Risk Factors of FEV1/FVC Decline in COPD Patients

Authors: Na Young Kim, Deog Kyeom Kim, Shinhee Park, Yong Il Hwang, Hyewon Seo, Dongil Park, Seoung Ju Park, Jin Hwa Lee, Kwang Ha Yoo, Hyun Woo Lee

Abstract

Background

Factors influencing the decline in forced expiratory volume in one second (FEV1)/forced vital capacity (FVC) for chronic obstructive pulmonary disease (COPD) progression remain uncertain. We aimed to identify risk factors associated with rapid FEV1/FVC decline in patients with COPD.

Methods

This multi-center observational study was conducted from January 2012 to December 2022. Eligible patients were monitored with symptoms, spirometric tests, and treatment patterns over 3 years. Rapid FEV1/FVC decliners were defined as the quartile of patients exhibiting the highest annualized percentage decline in FEV1/FVC.

Results

Among 1,725 patients, 435 exhibited rapid FEV1/FVC decline, with an annual change of −2.5%p (interquartile range, −3.5 to −2.0). Rapid FEV1/FVC decliners exhibited lower body mass index (BMI), higher smoking rates, elevated post-bronchodilator (BD) FEV1, higher post-BD FEV1/FVC, and a lower prevalence of Staging of Airflow Obstruction by Ratio (STAR) stage IV. Rapid FEV1/FVC decline was not linked to the annual exacerbation rate, but there was an association with symptom deterioration and FEV1 decline. In multivariable analyses, low BMI, current smoking, increased modified Medical Research Council dyspnoea score, low post-BD FEV1, low STAR stage, high forced mid-expiratory flow (FEF25-75%), accelerated FEV1 decline, and not initiating dual BD therapy were identified as independent risk factors for rapid FEV1/FVC decline.

Conclusion

We identified the risk factors for rapid FEV1/FVC decline, including BMI, smoking, symptoms deterioration, FEV1 decline, and adherence to standard inhaler treatment. Our findings underscore the potential benefits of maintaining consistent use of long-acting beta-agonist/long-acting muscarinic antagonist even in the presence of worsening symptoms, in attenuating FEV1/FVC decline.

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is a heterogeneous airway disease characterized by irreversible airflow limitation.1 Airflow obstruction has been typically measured by the forced expiratory volume in one second (FEV1)/forced vital capacity (FVC) ratio. In addition, the severity of COPD is usually evaluated using post-bronchodilator (BD) percentage predicted of FEV1 (ppFEV1), which has significant correlations with symptoms,23 acute exacerbations,4 and mortality.5 The ppFEV1 can be estimated by age, height, and gender-adjusted reference values.6 However, the use of ppFEV1 has limitations due to factors like obesity,7 genetics,8 and lung volume variations.9 Therefore, alternative approaches for assessing the severity of airway obstruction such as the FEV1/FVC ratio, have been proposed for COPD patients.1011 In the COPDgene study, the Staging of Airflow Obstruction by Ratio (STAR) stage, based on the FEV1/FVC ratio, demonstrated an enhanced performance for predicting mortality, especially in mild COPD patients, compared to conventional FEV1 severity assessments.11

Physiological decline in lung function begins in early adulthood and progresses with aging; however, COPD patients experience faster and more heterogenous lung function decline compared to the general population.12 It is well known that a rapid decline in lung function is associated with an increased risk of COPD-related hospitalization13 and death.14 Recently, the clinical significance of the annual FEV1/FVC decline rate has been evaluated in the general population.15 This study suggested that a rapid FEV1/FVC decline in the general population is linked to the development of obstructive lung disease and higher mortality rates.15 Certain studies showed that occupational dust exposure16 and narcotic use17 were associated with a decrease in FEV1/FVC. However, previous studies were conducted in the general population, and there has been a scarcity of studies examining the association between FEV1/FVC decline specifically in COPD patients.

The identification of risk factors for COPD progression stands as a pivotal role in the strategic landscape of COPD management. Most studies on the risk factors for COPD progression have predominantly focused on changes in FEV1, leaving the investigation of rapid declines in FEV1/FVC relatively underexplored. Hence, our study aimed to identify independent risk factors associated with a rapid decline in post-BD FEV1/FVC (%p/year).

METHODS

Our study adhered to the guidelines outlined in the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.18

Study design and eligibility criteria

Our nested case-control study used database of the Korea COPD Subgroup Study (KOCOSS) cohort (NCT02800499), which prospectively enrolled individuals diagnosed with COPD from January 2012 to December 2022 at 54 medical institutions in South Korea. Detailed information on study design and methodology is available in a prior study.19 In KOCOSS cohort, COPD patients ≥ 40 years and post-BD FEV1/FVC ratio < 0.7 were included.20 In our study, inclusion criteria required patients to have baseline and at least one follow-up post-BD spirometric assessment over the 3-year follow-up period. Regarding exclusion criteria, patients with statistically significant outliers in annual changes of post-BD FEV1/FVC were excluded, given that their spirometric assessments were considered unreliable, unrepeatable, or non-predictable.

Baseline variables

In the baseline assessment, we compiled demographic details from eligible patients, covering age, sex, body mass index (BMI), smoking habits, respiratory and non-respiratory comorbidities, as well as a history of previous exacerbations. We acquired information on baseline symptoms, including the modified Medical Research Council (mMRC) dyspnoea scale, the COPD assessment test (CAT) score, and St. George’s Respiratory Questionnaire for COPD Patients (SGRQ-C) score. Additionally, we checked the current inhaled therapies and oral medications being prescribed at the baseline.

Baseline spirometric data included post-BD FEV1, FVC, FEV1/FVC, diffusing capacity for carbon monoxide (DLCO), DLCO/alveolar volume (VA), and total lung capacity. Evaluation of exercise capacity involved the assessment of the 6-minute walking distance (6MWD). For laboratory analysis, the information on complete blood cells with differential counts was collected. Radiological assessment entailed an exploration of structural abnormalities, including emphysema, bronchiectasis, and tuberculosis-induced destructed lung, based on chest computed tomography (CT) scans.

Dynamically changing variables

We investigated dynamically changing variables over a 3-year period, including moderate-to-severe or severe exacerbation rates, mMRC change, CAT score change, SGRQ-C score change, 6MWD change, annual FEV1 change, annual FVC change, and inhaled therapy treatment patterns, to investigate their association with a rapid decline in FEV1/FVC. The analyses for dynamically changing variables were conducted with available data, excluding cases with missing values. Annual FEV1 change (mL/yr), annual FVC change (mL/yr), and annual FEV1/FVC change (%p/yr) were estimated for each patient by dividing the difference between baseline lung function and the last follow-up lung function values by the duration of the follow-up period.

Definition of rapid decliners

We employed the definition of rapid FEV1/FVC decline based on quartiles of the annual changes in FEV1/FVC, as introduced in a previous study.15 Patients with the most negative change in FEV1/FVC (1st quartile) were categorized as rapid FEV1/FVC decliners, while patients with the remaining quartiles (2nd, 3rd, and 4th quartiles) were designated as non-rapid FEV1/FVC decliners.

Statistical analysis

We used Student’s t-test for the comparison of continuous variables and the χ^2^ test for categorical variables between rapid and non-rapid FEV1/FVC decliners. Univariable and multivariable logistic regression analyses were performed to identify significant risk factors associated with rapid FEV1/FVC decliners. Significant covariates identified from the univariable analysis as risk factors for rapid FEV1/FVC decline were included in the multivariable analysis. We assessed significant multicollinearity with a variance inflation factor > 4.0. Variables exhibiting multicollinearity were analyzed in separate logistic regression models. For multivariable analyses, handling missing values included the use of the ‘mice’ R package, leveraging associations with other variables for imputation. We thoroughly checked the similarity in data distribution and representative values before and after the imputation process. Statistical significance was determined at a P < 0.05. R statistical software, version 4.3.2 (R Core Team [2021], Vienna, Austria), was employed for all statistical analyses.

Ethics statement

The study protocol received approval from the Institutional Review Board (IRB) at each hospital (Seoul Metropolitan Government-Seoul National University Boramae Medical Center, IRB No. 06-2012-36). All participants provided written informed consent at the time of enrollment. This study adhered to the principles outlined in the Declaration of Helsinki.

RESULTS

Among the 3,477 patients with COPD, 1,992 underwent both baseline and follow-up pulmonary function tests over a 3-year observation period (Fig. 1). We included the remaining eligible 1,725 patients with a reliably estimated post-BD FEV1/FVC decline rate. The median annual post-BD FEV1/FVC change rate was 0%p/year (−1.3 to 1.5) (Supplementary Fig. 1). The patients were categorized according to the quartiles of the annual post-BD FEV1/FVC change rate. The 1st quartile, representing the fastest FEV1/FVC decline, was defined as the rapid FEV1/FVC decliners (n = 435, 25.2%). The median change of FEV1/FVC in rapid decliner was −2.5 (interquartile range [IQR], −3.5 to −2.0) (Supplementary Fig. 2).

Demographic and clinical features

Rapid FEV1/FVC decliners had lower BMI and exhibited a higher proportion of current smokers (Table 1). There were no significant differences in terms of comorbidities, exacerbations, symptoms, or treatments between the rapid and non-rapid FEV1/FVC decliners. In the baseline spirometric assessment, rapid FEV1/FVC decliners showed higher values in post-BD FEV1 (%), post-BD FEV1/FVC (%), and forced mid-expiratory flow (FEF25-75%), and there was a significantly lower prevalence of STAR stage IV (Table 2). The percentage of blood eosinophil counts (BECs) exceeding 300 /uL did not show a significant association with a rapid FEV1/FVC decline. Additionally, radiologic findings in CT showed no significant differences between the two groups.

FEV1/FVC decline and change of clinical features

The annual change was −2.5%p (IQR, −3.5 to −2.0) in rapid FEV1/FVC decliners, while it increased by 0.6%p (IQR, −0.3 to 2.0) annually in non-rapid FEV1/FVC decliners (Supplementary Fig. 3). Rapid FEV1/FVC decline was not associated with the annual exacerbation rate, but was associated with the worsening of symptoms and FEV1 decline (Table 3). A significant increase in the deterioration of patient reported outcomes (PROs) beyond the minimum clinically important difference (MCID), including mMRC, CAT score, and SGRQ-C, was observed in rapid FEV1/FVC decliners during the 3-year follow-up. In addition, the annual change in FEV1 for rapid FEV1/FVC decliners was −71.0 mL/yr (IQR, −122.5 to −12.5), which deteriorated at a faster rate compared to non-rapid FEV1/FVC decliners with −7.0 mL/yr (IQR, −71.75 to 46.75). A higher proportion of rapid FEV1/FVC decliners experienced a decline in FEV1 ≥ 60 mL/yr or 100 mL/yr. However, 20% of rapid FEV1/FVC decliners had an annual FEV1 decline rate slower than the median value of −10 mL/yr, and notably, these patients commonly showed an increasing trend in annual FVC (Supplementary Fig. 4).

FEV1/FVC decline and treatment change

At the baseline, the rapid FEV1/FVC decliners predominantly were treated with triple therapy, with no significant differences in treatment compared to the non-rapid FEV1/FVC decliners. Over the 3-year tracking period of inhaler therapy usage, it was observed that the use of long-acting beta-agonist (LABA)/long-acting muscarinic antagonist (LAMA) in the non-rapid FEV1/FVC decliners increased over time, and the difference compared to the rapid FEV1/FVC decliners became more pronounced (P < 0.05) (Fig. 2, Supplementary Fig. 5). There was a tendency for more inhaled corticosteroid (ICS)/LABA use in the rapid FEV1/FVC decliners over the 3-year period, although a consistent statistical significance was not observed. Additionally, in the patients with BEC < 100 /uL, there was a significantly higher proportion of rapid FEV1/FVC decliners using ICS-containing medications at the 1-year follow-up. (Supplementary Fig. 6).

Risk factors related to rapid FEV1/FVC decline

In univariable logistic regression analysis, BMI, current smoker status, cardiovascular disease, post-BD FEV1 (%), STAR stage IV, FEF25-75%, and annual FEV1 change were significantly associated with rapid FEV1/FVC decline. Change of PROs (mMRC, CAT score, and SGRQ-C) and change to dual-BD were also found to be factors associated with the rapid FEV1/FVC decline (Table 4). In multivariable analyses, model 1 showed low BMI (adjusted odds ratio [aOR], 0.946; 95% confidence interval [CI], 0.909–0.984), current smoking (aOR, 1.850; 95% CI, 1.118–3.062), increased mMRC (aOR, 1.155; 95% CI, 1.037–1.286) over 3 years, low post-BD FEV1 (%) (aOR, 0.972; 95% CI, 0.959–0.985), low STAR stage (STAR stage IV: aOR, 0.361; 95% CI, 0.204–0.639), high FEF25-75% (%) (aOR, 1.021; 95% CI, 1.006–1.037), accelerated FEV1 decline (mL/yr) (aOR, 0.992; 95% CI, 0.991–0.993), not initiating dual BD therapy (aOR, 0.526; 95% CI, 0.322–0.859) were identified as independent risk factors for rapid FEV1/FVC decline. In Model 2, CAT score was included instead of mMRC, showing no significant association with rapid FEV1/FVC decline. In Model 3, SGRQ-C was included, revealing a significant association with rapid FEV1/FVC decline (aOR, 1.006; 95% CI, 1.001–1.012). Especially, in the subgroup patients who showed mMRC, CAT score, or SGRQ-C deterioration beyond the MCID over 3 years, the patients who continued using dual BDs exhibited a lower risk of rapid FEV1/FVC decline (aOR, 0.708; 95% CI, 0.508–0.987), while those who discontinued using dual BDs showed a higher risk of rapid FEV1/FVC decline (aOR, 1.526; 95% CI, 1.191–1.955) (Table 5).

DISCUSSION

In this observational study, we identified the risk factors associated with rapid FEV1/FVC decline in COPD patients. Our study showed that annual change of FEV1/FVC in COPD patients was 0%p/year, which was attenuated compared to −0.3%p/year in the general population.15 However, rapid FEV1/FVC decliners (1st quartile) in COPD patients exhibited a greater decline of −2.5%p/year compared to those in general population, which showed −0.8%p/year.15 Rapid FEV1/FVC decline was associated with worsening symptoms and accelerated FEV1 decline. The risk factors for rapid FEV1/FVC decline included low BMI, current smoking status, increased mMRC score, low post-BD FEV1, low STAR stage, high FEF25-75%, accelerated FEV1 decline, and not initiating dual BD therapy. Our study found that rapid FEV1/FVC decline was not associated with annual rate of acute exacerbation. This may be because if an exacerbation causes a reduction in both FEV1 and FVC, the patient would be classified as a non-rapid decliner.21 As a result, there would be no significant difference in exacerbation rates between rapid and non-rapid decliners. There was a significantly lower usage of LABA/LAMA in the rapid FEV1/FVC decliners during follow-up period, while there was no meaningful correlation with ICS usage.

The concept of pre-disease stages such as prediabetes and precancer has emerged in COPD to identify and intervene early-stage patients for disease modification, giving rise to the concepts of pre-COPD including preserved ratio impaired spirometry.22 Given that the decline in lung function is faster in early-stage disease than in severe cases,23 it is crucial to recognize high-risk groups prone to the rapid lung function deterioration at the pre-COPD or mild COPD stage. Similarly, in our study, rapid declines in FEV1/FVC were more frequently observed in less advanced stages of COPD, characterized by higher ppFEV1 and FEV1/FVC ratios. Interestingly, both in COPD patients and the general population, rapid FEV1/FVC decliners had higher baseline FEV1/FVC ratios and FEF25-75% values.15 This finding suggests that rapid FEV1/FVC decline may occur before the onset or progression of small airway disease. Therefore, identifying risk factors associated with rapid lung function decline before significant impairment occurs holds significant value. In a similar population, male sex, current smoking, BEC < 150 /µL, and high FVC were associated with an increased risk of rapid FEV1 decline.24 Among these risk factors, current smoking is a common modifiable factor for both rapid FEV1/FVC and FEV1 decline. Cigarette smoking significantly contributes to the development and progression of COPD.2526 Therefore, smoking cessation can have benefits in preserving FEV1/FVC decline rate and preventing the progression of COPD.

Recently, the COPDgene group introduced a new severity stage for COPD.11 In our study, the proportion of COPD patients classified as severe grade according to STAR stage (IV, FEV1/FVC < 40%) was unexpectedly lower among rapid FEV1/FVC decliners, suggesting a reduced risk of rapid FEV1/FVC decline in STAR stage IV. In COPDgene cohort findings,11 the greatest decline in FEV1 occurred in Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2 and STAR 3, with even less decline in stage 4 of both severity categories. Other recent cohort studies suggest that individuals with mild airflow obstruction experience a faster decline in FEV1 compared to those with more severe disease.2327 This phenomenon is explained by the idea that individuals with relatively preserved lung function have a greater potential for decline.

Reduced lung function is significantly associated with symptoms, quality of life, and exercise capacity in COPD patients.2 Rapid FEV1 decline and rapid FVC decline are associated with respiratory symptoms and an increased risk of acute exacerbations, garnering attention as prognostic factors for COPD progression.28293031 However, a single indicator such as FEV1 or FVC may be impacted by several factors such as lung volume and restrictive lung disease,79 beyond lung function, thereby leading to limitations. In addition, the decline in FEV1 can be influenced by a decrease in FVC. Previous studies have shown that FEV1 decline can be influenced by FVC (L) in COPD patients2432 and general population.33 On the other hand, our study and another previous study showed that FEV1/FVC decline rate was not affected by FVC (L) in COPD patients and in general population.15 Therefore, FEV1/FVC decline rate may provide a more reliable clinical indicator for assessing disease progression in COPD patients, particularly when FVC is impaired. Indeed, in a previous study, the decline in FEV1/FVC was more prominent in the emphysema-predominant group, while the decline in FEV1 and FVC in pre-COPD or GOLD 1 stage COPD was more pronounced in the airway-predominant group.34 Therefore, FEV1/FVC decline rate may serve as an indicator of the progression of parenchymal destruction in COPD patients. Considering the association of STAR stage with the prognosis of COPD patients,11 the decline in FEV1/FVC can be considered as an important indicator of COPD progression.

The previous Swedish general population study demonstrated a continuous decrease in the odds of any respiratory symptoms with an increase in FEV1/FVC.35 Similarly, SGRQ and mMRC deteriorated with the progression of STAR stage in COPDgene cohort.11 We also observed a significant increase in mMRC, CAT score, and SGRQ-C beyond MCID in rapid FEV1/FVC decliners. In addition to PROs, we observed a progressive disparity in the proportion of LABA/LAMA usage between rapid and non-rapid FEV1/FVC decliners throughout the 3-year observation period. There was a gradual and significant increase in LABA/LAMA usage among non-rapid FEV1/FVC decliners, and a transition to dual bronchodilators was associated with a potentially reduced risk of rapid FEV1/FVC decline. This suggests that dual bronchodilation, compared to monotherapy, significantly improves FEV1 36 and lowers the risk of FEV1/FVC decline. On the other hand, the impact of ICS on the annual change in FEV1/FVC was not clear. The proposed criteria for recommending ICS use, such as previous exacerbation history and an elevated BEC,37 did not differ between the two groups in baseline characteristics. However, within the BEC <100 group, there was a higher usage of ICS among rapid FEV1/FVC decliners. Therefore, it is speculated that both transitioning to dual bronchodilators and inappropriate use of ICS significantly contribute to the notable differences in FEV1/FVC decline rate. Our findings underscores the role of appropriate inhaled medication selection and adherence in attenuating the lung function decline.

Our study has several limitations. First, air trapping and emphysema can reduce FVC, potentially leading to changes in the FEV1/FVC ratio.38 Because of lack of longitudinal information for progression of small airway disease and parenchymal destruction, and therefore, we could not assess the impact of these structural changes on lung function decline. Second, there is a lack of information on comorbid lung disease39 and environmental40 or occupational exposures41 that may influence FEV1 or FVC, and these factors were not included in the analysis as potential risk factors. Third, some patients were excluded from the present analysis due to markedly aberrant changes observed in their FEV1/FVC ratios, which were considered statistically as outliers. Within the cohort dataset, the causes for these abrupt fluctuations could not be clarified. These outliers were likely attributed to various factors, such as suboptimal patient conditions during measurement, procedural errors during pulmonary function testing, or inaccuracies in data entry. Regardless of the underlying cause, the extreme variability in these measurements rendered them unreliable for inclusion in the analysis. Fourth, our cohort consisted of patients recruited from teaching hospitals, which might have led to a higher likelihood of enrolling high-risk patients compared to the general COPD population.

In conclusion, our study identified a significant correlation between rapid decline in post-BD FEV1/FVC and specific baseline characteristics, such as low BMI and current smoking, as well as dynamic changes in clinical parameters, including worsened symptoms, accelerated FEV1 decline, and non-compliance with dual BD therapy in COPD patients. Particularly, noteworthy is the observation that individuals who failed to consistently use LABA/LAMA despite experiencing worsening symptoms were more prevalent among rapid FEV1/FVC decliners.