The subjects were recruited during the community screening phase of the Early Chronic Obstructive Pulmonary Disease (ECOPD) study, a multicentre, community-based cohort study in China.21 The ECOPD study, approved by the Ethics Committee of the First Affiliated Hospital of Guangzhou Medical University (No. 2018-53), aims to longitudinally track individuals with and without COPD to identify parameters that may predict disease progression in early-stage COPD. Written informed consent was obtained from all subjects. Non-COPD subjects and mild-to-moderate COPD patients were included in this study.

The inclusion criteria were age 40–80 years with complete questionnaire, lung function, and CPET data. The exclusion criteria were: (1) acute exacerbation within past 4 weeks; (2) severe/very severe COPD defined as postbronchodilator FEV1/forced vital capacity (FVC)<0.70 and FEV1<50% predicted; (3) lobectomy history; (4) active cancer treatment; (5) active pulmonary tuberculosis, silicosis, extensive bronchiectasis, or other serious lung conditions; or (6) CPET contraindications and severe cardiovascular disease affecting ventilatory efficiency, including pulmonary hypertension, heart failure, pulmonary embolism, and severe coronary heart disease. To establish the independent prognostic value of impaired ventilatory efficiency, we further excluded subjects meeting predefined high-risk criteria: (1) CAT≥10, (2) mMRC≥2, (3) postbronchodilator FEV1<60% predicted, or (4) frequent exacerbations (≥2 moderate or ≥1 severe events in the year prior to baseline).11,12

Demographic information, respiratory-related risk factors, chronic respiratory symptoms, comorbidities, medication history, and exacerbations experienced in the year before baseline were collected.22,23 Respiratory-related risk factors included smoking history, passive smoke exposure, biomass exposure, occupational exposure, and a family history of respiratory disease. Chronic respiratory symptoms included dyspnea, chronic cough, chronic sputum, and wheezing.

Prebronchodilator and postbronchodilator lung function tests used portable spirometers (CareFusion, Yorba Linda, CA, USA), adhering to American Thoracic Society and European Respiratory Society guidelines for standard operating procedures and quality control.24,25 We used the reference values for lung function provided by the 1993 European Community for Steel and Coal, and subsequently adjusted the predicted FEV1 values using correction factors tailored to the characteristics of the Chinese population (0.95 for men, 0.93 for women).26,27 Non-COPD was defined as postbronchodilator FEV1/FVC ratio≥0.70, while mild-to-moderate COPD was defined as postbronchodilator FEV1/FVC ratio<0.70 and FEV1≥50% predicted.1

The subjects underwent a maximal incremental CPET on a calibrated cycle ergometer (Quark PFT Ergo Bp900; COSMED, Rome, Italy). They were encouraged to maintain a pedaling speed of 55–65rpm during the exercise phase until maximum exertion or limiting symptoms occurred, transitioning then to the recovery phase.28 Ventilation flow and CO2 concentrations were measured via breath-by-breath analysis during CPET. V˙CO2 was calculated as the product of CO2 concentration and V˙E. The nadir V˙E/V˙CO2 represents the lowest 30-s average ratio during exercise, typically occurring around the ventilatory compensation point, and is neither measured at maximal exertion nor protocol-dependent.16 Ventilatory efficiency in this study was assessed by the nadir V˙E/V˙CO2 because its high reproducibility in both healthy individuals and COPD patients.29 Impaired ventilatory efficiency was considered if the nadir V˙E/V˙CO2 exceeded the ULN as per the Wasserman and Sun equation.16

Thoracic CT imaging was acquired using 128-slice multi-detector scanners (Siemens/United-imaging) during maximal inspiration and expiration. Quantitative analyses through 3D Slicer's Chest Imaging Platform included: total lung capacity and residual volume measurements; emphysema quantification via inspiratory low-attenuation areas<−950Hounsfield units; air trapping assessment using expiratory attenuation thresholds<−856Hounsfield units; along with three-dimensional vascular modeling to determine total intraparenchymal vessel volume and small vessel fraction.30–32

The study outcomes of interest were lung function decline, exacerbation, and respiratory symptoms. The subjects underwent annual prebronchodilator and postbronchodilator lung function tests, with procedures and quality control consistent over the follow-up visits. In cases of exacerbation during follow-up, lung function data collection was delayed until 4 weeks after resolution. Exacerbation, including acute exacerbation for COPD patients and acute respiratory events for non-COPD subjects, was defined as the presence/worsening of ≥2 of the following symptoms: cough, sputum production, purulent sputum, wheezing, or dyspnea lasting>48h after excluding congestive heart failure, pulmonary embolism, pneumothorax, pleural effusion, and arrhythmia.4,5,33 Moderate-to-severe exacerbations required outpatient/emergency/hospital care with antibiotics and/or systemic corticosteroids. Subjects were given the contact information of the researcher and instructed to promptly report any respiratory symptom deterioration. The research team documented and evaluated all exacerbations details.

This study was an exploratory analysis, so the sample size was not calculated previously and no pairwise adjustment for multiple comparisons was performed. The calculated power for the primary outcome was 87% in this study based on the following data: a 27mL/year difference in the annual decline of postbronchodilator FEV1 between non-COPD subjects (n=465) and patients with impaired ventilatory efficiency (n=84), with a standard deviation of 90mL/year and a two-tailed significance level of 5% (PASS 23.0.2).6

Baseline characteristics were compared between the groups by analysis of variance for continuous variables and the Chi-square test or Fisher's exact test for categorical variables. Random coefficient regression models were used to compare the annual decline in lung function among the three groups. Mixed-effects models for repeated measures were used to identify differences in lung function and CAT score between the groups across multiple visits. Least-squares mean estimates were used to determine the changes of lung function at each time point relative to baseline. Exacerbations were evaluated using a negative binomial model, and the natural log-transformed follow-up duration was considered as an offset variable. The chi-square test was used to identify differences in mMRC dyspnea score between groups at each visit.

In analyses of lung function decline, exacerbations, and respiratory symptoms, we adjusted for potential confounders including age, sex, body mass index, smoking status, smoking index, passive smoking at home, biomass exposure, occupational exposure, and family history of respiratory diseases. Longitudinal lung function models further incorporated baseline parameter values, while exacerbations analyze further accounted for pre-baseline exacerbation frequency. As severe cardiovascular diseases affecting ventilatory efficiency were exclusion criteria, cardiovascular comorbidities with comparable prevalence across groups were not included in adjustment models.

To validate the robustness of the results, we conducted four sensitivity analyses. First, we analyzed the study outcome without excluding subjects meeting predefined high-risk criteria. Second, we used the lower limit of normal for FEV1/FVC ratio reference values and predicted FEV1 values obtained from the healthy Chinese population to diagnose and grade COPD.34 Third, we defined impaired ventilatory efficiency using an absolute cut-off value (nadir V˙E/V˙CO2>34) based on the previous study.16 Finally, we analyzed subjects who reached the ventilation compensation point.

Subjects with at least one follow-up data point for each outcome were included in longitudinal analyses. Minimal missing data in key variables were assumed missing at random, and analyses conducted on available cases. Statistical analyses were performed using SPSS 24.0 software (IBM Corp., Armonk, NY, US) and SAS 9.4 software (SAS, Cary, NC). Two-sided P<0.05 was considered significant.

The study flowchart is shown in Fig. 1. From 1004 subjects completing baseline CPET with valid questionnaires and lung function data, 909 were non-COPD or mild-to-moderate COPD qualified for nadir V˙E/V˙CO2 analysis. After excluding 129 subjects meeting predefined high-risk criteria, the final cohort (n=780) comprised three subgroups: 516 non-COPD subjects, 173 patients with normal ventilatory efficiency, and 91 patients with impaired ventilatory efficiency. The cohort maintained 88% (684/780) retention rate at 3-year follow-up.

Fig. 1.

Flowchart of patient eligibility, screening, and follow-up. * To establish the independent prognostic value of impaired ventilatory efficiency, we further excluded subjects meeting predefined high-risk criteria requiring treatment: (1) CAT≥10; (2) mMRC≥2; (3) post-bronchodilator FEV1<60% predicted; or (4) frequent exacerbations (≥2 moderate or ≥1 severe exacerbations in the year prior to baseline). No subjects in this community-based study had frequent exacerbations before enrolment. † Subjects with at least one follow-up data point for each outcome (lung function decline, exacerbations, respiratory symptoms) were included in longitudinal analyses. Abbreviations: BD=bronchodilator, COPD=chronic obstructive pulmonary disease, CPET=cardiopulmonary exercise testing, CAT=COPD Assessment Test, mMRC=modified Medical Research Council, FEV1=forced expiratory volume in 1 second; FVC=forced vital capacity, ULN=upper limit of normal, V˙E/V˙CO2 = ventilatory equivalent for carbon dioxide production.

(0.3MB).

Patients with impaired ventilatory efficiency (mean age 64.1 years, 98.9% male) demonstrated higher smoking exposure (55.0±33.4 pack-years), increased prevalence of chronic bronchitis and respiratory symptoms, and elevated CAT/mMRC dyspnea scores compared to non-COPD subjects. Among patients with impaired ventilatory efficiency, 21.1% had previously used respiratory medication, with 4.4% using long-acting bronchodilators (Table 1). They also had poorer lung function and exercise tolerance, higher RV/TLC, and more severe lung structural changes such as emphysema and air trapping than non-COPD subjects and patients with normal ventilatory efficiency (Table 2).

Subjects without lung function follow-up data had fewer males, more never smokers, lower smoking index, less dyspnea symptom, lower proportion of mMRC≥1, and more diabetes than those with follow-up data (e-Table 1). Additionally, subjects without follow-up data on other respiratory outcomes (exacerbations, CAT, or mMRC) also had fewer males and more never smokers than those with corresponding follow-up data (e-Tables 2–4).

At the 3-year follow-up, COPD patients with impaired ventilatory efficiency displayed a greater annual decline in postbronchodilator FEV1 (54 [95% confidence interval [CI]: 32–76]mL/year) than patients with normal ventilatory efficiency (31 [95% CI: 15–47]mL/year, adjusted mean difference [AMD]=25 [95%CI: 7–44], P=0.008) and non-COPD subjects (31 [95% CI: 22–40]mL/year, AMD=27 (95% CI: 11–44), P=0.001). However, there was no significant difference between patients with normal ventilatory efficiency and non-COPD subjects (AMD=2 [95%CI: −11 to 15], P=0.756). Similar findings were observed for prebronchodilator FEV1% of predicted, postbronchodilator FEV1% of predicted, and postbronchodilator FEV1/FVC ratio (Fig. 2 and e-Table 5). The adjusted least squares mean for pre- and post-bronchodilator FEV1, FEV1% predicted, FEV1/FVC ratio, and FVC at each visit were additionally presented in e-Table 6.

Fig. 2.

Longitudinal change in lung function throughout the study. Data are shown as the mean (95%CI). Mixed-effects models were used to identify differences in lung function among groups across multiple visits. Least-squares mean estimates were used to determine the changes at each time point relative to baseline. The model was adjusted for confounding factors including age, sex, body mass index, smoking status, smoking index, passive smoking at home, biomass exposure, occupational exposure, family history of respiratory diseases, and baseline spirometric values (pre or postbronchodilator FEV1 or FEV1% predicted). Abbreviations: COPD=chronic obstructive pulmonary disease, FEV1=forced expiratory volume in 1second, CI=confidence interval.

(0.34MB).

Patients with impaired ventilatory efficiency experienced significantly more exacerbations than non-COPD subjects, with total exacerbations rates of 0.58 vs. 0.34 per patient-year (adjusted relative risk [RR]=1.64, 95% CI: 1.19–2.30, P=0.003) and moderate-to-severe exacerbations rates of 0.35 vs. 0.18 per patient-year (adjusted RR=1.80, 95% CI: 1.23–2.63, P=0.002). Patients with normal ventilatory efficiency exhibited total and moderate-to-severe exacerbation rates of 0.41 and 0.22 per patient-year, respectively, both lower than those with impaired ventilatory efficiency (Total: adjusted RR=1.38, 95% CI: 0.97–1.97, P=0.073; Moderate-to-severe: adjusted RR=1.56, 95% CI: 1.04–2.35, P=0.031). No significant differences in exacerbation rate were observed between patients with normal ventilatory efficiency and non-COPD subjects (Total: adjusted RR=1.20, 95% CI: 0.92–1.55, P=0.184; Moderate-to-severe: adjusted RR=1.15, 95% CI: 0.84–1.57, P=0.381) (Fig. 3).

Fig. 3.

The frequency of exacerbation/acute respiratory events. Data are shown as the mean (95% CI). The frequency of exacerbations was evaluated for the relative risk using a negative binomial model over the 3-year follow-up period. The total events occurrences served as the response variable, while the natural log-transformed follow-up duration was considered as an offset variable. The analysis adjusted for potential confounders, including age, sex, body mass index, smoking status, smoking index, passive smoking at home, biomass exposure, occupational exposure, family history of respiratory diseases, and number of exacerbations during the year prior to baseline. Abbreviations: COPD=chronic obstructive pulmonary disease, RR=relative risk, CI=confidence interval.

(0.3MB).

COPD patients with impaired ventilatory efficiency showed significant differences in CAT scores and mMRC dyspnea scores compared with non-COPD subjects and patients with normal ventilatory efficiency starting at the 2-year follow-up. In contrast, minimal differences were observed in both scores between non-COPD subjects and patients with normal ventilatory efficiency subjects by the 3-year follow-up (Fig. 4, e-Tables 7, and 8).

Fig. 4.

Longitudinal change in CAT and mMRC dyspnea score throughout the study. Data are shown as the mean (95%CI) or n (%), as appropriate. Participants had at least one follow-up data point were included in the analysis for CAT and mMRC. Mixed-effects models were used to identify differences in CAT score between groups across multiple visits. The Chi-square test was used to identify differences in mMRC dyspnea score between groups across multiple visits. Abbreviations: COPD=chronic obstructive pulmonary disease, CAT=COPD Assessment Test, mMRC=modified Medical Research Council.

(0.17MB).

A sensitivity analysis of 909 subjects, including those meeting predefined high-risk criteria, aligned with the primary analysis (e-Tables 9–13). Whether using the latest lung function reference values for Chinese population (e-Tables 14–18), defining impaired ventilatory efficiency as a nadir V˙E/V˙CO2>34 (e-Tables 19–23) or including only subjects who reached the ventilation compensation point (e-Tables 24–28), patients with impaired ventilatory efficiency experienced faster lung function decline and a higher risk of moderate-to-severe exacerbations than non-COPD subjects. All above respiratory health outcome were similar between patients with normal ventilatory efficiency and non-COPD subjects.