Rationale: Chronic obstructive pulmonary disease (COPD) can develop not only through a lung function trajectory dominated by an accelerated decline of FEV1 from normal maximally attained FEV1 in early adulthood (normal maximally attained FEV1 trajectory) but also through a trajectory with FEV1 below normal in early adulthood (low maximally attained FEV1 trajectory).
Objectives: To test whether the long-term risk of exacerbations and mortality differs between these two subtypes of COPD.
Methods: The cohort included 1,170 young adults enrolled in the Copenhagen City Heart Study during the 1970s and 1980s. In 2001–2003, which served as the baseline for the present analyses, 79 participants had developed COPD through normal maximally attained FEV1 trajectory, 65 had developed COPD through low maximally attained FEV1 trajectory, and 1,026 did not have COPD.
Measurements and Main Results: From 2001 until 2018, we observed 139 severe exacerbations of COPD and 215 deaths, of which 55 were due to nonmalignant respiratory disease. In Cox models, there was no difference with regard to risk of severe exacerbations between the two trajectories, but individuals with normal maximally attained FEV1 had an increased risk of nonmalignant respiratory disease mortality (using inverse probability of censoring weighting with adjusted hazard ratio [HR], 6.20; 95% confidence interval [CI], 2.09–18.37; P = 0.001) and all-cause mortality (adjusted HR, 1.93; 95% CI, 1.14–3.26; P = 0.01) compared with individuals with low maximally attained FEV1.
Conclusions: COPD developed through normal maximally attained FEV1 trajectory is associated with an increased risk of respiratory and all-cause mortality compared with COPD developed through low maximally attained FEV1 trajectory.
A previous study has suggested that approximately half of COPD cases evolve from a normal FEV1 level in early adulthood (normal maximally attained FEV1 trajectory), whereas individuals belonging to the other half present with reduced lung function already in their early adulthood (low maximally attained FEV1 trajectory). The present study is the first to investigate the risk of severe exacerbations and survival, comparing individuals belonging to the two types of FEV1 trajectories leading to COPD.
In this prospective cohort study, individuals from the general population were followed for 42 years. Similar numbers of participants developed COPD, either from the normal maximally attained FEV1 or the low maximally attained FEV1, and presented with a similar degree of airflow limitation and symptomatology in 2001–2003. During the next 17 years, COPD that developed through normal maximally attained FEV1 trajectory was associated with an increased risk of both respiratory disease mortality and all-cause mortality compared with COPD that developed through low maximally attained FEV1 trajectory. This study shows a different risk of mortality in individuals with COPD with different previous courses of lung function leading to their disease and contributes to a better understanding of the heterogeneity of COPD.
Chronic obstructive pulmonary disease (COPD) can develop through different trajectories of FEV1. The trajectory that attracted most interest in the past was described by Fletcher and colleagues in 1977 and implicates an accelerated FEV1 decline from a normal level obtained in early adulthood (normal maximally attained FEV1 trajectory) (1). Another trajectory hypothesized by Burrows and colleagues is characterized by FEV1 below normal already in early adulthood (2). A recent study estimated that each of these two trajectories is responsible for approximately half of the cases of COPD in the general population (3). In both trajectories, smoking plays a pivotal role, but it is unknown whether individuals with COPD following these two lung function trajectories differ with regard to long-term prognosis.
In the present study, we identified middle-aged adults with COPD through spirometry in 2001–2003 in an ongoing survey of the general population in Copenhagen, Denmark. All of them had previously attended examination rounds of this survey during the 1970s and 1980s, when aged 20–40 years, allowing us to determine the FEV1 trajectory that led to their COPD. We investigated the long-term prognosis, including risk of severe exacerbations of COPD and respiratory disease mortality and all-cause mortality, from the baseline in 2001–2003 to 2018. We hypothesized that COPD developed through the normal maximally attained FEV1 trajectory is associated with higher long-term morbidity and mortality due to an ongoing destructive process in the lungs compared with COPD developed through the low maximally attained FEV1 trajectory, in which substantial damage most likely took place already during childhood and adolescence or even before birth.
All individuals included in the present analyses participated in the Copenhagen City Heart Study, an ongoing survey of the general population in Denmark (4). The study was approved by an institutional review board and a Danish ethics committee (approval KF-V.100.2039/91) and was conducted according to the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants. In 1976–1978, a sample of 19,698 subjects aged 20–100 years from the inner city of Copenhagen was randomly selected and invited for the initial survey. A total of 14,223 subjects participated and were reinvited to complete four subsequent surveys.
In the present analyses, we included 1,170 individuals who attended the examination round, which took place from 2001 to 2003, the baseline for the present analyses, and who had spirometry data from their initial examinations performed in either 1976–1978 or 1981–1983, when aged 20–40 years. At each examination, the participants answered an extensive questionnaire concerning lifestyle, health topics, symptoms, and exposures and underwent a physical health examination (3).
COPD was defined as an FEV1/FVC ratio less than the lower limit of normal (LLN) and FEV1 less than 80% of the predicted value (5). LLN was defined as the fifth percentile for FEV1/FVC, calculated as the mean value minus 1.645 SD (5). On the basis of spirometry data, participants were assigned to one of three mutually exclusive clinical subgroups (Figure 1):
COPD developed through normal maximally attained FEV1 trajectory: Defined as FEV1/FVC less than the LLN and FEV1 below 80% of the predicted value at the baseline examination in 2001–2003 and FEV1 greater than or equal to 80% of the predicted value at the study enrollment in 1976–1978 or in 1981–1983.
COPD developed through low maximally attained FEV1 trajectory: Defined as FEV1/FVC less than the LLN and FEV1 below 80% of the predicted value at the baseline examination in 2001–2003 and FEV1 below 80% of the predicted value at the study enrollment in 1976–1978 or in 1981–1983.
No COPD: Defined as FEV1/FVC greater than or equal to the LLN and FEV1 greater than or equal to 80% of the predicted value at the baseline examination in 2001–2003 and FEV1 greater than or equal to 80% of the predicted value at the study enrollment in 1976–1978 or in 1981–1983.
After the baseline examination in 2001–2003, we recorded severe exacerbations of COPD and respiratory disease mortality and all-cause mortality until 2018. Severe exacerbations were defined as hospitalizations due to exacerbation of obstructive lung disease as the main or secondary discharge diagnosis on the basis of information obtained from the Danish National Patient Registry (6), recorded until December 7, 2018.
Respiratory disease mortality was defined as cause of death containing lung disease but not lung cancer (i.e., nonmalignant respiratory disease) on the basis of information obtained from the national Danish Register of Causes of Death (7). Information on all-cause mortality was available from the national Danish Civil Registration System, recorded until December 13, 2018.
The annualized decline in FEV1 was calculated as the slope between the study enrollment in 1976–1978 or in 1981–1983 and the baseline examination in 2001–2003. For the follow-up examination in 2011–2015, we estimated the decline in FEV1 over a period of approximately 10 years from the baseline examination in 2001–2003.
Information from the Danish National Health Service Prescription Database was used to monitor use of inhaled corticosteroids and inhaled long-acting bronchodilators (β2-agonists and antimuscarinic agents), which were grouped together as maintenance medications (8).
For demographics, ANOVA for comparison of means or the Kruskal-Wallis test for comparison of medians was used for continuous variables, and Fisher’s exact test was used for categorical variables. Death caused by nonrespiratory disease was considered a competing risk for respiratory disease mortality, whereas death of all causes was considered a competing risk for severe exacerbations of COPD. The inverse probability of censoring weighting (IPCW) estimator can correct for informative censoring by giving extra weight to subjects who are not censored (9). These weights are chosen in such a way that individuals who best match the censored subject will have higher weights. At all observed time points, the individuals receive weights that are inversely proportional to the estimated probability of having remained uncensored until that time. The estimated probabilities are based on a Cox proportional hazards regression analysis with adjustment for age, sex, and phenotype trajectory. This IPCW methodology approximated marginal event rates that would be observed if the competing events were not censoring the outcomes of interest. The inverse Kaplan-Meier curves with IPCW depicts these marginal event rates. For the outcome of all-cause mortality, the unweighted inverse Kaplan-Meier estimator and the log-rank test were performed. The association between the different trajectories and all-cause mortality was analyzed with unweighted Cox proportional hazards regression, whereas weights from IPCW were applied in the setting of competing risks, both adjusted for age and sex. Additional adjustment included FEV1% predicted value when comparing the two COPD trajectories. A sensitivity analysis was performed by excluding subjects with an FEV1 decline of <40 ml/yr from the normal maximally attained FEV1 trajectory and another by excluding subjects with FEV1/FVC less than 0.7 at study enrollment.
The proportionality assumption in the Cox regression models was tested with the Lin, Wei, and Ying score process test, and the misspecifications of the functional form were tested by cumulative residuals (10). No adjustment for multiple comparisons was made. All analyses were performed using R version 3.5.2 software (R Foundation for Statistical Computing). A two-sided P value less than 0.05 was considered significant.
The flowchart in Figure E1 in the online supplement shows the study population according to the inclusion criteria of spirometry and age 20–40 years at study enrollment (1976–1983) and baseline examination 25 years later (2001–2003). The clinical characteristics and smoking history of the 1,170 participants according to the three FEV1 trajectories are given in Table 1. Both when aged 20–40 years and 25 years later, the individuals with COPD in the two COPD trajectories did not differ with regard to age, smoking habit, or prevalence of self-reported asthma. FEV1 decline was approximately twice as high in individuals with COPD developed through normal maximally attained FEV1 trajectory as the decline observed in those with low maximally attained FEV1 (61 ml/yr vs. 29 ml/yr; P < 0.001). From 1976–1978 or 1981–1983 to 2001–2003, this resulted in an average absolute loss in FEV1 of 1.5 L in the first group versus 0.7 L in the latter group. In 2001–2003, the individuals in the two COPD trajectories presented with similar features characterizing their COPD, including FEV1, blood eosinophil count, dyspnea score, prevalence of chronic mucus hypersecretion, body mass index, history of COPD exacerbations in the previous year, and use of maintenance COPD medications. Individuals in both COPD trajectories showed a very high persistence of smoking that was similar, but it was approximately twice as high as in those without COPD (Table 1). There were 12 individuals in the low maximally attained FEV1 trajectory and 3 individuals in the normal maximally attained FEV1 trajectory with an FEV1/FVC ratio less than 0.7 at study enrollment in 1976–1978 or in 1981–1983.
|Characteristics||No COPD at Baseline Examination (n = 1,026)||COPD at Baseline Examination||P Value*|
|Low Maximally Attained FEV1 Trajectory (n = 65)||Normal Maximally Attained FEV1 Trajectory (n = 79)|
|At study enrollment in 1976–1978 or in 1981–1983|
|Sex, M||511 (50%)||29 (45%)||49 (62%)||0.04|
|Mean ± SD||33 ± 6||33 ± 5||34 ± 5||0.32|
|Mean ± SD, L||3.7 ± 0.8||2.7 ± 0.5||3.6 ± 0.8||<0.001|
|Percent predicted value||95 ± 10||69 ± 7||90 ± 8||<0.001|
|FEV1/FVC, %||86 ± 7||75 ± 9||84 ± 8||<0.001|
|Never smoker||327/1,023 (32%)||8/64 (13%)||5/79 (6%)|
|Former smoker||157/1,023 (15%)||2/64 (3%)||3/79 (4%)|
|Current smoker||539/1,023 (53%)||54/64 (84%)||71/79 (90%)|
|Smoking onset before age 14 yr||46/682 (7%)||11/57 (19%)||10/74 (14%)||0.47|
|Asthma||11/1,003 (1%)||2/64 (3%)||3/77 (4%)||>0.99|
|Height, cm||172 ± 9||171 ± 9||173 ± 10||0.38|
|Body mass index, kg/m2||23 ± 3||23 ± 3||24 ± 4||0.28|
|At baseline examination in 2001–2003|
|Mean ± SD||57 ± 7||58 ± 6||59 ± 6||0.24|
|Mean ± SD, L||3.1 ± 0.7||1.9 ± 0.6||2.1 ± 0.6||0.10|
|Percent predicted value||98 ± 13||63 ± 12||66 ± 13||0.13|
|FEV1/FVC, %||78 ± 5||63 ± 6||63 ± 7||0.59|
|Decline in FEV1|
|Mean ± SD, ml/yr||27 ± 19||29 ± 17||61 ± 22||<0.001|
|Median (IQR), ml/yr||26 (22)||27 (21)||61 (28)||<0.001|
|Percentage of baseline value per year||0.7 ± 0.4||1.1 ± 0.7||1.7 ± 0.5||<0.001|
|≥40 ml/yr||218 (21%)||15 (23%)||67 (85%)||<0.001|
|Current smoker||336/1,009 (33%)||39/63 (62%)||51/77 (66%)||0.60|
|Smoking history, median (IQR), pack-years||10 (30)||38 (35)||40 (35)||0.66|
|Dyspnea (mMRC scale score, ≥2)||71/1,024 (7%)||16/64 (25%)||21/79 (27%)||0.85|
|Previous exposure to dust and fumes||133 (13%)||15 (23%)||13 (16%)||0.40|
|Chronic mucus hypersecretion||118/1,024 (12%)||19/65 (29%)||21/79 (27%)||0.85|
|Asthma||46/1,025 (4%)||11/65 (17%)||16/79 (20%)||0.67|
|Body mass index, kg/m2||26 ± 4||26 ± 5||27 ± 5||0.48|
|Median (IQR), cells/μl||190 (150)||222 (144)||204 (190)||0.65|
|≥300 cells/μl||195/1,014 (19%)||15/63 (24%)||22/77 (29%)||0.57|
|Using maintenance COPD medication||39 (4%)||11 (17%)||14 (18%)||>0.99|
|Exacerbation within last year||13 (1%)||3 (5%)||5 (6%)||0.73|
In 2011–2015, we examined a small sample of survivors (Table 2). The prevalence of smoking was still approximately twice as high among individuals in both trajectories with COPD as in individuals without COPD. For these survivors, the FEV1 decline from baseline in 2001–2003 to 2011–2015 was similar in the two COPD trajectories.
|Characteristics at Follow-up Examination in 2011–2015||No COPD at Baseline Examination (n = 808)||COPD at Baseline Examination||P Value*|
|Low Maximally Attained FEV1 Trajectory (n = 34)||Normal Maximally Attained FEV1 Trajectory (n = 30)|
|Sex, M||411 (51%)||15 (44%)||19 (63%)||0.14|
|Age, yr||67 ± 7||67 ± 6||68 ± 6||0.61|
|Mean ± SD, L||2.8 ± 0.8||1.7 ± 0.7||2.0 ± 0.6||0.15|
|Percent predicted value||101 ± 18||63 ± 18||69 ± 19||0.24|
|FEV1/FVC, %||75 ± 7||57 ± 12||59 ± 12||0.48|
|Decline in FEV1|
|Mean ± SD, ml/yr||33 ± 35||32 ± 33||38 ± 56||0.58|
|Median (IQR), ml/yr||28 (29)||32 (40)||31 (50)||0.71|
|≥40 ml/yr||245/800 (31%)||14/34 (41%)||12/28 (43%)||>0.99|
|Current smoker||140/788 (18%)||11/34 (32%)||12/29 (41%)||0.60|
|Smoking history, median (IQR), pack-years||7 (23)||40 (42)||37 (16)||0.74|
|Dyspnea (mMRC scale score, ≥2)||71/807 (9%)||12/34 (35%)||12/30 (40%)||0.80|
|Asthma||37/792 (5%)||7/31 (23%)||5/29 (17%)||0.75|
|Body mass index, kg/m2||27 ± 5||26 ± 5||27 ± 4||0.56|
After the 2001–2003 baseline examination, we registered the use of maintenance medications for COPD in individuals not on maintenance medication at baseline and observed no difference in the use of these medications between the individuals in the two FEV1 trajectories leading to COPD (adjusted hazard ratio [HR], 1.15; 95% confidence interval [CI], 0.68–1.94; P = 0.61 for comparison between low maximally attained FEV1 (reference) and normal maximally attained FEV1 trajectory) (Figures 2A and 2C).
During the follow-up from 2001–2003 to 2018, we observed 139 severe exacerbations of COPD. Figure 2 shows the inverse Kaplan-Meier curve for the IPCW in the three trajectories in Figure 2B, and the results from adjusted Cox regression analysis with weights from IPCW in Figure 2C. In the latter analysis, we have assigned the low maximally attained FEV1 trajectory as the reference group to highlight the difference between this trajectory and the normal maximally attained FEV1 trajectory. As expected, the lowest risk of exacerbations was observed among individuals without COPD, although some of them did experience events labeled as exacerbations, which is compatible with a substantial smoking prevalence of 33% in 2001–2003 in this group. The risk of severe exacerbations was not significantly different in the two COPD trajectories. Similar results were observed in stratified analyses of current and former smokers (Figure E2).
In total we registered 215 deaths, including 55 deaths from nonmalignant respiratory disease and 76 deaths from cancer of which 30 were caused by lung cancer. Figure 3 shows the inverse Kaplan-Meier curve of death from nonmalignant respiratory disease for the IPCW (Figure 3A) and from all causes without IPCW (Figure 3B) and results from the multivariable adjusted Cox regression analysis both with IPCW and unweighted in Figure 3C, where the low maximally attained FEV1 trajectory is assigned as reference group. Risk of all-cause mortality was substantially higher in the two COPD trajectories compared with individuals without COPD. The mortality rates in the two COPD trajectories separated after approximately 10 years. Compared with individuals with COPD in the low maximally attained FEV1 trajectory, adjusted HR for all-cause mortality was 1.93 (95% CI, 1.14–3.26; P = 0.01) in individuals with COPD in the normal maximally attained FEV1 trajectory. The pattern regarding all-cause mortality was similar after stratification for smoking habits, but the differences between the two COPD trajectories did not reach statistical significance, owing to smaller groups (Figure E3).
Only 4 (7%) of the individuals in the low maximally attained FEV1 trajectory died of nonmalignant respiratory disease, whereas the corresponding number in the normal maximally attained FEV1 trajectory was 21 (34%). Compared with individuals with COPD in the low maximally attained FEV1 trajectory, adjusted HR for nonmalignant respiratory disease mortality was 6.20 (95% CI, 2.09–18.37; P = 0.001) in individuals with COPD in the normal maximally attained FEV1 trajectory (Figure 3C). After adjustment for age and sex, the risk of death from nonmalignant respiratory disease, although numerically higher, did not differ significantly between individuals in the low maximally attained FEV1 trajectory and those without COPD (Figure 3C). Similar results regarding respiratory disease mortality were observed in individuals with COPD who were smokers in 2001–2003: Those following the normal maximally attained FEV1 trajectory had an approximately five times higher risk than those with low maximally attained FEV1 trajectory (Figure E4). Interestingly, we did not observe any deaths caused by nonmalignant respiratory disease in participants in the low maximally attained FEV1 trajectory who stopped smoking before our 2001–2003 baseline examination, whereas those in the normal maximally attained FEV1 trajectory continued to die of respiratory diseases despite smoking cessation (Figure E4).
The normal maximally attained FEV1 trajectory included 12 subjects with an FEV1 decline of <40 ml/yr. In a sensitivity analysis, we excluded these subjects to see whether a criterion of FEV1 decline would impact the results. The adjusted HR for all-cause mortality was 1.87 (95% CI, 1.09–3.19; P = 0.02), and the adjusted HR for nonmalignant respiratory disease mortality was 6.33 (95% CI, 2.13–18.80; P < 0.001) in individuals with COPD in the normal maximally attained FEV1 trajectory compared with individuals with COPD in the low maximally attained FEV1 trajectory. As for the main analysis, there was no significant association between the two COPD trajectories and severe exacerbations (adjusted HR, 1.51; 95% CI, 0.90–2.54; P = 0.12).
In a sensitivity analysis, excluding individuals with FEV1/FVC ratio less than 0.7 at study enrollment in 1976–1978 or in 1981–1983, the adjusted HRs were 1.85 (95% CI, 1.06–3.24; P = 0.03) for all-cause mortality, 4.60 (95% CI, 1.59–13.28; P = 0.005) for nonmalignant respiratory disease mortality, and 1.25 (95% CI, 0.73–2.16; P = 0.42) for severe exacerbations when comparing individuals with COPD in the normal maximally attained FEV1 trajectory with those in the low maximally attained FEV1 trajectory.
The main finding of this study is that the long-term mortality, in particular mortality caused by respiratory disease, is higher in individuals who develop COPD through the normal maximally attained FEV1 trajectory than in those following the low maximally attained FEV1 trajectory. Although the risk of severe exacerbations did not differ significantly between these two trajectories, we cannot out rule a higher long-term risk in those in the normal maximally attained FEV1 trajectory, because the CIs were quite wide and the risk was numerically higher. This is the first study focusing on exacerbation risk and survival of individuals with COPD developed through different lung function trajectories. We acknowledge that grouping into two specific lung function trajectories is an oversimplification, because several other trajectories may exist (11–13).
The main advantage of the present study is its duration, including an initial 25 years of observation, allowing the allocation of the participants to a relevant FEV1 trajectory followed by more than 17 years of observation of morbidity and mortality. Our follow-up after the 2001–2003 examination is complete, and we are also confident that the case definition of severe COPD exacerbations is valid (14). Our findings seem robust because the observed associations were confirmed when we excluded subjects with an FEV1 decline of <40 ml/yr from the normal maximally attained FEV1 trajectory and when we excluded subjects with an FEV1/FVC ratio less than 0.7 at study enrollment.
The present study spanning over more than four decades underpins the devastating role of smoking for COPD development and progression. In total, less than 10% of the individuals with COPD (6% in the normal maximally attained FEV1 trajectory and 13% in the low maximally FEV1 attained trajectory) were never-smokers (Table 1). This underlines that although COPD, or rather chronic airflow limitation, can be seen in never-smokers (15), an overwhelming majority of individuals with COPD in a highly developed country, Denmark, had a heavy cumulative tobacco exposure of approximately 40 pack-years at approximately 60 years of age, regardless of the FEV1 trajectory. A previous Danish study of a different cohort found that COPD in never-smokers was associated with normal survival (16); therefore, we assume that the serious prognosis of the individuals with COPD in the present study was related to smoking.
In line with the current perception that asthmatic features may precede and contribute to the risk of developing COPD, we have deliberately not excluded individuals with asthma from our analyses (17). Previous studies have shown that childhood asthma is associated with impaired lung function growth, lower lung function in both early adulthood and middle age, and higher risk of COPD in middle age (13, 18, 19). In our study, the overall prevalence of self-reported asthma was 1.4% in the 1970s and 1980s and 3–4% in those who later developed COPD (Table 1). However, 20% of the individuals in the normal maximally attained FEV1 trajectory and 17% in the low maximally attained FEV1 trajectory reported asthma in 2001–2003 (Table 1). Individuals with concomitant asthma and COPD have previously been shown to have a high risk of exacerbations (20–23).
Although the concept that FEV1 decline is the hallmark of COPD has been widely accepted since the study by Fletcher and colleagues (1), relatively few studies have investigated the importance of fast FEV1 decline, as such, to morbidity and mortality. Using data from two different U.S. cohorts, Mannino and colleagues reported a modestly increased risk of death and hospital admissions in those with the fastest declines (24, 25). In these two studies, the participants were assigned to the most rapidly declining lung function group on the basis of a relatively short observation period of 3–4 years, and there was no information on the maximally attained lung function in early adulthood, which precludes a definition of and comparison between different long-term FEV1 trajectories. Low maximally attained lung function in early adulthood has recently been shown to be related to a higher risk of comorbidities and mortality later in life (26, 27), a finding in line with the so-called Barker hypothesis, emphasizing the importance of prenatal and early-life events for risk of diseases later in life (28). In the present study, we do not know the reasons for the FEV1 reduction already in young adulthood among the individuals in the low maximally attained FEV1 trajectory. These individuals still had slightly lower FEV1 in their 60s than those with normal maximally attained FEV1, of whom the majority had fast FEV1 decline, but, in the long run, they had a better survival and a noticeably lower risk of death caused by respiratory disease. We can only speculate whether these two COPD trajectories reflect different lung pathologies. In context of the recent findings in the two U.S. cohorts (the Normative Aging Study and the COPDGene [COPD Genetic Epidemiology] study) the findings from the ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints) cohort, and the earlier observations from the Tucson study, we hypothesize that the normal maximally attained FEV1 trajectory represents individuals with emphysema as a predominant pathological disease process, whereas the low maximally attained FEV1 trajectory mostly includes individuals with less emphysema (29–31).
Our study has some potential limitations. The assignment of the participants to the two trajectories was based on prebronchodilator spirometry at study enrollment in the 1970s or 1980s combined with a single measurement in 2001–2003 and can therefore be subject to bias. Another drawback is the rather small number of individuals in the two COPD trajectories. Yet, we found a difference with regard to mortality but not severe exacerbations, although the number of events was similar in the two COPD trajectories for these endpoints. Selection bias is always a concern. However, less than 10% of the eligible subjects aged 20–40 years at study enrollment died before the baseline examination 25 years later, and the proportion of subjects with reduced lung function at study enrollment only shifted from 29% to 26% due to dropout during these 25 years. These findings suggest that the study population is not a survivor population, and dropout between study enrollment and baseline examination does not appear to be related to reduced lung function. In our analyses, we have not included use of inhaled medications during the observation period after the 2001–2003 examination as a possible confounder. Although the evidence regarding the possible effect of treatment on survival of patients with COPD is inconclusive, both inhaled long-acting bronchodilators and inhaled corticosteroids can reduce the risk of exacerbations (32). Yet, because the use of these medications in the two lung function trajectories leading to COPD was similar both at baseline in 2001–2003 and during follow-up (Figures 2A and 2C), we do not believe that lack of adjustment for maintenance COPD treatment in our analyses had any significant effect on our findings. Finally, we acknowledge that the clinical characterization of the individuals with COPD in 2001–2003 is not exhaustive. Although we have data on lung function, dyspnea, chronic mucus hypersecretion, exacerbation history, and blood eosinophils, assessment of bronchial reversibility was not done. In addition, we have no imaging data or measurement of DlCO, which would have been desirable to assess the amount of emphysema and small airway disease.
On the basis of long-term observations of a cohort of individuals with COPD from the general population, we conclude that the course of lung function trajectory leading to COPD is important for the clinical prognosis. COPD developed through normal maximally attained FEV1 trajectory is associated with a poorer survival, in particular a higher risk of respiratory disease mortality, than COPD developed through low maximally attained FEV1 trajectory.
|1.||Fletcher C, Peto R. The natural history of chronic airflow obstruction. BMJ 1977;1:1645–1648.|
|2.||Burrows B, Knudson RJ, Lebowitz MD. The relationship of childhood respiratory illness to adult obstructive airway disease. Am Rev Respir Dis 1977;115:751–760.|
|3.||Lange P, Celli B, Agustí A, Boje Jensen G, Divo M, Faner R, et al. Lung-function trajectories leading to chronic obstructive pulmonary disease. N Engl J Med 2015;373:111–122.|
|4.||Copenhagen City Heart Study Group. The Copenhagen City Heart Study. Osterbroundersøgelsen: a book of tables with data from the first examination (1976-78) and a five year follow-up (1981-83). Scand J Soc Med Suppl 1989;41:1–160.|
|5.||Løkke A, Marott JL, Mortensen J, Nordestgaard BG, Dahl M, Lange P. New Danish reference values for spirometry. Clin Respir J 2013;7:153–167.|
|6.||Schmidt M, Schmidt SA, Sandegaard JL, Ehrenstein V, Pedersen L, Sørensen HT. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol 2015;7:449–490.|
|7.||Helweg-Larsen K. The Danish register of causes of death. Scand J Public Health 2011;39(Suppl):26–29.|
|8.||Johannesdottir SA, Horváth-Puhó E, Ehrenstein V, Schmidt M, Pedersen L, Sørensen HT. Existing data sources for clinical epidemiology: the Danish National Database of reimbursed prescriptions. Clin Epidemiol 2012;4:303–313.|
|9.||Willems S, Schat A, van Noorden MS, Fiocco M. Correcting for dependent censoring in routine outcome monitoring data by applying the inverse probability censoring weighted estimator. Stat Methods Med Res 2018;27:323–335.|
|10.||Lin DY, Wein LJ, Ying Z. Checking the Cox model with cumulative sums of martingale-based residuals. Biometrika 1993;80:557–572.|
|11.||Krishnan JK, Martinez FJ. Lung function trajectories and chronic obstructive pulmonary disease: current understanding and knowledge gaps. Curr Opin Pulm Med 2018;24:124–129.|
|12.||Agusti A, Faner R. Lung function trajectories in health and disease. Lancet Respir Med 2019;7:358–364.|
|13.||Bui DS, Lodge CJ, Burgess JA, Lowe AJ, Perret J, Bui MQ, 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.|
|14.||Thomsen RW, Lange P, Hellquist B, Frausing E, Bartels PD, Krog BR, et al. Validity and underrecording of diagnosis of COPD in the Danish National Patient Registry. Respir Med 2011;105:1063–1068.|
|15.||Salvi SS, Barnes PJ. Chronic obstructive pulmonary disease in non-smokers. Lancet 2009;374:733–743.|
|16.||Thomsen M, Nordestgaard BG, Vestbo J, Lange P. Characteristics and outcomes of chronic obstructive pulmonary disease in never smokers in Denmark: a prospective population study. Lancet Respir Med 2013;1:543–550.|
|17.||Martinez FJ, Han MK, Allinson JP, Barr RG, Boucher RC, Calverley PMA, et al. At the root: defining and halting progression of early chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2018;197:1540–1551.|
|18.||Svanes C, Sunyer J, Plana E, Dharmage S, Heinrich J, Jarvis D, et al. Early life origins of chronic obstructive pulmonary disease. Thorax 2010;65:14–20.|
|19.||McGeachie MJ, Yates KP, Zhou X, Guo F, Sternberg AL, Van Natta ML, et al. Patterns of growth and decline in lung function in persistent childhood asthma. N Engl J Med 2016;374:1842–1852.|
|20.||Postma DS, Rabe KF. The asthma–COPD overlap syndrome. N Engl J Med 2015;373:1241–1249.|
|21.||Lange P, Çolak Y, Ingebrigtsen TS, Vestbo J, Marott JL. Long-term prognosis of asthma, chronic obstructive pulmonary disease, and asthma–chronic obstructive pulmonary disease overlap in the Copenhagen City Heart study: a prospective population-based analysis. Lancet Respir Med 2016;4:454–462.|
|22.||Hardin M, Silverman EK, Barr RG, Hansel NN, Schroeder JD, Make BJ, et al.; COPDGene Investigators. The clinical features of the overlap between COPD and asthma. Respir Res 2011;12:127.|
|23.||de Marco R, Marcon A, Rossi A, Antó JM, Cerveri I, Gislason T, et al. Asthma, COPD and overlap syndrome: a longitudinal study in young European adults. Eur Respir J 2015;46:671–679.|
|24.||Mannino DM, Davis KJ. Lung function decline and outcomes in an elderly population. Thorax 2006;61:472–477.|
|25.||Mannino DM, Reichert MM, Davis KJ. Lung function decline and outcomes in an adult population. Am J Respir Crit Care Med 2006;173:985–990.|
|26.||Agustí A, Noell G, Brugada J, Faner R. Lung function in early adulthood and health in later life: a transgenerational cohort analysis. Lancet Respir Med 2017;5:935–945.|
|27.||Vasquez MM, Zhou M, Hu C, Martinez FD, Guerra S. Low lung function in young adult life is associated with early mortality. Am J Respir Crit Care Med 2017;195:1399–1401.|
|28.||Barker DJ. The fetal and infant origins of adult disease. BMJ 1990;301:1111.|
|29.||Ross JC, Castaldi PJ, Cho MH, Hersh CP, Rahaghi FN, Sánchez-Ferrero GV, 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.|
|30.||Vestbo J, Edwards LD, Scanlon PD, Yates JC, Agusti A, Bakke P, et al.; ECLIPSE Investigators. Changes in forced expiratory volume in 1 second over time in COPD. N Engl J Med 2011;365:1184–1192.|
|31.||Burrows B, Bloom JW, Traver GA, Cline MG. The course and prognosis of different forms of chronic airways obstruction in a sample from the general population. N Engl J Med 1987;317:1309–1314.|
|32.||Global Initiative for Chronic Obstructive Lung Disease. Global strategy for the diagnosis, management and prevention of chronic obstructive pulmonary disease: 2020 report [accessed 2020 Feb 4]. Available from: https://goldcopd.org/wp-content/uploads/2019/11/GOLD-2020-REPORT-ver1.0wms.pdf.|
Supported by the National Institute for Health Research Manchester Biomedical Research Centre (J.V.). The Copenhagen City Heart Study was supported by the Capital Region of Copenhagen, the Danish Heart Foundation, the Danish Lung Association, the Velux Foundation, and the Lundbeck Foundation. The present analyses were supported by a grant from GlaxoSmithKline (grant eTrack no. 8898) and by the Danish Lung Association.
Author Contributions: Study concept and design: J.L.M. and P.L. Acquisition of data: J.L.M. and P.L. Analysis and interpretation of data: J.L.M., T.S.I., Y.Ç., J.V., and P.L. First draft of the manuscript: J.L.M. and P.L. Critical revision of the manuscript for important intellectual content: J.L.M., T.S.I., Y.Ç., J.V., and P.L. Statistical analysis: J.L.M. Obtained funding: J.L.M., J.V., and P.L. Administrative, technical, and material support: J.L.M. and P.L. Study supervision: P.L. Full access to all data in the study and final responsibility for the decision to submit for publication: J.L.M. and P.L.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Originally Published in Press as DOI: 10.1164/rccm.201911-2115OC on April 14, 2020