Airway Mucus Plugs on Chest Computed Tomography Are Associated with Exacerbations in Chronic Obstructive Pulmonary Disease

Abstract

Rationale

Acute exacerbations (AEs) of chronic obstructive pulmonary disease (COPD) are associated with significant morbidity and mortality. Whether mucus plugs are associated with prospective exacerbations has not been examined extensively.

Objective

To examine associations between mucus plugs on chest computed tomography (CT) and future moderate-to-severe AEs in two independent cohorts with spirometrically-confirmed COPD.

Methods

Mucus plugs were visually identified on baseline chest computed tomography scans from smokers with Global Initiative for Chronic Obstructive Lung Disease grade 2–4 COPD enrolled in two multicenter cohort studies: ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints) and COPDGene (Genetic Epidemiology of COPD). Associations between ordinal mucus plug score categories (0, 1–2, and ≥3) and prospectively ascertained AEs, defined as worsening respiratory symptoms requiring systemic steroids and/or antibiotics (moderate to severe) and/or emergency room visit or hospitalization (severe), were assessed using multivariable-adjusted zero-inflated Poisson regression; subjects were exacerbation-free at enrollment.

Measurements and Main Results

Among 3,250 participants in COPDGene (mean age ± SD, 63.7 ± 8.4 yr; FEV₁, 50.6 ± 17.8% predicted; 45.1% female) and 1,716 participants in ECLIPSE (age, 63.3 ± 7.1 yr; FEV₁, 48.3 ± 15.8% predicted; 36.2% female), 44.4% and 46.0% had mucus plugs, respectively. The incidence rates of AEs were 61.0 (COPDGene) and 125.7 (ECLIPSE) per 100 person-years. Relative to those without mucus plugs, the presence of 1–2 and ≥3 mucus plugs was associated with increased risk (adjusted rate ratio [aRR], 1.07 [95% CI, 1.05–1.09] and 1.15 [1.1–1.2] in COPDGene; aRR, 1.06 [95% CI, 1.02–1.09] and 1.12 [1.04–1.2] in ECLIPSE, respectively) for prospective moderate to severe AEs. The presence of 1–2 and ≥3 mucus plugs was also associated with increased risk for severe AEs during follow-up (aRR, 1.05 [95% CI, 1.01–1.08] and 1.09 [1.02–1.18] in COPDGene; aRR, 1.17 [95% CI, 1.07–1.27] and 1.37 [1.15–1.62] in ECLIPSE, respectively).

Conclusions

Computed tomography–detected mucus plugs are associated with an increased risk for future COPD AEs.

At a Glance Commentary

Scientific Knowledge on the Subject

Acute exacerbations (AEs) are associated with increased morbidity and mortality in chronic obstructive pulmonary disease (COPD), yet relatively few consistent risk factors beyond worse airflow limitation and a personal history of prior AEs have been identified. Mucus plugs on chest computed tomography have been associated with worse spirometric lung function and quality of life and increased mortality in COPD, but their association with future AEs has been incompletely explored.

What This Study Adds to the Field

In two large, independent cohorts of individuals with Global Initiative for Obstructive Lung Disease spirometry grade 2–4 COPD, visually assessed mucus plugs on enrollment chest computed tomography were associated with a 5–15% increased risk of moderate to severe AEs and a 5–37% increased risk of severe AEs during follow-up.

Acute exacerbations (AEs) are significant events in the natural history of chronic obstructive pulmonary disease (COPD) (1, 2) that are associated with decline in lung function and increased mortality (1–3). AEs, defined as an increase in cough, phlegm, and/or dyspnea beyond that expected with day-to-day variation in symptoms, are frequently triggered by bacterial and viral infections and air pollution exposure. It has been postulated that occlusion of airways by mucus plugs may serve as a nidus of inflammation and infection (4), which can potentially trigger the occurrence of these clinical events. However, the hypothesis has not been comprehensively examined in persons with COPD.

Mucus plugs are an airway pathologic process in COPD that can be detected and quantified on computed tomography (CT) scans. CT-based mucus plugs are observed in 41–67% of patients with COPD and are associated with lower spirometric lung function, worse quality of life, and increased mortality (5–8). Because mucus plugs are potentially reversible with medical therapy (9–11), exploring the impact of airway-occluding mucus plugs on future AEs is of clinical relevance. A previous study in smokers (N = 400) demonstrated that individuals with high mucus plug burden on CT (five or more pulmonary segments with mucus plugs) had increased exacerbations in the year after enrollment relative to those with low mucus plug burden (i.e., fewer than five) (6); however, power in that analysis was limited, and the impact of potential confounders and subgroups could not be assessed. Given these knowledge gaps, we aimed to test the hypothesis that airway mucus plugs on chest CT are associated with future AEs of COPD by using data from two well-characterized cohorts: COPDGene (Genetic Epidemiology of COPD; clinicaltrials.gov identifier, NCT00608764) (12) and ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints; NCT00292552) (13). Some of the results have been previously reported as abstracts (14, 15).

Methods

Study Populations

COPDGene (12) is an ongoing multicenter observational prospective study that enrolled participants who self-identified as non-Hispanic Black and non-Hispanic White, were aged 45–80 years, and had a ⩾10–pack-year history of cigarette smoking. For the baseline visit, participants were enrolled across 21 U.S. centers between November 2007 and April 2011; additional inclusion and exclusion criteria have been previously published (12). Nonsmoking control subjects were recruited, but are not included in the present analysis.

ECLIPSE was designed as a 3-year, multicenter observational study that enrolled participants aged 40–75 years old from 12 countries who were current or former smokers with a ⩾10–pack-year history and Global Initiative for Chronic Obstructive Lung Disease (GOLD) grades 2–4 COPD on postbronchodilator spirometry (13). Smoking and nonsmoking control subjects were also recruited for ECLIPSE but are not included in the present analysis. The institutional review boards at the participating clinical centers for COPDGene and ECLIPSE approved the respective protocols, and all participants gave written informed consent (12, 13).

All participants completed questionnaires (respiratory symptoms, demographic characteristics, and medical history), pre- and postbronchodilator spirometry, and chest CT imaging at the enrollment visit. Spirometric measures of lung function were performed before and after the use of albuterol, according to prevailing guidelines at the time of enrollment (16). Postbronchodilator FEV₁ and FVC are expressed as percentages of predicted values (17). COPD was defined as FEV₁/FVC ratio <0.7 after the administration of albuterol, with severity classified based on spirometric GOLD grades 1–4 (18); subjects with GOLD grade 2–4 obstruction were included in the present analysis. A history of exacerbations in the year before study enrollment was self-reported and dichotomized (i.e., yes/no) in both cohorts.

Outcome

The primary outcome was the total number of AEs, defined as new or increased respiratory symptoms (cough, phlegm, dyspnea) that required treatment with systemic steroids and/or antibiotic therapy in the outpatient, emergency room, or inpatient setting over time. The secondary outcome was severe AEs, defined as AEs that required treatment in an emergency room or hospitalization. In the COPDGene cohort, AEs were assessed biannually through a combination of automated telephony and web-based questionnaires through the Longitudinal Follow-Up program (19). In the ECLIPSE cohort, AEs were assessed via monthly telephone calls and at in-person study visits (13). In both cohorts, AEs were self-reported and quantified based on responses to standardized questions (20, 21).

CT Assessments

The COPDGene and ECLIPSE imaging acquisition protocols have been previously published (12, 13); noncontrast chest CT scans at full inspiration were used. Quantitative CT measurements of emphysema and airways were performed with Thirona (COPDGene) and VIDA (ECLIPSE) software. Emphysema was defined as the percentage of low-attenuation areas (lower than −950 HU), whereas airway wall thickness was defined as the square root of wall area of a hypothetical airway with an inner diameter of 10 mm (Pi10).

Mucus plug assessment was performed on volumetric CT scans with a submillimeter slice thickness, providing a detailed visualization of the bronchial tree. Readers with at least 2 years of experience in lung imaging were trained to identify and score airway mucus plugs on baseline inspiratory CT scans using a window width of 1,400 and level of −500 HU while blinded to clinical information (8). We used a sequential reading system as previously described (8) whereby all CT scans were rated for airway mucus plugs by a first reader. All positive and 20% of negative CT images were then scored by a second reader. CT images with discrepant mucus plug scores were sent to a third reader (8). We used a scoring system based on the Netter’s bronchial anatomy nomenclature, employing 18 bronchopulmonary segments (22). A mucus plug was defined as an opacity that completely occluded the lumen of an airway, regardless of the airway size or generation (Figure 1). Medium- to large-sized airways (i.e., 2–10-mm lumen diameter) were surveyed (5). The lung zone within 2 cm from the costal or diaphragmatic pleura was excluded because the airways in that zone are too small to accurately ascertain a complete occlusion by luminal plugs. A score was generated for each CT scan as an aggregation of the number of pulmonary segments with mucus plugs, ranging from 0 to 18, with scans without mucus plugs scored as 0. For CT scans with more than one reading, the scores from two or three readers were averaged (5). The interreader agreement for mucus plug score was 0.77 (5). See online supplement for additional details.

Figure 1.

Example of a mucus plug on chest computed tomography imaging. Computed tomography axial (A) and sagittal (B) sections showing a mucus plug (encircled) occluding the lumen of an airway in the left lower lobe of a COPDGene (Genetic Epidemiology of COPD) participant with COPD. COPD = chronic obstructive pulmonary disease.

Statistical Analysis

Data are presented as mean ± standard deviation or numbers and proportions (percentages) for continuous and categorical variables, respectively. The incidence rate of AEs in each cohort was calculated as the sum of AEs divided by the sum of the duration of follow-up for each subject at risk. The association between mucus plug score and total number of exacerbations was assessed with zero-inflated Poisson regression analysis with log-link function and log of follow-up time as the offset; results are expressed as the adjusted rate ratio (aRR). Each model was adjusted for covariates that were determined a priori: age, sex, race, body mass index, pack-years of smoking, smoking status (current vs. former), history of exacerbations in the year before enrollment (yes/no), postbronchodilator FEV₁% predicted, clinical center/country, and CT measures of airway wall thickness and emphysema; FEV₁% predicted was also included in the zero component of the model. Separate models for the COPDGene and ECLIPSE cohorts were constructed. Directed acyclic graphs illustrating the relationship between covariates and the primary predictor (mucus plug score) and outcome (prospective AEs) are shown in Figure E1 in the online supplement. Because the overall missingness for both analysis cohorts was <2%, all models were based on complete case analysis.

Primary analyses were performed with ordinal mucus plug score groups (score of 0, 1–2, or ≥3); secondary analyses examined continuous mucus plug score and dichotomized mucus plug score (0 vs. >0). Sensitivity analyses were performed including additional adjustment for chronic bronchitis (yes/no) and excluding adjustment for radiographic variables (percent emphysema and Pi10) and adjusting for absolute FEV₁ (in liters) instead of FEV₁% predicted. Subgroup analyses by clinically relevant features (e.g., sex, smoking status, obesity, GOLD spirometry grade, and history of AEs in the 1 year before enrollment) were also performed using the dichotomized mucus plug score (0 vs. >0) and the primary zero-inflated Poisson model outlined above. All analyses were conducted using R software (version 4.2.3). A two-sided P value <0.05 was considered significant.

Results

Cohort Characteristics

Flow diagrams for participants from each cohort retained for analysis are shown in Figure 2. Of the 10,198 ever-smoking participants in the COPDGene cohort, 3,696 had GOLD grade 2–4 COPD. After exclusion of participants with missing or poor-quality CT scans and/or those with missing exacerbation data, 3,250 participants were included for analysis. Among 2,163 subjects in ECLIPSE with GOLD grade 2–4 COPD on postbronchodilator spirometry, 446 were excluded because of inadequate imaging quality for mucus plug assessment or missing outcome data. The baseline characteristics of participants included for analysis in each cohort are shown in Table 1. Subjects excluded from analysis were more likely to be male; in the COPDGene cohort, excluded subjects were younger, more likely to be of non-European ancestry, and included a higher proportion of current smokers (Table E1).

Figure 2.

Flow chart of subjects included in (A) COPDGene (Genetic Epidemiology of COPD) and (B) ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints). AE = acute exacerbation; COPD = chronic obstructive pulmonary disease; CT = computed tomography; GOLD = Global Initiative for Chronic Obstructive Lung Disease.

Table 1.

Baseline Characteristics by Cohort

Characteristic	COPDGene (N = 3,250)	ECLIPSE (N = 1,716)
Age, yr	63.7 ± 8.4	63.3 ± 7.1
Sex
Male	1,784 (54.9%)	1,094 (63.7%)
Female	1,466 (45.1%)	622 (36.2%)
European ancestry	2,580 (79.4%)	1,680 (97.9%)
BMI	28.2 ± 6.2	26.4 ± 5.6
Current smoker	1,245 (38.3%)	638 (37.2%)
Pack-years of smoking	52.9 ± 27.2	49.1 ± 27.3
Exacerbation in 1 yr before enrollment	1,257 (38.7%)	782 (45.6%)
Chronic bronchitis	894 (27.5%)	560 (32.6%)
FEV1, L	1.45 ± 0.62	1.34 ± 0.52
FEV1, % predicted	50.6 ± 17.8	48.3 ± 15.8
Percent emphysema on CT*	13.5 ± 12.9	17.8 ± 12.1
Pi10†	2.73 ± 0.56	3.95 ± 0.20
Years of follow-up	8.1 ± 4.3	2.7 ± 0.7
Total prospective acute exacerbations	4.9 ± 7.4	3.4 ± 3.8
Mucus plug score	1.4 ± 2.5	1.6 ± 2.6

Definition of abbreviations: BMI = body mass index; COPDGene = Genetic Epidemiology of COPD; COPD = chronic obstructive pulmonary disease; CT = computed tomography; ECLIPSE = Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints; Pi10 = square root of wall area of a hypothetical airway with an inner diameter of 10 mm.

Data are presented as mean ± SD where applicable.

^*n = 3,100 in COPDGene; n = 1,547 in ECLIPSE.

^† n = 3,100 in COPDGene; n = 1,555 in ECLIPSE.

Association between Mucus Plug Score and AEs

The prevalences of mucus plugs were 44.4% in COPDGene and 46.0% in ECLIPSE; the distribution of mucus plug score by cohort, GOLD spirometry grade, and FEV₁% predicted are shown in Figures E2–E4. The prevalence of a mucus plug score >0 increased with GOLD spirometry grade (i.e., worse airflow obstruction; see Figure E3). The prevalence of a mucus plug score >0 was also higher among those with a history of AEs in the year before enrollment (50.1% vs. 40.7% in COPDGene and 59.1% vs. 41.1% in ECLIPSE). During a median follow-up of 8.6 (IQR, 7.8) and 3 (0.1) years, 2,166 (66.6%) and 1,315 (76.6%) of participants in COPDGene and ECLIPSE, respectively, experienced at least one moderate to severe AE. In the primary analysis of ordinal mucus plug groups (zero, 1–2, and ≥3), a significantly increased risk for future moderate to severe AEs was observed with a higher mucus plug score in both cohorts, with aRRs of 1.06–1.07 (for scores of 1–2 vs. 0) and aRRs of 1.12–1.15 (for scores of ⩾3 vs. 0) (Table 2).

Table 2.

Association between Ordinal Mucus Plug Score Groups and AEs

	COPDGene			ECLIPSE
	0	1–2	⩾3	0	1–2	⩾3
No. of patients	1,808	753	689	926	373	417
Moderate to severe AEs	Ref.	1.070 (1.048–1.093)	1.145 (1.098–1.195)	Ref.	1.056 (1.019–1.094)	1.115 (1.039–1.197)
Severe AEs	Ref.	1.045 (1.008–1.084)	1.092 (1.016–1.175)	Ref.	1.169 (1.072–1.273)	1.365 (1.15–1.622)

Definition of abbreviations: AE = acute exacerbation; COPDGene = Genetic Epidemiology of COPD; COPD = chronic obstructive pulmonary disease; ECLIPSE = Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints.

Data are presented as adjusted rate ratios (95% confidence intervals). Moderate to severe AEs were considered as increased respiratory symptoms requiring systemic steroids and/or antibiotic treatment; severe AEs were those that required an emergency room visit and/or hospitalization. The model used was a zero-inflated Poisson model with ordinal mucus plug categories (primary predictor) and covariate adjustment for age at enrollment, sex, race, body mass index, current smoking, pack-years of smoking, FEV₁% predicted, history of AE in the 12 months before enrollment, percent emphysema, square root of wall area of a hypothetical airway with an inner diameter of 10 mm, and clinical center/country. Log follow-up time was included as the offset. The zero component was adjusted for FEV₁% predicted.

During the same follow-up period, 1,499 (46.1%) and 540 (31.5%) participants in COPDGene and ECLIPSE, respectively, experienced one or more severe exacerbation, defined as an AE requiring emergency room treatment and/or hospitalization. Similarly, a significantly increased risk for future severe AEs was observed with a higher ordinal mucus plug score, with aRRs of 1.05–1.17 (for scores of 1–2 vs. 0) and 1.09–1.37 (for scores of ⩾3 vs. 0) (Table 2).

Secondary Analyses

Analysis of mucus plug score as a continuous predictor demonstrated a significant association between higher score and increased risk for future AEs, with an aRR of 1.03 (in both COPDGene and ECLIPSE) for moderate to severe AEs and an aRR range of 1.02–1.05 for severe exacerbations for each additional bronchopulmonary segment with mucus plugs (Table E2). Analysis of dichotomized mucus plug score (0 vs. >0) demonstrated an 8–11% increased risk for moderate to severe AEs and a 7–41% increased risk for severe AEs among those with a mucus plug score >0 relative to those with no mucus plugs.

Sensitivity Analyses

Associations between AEs and ordinal mucus plug categories were similar in sensitivity analyses with additional adjustment for chronic bronchitis (Table E3), excluding adjustment for percent emphysema and Pi10 (Table E4), and adjusting for absolute FEV₁ (in liters) instead of FEV₁% predicted (Tables E5 and E6).

Subgroup Analyses

Subgroup analyses by sex, obesity (body mass index ⩾30 vs. <30 kg/m²), current versus former smoking status, history of AEs in the year before enrollment (yes vs. no), and GOLD spirometry grade 2 versus grade 3/4 in each cohort, using the dichotomized mucus plug score (0 vs. >0) as the primary predictor, are shown in Figure 3. An increased risk for future AEs among former versus current smokers with mucus plugs was observed in both cohorts. In contrast to COPDGene, in ECLIPSE, a higher risk for prospective AEs was observed among men, nonobese subjects, those with a history of AEs in the year before enrollment, and GOLD grade 2 (relative to grade 3/4) spirometry.

Figure 3.

Subgroup analyses in (A) COPDGene (Genetic Epidemiology of COPD) and (B) ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints). Subgroup analyses by sex, obesity, current/former smoking, history of acute exacerbation (AE) in the year before enrollment, and Global Initiative for Chronic Obstructive Lung Disease (GOLD) spirometry grade 2 (moderate) versus 3 and 4 (severe and very severe) obstruction. The model used was a zero-inflated Poisson model with dichotomized mucus plug score (0 vs. >0) as the primary predictor and the number of moderate to severe AEs as the outcome, with adjustment for age, sex*, non-European ancestry, body mass index (BMI)*, current/former smoking*, pack-years of smoking history, FEV₁% predicted, history of AEs in the 12 months before enrollment*, percent emphysema, square root of wall area of a hypothetical airway with an inner diameter of 10 mm, and clinical center/country. Log follow-up time was used as the offset. Zero component adjusted for FEV₁% predicted. Asterisks indicate covariates that were excluded if this is the subgroup being tested (e.g., for subgroup analyses by sex, sex would be excluded from the primary model).

Discussion

In our study, based on data from two independent cohorts with rigorously ascertained baseline demographic and radiographic data, as well as prospective exacerbation data, CT-based mucus plugs on imaging at enrollment were associated with increased moderate to severe and severe AEs of COPD during follow-up. Our results remained robust on multiple sensitivity and secondary analyses and support the viability of CT-detected mucus plugs as a novel, potentially treatable trait in individuals with COPD (23).

AEs, particularly severe AEs, contribute to morbidity and mortality and account for a significant proportion of the direct costs associated with the care of individuals with COPD (1). However, despite substantial effort to identify determinants of AE susceptibility, relatively few consistent risk factors beyond worse spirometric lung function (i.e., lower FEV₁%) and a personal history of prior AEs have been identified (24, 25). Notably, the associations observed in our analyses remained robust after adjusting for multiple established risk factors, including age, sex, FEV₁% predicted, smoking status, and a history of prior AEs, as well as quantitative CT metrics of COPD (i.e., percent emphysema and airway wall thickness); our results support that CT-detected mucus plugs represent an independent risk factor for AEs. Given that medical therapy for COPD has historically relied on bronchodilators and corticosteroids, exploring the efficacy of enhancing mucus clearance, whether by pharmacological (9, 10) or mechanical strategies (11, 26), among individuals with COPD and mucus plugs on chest CT may be warranted.

Early work in lung tissue demonstrated that increased airway wall thickness and occlusion of small airways (i.e., <2-mm lumen diameter) by mucus on histology was associated with lower lung function and increased risk of mortality in persons with severe COPD (27, 28). Our findings build on those studies by demonstrating that the occlusion of medium to large airways by mucus plugs, which are visually assessable using standard CT imaging techniques (8), is associated with an increased number of exacerbations of COPD over time. The increased risk for future AEs observed with continuous mucus plug score and higher ordinal mucus plug score group (zero, one or two, three or more) support a possible dose–response relationship; thus, CT-based mucus plugs, beyond serving as a biomarker for increased AE risk, may directly contribute to the pathogenesis of AEs.

Normal mucus production and mucociliary clearance are critical protective processes in the respiratory tract; perturbations in the composition or transport of mucus have been postulated to result in abnormal sputum production and mucus plug formation (4). Although total mucus plug scores have demonstrated stability over time in selected subgroups (6), individual mucus plug formation is likely dynamic and may serve as a marker for increased susceptibility to infections. Conversely, pathophysiological processes that may directly contribute to an increased susceptibility to AEs include mechanical obstruction to laminar air flow and the creation of regional hypoxic environments, which may facilitate local inflammation and airway dysbiosis (4, 29). Further studies are warranted to examine the role of mucus pathology in the cascade of events leading to exacerbations in people with COPD.

Although the association between CT-detected mucus plugs and increased risk of future AEs was consistent in the COPDGene and ECLIPSE cohorts, differences in effect estimates between cohorts, as well as in subgroup analyses, were observed. For example, differential risks for AEs by sex, obesity, history of prior AEs, and GOLD spirometry grade group were observed in ECLIPSE but not in COPDGene. These may be attributable to differences in the size of each cohort, populations enrolled (e.g., United States–based vs. multinational, proportion of male and female patients, proportion with non-European ancestry), or outcomes assessment methodology (e.g., monthly vs. twice-yearly reporting of AEs). Similarly, in contrast to the relatively consistent risk estimates for moderate to severe AEs, the aRRs for severe exacerbations were consistently higher in the ECLIPSE cohort. Given that the rate of severe AEs may vary over time (30), as well as the shorter duration of follow-up in ECLIPSE, whether CT-detected mucus plugs are associated with differential AE risk in short- versus long-term follow-up should be explored in independent cohorts. Similarly, whether longitudinal change in mucus plug location or overall burden is associated with differential COPD AE risk remains to be explored.

The strengths of our study include the use of data from two large independent, well-characterized COPD cohorts and robust findings after adjustment for established risk factors for AEs. Despite this, we acknowledge the following limitations. First, as a result of the observational nature of both cohorts, conclusions regarding the causality of airway mucus plugs and COPD exacerbations cannot be established. Second, even though approximately 20% of the COPDGene cohort was non-Hispanic Black, the majority of individuals in both cohorts were of European ancestry, which may limit the generalizability of our findings to other racial and ethnic groups. Third, both cohorts comprised individuals with a ⩾10–pack-year smoking history and GOLD spirometry grade 2–4 COPD; extrapolation of our findings to individuals with pre-COPD, mild COPD (i.e., GOLD grade 1), or smokers with a history of <10 pack-years should be performed with caution (31). Future studies of CT-detected mucus plug burden and AEs in these subgroups are needed (32). Fourth, the visually assessed anatomical lung segment–based score may have underestimated the true burden of mucus plugs, and the interreader agreement was moderate; we assert that both of these would have likely resulted in a bias toward the null, although future studies quantifying regional burden, as well as the use of automated algorithms for mucus plug detection, will be informative. Fifth, even though we adjusted for multiple known covariates, residual confounding relevant to the association between mucus plugs and AEs may exist. We performed a post hoc E-value analysis (33), which supports that the observed aRRs reported in our primary analysis could be explained by unmeasured confounders associated with mucus plug score and AEs with rate ratios ranging from 1.26 to 2.07 (Table E7). Despite these limitations, we contend that CT-based mucus plugs represent a novel predictor of clinically significant AEs among individuals with COPD; future studies to explore the pathophysiology underlying this association and potential therapeutic strategies are needed.