Discussion
The main result of our observationnal study is to confirm the result of the CAMP study showing that about a quarter of asthmatic children participating in out-of-hospital follow-up exhibited loss of lung function during childhood and adolescence.
We have previously shown, in a retrospective study, that a significant increase in prebronchodilator sRaw was observed in 17% of the children (mainly boys with a lower initial and higher final specific resistance) who suffered from persistent asthma.12 Nevertheless, the effect of lung growth (dysanaptic or isotropic, modifying sRaw via thoracic gas volume) was identified as a potential source of bias. The CAMP study used a predefined criterion of loss of lung function, which was at least 1% loss in postbronchodilator FEV1% predicted per year.4 Using a statistical definition of the loss of lung function, we were able to confirm the results of the CAMP study (25.7% versus 27.8%). It should be emphasized that our three groups were differentiated based on prebronchodilator FEV1 slopes. In most cohorts, the evaluation of longitudinal lung function is based on prebronchodilator values of FEV1 with bronchodilator withdrawal on the day of testing.13,14 This may introduce a source of bias as some asthmatic patients may occasionally exhibit some degree of airflow limitation and a positive bronchodilator response. In order to reduce the effect of occasional FEV1 variability, we calculated slopes using at least 10 spirometry results, as persistent significant bronchodilator response is a rare endotype of asthma (5%).15 Moreover, Bui and colleagues have recently shown that baseline asthma was associated with accelerated decline in both pre- and post-bronchodilator FEV1.16
The observation that children with better initial lung function at study entry exhibited a loss of lung function is similar to other studies.4,7,11 Their frequency of severe exacerbation leading to emergency department visit was not different, and they had a better level of control at their last visit, than children with tracking of lung function. Thus, asthma control does not seem to be a risk factor of loss of lung function in our study. Among risk factors of loss of lung function, elevated blood eosinophils have been associated with an accelerated decline in FEV1 and vital capacity compared to normal blood eosinophils in the younger asthmatic subjects in longitudinal studies.17 Overall, the inflammatory phenotype in asthma has prognostic relevance since the annual decline in FEV1 can also be predicted by the bronchial CD8+ cell infiltrate.18 In our study, sex, age at first symptoms and atopy status were not significantly different in children exhibiting loss of lung function while asthma severity was different, which seems overall consistent with the overview made by Ulrik.19
Recently, Denlinger and colleagues evaluated corticosteroid response endotypes as longitudinal predictors of lung decline of adults in the NHLBI Severe Asthma Research Program.11 The odds ratios of BMI (for 5-unit decrease) and baseline FEV1% predicted (for a 10-unit increase) for predicting decline were of borderline significance (1.06 [0.95, 1.19] and 1.07 [0.96, 1.19], respectively), which may be consistent with our results. Of note, the effect of BMI in these two studies may seem opposite to the expected effect as overweight has been associated with reduced FEV1/FVC z-score.20 Accordingly with our results, Graff and colleagues recently showed that a lower BMI is associated with lung function decline and irreversible airflow obstruction in adult asthma.21 Overall, why a better initial lung function is a risk factor of subsequent loss of lung function, which has consistently been found, remains unexplained but cannot seem ascribable to regression toward the mean. This phenomenon arises if a sample point of a random variable is extreme, a future point is likely to be closer to the mean or average; thus, in our study the slope criteria were defined a priori and were calculated based on at least 10 measurements.
An original finding of the current study is that less variability in prebronchodilator FEV1 values over time is associated with loss of lung function. Tantisira and colleagues showed that a single higher bronchodilator response at inclusion was an independent predictor of higher prebronchodilator FEV1 after four years in the CAMP study.22 Thus, occasional FEV1 variability could be associated with a better prognosis, which may explain that absence of variability is associated with a worst functional prognosis.
The European Respiratory Society /American Thoracic Society guidelines define severe asthma not only as asthma that remains uncontrolled despite aggressive drug therapy, but also one that requires aggressive therapy to prevent from becoming uncontrolled.23Airflow limitation (after appropriate bronchodilator withhold FEV1 <80% predicted, in the face of reduced FEV1/FVC defined as less than the lower limit of normal) that is one of the four criteria of uncontrolled asthma was quite frequent (15%) and logically more frequent in children with loss of function. Lung function may continue to decline in severe asthmatics despite high-intensity treatment and improved asthma control is not sufficient to prevent such progressive deterioration,24 as suggested by our results. It highlights that the future risks of severe asthma are poorly recognized by both patients and physicians, may be because patients with loss of function have the best initial lung functions. In the study of McGeachie et al. including 684 study participants of the CAMP trial, (mean age, 26.0±1.8 years), a total of 23% were classified as having reduced growth without an early decline while 26% were classified as having reduced growth and an early decline.14 Therefore, reduced growth (loss of lung function during childhood) may further affect the prevalence of chronic obstructive pulmonary disease that is an important message. Smoking prevention in these patients is mandatory.
An unexpected finding was the discovery of a small subset (3%) of asthmatic children who exhibited a gain in lung function. This result is in line with individual data of the CAMP study (see appendix of14), in which some children clearly exhibited a gain in lung function during childhood or adolescence based on prebronchodilator FEV1% predicted. This result is also consistent with the pattern of early low, accelerated growth, normal decline (8% of participants) evidenced in the Tasmanian Longitudinal Health Study (16% of participants with ever asthma and 4% with persistent asthma) which modelled lung function trajectories measured at 7, 13, 18, 45, 50 and 53 years based on prebronchodilator FEV1 z -scores.25
Our study has some limitations, such as its monocenter design. The children were in an open cohort; thus, there was a selection bias as asthmatic children who were followed up for years were probably more symptomatic. The percentage of children exhibiting loss of lung function may have been overestimated. Nevertheless, this percentage was similar to that observed in the CAMP study that included mild to moderate persistent asthmatic children, which deserved to be confirmed in an out-of-hospital setting.
In conclusion, we confirm the results of the CAMP study showing that a significant proportion (27.8%) of asthmatic children exhibits a loss of lung function during childhood and adolescence, additionally a small subset of asthmatic children exhibits a gain in lung function.
Author Contributions: Conceptualization: BM and CD2; Formal analysis: PB, CD2; Investigation: BM; Methodology: PB, NB; Project administration: CD2, NB; Supervision: CD2, NB; Validation: BM, CD2; Roles/Writing - original draft: CD2; Writing - review &editing: BM, PB, NB.