Detecting Impaired Lung Growth Using the Conditional Change Score of FEV₁ in an Open Cohort of Children With Asthma

ABSTRACT

Background

It has been suggested that a quarter of children with asthma show impaired lung growth, with better initial function as a risk factor. A conditional change score (CCS) for FEV₁ that is independent of the initial FEV₁ and that takes into account the time interval between the measurements has been proposed, a value < −1.96 being predictive of impaired growth.

Objective

To evaluate whether the use of the CCS allows the identification of a subgroup of children with asthma that has an impaired lung growth, and to identify the risk factors of impaired growth.

Methods

We reanalyzed the data of 295 children with confirmed asthma (199 boys) who had undergone at least 10 spirometry tests from the age of 8 who were selected from a single‐center open cohort. The annualized rate of change (slope) for prebronchodilator FEV₁ (percent predicted) was estimated for each participant. Using the first and last visit of follow‐up, the CCS was calculated.

Results

In total, 46 children (16%, 95% confidence interval: 12–20) followed up for a median duration of 6.5 years [interquartile: 5.7–7.3] exhibited an impaired lung growth. The CCS correlated with the standard deviation of individual slopes (FEV₁ variability criterion: R = 0.21; p < 0.001), the z‐score of FEV₁ at first visit (R = −0.22; p < 0.001), and the body mass index at final visit (R = 0.16; p = 0.005), which remained independently associated with the CCS.

Conclusion

We confirm that some children with asthma (16%) show impaired lung growth with better initial lung function and decreased FEV₁ variability as risk factors.

1. Introduction

Recent ERS/ATS technical standard on interpretive strategies for routine lung function tests stated that “interpretation of serial measurements relies on accurate limits of reproducibility of a pulmonary function test index, including the natural changes over time and the changes that would be considered outside the normal biological variability over both short and long periods of time” [1]. The methodological study of Stanojevic and colleagues (47,938 FEV₁ measures from 7885 healthy children 6–18 years of age) showed that there was a correlation between the first (z‐score of initial FEV₁) and the difference between two measurements [2]. A healthy individual with a higher initial value is more likely to have a lower measurement at the second visit. Thus, they derived a conditional change score for FEV₁ that is independent of the initial FEV₁ value and that also takes into account the time interval between the two measurements [2].

At least two studies have shown that children with asthma can exhibit impaired lung growth based on spirometry [3, 4]. In the first study, a reduction in postbronchodilator FEV_1% predicted was observed in 25.7% of children with mild to moderate asthma who were enrolled in the Childhood Asthma Management Program [3]. Factors associated with a reduction in postbronchodilator FEV_1% predicted included a younger age at enrollment, male sex, and higher postbronchodilator FEV_1% predicted at enrollment. In the second study, from our group, the annualized rate of change (slope) for prebronchodilator FEV₁% predicted (at least 10 measures) was estimated using standard least squares linear regression models in 295 children with asthma [4]. The children characterized by impaired lung growth (82/295, 27.8%) showed a better initial function and a lower FEV₁ variability during their follow‐up than children with tracking (68.8%) or increased lung growth (3.4%). Thus, both studies showed that better lung function was a risk factor for subsequent impaired lung growth and showed similar percentages of impaired lung growth in young asthmatics.

One may wonder whether the results of these previous studies in children with asthma [3, 4] may have been explained by the bias evidenced by Stanojevic et al. [2] Therefore, the objective of our new retrospective study was to assess whether the calculation of conditional change score of FEV₁ in our cohort of 295 children with asthma gave new insights in pediatric lung function trajectories [4] since there is a need to identify children with asthma at higher risk of poor lung growth.

2. Methods

The La Berma open cohort enrolls since 2009 asthmatic children who are followed up in an out‐of‐hospital setting by a pediatric pulmonologist (B. M.). For this study, selection was made in children with confirmed asthma: suggestive symptoms and a significant bronchodilator response (either decrease in sRaw ≥ 42% from baseline or increase in FEV₁ ≥ 12% from baseline) or by an asthma exacerbation diagnosed and treated in a hospital Emergency Department or both. Preterm birth (gestational age < 37weeks) was a noninclusion criterion as it is a well‐known risk factor for the development of persistent airflow limitation. The criterion reducing the number of children was the availability of at least 10 pulmonary function tests after 8 years of age, leaving 295 children having had 3812 spirometric sessions (prebronchodilator values). The characteristics of each visit have been standardized, as previously described [4, 5].

The annualized rate of change (slope) for prebronchodilator FEV₁ percent predicted was estimated for each participant using standard least squares linear regression models, as performed by others [6]. The standard deviation (SD) of each individual slope was recorded as a criterion of FEV₁ variability.

Reference values of spirometry were those of GLI‐2012 [7], as recommended for describing the progression of pulmonary function [8]. The z‐scores of FEV₁ allowed the calculation of the conditional change score: [FEV₁ z‐score at Time 2 − (r × FEV₁ z‐score at Time 1)]/$\sqrt{1-r^2}$ with r = 0.642 − 0.04 × Period between the two visits (years) + 0.020 × Age at Time 1 (year) [2]. Time 1 was the visit of inclusion of the participant in the cohort and Time 2 was the last follow‐up visit. Changes within ±1.96 change scores are considered within the normal limits [1], those < −1.96 were labeled as reduced growth and those > 1.96 were labeled as enhanced growth.

This cohort was registered to our regulatory agency for electronic data collection (Commission Nationale Informatique et Libertés, No. 1408710). Approval from the Ethics Committee of the French learned Society of Pulmonology (SPLF) was obtained (CEPRO 2009/019). All children and parents were informed of the prospective recording of clinical and physiological data and could request to be exempted from the study in accordance with French law regarding noninterventional research.

Our study complied with STROBE guidelines of cohort studies.

3. Statistical Analyses

The results were expressed as median (25th and 75th percentile) as some indices did not conform to a normal distribution, and the enhanced growth group was small. Comparisons of continuous variables between the three groups of children were performed using the Kruskal–Wallis test, and subsequent intergroup comparisons were performed using the Mann–Whitney U test. Categorical variables were compared using the χ ² test or Fisher's exact test where appropriate. Linear correlations were evaluated using the Pearson correlation coefficient. The other statistical analyses are described in the text. A p value < 0.05 was deemed statistically significant. All statistical analyses were performed with StatView 5.0 software (SAS Institute, Cary, NC, USA).

4. Results

The characteristics of the 295 participants are reported in Table 1, according to the three classes of their conditional change scores. As compared to the two other groups, the asthmatic children with impaired lung growth were characterized by a best FEV₁ reversibility ever and a lower level of lung function at the last visit.

Table 1.

Characteristics (number or median [25th; 75th percentile]) of the 295 children with asthma according to the three groups of scores of conditional change of FEV₁.

Characteristics N (% population, [95% CI])	CC score > +1.96 6 (2%, [1–4]) Enhanced growth Group 1	−1.96 ≤ CC score ≥ +1.96 243 (82%, [78–87]) Normal lung growth Group 2	CC score < −1.96 46 (16%, [12–20]) Reduced growth Group 3	p valued	Intergroup comparisonse
Sex, male (%)	5 (83)	165 (68)	29 (63)	0.572
Age at first symptoms, years	0 [0; 3]	1 [0; 3]	0 [0; 4]	0.853
Atopy status				0.891
Negative skin prick tests, n	0	25	4
One positive prick test, n	1	51	11
More than one positive prick test, n	5	167	31
Parental atopy, n	4	115	28	0.099
Parental asthma, n	2	99	17	0.842
Longitudinal PFT characteristics
Duration of follow‐up, years	5.9 [5.4; 6.3]	6.5 [5.7; 7.3]	6.6 [5.9; 7.3]	0.352
ED Hospitalization frequencya	0.000 [0.000; 0.036]	0.071 [0.000; 0.216]	0.000 [0.000; 0.119]	0.205
Number of spirometric sessions	11 [10; 12]	12 [11; 14]	13 [11; 15]	0.082
Best sRaw reversibility in the past	53 [52; 64]	54 [46; 62]	55 [49; 62]	0.665
Best FEV1 reversibility in the past	14 [10; 26]	20 [15; 29]	25 [19; 36]	0.028	1,2 < 3
FEV1 slope, % per year	+3.30 [+1.39; +4.15]	−0.78 [−1.91; 0.28]	−2.94 [−3.88; −2.18]	< 0.001	1 > 2 > 3
SD of FEV1 slope	1.30 [1.06; 1.51]	1.01 [0.75; 1.38]	0.96 [0.80; 1.42]	0.192
Spirometry at inclusion
Age, years	9.0 [8.1; 9.5]	8.5 [8.2; 9.3]	8.4 [8.2; 9.1]	0.262
FEV1, z‐score	−1.58 [−2.29; −1.33]	−0.55 [−1.38; 0.18]	0.05 [−0.85; 0.77]	< 0.001	1 < 2 < 3
FVC, z‐score	−0.73 [−1.83; −0.30]	0.26 [−0.38; 1.06]	0.52 [−0.11; 1.23]	0.008	1 < 2,3
FEV1/FVC, z‐score	−1.40 [−1.78; −1.10]	−1.27 [−2.03; −0.45]	−0.99 [−1.64; −0.27]	0.252
Obstructive defect, n (%)	3 (50)	99 (41)	11 (24)	0.083
Spirometry at last follow‐up
FEV1, z‐score	1.11 [0.47; 1.26]	−0.77 [−1.36; −0.08]	−2.14 [−2.66; −1.58]	< 0.001	1 > 2 > 3
FVC, z‐score	1.19 [0.00; 1.78]	−0.07 [−0.58; 0.76]	−0.79 [−1.31; −0.19]	< 0.001	1,2 > 3
FEV1/FVC, z‐score	−0.88 [−0.94; −0.28]	−1.25 [−1.93; −0.49]	−2.13 [−2.99; −1.64]	< 0.001	1,2 > 3
Obstructive defectb, n (%)	0 (0)	85 (35)	34 (74)	< 0.001	1,2 < 3
Decreased FEV1 (z‐score < −1.645), n (%)	1 (17)	36 (15)	34 (74)	< 0.001	1,2 < 3
Clinical characteristics at last follow‐up
Age, years	14.8 [14.3; 15.4]	15.4 [14.8; 16.0]	15.4 [14.8; 16.0]	0.262
BMI, kg.m−2	19.1 [18.1; 21.6]	20.1 [18.5; 22.0]	19.1 [18.2; 20.7]	0.088
SABA on demand, n	3	102	15	0.443
ICS only, n	0	10	0	0.330
ICS dose, BED µg/dayc	400 [175; 400]	400 [200; 400]	400 [200; 400]	0.510
ICS‐LABA, n	3	131	31	0.230
Partially or uncontrolled last 3 months, n (%)	0 (0)	106 (44)	17 (37)	0.079
Days, symptoms within last 3 months	0 [0; 0]	1 [0; 7]	0 [0; 5]	0.069
Severe exacerbation within last 3 months, n	0	22	3	0.642

Note: Categorical variables were compared using the χ ² test or Fisher's exact test where appropriate.

Abbreviations: BED, beclomethasone‐equivalent daily dose; BMI, body mass index; CC, conditional change; CI, confidence interval; ICS, inhaled corticosteroid; LABA, long‐acting beta agonist; SABA, short‐acting beta agonist treatment; SD, standard deviation.

^a Rate per year of hospitalization in a hospital Emergency Department (ED) for severe asthmatic exacerbation, calculated since 2 years of age.

^b An obstructive defect is defined by a z‐score of FEV₁/FVC < −1.645.

^c Prescribed ICS dose.

^d Using the Kruskal–Wallis test.

^e Using the Mann–Whitney U test.

The correlates of the conditional change score were the slope of annualized rate of change in FEV₁ (R = 0.74; p < 0.001), the SD of each individual slope (FEV₁ variability criterion: R = 0.21; p < 0.001), the z‐score of FEV₁ at first visit (R = −0.22; p < 0.001) and the BMI at final visit (R = 0.16; p = 0.005).

In a multiple regression analysis with the conditional change score as dependent variable and the SD of individual slope, the z‐score of FEV₁ at first visit and the BMI at final visit (the slope of annualized rate of change in FEV₁ was not included due to collinearity) as independent variables, these three variables remained independently associated with the conditional change score (adjusted r ² of the model = 0.11, p < 0.001): SD of individual slope, estimated effect: 0.372, standard error: 0.104, p < 0.001; z‐score of FEV₁ at first visit: estimated effect: −0.214, standard error: 0.057, p < 0.001; BMI at final visit: estimated effect: 0.083, standard error: 0.025, p = 0.001.

Thus, we observed a strong positive correlation between the new conditional change score and our annualized rate of change in FEV₁, with a substantial agreement (percent overall agreement: 75%, free‐marginal kappa: 0.62; 95% confidence interval [0.54, 0.69]).

5. Discussion

The main result of our study is to confirm that a significant percentage (16%) of children with asthma exhibit poor lung growth with better initial lung function, decreased FEV₁ variability, and underweight as risk factors.

The conditional change score identified 16% (95% confidence interval: 12%–20%) of our study population who had a greater lung function deterioration than what would be expected otherwise in health. This percentage was lower than that found in our previous study (16% vs. 28%, p < 0.001) or in the Childhood Asthma Management Program (26%) [3], which is important to highlight. Nevertheless, the substantial agreement between the conditional change score and our annualized rate of change in FEV₁ suggests that the conditional change score obtained from only two measures is an easier mean to identify children with asthma at higher risk of poor lung growth. Our percentage of children with impaired lung growth is consistent with the 16% observed in adults with asthma who had irreversible airway obstruction after 21–33 years in a clinic in Groningen, the Netherlands [9] or with the 23% observed in adults with moderate to severe asthma who had irreversible airflow obstruction in a 10‐year follow‐up study [10]. Most of our children with impaired lung growth had an obstructive prebronchodilator defect at their last visit (74%), and could be at risk of subsequent chronic obstructive pulmonary disease [11]. Both the Melbourne (Australia) and Dunedin (New Zealand) longitudinal studies of the natural history of asthma show that impaired lung function in adult asthmatics often has its origins in childhood [12, 13].

What are the risk factors for impaired lung function with time in childhood?

We showed that the z‐score of FEV₁ at first visit remained correlated with the conditional change score, demonstrating that a higher level of lung function truly constitutes a risk factor of subsequent reduced growth in childhood asthma, which has previously been shown using specific airway resistance measurement [14]. This result seems specific to the lung growth period since the studies in adult asthma dealing with lung function decline did not find such a result [15]. Interestingly, it has been shown that some children with asthma have a very high lung distensibility and normal airway function (resulting in better lung function), whereas others have nearly normal lung elasticity but grossly reduced airway distensibility [16]. The former endotype could be at risk of impaired lung growth.

Overall, the main results of our previous analyses [4] were confirmed, which is an important result and are quite consistent with those of Gaffin and colleagues who determined phenotypic factors associated with FEV₁ throughout childhood in the Severe Asthma Research Program (SARP) 3 pediatric cohort [17]. The effect of baseline maximal bronchodilator FEV₁ reversibility was an inverse association with predicted FEV₁; accordingly, the best reversibility ever achieved was higher in our children with asthma with subsequent impaired lung growth. Similarly, increasing degree of bronchodilator reversibility at enrollment in the longitudinal study of Ulrik and Backer was associated with an increased risk for the presence of irreversible airflow obstruction [10]. Asthma severity did not remain statistically significant in the multivariate models of Gaffin and colleagues and there was no relationship of inhaled corticosteroid dose with lung function outcomes in unadjusted models [17], as observed in our study at the last visit. We confirm that less variability in prebronchodilator FEV₁ values over time is associated with reduced lung growth, which has previously been discussed [4] and that probably reflects the constitution of an irreversible airway obstruction. Finally, underweight was also a risk factor of impaired lung growth and/or overweight was a risk factor of enhanced lung growth. Along this line, it has been demonstrated that z‐scores of FEV₁ and FVC had similarly increasing slopes with z‐score of BMI in children [18], suggesting an association between body mass and lung growths.

Our study has limitations, first related to its monocenter design. Second, changes in prebronchodilator FEV₁, although influenced by remodeling, can also be attributable to reversible factors (e.g., acute bronchoconstriction) that is dampened by the high number of spirometric sessions per child included in our study that were used for slope calculation only. Nevertheless, the risk of chronic obstructive pulmonary disease would be better defined using postbronchodilator FEV₁ values. Third, the generalizability of the findings may be limited to children with asthma with an out‐of‐hospital follow‐up.

Our study has some practical consequences; the calculation of the conditional change score of FEV₁ should be implemented in daily practice to allow the early identification of children with asthma who are at increased risk of loss of lung function. A normal initial lung function z‐score (Table 1) should not be regarded as a healthier lung function since after 6–7 years of follow‐up, three quarters of these children will exhibit airflow limitation with reduced FEV₁.

In conclusion, using the conditional change score of FEV₁, we confirm that some children with asthma (16%) show impaired lung growth during childhood and a normal initial lung function is a risk factor of this impaired growth.

Author Contributions

Bruno Mahut: investigation, writing – original draft, validation, visualization. Plamen Bokov: writing – original draft, methodology, validation, visualization, software, formal analysis. Nicole Beydon: writing – original draft, methodology, validation, visualization, data curation. Christophe Delclaux: conceptualization, writing – original draft, methodology, formal analysis, project administration. All authors reviewed the results and approved the final version of the manuscript.

Ethics Statement

Approval from the Ethics Committee of the French learned Society of Pulmonology (SPLF) was obtained (CEPRO 2009/019). All children and parents were informed of the prospective recording of clinical and physiological data and could request to be exempted from the study in accordance with French law regarding noninterventional research.

Conflicts of Interest

The authors declare no conflicts of interest.

Mahut B., Bokov P., Beydon N., and Delclaux C., “Detecting Impaired Lung Growth Using the Conditional Change Score of FEV₁ in an Open Cohort of Children With Asthma,” Pediatric Pulmonology 60 (2025): 1–6, 10.1002/ppul.71295.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.