Simple Breathing Test Could Reveal Which Scarring Lung Diseases Are About to Worsen

A quiet breath, a gentle puff of pressure, and a set of numbers that may foretell the future of a failing lung. Researchers in China report that an unusual breathing test called impulse oscillometry, which measures how stiffly the lungs resist and rebound tiny pressure waves during ordinary breathing, appears to flag which patients with fibrotic interstitial lung disease are most likely to deteriorate over the following year. The findings, published in Frontiers in Physiology, come from a cohort of 255 adults tracked at a single hospital in Xuzhou, and they point to two oscillometry measures — reactance at 5 hertz, known as X5, and the reactance area, or AX — as strikingly powerful signals of coming disease progression.

Interstitial lung disease is an umbrella term for more than two hundred disorders in which the delicate tissue between the air sacs becomes inflamed and scarred. As fibrosis accumulates, the lungs stiffen, oxygen transfer falters, and breathlessness deepens. Some patients remain stable for years while others slide rapidly toward respiratory failure, and distinguishing these trajectories early is one of the most urgent unsolved problems in pulmonary medicine. The concept of progressive pulmonary fibrosis was introduced to describe the non-idiopathic pulmonary fibrosis patients whose disease advances despite treatment, and international guidelines now define progression through three domains: worsening respiratory symptoms, declining lung function, and advancing scarring on high-resolution computed tomography.

Because the Xuzhou cohort included both idiopathic pulmonary fibrosis and other fibrotic ILD subtypes, the research team used the term fibrotic disease progression for a progression-like composite outcome adapted from those criteria. Patients were considered progressors if at least two of the three domains worsened within a nine-to-fifteen-month window around the twelve-month follow-up point. Two interstitial lung disease specialists who were blinded to the baseline oscillometry results adjudicated each outcome, and a thoracic radiologist blinded to all physiological data assessed the CT scans. Independent blinding of outcome assessors is a methodological detail that matters enormously in studies like this one, because expectations can otherwise quietly shape how ambiguous changes are judged.

The study was ambispective, meaning it combined retrospectively assembled routine clinical data with prospective follow-up after formal enrollment. Of 348 screened patients, 93 were excluded for reasons including airway-dominant diseases such as COPD and asthma, acute severe illness, malignancy, missing records, or absence of definite fibrosis on baseline CT. The final 255 patients underwent spirometry, diffusing capacity testing and impulse oscillometry at baseline. The technique itself is elegantly simple: while the patient breathes tidally through a mouthpiece, a loudspeaker superimposes small mult-frequency pressure pulses, and the device separates how much of each wave the respiratory system resists, captured as resistance measures R5 and R20, from how elastically it stores and returns energy, captured as reactance X5, resonant frequency Fres and the reactance area AX.

Over the follow-up window, 41 of the 255 patients — 16.1 percent — met the criteria for fibrotic disease progression. At baseline, the progressors already looked mechanically different: their X5 values were more negative, their AX values larger, and their resonant frequencies higher than those of patients who remained stable. By contrast, differences in R20, the resistance measured in larger central airways, were negligible, hinting that the informative signal lives in the lung periphery rather than the conducting airways. When the team modeled each parameter continuously using Firth penalized logistic regression — a method chosen because the number of events was limited and X5 showed near-complete statistical separation — a one-standard-deviation worsening in X5 carried an adjusted odds ratio of 8.89 for subsequent progression, with a confidence interval of 4.66 to 19.47.

AX, a composite measure integrating abnormal reactance across the breathing frequency spectrum, was associated with progression at an adjusted odds ratio of 1.67 per standard deviation, and resonant frequency at 1.61. The resistance measures R5, R20 and the device-reported mean R5–R20 did not survive adjustment and false-discovery-rate correction. Scaled differently, each 0.01 kilopascal-second-per-liter increase in X5, representing a less negative and therefore healthier reactance, was associated with 33 percent lower odds of progression. Yet the story is not a straight line: restricted cubic spline analysis revealed significant nonlinearity, meaning the relationship between X5 and risk changes shape across its range and cannot be compressed into a single constant effect.

The association survived a series of robustness checks. When the 87 patients with idiopathic pulmonary fibrosis were excluded, leaving 168 patients with 28 progression events, X5 remained strongly associated with progression at an adjusted odds ratio of 8.55, and AX stayed significant at 1.47. A sensitivity analysis restricted to patients assessed within a narrower ten-to-fourteen-month window reproduced the same directional pattern. This resilience matters because the formal progressive pulmonary fibrosis construct applies to non-IPF disease, so the signal is not merely an artifact of the most notoriously aggressive fibrotic subtype.

What makes the result scientifically interesting is what X5 and AX actually represent. Although oscillometry is often described as a small-airway test, in fibrotic lung disease the reactance parameters are influenced not only by the caliber of peripheral airways but also by parenchymal stiffness, elastic recoil and the mechanical load of scarred, heterogeneous tissue. In other words, a more negative X5 may reflect the lung as a whole becoming a less compliant, less elastic organ — a direct mechanical signature of fibrosis tightening its grip. Spirometry, the workhorse of pulmonary function testing, requires forceful maximal efforts and mainly indexes larger airways, while oscillometry requires only quiet tidal breathing, which could eventually make it valuable for patients unable to perform demanding maneuvers.

The predictive performance analyses, though explicitly exploratory, were eye-catching. A clinical base model containing age, sex, height and baseline forced vital capacity percent predicted achieved an optimism-corrected area under the curve of only 0.585 — barely better than a coin flip. Adding AX raised it to 0.703, and adding X5, whether modeled linearly or with a spline, lifted it to roughly 0.92, an incremental gain of about 0.34 with confidence intervals comfortably above zero. Brier scores, which penalize prediction error, improved correspondingly. With the internally derived X5 cutoff, no progression events occurred among patients classified as having normal reactance, a complete-separation pattern the authors interpret cautiously as cohort-specific rather than as proof of zero risk.

That caution is warranted, and the researchers are candid about the limitations. The study was conducted at a single center, progression events numbered only 41, and the receiver-operating-characteristic thresholds derived from the data shifted substantially across 500 bootstrap resamples — X5 cutoffs ranged from minus 0.130 to minus 0.050 kilopascal-second per liter — so no clinically applicable cutoff can yet be recommended. Follow-up CT was not uniformly available and was more often obtained in progressors, potentially biasing ascertainment, and treatment changes during follow-up were not modeled. Healthy controls were substantially younger than the patient cohort and served only as a descriptive reference. The authors accordingly frame their conclusions carefully: continuous oscillometric abnormalities, particularly X5 and AX, were associated with subsequent fibrotic progression and merit evaluation in larger, externally validated cohorts, but these results do not yet establish bedside predictive thresholds. If independent teams reproduce the signal, a test that takes minutes and asks nothing more of the patient than relaxed breathing could become a quiet early-warning system for one of medicine’s most relentless disease processes.

Subject of Research: Impulse oscillometry-derived respiratory mechanics as predictors of fibrotic progression in interstitial lung disease

Article Title: Impulse oscillometry-derived respiratory mechanical abnormalities and fibrotic disease progression in interstitial lung disease: a single-center cohort study

Article References: Zhang, C., Wang, Y., Feng, Z., Sun, Y., Zhang, M., Liu, Y., & Chen, B. (2026). Impulse oscillometry-derived respiratory mechanical abnormalities and fibrotic disease progression in interstitial lung disease: a single-center cohort study. Frontiers in Physiology, 17, Article 1962001. https://doi.org/10.3389/fphys.2026.1962001

Image Credits: AI Generated

DOI: 10.3389/fphys.2026.1962001

Keywords: interstitial lung disease, pulmonary fibrosis, impulse oscillometry, X5 reactance, reactance area, respiratory mechanics, progressive pulmonary fibrosis, spirometry, disease progression, lung function testing, Firth penalized regression, predictive modeling