Analysis of the correlations and inconsistencies between spirometry and impulse oscillometry in the diagnosis of small-airway dysfunction

Abstract

Objective

Currently, there has been no gold standard for diagnosing small airway dysfunction (SAdf). This study aimed to evaluate the correlation between small airway parameters derived from spirometry and oscillometry in hospitalized patients, assessing the potential of oscillometry as an alternative diagnostic tool for SAdf. Additionally, this study explored the inconsistencies and influencing factors related to spirometry and oscillometry in diagnosing SAdf, conducting a preliminary assessment of these factors.

Methods

A retrospective study was conducted involving data collection from patients who underwent both spirometry and oscillometry between June 1, 2022, and September 1, 2023, at Chengdu Third People’s Hospital was conducted. Initially, 1,771 patients were considered, with 1,446 meeting the inclusion and exclusion criteria. The clinical characteristics and correlations between small airway parameters from the two methods were analyzed based on different lung function data groups. Besides, this study explored the inconsistency between the two pulmonary function tests in diagnosing SAdf in hospitalized patients. Multivariate logistic regression was employed to investigate the factors contributing to these inconsistencies.

Results

Significant correlations were identified between parameters (reactance area [AX], resonant frequency [Fres], reactance at 5 Hz [X5], difference between resistance at 5 Hz and R20 [R5-R20]) and the forced expiratory flow (FEF) metrics (FEF25%-75%, FEF50%, and FEF75%). Among these, AX showed the strongest correlation, regardless of the severity of forced vital capacity (FVC), forced expiratory volume in one second (FEV1), and the FEV1/FVC ratio. Diagnostic inconsistencies were influenced by factors such as sex, body mass index (BMI), and sputum production. Females, individuals with a high BMI, and those with less sputum were linked to oscillometry-only SAdf, while males, individuals with alow BMI, and those with more sputum were linked to spirometry-only SAdf.

Conclusion

In hospitalized patients, oscillometry could serve as an effective alternative or complement to spirometry for diagnosing SAdf. A greater degree of lung function impairment was correlated with small airway parameters between the two tests. The oscillometry might detect SAdf more sensitively in patients with normal pulmonary function as measured by spirometry. Ultimately, we recommend the combined use of spirometry and oscillometry in hospitalized patients based on its comprehensive assessment of small airway function and potential in timely intervention.

Background

Small airways are defined as distal airways with a diameter of ≤ 2 mm, originating from the 4th to the 14th generation of the tracheobronchial tree, with the trachea as the 0th generation and the alveolar sac as the 23rd, Typically, these airways are observed around the 8th generation of the airway structure [1]. Small airway dysfunction (SAdf) encompasses pathological changes such as congestion, edema, smooth muscle spasm, and airway remodeling in these small airways due to various noxious stimuli. These changes lead to increased airway resistance and functional abnormalities [2]. Small airway disease (SAD) must be distinguished from small airway dysfunction (SAdf) from the physiological point of view. SAdf refers only to a single functional property of a parameter, while SAD (small airway disease) indicates that conditionally several functional components of airway destruction prove and small airway dysfunction (SAdf) is a specific pathophysiological component of SAD [3]. SAdf is prevalent not only in patients with asthma and chronic obstructive pulmonary disease (COPD) but also in individuals without COPD [4], serving as an early indicator that may progress to chronic respiratory diseases [5]. Due to the anatomical characteristics of small airways, accurate assessment methods are limited to histological examinations and micro-computed tomography (micro-CT), both of which are invasive and not widely applicable in clinical settings [6, 7]. Consequently, there has been no universally accepted gold standard for diagnosing SAdf [8]. Various techniques can evaluate SAdf, including spirometry, oscillometry, body plethysmography, chest CT, and multiple-breath nitrogen washes [9]. Among these methods, spirometry is the most common due to its relative ease of performance and straightforward measurement equipment [10]. However, spirometry has inherent limitations. It can be influenced by high airway resistance and the effort-dependent nature of forced vital capacity (FVC), thereby undermining its reliability and sensitivity in reflecting the abnormalities in small airway function [11] and there are some concerns exist regarding the reliability of parameters for assessing small airways. Moreover, spirometry requires active cooperation for regular, forceful breathing, which may be challenging for children, elderly individuals, or those who are severely debilitated. Thus, alternative tools are needed to effectively assess lung function in patients who cannot undergo spirometry or have contraindications.

Oscillometry is a forced oscillation technique that measures total airway resistance, proximal airway resistance, and peripheral airway resistance [12]. Typically, resistance at 5 Hz (R5) represents total airway resistance, while resistance at 20 Hz (R20) indicates central airway resistance. The difference between R5 and R20 (R5-R20) corresponds to peripheral airway resistance, indicating small airway resistance. Additionally, reactance at 5 Hz (X5), the reactance area (AX), and the resonant frequency (Fres) provide further insights into peripheral airways [13]. Compared to spirometry, the oscillometry requires only calm and natural breathing from the patient. Thus, it is suitable for a wider range of individuals, such as children and extremely frail patients, indicating a broader range of clinical applications.

Previous studies have suggested that the oscillometry can supplement or potentially replace spirometry [14]. However, these studies often involve small sample sizes, and there has been limited research on the correlation between small airway parameters obtained from both spirometry and oscillometry in hospitalized patients. Furthermore, inconsistencies often arise in clinical practice, where spirometry identifies SAdf while the oscillometry does not (spirometry-only SAdf) or where the oscillometry detects SAdf while spirometry does not (oscillometry-only SAdf). This study aimed to examine the correlation between small airway parameters measured by spirometry and oscillometry in hospitalized patients and investigate the factors contributing to these inconsistencies between the two methods in the diagnosis of SAdf.

Subjects

This retrospective observational study included 1,771 patients who underwent both spirometry and oscillometry for acute exacerbation of respiratory symptoms (cough, sputum production, wheezing, and dyspnea) during hospitalization at Chengdu Third People’s Hospital between June 1, 2022 and September 1, 2023. Patients identified by seasoned clinicians as suffering from acute asthma exacerbations, chronic obstructive pulmonary disease (COPD) with acute exacerbations, bronchiectasis, interstitial lung disease (ILD), or pneumonia. The inclusion criteria were as follows; (1) Patients with complete and reliable data for both pulmonary function tests; (2) Patients with both tests on the same day; and (3) Patients with no contraindications for spirometry. The exclusion criteria were; (1) Patients with advanced lung cancer or active pulmonary tuberculosis; (2) Patients with a history of lung resection; and (3) Patients with unreliable pulmonary function data. A total of 1,446 patients met the criteria for analysis. Figure 1 illustrates the inclusion criteria. This study was reviewed and approved by the Ethics Committee of Chengdu Third People’s Hospital (Project No. Ethics Committee of Chengdu Third People’s Hospital [2024]-S-85).

Spirometry-diagnosed SAdf was defined as having at least two out of the three forced expiratory flow (FEF) metrics (FEF25%-75%, FEF50%, and FEF75%) with values less than 65% of the predicted value [15, 16] based on a multicenter study of spirometry reference values in a healthy Chinese population [16]. Among these metrics, FEF25%-75% was the most sensitive indicator. In the oscillometry, the SAdf would be identified when R5-R20 reached up to ≥ 0.07. In fact, the residual volume to total lung capacity (RV/TLC) ratio has been proposed as a potential marker for small airway function [17,18,19,20]. Some studies have used a RV/TLC ratio of > 40% as a criterion for diagnosing SAdf [21, 22]. For Chinese adults, the lower limits of normal have been defined as 80% for forced expiratory volume in one second (FEV1%) and 65% for FEF25%-75% [16]. An FEV1% <80% is indicative of large airway abnormalities, whereas an FEF25%-75% <65% suggests small airway dysfunction. Our study applied clinical diagnostic criteria for analysis (at least two out of the FEF25%-75%, FEF50%, and FEF75% values less than 65% of the predicted value). Therefore, for the correlation analysis, the patients were categorized into three groups on the basis of spirometry results: no-SAdf (normal small airways dysfunction), SAdf (small airway dysfunction without large airway abnormality), and LSAdf (both small and large airway dysfunction); For the inconsistency analysis, patients were grouped as follows: both no-SAdf, spirometry-only SAdf, oscillometry-only SAdf, and both SAdf.

Fig. 1

Enrollment of study population

Methods

Baseline demographic variables, including gender, age, body mass index (BMI), smoking status (never, current, former), smoking index (pack-years), respiratory symptoms (wheezing, cough, sputum, dyspnea), and history of chronic diseases (diabetes, hypertension, hyperlipidemia, heart disease), were collected through a comprehensive review of medical records. All patients underwent spirometry (Jaeger MasterScreen, Germany) and The Impulse Oscillometry System (MS-IOS Jaeger), which were performed by trained operators. Each measurement performed in triplicate. The pre-bronchodilator parameters included FVC percentage (FVC%), FEV1 to FVC ratio (FEV1/FVC), FEV1%, FEF25%-75%, FEF50%, FEF75%, RV/TLC, R5, R20, X5, Fres, R5-R20, and AX. The procedure was performed following the American Thoracic Society and European Respiratory Society recommendations [23, 24].

Statistical analysis

Statistical analyses were performed using SPSS version 26.0 software. Normally distributed continuous data were expressed as means ± standard deviations, whereas non-normally distributed data were expressed as medians and interquartile ranges. For group comparisons, independent samples t-tests or Wilcoxon-Mann-Whitney tests were used for continuous variables, while chi-square tests or Fisher’s exact tests were used for categorical variables. Correlation analyses utilized Spearman correlation coefficients, with the results depicted in scatter plots. Multivariate logistic regression was applied to analyze the factors influencing inconsistencies in the diagnosis of SAD between spirometry and oscillometry. Correlation strength was categorized as follows: 0 < |r| < 0.3 = weak correlation; 0.3 < |r| < 0.7 = moderate correlation; and |r| > 0.7 = strong correlation [25]. A p-value < 0.05 was considered statistically significant.

Results

Demographic and clinical characteristics of patients

A total of 1,446 patients were included, with 352, 319, and 775 patients in the no-SAdf, SAdf, and LSAdf groups, respectively. The demographic characteristics of these three groups are detailed in Table 1. There were significant differences in age, sex, wheezing, hypertension, and heart disease between the no-SAdf group and the SAdf group (p < 0.05). Similarly, the LSAdf group demonstrated significant differences in age, sex, BMI, smoking index, sputum production, wheezing, dyspnea, hypertension, and heart disease compared to the no-SAdf group (p < 0.05). The LSAdf group also exhibited significant differences in sex, BMI, smoking index, sputum production, wheezing, dyspnea, heart disease, and hyperlipidemia compared to the SAdf group (p < 0.05).

Table 1 Comparison of demographic and clinical characteristics among no-SAdf and SAdf and LSAdf groups

Analysis of parameters of spirometry and oscillometry

Table 2 presents significant differences in spirometry parameters (FEV1%, FEV1/FVC, FEF25%-75%, FEF50%, FEF75%, and RV/TLC%) and oscillometry parameters (R5, R5-R20, AX, Fres, and X5) among the three groups (p < 0.001). Pulmonary function decreased, and airway resistance increased progressively from the no-SAdf group to the SAdf group, and further to the LSAdf group. In the no-SAdf group, most patients exhibited normal pulmonary function (FEV1 ≥ 80%, FEF25%-75% ≥ 65%) with no significant airflow limitation (FEV1/FVC > 70%). The SAdf group also generally had an FEV1/FVC > 70%, indicating predominantly normal airflow. However, the LSAdf group exhibited significantly worse pulmonary function and higher airway resistance compared to the SAdf group. Notably, 115 patients in the no-SAdf group were diagnosed with SAdf based on the oscillometry alone. Figure 2 illustrates the variation in airway resistance across the three groups.

Table 2 Lung function features among no-SAdf and SAdf and LSAdf groups

Fig. 2

Airway resistance in the subgroups

Analysis of correlations in small-airway parameters measured by spirometry and oscillometry

Figure 3a depicts the correlations among small-airway parameters derived from spirometry. Significant and strong correlations were observed among FEF25%-75%, FEF50%, and FEF75%, with the highest correlation found between FEF25%-75% and FEF50% (r = 0.977, p < 0.001). Figure 3b illustrates the correlations among small-airway parameters measured by the oscillometry. AX, Fres, and X5 showed strong correlations with R5-R20, with the most significant correlation between AX and R5-R20 (r = 0.959, p < 0.001). Table 3 details the correlations between small-airway parameters from spirometry and oscillometry. Among the 1,446 patients, significant correlations were found between the parameters measured by oscillometry (R5-R20, Fres, AX, X5) and the parameters measured by spirometry (FEF25%-75%, FEF50%, FEF75%, RV/TLC) (p < 0.001). AX, reflecting peripheral airway function, demonstrated the strongest correlations with FEF25%-75%, FEF75%, and FEF50% (r = -0.756, r = -0.752, and r = -0.755, respectively; all p < 0.001). R5-R20 showed moderate correlations with these spirometry parameters, and the weakest correlation (r = -0.696, r = -0.697, r = -0.699, all p < 0.001). Table 4 indicates that in groups with airflow limitation (FEV1/FVC < 70%), restrictive ventilation (FVC% < 80%), and large airway abnormalities (FEV1% < 80%), the small-airway parameters measured by oscillometry had better correlations with the small-airway parameters measured by spirometry compared to the groups without these conditions. Among these parameters, AX exhibited the strongest correlation with small-airway parameters measured by spirometry. Table 5 presents the correlations between small-airway parameters measured by oscillometry and spirometry in the SAdf and LSAdf groups. The study revealed that in the LSAdf group, the correlation between the small-airway parameters measured by oscillometry and spirometry was significantly greater than that in the SAdf group.

Table 3 Analysis of correlations between spirometry and oscillometry parameters in all subjects (n = 1446)

Table 4 Analysis of correlations between spirometry and oscillometry parameters in all subjects in different groups based on the values of FEV1%, FEV1/FVC, and FVC%

Table 5 Analysis of correlations between spirometry and oscillometry parameters in SAdf group and LSAdf group

Fig. 3

a Correlation between traditional pulmonary function small airway parameters via Spearman correlation analysis. b Correlation between small-airway parameters and R5-R20 in The Impulse Oscillometry System by Spearman correlation analysis

Analysis of inconsistencies in diagnosing SAdf using spirometry and oscillometry

Tables 6 and 7 present the clinical characteristics and pulmonary function features across the different subgroups. Among the 1,446 patients, 237 (16.4%) had both no-SAdf, 215 (14.8%) had spirometry-only SAdf, 115 (7.9%) had oscillometry-only SAdf, and 879 (62.0%) had both SAdf. Compared to the oscillometry-only SAdf group, the spirometry-only SAdf group exhibited a higher proportion of males, a lower BMI, increased sputum production, more severe airflow limitation, and greater impairment in lung function, although they demonstrated lower airway resistance. Figure 4a and b illustrate the changes in pulmonary function features across the four groups, indicating a sequential decrease in lung function.

Table 6 Comparison of demographic and clinical characteristics in all subjects from different subgroups

Table 7 Lung function features in all subjects from different subgroups

Fig. 4

a Lung function features in the four subgroups. b Changes in small airway parameters in four subgroups

Multivariate logistic regression analysis of factors influencing inconsistencies in the diagnosis of SAdf

Table 8 displays the results of the adjusted analysis, comparing the subjects with oscillometry-only SAdf to those with spirometry-only SAdf, based on the factors of sex, BMI, smoking status, and smoking index. The analysis revealed that a higher BMI (odds ratio [OR] = 1.12, 95% confidence interval [CI] [1.01–1.24], p < 0.001) was an independent risk factor for oscillometry-only SAdf compared to the spirometry-only SAdf group. Conversely, being female (OR = 0.39, 95% CI [0.16–0.94], p < 0.001) and increased sputum production (OR = 0.24, 95% CI [0.08–0.77], p = 0.03) were associated with spirometry-only SAdf.

Table 8 Factors associated with inconsistencies were analyzed by multivariable logistic regression, comparing with spirometry-only SAdf, adjusted for sex, BMI, smoking status and smoking index

Discussion

This study systematically examined the correlation between small airway parameters obtained from spirometry and oscillometry in hospitalized patients, along with the factors influencing inconsistencies in the diagnosis of SAdf. The following conclusions were drawn from the correlation analysis: Firstly, significant correlations were observed between small airway parameters from oscillometry (AX, Fres, X5, R5-R20) and those from spirometry (FEF25%-75%, FEF50%, FEF75%). This finding suggests that despite differing detection principles, both methods effectively reflected SAdf. Notably, AX exhibited the strongest correlation, regardless of the presence of restrictive ventilation or large airway abnormalities. Secondly, subgroup correlation analysis revealed that in the LSAdf group, AX, Fres, X5, and R5-R20 had moderate correlations with FEF25%-75%. Conversely, correlations in the SAdf group were weak. We speculate that greater lung impairment corresponds with stronger correlations between small airway parameters measured by oscillometry and spirometry, highlighting a potential area for further research. Thirdly, the SAdf group predominantly exhibited no significant airflow limitation (FEV1/FVC > 70%), suggesting that SAdf may precede detectable airflow limitation. A study by Nowaf Y Alobaidi et al. confirmed small airway impairment to be a constant feature in patients with airflow limitation, with or without correction for FVC, and there was a significant reduction in FEF25%­75% even in patients with mild airflow limitation [26]. Fourthly, the presence of SAdf in 115 patients diagnosed via oscillometry from the no-SAdf group, underscored the inconsistency between spirometry and oscillometry. This finding indicated that oscillometry might be more sensitive than spirometry in detecting SAdf in patients with normal or mildly impaired lung function as measured by spirometry. We speculated that these patients might represent an early stage of small airway abnormalities. Li L Y et al. concluded that oscillometry was more sensitive than spirometry for detecting SAdf in subjects with chronic respiratory symptoms, suggesting its utility as an additional method for early-stage detection of SAdf [27]. Conversely, other studies suggest that spirometry may be more sensitive than oscillometry for diagnosing SAdf in patients with severe lung impairment [10, 27, 28]. This highlighted the importance of combining both indicators in the evaluation of SAdf, given their respective strengths and limitations.

The inconsistency analysis indicated that the oscillometry-only SAdf group exhibited less severe lung function impairment via spirometry but greater airway resistance measured by oscillometry compared to the spirometry-only SAdf group. This suggests that oscillometry can detect changes in airway resistance even in patients with normal or mildly impaired lung function, emphasizing the need for its inclusion in early disease stages. The combination of spirometry and oscillometry provided a more comprehensive view of lung function, facilitating the detection of disease progression and patient prognosis. Multivariate logistic regression revealed that being female, having a high BMI, and low sputum production were factors contributing to oscillometry-only SAdf. Females are more likely to be diagnosed with SAdf via oscillometry, possibly due to the reliance of FEF25%-75% accuracy on FVC, which can be less reliable in females with lower lung capacity [29].Female patients often have lower lung capacity, leading to unreliable FVC measurements, making it difficult to detect small-airway abnormalities via spirometry. High BMI is a risk factor for oscillometry-only SAdf. This finding also reported in another study [28]. It possibly due to mild inflammation and increased chest mechanical load in obese patients, increasing airway resistance [30]. In contrast, males and higher sputum production were factors associated with spirometry-only SAdf.

However, concerns exist regarding the reliability of FEF25%-75% for assessing small airways.

A recent article [3], concluded that small airway disease(SAD) must be distinguished from small airway dysfunction (SAdf) from the physiological point of view, stating that the diagnosis small airway disease cannot rely on a single lung function parameter, particularly FEF25%-75%. They recommend using at least three spirometric or plethysmographic parameters to predict small airway disease in populations with clinically stable moderate-to-severe asthma, asthma-COPD overlap, or COPD. Other studies have also indicated that FEF25%-75% by spirometric measures does not provide additional clinical information in large cross-sectional studies [31]. Nonetheless, Kwon et al. described that patients with an FEF25%-75% (expressed as a z-score) may have small airway impairment that could worsen over time with ongoing exposure to risk factors. A study by NY Alobaidi et al. supported the use of FEF25%-75% in assessing patients at risk of developing COPD, suggesting that those with an z-score < − 0.8453 may represent a phenotypic group reflecting early small airway impairment. This underscored the potential significance of FEF25%-75% as a marker for small airway impairment [26]. The utility of R5-R20 in IOS for assessing small-airway involvement is not universally accepted [32]. Cottini et al. [33] demonstrated that R5-R20 correlated closely with other markers of peripheral airway abnormalities, supporting its use as a surrogate measure of peripheral airway mechanics. Domestic and foreign scholars have made continuous efforts in exploring small airway dysfunction, hoping to find the most accurate method for diagnosing small airway function. At present, given the absence of a consensus on diagnosing SAdf, we advocate for the combined use of spirometry and oscillometry for comprehensive evaluation.

Additionally, the 2021 guidelines from the American Thoracic Society (ATS) and European Respiratory Society (ERS) recommend the use of the Global Lung Function Initiative (GLI) equations, which are the most widely adopted predictive models for spirometry, diffusing capacity of the lung for carbon monoxide (DLCO), and lung volume measurements, and they also strongly advocate expressing the degree of airflow limitation using z-scores for parameters such as FEV1, FVC, and the FEV1/FVC ratio [34]. However, lung function is influenced by a multitude of factors such as age, height, gender, race, and environment, and there is currently no single predictive equation that is universally applicable. Lacking prediction equations suitable for the Chinese population, we cannot readily convert to z-scores at present. In 2017, China updated the normal predicted values for spirometry based on a sample of 7115 individuals aged 4 to 80 from six administrative regions of the country, which is better suited for interpreting pulmonary function results of Chinese individuals. Thus, the predicted percentage ratio we have adopted may be more suitable for Chinese people. In the future, we will endeavor to find or develop predictive equations for lung function parameters that are applicable to Chinese individuals, so as to more accurately convert the lung function test results to z-scores. We advocate for more studies that can provide the necessary data to create these equations.

This study had several strengths: it analyzed the correlation of small airway parameters measured by spirometry and oscillometry while providing insights across different subgroups. These findings offered valuable guidance for clinicians in diagnosing SAdf and understanding lung function of patients. Additionally, the study investigated inconsistencies in the diagnoses of SAdf and explored the factors influencing these discrepancies.

However, there were limitations to this study. Firstly, it was a cross-sectional, single-center study predominantly involving patients with COPD. Secondly, there have been currently no uniform diagnostic criteria for SAdf. We used at least two out of the three indicators (FEF25%-75%, FEF50%, and FEF75%) and R5-R20 as diagnostic criteria for SAdf, which may not be applicable to patients with differing diagnostic criteria. Increasing evidence suggests that FEF25%-75% does not provide additional information on small airways, and R5-R20 has yet to be widely accepted for diagnosing SAdf, with its threshold undetermined. Thirdly, we employed fixed ratios of FEV1, FVC, and FEV1/FVC instead of z-scores which could potentially impact the inclusion criteria, disease progression assessment, and risk prediction models in clinical trials. Fourthly, using subjective categorical variables rather than the Modified Medical Research Council (mMRC) dyspnea scale and COPD assessment test (CAT) scores for assessing respiratory symptoms might have influenced the results. Finally, the study did not include more objective CT imaging results or plethysmographic parameters for a thorough analysis of small airways.

Conclusion

This study demonstrated that oscillometry could serve as a supplementary or alternative method for diagnosing SAdf, particularly in patients with normal or mildly impaired lung function measured by spirometry. Additionally, oscillometry might reflect the early stages of abnormalities in small airways. Given the absence of a gold standard for diagnosing SAdf, we emphasized the combined use of spirometry and oscillometry for a comprehensive evaluation of SAdf, facilitating timely interventions. Furthermore, there has been currently insufficient understanding of the z-score in spirometry among healthcare providers, and it has not been widely utilized in clinical practice. Clarification of this indicator and its clinical significance would be needed. We suggest the concurrent use both methods for subsequent evaluations in Chinese population.