Heterogeneity of reduced FEV₁ in early adulthood: A looking forward, looking backwards analysis

Abstract

Background

Some individuals never achieve normal peak FEV₁ in early adulthood. It is unknown if this is due to airflow limitation and/or lung restriction.

Methods

To investigate this, we: (1) looked forward in 19,791 participants in the Dutch Lifelines general population cohort aged 25–35 years with 5‐year follow‐up; and (2) looked backwards in 2032 participants in the Swedish BAMSE birth cohort with spirometry at 24 years of age but also at 16 and/or 8 years.

Results

(1) In Lifelines 8.5% of participants had reduced FEV₁ at 25–35 years, 68% due to Preserved Ratio Impaired Spirometry (‘PRISm’) and 32% to airflow limitation (‘low‐limited’); besides, 3.8% participants with normal FEV₁ showed airflow‐limitation (‘normal‐limited’). Low‐limited and normal‐limited, but not PRISm, reported higher smoking exposures and asthma diagnosis than normal (p < 0.05). At 5‐year follow‐up, 91.2% of participants remained in the same group, and FEV₁ decline was similar in normal and normal‐limited participants, but statistically smaller (p < 0.05) in PRISm and low‐limited; (2) these observations were largely reproduced in BAMSE at 24 years of age; and, (3) in BAMSE, low‐limited or PRISm individuals were already identifiable at 8–16 years of age.

Conclusion

Low peak FEV₁ in early adulthood is most often due to PRISm and results in a significant burden of respiratory symptoms. Only low‐limited and normal‐limited, but not PRISm, associate with a doctor diagnosis of asthma, and FEV₁ decline was statistically different in PRISm indicating a need for differentiated clinical approaches. These spirometric abnormalities can be already identified in childhood and adolescence.

Short abstract

Low peak FEV₁ in early adulthood is most often due to PRISm, and to a less extend to airflow limitation. These abnormal spirometric patterns are associated with significant burden of respiratory symptoms, different rates of lung function decline with age and, importantly, can already be identified in childhood and adolescence.

See related editorial

INTRODUCTION

A sizable proportion (4%–12%) of young adults (25–35 years of age) in the general population have an abnormally low FEV₁ value (i.e., the volume of gas expired in the 1st second of a forced spirometric manoeuvre).^{1, 2} This is clinically relevant because, compared to those of similar age but normal peak FEV₁, individuals with low FEV₁ value: (1) exhibit a higher prevalence and a decade earlier incidence of respiratory, cardiovascular and metabolic disorders (1); and, (2) have an increased risk dying prematurely.^{1, 3, 4, 5}

A low FEV₁ value can be due to two different pathophysiologic mechanisms: (1) airflow limitation, as indicated by a reduced FEV₁/FVC (Forced Vital Capacity) ratio; and/or (2) reduced lung volume (low FVC and/or FEV₁ with normal FEV1/FVC ratio, also referred to as PRISm; Preserved Ratio Impaired Spirometry). These two spirometry patterns have been reported by some previous life‐long longitudinal studies.^{6, 7} While previous studies have identified low FEV₁ as a risk factor for early morbidity, and have described the childhood factors for development of irreversible airflow limitation in young adults (childhood respiratory tract infections and asthma, and exposure to air pollution),^{8, 9} the specific roles of PRISm and airflow limitation in this context have not been thoroughly investigated, leaving a critical gap in our understanding. We hypothesized that both mechanisms contribute in different proportions to low peak FEV₁ in young individuals, and that these differential spirometry patterns are associated with different clinical characteristics and biomarkers. Further, given that it is now well‐established that abnormal lung development early in life can track to adulthood,^{10, 11, 12} we also hypothesized that these two spirometric patterns of reduced peak FEV₁ in young adulthood can already be identified in childhood and adolescence. We explored these hypotheses using a combined looking forward—looking backwards strategy. The former was investigated in the Dutch Lifelines cohort ¹³ (that includes young individuals from the general population followed prospectively for 5 years). The latter was investigated in the Swedish birth‐cohort BAMSE^{14, 15} with spirometric measurements at 24 years of age but also in childhood and infancy.

METHODS

Study design, population and ethics

Lifelines

The design of the Lifelines study has been published in detail elsewhere, ¹³ and is detailed in Appendix S1 in the Supporting Information. In short, Lifelines is a large population‐based cohort study that investigates the role of complex interactions between environmental, phenotypic and genomic factors in the development of chronic diseases and healthy ageing. ⁴ Between 2006 and 2013, baseline data was collected in 167,729 inhabitants of the Northern part of The Netherlands (Friesland, Groningen and Drenthe), aged 6 months to 93 years, and follow‐up visits were scheduled every 5 years. ¹⁰ For the current analysis we selected participants who, at recruitment, were between 25 and 35 years of age (n = 19,791).

BAMSE

The design of the Swedish population‐based birth cohort BAMSE has been published elsewhere.^{15, 16} In brief, between 1994 and 1996 BAMSE recruited 4089 infants from inner‐city, urban and suburban districts of Stockholm (Sweden) and followed participants until around 22–24 years.^{15, 16} BAMSE was approved by the Regional Ethical Review Board in Stockholm (Ref 2016/1380‐31/2). Informed study consent was obtained from all parents (at inclusion, 4 and 8 years) and participants (at age 16 and 24 years), under the Helsinki Declaration. For the current analysis, we used data from 24‐, 16‐ and 8‐years of age.

Spirometry and study groups

In both cohorts forced spirometry was measured according to international guidelines. ¹⁷ For analysis, we used values determined before bronchodilation. Reference values were those of the Global Lung function Initiative (GLI).^{17, 18} Participants were initially stratified according to their pre‐bronchodilator FEV₁ value (≥ or < than their lower limit of normal—LLN) according to the GLI reference equations. ¹⁷ Afterwards, individuals were further subdivided by the presence or absence of airflow limitation (FEV₁/FVC ratio < or ≥0.7, respectively). ¹² This classification led to four study groups: Normal (FEV₁ ≥ LLN and FEV₁/FVC ≥0.7), Normal Limited (FEV₁ ≥ LLN and FEV₁/FVC <0.7), PRISm (FEV₁ < LLN and FEV₁/FVC ≥0.7), Low Limited (FEV₁ < LLN and FEV₁/FVC <0.7). In Lifelines, sensitivity analyses were performed with FEV₁/FVC < LLN and after excluding patients with a diagnosis of asthma. FEV₁/FVC was analysed with the fixed ratio (<0.7) to align with the GOLD COPD criteria.

Statistical analysis

Descriptive statistics include number and proportion (for categorical variables) or mean ± standard deviation (SD) (for continuous variables except biochemical measurements which are shown as median ± Interquartile range [IQR]). Groups were compared using χ ² test (categorical variables) or with a one‐way Anova and a Tukey post‐hoc test (continuous variables). To identify factors independently associated with the study groups, several baseline variables selected by expert clinicians, on the basis of being described in the literature as factors associated to low FEV₁, and/or their nominal associations to FEV₁ in lifelines (Table S2 in the Supporting Information). These variables (Table S1 in the Supporting Information) were included in a multivariate logistic regression of each group (normal‐limited, PRISm and low‐limited) versus normal and a multinomial analysis was done for validation. The annual rate of lung function decline was calculated from the absolute values of FEV₁ and FVC as: (value at follow‐up − baseline value)/years of follow‐up. A p value <0.05 was considered statistically significant. All analyses were performed using the statistical software R, version 4.2.0 (www.r–project.org).

RESULTS

A flowchart of the study cohorts is included as Figure S1 in the Supporting Information.

Lifelines

Baseline data

Lifelines included 19,791 participants aged 25–35 years with valid spirometry. FEV₁ was ≥LLN in 18,115 (91.5%; ‘Normal’). Of note, however, 696 of them (3.8%) had airflow limitation (FEV₁/FVC <0.7), from now on labelled ‘Normal‐limited’. On the other hand, FEV₁ was <LLN in 1676 participants (8.5%); 68% of whom had PRISm (FEV₁/FVC ≥0.7 and FEV₁ < LLN) (4), and 32% of whom had airflow limitation (FEV₁/FVC <0.7 and FEV₁ < LLN) from now on labelled ‘Low‐limited’. Table 1 contrast the main clinical characteristics of these four groups (Table S2, in the Supporting Information, presents the full dataset). By design, spirometry values were different across the four study groups, but FVC (% ref.) was higher in normal and normal‐limited than in PRISm and low‐limited (99.4 ± 9.6, 108.4 ± 7.8, 81.3 ± 6.7 and 92.1 ± 8.62; p < 0.05) (Figure S2 in the Supporting Information). Never smokers were more prevalent in normal and PRISm (53.5%, 42.6%, 52.4% and 44.1%, p < 0.05), while current smokers were more prevalent in normal‐limited and low‐limited. Cumulative smoking (packs‐year) (2.9 ± 4.6, 4.3 ± 5.5, 3.6 ± 5.6, 5.0 ± 6.8, p < 0.05) and accumulated life course exposures were higher in low‐limited and normal‐limited (Table S2 in the Supporting Information). Compared to normal, a self‐reported previous doctor diagnosis of asthma starting below the average age of the population, was more prevalent in all three other groups (4.8%, 11.0%, 8.3%, 15.2%, p < 0.05) (Table S2 and Figure S3 in the Supporting Information). Nominal and cumulative respiratory symptoms were different across groups, being highest in low‐limited (Figure S2 in the Supporting Information), the most frequently reported symptom was wheezing statistically higher all groups versus normals, (18.6%, 31.7%, 28.9%, 49.4%, p < 0.05). Eosinophil levels were higher in all groups versus normal (2.4 ± 2,2.7 ± 2.3, 2.5 ± 1.9, 2.9 ± 2.2, p < 0.05) but lower in those with PRISm versus low‐limited and normal‐limited (Table 1 and Table S2 in the Supporting Information). Finally, in relation to participants with normal lung function, PRISm and low‐limited showed higher glycated haemoglobin (HbA1c) (Table 1). In the first sensitivity analysis using FEV₁/FVC < LLN instead of a fixed value of 0.7 for the definition of airflow limitation, results were by and large similar (Table S3 in the Supporting Information), albeit the prevalence of normal‐limited was higher (6.5% vs. 3.5%). In the second sensitivity analysis, when individuals ever diagnosed of asthma were excluded (Table S4 in the Supporting Information), the prevalence of the four groups was maintained: normal 89%, normal‐limited 3.2%, PRISm 5.5% and low‐limited 2.1%; and by and large, their main characteristics were also unchanged.

TABLE 1.

Key baseline characteristics of the four study groups in lifelines at 25–35 years of age.

	FEV1 ≥ LLN		FEV1 < LLN		Anova or chi‐squared p‐value
	FEV1/FVC ≥0.7	FEV1/FVC <0.7	FEV1/FVC ≥0.7	FEV1/FVC <0.7
	Normal	Normal limited	PRISm	Low limited
	Mean ± SD or n (%)	Mean ± SD or n (%)	Mean ± SD or n (%)	Mean ± SD or n (%)
Prevalence n, % from total young	17,419 (88%)	696 (3.52%)	1140 (5.76%)	536 (2.71%)
Prevalence, % from FEV1 group	96.16%	3.84%	68.02%	31.98%
Demographics
Age	30.28 ± 3.12	31.14 ± 2.96***,‡‡‡	29.89 ± 3.17***	30.37 ± 3.06‡‡,§§§	2.9E−15
Sex Male	7143 (41.01%)	417 (59.91%)***,‡‡‡	440 (38.6%)	285 (53.17%)‡‡‡,***,§	2.5E−28
Body Mass Index (kg/M2)	25.07 ± 4.15	25.09 ± 3.63‡‡	25.69 ± 5.26***	25.56 ± 4.65,*	2.5E−06
Lung function
FEV1% ref.	97.01 ± 9.24	88.36 ± 6.07***,‡‡‡	75.6 ± 4.41***	71.92 ± 7.26‡‡‡,***,§§§	0.0E+00
FVC % ref.	99.36 ± 9.59	108.37 ± 7.77***,‡‡‡	81.27 ± 6.71***	92.1 ± 8.62‡‡‡,***,§§§	0.0E+00
FEV1/FVC	0.81 ± 0.05	0.67 ± 0.02***,‡‡‡	0.78 ± 0.05***	0.65 ± 0.05‡‡‡,***,§§§	0.0E+00
Exposures
Smoking					2.1E−18
Never smoker	8774 (53.48%)	276 (42.66%)	559 (52.39%)	220 (44.09%)
Ex‐smoker	3416 (20.82%)	141 (21.79%)	197 (18.46%)	76 (15.23%)
Current smoker	4217 (25.7%)	230 (35.55%)***,‡‡	311 (29.15%)**	203 (40.68%)‡‡‡,***,§
Pack‐years	2.89 ± 4.64	4.32 ± 5.47***,‡‡	3.64 ± 5.56***	5.03 ± 6.79‡‡‡,***	1.9E−34
Hours per day are you exposed to the tobacco smoke of others	0.63 ± 1.65	0.95 ± 2.14***	1.02 ± 2.42***	1.11 ± 2.36,***	1.4E−16
Early life
Prematurity (<37 weeks)	725 (5.23%)	38 (7.22%)	58 (6.47%)	27 (6.73%)	5.6E−02
Education level					2.0E−16
Low education (0–3)	873 (5.02%)	62 (8.93%)***	96 (8.47%)***	46 (8.61%),***
Medium education (4–6)	8616 (49.56%)	365 (52.59%)	608 (53.62%)	300 (56.18%)
High education (7, 8)	7721 (44.41%)	264 (38.04%)	414 (36.51%)	182 (34.08%)
Respiratory symptoms
Ever wheezing	3232 (18.63%)	220 (31.75%)***	327 (28.86%)***	262 (49.43%)‡‡‡,***,§§§	6.4E−89
At times breathing problems	4024 (23.37%)	226 (32.94%)***	370 (33.01%)***	226 (43.46%)‡‡‡,***,§§§	5.3E−39
Cough in winter when getting up	1147 (6.62%)	72 (10.42%)***	110 (9.77%)***	81 (15.34%)‡‡,***,§	6.5E−18
At times wake up at night with shortness of breath	743 (4.28%)	53 (7.65%)***	73 (6.45%)**	73 (13.77%)‡‡‡,***,§§	1.7E−26
Cardiovascular measures
Systolic Blood pressure in mm Hg	121.04 ± 12.53	123.35 ± 12.51***,‡‡	121.62 ± 13.3	122.31 ± 12.48	2.7E−06
T‐axis	46.13 ± 22.56	49.46 ± 24.23***,‡‡‡	43.57 ± 20.28**	47.57 ± 21.81‡‡	4.6E−07

Biochemistry#	Median ± IQR	Median ± IQR	Median ± IQR	Median ± IQR
Leukocytes (10E9/L)	5.8 ± 1.9	5.8 ± 2‡‡‡	6.1 ± 2.1***	6.1 ± 2.1**,§	2.0E−11
Eosinophil Granulocytes (%)	2.4 ± 2	2.7 ± 2.3***,‡‡	2.5 ± 1.9	2.9 ± 2.2‡‡‡,***	1.4E−10
HbA1c (%)	5.4 ± 0.3	5.4 ± 0.4	5.4 ± 0.4***	5.4 ± 0.3‡‡,***,§§§	7.9E−15
Creatinine (μmol/L)	72 ± 16	75 ± 16***,‡‡‡	71 ± 16***	74 ± 17‡‡‡,*,§	1.7E−14

Note: The exact p values of the comparison between groups are provided in detail in Table S2 in the Supporting Information. In the current table, the symbols refer to the p values of the following Tukey post‐hoc tests: (1) *** p < 0.001 versus normals, ** p < 0.01 versus normals, * p < 0.05 versus normals; (2) ‡‡‡ p < 0.001 versus PRISm, ‡‡ p < 0.01 versus PRISm, ‡ p < 0.01 versus PRISm, (3) §§§ p < 0.001 normal limited versus low limited, §§ p < 0.01 normal limited versus low limited, § p < 0.05 normal limited versus low limited. # Outcomes shown as median ± Interquartile range (IQR), compared with Mann‐Whitney‐Wilcoxon Test. Bold‐italic outcomes represent Anova or Chi‐squared p‐values < 0.05.

Next, using a multivariate logistic regression, we identified the factors associated with the different study groups (vs. normal) (Figure 1 and Table S5A in the Supporting Information). We found that: (1) for normal‐limited, age, higher packs‐year, asthma diagnosis and ever wheezing were significantly associated risk factors, while being a female and having lower BMI were significantly associated with lower risk; (2) for PRISm, ever wheezing, breathing problems, glycated haemoglobin and haematocrit were associated risk factors, while younger age, higher education and creatinine level were associated with lower risk and (3) for low‐limited, higher packs‐year, asthma diagnosis, ever wheezing, breathing problems, glycated haemoglobin and haematocrit were associated risk factors. The multinomial analysis across groups presented similar results (Table S5B in the Supporting Information).

FIGURE 1.

Forest plots of the multivariate logistic analysis for normal‐limited versus normal (grey), PRISm versus normal (orange) and low‐limited versus normal (red). The odds ratio of each factor is indicated, the confidence interval and the p values are provided in Table S5 in the Supporting Information.

Looking forward: Follow‐up data

Spirometry could be determined again after a 5‐years follow‐up in 4649 of the 19,791 individuals studied at recruitment (23.5%). In this subpopulation, the prevalence of the four groups was very similar to that observed in the entire population: normal 89.2%, normal‐limited 3.3%, low‐limited 2.7% and PRISm 4.8%. At the time of the second evaluation, most participants (91.2%) remained in the same study group as at recruitment (Figure 2), the most frequent change was from normal to normal‐limited (2.5%), followed by normal to PRISm (2%). Table S6 in the Supporting Information contrasts the main characteristics of these 4649 participants at 5 years of follow‐up, stratified by baseline category. In agreement with baseline data, we observed that the groups with airflow‐limitation had higher FVC (both in absolute terms and as % reference).

FIGURE 2.

Changes of spirometry pattern of the individuals that have been followed up 5 years in Lifelines, shown as either alluvial plot (A) or the specific % in the table (B).

Figure 3 shows the changes of FEV₁ during follow‐up (expressed either in absolute values [litres] in the four groups defined at baseline in Lifelines [right plots]). The decline in each study group is depicted in Table 2, as expected, FEV₁ decline was within the normal range in normal individuals (about 25–30 mL/year). ¹⁷ It was also normal in normal‐limited participants, but much reduced in low‐limited and PRISm (Table 2). Finally, FVC declined in the groups with airflow‐limitation, but not in PRISm (Table 2).

FIGURE 3.

Mean ± SEM levels of FEV₁ litres of individuals stratified by the four study groups. Looking forward: The stratification was done at baseline (25–35 years) in Lifelines and is presented the evolution of the FEV₁ of only those individuals with the 5‐years follow‐up visit (30–35 years). Looking backwards: The stratification was done at 24 years in BAMSE, and the evolution of the FEV₁ was considered, in those individuals with a previous spirometry in childhood, at 16 and/or 8 years. Normal‐limited in grey, normal in green, low‐limited in red and PRISm in orange.

TABLE 2.

Change (mean ± SD) in FEV₁ (litres) and FVC (litres) during the 5‐year‐follow‐up in lifelines.

Lung function change	Normal	Normal limited	PRISm	Low‐limited	Anova p‐value
Change FEV1 (mL)	−24.4 ± 45.6	−20.6 ± 57.9	−5.9 ± 46.8***	−8.0 ± 60.5 ‡‡	3.0E−10
Change FVC (mL)	−8.5 ± 53.3	−27.7 ± 67 ¥¥¥	7.1 ± 57.4***	−16.2 ± 58.5	5.9E−09

^*** p < 0.001 versus normals.

^‡‡ p < 0.01 versus normals.

^¥¥¥ p < 0.001 versus normals.

BAMSE

In BAMSE participants (n = 2034) we observed that (Table S7 in the Supporting Information): (1) at 24 years of age, FEV₁ was ≥LLN in 1935 (95%; ‘Normal’) and of them, 28 (1.4%) were normal‐limited; (2) the prevalence of low FEV₁ was 5%, and in keeping with our observations in Lifelines, most of those with FEV₁ < LLN had PRISm (79.4%) and only 20.6% of them were low‐limited; (3) low‐limited and normal‐limited participants, but not PRISm, presented more wheezing and cough than normal; (4) low‐limited and normal‐limited but not PRISm had been more often diagnosed of chronic bronchitis/irreversible airflow limitation, and only low‐limited had more diagnosis of asthma and higher eosinophil levels than normals and, (5) interestingly, as illustrated in Figure 3, these four groups (stratified at the age of 24 years) could be already identified during the period of lung growth (i.e., at 8 and 16 years of age). The prevalence of the different spirometry patterns is depicted in Table 3, and low‐limited participants showed FEV₁ decline starting from 16 years of age (Figure 3).

TABLE 3.

Prevalence of the four study groups in BAMSE at different ages.

	24 years	16 years	8 years
	n and (%)	n and (%)	n and (%)
Normal	1907 (93.8%)	1271 (93.7%)	1036 (93.8%)
Normal‐limited	28 (1.4%)	19 (1.4%)	15 (1.4%)
PRISm	77 (3.8%)	53 (3.9%)	41 (3.7%)
Low‐limited	20 (1%)	13 (1%)	12 (1.1%)

DISCUSSION

The main findings of this analysis in two large independent European cohorts are that: (1) the prevalence of reduced FEV₁ at 24–35 years of age oscillates between 4.8% (BAMSE) to 8.5% (Lifelines), and that it is most often due to PRISm (68% [Lifelines] and 79.4% [BAMSE]); (2) both normal‐limited and low‐limited, but not PRISm, were associated to starting smoking at a younger age and with a doctor diagnosis of asthma; but PRISm and low‐limited were both associated to higher HbA1c; (3) looking forward, the different spirometric groups identified at 24 years in the Lifelines are associated with different rates of FEV₁ decline during follow‐up, being it remarkably small in individuals with PRISm; and, (4) looking backwards in the BAMSE cohort, participants with reduced FEV₁ (with or without airflow limitation (i.e., PRISm)) at 24 years of age were already identifiable at 8 and 16 years of age.

The observed prevalence of FEV₁ < LLN at 24–25 years of age is in line with previous reports in young individuals in other the general population cohorts (4%–12%).^{12, 19} By definition in a general population, as those used to calculate the GLI LLN, we expect 5% of the population to be below, and is also expected this 5% is constant in the general population. In our study we find, 4.8% (BAMSE) and 8.5% (Lifelines) of the population below the LLN of FEV₁, our findings suggest that specific cohort effects might influence the prevalence of LLN. Likewise, our observations in Lifelines (Table S8 in the Supporting Information) are also in keeping with previous studies showing that young adults with low FEV₁ suffer a higher burden of respiratory and non‐respiratory conditions, ² hence further supporting the notion that spirometry is a clinically relevant global health marker.^{3, 19} On the other hand, among novel observations, to our knowledge our study is the first to explore the prevalence, characteristics and longitudinal lung function in young adults with reduced peak FEV₁, with and without airflow limitation. It is also the first to contrast data collected in young adults in Lifelines with a birth cohort (BAMSE) with measurements obtained during childhood (8 years of age), adolescence (16 years of age) and early adulthood (24 years of age). These novel observations are discussed below.

Reduced FEV₁ can be due to either airflow limitation and/or lung volume restriction.^{12, 20} We found that PRISm was the more frequent cause of low FEV₁ in young adults (68% in Lifelines and 79.4% in BAMSE). The mechanisms leading to PRISm are not well understood and can include impaired lung growth and obesity, among others.^{21, 22} Of note, in our study we observed that wheezing, but not a diagnosis of asthma, low education, higher HbA1c % and haematocrit levels were associated to PRISm. Likewise, the low creatinine levels observed in PRISm are in keeping with previous findings by Guerra et al in individuals with ‘asthma’, ²³ but the distinct clinical characteristics observed in the PRISm group in the current study, including lower levels of eosinophils and higher glycated haemoglobin, suggest a metabolic rather than inflammatory pathway may be predominant in this phenotype. Additionally, the majority of PRISm were not smokers suggesting either that the underlying cause is associated to structural abnormalities (e.g., dysanapis), or that an inflammation (e.g., such as that associated to asthma) may lead to airway remodelling, or that exposure to various air pollutants contributes to small airway dysfunction. In any case, the pathophysiologic mechanisms underlying PRISm still require research.

On the other hand, airflow limitation was present in 32% of young individuals with low FEV₁ in Lifelines and 20% in BAMSE. In all cohorts, results showed that low‐limited was associated with higher smoking exposure starting at a younger age, HbA1c % and haematocrit, wheezing and a doctor diagnosis of asthma (Figure 1). These observations expand previous findings of elevated insulin in asthmatics, to individuals with low peak lung function with airflow limitation. ²⁴ Importantly, low‐limited individuals had a higher prevalence of symptoms, previous respiratory diagnoses and use of respiratory medications, indicating the burden of disease was already substantial in these young individuals and highlighting an opportunity for prevention and early diagnosis.

The second and novel important observation of our study was that lung function changes over time were remarkably different in the four groups identified in Lifelines (Figure 3). Whereas FEV₁ decline was, as expected, within the normal range in participants with normal peak FEV₁ (even in normal‐limited), it was remarkably reduced in PRISm or low‐limited participants, suggesting that these individuals may be already travelling in different lung function trajectories that may have started earlier in life. ²⁵ In fact, most of the individuals (91.2%) remained in the same group at 5 years of follow‐up, and in agreement with the FEV₁ decline observed, most changes if happening were from normal to normal‐limited. Previous works have identified as risk factors of accelerated decline in young adults smoking, diagnosis of asthma, chronic bronchitis and alterations in the CT scan (as small airway disease and ground‐glass opacification).^{26, 27} In agreement with the previous findings, we observe in peak lung function (prior to the decline phase) a higher smoking exposure (packs‐year) and diagnosis of asthma in Low‐limited and normal‐limited but not in PRISm, showing that the same factors associate to early decline and low peak. Our work also suggests, that early interventions on factors related to previous respiratory conditions (e.g., asthma) and structural or functional pulmonary impairments, could help clarify the mechanisms by which high‐risk young individuals follow distinct clinical trajectories. Our observations in BAMSE show that these different lung function spirometry patterns were already apparent during childhood and adolescence (Figure 2). Collectively, therefore, these observations strongly support the critical relevance of early life events in determining respiratory health later in life^{28, 29} and the importance of measuring spirometry in children and adolescents as a global health marker.^{3, 29} Additionally, our findings suggest that biomarkers such as eosinophil counts and FEV₁ measurements may serve as valuable tools for stratifying risk profiles, thus enabling a more targeted therapeutic approach.

Finally, the current study in Lifelines and our previous publication in BAMSE show that, both in childhood/adolescence and early adulthood, most lung function trajectories remain stable over time but there is some plasticity at the individual level, so few individuals can improve or worsen their functional status. This opens opportunities for prevention and/or treatment. Importantly, the prevalence of reduced FEV₁ due to either PRISm or airflow limitation was largely unaltered if the LLN was used to define airflow limitation instead of a fixed ratio value lower than 0.7.

A clear strength of our study is that, to our knowledge, it is the first to characterize, cross‐sectionally and longitudinally, both looking forward and backwards, the spirometry patterns of peak FEV₁ in two European independent, population‐based, large cohorts of young adults. Among potential limitations we acknowledge that spirometric measurements are pre‐bronchodilator, that may lead to a misclassification as some individuals may have a reversible form of airflow limitation that indicates asthma. Results, however, were largely unchanged after excluding participants with a diagnosis of asthma. Additionally, previous population based studies have used prebronchodilator airflow limitation to define COPD, and in COPD cohorts, prebronchodilator airflow limitation has been shown to identify individuals at high‐risk to develop postbronchodilator airflow limitation (COPD). ³⁰ To identify airflow limitation we used the fixed ratio, but previous studies using FEV₁/FVC < LLN instead of the 0.7 cutoff showed comparable results suggesting minimal impact on outcomes. ³¹ Additionally, potential bias might arise from lung function technical measurements performed in childhood and adults. Future analysis should compare the factors associated to lung function decline in this young population, to those factors previously associated in the literature to the decline in young COPD, and to different lung function trajectories associated to COPD. Finally, the prevalence of PRISm varies greatly by race/ethnicity. Most participants in our study were of Caucasian origin our results will have to be validated in other ethnias to explore the potential influence of genetics, different environmental exposures, different prevalences of comorbidities and/or socio‐economic status effects, among others.

In conclusion, low peak FEV₁ in early adulthood is most often due to PRISm and results in a significant burden of respiratory symptoms. However, only low‐limited and normal‐limited, not PRISm, associate with a doctor diagnosis of asthma. Importantly, all abnormal spirometry patterns can be already identified in childhood and adolescence and persist post‐peak. Collectively, these observations indicating a need for differentiated clinical approaches and open new opportunities for prevention, early intervention and promotion of healthier ageing.

AUTHOR CONTRIBUTIONS

Nuria Olvera: Formal analysis (lead); methodology (lead); software (lead); writing – original draft (equal); writing – review and editing (equal). Alvar Agusti: Conceptualization (equal); formal analysis (equal); writing – original draft (equal); writing – review and editing (equal). Judith M. Vonk: Data curation (equal); formal analysis (supporting); methodology (supporting); writing – review and editing (equal). Gang Wang: Data curation (equal); formal analysis (equal); writing – review and editing (equal). Jenny Hallberg: Data curation (equal); formal analysis (supporting); writing – review and editing (equal). H. Marike Boezen: Conceptualization (supporting). Maarten van den Berge: Conceptualization (equal); formal analysis (supporting); methodology (supporting); supervision (supporting); writing – review and editing (equal). Erik Melén: Conceptualization (equal); data curation (supporting); formal analysis (supporting); supervision (supporting); writing – review and editing (equal). Rosa Faner: Conceptualization (lead); formal analysis (lead); methodology (equal); supervision (lead); writing – original draft (lead); writing – review and editing (equal).

FUNDING INFORMATION

This work was supported in part by Instituto de Salud Carlos III (ISCIII) (PI21/00735), SEPAR and RF is a Serra Hunter program. Funded by the European Union (ERC, PredictCOPD, 101044387 and TRIBAL, 757919). NO has PhD fellowship FI‐AGAUR (2022FI_B00371) from Generalitat de Catalunya and European Social Fund. BAMSE was also supported by the Swedish Heart‐Lung Foundation, the Swedish Research Council and ALF Region Stockholm. Views and opinions expressed are however those of the authors only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.

CONFLICT OF INTEREST STATEMENT

MvdB is an Editorial Board member of Respirology and a co‐author of this article. He was excluded from all editorial decision‐making related to the acceptance of this article for publication.

RF is supported by the Serra Húnter program, holds grants from the ISC‐III and EU, has received lecture fees or consulting fees from AZ and Chiesi outside the submitted work, and has industry grants from AZ, Menarini and Sanofi for unrelated research.

EM has received lecture fees or consulting fees from ALK, AstraZeneca, Chiesi and Sanofi outside the submitted work.

AA reports consulting fees and/or lectures fees and/or research grants from GSK, AZ, Chiesi, Menarini, Zambon, MSD, Sanofi, and reports being the Chairman of Board of Directors of Gold, all activities outside the submitted work.

HUMAN ETHICS APPROVAL DECLARATION

This study was approved by the Ethics Committees at The University Medical Center Groningen (METC UMCG METc 2007/152) in The Netherlands and the Karolinksa Institute (Ref 2016/1380‐31/2) in Stockholm, Sweden. All participants (or parents) signed their informed consent for participation in this research study.

Supporting information

Data S1. Supporting Information.

Table S1. Clinical variables selected by experts for the multivariate regressions and accumulation symptoms.

Table S2. Lifelines 25–35 years baseline: four groups based on baseline FEV₁ and FEV₁/FVC <0.7.

Table S3. Lifelines 25–35 years baseline: four groups based on baseline FEV₁ and FEV₁/FVC < LLN.

Table S4. Lifelines 25–35 years baseline: four groups based on baseline FEV₁ and FEV₁/FVC <0.7 no doctor diagnosis of asthma.

Table S5. Lifelines 25–35 years baseline: multivariable and multinomial regressions.

Table S6. Lifelines 5 years follow‐up: four groups based on baseline FEV₁ and FEV₁/FVC <0.7.

Table S7. BAMSE 24 years: four groups based on FEV₁ and FEV₁/FVC <0.7.

Table S8. Lifelines 25–35 years FEV₁ > or < LLN.

Olvera N, Agusti A, Vonk JM, Wang G, Hallberg J, Boezen HM, et al. Heterogeneity of reduced FEV₁ in early adulthood: A looking forward, looking backwards analysis. Respirology. 2025;30(4):326–334. 10.1111/resp.14876

DATA AVAILABILITY STATEMENT

Data is available following the established procedures, see: https://www.lifelines-biobank.com/.