Discrepancy between airwave and impulse oscillometry measurements of impedance in healthy children

Introduction: Respiratory oscillometry (OSC) is a pulmonary function test that measures respiratory system impedance and is well-suited for young children. Previous studies have suggested that there may be differences between impedance values obtained on different instruments, however, no studies comparing instruments have been done in healthy children. Methods: We performed both impulse oscillometry (IOS) and airwave oscillometry (AOS), in random order, in a cohort of healthy children with no history of cardiopulmonary disease or recent illness. We used linear regression analysis and Bland Altman plots to compare results. Results: Eighty children ages 4 – 18 years completed study visits. Compared with AOS, the IOS mean and median values for resistance at 5 Hz (R5) were higher (p < 0.001), reactance at 5 Hz (X5) mean and median values were less negative (p < 0.001), and both the area under the reactance curve (AX) and resonant frequency (Fres) were lower (p < 0.001). Bland Altman analysis revealed proportional bias between the two systems. Conclusions: Our results in healthy children show differences between results obtained on different OSC instruments. This may be due to differences in the type of pressure signal produced by the instrument, or, in the case of IOS, may be related to a stress-relaxation response. Our results suggest that normal values obtained on one instrument may not be comparable to results obtained on another. Discrepancy between airwave and impulse oscillometry measurements of impedance in healthy children Heather Boas, MD1, 2 , Katharine Tsukahara, MD3, Joseph McDonough, MS 1, 2 , Tryce Scully, MS4, Cather- ine Qiu, PSM 4, Laurie Travaglini, BS 5, Kristina Ziolkowski, BS 5, Maureen Josephson, DO 1, 2 , Clement L. Ren, MD1, 2 , Sara B. DeMauro, MD 5, 2 , Samuel L. Goldfarb, MD 6, Julian L. Allen, MD 1, 2 1 Division of Pulmonary and Sleep Medicine, Children’s Hospital of Philadelphia, Philadelphia, PA, USA 2 Department of Pediatrics, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA 3 Division of Pediatric Pulmonary and Sleep Medicine, University of Utah, Salt Lake City, UT, USA 4 Biostatistics and Data Management Core, Children’s Hospital of Philadelphia, Philadelphia, PA, USA 5 Division of Neonatology, Children’s Hospital of Philadelphia, Philadelphia, PA, USA 1 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. 6 Division of Pulmonary and Sleep Medicine, Department of Pediatrics, University of Minnesota, Masonic Children’s Hospital, Minneapolis, MN, USA Author emails: HB: [email protected]; KT: [email protected]; JD: mc- [email protected]; TS: [email protected]; CQ: [email protected]; LT: travaglinilau- [email protected]; KZ: [email protected]; CLR: [email protected]; SBD: [email protected]; MJ: [email protected]; SLG: [email protected]; JLA: [email protected] Corresponding Author: Heather Boas, MD Dartmouth-Hitchcock Medical Center One Medical Center Drive Faulkner Building Reception 6M Lebanon, NH 03756 [email protected] (610) 909-3674 (p) Key words: pulmonary function tests, reference values, impedance, pediatric F unding: HB reports grant funding from the Cystic Fibrosis Foundation (BOAS23D0). KT reports grant funding by the National Human Genome Research Institute (T32-HG009495). SBD reports grant funding from the National Institute of Health (UG3HL137872). SLG reports grant funding from the National Heart, Lung, and Blood Institute (NCT04098445). JLA received funding from National Heart, Lung, and Blood Institute (NCT04098445), the Morse Family Research Fund, and the Capek Foundation. The authors have no conflicts of interest to declare. Ethics statement: This study was reviewed by the Children’s Hospital of Philadelphia Institutional Review Board (IRB 20-018357 PERC), and procedures were done in accordance with the Declaration of Helsinki and federal regulations for protection of human subjects (45 CFR 46). Parents or guardians provided written informed consent, and children aged 7 years and older provided assent for study participation. Author contributions statement: HB contributed to investigation, funding acquisition, writing of the original draft, methodology, visualization, writing – reviewing and editing, formal analysis, and project administration. KT contributed to conceptualization, investigation, funding acquisition, and writing – re- viewing and editing. JM contributed to conceptualization, investigation, methodology, writing – reviewing and editing, software, formal analysis, and resources. TS contributed to writing – original draft, validation, visualization, writing – reviewing and editing, formal analysis, and data curation. CQ contributed to vali- dation, writing – review and editing, formal analysis, and data curation. LT contributed to investigation, writing – review and editing, and project administration. KZ contributed to investigation, writing – review- ing and editing, project administration, and resources. MJ contributed to conceptualization and writing – reviewing and editing. CLR contributed to writing – reviewing and editing, supervision, and resources. SBD contributed to conceptualization, funding acquisition, writing – reviewing and editing, supervision, and resources. SLG contributed to conceptualization, funding acquisition, and writing – reviewing and editing. JLA contributed to conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, writing – reviewing and editing, analysis, project administration, and supervision. Preliminary data from this work was initially presented in 2023 at the American Thoracic Society In- ternational Conference during RAPiD poster session “Evolving Concepts in Lung Function Testing and Monitoring.” 2 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary.

Introduction: Respiratory oscillometry (OSC) is a pulmonary function test that measures respiratory system impedance and is well-suited for young children. Previous studies have suggested that there may be differences between impedance values obtained on different instruments, however, no studies comparing instruments have been done in healthy children.

We performed both impulse oscillometry (IOS) and airwave oscillometry (AOS), in random order, in a cohort of healthy children with no history of cardiopulmonary disease or recent illness. We used linear regression analysis and Bland Altman plots to compare results.

Eighty children ages 4 – 18 years completed study visits. Compared with AOS, the IOS mean and median values for resistance at 5 Hz (R5) were higher (p < 0.001), reactance at 5 Hz (X5) mean and median values were less negative (p< 0.001), and both the area under the reactance curve (AX) and resonant frequency (Fres) were lower (p < 0.001). Bland Altman analysis revealed proportional bias between the two systems.

Our results in healthy children show differences between results obtained on different OSC instruments. This may be due to differences in the type of pressure signal produced by the instrument, or, in the case of IOS, may be related to a stress-relaxation response. Our results suggest that normal values obtained on one instrument may not be comparable to results obtained on another.

Objective measurement of lung function is essential in the diagnosis, evaluation, and monitoring of respiratory symptoms in chronic lung diseases of childhood such as asthma, cystic fibrosis, and chronic lung disease following preterm birth. Spirometry is the most commonly used test of lung function in the clinical setting. It can be helpful in discerning obstructive from restrictive pulmonary diseases and in monitoring disease progression. [1] Spirometry can reliably be performed in children as young as 6 years, and with good coaching in children as young as 3 years. [2] However, given its volitional nature requiring a correctly performed forced expiratory maneuver, it can be difficult for young children or those with limited cognitive ability to perform spirometry successfully. Spirometry is also not a sensitive test to detect early lung disease in certain conditions. [3] Therefore, there is a need to adopt additional testing techniques that are easily performed in the clinical setting, can be reliably performed in young children, and are sensitive to early lung disease. Respiratory oscillometry (OSC) measures respiratory system impedance by applying a small, high-frequency pressure wave at the airway opening. [4] The test is performed during quiet breathing and can be reliably performed in children as young as 3 years old. [5-7] In children, OSC is useful in the evaluation of asthma and in the assessment of bronchodilator responsiveness. [4, 5, 8, 9] Studies have demonstrated utility in asthma and cystic fibrosis, especially when performed in conjunction with other tests of lung function, [10, 11] and in children with history of premature birth. [12-15] Several devices are commercially available, and while all oscillometry systems rely on the same basic princi- ples, there are differences between systems and devices, most notably in the pressure waves produced. For example, impulse oscillometry (IOS) produces an impulse pressure wave, what is commonly referred to as a “square wave”, while the pressure wave produced in airwave oscillometry (AOS) is sinusoidal pseudorandom noise. [16] Results of previous studies suggest that there may be differences in results derived from different devices. This has been shown using mechanical test loads [17], airway models [18, 19], and in adults with respiratory disease. [18, 20] Ducharme et al found proportional bias between IOS and AOS systems in chil- dren with asthma. [6] However, the possible differences between these two devices have not been described in healthy children. Moreover, while multiple sets of reference equations have been validated, [16, 21-24] it is not known whether reference equations derived from one device can be applied to another. The objective of this study was to assess the comparability of AOS and IOS in a cohort of healthy children to aid in the determination of normal values. Because of the different types of pressure waves applied from 3 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. each instrument, we hypothesized that results and values obtained from each instrument would differ.

and Materials Approval was obtained by the Children’s Hospital of Philadelphia Institutional Review Board (IRB 20- 018357 PERC). Procedures were done in accordance with the Declaration of Helsinki and federal regulations for protection of human subjects (45 CFR 46). Parents or guardians provided written informed consent, and children aged 7 years and older provided assent for study participation. All study visits were completed at Children’s Hospital of Philadelphia (CHOP) in Philadelphia, PA. Recruitment and eligibility Participants were recruited by flyers distributed at CHOP primary care offices and from letters through the Pediatric Research Consortium (PeRC). Flyers and letters were given to families of children who were identified through PeRC as eligible based on age and health history. Screening for inclusion and exclusion criteria were done by chart review and through a screening questionnaire that was reviewed over the phone. We included males and females aged 4 – 18 years old, with no history of physician diagnosed asthma and a negative asthma screen on modified Global Initiative for Asthma (GINA) Guidelines Questionnaire for Pediatrics (MGGQ-P) questionnaire. [25] We excluded children born prematurely ( < 37 weeks gestational age), had viral symptoms within a month of the study visit, had a previous diagnosis of asthma, personal history of vaping or smoking, or had chronic conditions such as chronic lung disease (e.g., cystic fibrosis, primary ciliary dyskinesia), airway anomaly, history of tracheostomy, neuromuscular disease, congenital heart disease, immunodeficiency, or chest wall deformity. A second pre-visit screen was performed via phone approximately 48 hours prior to all study visits to ensure that participants had not experienced any respiratory illness or symptoms within the four weeks leading up to the study date. Outcomes The primary OSC outcomes included resistance and reactance at 5Hz (R5 and X5, respectively), 10 Hz for IOS or 11 Hz AOS (R10/R11, X10/X11 respectively), resonant frequency (Fres), and area under the reactance curve (AX). Statistical methods We conducted a power analysis to determine our sample size for detecting a small effect (Cohen’s d = 0.3) and determined that a total sample size of 90 was necessary for an estimated 80% power and significance level of 0.05. We conducted univariate analyses across all data. Significance testing for groups of interest were completed including parametric tests comparing AOS vs. IOS data. Pairwise comparisons of AOS and IOS primary outcomes for the study sample were performed using the Wilcoxon signed rank test. We used linear mixed effects models to account for the dependency between observations within individuals. Simple linear re- gression of each oscillometry technique was completed to detail the effect of height on the outcomes (R5, R10/11, X5, X10/11, AX, Fres). A p-value of [?] 0.05 was considered statistically significant. Bland Altman analysis was performed to describe agreement between AOS and IOS measurements. The statistical limits of agreement were calculated using the mean and the standard deviations of the differences between two measurements.

A total of 189 children were screened for participation. Forty-two were ineligible; of the 147 children who met eligibility criteria, 67 chose not to enroll and 80 successfully completed study visits (Figure 1). Participants’ characteristics are summarized in Table 1. There were 38 females (47.5 %) and 42 males (52.5%), with ages ranging from 4 – 18 years (mean age 9.8 years). The majority of participants (70.0%) were White, with 21.3% Black or African American, and 12.5% Asian, with some identifying as multiple races. Females tended 4 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. to be smaller than males, with slightly lower median height and weight, but slightly higher BMI. Respiratory rates and pulse oximetry were normal for all subjects and were similar between females and males. Significant differences in outcomes were seen between AOS and IOS (Table 2). Compared with AOS, the IOS mean and median values for R5 were higher (p < 0.001), X5 mean and median values were less negative (p < 0.001), and both AX and Fres were lower (p < 0.001). Linear regression analysis (Figure 2) also demonstrated that IOS had higher R5, less negative X5, and lower AX and Fres as compared to AOS. Furthermore, there were greater discrepancies between the results at smaller heights, most notable in X5, AX, and Fres (Figure 2). Bland Altman analysis (Figure 3) demonstrated a mean bias between instruments for all outcomes as well as proportional bias between the instruments, with shorter/younger subjects showing greater discrepancy between the two instruments.

In this study of healthy children performing both AOS and IOS, we found that there were significant differences in the results obtained between the devices for R5, X5, AX, and Fres. Both commercial devices also showed a decrease in Fres with increasing height, similar to results shown in prior studies. However, as compared to IOS, the AOS system yielded lower values for R5 and more negative values for X5, with higher values for both AX and Fres. We also found age/height dependent differences and proportional bias between the two systems, with greater differences seen between the systems in smaller children. Previous studies also found differences in results obtained on different OSC devices. Using mechanical test loads, Dandurand, et al found substantial differences in impedance measurements using five different OSC devices, including IOS and AOS systems. Others have specifically looked at differences between IOS and AOS and found both differences in the results as well as proportional bias between the instruments. [18, 19] Interestingly, Soares et al found in their study population of adults, including healthy controls as well as those with asthma and history of smoking, that resistance was higher as measured by IOS (Jaeger MasterScope CT IOS) compared to AOS (TremoFlo C-100), and that IOS reactance was more negative as compared to AOS. [18] Similarly, Tamimura, et al used phantom models to compare IOS and MasterScreen, and also found higher resistance in the IOS devices. [19] These findings differ from our results in healthy children, where we found the opposite to be true. However, our data agree with those of Ducharme, et al, who performed IOS and AOS in a cohort of 50 children aged 3 – 17 years with asthma, and showed resistance and Fres to be higher and reactance to be more negative using AOS, also using TremoFlo and Jaeger devices. They also found proportional bias between the devices for both resistance and reactance. [6] It is possible that these findings in children with asthma and our findings in healthy children differ from the results of adults and models due to differences in respiratory physiology in childhood. Notably, younger children with smaller airways have higher respiratory system resistance and elastance as compared to adults. This is reflected in normal OSC values with decreasing resistance with height, increasing (less negative) reactance, and decreasing Fres. [16] We also found greater discrepancies between the two instruments in younger children, particularly in Fres and X5 (Figure 2). These data suggest that normal or predicted values derived on one instrument may not be suitable to use as

values for another. The reason for the different impedance results obtained by different devices is unclear, though it may be related to the different types of pressure waveforms generated by each instrument. All OSC devices utilize pressure waves at different harmonic frequencies to measure the changes in respiratory impedance at different frequencies. The IOS system uses a loudspeaker to create an “impulse” wave, though it more closely resembles a “Dirac Delta” function. [26] These impulses are delivered every 0.2 seconds, and the amplitudes of the harmonics are not controlled by the device. On the other hand, TremoFlo, which is an AOS system, produces a sinusoidal pseudorandom noise waveform, which is a pre-recorded signal that encompasses a range of frequencies that are continuously delivered, with the amplitude of the harmonics pre-determined. [Lennart Lundblad, personal communication] It is quite possible that these differences in the types of waveforms and harmonics generated by each device lead to systemic inter-device differences in 5 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. results. We speculate that a possible explanation for the lower Fres and AX, and less negative X5 seen in IOS compared to a pseudorandom noise AOS system could be stress-relaxation of respiratory tissues of the pulmonary parenchyma and/or the chest wall. This time-dependent phenomenon occurs in visco-elastic tissue and has been described in both human and animal models. [27] It occurs when an abrupt step change in volume causes an abrupt rise in pressure, followed by a slow decrease to a pressure plateau (Figure 4). When a viscous element is in series with an elastic element, e.g. a spring, the sudden distension of that spring can be lessened as the viscous element slowly expands. It is reasonable to surmise that the impulse waveform produced by IOS may induce stress-relaxation in the respiratory system, which in turn could impact impedance measurements. While the differences seen between instruments were modest, they were statistically significant and of po- tential clinical significance. For example the differences seen between AOS and IOS in our younger subjects were of similar magnitude to differences in AX and Fres seen after bronchodilator administration in a study of normal and asthmatic preschoolers. [28] Such differences should be considered when comparing OSC

obtained on different instruments. This is the first study to examine differences between OSC devices in a cohort of healthy children. One

of this study is that it is a relatively small, single center study, with all participants being local to the Philadelphia region. Differences in race and ethnicity compared to other populations, however, have not been shown to impact results. [16] In addition, by using each subject as their own control, such differences would not explain any inter-instrumental differences. Strengths of the study include that it was sufficiently powered to detect differences between the instruments. We also performed extensive screening via chart review and questionnaires, as well as confirmed absence of upper respiratory symptoms in the weeks leading up to the visit, so we are confident that we captured a healthy population of children. In conclusion, in this study of healthy children performing both AOS and IOS, we found differences in impedance measurements between the two devices. We demonstrated proportional bias between the instru- ments, most notably in Fres and AX, with greater differences in Fres, AX, and X5 in younger children. Our data agree with some previous studies that demonstrate inter-device variation. While the reasons for these differences are not well understood, we propose that different input waveforms may have different effects on mechanical outputs such as stress relaxation. It is important to recognize these inter-instrument differences when applying reference equations and predicted values or comparing an individual’s results obtained on different devices, as predicted values may not be interchangeable between instruments. T able 1: Baseline characteristics of the study sample Overall Study Sample, mean (SD) (n=80) Male, mean (SD) (n = 42) F emale, mean (SD) (n = 38) Characteristic, mean (SD) Age (years) 6 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. 9.8 (3.6) 9.6 (3.3) 9.9 (4.0) Height (cm) 140.0 (18.8) 141.0 (19.4) 139.0 (18.4) Weight (kg) 39.9 (18.2) 40.3 (21.4) 39.5 (14.0) BMI 19.4 (5.0) 19.1 (5.5) 19.8 (4.4) Heart Rate (beats/min) 88.8 (12.8) 87.8 (11.5) 89.9 (14.2) Respiratory Rate (breaths/min) 20.1 (2.4) 20.3 (2.4) 19.9 (2.4) Race and Ethnicity , n (%)* Asian 10 (12.5) 8 (19.0) 2 (5.3) Black 17 (21.3) 6 (14.3) 11 (28.9) White 56 (70.0) 7 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. 29 (69.0) 27 (71.1) Not Reported 1 (1.3) 0 (0) 1 (2.6) Hispanic or Latino 5 (6.3) 2 (4.8) 3 (7.9) *Categories were not mutually exclusive. T able 2: Pairwise comparisons of AOS and IOS primary outcomes for the overall study sample (n = 80). AOS IOS p-value R5* Mean (SD) 5.76 (2.00) 6.20 (2.04) < 0.001 Median [Min, Max] 5.52 [2.02, 10.50] 6.01 [2.70, 10.70] R10/11* Mean (SD) 5.52 (1.87) 5.45 (1.61) 0.079 Median [Min, Max] 5.38 [2.04, 9.81] 5.39 [2.39, 8.86] X5* Mean (SD) -2.07 (0.92) -1.84 (0.04) < 0.001 Median [Min, Max] -1.97 [-4.40, -0.42] -1.60 [-3.94, -0.36] X10/11* Mean (SD) -0.87 (0.87) -0.80 (0.81) 0.059 Median [Min, Max] -0.728 [-3.12, 0.47] -0.60 [-2.78, 0.53] AX+ Mean (SD) 17.10 (16.90) 12.10 (11.50) < 0.001 Median [Min, Max] 10.80 [0.61, 73.00] 6.84 [0.26, 45.10] F res++ Mean (SD) 19.10 (7.63) 16.90 (5.96) < 0.001 Median [Min, Max] 17.00 [7.85, 35.00] 16.70 [6.62, 29.70] * cmH2O/L/s + cmH2O/L ++ Hz 8 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Figure 1 : Eligibility and enrollment *Other reasons for exclusion included scoliosis/chest wall deformities, history of congenital heart disease, abnormal chest radiograph, history of vaping or smoking. 9 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Figure 2 : Linear regression analysis for AOS (black) and IOS (gray) vs height. (A) R5, (B) R10/11, (C) X5, (D) X10/11, (F) Fres. 10 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Figure 3 : Bland Altman plots demonstrating proportional bias between instruments. The shaded regions represent the mean bias confidence interval (center) and the upper = and lower 95 % limits of agreement. 11 Posted on 18 Aug 2025 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/au.175554196.62476180/v1 — This is a preprint and has not been peer-reviewed. Data may be preliminary. Figure 4: Stress relaxation (SR). An abrupt step change in volume (V) leads to an abrupt rise in pressure (P) followed by a slow decrease to a pressure plateau. 12