A Quantitative Lung Mucin Score to Identify Chronic Bronchitis

Abstract

BACKGROUND

We previously demonstrated that sputum total mucin concentration is an objective marker for chronic bronchitis (CB). This current study introduces a novel Mucin Quantitative Score (MUCQ) that combines total mucin concentration and mucin composition to improve the assessment of risk, onset of clinically diagnosed disease, and progression of muco-obstructive lung diseases.

METHODS

Patients from the SPIROMICS (SubPopulations and InteRmediate Outcome Measures in COPD Study) cohort were classified as having CB, or not, based on clinical questionnaires. Using the measured total mucin, MUC5AC, and MUC5B concentrations in sputum samples, we calculated MUCQ as [Total mucin]×([MUC5AC]÷[MUC5B])÷100 μg/ml, which is a unitless, weighted concentration score. Our primary outcome was the net reclassification of patients with a diagnosis of CB, or not, based on total mucin concentrations in their sputum compared with using the MUCQ score. Participants were first classified as CB-positive or -negative using a total mucin concentration threshold of 2306 μg/ml, then reclassified using the MUCQ threshold of 4.30. Associated z statistics and a P value for the primary outcome are reported.

RESULTS

Among 164 patients in the SPIROMICS cohort with clinically defined CB, using the MUCQ score up-classified 18 patients who were currently smoking to a diagnosis of CB and down-classified 5 patients who were currently smoking and 3 control participants who had never smoked, compared with the classification of CB was based on total mucin concentrations alone (P=0.001). In addition, MUCQ correlated with other clinical and pathological indices of chronic airway disease and airway obstruction.

CONCLUSIONS

The MUCQ metric was superior in distinguishing patients with CB compared to a total mucin concentration. Trials are needed to ascertain the prospective use of MUCQ metrics in research and clinical settings for assessment, management, and tracking therapeutic responses in CB and potentially other muco-obstructive conditions. (Funded by the National Institutes of Health and others.)

Introduction

Chronic muco-obstructive lung diseases, characterized by airway obstruction, persistent and productive cough, and infection and inflammation, affect more than 45 million clinically diagnosed individuals in the United States and many more who are undiagnosed or at risk.¹ Chronic bronchitis (CB), a subtype of chronic obstructive pulmonary disease (COPD), affects 10 million people in the United States alone.² The diagnosis of CB currently relies on patients’ reports of cough and sputum history.³ Notably, there are few biological metrics to gauge the risk of developing muco-obstructive lung disease, monitor the course of disease, or predict the effectiveness of the therapeutics targeting mucus abnormalities.

We have previously shown that total mucin concentrations increase in CB and are closely related to all manifestations of COPD, including small airways dysfunction and future exacerbation frequencies.⁴ We have also shown that both the concentration and the type of mucins, especially MUC5AC, are closely associated with the progression and symptoms of muco-obstructive lung diseases, including CB.⁵

Total mucin concentrations describe the quantity of the large polymers that dominate the osmotic pressure of the mucus layer and govern the mucus transport on airway surfaces.^6,7 Two major secreted mucins, MUC5AC and MUC5B, contribute to total airway mucin concentrations, and each mucin exhibits a distinct biology.⁸ Accordingly, measurement of each mucin provides additional information describing the state of the lung in health and disease. Our previous studies indicated that MUC5B is the dominant mucin in induced sputum from healthy individuals and its concentration increases two- to eightfold in muco-obstructive lung diseases. In contrast, MUC5AC is the minor mucin in the sputum of healthy participants, but it has fractional increases that are disproportionate to those of MUC5B (e.g., 10- to 30-fold) in bronchitic and bronchiectatic diseases.^5,9,10

Building on these findings, we developed a novel quantitative metric, termed the MUCin Quantitative Score (MUCQ), that combines total airway mucin concentrations with the ratio of the two dominant airway mucins, MUC5AC and MUC5B, combining mucin concentration and mucin composition. The primary goal of this study was to evaluate whether the aggregate MUCQ score improved diagnostic classification of CB compared to total mucin scores alone. We also investigated the relationship between MUCQ and other measures commonly assessed in people with obstructive airway disease.

Materials and Methods

STUDY DESIGN AND COHORTS

MUCQ score calculations were performed using mucin data from the SPIROMICS (SubPopulations and InteRmediate Outcome Measures in COPD Study) observational study of 2,981 people followed for more than 10 years with medical history data, sputum collection, imaging studies, and lung function tests.¹¹ Induced sputum was collected from 1011 participants at the baseline visit, and 917 sputum samples were used in our previous total mucin concentration study.⁴ Of these samples, 331 sputum samples were available for measurement of MUC5AC and MUC5B concentrations and study cohort analyses.⁵ A Consolidated Standards Of Reporting Trials (CONSORT) diagram describing SPIROMICS sputum sample acquisition and participant measurements is shown in Figure 1. The population used for MUCQ score calculation constitutes the convenience subset of 331 participants from the SPIROMICS cohort (Table S1 in the Supplementary Appendix) for whom adequate sputum specimens were available. Information on the data, and how they were collected, is available in SPIROMICS publications.^4,11 Based on our prior work with this population,⁴ we calculated a total mucin “cutoff” for the diagnosis of CB of 2306 μg/ml.

Figure 1.

Participant Flow Diagram for Total Mucin and MUC5AC and MUC5B Concentrations and MUCQ Score Calculations.

MUCQ denotes Mucin Quantitative Score; NEJM, New England Journal of Medicine; RM, Respiratory Medicine; SEC-MALLS/dRI, size-exclusion chromatography coupled with multiangle light scattering and differential refractometry; and UNC, University of North Carolina.

The clinical diagnosis of CB was assigned based on the response to a questionnaire administered to participants that reflected the definition of the disorder at the time of sample acquisition (section II in the Supplementary Appendix). Among the participants for whom a sputum sample was available, an assignment of the levels of their symptomatic involvement was based on their score on the COPD Assessment Test (CAT range 0–40, with higher scores indicating more severe symptoms; a change of 2 points is considered the minimally clinically important difference [MCID];¹² scores of 10 or greater are consistent with clinically manifest symptoms) or the St. George’s Respiratory Questionnaire (SGRQ range 0–100, with higher scores indicating worse respiratory health, with a change of 4 points considered to be the MCID¹³; an SGRQ score of 25 or higher is consistent with clinically manifest disease).¹⁴

A table of representativeness comparing the SPIROMICS mucin substudy CB+ population with the entire SPIROMICS CB+ population and with other CB cohorts is provided in section III-1 of the Supplementary Appendix. A description of the demographic questions and methods used to collect age, sex/gender, race, and ethnicity is also included in that section.

MEASUREMENTS OF MUCQ SCORE COMPONENTS AND CALCULATION

Using the induced sputum samples collected at SPIROMICS visit 1 (baseline), total mucin concentrations were measured using size-exclusion chromatography coupled with multiangle light scattering and differential refractometry (SEC-MALLS-dRI),^4,15 and MUC5AC and MUC5B concentrations were measured using labeled mass spectrometry.^4,5 Mucin concentrations were measured once per participant, and no repeated sputum measurements were obtained on the same individuals. Participants’ MUCQ indices were calculated as MUCQ=[Total mucin] μg/ml×([MUC5AC (pmol/ml)]÷[MUC5B (pmol/ml)])÷100 μg/ml. Although this formulation formally carries the same units as total mucin concentration (μg/ml), the÷100 μg/ml scaling step normalizes the metric into a relative, weighted/adjusted concentration and for simplicity and interpretive consistency.

PRIMARY OUTCOME

The primary outcome of this study was to assess whether the MUCQ score provides superior discrimination between health and airway disease (CB) in current smokers as compared to total mucin concentration. Other data are presented without hypothesis testing to show relationships among MUCQ and common measures of muco-obstructive disease impact.

STATISTICAL ANALYSES

MUCQ scores were log₁₀-transformed for statistical analysis for linear modeling, and analysis of variance (ANOVA) and multivariate analyses were used to reduce skewness. Smoking status was categorized as current, former, or never-smokers based on self-reported information. Associations between MUCQ and clinical variables were evaluated using linear regression and ANOVA. Due to the high variability in the data and weak linear associations, participants were divided into terciles of raw MUCQ values to identify trends (MUCQ: low ≤0.95, mid 0.96–5.71, high >5.71). Because terciles were treated as ordinal variables, linear trend analyses were conducted using GraphPad Prism (version 10.3.1). Sensitivity analyses using linear models were performed and detailed in section III-3 in the Supplementary Appendix. Results are illustrated using scatter plots with all data points shown. Missing values were considered to have occurred completely at random. Multivariable analyses were performed using JMP 17 Pro software with statistical significance set at α=0.05. The model included age, sex, race, ethnicity, body-mass index (BMI), smoking status, smoking-pack history, and asthma status as covariates. Data are presented as mean±95% confidence intervals (CIs). Also, mean±standard deviation values are reported in section IV-2 and Fig. S2 in the Supplementary Appendix.

To evaluate the primary outcome we calculated Net Reclassification Improvement (NRI)^16–18 comparing total mucin concentrations to MUCQ indices as our primary analytical and statistical tool. NRI quantifies the net proportion of individuals correctly reclassified, upward for patients with CB and downward for non-CB control participants, when using MUCQ instead of total mucin. In this analysis, the ground reference for CB diagnosis was based on questionnaire-defined clinical CB (section II in the Supplementary Appendix). Participants were first classified as CB- positive or -negative using a total mucin concentration threshold of 2306 μg/ml, then reclassified using the MUCQ threshold of 4.30. Associated z statistics and a P value for the primary outcome were reported.

Receiver operating characteristic (ROC) curves were generated using JMP Pro 17 for MUCQ, total mucin concentration, and the MUC5AC/MUC5B ratios. The area under the curve (AUC) was used to quantify diagnostic performance, including specificity and sensitivity. AUC comparisons between MUCQ and total mucin were performed using the DeLong method in MATLAB.¹⁹

Correlations of MUCQ scores with CB and COPD status and symptoms, COPD impact and severity, lung function, small airway dysfunction, smoking duration and status, and exacerbations noted after MUCQ data accrual were also investigated. Because there was no control for multiple comparisons, a P value is only given for the primary outcome; confidence intervals have not been adjusted for multiplicity and should not be used for clinical decision making. See Supplementary Methods for additional participants details and analytical information including statistical analyses in sections II and III in the Supplementary Appendix.

Results

SPIROMICS participants were categorized into three major groups, each with relevant subgroups. Healthy controls were never-smokers without respiratory symptoms. The at-risk group included participants who were smoking currently, or not, without spirometric airflow obstruction, further subdivided into asymptomatic (SGRQ <25) and symptomatic (SGRQ ≥25, those reporting cough, phlegm, or other respiratory symptoms) groups. The COPD group included participants who were smoking currently or not with spirometric evidence of airflow obstruction (COPD), categorized by Global Initiative for Chronic Obstructive Lung Disease (GOLD) stages GOLD 1 (mild), GOLD 2 (moderate), and GOLD 3 (severe) (see Table 1 and Supplementary Appendix, section I).

Table 1.

Study Population Characteristics.^*

Characteristic	Healthy Never-Smoker (n=40)	At-Risk (n=89)	GOLD 1 (n=59)	GOLD 2 (n=70)	GOLD 3 (n=73)
Age — yr	58.7±9.94	60.3±10.23	65.3±8.43	63.8±8.33	66.5±7.63
Male sex — no. (%)	21 (52.5)	42 (46.6)	40 (67.8)	43 (61.4)	39 (54.16)
Current smokers — no. (%)	0 (0)	52 (57.7)	28 (47.4)	32 (45.7)	23 (31.9)
Emphysema — no. (%)	2 (5)	12 (13.3)	17 (28.8)	24 (34.3)	54 (75.0)
CB — no (%)*	2 (5)	39 (43.8)	31 (52.5)	51 (72.8)	41 (56.2)
Current asthma — no. (%)	0 (0)	10 (11.1)	8 (13.5)	12 (17.1)	7 (9.7)
FEV1%	101.1±1.93	96.94±1.48	90.02±1.27	65.74±1.10	43.46±0.47
FEV1/FVC	0.80±0.04	0.78±0.05	0.63±0.05	0.56±0.08	0.42±0.08
CAT score†	2.00±5.00	10.00±11.00	10.50±12.0	15.50±13.75	15.00±9.50
SGRQ score†	4.1±6.00	15.7±16.4	15.7±15.7	23.95±18.5	26.8±17.2

^*See section I of the Supplementary Appendix for detailed information and definitions. Chronic bronchitis (CB) diagnosed with either classic or Saint George’s Respiratory Questionnaires (SGRQ). FEV₁ denotes forced expiratory volume in 1 second; FVC, forced vital capacity; and GOLD, Global Initiative for Chronic Obstructive Lung Disease.

^†For COPD Assessment Test (CAT) (range 0–40) and SGRQ (range 0–100) scores, higher scores indicate more severe symptoms; a change of 2 points for CAT, and 4 points for SGRQ is considered the minimally clinically important difference.^12,13,14

PRIMARY OUTCOME

MUCQ Discrimination Between Healthy People and Patients with Airway Disease Compared with Measures of Total Mucin and MUC5AC/MUC5B

We generated ROC curves to evaluate sensitivity and specificity and to define the optimal diagnostic threshold value for MUCQ. Using the Youden J statistics applied to the ROC curve, comparing healthy never-smokers to ever-smokers with CB, we identified 4.30 as the MUCQ “cutoff.” The AUC was 0.81 (95% CI, 0.72–0.89) for MUCQ, 0.79 (95% CI, 0.68–0.84) for total mucin, and 0.75 (95% CI, 0.63–0.83) for the MUC5AC/MUC5B ratio. For comparisons of current smokers with CB, AUCs were 0.87 (95% CI, 0.80–0.92) for MUCQ and 0.81 (95% CI, 0.74–0.89) for both total mucin and MUC5AC/MUC5B ratio (Fig. S1). At diagnostic thresholds for CB, specificity was 97% for both MUCQ comparisons of healthy never-smokers versus ever-smokers with CB and healthy never-smokers versus current smokers with CB and 91% for total mucin in the same comparisons. Sensitivity was 51% and 63% for MUCQ for ever-smokers and current smokers, respectively, and 48% in both groups for total mucin (Fig. S1).

We first used a value of 2306 μg/ml for total mucin concentration to divide patients into CB, or not, and compared this classification with the clinical, questionnaire-based classification. This was followed by classification into CB, or not, using the 4.30 MUCQ score and then compared the relative ability of the two methods to classify people compared to their clinical, questionnaire-based, CB classification using NRI.

Among 94 current smokers with clinically defined CB, MUCQ reclassified 18 patients upward to CB and 5 patients downward to non-CB; among never-smokers healthy controls MUCQ reclassified 3 healthy controls downward to non-CB. This yielded a net gain of 16 correctly classified current smokers with CB, corresponding to an NRI of 0.22 (22%) compared with using total mucin concentrations alone (z=2.57, P=0.001) (Fig. 2). Among 164 ever-smokers, MUCQ reclassified 21 patients upward to CB and 15 downward to non-CB, yielding a net gain of 9 (NRI: 0.12, 12%) correctly classified ever-smokers with CB compared to using total mucin concentrations alone (Fig. 2).

Figure 2.

Net Reclassification Improvement (NRI) of Chronic Bronchitis (CB) Using the MUCQ Score Compared with Total Mucin Concentration.

(Panel A) Bars show the number of individuals reclassified when MUCQ score is used instead of total mucin concentration, using the questionnaire-defined CB status as the reference. Reclassifications that improve agreement with questionnaire-defined CB status contribute positively to the NRI and are shown as positive bars, whereas reclassifications that do not improve agreement contribute negatively and are shown as negative bars. The net reclassified (gain) reflects the overall improvement in classification accuracy of MUCQ relative to total mucin concentration. Net reclassified=(UP_CB+−DOWN_CB+)+(DOWN_HC−UP_HC). (Panel B) The inset table shows participants with CB and without CB by reclassification direction. NRI, Z-score, and associated P-values are presented. HC=Healthy Controls (never-smokers with no CB, no emphysema); “Reclassified Up”=participants correctly reclassified as CB+ by MUCQ. “Reclassified Down”=participants reclassified as CB−by MUCQ. *The threshold used for MUCQ is 4.30 and for total mucin concentration is 2306 μg/ml. NRI=(UP_CB+−DOWN_CB+)/N_CB++(DOWN_HC−UP_HC)/N_HC.

THE RELATIONSHIP BETWEEN MUCQ AND OTHER CLINICAL OUTCOMES

Associations Between MUCQ Score and GOLD Category

We next tested the relationship between MUCQ scores and disease severity in the SPIROMICS dataset. The mean MUCQ score was 1.8 (95% CI, 0.9–2.5) in participants who had never smoked and 7.5 (95% CI, 4.2–10.3) in the at-risk group (ever-smokers without airflow obstruction) (Fig. 3A). The mean MUCQ score was 19.8 (95% CI, 12.8–26.7) in the COPD group, that is, all GOLD categories considered together (Fig. 3A). Among COPD GOLD categories, the mean MUCQ scores were 11.8 (95% CI, 5.2–18.5) in mild COPD (GOLD 1), 26.5 (95% CI, 15.0–38.0) in moderate COPD (GOLD 2), and 17.4 (95% CI, 6.2–28.5) in severe COPD (GOLD 3) (Fig. S2A). In the ordinal regression of COPD severity cohorts (healthy→at-risk→GOLD 1–3), the likelihood ratio χ² and the index of variation, RSquare (U), was 22.2 and 0.036, respectively for MUCQ, 11.2 for total mucin concentration, and 5.3 for the MUC5AC/MUC5B, whereas the index of variation, RSquare (U), was 0.039 (Fig. S3).

Figure 3.

MUCQ Scores across Clinical Groups and Relationship with Lung Function and Future Exacerbations.

(Panel A) MUCQ in healthy never-smokers (NS, N=40), at-risk people with a history of smoking at any time without airflow obstruction (N=89), and COPD (GOLD 1–3; N=202). Mean MUCQ (95% confidence interval [CI]): 1.8 (0.9–2.5), 7.5 (4.2–10.3), and 19.8 (12.8–26.7), respectively. (Panels B and C) The association between spirometrically defined lung function (FEV₁% predicted, FEF_25/75% predicted) and MUCQ scores. Terciles of raw MUCQ (low=≤0.95, N=109; mid=0.96–5.71, N=106; and high=>5.71, N=115) in SPIROMICS participants. FEV₁ % predicted (B): Mean (95% CI): 81.9 (77.1–86.7), 80.1 (75.5–84.6), 72.2 (67.9–76.4) for low, mid, and high MUCQ, respectively. FEF_25/75% predicted (C): Mean (95% CI): 79.3 (68.8–89.8), 64.7 (56.8–72.4), and 50.9 (43.8–57.9) for low, mid, and high MUCQ, respectively. (Panel D) Prospective exacerbations through 4-year follow-up: 0 events (N=260), mean MUCQ 12.3 (8.3–16.2); one event (N=34), 11.2 (2.9–19.6); two or more events (N=22), 28.7 (0.1–57.2). For Panels A and D MUCQ values are plotted as raw (non-log-transformed) values on a log₁₀-scaled y axis. Confidence intervals have not been adjusted for multiple comparisons and should therefore not be used for clinical decision making. COPD denotes chronic obstructive lung disease; FEF_25/75, forced expiratory flow rate over the middle 50% of the vital capacity; FEV₁, forced expiratory volume in the first second; GOLD, Global Initiative for Chronic Obstructive Lung Disease; and MUCQ, Mucin Quantitative Score.

In the at-risk group, participants were further categorized based on symptoms using the SGRQ symptom score (<25 vs. ≥25). The mean MUCQ score was 10.8 (95% CI, 5.7–16.3) in the symptomatic at-risk group, compared to 4.2 (95% CI, 0.6–7.7) in the asymptomatic at-risk group (Fig. S2B).

Associations between MUCQ and lung function were evaluated using raw MUCQ values divided into terciles (MUCQ: low=≤0.95, mid=0.96–5.71, and high=>5.71). Participants in the high MUCQ tercile exhibited forced expiratory volume in 1 second (FEV₁, percent predicted) values of 72.2% (95% CI, 67.9–76.4) compared to patients in the low or mid terciles, which had mean FEV₁ values of 81.9% (95% CI, 77.1–86.7) and 80.1% (95% CI, 75.5–84.6), respectively (Fig. 3B and Fig. S4A). The associations between MUCQ score versus total mucin and MUC5AC/MUC5B ratio and FEV₁ percent predicted in regression analyses after adjusting for age, sex, BMI, race, ethnicity, asthma history, smoking status (never, current, past), and smoking pack-year are shown in Table S1. Comparisons of MUCQ scores with either total mucin concentration or with the MUC5AC/MUC5B ratio across healthy controls, at-risk individuals, and COPD groups are presented in Figure S2C–S2E.

MUCQ Scores and FEF_25/75%, RV/TLC, and Computed Tomography Image Analysis

We also assessed the association between MUCQ scores and FEF_25/75% (forced expiratory flow over the middle half of the forced vital capacity), RV/TLC²⁰ (residual volume/total lung capacity), and PRM^fSAD (parametric response mapping functional small airways disease, an analysis of computed tomography scans of the lungs purported to identify small airway disease)²¹ using Tukey–Kramer tests for terciles analysis, Spearman correlations, and ANOVA. Tercile analyses demonstrated that FEF_25/75% was inversely related to MUCQ (Fig. 3C). Correlation coefficients were r=−0.19 for FEF_25/75%, r=0.21for RV/TLC, and r=0.14 for PRM^fSAD (Fig. S4B–S4D).

The association between MUCQ score and FEF25/75% in regression analyses after adjusting for age, sex, BMI, race, ethnicity, asthma history, smoking status, and smoking pack-year included as covariates is shown in Table S2.

MUCQ Scores and COPD Exacerbations

Relationships between MUCQ scores and annualized prospective exacerbations frequencies were examined in the SPIROMICS cohort. Mean MUCQ scores were 12.3 (95% CI, 8.3–16.2) for participants with no exacerbations, 11.2 (95% CI, 2.9–19.6) for participants with one exacerbation per year, and 28.7 (95% CI, 0.1–57.2) for participants who experienced two or more exacerbations per year (Fig. 3D Similar relationships were observed for exacerbations requiring health care use (Fig. S5).

MUCQ and Smoking Status and Cumulative Smoking History

Associations between MUCQ scores and smoking-pack history were assessed using the cigarette pack-year variable divided into terciles (low=20–31, mid=32–60, high=More than 60 pack-year). The mean MUCQ score was 9.1 (95% CI, 3.2–14.9) in the low smoking pack-year tercile, 19.1 (95% CI, 11.7–26.4) in the mid tercile, and 20.0 (95% CI, 6.2–33.71) in the high tercile (Fig. 4A). Among at-risk (ever-smokers with no airflow obstruction), current smokers had a mean MUCQ score of 10.3 (95% CI, 5.3–15.2), whereas those who had not smoked for at least 1 year in the at-risk group had a score of 3.5 (95% CI, 0.3–6.6) (Fig. 4B). In GOLD 1–3 COPD participants, mean MUCQ scores were 15.5 (95% CI, 7.8–23.1) in former smokers and 24.1 (95% CI, 14.4–33.8) in current smokers (Fig. 2B). MUCQ scores in former-smoker COPD participants compared to never-smoker healthy participants are shown in Figure 4B.

Figure 4.

Association of Smoking Status, History, and Cessation on MUCQ Scores.

(Panel A) MUCQ scores compared across smoking pack-year terciles (0, 20–31, 32–60, and>60). Mean MUCQ (95% confidence interval [CI]): 1.8 (0.9–2.5), 9.1 (3.2–14.9), 19.1 (11.7–26.4), and 20.0 (6.2–33.71), respectively. (Panel B) MUCQ and smoking status within at-risk people with a history of smoking at any time without airflow obstruction and within COPD (GOLD 1–3). At-risk: current 10.3 (5.3–15.2), former 3.5 (0.3–6.6). COPD: current 24.1 (14.4–33.8), former 15.5 (7.8–23.1). (Panel C) MUCQ by years since smoking cessation (terciles): 6 years or less, 6–21 years, more than 21 years. Mean MUCQ (95% CI): 21.8 (6.7–36.8), 11.8 (2.9–20.7), and 4.0 (0.9–7.1), respectively. The mean MUCQ for healthy nonsmokers (NS) is marked with the dotted line. MUCQ values are plotted as raw (non-log-transformed) values on a log₁₀-scaled y axis. Confidence intervals have not been adjusted for multiple comparisons and should therefore not be used for clinical decision making. COPD denotes chronic obstructive lung disease; GOLD, Global Initiative for Chronic Obstructive Lung Disease; and MUCQ, Mucin Quantitative Score.

Associations between MUCQ scores and smoking cessation history were contrasted using the “years after quit smoking” variable divided into terciles: low (<6 years), mid (6–21 years), and high (>21 years) (Fig. 4C). The mean MUCQ scores were 21.8 (95% CI, 6.7–36.8) in the short-term cessation group, 11.83 (95% CI, 2.9–20.7) in the mid-term group, and 4.0 (95% CI, 0.9–7.1) in those who quit more than 21 years ago (Fig. 4C).

MUCQ and Symptoms of CB with or Without Airflow Obstruction

Associations between MUCQ scores and SGRQ scores were assessed using both terciles and linear regression analyses. SGRQ_symptom scores across MUCQ terciles were 34.4 (28.9–39.9) for mean (95% CI) in the low tercile, 41.9 (36.0–47.7) in the mid tercile, and 53.3 (48.7–58.4) in the high tercile (Fig. 5A). SGRQ_total and SGRQ_symptom scores showed associations with MUCQ scores by regression analyses (r=0.27 and r=0.31, respectively) (Fig. S6). The associations between MUCQ score and SGRQ_symptom and SGRQ_total scores were similar in regression analyses after age, sex, BMI, race, ethnicity, asthma history, and smoking status were included as covariates (Table S3).

Figure 5.

Associations between MUCQ Scores and Questionnaire-Defined Chronic Bronchitis (CB) and CB Symptoms Such as Cough, Phlegm Production, and Wheezing/Whistling in the Chest.

(Panel A) SGRQ symptoms and MUCQ terciles (low ≤0.95; mid 0.95–5.71; high >5.71). Mean (95% confidence interval [CI]): 34.4 (28.9–39.9), 41.9 (36.0–47.7), and 53.5 (48.7–58.4) for low, mid, and high MUCQ terciles, respectively. (B) Mean MUCQ (95% CI) in participants classified as CB+ was 15.7 (9.9–21.4) using the classic definition and 18.4 (12.2–24.6) using the SGRQ-based definition. (Panel C) Within ever-smokers, mean MUCQ (95% CI) was 18.4 (12.3–24.3) in CB+ and 13.0 (6.5–19.7) in CB− participants. (Panels D–F) Symptom-specific SGRQ items: The MUCQ values from participants reporting cough, phlegm, or wheeze compared to those without these symptoms. Mean MUCQ (95% CI): cough+ 20.0 (12.8–27.1) versus cough− 8.9 (3.7–14.0); phlegm+ 17.4 (11.6–23.2) versus phlegm− 12.7 (5.1–20.1); wheeze+ 18.6 (11.4–25.8) versus wheeze− 10.9 (5.4–16.4). For B–D MUCQ values are plotted as raw (non-log-transformed) values on a log₁₀-scaled y axis. Confidence intervals have not been adjusted for multiple comparisons and should therefore not be used for clinical decision making. For SGRQ (range 0–100) higher scores indicating worse respiratory health; a change of 4 points is considered the minimally clinically important difference. MUCQ denotes Mucin Quantitative Score; and SGRQ, St. George’s Respiratory Questionnaire.

The mean MUCQ score in participants with a diagnosis of CB using both the classic and SGRQ definitions were 15.7 (95% CI, 9.9–21.4) (Fig. 5B) and 18.4 (95% CI, 12.2–24.6), respectively, versus the group of “healthy controls” whose values were 1.8 (95% CI, 0.9–2.5). Within the ever-smoker population, the mean MUCQ score was 18.4 (95% CI, 12.3–24.3) in smokers who met the case definition for CB and was 13.0 (95% CI, 6.5–19.7) in smokers who did not meet the CB case definition (Fig. 5C) versus “healthy controls” whose values were 1.8 (95% CI, 0.9–2.5).

We also examined specific symptoms related to CB captured in the SGRQ instrument, for example, cough and phlegm. MUCQ scores in the participants who experienced cough and phlegm production and chest wheezing compared to those without cough or phlegm or wheezing were as follows: For the mean MUCQ (95% CI) for patients with cough, MUCQ values were 20.0 (12.8–27.1) versus those without cough of 8.9 (3.7–14.0); for patients with phlegm MUCQ values were 17.4 (11.6–23.2) versus those without phlegm of 12.7 (5.1–20.1); and for patients with wheezing MUCQ values were 18.6 (11.4–25.8) versus those without wheezing of 10.9 (5.4–16.4) (Fig. 5D–5F).

Discussion

In this study, we tested whether a composite mucin-based metric, MUCQ score, improved the discrimination between healthy people and patients with chronic airway disease compared with total sputum mucin concentrations and ratio of major airway mucins MUC5AC and MUC5B. Using the SPIROMICS cohort, NRI analysis demonstrated that patients who met the clinical case definition of CB were more likely to be classified correctly when using MUCQ than total mucin concentration, especially in current smokers. These findings, and the findings with respect to common clinical and laboratory measures of disease state in COPD, are consistent with the idea that integrating mucin concentration and compositional information into the MUCQ score captures disease-relevant mucus features and improves classification performance, particularly in distinguishing smoking-related disease risk, compared with total mucin or MUC5AC/MUC5B alone.

Mucin hyperconcentration is a hallmark of chronic lung disease resulting from persistent airway stressors such as cigarette smoke (CB, COPD),^1,4,5 allergens (asthma),²² genetic etiologies associated with persistent infection, such as cystic fibrosis, and primary ciliary dyskinesia^9,15 and non-cystic fibrosis bronchiectasis.^10,23 The rationale for developing a composite mucin score for muco-obstructive lung diseases stem from two complementary hypotheses regarding airway mucus pathology. First, the “two-gel hypothesis” posits that mucociliary transport rates are determined by total mucin concentration-dependent osmotic pressure on airway surfaces, regardless of the mucin subtype composition.^6,7 Even small increases in total mucin concentrations may be pathophysiologically important, slowing mucus clearance and promoting obstruction.^4,7 Second, prior data showing that increased sputum MUC5AC (but not MUC5B) concentrations and higher MUC5AC/MUC5B ratios are linked to COPD initiation, progression, symptom severity, viral load and exacerbations,^5,24 together with our biophysical data demonstrating MUC5AC is more adhesive and viscoelastic than MUC5B,⁸ led to a complementary MUC5AC hypothesis that disproportionate increases in MUC5AC drive dysfunctional mucus formation in CB, COPD, and other muco-obstructive lung diseases.^5,9,10,24,25 Thus, we developed a novel score that incorporates both total mucin concentrations and the role of MUC5AC in disease pathogenesis, termed the MUCQ.

The MUCQ score performed well in the context of established COPD. For example, we found that MUCQ scores were associated with FEV₁, small airway dysfunction, and exacerbation rate. The association between elevated MUCQ scores and increased exacerbation rates, as well as higher healthcare use, in particular highlight the potential clinical relevance of MUCQ as a predictive marker for disease severity and progression. Based on our data we speculate that in patients with CB/COPD, even before spirometric abnormalities are manifest, mucin hypersecretion and the types of mucin in the sputum, that is, MUC5AC and MUC5B, may be present.^4,5 This phenotype, commonly referred to as pre-COPD (at-risk, formerly known as GOLD 0), includes ever-smokers with respiratory symptoms and/or other physiological or imaging-based lung abnormalities despite preserved spirometry.²⁶ In the SPIROMICS cohort, this group represented an at-risk group with persistent symptoms, including cough and sputum production with normal spirometry.^27,28 This population differs from COPD in the absence of airflow obstruction but still exhibits a substantial both present and future disease burden.²⁸ Importantly, the MUCQ identified this group by defining the presence of mucus abnormalities versus nonsymptomatic at-risk participants, suggesting a potential role for MUCQ in early disease detection.

We acknowledge perceived limitations for broader applicability of the MUCQ score owing to the difficulty of the required physicochemical measures. However, total mucin concentrations are measured using size-exclusion chromatography coupled with a refractive index detector, a well-established and widely available and validated²⁹ technique in research and industrial settings that could be adapted for clinical study use. The MUC5AC and MUC5B concentrations are measured by labeled mass spectrometry, a highly accurate and sensitive method. Mass spectrometry (MS), including labelled or label-free tandem MS (MS/MS), is increasingly integrated into clinical settings, replacing traditional immunoassays for biomarker detection in biofluids. Therefore, should the MUCQ score prove to have value in the diagnosis and management of COPD through further research, the potential for its clinical incorporation in such settings is likely within reach.

In summary, by combining three key mucin measurements, total mucins, MUC5AC and MUC5B, we derived a sputum mucin score that reflects independent yet complementary properties of pathological airway mucus. We showed that this MUCQ score improved the ability of an objective laboratory test to establish a clinical diagnosis that had previously rested on historical findings and is subject to substantial recall bias. Objective classification of patients with CB, and the linking of MUCQ to many measures of lung function, sets the stage for its use to be tested use in clinical and research settings.