One Equation to Rule Them All: New Meta-Model Standardizes Children’s Lung Sound-Wave Tests

For more than half a century, doctors have known that the lungs can be probed with sound waves. The forced oscillation technique, introduced by A. B. DuBois in 1956, superimposes gentle pressure oscillations on a patient’s tidal breathing and reads back a single number: respiratory impedance. That number splits into resistance, the frictional energy lost as air rushes through the tracheobronchial tree, and reactance, the energy stored in lung tissue and the elastic recoil of the chest. Because the method requires no forced exhalation manoeuvres, it is ideal for young children who cannot perform conventional spirometry reliably. Yet despite its elegance, oscillometry has never achieved the clinical ubiquity its advocates hoped for, and the reason is deceptively simple: nobody agrees on what normal looks like.

Now a team of biomedical engineers at Adelaide University, working with collaborators in Bangladesh and Finland, has taken a strikingly unconventional route to solving that problem. Anuva Chowdhury, Md. Mokammel Haque, and Mathias Baumert report in Annals of Biomedical Engineering that they synthesized 70 virtual children and adolescents, aged 2 to 19 years, from published growth charts and then distilled 25 separate paediatric reference equations, reported since 2000, into two unified meta-model equations for respiratory resistance and reactance at 5 Hz. The work, published as an open-access article on 18 October 2026, offers what the authors describe as a standardized framework for paediatric respiratory impedance, and it arrives at a moment when the European Respiratory Society has explicitly flagged the absence of globally validated paediatric reference values as the critical barrier to broader clinical adoption of oscillometry.

The physics behind the measurement explains why the 5 Hz frequency matters so much. At low oscillation frequencies, the pressure signal penetrates deep into the peripheral lung regions and distal airways. Resistance at 5 Hz, denoted R5, is therefore sensitive to alterations in both central and peripheral airway mechanics and rises when distal airways narrow, as in asthma or other obstructive conditions. Reactance at 5 Hz, denoted X5, is the difference between inertive reactance, which grows with frequency and airway inertance, and compliant reactance, which shrinks with frequency and reflects respiratory compliance. By convention, compliant reactance is negative and inertive reactance is positive, so X5 is negative at low frequencies and becomes positive at high ones. Conditions that reduce respiratory system compliance, such as fibrosis, interstitial disease, or hyperinflation, drive X5 increasingly negative, making it a sensitive marker of small airway dysfunction.

To build their meta-models, the researchers first systematically reviewed 25 cross-sectional studies spanning Europe, North America, Australia, Asia, Africa, the Middle East, and South America, encompassing paediatric populations aged 2 to 19 years. Healthy participants were defined with strict criteria: not born prematurely before 36 weeks, no oxygen supplementation at birth, no history of chronic respiratory diseases such as asthma or allergies, no recent respiratory infections within two to three weeks of testing, and minimal exposure to environmental risk factors like parental tobacco smoke and elevated air pollution. The team then simulated all published equations in MATLAB R2024b, using representative anthropometric combinations drawn from normal paediatric growth trajectories reported by Stanford Medicine Children’s Health, with heights ranging from 90 to 187 centimetres for boys and 86 to 173 centimetres for girls.

The comparative analysis revealed a remarkable degree of underlying consensus despite methodological heterogeneity. All resistance equations showed strong mutual agreement, with Pearson correlation coefficients no lower than 0.95, and all reactance equations exceeded 0.92. When the team selected the study with the largest population, a 2022 reference equation for healthy Chinese children aged 4 to 17 years, as the benchmark and applied a plus-or-minus 30 percent tolerance interval, almost all published equations remained within acceptable bounds across the evaluated height range of 90 to 150 centimetres. Predicted resistance values fell from approximately 1.0 to 1.4 kilopascals per litre per second at 90 centimetres to roughly 0.1 to 0.80 at 150 centimetres, while reactance shifted from approximately negative 0.52 to negative 0.28 at 90 centimetres to between negative 0.20 and positive 0.21 at 150 centimetres. A handful of equations, including those from Iranian and Australian cohorts, showed noticeable discrepancies, with root-mean-square errors exceeding 0.3 for resistance, but the overall picture was one of broad comparability.

The physiological trends were equally consistent. Resistance decreased as age, height, and weight increased, reflecting lung development during growth: as children mature, their lungs expand, bronchi and bronchioles widen, and overall airway resistance falls. Reactance became progressively less negative with growth, mirroring thoracic and alveolar maturation. During infancy, the chest wall is highly compliant, with chest wall compliance reported to be nearly three times greater than lung compliance, partly attributable to more horizontally oriented ribs and reduced thoracic expansion during rib motion. Most geometric and biomechanical maturation occurs within the first two years of life, after which the mechanical relationship between chest wall and lungs approaches that observed in adulthood, with nearly equal compliance. These developmental changes explain why negative reactance values diminish with age.

Sex-related differences emerged as a more nuanced story. Several studies estimated higher resistance values for boys than for girls, while others reported the opposite, and most studies that identified sex-based differences involved teenagers rather than young children. The airway-parenchymal growth relationship appears to differ between sexes: in boys, parenchymal growth is the primary determinant of airway growth, whereas in girls, additional factors, possibly genetic, may influence airway size beyond parenchymal development. No sex-related differences in airway dimensions are evident between ages 1 and 12, but starting at 14 years and older, boys display larger airway cross-sectional areas than girls even after adjusting for height. Girls tend to show superior lung function before puberty, whereas lung function becomes higher in boys after puberty, likely because greater lean body mass in males enhances ventilatory capacity.

With the comparative landscape mapped, the team turned to statistical modelling. Pearson correlation analysis on the synthetic cohort revealed that height, age, and weight were all significantly associated with both resistance and reactance, but variance inflation factor analysis confirmed severe multicollinearity among these growth variables, with values of 30.785 for age, 23.247 for height, and 17.382 for weight. Sex had a low variance inflation factor of 1.105, suggesting independence from other variables. Age showed the strongest negative correlation with resistance at negative 0.88, explaining 78 percent of its variance, while height showed the strongest positive correlation with reactance at 0.80, explaining 65 percent of its variance. Given the multicollinearity, stepwise multivariable regression was employed to identify the best-fitting models from linear, logarithmic, and exponential candidates. The logarithmic equations for both resistance and reactance yielded the highest coefficients of determination, approximately 0.84 for both indices at 5 Hz.

The resulting meta-model equations are strikingly compact. Resistance at 5 Hz equals 6.8545 minus 1.2871 times the natural logarithm of height in centimetres, plus 0.15443 times a sex term where 1 denotes boys and 0 denotes girls. Reactance at 5 Hz equals negative 4.6684 plus 0.9602 times the natural logarithm of height, minus 0.0059 times weight in kilograms. The inclusion of sex as a predictor significantly improved the model fit for resistance, indicating a measurable sex-related effect on airway resistance, whereas no significant association was observed between sex and reactance. Predicted resistance values from the meta-model fall from approximately 1.2 kilopascals per litre per second at 90 centimetres to about 0.6 at 150 centimetres, while predicted reactance rises from negative 0.41 to about negative 0.12 across the same height range, closely following the general growth-related trends reported in previously published paediatric reference equations.

Crucially, the meta-models were evaluated against condition-specific regression equations for probable small airway impairment, small airway impairment, and asthma, restricted to participants aged 5 to 13 years to avoid extrapolation beyond applicable ranges. The proposed equations preserved the progressive increase in resistance from healthy participants through probable small airway impairment, small airway impairment, and asthma groups, and demonstrated the same progressive trend toward increasingly negative reactance values across disease categories. The models also captured bronchodilator-induced changes, with post-bronchodilator resistance distributions showing clear separation from baseline values and post-bronchodilator reactance distributions shifting toward less negative values with minimal overlap, suggesting the framework could identify clinically meaningful variations in airway impairment and treatment responsiveness. The authors acknowledge that their synthetic approach may not fully capture real-world variability in regional growth patterns, environmental exposures, socioeconomic factors, and device calibration, and that external validation in large, multicentre paediatric cohorts remains essential before clinical implementation. To the best of their knowledge, this is the first study to comprehensively synthesize paediatric resistance and reactance reference equations into generalized meta-models for children and adolescents aged 2 to 19 years, and future studies should focus on collecting standardized oscillometric data from geographically and clinically diverse paediatric populations to further refine and validate the proposed models.

Subject of Research: Meta-modelling of paediatric respiratory impedance reference equations for oscillometry in children and adolescents

Article Title: Respiratory Impedance Prediction in Children and Adolescents: A Meta-Modelling Approach

Article References: Chowdhury, A., Haque, M. M., & Baumert, M. (2026). Respiratory Impedance Prediction in Children and Adolescents: A Meta-Modelling Approach. Annals of Biomedical Engineering. https://doi.org/10.1007/s10439-026-04323-2

Image Credits: AI Generated

DOI: 10.1007/s10439-026-04323-2

Keywords: respiratory oscillometry, respiratory impedance, paediatric lung function, forced oscillation technique, meta-modelling, small airway obstruction, asthma, bronchodilator response, reference equations, biomedical engineering, Respiratory, Impedance