The Target Trial Emulation of Muco-active Agent Use and Clinical Outcomes After Lung Cancer Diagnosis Among Patients With Chronic Obstructive Pulmonary Disease (TAME-CLC)
The Target Trial Emulation of Muco-active Agent Use and Clinical Outcomes After Lung Cancer Diagnosis Among Patients With Chronic Obstructive Pulmonary Disease (TAME-CLC)
The Target Trial Emulation of Muco-active Agent Use and Clinical Outcomes After Lung Cancer Diagnosis Among Patients With Chronic Obstructive Pulmonary Disease (TAME-CLC) study is a retrospective observational cohort study using existing real-world data from the Korean Cancer Data Center (K-CURE) database to evaluate whether use of muco-active agents is associated with clinical outcomes after lung cancer diagnosis among adults with chronic obstructive pulmonary disease (COPD) and localized-stage lung cancer.
The investigators will not assign any treatment or medication. Instead, the study will emulate a hypothetical target trial using routinely collected retrospective data. Eligible patients will be adults aged 40 to less than 80 years who have COPD before lung cancer diagnosis and meet prespecified diagnostic and treatment-based criteria. Patients will be classified according to muco-active agent treatment strategies based on prescription records after the first lung cancer diagnosis.
The emulated treatment strategy is use of any muco-active agent within a prespecified grace period after lung cancer diagnosis. The comparator strategy is no muco-active agent use during the same grace period. Muco-active agents include mucolytics, mucoregulators, expectorants, and mucokinetics, such as N-acetylcysteine, erdosteine, carbocysteine, guaifenesin, ivy-leaf extract, ambroxol, and bromhexine.
The primary outcome is time to moderate-to-severe COPD exacerbation after lung cancer diagnosis. Secondary outcomes include all-cause mortality, cancer-related mortality, and respiratory disease-related mortality. A clone-censoring-weighting approach will be used within a target trial emulation framework to estimate the modified intention-to-treat effect of muco-active agent use while adjusting for measured baseline differences between treatment strategies.
Patients with chronic obstructive pulmonary disease (COPD) and lung cancer frequently experience mucus hypersecretion, impaired mucociliary clearance, respiratory symptoms, COPD exacerbations, and poor clinical outcomes. COPD is also closely linked to lung cancer through shared risk factors and potentially overlapping biological mechanisms, including chronic airway inflammation, oxidative stress, airway epithelial injury, and mucus plugging. Muco-active agents are commonly prescribed in clinical practice to improve mucus clearance and reduce airway mucus burden. However, whether muco-active agent use is associated with clinical outcomes after lung cancer diagnosis among patients with COPD remains uncertain.
This study will use retrospective real-world data from the Korean Cancer Data Center (K-CURE) database. The study period will extend from January 1, 2002, to December 31, 2023, subject to data availability. The source population will include adults aged 40 to less than 80 years with COPD and newly diagnosed localized-stage lung cancer. COPD will be identified using prespecified diagnosis codes and treatment-based criteria before lung cancer diagnosis. Lung cancer diagnosis, stage, histology, treatment, medication prescriptions, comorbidities, COPD exacerbations, and mortality outcomes will be identified from the K-CURE database and linked clinical data where available.
The study will apply a target trial emulation framework to compare the following treatment strategies: initiation or use of any muco-active agent after lung cancer diagnosis versus no muco-active agent use. The index date will be the date of first lung cancer diagnosis among eligible patients. To support a new-user design and reduce bias from prevalent use, patients with muco-active agent use during the 12-week washout period before lung cancer diagnosis will be excluded according to the prespecified protocol.
Treatment assignment will be defined using prescription records during a 4-week grace period after the index date. For sensitivity analyses, we will additionally evaluate 2-, 6-, and 8-week grace periods after the index date. Patients assigned to the muco-active agent strategy will be those who receive at least one prescription for any prespecified muco-active agent during the grace period. Patients assigned to the comparator strategy will be those who do not receive a prespecified muco-active agent during the same grace period. Muco-active agents will include mucolytics such as N-acetylcysteine and erdosteine, mucoregulators such as carbocysteine, expectorants such as guaifenesin and ivy-leaf extract, and mucokinetics such as ambroxol and bromhexine.
Because treatment is not randomized in the observed data, randomization will be emulated using a clone-censoring-weighting approach. Each eligible patient may be cloned into treatment strategy groups. Clones will be artificially censored when their observed treatment pattern becomes inconsistent with the assigned strategy. Inverse probability of censoring weights will be used to adjust for selection bias introduced by artificial censoring. The primary causal contrast will be the modified intention-to-treat effect of muco-active agent use compared with no muco-active agent use.
Follow-up will begin according to the prespecified emulation protocol after treatment strategy assignment and will continue until the earliest occurrence of the outcome of interest, death, loss of eligibility, the end of the prespecified follow-up period, or administrative censoring. The primary follow-up period for the emulated target trial will be 60 months after lung cancer diagnosis or treatment strategy assignment, with longer follow-up windows evaluated in secondary or sensitivity analyses where data are available.
The primary outcome is time to moderate-to-severe COPD exacerbation. Moderate exacerbation will be defined using outpatient treatment records indicating systemic corticosteroid or antibiotic use for COPD exacerbation, according to the prespecified algorithm. Severe exacerbation will be defined using hospitalization or emergency department visit records for COPD exacerbation. Secondary outcomes include time to all-cause mortality, cancer-related mortality, and respiratory disease-related mortality.
Baseline covariates will be assessed before the index date and during the prespecified baseline period. Covariates may include age, sex, body mass index, smoking status, physical activity, prior COPD exacerbation history, inhaled COPD therapy, lung cancer histology, stage at diagnosis, lung cancer treatment, respiratory comorbidities, medical comorbidities, insurance status, income level, and healthcare utilization. Key adjustment variables in the primary emulation analysis will include age, smoking status, respiratory comorbidities such as chronic bronchitis, and prior COPD exacerbation history.
The primary analysis will estimate hazard ratios using weighted Cox proportional hazards models under the clone-censoring-weighting framework. Additional estimands may include differences in survival probabilities at prespecified time points and restricted mean survival time, where appropriate. Sensitivity analyses may evaluate alternative grace periods, alternative exposure definitions, duration or cumulative use of muco-active agents, and drug class-specific strategies such as N-acetylcysteine, erdosteine, and carbocysteine. Subgroup analyses may be performed by COPD-related and lung cancer-related clinical characteristics including age, sex, body mass index, physical activity, smoking status, previous COPD exacerbation history, comorbidities, lung cancer histology, lung cancer stage, lung cancer treatment, and COPD treatment.
This is an observational study using existing retrospective data. The investigators will not assign muco-active agents, cancer treatment, COPD treatment, or any other intervention. No additional study visits, procedures, or medication changes will be required for participants.
Inclusion Criteria:
Exclusion Criteria: