Methods
The study was approved by the Institutional Ethics Committee and signed informed consent was obtained from parents/guardians along with the assent of children/adolescents.
Study population
Patients aged 10 to 20 years were recruited from the pediatric allergy and immunology clinic of the Hospital das Clínicas of the Universidade Federal de Pernambuco, in Recife – Brazil between 2017 and 2019. All had received a specialist diagnosis of asthma based on GINA criteria25.
Individuals were excluded if they had a recent history (less than 6 weeks) of asthma exacerbation or acute respiratory infection symptoms, were regularly using inhaled steroids, were smokers or ex-smokers, had any other known pulmonary diseases, were unable to perform adequate forced expiratory maneuvers, or had a baseline FEV1< 60% of predicted FEV1.
Procedures
All subjects were required to attend the laboratory on two occasions separated by a period of 2 to 4 days at a similar time of day (+2 hours), in the morning. On the first visit, anthropometric measurements were taken and patients responded to the Asthma Control Test (ACT) (Portuguese/Brazilian version)26. They then underwent spirometry to determine baseline FEV127 using a daily calibrated spirometer (KitMicro Cosmed, Rome, Italy). The best of three acceptable maneuvers was chosen. If the value was > 60% of that predicted (for the Brazilian population28), individuals proceeded to the EVH challenge. On the second visit, baseline FEV1 was measured and, if the best value lay within 15% of that of the first day (but higher than 60% of predicted), patients proceeded to the EVH challenge. The laboratory room temperature was controlled and air humidity was measured using a thermo hygrometer (Incoterm, Brazil).
Patients were instructed to abstain from caffeine and exercise on test days and from using short- and long-acting beta-2 agonists for 12 and 48 hours, respectively. It was recommended that patients not use inhaled corticosteroids on the test day13.
Eucapnic voluntary hyperpnea and spirometry
Patients breathed a mixture of dry room-temperature air with 5% added carbon dioxide (CO2) (White-Martins, Recife, PE - Brazil) collected in a Douglas balloon, through the mouth, with nose clipped using a one-way low-resistance valve (Laerdal, Copenhagen -Denmark). The test lasted 6 min and the target ventilation per minute was set at 21 times that of the personal baseline FEV1 (equivalent to 60% of predicted maximal voluntary ventilation)13. The ventilation rate per minute was monitored using an analog ventilometer (Wright Mark 8 NSPIRE Health, Colorado - USA) and subjects were coached every 30 seconds to maintain the target ventilation level. The ventilation rate was recorded every minute for six minutes and the mean expressed as a percentage of the calculated target ventilation. Spirometry for FEV1 measurement was performed in duplicate 5, 15 and 30 minutes after EVH and the highest value recorded. The % drop in FEV1 was calculated using the formula (baseline FEV1– post-EVH FEV1/baseline FEV1.100) and the maximum decrease in FEV1 at any of the three points in time was recorded (FEV1fallmax%)13. Subjects were considered positive for EIB diagnosis if a decrease of> 10% in FEV1 was observed at any evaluation time-point on at least one of the test days.
Individuals were divided into three groups: those with a FEV1fallmax% > 10% on both test days, those with FEV1fallmax% > 10% on one day only (divergent group) and those with FEV1fallmax% <10% on both test days. The severity of EIB was graded as mild, moderate, or severe if the percentage decrease in FEV1 from the pre-exercise level was > 10%< 25%,> 25%< 50%, and> 50%, respectively13.
Statistical analysis
Data were processed and analyzed using Statistical Package for the Social Sciences (SPSS), version 20.0 and figures produced using GraphPadPrism® version 6.00 for Windows. Normally distributed data (Kolmogorov-Smirnov) are expressed as means + SD and/or 95% confidence interval (CI) and non-normal data as median and interquartile range (IQR). Agreement was evaluated as a binary outcome considering the proportion of positive and negative response on both tests (proportion of agreement). Reproducibility (repeatability) and limits of agreement (LOAs) were assessed for FEV1fallmax% and AUC0-30min using the method described by Bland & Altman29. Individuals who tested positive on both days with a FEV1falmax% > 15% and those with > 20% were also analyzed separately for bias and LOAs in these sub-groups. The area under the curve up to the thirtieth minute (AUC 0-30min) was calculated by the trapezoidal method using the percentage of FEV1fallmax% at each point in the evaluation. For the calculation of AUC0-30min, positive variations in FEV1after EVH were considered zero. Multivariate analysis of variance (MANOVA) was used to compare baseline FEV1 as a percentage of predicted and the achieved ventilation rate as a percentage of the target calculated by group (with and without BIE) and visit. The Kruskal-Wallis test was performed to assess the differences in ACT between the groups. The intra-class correlation coefficient (ICC) for the FEV1falmax% and AUC0-30min was calculated using a two-way mixed-effect model with the mean single measurement reported to evaluate relative reliability. The alpha error probability was set at 0.05.