Chronic obstructive pulmonary disease (COPD) ranks as the third leading cause of death globally, following ischemic heart disease and stroke.1 Characteristics of COPD are persistent airflow limitation and corresponding respiratory symptoms such as dyspnea, cough, and sputum production. COPD is usually diagnosed at an advanced stage, resulting in a heavy disease burden with high rates of mortality and disability. With the aging population increasing in high-income countries and rising smoking rates in developing nations, the number of deaths due to COPD is projected to exceed 5.4 million by 2060.2 A national cross-sectional study in China reported that approximately 100 million individuals have COPD, and prevalence was higher in men than in women (11.9% vs 5.4%).3 COPD in women has distinct characteristics expressed differently compared to men in terms of respiratory symptoms and disease outcomes. Sex differences in COPD management and therapy also warrant attention.4–6 The Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines recommend diagnosing COPD with a post-bronchodilator forced expiratory volume in 1s (FEV1)/forced vital capacity (FVC) ratio <0.70 and using FEV1% predicted to assess the severity of the obstruction.1 FEV1% predicted was associated with increased symptoms and long-term mortality.7,8 Recently, some studies have proposed new classifications of severity of the obstruction, including FVC-adjusted FEV1% predicted, the FEV1/FVC ratio, and FEV1z-scores.7,9–11 However, regardless of the obstruction standard employed, these measures are all based on spirometry tests. Spirometry tests require patients to exert maximal effort in exhalation, and certain clinical patients with severe dyspnea, may not complete spirometry tests. Consequently, clinicians faced significant challenges in assessing the severity of COPD to guide pharmacological treatment. Therefore, there is an urge to find supplementary methods to indirectly evaluate the severity of COPD.

Impulse oscillometry (IOS) is a technique to measure airway resistance during calm breathing but not require patient effort. IOS plays a crucial role in assessing small airway dysfunction, expiratory flow limitation, monitoring disease recovery, and disease progression.12–16 Recently, Liang et al. proposed new obstruction severity classifications using IOS z-scores, establishing cutoff values for grading the severity of COPD using IOS parameters[resistance at 5Hz (R5), difference between resistance at 5Hz and 20Hz (R5–R20), reactance at 5Hz (X5) z-scores]: stage I, II, III, IV (staging of airflow obstruction by IOS, SAIO).17 However, it remained unclear whether SAIO grades were correlated with respiratory symptoms, image abnormalities, and prognosis and had similar results to the GOLD classifications.

The objective of our study was to identify a supplementary method to assess COPD severity and guide treatment of patients with COPD. We hypothesized that SAIO grades may correlate with respiratory symptoms, image abnormalities, and prognosis similarly to GOLD classifications. To test this hypothesis, we conducted a prospective cohort study in China to investigate the associations between SAIO grades and symptom scores, imaging abnormalities, acute exacerbations (AE) and lung function decline at a 2-year follow-up.

Patients with COPD who had acceptable IOS data were enrolled in our study, of 833 patients. Table 1 presents the demographic characteristics of patients with COPD stratified by sex. Of the patients, 91.6% were male and 8.4% were female. Among males, 94.6% were current or former smokers, while 44.3% of females had a history of biomass exposure. The mean±standard deviation (SD) FEV1 of male patients was 2.01±0.61L. Table S1 illustrates the categorization of airflow obstruction using SAIO and GOLD stages. The proportion of grade I–IV assessed by R5 z-score was 42.3%, 11.6%, 19.8%, 26.3%; assessed by R5–R20 was 38.7%, 15.7%, 16.8%, 28.8%; assessed by X5 was 48.7%, 25.5%, 14.2%, 11.6%; assessed by FEV1%pred was 44.1%, 43.5%, 10.6%, 1.9%. Most subjects who changed categorization shifted from a less severe FEV1% predicted category to a more severe R5 z-score or R5–R20 z-score (Fig. 1).

Fig. 1.

Airflow obstruction categorization for SAIO and GOLD stages. R5: resistance at 5Hz; R5–R20: difference from R5 to R20; X5: reactance at 5Hz; FEV1: forced expiratory volume in 1s; FVC: forced vital capacity.

(0.17MB).

Concordance Between GOLD and SAIO Classes

Fig. 2 illustrates the Bangdiwala plots depicting agreement between GOLD and R5 z-scores, R5–R20 z-scores, and X5 z-scores. The agreement, measured by the weighted Bangdiwala B, between GOLD stages and R5 z-scores, R5–R20 z-scores, and X5 z-scores was 0.68, 0.70, and 0.83, respectively.

Fig. 2.

Bangdiwala agreement charts comparing classification of severity of airflow obstruction using GOLD and SAIO severity schema. R5: resistance at 5Hz; R5–R20: difference from R5 to R20; X5: reactance at 5Hz; GOLD: Global Initiative for Chronic Obstructive Lung Disease.

(0.21MB).

Symptom score, Emphysema, Air trapping Among SAIO Stages and GOLD Stages

Fig. 3 displays the difference of symptom score, emphysema, air trapping across SAIO stages. After adjusting for age, sex, smoking status, pack-years, BMI, family history of respiratory diseases, occupational exposure, and biomass exposure, the analysis of covariance indicated that SAIO stages III and IV were associated with higher mMRC and CAT scores, as well as more severe emphysema and air trapping, compared to stage I. However, no significant differences were observed between SAIO stage I and II in symptom scores, emphysema, and air trapping. Meanwhile, the GOLD stages demonstrated significant differentiation in terms of stages I–IV in symptom scores, emphysema, and air trapping (Fig. S1).

Fig. 3.

The difference among SAIO stages in symptom score, emphysema, and air trapping. Data are mean±SE. R5: resistance at 5Hz; R5–R20: difference from R5 to R20; X5: reactance at 5Hz; LAA−950: low-attenuation area of the lung with attenuation values below −950 Hounsfield units; LAA−856: low-attenuation area of the lung with attenuation values below −856 Hounsfield units; mMRC: modified British medical research council score; CAT: COPD assessment test. ***P<0.001; **P<0.05.

(0.21MB).

AE

The SAIO stages exhibited similar discriminatory power to GOLD stages in predicting AE at 2-year follow-up between stage I and stage II–IV. After adjusting for age, sex, smoking status, pack-years, BMI, family history of respiratory diseases, occupational exposure, biomass exposure and follow-up time. Compared with SAIO stage I, the incidence rate ratio (IRR) for the presence of SAIO stages assessed by R5 z-scores II through IV airflow obstruction were 1.33 (95% CI 1.03–1.71), 1.30 (95% CI 1.05–1.60), and 1.91 (95% CI 1.60–2.28). The IRR for the presence of SAIO stages assessed by R5–R20 z-scores II through IV airflow obstruction were 1.38 (95% CI 1.09–1.74), 1.38 (95% CI 1.10–1.74), and 2.02 (95% CI 1.68–2.43); The IRR for the presence of SAIO stages assessed by X5 z-scores II through IV airflow obstruction were 1.04 (95% CI 0.86–1.26), 1.79 (95% CI 1.48–2.18), and 1.38 (95% CI 1.10–1.73). Similarly, compared with GOLD stage I, IRR for the presence of GOLD stages II through IV airflow obstruction were 1.61 (95% CI 1.37–1.90), 1.47 (95% CI 1.14–1.90), and 2.08 (95% CI 1.29–3.37) (Fig. 4). We additionally analyzed the differences in moderate-to-severe AE among the SAIO stages and GOLD stages. Our results showed that SAIO stages III–IV had higher risk of moderate-to-severe AE than stage I, but no differences in moderate-to-severe AE were observed between SAIO stages I and II (Fig. S2). Then subgroup analysis stratified by sex was conducted, our results showed that significant differences in AE or moderate-to-severe AE between SAIO stages III–IV and stages I was observed in male individuals but hardly any in female individuals (Figs. S3–S6).

Fig. 4.

The difference among SAIO stages in acute exacerbations in overall individuals. Data are mean±SE. AE: acute exacerbations; R5: resistance at 5Hz; R5–R20: difference from R5 to R20; X5: reactance at 5Hz; GOLD: Global Initiative for Chronic Obstructive Lung Disease; IRR: the incidence rate ratio.

(0.3MB).

Our study presents a novel severity classification based on IOS parameters (SAIO stages) and investigates their associations with symptoms, imaging abnormalities, AE and lung function decline. Our study has reported several significant findings. Initially, SAIO stages evaluated by X5 z-scores demonstrated a stronger agreement with GOLD stages in severity classifications compared to other parameters. Secondly, the SAIO grades identify a higher proportion of patients in stages III–IV compared to the GOLD grades. Finally, SAIO grade I and III–IV exhibited similar discrimination in symptom scores, emphysema, air trapping symptoms, and AE compared to GOLD grades.

The SAIO stages were defined based on severity classifications of COPD using FEV1z-scores.17 However, the clinical significance of the SAIO stages in COPD remains unclear. In this study, we utilized FEV1%pred instead of FEV1z-score to evaluate COPD severity. Several factors were considered: (1) the ATS/ERS and GOLD guidelines recommend using FEV1%pred for severity classifications. (2) Although the FEV1z-score can eliminate biases related to age, sex, height, and race,27 in terms of mortality, a study by Hegewald et al. indicated that FEV1%pred was more accurate in predicting the risk of 5-year mortality than FEV1z-score.7 Therefore, we focused on the agreement between the GOLD stages as assessed by FEV1%pred and SAIO stages. We found that the SAIO stages assessed by X5 z-scores demonstrated better agreement with GOLD stages in severity classifications compared to other parameters. This result indicated that the reactance value was more appropriate than the resistance values for reflecting GOLD stages using FEV1%pred. We further analyzed the distributions of SAIO and GOLD stages. Compared to the GOLD classifications, the SAIO classifications, as defined by R5 and R5–R20, divided the GOLD stage II into SAIO III–IV. We believed that this phenomenon was attributable to the different physiological mechanisms of each indicator. Firstly, Rrs (R5 and R5–R20) primarily reflects changes in airway diameter, particularly the degree of small airway stenosis.28 Small airway disease represented an early lesion in the progression to COPD.29 Conversely, FEV1%pred primarily reflected the obstruction of large airway, and may not be sensitive to small airway lesions.30,31 Thus, small airway disease may occur before abnormalities in FEV1 become evident. This result also explained why the SAIO classifications defined by R5 and R5–R20 divided GOLD stage II into SAIO III–IV.

The strength of our study lies in exploring the distribution of the new SAIO classifications in COPD and analyzing the differences in physiological indicators and AE among SAIO stages. Clinically, some patients cannot cooperate with spirometry due to the exertion of maximal effort in exhalation, which makes it challenging for clinicians to assess disease severity in these cases. Our study aimed to provide an indirect, effortless tool for evaluating the severity of COPD patients who cannot cooperate with lung function tests. We further analyzed the discriminatory ability of SAIO grades in assessing severity through symptom scores and imaging. Our study showed that SAOI stage III–IV had higher mMRC and CAT scores, more severe emphysema, and air trapping than those with stage I. Similar findings were also observed in GOLD grades. Meanwhile, GOLD stage II showed higher mMRC and CAT scores, more severe emphysema, and air trapping compared to GOLD stage I. This finding was consistent with that reported by Bhatt et al., who found that mMRC score and emphysema worsen with increasing GOLD grade severity in the Pittsburgh Cohort.9 However, no significant differences were observed in symptom scores, emphysema, or air trapping between SAIO grade I and II. This is not surprising. Patients with GOLD grade I and II have already exhibited imaging abnormalities (emphysema, air trapping), whereas SAIO grade I and II only exhibit early lesions (such as SAD), and these patients may not present with specific imaging abnormalities or obvious respiratory symptoms. Of course, our results showed that SAIO classifications may be more effectively used to assist in the evaluation of patients with GOLD III to IV who cannot cooperate with lung function. Due to severe dyspnea typically occurred in these patients. It was difficult for the patients to complete spirometry. IOS only requires calm breathing and has a similar distinction of symptoms, imaging, and AE as spirometry. Therefore, IOS can be used to assist in evaluating the severity of patients with GOLD III–IV to guide treatment.

The goal of COPD treatment is to alleviate current symptoms and reduce the risk of future AE.1 We analyzed the differences in the number of AE between GOLD and SAIO classifications based on 2-year follow-up data. We found that GOLD grades demonstrated results similar to SAIO grades in predicting AE. Patients with grade II–IV experienced a higher number of AE over the following two years compared to those with grade I. This result indicated that both the GOLD and SAIO grades were effective in predicting the risk of AE, suggesting their potential as alternative tools in clinical practice. Identifying high-risk patients with grades II–IV can assist clinicians in developing personalized treatment plans that emphasize aggressive intervention and management to mitigate the occurrence of AE. Interestingly, it has been observed that patients classified in stage III of the X5 parameter presented a higher number of exacerbations than those in stage IV. We considered possible explanations, such as limited sample size or patient heterogeneity across these stages. Additionally, it would be beneficial to reflect on the need for future studies with larger samples to verify this trend and gain a better understanding of the behavior of the X5 parameter in COPD. Finally, in lung function decline, we only found GOLD grade I had a faster decline in FEV1 than grade IV. The result was consistent with previous studies that have reported faster lung function decline in patients with GOLD I compared to patients with advanced COPD.32,33 However, no differences were found in the FEV1 decline among SAIO grades. We also guessed that the distribution gap of SAIO and GOLD grades could play an important role, especially as some patients with GOLD grade II were classified as SAIO grade I, III–IV. Lung function decline was measured on the basis of FEV1, and there was a so-called “horse race effect” that patients with advanced COPD had slower lung function decline than mild to moderate COPD because of a lower baseline FEV1. Inconsistent GOLD classifications exist in all SAIO grades, which also affects disease prognosis to some extent. Long-term follow-up is needed in the future to explain this potential mechanism.

There were some limitations to our research. First of all, due to a small number of patients with GOLD grade III–IV in our study, in the more advanced stages, where variability in exacerbations could influence the results. Future large cohort studies will be necessary to further elucidate the relationship between SAIO classifications and the prognosis of COPD. Secondly, our study only demonstrated that SAIO classifications and GOLD classifications had similar effects in predicting AE, but we found that SAIO classifications do not offer advantages in distinguishing certain physiological reactions than GOLD classifications in patients with grade I and II. Therefore, new physiological indicators are needed to explore the advantages of SAIO grades. Thirdly, all-cause mortality is an important indicator for assessing disease prognosis.2 Since our cohort was followed up for only 2 years and there have been few fatalities, we could not compare differences in all-cause mortality between SAIO and GOLD classifications. Furthermore, the SAIO classifications were based on the IOS parameter z-scores, which were affected by race,28 and lacked validation of results in other countries.34,35 Therefore, the generalizability of our study's findings was limited. Moreover, ΔX5, the difference between inspiratory and expiratory reactance at 5Hz, was an indicator of assessing expiratory flow limitation in COPD. ΔX5 showed significantly correlated with FEV1, and reflected the severity of COPD to some extent.36,37 Therefore, exploring whether ΔX5 can be used as a new classification of COPD is worth considering. However, we currently lack a cutoff value for ΔX5 to effectively assess the severity classifications.

The new severity classifications, SAIO grades, provide discrimination between the stage I and stage III, IV for symptom score, emphysema, air trapping, and AE similar to the GOLD classifications. The SAIO classifications method offer an alternative assessment for patients unable to complete spirometry, thereby assisting clinicians in gaining a comprehensive understanding of a patient's disease status, leading to more accurate diagnosis and treatment decisions. With further research and clinical validation, the SAIO classifications are expected to become a crucial tool in the management of COPD.

P.X.R., Y.M.Z., and L.F.L. designed the project and planned the statistical analysis. L.F.L. drafted and revised the paper. L.F.L., Q.W., G.Y.T., F.W., Z.S.D., J.Q.P., C.Q.D., K.N.Z., X.H.W., S.Q.Y., Y.Q.H., C.L.Y., S.T.C., P.X.R., and Y.M.Z. collected and monitored the data collection. All authors approved the final draft of the manuscript for publication. L.F.L. take responsibility for the integrity of the data and the accuracy of the data analysis. L.F.L. is the study guarantors.

This study adhered to the ethical guidelines outlined in the Declaration of Helsinki. The research protocol received approval from the Ethics Committee of the First Affiliated Hospital of Guangzhou Medical University (Approval No. 2018-53) prior to initiation. Written informed consent was obtained from all participants prior to their enrollment in the study.

Not applicable.

This work was supported by Foundation of Guangzhou National Laboratory (SRPG22-018 and SRPG22-016), the National Natural Science Foundation of China (81970045, 81970038, 82270043), the Clinical and Epidemiological Research Project of State Key Laboratory of Respiratory Disease (SKLRD-L-202402), and Major Clinical Research Project of Guangzhou Medical University's Scientific Research Capability Improvement Plan (GMUCR2024-01012) the Guangzhou Science and Technology Plans (No. 202201020372), the grant of State Key Laboratory of Respiratory Disease (SKLRD-Z-202315), and Research Capability Strengthening and Fundamental Improvement Project of Guangzhou Medical University (2024SRP080).