Discussion
In preschool age the diagnosis of asthma is particularly challenging due to both the heterogeneity of wheezing and to the difficulties of performing acceptable and reproducible spirometric manoeuvres. However, objectifying the lung function of these patients remains crucial to reduce the risk of misdiagnosis, over and under treatment of patients. In this context, the IOS represents a useful method for assessing lung function and BDR in young children. Several studies have demonstrated that IOS is capable of identifying airway obstruction and response to bronchodilators and bronchoconstrictors [17,19,23,24,26,27,30].
The presented study was aimed to identify the optimal oscillometric cut-offs of the BDR in preschool children with history and symptoms consistent with asthma. Regarding baseline measurements, we found a significant good correlation between FEV1 and IOS indices R5 (ρ=-0.67), X5 (ρ=0.62) and AX (ρ=0.6). Previously, Carvalho et al. found a significant weak to moderate associations only between FEV1 and IOS parameters that reflect small airway obstruction (Di5-20 and AX), whereas the correlation between FEV1 and R5 was minimal [31].
However, after BPD, a significant good correlation with FEV1 was demonstrated only for R5 (ρ=-0.64) whereas X5 and AX exhibited a significant moderate correlation (ρ=0.42 and ρ=-0.44, respectively). These data are consistent with those published by Olaguíbel et al., who demonstrated that R5 was correlated with FEV1 at both baseline and post-bronchodilator in 33 asthmatic preschoolers [32]. Moreover, among IOS parameters, R5 and X5 differed significantly between patients with positive and negative BDR. Specifically, the magnitude of this variation was of 26.82% ± 6.58 vs. 18.18% ± 10.08 for R5 and of 34.3% ± 9.83 vs. 20.59% ± 18.43 for X5 in patients with positive and negative BDR test, respectively. Considering that a positive BDR in patients with a symptom pattern consistent with asthma configures these patients as asthmatic and that in older children IOS is more sensitive than spirometry in assessing BDR, this finding is not surprising [10,22].
Contrary to findings by Malmberg et al., in our study the BDR was not related to baseline lung function since no significant difference in baseline spirometry and IOS parameters was found between patients with positive- and negative BDR [33].
Regarding to the optimal cut-offs for establishing a positive BDR with IOS in preschool children, we found that a decrease in R5 of 25.7% and an increase in X5 of 25.7% on baseline exhibited the best combination of sensitivity and specificity to detect an increase of FEV1 ≥12% and/or ≥200 mL with an AUC of 0,77 and 0,75, respectively. A recent ERS technical standard stated that recommended IOS thresholds that define BDR for both children and adults are −40% in R5, +50% in X5 and −80% in AX [34]. However, in preschoolers, there is still no consensus in the literature regarding standardization of oscillometric cut-offs for BDR.
Our findings are in close agreement with those obtained by Gleason et al. and Bisgaard and Nielsen, who suggested as cut-off a decrease in R5 of respectively 24% and 29% in children aged 2 to 6 years [35,36]. Similarly, Marotta et al. proposed to consider a decrease between 20 and 25% in R5 as a positive IOS BDR and this range was confirmed by Konstantinou et al. and Shin et al., who found a cut-off of 20.5% and 19% respectively [23,30,37]. In contrast, different results were obtained by other studies, that proposed a decrease in R5 between 37% and 43% as threshold to define a positive BDR [24-26].
In reference to X5 parameter, Shin et al. reported results comparable with those obtained in our work since they considered an increase of 24% as a positive BDR [30]. Nielsen et al. and Thamrin et al., on the other hand, fixed this threshold at 42 and 61%, respectively [25, 34, 35].
Finally, in our study, AX was not able to discriminate between a positive and negative BDR test (AUC 0,53; p = 0,83) although the work of Oostveen et al. suggested that a decrease on baseline of 81% is indicative for a positive BDR [24].
The large discrepancies detectable between the previously mentioned studies might be attributable to differences in IOS technique, populations, and study design. In fact, many of the reported thresholds for different oscillometric parameters are based on varying, heterogeneous criteria for the differentiation between asthmatic and non-asthmatic subjects (e.g., clinical diagnoses, questionnaire-based diagnoses). This fact, combined with the wide heterogeneity of preschool wheezing that complicates the diagnostic work-up of asthma in these patients, might explain the different results of various studies.
Our threshold values for oscillometric parameters were based on an objective criterion, such as a 12% increase in FEV1 and/or ≥ 200 mL in patients able to perform acceptable and reproducible spirometry and high quality oscillometric measurements. This explains the small number of patients analyzed (n=36) despite the larger number of patients initially examined (n=72).
Nevertheless, the small number of evaluated patients and the retrospective nature of the study remain a limitation of the presented work. Furthermore, the approach to derive oscillometric thresholds from a less sensitive technique such as spirometry remains a limitation. However, BDR in spirometry still presents an objective consensus technique to diagnose asthma and might therefore be superior to discriminating thresholds by clinical parameters.
In conclusion, the IOS represents a useful technique since it requires only minimal collaboration of the patient and provides additional information to the spirometry; moreover, the BDR test is a further tool able to improve the sensitivity of IOS in detecting patients with small airway obstruction. Our data demonstrated a positive BDR at a 26% decrease in R5 and a 26% increase in X5 in preschoolers with symptoms consistent with asthma.