% CI 0.69–0.98) were protective trajectory clusters. Clusters were characterised by differences in genetic predisposition (polygenic scores of FEV1 and FEV1/FVC), demographics, cigarette smoking, respiratory symptoms (chronic cough, wheezing and dyspnoea), cardiovascular factors (body mass index, hypertension and heart failure) and serum C-reactive protein levels. Frailty, weight changes and the development of respiratory symptoms, COPD and PRISm were significantly associated with trajectory clusters.

Conclusions This study reveals clinically relevant lung function trajectory clusters in older adults of the general population.

Tweetable abstract

Middle-aged and older adults exhibit eight distinct lung function trajectory clusters, which are differentially associated with all-cause mortality and incident morbidity https://bit.ly/40oFUDC

Introduction

Lung function is a marker of general health and survival [1]. Achieving and maintaining normal lung function is challenged in two ways. First, early-life lung development and growth can be impaired, leading to a suboptimal maximal lung function at young adulthood. Second, age-related decline of lung function may be further accelerated during adulthood. These phenomena probably result from early-life events (e.g. preterm birth, lower respiratory tract infections and allergen exposure) as well as gene–environment interactions upon exposure to noxious gases and particles (e.g. cigarette smoke, air pollution and occupational hazards) [2–6]. Consequently, individuals have different patterns of lung function over the life-course, referred to as the lung function trajectome [7]. It is hypothesised that individual lung function trajectories can be used to better phenotype complex respiratory diseases, such as COPD, given that different trajectories may lead to COPD and reflect different underlying mechanisms which may require differential treatment [7].

Several studies on lung function trajectories in the general population explored trajectories in childhood or young adulthood [3]. Fewer studies included follow-up data of older persons or started modelling at an older age [8–11]. Birth cohorts revealed growth failure trajectories in children [12] and studies in young adults showed the existence of rapid decline trajectories with a prevalence between 1% and 6% [4, 8, 13]. Furthermore, Lange et al. [10]. demonstrated that COPD in late adulthood is driven by both low peak lung function and accelerated forced expiratory volume in 1 s (FEV1) decline. In patients with established COPD, several studies have shown substantial variability in lung function decline [14–16].

Various factors have been associated with accelerated FEV1 and forced vital capacity (FVC) decline in the general population including smoking exposure, occupational exposures (e.g. aromatic solvents), air pollution (e.g. nitrogen dioxide exposure), socioeconomic status, physical activity level, body mass index (BMI), hypertension, and glutathione-S-transferase deficiency [3, 17–21]. Additional risk factors for FEV1 decline include asthma [3, 17] and single-nucleotide polymorphisms in the ADAM33 gene [22].

Yet, overall, the prevalence, characteristics and clinical importance of lung function trajectories starting in middle-aged and older adults from the general population remain uncertain. This is especially relevant since older adults are expected to have the highest short-term risk of morbidity and mortality related to harmful trajectories. Additionally, the study of individual trajectories remains challenging and mainly focused on modelling FEV1, FVC or ratio separately. Our study clustered within-person lung function trajectories in middle-aged and older adults from the general population using all three indices simultaneously, which may provide more granular trajectories and shed light on the associated factors. We aimed 1) to examine the prevalence and characteristics of such trajectory clusters; 2) to compare all-cause mortality between clusters; and 3) to compare within-trajectory health outcomes between clusters including the incidence of respiratory symptoms and diseases.

Methods

Study population

This study was embedded in the Rotterdam Study, a prospective population-based cohort study of individuals aged ≥45 years. Detailed information on the Rotterdam Study has been published [23] and is summarised in the supplementary material. The study population of this analysis included participants with repeated spirometry at different study visits between 2002 and 2016 (supplementary figure S1). Pre-bronchodilator spirometry was performed according to the American Thoracic Society (ATS)/European Respiratory Society (ERS) guidelines by trained paramedical personnel. Participants had two or three spirometric measurements, in which baseline was defined as the participant's first study visit including spirometry.

Trajectory clustering

Z-scores of FEV1, FVC and FEV1/FVC were calculated using the Global Lung Initiative (GLI) reference equations [24] and represent departures from the expected lung function for a given age, sex, height and ethnicity (supplementary figure S2). Model-based clustering was applied on the baseline value (z-score) and annualised rate of decline (scaled z-score per year) of FEV1, FVC and FEV1/FVC to cluster individuals with a phenotypically similar lung function trajectory (supplementary table S1). Annualised rate of decline was calculated as the average decline per year using the first and last spirometry. Briefly, finite Gaussian mixture modelling was applied using an expectation-maximisation algorithm, which was initialised by complete-linkage hierarchical clustering to avoid local maxima. One to 10 clusters using spherically distributed models with equal variance per component were considered. The optimal number of clusters was selected based on maximal integrated complete-data likelihood and confirmed using bootstrap sequential likelihood ratio test (LRT) [25]. Robustness was evaluated using different initialisation and linkage methods (supplementary table S2).

Characteristics and outcomes

Detailed information on spirometry, baseline characteristics (supplementary figure S3) and outcomes are provided in the supplementary material. The longitudinal health outcomes were all-cause mortality (post-trajectory data) and the incidence of COPD, preserved ratio impaired spirometry (PRISm), respiratory symptoms, smoking cessation and weight changes (within-trajectory data). An additional cross-sectional analysis was performed to assess the association with frailty status (at the last study visit with spirometry). COPD was defined by a pre-bronchodilator FEV1/FVC ratio <0.7 in participants without asthma. Asthma was defined by an ever physician diagnosis of asthma retrieved from medical records. PRISm was defined as FEV1/FVC ≥0.7 and FEV1 or FVC <80%. Polygenic scores (PGS) of FEV1 and FEV1/FVC were based on Moll et al. [26] and calculated using the PGS Catalog [27].

Statistical analysis

An α-level of 0.05 was considered statistically significant. Normally distributed continuous variables were expressed as mean±sd. Non-normally distributed continuous variables were expressed as median and interquartile range (IQR). Mean differences were tested using Welsh's two-sample t-test and the two-sample z-test for equality of proportions, as appropriate. Logistic regression models were fitted for every cluster individually using the variables that were significantly different at baseline compared with the reference cluster. Incidences rates (new onset) were expressed per 1000 person-years and differences were tested by exact rate ratio test. Differences in all-cause mortality were evaluated by Cox proportional hazard models and adjusted for age, sex, BMI, education level, smoking status, COPD, asthma, hypertension and treated heart failure (supplementary figure S4). Proportionality was checked using Schoenfeld residuals and time-dependencies were addressed by a step function [28]. Statistical analyses were performed in R 4.1.1 (Vienna, Austria) using the “mclust”, “tidyverse”, “survival” and “epiR” packages [25, 29–31].

Results

Study population

The study population (n=3884) consisted of middle-aged to older adults (mean±sd age 64.7±8.9 years). Participants had two (n=2992; 77.0%) or three (n=892; 23.0%) spirometry visits over a median (IQR) period of 5.9 (4.5) years, with a maximum period of 13.3 years (supplementary tables S3 and S4). A total of 35 outliers were excluded from the analysis based on the 99.999th percentile (n=3849). Compared to individuals with only one spirometry measurement (excluded), individuals with repeated spirometry (study population) were significantly younger; had a higher education level; were less often smokers; had less often respiratory symptoms, COPD, PRISm, hypertension or heart failure; and showed lower serum C-reactive protein (CRP) levels (supplementary table S5).

Trajectory clusters

A model consisting of eight trajectory clusters was selected based on integrated completed likelihood maximisation (supplementary figure S5). Bootstrap sequential LRT confirmed that a model of eight clusters showed a significantly better model likelihood than a model of seven clusters (p<0.001). Trajectory clusters were labelled by their main feature in terms of FEV1 and FVC patterns (figure 1). The prevalence and progression patterns are presented in table 1 and in supplementary figures S6 and S7. The largest trajectory cluster, comprising 42.8% of participants, exhibited persistently normal spirometry and was defined as the reference group for all comparisons. This cluster showed a median (IQR) decline of −240 (−380–−100) mL in FEV1 and −190 (−360–−10) mL in FVC over a median period of 6.0 years (supplementary table S4).

FIGURE 1
FIGURE 1

Graphical representation of lung function trajectory clusters in middle-aged and older adults. Baseline lung function in Global Lung Initiative z-scores; change in lung function calculated as annualised rate of decline (z-score per year). FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; +: higher baseline lung function; −: lower baseline lung function; ↑: attenuated decline or improvement; ↓: accelerated decline; no symbol: average z-score between −1.00 and +1.00; one symbol: average z-score between −1.65 and −1.00 or between +1.00 and +1.65; two symbols: average z-score lower than −1.65 or higher than +1.65.

View this table:
TABLE 1

Prevalence and progression patterns of lung function trajectory clusters

Baseline characteristics

Baseline characteristics of the trajectory clusters are presented in table 2. In terms of general characteristics, the persistently low FEV1 cluster showed the highest baseline age, lowest education level, most current smokers and pack-years, and fewest never-smokers. Conversely, the persistently high FEV1 cluster was the youngest, had the most never-smokers, and the fewest current smokers and pack-years. The average BMI was highest in low-start improvers. In terms of age distribution, FVC decliners and FVC improvers consisted of a higher proportion of middle-aged individuals compared to older adults (supplementary table S6).

View this table:
TABLE 2

Baseline characteristics of lung function trajectory clusters

Respiratory symptoms (chronic cough, chronic sputum production, wheezing and dyspnoea) were present in 6–22% of participants in the reference cluster. Overall, the highest prevalence of symptoms (18–51%) was observed in the persistently low FEV1 cluster. The persistently high FEV1 cluster showed the lowest prevalence of wheezing (13.8%) and dyspnoea (4.9%), while FVC improvers showed the lowest prevalence of chronic cough (4.7%) and chronic sputum production (2.7%).

Individuals with persistently low FEV1 demonstrated the highest prevalence of COPD (64%), while the lowest prevalence of COPD (0.4%) was observed in FVC improvers followed by the persistently high FEV1 cluster. PRISm and hypertension were most prevalent in low-start improvers and least prevalent in high-start rapid decliners. The highest prevalence of treated heart failure was observed in rapid FEV1 decliners.

Regarding genetic predisposition and biomarkers, individuals with persistently low FEV1 and FVC decliners demonstrated a significantly higher PGS for an impaired ratio (FEV1/FVC). The persistently low FEV1 cluster also showed the highest CRP levels. In contrast, those with persistently high FEV1 had a lower PGS for impaired FEV1 and had lower CRP values.

In a subgroup with diffusion capacity testing, the persistently low FEV1 group exhibited significantly lower DLCO values, whereas persistently high FEV1 showed the highest DLCO values. Carbon monoxide transfer coefficient values were lowest in high-start rapid decliners. Associations with diffusion data remained statistically significant after adjusting for smoking status (supplementary table S8).

Independent risk factors for persistently low FEV1 included older age, lower education level, current smoking, higher pack-years, wheezing, dyspnoea and a higher PGS for impaired FEV1/FVC (supplementary table S9). Risk factors for the rapid FEV1 decliners were chronic cough and treated heart failure. A higher PGS for impaired FEV1/FVC, current smoking and the absence of hypertension were identified as independent factors for FVC decliners. Similarly, the absence of hypertension was associated with a higher risk for being a high-start rapid decliner. Younger age was independently associated with FVC improvers, while higher BMI was associated with low-start improvers. Lastly, the persistently high FEV1 cluster was independently associated with a lower PGS for impaired FEV1, the absence of current smoking, wheezing and hypertension, as well as lower CRP levels.

Mortality

Participants with persistently low FEV1, rapid FEV1 decliners and FVC decliners were at a significantly higher risk of all-cause mortality, adjusted for age and sex, compared to the reference cluster (table 3). In contrast, FVC improvers and participants with persistently high FEV1 showed a lower mortality hazard (table 3). These associations were driven by participants aged ≥65 years (supplementary table S10) and remained statistically significant when additionally adjusting for BMI, education level, smoking status, COPD, asthma, hypertension and heart failure. Furthermore, there was no evidence for higher mortality in low-start improvers, despite this group having the lowest baseline FVC. The relationship between all-cause mortality and lung function trajectory clusters is plotted in figure 2.

View this table:
TABLE 3

All-cause mortality of trajectory clusters

FIGURE 2
FIGURE 2

Lung function trajectory clusters and all-cause mortality. Baseline lung function expressed in Global Lung Initiative z-scores and change calculated as an annualised rate of decline (z-score per year). Cox proportional hazards model was adjusted for age and sex (model 1). FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; HR hazard ratio.

Morbidity

Participants with persistently low FEV1 demonstrated a significantly higher incidence rate of new onset chronic cough, wheezing and dyspnoea. Additionally, this cluster was associated with a higher incidence of clinically relevant weight loss and chronic sputum production, and had the highest prevalence of frailty at the last spirometry visit (table 4).

View this table:
TABLE 4

Within-trajectory morbidity outcomes

All declining clusters (rapid FEV1 decliners, high-start rapid decliners and FVC decliners) showed a higher incidence of weight gain. Rapid FEV1 decliners additionally showed the highest incidence of COPD, chronic cough and sputum production, and were at a higher risk for incident wheezing and PRISm. FVC decliners also demonstrated a higher incidence of PRISm, whereas they had the lowest incidence of COPD.

Low-start improvers showed the highest incidence of weight loss and, together with FVC improvers, a higher incidence of PRISm resolving to normal spirometry.

Persistently high FEV1 was protective against incident wheezing, COPD and PRISm, and showed a lower prevalence of frail individuals at the last spirometry visit. None of the clusters showed a significant association with the incidence of smoking cessation, although the percentage of quitters was borderline significantly higher in the persistently high FEV1 cluster (+12.4 percentage points, 95% CI −0.5–+25.3 percentage points).

Discussion

Eight distinct lung function trajectory clusters were identified in middle-aged and older adults based on longitudinal patterns of FEV1, FVC and FEV1/FVC. Our findings show that the progression of lung function over time is heterogeneous and informative for respiratory and overall health beyond disease labels in older adults from the general population. Independent risk factors included genetic predisposition for impaired FEV1 and FEV1/FVC, demographics (age and education level), cigarette smoking, respiratory symptoms (chronic cough, wheezing and dyspnoea), cardiovascular factors (BMI, hypertension and heart failure) and inflammation markers (CRP). Secondly, three lung function trajectory clusters were at a higher mortality risk: 1) individuals with persistently low FEV1; 2) rapid FEV1 decliners; and 3) FVC decliners. Two trajectory clusters were protective for mortality: 1) FVC improvers and 2) individuals with persistently high FEV1. Thirdly, trajectory clusters were differentially associated with frailty, weight changes, and the development of respiratory symptoms, COPD and PRISm over time. Overall, our results suggest that in a general population of older adults, two spirometric measurements over a period of 5 years may describe clinically relevant short-term lung function trajectories.

A novel phenotype of FVC improvers was observed in ∼7% of middle-aged and older adults. Participants in this cluster demonstrated a median improvement of +345 mL FVC over 5.8 years. Promisingly, this trajectory cluster showed lower all-cause mortality than the reference cluster. In line with the overall importance of FVC on survival [32], these results show that resilience toward FVC decline is a protective factor for overall health in middle-aged and older adults.

Approximately 3% of the study population exhibited low baseline spirometry levels followed by improvements in both FEV1 and FVC (low-start improvers). Even though this cluster had the lowest baseline FVC values, it did not show higher mortality, nor an increased development of symptoms or diseases, in sheer contrast to those with persistently low FEV1, who had similar low baseline spirometric values. ∼60% of low-start improvers had PRISm at baseline, which was generally followed by a reversal to normal spirometry in subsequent years. This PRISm-to-normal phenotype has been described previously in both young and older adults [33, 34].

Low FVC at baseline may be the result of a myriad factors, including an early-life trajectory of reduced lung growth and development [11] or air trapping [35] (e.g. due to small airway disease). Low-start improvers showed the largest incidence of weight loss and a significantly lower prevalence of frailty (−11.0 percentage points, 95% CI −17.0–−4.7 percentage points) compared to individuals with persistently low FEV1, indicating potential opportunities for improving lung health. This is consistent with previous studies showing beneficial effects of weight loss in obese individuals [17, 21]. Other mechanisms explaining lung function improvements may include incident bronchodilator use, reversal of air trapping, improving health status, increased physical activity and heart failure control. Smoking cessation was not associated with improving clusters over a relatively short-term period, yet previous studies showed that its effects are more apparent as long-term attenuation of lung function decline in sustained quitters [36].

Furthermore, we identified two trajectory clusters leading to both a low FEV1/FVC ratio and low FVC in older adults: one consisting of individuals with persistently low FEV1 (∼7%) who also showed lower FVC values, and rapid FEV1 decliners (∼5%) who concomitantly showed accelerated FVC decline, albeit to a lesser extent. Such mixed spirometry patterns were also observed recently in a younger age range (7–53 years) of the general population [11]. Both trajectory clusters were associated with higher mortality and showed the highest incidence of respiratory symptoms, COPD and PRISm. Persistently low FEV1 was also associated with frailty suggesting poor resilience among these persons. Most notably, this cluster was significantly more genetically predisposed to impaired FEV1/FVC while this was not the case in rapid FEV1 decliners. This supports that the genetic variants included in the FEV1/FVC PGS [26] (∼1.2 million) are associated with COPD age of onset [37] and that these variants might be more informative for incident airflow obstruction before the age of 65 years. Risk factors for rapid FEV1 decline included chronic cough, in line with previous literature [38], and heart failure. Rapid FEV1 decline is a known risk factor for incident heart failure [39], yet our study adds that this acceleration may persist in individuals with prevalent heart failure.

Additionally, our study provides evidence for a trajectory of FVC decline (∼4%), which was associated with higher mortality and PRISm development. Interestingly, we also observed a cluster of high-start rapid decliners (∼3%). This latter group did not show increased morbidity over time, nor higher mortality. This may indicate that their supranormal decline did not persist in the long term and did not compromise their functional capacities due to their protective baseline values. Smoking, genetic predisposition for low FEV1/FVC and the absence of (treated) hypertension were independent risk factors for FVC decline in older adults of the general population.

Together, ∼11% of this population demonstrated rapid FEV1 and/or FVC decline, exceeding the proportion of rapid decliners identified in other studies [8, 11, 13, 40]. This may be explained by the fact that our study clustered (short-term) individual trajectories and did not investigate life-long trajectories, therefore also capturing more extreme trajectories. Indeed, rapid decline which persists over several decades may be relatively rare due to incompatibility with life and is therefore harder to observe due to loss-to-follow-up [1].

Study strengths include 1) generalisability of our findings, as no exclusion criteria were defined; 2) longitudinal morbidity and mortality data in a large cohort of middle-aged and older adults; and 3) the data-driven way of phenotyping participants based on the progression of FEV1, FVC and FEV1/FVC simultaneously, which led to a comprehensive set of distinct trajectory clusters.

Limitations of our study include 1) selection bias due to the repeated spirometry inclusion criterion, initial cohort response and attrition before the first research visit including spirometry; 2) the use of pre-bronchodilator spirometry; and 3) the lack of post-trajectory morbidity data. Furthermore, as spirometry is effort-dependent, an underestimation of FVC due to poor effort cannot be ruled out. Future studies should investigate whether case-finding by repeated spirometry is justified in this population and disentangle the mechanisms which drive improving trajectory clusters.

To conclude, we demonstrated that middle-aged and older adults show distinct clusters of lung function trajectories, which appear to be clinically relevant.

Supplementary material

Supplementary Material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material 00793-2023.SUPPLEMENT

Acknowledgements

The Rotterdam Study is funded by Erasmus Medical Center and Erasmus University, Rotterdam; Netherlands Organisation for Health Research and Development (ZonMw); the Research Institute for Diseases in the Elderly (RIDE); the Ministry of Education, Culture and Science, the Ministry for Health, Welfare and Sports; the European Commission (DG XII); and the Municipality of Rotterdam. The authors are grateful to the study participants, the staff from the Rotterdam Study and the participating general practitioners and pharmacists.

The generation and management of genome-wide associated study (GWAS) genotype data for the Rotterdam Study (RS-I, RS-II and RS-III) was executed by the Human Genotyping Facility of the Genetic Laboratory of the Department of Internal Medicine, Erasmus MC, Rotterdam, the Netherlands. The GWAS datasets are supported by the Netherlands Organisation of Scientific Research (NOW) Investments (number 175.010.2005.011, 911–03–012); the Genetic Laboratory of the Department of Internal Medicine, Erasmus MC; RIDE (014–93–015; RIDE2); and the Netherlands Genomics Initiative/NWO Netherlands Consortium for Healthy Aging (project number 050–060–810). We thank Pascal Arp, Mila Jhamai, Marijn Verkerk, Lizbeth Herrera, Marjolein Peters and Carolina Medina-Gomez (Erasmus MC, Rotterdam, the Netherlands) for their help in creating the GWAS database. This study makes use of sequence reference data generated by the UK10K Consortium, derived from samples from the ALSPAC and TwinsUK datasets. A full list of the investigators who contributed to the generation of the data is available from www.UK10K.org. Funding for UK10K was provided by the Wellcome Trust under award WT091310. We thank Jie Huang at the Wellcome Trust's Sanger Institute, Hinxton, UK for the creation of imputed data, with the support of Marijn Verkerk, Carolina Medina-Gomez and Anis Abuseiris (Erasmus MC) and their input for the analysis setup.

Footnotes

  • Provenance: Submitted article, peer reviewed.

  • Ethics statement: The medical ethics committee of the Erasmus Medical Centre (Rotterdam, the Netherlands) and the review board of the Netherlands Ministry of Health, Welfare and Sports (1068889-159521-PG) approved the study.

  • Author contributions: X. Bertels designed and planned this analysis guided by L. Lahousse. Methodological advice was provided by L. Lahousse, J.C. Ross, M.H. Cho and R. Faner. Formal analysis and drafting of manuscript was performed by X. Bertels. Critical revision and expert feedback were received from J.C. Ross, M.H. Cho, R. Faner, M.A. Ikram, G.G. Brusselle and L. Lahousse.

  • Conflict of interest: X. Bertels reports BOF.SIP.2020 funding for international mobility in relation to the submitted work.

  • Conflict of interest: R. Faner reports grants from the Serra Húnter Program, Instituto de Salud Carlos III (PI21/00735) and European Research Council under the Horizon Europe research and innovation programme (101044387) outside the submitted work.

  • Conflict of interest: M.H. Cho reports NHLBI funding (R01HL153248, R01HL149861 and R01HL147148), grants from Bayer and GSK, consulting fees from Genentech, AstraZeneca and Illumina, and honoraria for lectures from Genentech, AstraZeneca and Illumina outside the submitted work.

  • Conflict of interest: G.G. Brusselle reports fees for advisory boards and/or lectures from AstraZeneca, Boehringer Ingelheim, Chiesi, GSK, Merck Sharp & Dohme, Novartis and Sanofi Regeneron outside the submitted work.

  • Conflict of interest: L. Lahousse reports BOF.SIP.2020 funding for international mobility in relation to the submitted work, consulting fees from AstraZeneca, and speaking/lecture fees from Chiesi and IPSA (non-profit) outside the submitted work.

  • Conflict of interest: J.C. Ross and M.A. Irfan declare no conflicts of interest.

  • Support statement: Mobility of X. Bertels and L. Lahousse was supported by BOF.SIP.2020. M.H. Cho was supported by R01HL153248, R01HL149861 and R01HL147148. R. Faner was supported by the Serra Húnter Program, PI21/00735 and the European Research Council under the Horizon Europe research and innovation programme (101044387). Funding information for this article has been deposited with the Crossref Funder Registry.

  • Received October 20, 2023.
  • Accepted October 27, 2023.
http://creativecommons.org/licenses/by-nc/4.0/

This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org

References

  1. ↵
    1. Papi A,
    2. Beghé B,
    3. Fabbri LM
    . Rate of decline of FEV1 as a biomarker of survival? Am J Respir Crit Care Med 2021; 203: 663–665. doi:10.1164/rccm.202010-3784ED
    OpenUrl
  2. ↵
    1. Cho MH,
    2. Hobbs BD,
    3. Silverman EK
    . Genetics of chronic obstructive pulmonary disease: understanding the pathobiology and heterogeneity of a complex disorder. Lancet Respir Med 2022; 10: 485–496. doi:10.1016/S2213-2600(21)00510-5
    OpenUrl
  3. ↵
    1. Okyere DO,
    2. Bui DS,
    3. Washko GR, et al.
    Predictors of lung function trajectories in population-based studies: a systematic review. Respirology 2021; 26: 938–959. doi:10.1111/resp.14142
    OpenUrl
  4. ↵
    1. Bui DS,
    2. Lodge CJ,
    3. Burgess JA, et al.
    Childhood predictors of lung function trajectories and future COPD risk: a prospective cohort study from the first to the sixth decade of life. Lancet Respir Med 2018; 6: 535–544. doi:10.1016/S2213-2600(18)30100-0
    OpenUrl
    1. Martinez FD
    . Early-life origins of chronic obstructive pulmonary disease. N Engl J Med 2016; 375: 871–878. doi:10.1056/NEJMra1603287
    OpenUrlCrossRefPubMed
  6. ↵
    1. Fletcher C,
    2. Peto R
    . The natural history of chronic airflow obstruction. Br Med J 1977; 1: 1645–1648. doi:10.1136/bmj.1.6077.1645
    OpenUrlAbstract/FREE Full Text
  7. ↵
    1. Martinez FJ,
    2. Agusti A,
    3. Celli BR, et al.
    Treatment trials in young patients with chronic obstructive pulmonary disease and pre-chronic obstructive pulmonary disease patients: time to move forward. Am J Respir Crit Care Med 2022; 205: 275–287. doi:10.1164/rccm.202107-1663SO
    OpenUrlPubMed
  8. ↵
    1. van Oostrom SH,
    2. Engelfriet PM,
    3. Verschuren WMM, et al.
    Aging-related trajectories of lung function in the general population – the Doetinchem Cohort Study. PLoS One 2018; 13: e0197250. doi:10.1371/journal.pone.0197250
    OpenUrl
    1. Ross JC,
    2. Castaldi PJ,
    3. Cho MH, et al.
    Longitudinal modeling of lung function trajectories in smokers with and without chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2018; 198: 1033–1042. doi:10.1164/rccm.201707-1405OC
    OpenUrl
  10. ↵
    1. Lange P,
    2. Celli B,
    3. Agustí A, et al.
    Lung-function trajectories leading to chronic obstructive pulmonary disease. N Engl J Med 2015; 373: 111–122. doi:10.1056/NEJMoa1411532
    OpenUrlCrossRefPubMed
  11. ↵
    1. Dharmage SC,
    2. Bui DS,
    3. Walters EH, et al.
    Lifetime spirometry patterns of obstruction and restriction, and their risk factors and outcomes: a prospective cohort study. Lancet Respir Med 2023; 11: 273–282. doi:10.1016/S2213-2600(22)00364-2
    OpenUrl
  12. ↵
    1. Wang G,
    2. Hallberg J,
    3. Faner R, et al.
    Plasticity of individual lung function states from childhood to adulthood. Am J Respir Crit Care Med 2023; 207: 406–415. doi:10.1164/rccm.202203-0444OC
    OpenUrl
  13. ↵
    1. Washko GR,
    2. Colangelo LA,
    3. Estepar RSJ, et al.
    Adult life-course trajectories of lung function and the development of emphysema: the CARDIA lung study. Am J Med 2020; 133: 222–230.e11. doi:10.1016/j.amjmed.2019.06.049
    OpenUrl
  14. ↵
    1. Vestbo J,
    2. Edwards LD,
    3. Scanlon PD, et al.
    Changes in forced expiratory volume in 1 second over time in COPD. N Engl J Med 2011; 365: 1184–1192. doi:10.1056/NEJMoa1105482
    OpenUrlCrossRefPubMed
    1. Casanova C,
    2. de Torres JP,
    3. Aguirre-Jaíme A, et al.
    The progression of chronic obstructive pulmonary disease is heterogeneous: the experience of the BODE cohort. Am J Respir Crit Care Med 2011; 184: 1015–1021. doi:10.1164/rccm.201105-0831OC
    OpenUrlCrossRefPubMed
  16. ↵
    1. Celli B,
    2. Locantore N,
    3. Yates JC, et al.
    Markers of disease activity in COPD: an 8-year mortality study in the ECLIPSE cohort. Eur Respir J 2021; 57: 2001339. doi:10.1183/13993003.01339-2020
    OpenUrlAbstract/FREE Full Text
  17. ↵
    1. Bui DS,
    2. Perret JL,
    3. Walters EH, et al.
    Lifetime risk factors for pre- and post-bronchodilator lung function decline. A population-based study. Ann Am Thorac Soc 2020; 17: 302–312. doi:10.1513/AnnalsATS.201904-329OC
    OpenUrlPubMed
    1. Ramsay SE,
    2. Whincup PH,
    3. Lennon LT, et al.
    Longitudinal associations of socioeconomic position in childhood and adulthood with decline in lung function over 20 years: results from a population-based cohort of British men. Thorax 2011; 66: 1058–1064. doi:10.1136/thoraxjnl-2011-200621
    OpenUrlAbstract/FREE Full Text
    1. Garcia-Aymerich J,
    2. Lange P,
    3. Benet M, et al.
    Regular physical activity modifies smoking-related lung function decline and reduces risk of chronic obstructive pulmonary disease: a population-based cohort study. Am J Respir Crit Care Med 2007; 175: 458–463. doi:10.1164/rccm.200607-896OC
    OpenUrlCrossRefPubMed
    1. Miele CH,
    2. Grigsby MR,
    3. Siddharthan T, et al.
    Environmental exposures and systemic hypertension are risk factors for decline in lung function. Thorax 2018; 73: 1120–1127. doi:10.1136/thoraxjnl-2017-210477
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Peralta GP,
    2. Marcon A,
    3. Carsin AE, et al.
    Body mass index and weight change are associated with adult lung function trajectories: the prospective ECRHS study. Thorax 2020; 75: 313–320. doi:10.1136/thoraxjnl-2019-213880
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. van Diemen CC,
    2. Postma DS,
    3. Vonk JM, et al.
    A disintegrin and metalloprotease 33 polymorphisms and lung function decline in the general population. Am J Respir Crit Care Med 2005; 172: 329–333. doi:10.1164/rccm.200411-1486OC
    OpenUrlCrossRefPubMed
  23. ↵
    1. Ikram MA,
    2. Brusselle G,
    3. Ghanbari M, et al.
    Objectives, design and main findings until 2020 from the Rotterdam Study. Eur J Epidemiol 2020; 35: 483–517. doi:10.1007/s10654-020-00640-5
    OpenUrlCrossRefPubMed
  24. ↵
    1. Quanjer PH,
    2. Stanojevic S,
    3. Cole TJ, et al.
    Multi-ethnic reference values for spirometry for the 3–95-yr age range: the global lung function 2012 equations. Eur Respir J 2012; 40: 1324–1343. doi:10.1183/09031936.00080312
    OpenUrlAbstract/FREE Full Text
  25. ↵
    1. Scrucca L,
    2. Fop M,
    3. Murphy TB, et al.
    mclust 5: clustering, classification and density estimation using Gaussian finite mixture models. R J 2016; 8: 289–317. doi:10.32614/RJ-2016-021
    OpenUrlCrossRefPubMed
  26. ↵
    1. Moll M,
    2. Sakornsakolpat P,
    3. Shrine N, et al.
    Chronic obstructive pulmonary disease and related phenotypes: polygenic risk scores in population-based and case–control cohorts. Lancet Respir Med 2020; 8: 696–708. doi:10.1016/S2213-2600(20)30101-6
    OpenUrl
  27. ↵
    1. Lambert SA,
    2. Gil L,
    3. Jupp S, et al.
    The Polygenic Score Catalog as an open database for reproducibility and systematic evaluation. Nat Genet 2021; 53: 420–425. doi:10.1038/s41588-021-00783-5
    OpenUrl
  28. ↵
    1. Zhang Z,
    2. Reinikainen J,
    3. Adeleke KA, et al.
    Time-varying covariates and coefficients in Cox regression models. Ann Transl Med 2018; 6: 121. doi:10.21037/atm.2018.02.12
    OpenUrl
  29. ↵
    1. R Core Team
    . R: A Language and Environment for Statistical Computing. Vienna, R Foundation for Statistical Computing, 2022.
    1. Wickham H,
    2. Averick M,
    3. Bryan J, et al.
    Welcome to the Tidyverse. J Open Source Softw 2019; 4: 686. doi:10.21105/joss.01686
    OpenUrl
  31. ↵
    1. Therneau TM
    . A Package for Survival Analysis in R. 2023. https://cran.r-project.org/web/packages/survival/vignettes/survival.pdf
  32. ↵
    1. Burney PG,
    2. Hooper R
    . Forced vital capacity, airway obstruction and survival in a general population sample from the USA. Thorax 2011; 66: 49–54. doi:10.1136/thx.2010.147041
    OpenUrlAbstract/FREE Full Text
  33. ↵
    1. Marott JL,
    2. Ingebrigtsen TS,
    3. Çolak Y, et al.
    Trajectory of preserved ratio impaired spirometry: natural history and long-term prognosis. Am J Respir Crit Care Med 2021; 204: 910–920. doi:10.1164/rccm.202102-0517OC
    OpenUrlCrossRefPubMed
  34. ↵
    1. Wijnant SRA,
    2. De Roos E,
    3. Kavousi M, et al.
    Trajectory and mortality of preserved ratio impaired spirometry: the Rotterdam Study. Eur Respir J 2020; 55: 1901217. doi: 10.1183/13993003.01217-2019
    OpenUrlPubMed
  35. ↵
    1. Torén K,
    2. Olin AC,
    3. Lindberg A, et al.
    Vital capacity and COPD: the Swedish CArdioPulmonary bioImage Study (SCAPIS). Int J Chron Obstruct Pulmon Dis 2016; 11: 927–933. doi:10.2147/COPD.S104644
    OpenUrlPubMed
  36. ↵
    1. Godtfredsen NS,
    2. Lam TH,
    3. Hansel TT, et al.
    COPD-related morbidity and mortality after smoking cessation: status of the evidence. Eur Respir J 2008; 32: 844–853. doi:10.1183/09031936.00160007
    OpenUrlAbstract/FREE Full Text
  37. ↵
    1. Zhang J,
    2. Xu H,
    3. Qiao D, et al.
    A polygenic risk score and age of diagnosis of COPD. Eur Respir J 2022; 60: 2101954. doi:10.1183/13993003.01954-2021
    OpenUrlAbstract/FREE Full Text
  38. ↵
    1. Abozid H,
    2. Kirby M,
    3. Hartl S, et al.
    Is chronic cough an additional marker for lung function decline? Eur Respir J 2022; 60: 1816. DOI: 10.1183/13993003.congress-2022.1816
    OpenUrlCrossRef
  39. ↵
    1. Silvestre OM,
    2. Nadruz W Jr.,
    3. Querejeta Roca G, et al.
    Declining lung function and cardiovascular risk: the ARIC study. J Am Coll Cardiol 2018; 72: 1109–1122. doi:10.1016/j.jacc.2018.06.049
    OpenUrlFREE Full Text
  40. ↵
    1. Berry CE,
    2. Billheimer D,
    3. Jenkins IC, et al.
    A distinct low lung function trajectory from childhood to the fourth decade of life. Am J Respir Crit Care Med 2016; 194: 607–612. doi:10.1164/rccm.201604-0753OC
    OpenUrlPubMed
Back to top
Email
Print
Citation Tools
Clinical relevance of lung function trajectory clusters in middle-aged and older adults
Xander Bertels, James C. Ross, Rosa Faner, Michael H. Cho, M. Arfan Ikram, Guy G. Brusselle, Lies Lahousse
ERJ Open Research Jan 2024, 10 (1) 00793-2023; DOI: 10.1183/23120541.00793-2023
del.icio.us logo Technorati logo Twitter logo Connotea logo Facebook logo Mendeley logo
Full Text (PDF)

Jump To

Subjects