Materials and Methods
Study design
This cohort study complied with STROBE guidelines and with the guidance provided by the editors of respiratory, critical care, and sleep journals.5
The La Berma open cohort enrolls asthmatic children since 2009. This open cohort is constituted of 7817 children with asthmatic symptoms (with or without confirmed variable expiratory flow limitation). For this study, selection was made in children with confirmed asthma: suggestive symptoms and (1) a significant bronchodilator response (either sRaw or FEV1, n=1152) or (2) by an asthma exacerbation diagnosed and treated in a hospital Emergency Department (n=1295) or (3) both (n=777). Preterm birth (gestational age <37 weeks) was a non inclusion criterion as it is a well-known risk factor for the development of persistent airflow limitation.6 Only children who had at least 10 pulmonary function tests after 8 years of age were selected to ensure the quality of spirometry, leaving 295 children with confirmed asthma. The characteristics of each visit have been standardized,7 as described in Table 1.
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 non-interventional research.
Pulmonary function tests
Spirometry (MasterScreen Body; Jaeger, CareFusion, San Diego, CA, USA) was performed without inhaled treatment (bronchodilator or LABA/ICS association) on the day of measurement by the same operator (BM),7 according to international guidelines.8 Reference values were those of GLI-2012,9 as recommended for describing the progression of pulmonary function.10
Classification of lung function patterns
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.11 Three patterns were defined. A positive slope was defined by a significantly positive value of the slope, whereas a negative slope was defined by a significantly negative value of the slope according to the P-value obtained in the Pearson test. When the P-value was not significant, the slope was considered null. The standard deviation (SD) of each individual slope was recorded as a criterion of FEV1 variability.
Statistical analyses
The results were expressed as median (25th and 75th percentile) as some indices did not conform to a normal distribution, and the positive slope 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 chi-square test or Fisher’s exact test where appropriate. A P-value <0.05 was deemed statistically significant. All statistical analyses were performed with Statview 5.0 software (SAS institute, Cary, NC, USA).