Airwave oscillometry and spirometry in children with asthma or wheeze

Abstract

Objective

Lung function testing is used in diagnosing asthma and assessing asthma control. Spirometry is most commonly used, but younger children can find performing this test challenging. Non-volitional tests such as airwave oscillometry (AOS) may be helpful in that population. We compared the success of spirometry and AOS in assessing bronchodilator responsiveness in children.

Methods

AOS was conducted alongside routine lung function testing. Resistance at 5 Hz (R5), the difference between the resistance at 5 and 20 Hz (R5-20) and the area under the reactance curve (AX) were assessed. Patients between 5 and 16 years old attending clinic with wheeze or asthma were assessed. Patients performed AOS, followed by spirometry and were then given 400 µg salbutamol; the tests were repeated 15 minutes later.

Results

Lung function testing was performed in 47 children of whom 46 (98%) and 32 (68%) performed acceptable baseline oscillometry and spirometry, respectively (p < 0.001). Children unable to perform acceptable spirometry were younger (7.35, range: 5.4–10.3 years) than those who could (10.4, range: 5.5–16.9 years), p < 0.001. The baseline z-scores of AOS R5 correlated with FEV₁ (r = 0.499, p = 0.004), FEF₇₅ (r = 0.617, p < 0.001), and FEV₁/FVC (r = 0.618, p < 0.001). There was a positive bronchodilator response assessed by spirometry (change in FEV₁ ≥ 12%) in eight children which corresponded to a change in R5 of 36% (range: 30%–50%) and a change in X5 of 39% (range: 15%–54%).

Conclusions

Oscillometry is a useful adjunct to spirometry in assessing young asthmatic children’s lung function. The degree of airway obstruction, however, might affect the comparability of the results of the two techniques.

Keywords:

• Lung function
• bronchodilator responsiveness
• forced expiratory volume in 1 second (FEV1)
• resistance at 5Hz (R5)
• pediatrics

Introduction

Asthma is the most common chronic respiratory disease worldwide (1). In the UK, one in 11 children has a diagnosis of asthma (2), which is associated with significant morbidity as indicated by 2.8 million school days lost annually (3). The most frequent signs are cough, wheeze, and breathlessness (4). Reversible airflow obstruction, as assessed by spirometry, is an objective assessment to support a diagnosis of asthma in children and young people aged 5–16 years with suggestive signs (5). The European Respiratory Society (ERS) Task Force for pediatric asthma recommends to achieve asthma control with minimization of asthma manifestations (6,7), the key monitoring tools are symptomology and lung function testing (7). The Global Initiative for Asthma recommends spirometry as the gold standard lung function test (8). Spirometry, however, can be challenging in younger children because of the requirements to make a maximal effort and to follow instructions to generate acceptable maneuvers (9). A paucity of trained respiratory technicians to supervise spirometry measurements is another obstacle to its use (10). An alternative to spirometry which may be particularly useful in younger children is the non-volitional forced oscillation technique (FOT), in which oscillations are superimposed on passive tidal breathing to assess respiratory impedance (11). As well as a lower requirement for patient comprehension and operator-training, FOT is a non-aerosol generating procedure, which minimizes the potential transmission of SARS-CoV-2 (12). The oscillations of FOT can be delivered as pulses [impulse oscillometry (IOS)] or sinusoidal waves (13). Several groups have used FOT to successfully assess bronchodilator responsiveness and asthma control in children and young people (14–18). FOT could be a useful adjunct to spirometry in younger asthmatic children, who are unable to perform acceptable spirometry.

Developments in FOT technology have led to small, portable, and hand-held devices such as airwave oscillometry (AOS), which uses a vibrating mesh to generate a multifrequency sinusoidal pseudorandom noise (PRN) signal. The portability of AOS has benefits over other FOT devices as it can be used for home and clinic-based lung function testing (19). Importantly, more successful AOS than spirometry measurements were made in 5–17-year-old children with Down syndrome, likely due to it being a non-volitional test (20). The AOS markers, resistance at 5 Hz (R5), the difference between the resistance at 5 and 20 Hz (R5-20) and the area under the reactance curve (AX), measure peripheral airway disease in asthma (13,21). R5 correlated well with forced expiration volume in 1 s (FEV₁), another measure of airflow obstruction, in asthmatic children (22). In adult asthmatic patients, AOS was more sensitive than spirometry at detecting bronchodilator reversibility (23). Furthermore, changes in IOS results were more sensitive than changes in FEV₁ in diagnosing children with “asthmatic cough” (24). In vivo and in vitro studies have demonstrated differences in resistance (R_rs) and reactance (X_rs) results between AOS and IOS devices (25,26). The differences between AOS and IOS were seen in resistances greater or equal to 10 Hz and reactance in asthmatic children (27). The ERS, however, recently proposed a universal bronchodilator response threshold for all FOT devices of 40% of R5 results and 50% of reactance at 5 Hz (X5) results in children and adults (28). It should be noted that large reference datasets of healthy children have identified lower bronchodilator thresholds of a change of 32% at R5 (29).

The aim of our study was to compare the success of spirometry and AOS in assessing bronchodilator responsiveness in children with asthma or wheeze. Our secondary aims were to assess the correlation of spirometry and AOS results assessing airway obstruction and to review the ERS proposed bronchodilator response thresholds for the AOS device.

Methods

AOS assessments were conducted alongside routine lung function testing at the Amanda Smith Pulmonary Function Laboratory, King’s College Hospital NHS Foundation Trust (KCH), London, UK. Patients attending clinic aged between 5 and 16 years old, presenting with wheeze or asthma, were assessed. “Asthma” was defined as a clinical diagnosis of asthma before referral. “Wheezing” was defined as individuals with wheeze, who had not been given a clinical diagnosis at the time of referral. They were asked to withhold inhalers on the day of lung function testing until the assessments had taken place. Patients performed AOS, followed by spirometry and were then given 400 µg salbutamol via a spacer. Both tests were repeated in the same order 15 minutes after the bronchodilator challenge. AOS was performed before spirometry to avoid the potential impact of forced respiratory maneuvers on AOS results. Individuals who could not perform acceptable baseline spirometry did not undergo bronchodilator challenge. This study was approved as a quality improvement audit by KCH’s Pediatric Department (ref: CH119).

Airwave oscillometry

The Thorasys Tremoflo-C100 Airwave Oscillometry System™ (Thorasys, Montreal, QC, Canada) was used with Thorasys software. The device has a vibrating mesh to generate a multifrequency composite sinusoidal waveform from 5 to 37 Hz. Resistance (R5) and reactance (X5) at 5 Hz, the difference of resistance at 5 and 20 Hz (R5-20) and the area under the reactance curve (AX) were recorded. Verification of the device’s results was performed daily using a standard 15 cm H₂0 calibration load.

Patients were assessed in a seated upright position with a nose clip in place and supporting their cheeks with both hands. Children who could not tolerate the nose clip had their nose held by their accompanying parent/guardian. The participants breathed through a standard mouthpiece filter. A custom made 30-s template of pseudorandom noise was used, which is longer than the standard 20 s minimum duration of data acquisition (29). Testing was repeated until a minimum of three tests were obtained with coherence of variance (CoV) equal or less than 15% at R5, (29) up to a maximum of five trials, to be considered a “successful” result. The results were then expressed as z-scores using reference equations from a healthy population of children and young people which took into account the individual’s height, sex, and age (30,31). A positive bronchodilator response was defined as a decrease in R5 of 40% or an increase in X5 of 50% (28).

Spirometry

Spirometry was performed using the MasterScreen (MS-PFT PRO) spirometer and analyzed with SentrySuite software. Testing was performed with the participant in an upright seated position, with nose clip in place. Patients were coached to inhale to total lung capacity and then to undertake a forced, complete expiration. Verbal and visual prompts were used. Spirometry curves were reviewed by a pediatric respiratory physiologist and considered “successful” if three acceptable forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV₁) maneuvers were achieved as per the American Thoracic Society (ATS) and ERS guidelines (32). A maximum of 10 trials were attempted. Calibration of the device was performed twice daily using a 3 L syringe. The results were converted to z-scores using an appropriate age, sex, height, and ethnic appropriate reference range (33–35). FEV₁, the ratio of FEV₁ to FVC (FEV₁/FVC), and the forced expiratory flow at 75% of the vital capacity (FEF₇₅) results were recorded. A positive bronchodilator response was defined as an increase in FEV₁ of equal or greater than 12% (36).

Sample size

A previous study comparing lung function measurements in 7-year-old children found that 66.7% and 92.5% were able to perform acceptable spirometry and oscillometry testing, respectively (37). To detect such a difference (25.8%) in successful measurements with 90% power at the 5% significance level, a sample size of 47 patients was required (38).

Analysis

The data were tested for normality using the Shapiro–Wilk test and found not to be normally distributed. The chi-squared test was used to assess whether the proportions of successful measurements with spirometry or AOS differed significantly. The ages of children who could and could not perform spirometry were assessed for statistical significance using the Mann–Whitney U test. The agreement between R5 and FEV₁ was assessed using a Bland/Altman analysis (39). The data in the Bland Altman plot were analyzed by linear regression. Pearson’s two-tailed correlation coefficients were calculated to assess the strengths of correlations between spirometry and AOS results. Statistical analysis was conducted via SPSS software Version 28.

Results

Forty-seven children completed lung function testing (Table 1). Most patients presented with asthma (74%) and the remainder with wheeze (26%). Eighty-nine percent of the children were currently being treated with an inhaler: salbutamol (n = 41), a long-acting beta agonist (n = 42), and/or a corticosteroid (n = 39). In the past year, 16 of the children (34%) had received oral corticosteroids. A greater proportion of acceptable results were obtained with AOS (98%, 46/47) than with spirometry (68%, 32/47), p < 0.001. The median age of children who failed spirometry was lower (7.35, range: 5.4–10.3 years) than those who could perform spirometry (10.4, range: 5.5–16.9 years), p < 0.001. One patient was unable to perform either test.

Table 1. Demographics of study population.

Participants, n (male)	47 (26)
Age (years)	8.8 (5.4–16.9)
Height (cm)	134 (105–176)
Weight (kg)	32.3 (14.7–81.6)
Number < 12 years	36 (77%)
Number ≥ 12 years	11 (23%)

Data are presented as n (%) or median (range).

There was a significant regression slope for the z-scores of FEV₁ and R5 (mean Bland–Altman bias: −0.573 z-scores, p = 0.003; Figure 1). Linear regression analysis of the Bland-Altman data demonstrated a relatively good correlation (r = 0.503) and low variance (r² = 0.252), p = 0.003. Pre-bronchodilator R5 results correlated with FEV₁ (p = 0.004, r = 0.488), FEV₁/FVC (p < 0.001, r = 0.618), and FEF₇₅ (p < 0.001, r = 0.617) (Table 2). Baseline R5-20 correlated significantly with FEV₁ (p = 0.015, r = 0.419), FEV₁/FVC (p < 0.001, r = 0.568), and FEF₇₅ (p = 0.002, r = 0.529). Pre-bronchodilator AX results correlated with FEV₁ (p < 0.001, r = 0.68), FEV₁/FVC (p = 0.007, r = 0.574), and FEF₇₅ (p = 0.003, r = 0.618).

Figure 1. Bland-Altman plot for FEV1 z-scores paired with R5 z-scores, Regression equation: y = −0.73 + 1.6x.

Table 2. Correlation of AOS parameters and spirometry parameters at baseline.

Pre-bronchodilator AOS versus spirometry	R value	P value
R5 versus FEV1	0.488	0.004
R5 versus FEV1/FVC	0.618	<0.001
R5 versus FEF75	0.617	<0.001
R5-20* versus FEV1	0.419	0.015
R5-20* versus FEV1/FVC	0.568	<0.001
R5-20* versus FEF75	0.529	0.002
AX versus FEV1	0.68	<0.001
AX versus FEV1/FVC	0.574	0.007
AX versus FEF75	0.618	0.003

All variables are z-scores unless stated otherwise.

* Raw values used instead of z-scores.

Eight children (17%) had a positive bronchodilator response as determined by the results of spirometry. They had a median age of 8.75 years (6.8–16.9 years). The eight children had median changes in R5 of −36% (range: −30% to −50%) and in X5 of 39% (15%–54%). Only three children (6.4%) had a positive bronchodilator response as assessed by AOS using the ERS definition. All three children had positive bronchodilator responses as assessed by spirometry. The bronchodilator ΔR5 correlated with ΔFEV₁ (p < 0.001, r = −0.574), ΔFEV₁/FVC (p < 0.001, r = 0.–0.577), and ΔFEF₇₅ (p = 0.015, r = −0.424). The bronchodilator ΔAX correlated with ΔFEV₁ (p = 0.003, r = −0.512), ΔFEV₁/FVC (p = 0.005, r = 0.–0.487), and ΔFEF₇₅ (p = 0.034, r = −0.376; Table 3).

Table 3. Correlation of delta change in AOS and delta change in spirometry results.

ΔAOS versus Δspirometry measures	R value	P value
ΔR5 versus ΔFEV1	−0.574	<0.001
ΔR5 versus ΔFEV1/FVC	−0.577	<0.001
ΔR5 versus ΔFEF75	−0.424	0.015
ΔR5-20 versus ΔFEV1	−0.178	0.331
ΔR5-20 versus ΔFEV1/FVC	−0.391	0.027
ΔR5-20 versus ΔFEF75	0.018	0.924
ΔAX versus ΔFEV1	−0.512	0.003
ΔAX versus ΔFEV1/FVC	−0.487	0.005
ΔAX versus ΔFEF75	−0.376	0.034

Discussion

We have demonstrated that more children had acceptable AOS than spirometry results. The children who failed spirometry were significantly younger than those who produced acceptable results. AOS parameters of peripheral airway obstruction (R5, R5-20 and AX) correlated significantly with spirometry parameters of obstructive airflow disease (FEV₁, FEF₇₅, and FEV₁/FVC) at baseline, further suggesting AOS assessments are useful in asthmatic children.

Our finding that oscillometry gave more technically acceptable results than spirometry is consistent with other studies in which oscillometry and spirometry were compared in children with Down syndrome (20,40) and those born prematurely (37,41,42). In this study, the children unable to perform spirometry were younger than those who could produce acceptable results. AOS perhaps then may be most useful in the youngest children. Indeed, oscillometry, has been used as a tool to investigate preschool wheeze (14,15,43). Technically acceptable AOS results were defined as three results with a coherence of variance (CoV) equal to or less than 15% at R5 (30). Cottee et al. have suggested a quality grading system (A–F) for oscillometry, with an “A” grade awarded to at least three trials with CoV equal or less than 10% at R5 (44), which is stricter than the criteria we applied. A grading system would be similar to the spirometry standard, in which an “A” grade is awarded to three acceptable trials within 0.15 L for people over 6 years old and 0.1 L for people 6 years old and younger (32).

Assessments of peripheral airway resistance, R5 and R5-20, correlated significantly with spirometry parameters of airflow obstruction, FEV₁, FEV₁/FVC, and FEF₇₅, at baseline. The Bland-Altman regression slope of z-scores of FEV₁ and R5 was significant, with a mean bias of −0.573 z-scores. Linear regression analysis of the Bland-Altman data showed a significant slope with greater discrepancy at more extreme values suggesting the degree of airway obstruction might affect the comparability of the results of the two techniques. Oscillometry, therefore, might underestimate severe airway disease and this should be considered when defining the upper and lower limits of “normal” for AOS. The bronchodilator changes in spirometry parameters most strongly correlated with ΔR5, rather than ΔR5-20. In peripheral airway obstruction, resistance is highest at low oscillation frequencies of 5 Hz (R5) and falls with increasing frequency; therefore, elevated R5 reflects peripheral obstructive airway disease (13). Computational modeling has supported R5-20 as a marker of small airway dysfunction in asthma (21). In a study of three to 6-year old asthmatic children, FEV₁ more strongly correlated with R5 (p = 0.006, r = −0.51) than R20 (p = 0.015, r = −0.46) (22). In that study, the authors reported that R5 had good between-test variability (0.92 for baseline-placebo) and within-test repeatability (4.1%). Further, in a study of prematurely born children who had bronchopulmonary dysplasia when assessed at 6–9 years old, FEV₁ better correlated with R5 and R10 (p < 0.002, r = −0.43–079) than values of reactance (45). We found the strongest correlation of baseline R5 with FEF₇₅ (R = 0.617, p < 0.001) and FEV₁/FVC (R = 0.618, p < 0.001). FEV₁/FVC has previously been shown to be more sensitive marker than FEV₁ in the detection of mild asthma in children (46). FEF₇₅ is a better marker of smaller peripheral airway caliber than FEV₁, which more assesses the larger central airway (47). A retrospective analysis of over 2000 spirometry results performed by asthmatic children aged 6–18 years found that FEF₇₅ was more sensitive than FEV₁ in detecting mild (33% versus 6.8%, p < 0.0001) and severe (71% versus 14.8%, p = 0.0001) airway obstruction (48).

We also observed significant correlations at baseline and following the bronchodilator challenge between spirometry measures of airway obstruction and the reactance measure of AX, a summative marker of peripheral airway obstruction, (13). The Pediatric Asthma Controller Trial (PACT) used IOS and spirometry in a double-blind randomized trial of three treatments for asthmatic children aged between 6 and 14 years (8). They found the changes in AX demonstrated continued improvement over a prolonged period. The improvements in FEV₁ and FEF_25–75 were limited to the first 12 weeks of therapy, whereas improvements in AX were seen for a further 36 weeks in the group receiving fluticasone. Subjects more likely to respond with improvements in AX had a higher level of peripheral blood eosinophils, a marker of severe asthma (49), which supports the idea that AX may be more sensitive to peripheral airway dysfunction than spirometry in asthmatic patients. Recently, a group showed that abnormality in the reactance parameters of X5 and AX correlated with poor asthma control in adults (p < 0.001 and p < 0.001, respectively) (50). This suggests that both resistance and reactance markers might be useful in detecting peripheral airway obstruction in asthmatic patients.

Our final aim was to investigate the bronchodilator response assessed by these two devices. Significant correlations were found between the post-bronchodilator changes in ΔR5 with ΔFEV₁ (p < 0.001, r = −0.574), ΔFEV₁/FVC (p < 0.001, r = 0.–0.577), and ΔFEF₇₅ (p = 0.015, r = −0.424) Short et al. compared the magnitude of beta blocker-induced bronchoconstriction and bronchodilation as assessed by IOS and spirometry in a randomized placebo-controlled crossover study of asthmatic adults (51). All the IOS changes were of larger magnitude than the changes in spirometry, the greatest changes were in R5-20 and AX. This suggests that oscillometry may be more sensitive than spirometry in detecting reversible airway obstruction.

A positive bronchodilator response was defined as an increase in FEV₁ of equal or greater than 12% (36). In our study, only eight children (17%) met this criteria with corresponding changes in R5 of −36% (−30% to −50%) and X5 of 39% (15%–54%). These are lower than the threshold values suggested by the ERS, of at least −40% in R5 and 50% in X5, which were based on data sets from healthy children (30,52–56). In the largest dataset (n = 508) of children aged between 2 and 13 years old, however, the cut off for the lowest frequency of resistance (R6) was −32%, and closer to our findings (30). Furthermore, two other studies used cutoff values for change in R5 of 29% (53) and 37% (54). In older children and young people, the highest threshold for adults was 32% (57). It should be noted that those studies investigated IOS, rather than AOS.

Previous work has demonstrated lower resistance values and more negative reactance values measured by AOS in comparison to IOS devices (25). Dandurand et al. assessed AOS and IOS devices using mechanical loads of known impedance (58). They found large variation between devices when higher impedance loads were used, which was greater with regards to reactance. Lundblad et al. performed a similar study using a reference respiratory phantom, healthy individuals and patients with COPD (59). They showed systematic differences between devices in vivo which were especially marked for measures of AX, but were significantly lower with IOS than AOS devices (p < 0.001). Testing using the respiratory phantom found close correlation between AOS results and test loads, whilst IOS results deviated significantly at higher loads (p < 0.0028). The authors suggested that varying calibration procedures might account for those differences. In view of their results, we suggest that device- and age-specific bronchodilator thresholds might be beneficial for deciding bronchodilator response assessed by oscillometry.

Our findings have clinical implications in that they suggest oscillometry could be used as a lung function test for children unable to perform acceptable spirometry. Schulze et al. conducted a post-hoc analysis of a pediatric asthma study with a 1-year follow up. In that study, using an area under the curve (AUC) analysis, R5 (AUC 0.8, p < 0.001) was superior to FEV₁ (AUC 0.62) in predicting adverse events in asthma patients, defined as the occurrence of asthma symptoms and the use of salbutamol (60). In another study, adults were classified to have mild, moderate persistent and severe persistent asthma according to NHLBI guidelines (61) and their lung function was then assessed using both oscillometry and spirometry (62). There were significant differences in the percentage predicted R5 (108.1% versus 130.1% versus 171.5% p < 0.05) and R5-R20 (1.2 versus 2.4 versus 3.3 hPa/l/s, p < 0.05) in the mild, moderate, and severe groups, respectively. This suggests that oscillometry can also be used to differentiate between clinically defined asthma severity.

Our study benefits from including a range of younger and older children. It was conducted in a real-world lung function clinic and so not subject to the possible biases of a research study. One of the limitations was that children who were unable to perform acceptable baseline spirometry did not go on to have bronchodilator challenge because of clinic time constraints. Therefore, children with positive bronchodilator responses as assessed by AOS might have been missed. Another limitation is that the reference datasets of Polish (31), Australian and Italian (30) children might not adequately represent the multi-ethnic population of children who were included in this report. Previous studies have also used other FOT device-based reference ranges (20) in accordance with ERS guidance (29). Unfortunately, until multi-ethnic reference datasets are available as they are for spirometry (33), little can be done to account for this shortcoming. Children with wheeze (probable asthma) and physician-diagnosed asthma were included in this report. Although this would give a more comprehensive dataset of AOS in healthy and disease states, it perhaps reduced the power of our study in detecting the bronchodilator response threshold for the AOS device, as only eight patients had positive responses. Our conclusions regarding the comparison of spirometry and AOS in assessing bronchodilator responsiveness are limited by the low proportion of children (17%) who had a positive bronchodilator response. That low rate could be due to the influence of the treatments the children were receiving, but they were asked withhold their inhalers on the morning of testing.

In conclusion, we have shown that oscillometry could be a useful adjunct to spirometry in lung function testing of asthmatic children. The benefits were most evident in younger children who were unable to perform spirometry. Future work should aim to correlate the bronchodilator response as assessed by AOS with the clinical severity of asthma. We also suggest that device- and age-specific bronchodilation thresholds may be useful when using oscillometry.

Authors’ contributions

SG collected the AOS data, analyzed the data and drafted the manuscript. MT collected the spirometry data. LW was involved in conceptualization of the project. JC was involved in conceptualization of the project. CH conceptualized the project and analyzed the data. AG conceptualized the project and edited the manuscript. All authors approved the final manuscript.

Ethics

This study was approved as a quality improvement audit by KCH’s Pediatric Department (ref: CH119).

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Additional information

Funding

This research was supported by the National Institute for Health Research (NIHR) Biomedical Research Center at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.