Comparison of Clinical Outcomes Among Different Fixed-Dose Combinations of Long-Acting Muscarinic Antagonists and Long-Acting β2-Agonists in Patients With COPD

BackgroundResearchers have yet to obtain conclusive evidence differentiating among fixed-dose combinations (FDCs) of long-acting muscarinic antagonists (LAMAs) and long-acting β2-agonists (LABAs) for COPD in terms of real-world clinical outcomes.Research QuestionWhat are the differences between available LAMA/LABA FDCs in the risk of acute exacerbation (AE) and cardiovascular events?Study Design and MethodsThis retrospective cohort study based on a national insurance claims database included patients with COPD ≥ 40 years of age who were newly prescribed glycopyrronium (GLY)/indacaterol (IND), umeclidinium (UMEC)/vilanterol (VI), or tiotropium (TIO)/olodaterol (OLO) FDC between January 1, 2015, and June 30, 2019. Propensity score matching and Cox regression models were used to compare outcomes of AE and cardiovascular events associated with LAMA/LABA FDC treatment.ResultsAmong the 44,498 patients identified and included, 15,586 received GLY/IND, 20,460 received UMEC/VI, and 8,452 received TIO/OLO. Baseline characteristics were well balanced after 1:1 matching of UMEC/VI and GLY/IND, 2:1 matching of UMEC/VI and TIO/OLO, and 2:1 matching of GLY/IND and TIO/OLO. Risk of severe AE was lower among patients treated with UMEC/VI or GLY/IND than among those who received TIO/OLO (UMEC/VI vs TIO/OLO: 17.85 vs 29.32 per 100 person-years; hazard ratio, 0.76; 95% CI, 0.68-0.84; GLY/IND vs TIO/OLO: 15.54 vs 25.53 per 100 person-years; hazard ratio, 0.77; 95% CI, 0.67-0.88). In addition, GLY/IND users tended to have a lower risk of cardiovascular events than TIO/OLO users, but the difference dissipated when restricting follow up to a shorter duration.InterpretationOur results revealed that the risk of severe AE was lower among patients with COPD receiving UMEC/VI or GLY/IND than among those receiving TIO/OLO, whereas the incidence of cardiovascular events was similar across groups but was slightly lower in GLY/IND users when compared with TIO/OLO users. Further research will be required to confirm these findings. Researchers have yet to obtain conclusive evidence differentiating among fixed-dose combinations (FDCs) of long-acting muscarinic antagonists (LAMAs) and long-acting β2-agonists (LABAs) for COPD in terms of real-world clinical outcomes. What are the differences between available LAMA/LABA FDCs in the risk of acute exacerbation (AE) and cardiovascular events? This retrospective cohort study based on a national insurance claims database included patients with COPD ≥ 40 years of age who were newly prescribed glycopyrronium (GLY)/indacaterol (IND), umeclidinium (UMEC)/vilanterol (VI), or tiotropium (TIO)/olodaterol (OLO) FDC between January 1, 2015, and June 30, 2019. Propensity score matching and Cox regression models were used to compare outcomes of AE and cardiovascular events associated with LAMA/LABA FDC treatment. Among the 44,498 patients identified and included, 15,586 received GLY/IND, 20,460 received UMEC/VI, and 8,452 received TIO/OLO. Baseline characteristics were well balanced after 1:1 matching of UMEC/VI and GLY/IND, 2:1 matching of UMEC/VI and TIO/OLO, and 2:1 matching of GLY/IND and TIO/OLO. Risk of severe AE was lower among patients treated with UMEC/VI or GLY/IND than among those who received TIO/OLO (UMEC/VI vs TIO/OLO: 17.85 vs 29.32 per 100 person-years; hazard ratio, 0.76; 95% CI, 0.68-0.84; GLY/IND vs TIO/OLO: 15.54 vs 25.53 per 100 person-years; hazard ratio, 0.77; 95% CI, 0.67-0.88). In addition, GLY/IND users tended to have a lower risk of cardiovascular events than TIO/OLO users, but the difference dissipated when restricting follow up to a shorter duration. Our results revealed that the risk of severe AE was lower among patients with COPD receiving UMEC/VI or GLY/IND than among those receiving TIO/OLO, whereas the incidence of cardiovascular events was similar across groups but was slightly lower in GLY/IND users when compared with TIO/OLO users. Further research will be required to confirm these findings. FOR EDITORIAL COMMENT, SEE PAGE 731Take-home PointsStudy Question: What are the differences between available long-acting muscarinic antagonist/long-acting β2-agonist fixed-dose combinations in the risk of acute exacerbation (AE) and cardiovascular events based on nationwide data in Taiwan?Results: Patients with COPD treated with umeclidinium (UMEC)/vilanterol (VI) or glycopyrronium (GLY)/indacaterol (IND) had a lower risk of severe AEs than patients who received tiotropium (TIO)/olodaterol (OLO) (UMEC/VI: hazard ratio, 0.76; 95% CI, 0.68-0.84; GLY/IND: hazard ratio, 0.77; 95% CI, 0.67-0.88). In addition, GLY/IND users tended to have a lower risk of cardiovascular events than TIO/OLO users, but the difference dissipated when restricting follow up to a shorter duration.Interpretation: Risk of severe AE was lower among patients with COPD receiving UMEC/VI or GLY/IND compared with patients receiving TIO/OLO, whereas the incidence of cardiovascular events was more similar. FOR EDITORIAL COMMENT, SEE PAGE 731 Study Question: What are the differences between available long-acting muscarinic antagonist/long-acting β2-agonist fixed-dose combinations in the risk of acute exacerbation (AE) and cardiovascular events based on nationwide data in Taiwan? Results: Patients with COPD treated with umeclidinium (UMEC)/vilanterol (VI) or glycopyrronium (GLY)/indacaterol (IND) had a lower risk of severe AEs than patients who received tiotropium (TIO)/olodaterol (OLO) (UMEC/VI: hazard ratio, 0.76; 95% CI, 0.68-0.84; GLY/IND: hazard ratio, 0.77; 95% CI, 0.67-0.88). In addition, GLY/IND users tended to have a lower risk of cardiovascular events than TIO/OLO users, but the difference dissipated when restricting follow up to a shorter duration. Interpretation: Risk of severe AE was lower among patients with COPD receiving UMEC/VI or GLY/IND compared with patients receiving TIO/OLO, whereas the incidence of cardiovascular events was more similar. The role of COPD in global morbidity and mortality is rapidly growing, and as reported by the World Health Organization, COPD was the third leading cause of death worldwide in 2019.1World Health OrganizationThe top 10 causes of death.https://www.who.int/news-room/fact-sheets/detail/the-top-10-causes-of-deathDate accessed: August 25, 2022Google Scholar The disease is characterized by airflow limitation and destruction of lung parenchyma leading to progressively worsening symptoms and lung function that can be accelerated by acute exacerbations (AEs).2Global Initiative for Chronic Obstructive Lung Disease, Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease 2022 report.https://goldcopd.org/2022-gold-reports/Date accessed: August 25, 2022Google Scholar Strong evidence supports the efficacy of bronchodilators in managing symptoms, improving quality of life, preventing AEs, and reducing mortality in patients with COPD, and treatment with long-acting muscarinic antagonists (LAMAs) and long-acting β2-agonists (LABAs) has been firmly established as the cornerstone of maintenance therapy for COPD.2Global Initiative for Chronic Obstructive Lung Disease, Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease 2022 report.https://goldcopd.org/2022-gold-reports/Date accessed: August 25, 2022Google Scholar, 3Chapman K.R. Rennard S.I. Dogra A. et al.Long-term safety and efficacy of indacaterol, a long-acting β₂-agonist, in subjects with COPD: a randomized, placebo-controlled study.Chest. 2011; 140: 68-75Abstract Full Text Full Text PDF PubMed Scopus (120) Google Scholar, 4Hanania N.A. Feldman G. Zachgo W. et al.The efficacy and safety of the novel long-acting β2 agonist vilanterol in patients with COPD: a randomized placebo-controlled trial.Chest. 2012; 142: 119-127Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar, 5Tashkin D. Celli B. Senn S. et al.A 4-year trial of tiotropium in chronic obstructive pulmonary disease (UPLIFT trial).Rev Port Pneumol. 2009; 15: 137-140Crossref PubMed Google Scholar, 6Calverley P.M.A. Anzueto A.R. Carter K. et al.Tiotropium and olodaterol in the prevention of chronic obstructive pulmonary disease exacerbations (DYNAGITO): a double-blind, randomised, parallel-group, active-controlled trial.Lancet Respir Med. 2018; 6: 337-344Abstract Full Text Full Text PDF PubMed Scopus (139) Google Scholar, 7Celli B. Crater G. Kilbride S. et al.Once-daily umeclidinium/vilanterol 125/25 mcg in COPD: a randomized, controlled study.Chest. 2014; 145: 981-991Abstract Full Text Full Text PDF PubMed Scopus (144) Google Scholar The Global Initiative for Chronic Obstructive Lung Disease2Global Initiative for Chronic Obstructive Lung Disease, Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease 2022 report.https://goldcopd.org/2022-gold-reports/Date accessed: August 25, 2022Google Scholar guidelines classify patients with COPD based on severity of airflow limitation, dyspnea, exacerbation history, and impact on disease-specific quality of life, and recommend initial maintenance treatment with an inhaled LAMA or LABA for patients with moderate or severe disease. Escalation to dual bronchodilator therapy is recommended for patients who remain highly symptomatic or experience AEs despite bronchodilator monotherapy; initial dual bronchodilation can also be considered for patients with severe disease at diagnosis.2Global Initiative for Chronic Obstructive Lung Disease, Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease 2022 report.https://goldcopd.org/2022-gold-reports/Date accessed: August 25, 2022Google Scholar LAMA and LABA have differing and synergistic mechanisms of action, and multiple studies have shown that combined LAMA/LABA treatment more effectively increases lung function, reduces AEs, and ameliorates symptoms compared with LAMA or LABA alone or combination LABA and inhaled corticosteroid (ICS).8Calzetta L. Matera M.G. Cazzola M. Pharmacological mechanisms leading to synergy in fixed-dose dual bronchodilator therapy.Curr Opin Pharmacol. 2018; 40: 95-103Crossref PubMed Scopus (47) Google Scholar, 9Han M.K. Ray R. Foo J. Morel C. Hahn B. Systematic literature review and meta-analysis of US-approved LAMA/LABA therapies versus tiotropium in moderate-to-severe COPD.NPJ Prim Care Respir Med. 2018; 28: 32Crossref PubMed Scopus (9) Google Scholar, 10Oba Y. Keeney E. Ghatehorde N. Dias S. Dual combination therapy versus long-acting bronchodilators alone for chronic obstructive pulmonary disease (COPD): a systematic review and network meta-analysis.Cochrane Database Syst Rev. 2018; 12: CD012620PubMed Google Scholar, 11Mammen M.J. Pai V. Aaron S.D. Nici L. Alhazzani W. Alexander P.E. Dual LABA/LAMA therapy versus LABA or LAMA monotherapy for chronic obstructive pulmonary disease. A systematic review and meta-analysis in support of the American Thoracic Society Clinical Practice Guideline.Ann Am Thorac Soc. 2020; 17: 1133-1143Crossref PubMed Scopus (23) Google Scholar However, concern has been raised regarding potentially increased risk of cardiovascular events with dual bronchodilator treatment, with some data suggesting increased risk, especially within 30 days of new use.12Wang M.T. Lai J.H. Tsai C.L. Liou J.T. Risk of adverse cardiovascular events with use of inhaled long-acting bronchodilators in management of chronic obstructive pulmonary disease.J Food Drug Anal. 2019; 27: 657-670Crossref PubMed Scopus (9) Google Scholar,13Wang M.T. Liou J.T. Lin C.W. et al.Association of cardiovascular risk with inhaled long-acting bronchodilators in patients with chronic obstructive pulmonary disease: a nested case-control study.JAMA Intern Med. 2018; 178: 229-238Crossref PubMed Scopus (89) Google Scholar Furthermore, patients with COPD are very often smokers, and both cigarette smoking and COPD are risk factors for cardiovascular disease.14Brassington K. Selemidis S. Bozinovski S. Vlahos R. Chronic obstructive pulmonary disease and atherosclerosis: common mechanisms and novel therapeutics.Clin Sci (Lond). 2022; 136: 405-423Crossref PubMed Scopus (7) Google Scholar Therefore, attention to cardiovascular risk is an important consideration in COPD treatment selection. In the treatment of COPD, fixed-dose combination (FDC) inhalers that simultaneously deliver multiple agents via a single inhaler device have been shown to improve medication adherence, prolong treatment persistence, and reduce health care costs compared with multiple single agent inhalers.15Bogart M. Stanford R.H. Laliberte F. Germain G. Wu J.W. Duh M.S. Medication adherence and persistence in chronic obstructive pulmonary disease patients receiving triple therapy in a USA commercially insured population.Int J Chron Obstruct Pulmon Dis. 2019; 14: 343-352Crossref PubMed Scopus (47) Google Scholar, 16Ismaila A.S. Risebrough N. Schroeder M. et al.Cost-effectiveness of once-daily single-inhaler triple therapy in COPD: the IMPACT trial.Int J Chron Obstruct Pulmon Dis. 2019; 14: 2681-2695Crossref PubMed Scopus (18) Google Scholar, 17Yu A.P. Guerin A. Ponce de Leon D. et al.Therapy persistence and adherence in patients with chronic obstructive pulmonary disease: multiple versus single long-acting maintenance inhalers.J Med Econ. 2011; 14: 486-496Crossref PubMed Scopus (113) Google Scholar Numerous LAMA/LABA FDCs are currently available, but there is a notable paucity of head-to-head comparison studies of the various FDCs. A network meta-analysis including 74 randomized controlled trials showed through indirect comparative analyses that LAMA/LABA was the treatment class with the highest ranking probability of improving lung function as assessed with FEV1; however, there were no significant differences among the different LAMA/LABA combinations.18Aziz M.I.A. Tan L.E. Wu D.B. et al.Comparative efficacy of inhaled medications (ICS/LABA, LAMA, LAMA/LABA and SAMA) for COPD: a systematic review and network meta-analysis.Int J Chron Obstruct Pulmon Dis. 2018; 13: 3203-3231Crossref PubMed Scopus (35) Google Scholar Randomized controlled trials directly comparing different LAMA/LABA FDCs did not show significant differences in effect on FEV1.19Kerwin E. Ferguson G.T. Sanjar S. et al.Dual bronchodilation with indacaterol maleate/glycopyrronium bromide compared with umeclidinium bromide/vilanterol in patients with moderate-to-severe COPD: results from two randomized, controlled, cross-over studies.Lung. 2017; 195: 739-747Crossref PubMed Scopus (31) Google Scholar, 20Maltais F. Ferguson G.T. Feldman G.J. et al.A randomized, double-blind, double-dummy study of glycopyrrolate/formoterol fumarate metered dose inhaler relative to umeclidinium/vilanterol dry powder inhaler in COPD.Adv Ther. 2019; 36: 2434-2449Crossref PubMed Scopus (18) Google Scholar, 21Muraki M. Kunita Y. Shirahase K. et al.A randomized controlled trial of long-acting muscarinic antagonist and long-acting β2 agonist fixed-dose combinations in patients with chronic obstructive pulmonary disease.BMC Pulm Med. 2021; 21: 26Crossref PubMed Scopus (5) Google Scholar Furthermore, statistically significant changes in FEV1 may not translate to noticeable clinical improvement.22Deas S.D. Huprikar N. Dual bronchodilator therapy for chronic obstructive pulmonary disease: evidence for the efficacy and safety of fixed dose combination treatments in the setting of recent guideline updates.Curr Opin Pulm Med. 2018; 24: 130-137Crossref PubMed Scopus (8) Google Scholar,23Jones P.W. Donohue J.F. Nedelman J. Pascoe S. Pinault G. Lassen C. Correlating changes in lung function with patient outcomes in chronic obstructive pulmonary disease: a pooled analysis.Respir Res. 2011; 12: 161Crossref PubMed Scopus (67) Google Scholar To further clarify the effects of different LAMA/LABA FDCs on clinically important outcomes in COPD, this study aimed to evaluate the risk of AEs and cardiovascular events after treatment initiation of three LAMA/LABA FDCs. This retrospective cohort study used claims data from the National Health Insurance Research Database and mortality data from the National Death Registry in Taiwan between 2010 and 2019. The National Health Insurance (NHI) is mandatory universal health insurance that covers > 99.8% of the Taiwanese population. Under the NHI, preventive, primary, specialist, hospital, emergency, and pharmacy services are highly accessible with affordable copayments. The approval of this study was obtained from the Research Ethics Committees of the National Taiwan University Hospital (No. 201706124RINA). During the study period, three LAMA/LABA FDCs were available in Taiwan: glycopyrronium (GLY)/indacaterol (IND) 50/110 μg/dose (Ultibro Breezhaler), umeclidinium (UMEC)/vilanterol (VI) 55/22 μg/dose (Anoro Ellipta), and tiotropium (TIO)/olodaterol (OLO) 2.5/2.5 μg/dose (Stiolto Respimat). These medications have been reimbursed without restrictions by the NHI since January 2015, May 2015, and November 2016, respectively. This study included all patients with COPD ≥ 40 years of age who received GLY/IND, UMEC/VI, or TIO/OLO between January 1, 2015, and June 30, 2019. Because our data were available until December 31, 2019, the included patients would have a follow-up duration of at least 6 months after initiation of LAMA/LABA FDC. Instances of COPD were identified by at least one inpatient-related or three outpatient-related claims with the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes of 491.x, 492.x, or 496.x or ICD-10-CM codes of J41.x-J44.x. Despite the current availability and reimbursement of multiple LAMA/LABA FDCs in Taiwan, the use of various combinations of dual LAMA and LABA monotherapy continues in clinical practice. We included only new users of LAMA/LABA FDCs, and we further excluded patients with any concomitant use of LAMA and LABA in the preceding 12 months (either through dual LAMA and LABA monotherapy or by adding an LAMA to ICS/LABA combination). The washout period was included as a measure to eliminate prevalent user bias, and reducing such confounding bias would allow a clearer observation of the outcomes and comparative effectiveness among LAMA/LABA FDCs. The NHI did not reimburse triple ICS/LABA/LAMA FDCs until March 2020. Patients who started ICS treatment on the same date as their first LAMA/LABA FDC prescription were also excluded. To better ensure that patients received LAMA/LABA as continuous maintenance treatment, at least two prescriptions for the same LAMA/LABA FDC were required with the interval between the first two prescriptions shorter than the daily supply of the first prescription. The start of follow up was the date of the second prescription (the index date), and this lag time was imposed as a means to reduce protopathic bias. To ensure the completeness of captured baseline characteristics, patients who were not continuously enrolled in the NHI for at least 2 years prior to the index date were excluded. The primary outcome of this study was severe AE, defined as hospitalization or ED visit with a primary diagnosis of COPD (ICD-9-CM codes: 491.x, 492.x, or 496.x; ICD-10-CM codes: J41.x-J44.x) or a secondary diagnosis of severe AE (ICD-9-CM code: 491.21; ICD-10-CM code: J44.1). The secondary outcome was cardiovascular event, defined as hospitalization with a primary diagnosis of acute myocardial infarction (ICD-9-CM code: 410.x; ICD-10-CM code: I21.x-I22.x), heart failure (ICD-9-CM code: 428.x; ICD-10-CM code: I50.x), or arrhythmia (ICD-9-CM code: 427.x; ICD-10-CM code: I46.x-I49.x) or cardiovascular disease-related death (ICD-10-CM code: I01.x-I99.x). In Taiwan, the diagnostic coding system transitioned from ICD-9-CM to ICD-10-CM in 2016; therefore, we further examined whether the coding transition could have impacted the identification of study events over time. The primary analysis was based on the as-treated approach. Specifically, patients were followed from the index date until the occurrence of the study outcome, discontinuation of LAMA/LABA FDC therapy, add-on or switching LAMA/LABA maintenance therapy (any product containing LAMA or LABA), disenrollment from the NHI, death, or the end of the study period. Discontinuation of LAMA/LABA FDC therapy was defined as a drug supply gap exceeding the daily supply of the prescription (ie, grace period), wherein the date of discontinuation was designated as the end date of the grace period. We added a 14-day follow-up period after switching the LAMA/LABA FDC or adding/switching LAMA/LABA maintenance therapy to account for potential carryover effects. Pairwise comparisons were performed on the three treatment groups: UMEC/VI vs GLY/IND, UMEC/VI vs TIO/OLO, and GLY/IND vs TIO/OLO. Propensity score matching (1:1 or up to 2:1) was used to minimize the impact of confounders. Multivariable logistic regression was used to estimate propensity scores based on the following baseline covariates (including all the covariates listed in Table 1): age, sex, index year, care setting, duration of COPD, history of influenza or pneumococcal vaccination, proxies of COPD severity, recent history of COPD medication use (measured mainly by the number of prescriptions in the past year prior to the index date), and comorbidities. Comorbidities were identified using records within 2 years prior to the index date (requiring at least two outpatient or one inpatient diagnosis). Patients were matched according to propensity score and the presence of cardiovascular disease at baseline (including ischemic heart disease, cardiac dysrhythmias, cerebrovascular disease, congestive heart failure, hypertension, or peripheral artery disease). Greedy nearest-neighbor matching was applied to propensity scores using calipers with a width of 0.03.24Austin P.C. Some methods of propensity-score matching had superior performance to others: results of an empirical investigation and Monte Carlo simulations.Biom. J. 2009; 51: 171-184Crossref PubMed Scopus (524) Google Scholar The groups were compared in terms of baseline characteristics using standardized differences with weighted proportions to account for the matched nature of the sample.25Austin P.C. Assessing balance in measured baseline covariates when using many-to-one matching on the propensity-score.Pharmacoepidemiol. Drug Saf. 2008; 17: 1218-1225Crossref PubMed Scopus (133) Google Scholar Generally, an absolute standardized difference of < 0.1 is considered negligible. Poisson regressions were used to estimate the incidence of severe AE and cardiovascular events, and Cox regression models were used to compare the risk of events between matched groups. We tested the assumption of proportional hazards and used the robust estimation method to account for clustering within matched sets. The cumulative incidences of severe AE and cardiovascular events between matched groups were calculated using the Kaplan-Meier method. All P values were two-sided, and the significance level was set at 0.05. Data analysis was performed using SAS version 9.4 software (SAS Institute).Table 1Baseline Characteristics of LAMA/LABA FDC Users After MatchingCharacteristicUMEC/VI (n = 13,734)GLY/IND (n = 13,734)Absolute Standardized DifferenceUMEC/VI (n = 10,407)TIO/OLO (n = 6,088)Absolute Standardized DifferenceGLY/IND (n = 7,525)TIO/OLO (n = 4,881)Absolute Standardized DifferenceAge, mean ± SD, y69.5 ± 10.869.5 ± 10.8< 0.0170.1 ± 10.770.7 ± 11.30.0269.3 ± 10.869.7 ± 11.20.01Male sex11,642 (84.8)11,616 (84.6)0.018,715 (83.7)5,077 (83.4)0.026,403 (85.1)4,119 (84.4)0.02Index y0.010.030.02 2015697 (5.1)690 (5.0)0 (0)0 (0)0 (0)0 (0) 20163,566 (26.0)3,550 (25.9)0 (0)0 (0)0 (0)0 (0) 20174,455 (32.4)4,459 (32.5)2,883 (27.7)1,500 (24.6)2,603 (34.6)1,425 (29.2) 20183,682 (26.8)3,746 (27.3)4,998 (48.0)3,006 (49.4)3,630 (48.2)2,454 (50.3) 20191,334 (9.7)1,289 (9.4)2,526 (24.3)1,582 (26.0)1,292 (17.2)1,002 (20.5)Time from first COPD diagnosis, y0.010.010.01 < 36,539 (47.6)6,530 (47.6)4,821 (46.3)2,781 (45.6)3,581 (47.6)2,279 (46.7) 3-52,255 (16.4)2,214 (16.1)1,432 (13.8)838 (13.8)1,049 (13.9)687 (14.1) > 54,940 (36.0)4,990 (36.3)4,154 (39.9)2,469 (40.6)2,895 (38.5)1,915 (39.2)No. of severe AEs in the past y prior to the index date0.010.020.01 011,006 (80.1)11,052 (80.5)7,975 (76.6)4,570 (75.1)6,057 (80.5)3,831 (78.5) 12,250 (16.4)2,217 (16.1)2,052 (19.7)1,273 (20.9)1,251 (16.6)890 (18.2) ≥ 2478 (3.5)465 (3.4)380 (3.7)245 (4.0)217 (2.9)160 (3.3)No. of moderate AEs in the past y prior to the index date0.010.010.01 09,739 (70.9)9,770 (71.1)7,637 (73.4)4,517 (74.2)5,509 (73.2)3,580 (73.4) 12,114 (15.4)2,073 (15.1)1,597 (15.4)892 (14.7)1,121 (14.9)727 (14.9) ≥ 21,881 (13.7)1,891 (13.8)1,173 (11.3)679 (11.2)895 (11.9)574 (11.8)Care setting0.010.010.02 Medical center2,845 (20.7)2,881 (21.0)2,383 (22.9)1,435 (23.6)1,573 (20.9)1,110 (22.7) Regional hospital4,286 (31.2)4,255 (31.0)3,424 (32.9)1,976 (32.5)2,346 (31.2)1,495 (30.6) District hospital or lower level6,603 (48.1)6,598 (48.0)4,600 (44.2)2,677 (44.0)3,606 (47.9)2,276 (46.6) Previous influenza or pneumococcal vaccination7,754 (56.5)7,821 (57.0)0.016,391 (61.4)3,785 (62.2)< 0.014,551 (60.5)2,990 (61.3)0.01Comorbidities Atherosclerosis372 (2.7)379 (2.8)< 0.01298 (2.9)183 (3.0)< 0.01194 (2.6)135 (2.8)< 0.01 Cardiac dysrhythmias1,964 (14.3)1,984 (14.5)< 0.011,578 (15.2)939 (15.4)< 0.011,080 (14.4)718 (14.7)0.01 Cerebrovascular disease1,681 (12.2)1,704 (12.4)0.011,411 (13.6)862 (14.7)0.01913 (12.1)611 (12.5)0.01 Congestive heart failure3,578 (26.1)3,578 (26.1)< 0.013,112 (29.9)1,872 (30.8)< 0.012,116 (28.1)1,411 (28.9)0.01 Hypertension7,911 (57.6)7,944 (57.8)< 0.016,126 (58.9)3,609 (59.3)< 0.014,329 (57.5)2,825 (57.9)0.01 Ischemic heart disease3,691 (26.9)3,726 (27.1)0.012,783 (26.7)1,690 (27.8)0.012,019 (26.8)1,353 (27.7)0.01 Asthma3,504 (25.5)3,486 (25.4)< 0.012,379 (22.9)1,381 (22.7)< 0.011,719 (22.8)1,108 (22.7)0.01 Bronchiectasis1,161 (8.5)1,156 (8.4)< 0.01848 (8.2)531 (8.7)< 0.01619 (8.2)415 (8.5)< 0.01 TB507 (3.7)505 (3.7)< 0.01383 (3.7)226 (3.7)0.01255 (3.4)174 (3.6)< 0.01 Diabetes3,355 (24.4)3,429 (25.0)0.012,729 (26.2)1,591 (26.1)0.011,888 (25.1)1,260 (25.8)0.01 Lung cancer731 (5.3)698 (5.1)0.01615 (5.9)356 (5.9)< 0.01385 (5.1)264 (5.4)< 0.01 Cancers (other than lung)1,873 (13.6)1,859 (13.5)< 0.011,230 (11.8)756 (12.4)< 0.01826 (11.0)570 (11.7)0.01 Immunodeficiency55 (0.4)52 (0.4)< 0.0145 (0.4)31 (0.5)0.0126 (0.4)18 (0.4)< 0.01 Dyslipidemia4,042 (29.4)4,073 (29.7)< 0.013,146 (30.2)1,751 (28.8)0.022,249 (29.9)1,450 (29.7)< 0.01 Rheumatoid arthritis199 (1.5)182 (1.3)0.01160 (1.5)89 (1.5)0.01106 (1.4)63 (1.3)0.01 Renal disease1,456 (10.6)1,421 (10.4)0.011,374 (13.2)840 (13.8)< 0.01888 (11.8)607 (12.4)< 0.01 Liver disease1,123 (8.2)1,113 (8.1)< 0.01826 (7.9)503 (8.3)< 0.01629 (8.4)411 (8.4)< 0.01 Osteoporosis609 (4.4)615 (4.5)< 0.01492 (4.7)287 (4.7)0.01334 (4.4)210 (4.3)0.01 Anxiety and depression2,226 (16.2)2,262 (16.5)0.011,709 (16.4)986 (16.2)0.011,218 (16.2)799 (16.4)< 0.01 Dementia592 (4.3)579 (4.2)< 0.01607 (5.8)407 (6.7)< 0.01361 (4.8)270 (5.5)< 0.01 Gastroesophageal reflux2,639 (19.2)2,643 (19.2)< 0.011,935 (18.6)1,155 (19.0)< 0.011,464 (19.5)988 (20.2)0.02Prescription No. of COPD medications in the past y prior to the index date SABA0.010.020.02DDD = 07,735 (56.3)7,819 (56.9)5,351 (51.4)3,051 (50.1)4,324 (57.5)2,702 (55.4)0 < DDD ≤ 302,391 (17.4)2,351 (17.1)2,284 (22.0)1,430 (23.5)1,349 (17.9)952 (19.5)30 < DDD ≤ 602,001 (14.6)1,982 (14.4)1,608 (15.5)915 (15.0)1,100 (14.6)728 (14.9)60 < DDD1,607 (11.7)1,582 (11.5)1,164 (11.2)692 (11.4)752 (10.0)499 (10.2) SAMA0.010.020.01DDD = 09,169 (66.8)9,243 (67.3)6,237 (59.9)3,505 (57.6)5,063 (67.3)3,130 (64.1)0 < DDD ≤ 302,956 (21.5)2,927 (21.3)2,812 (27.0)1,765 (29.0)1,697 (22.6)1,199 (24.6)30 < DDD ≤ 60904 (6.6)876 (6.4)869 (8.4)512 (8.4)457 (6.1)323 (6.6)60 < DDD705 (5.1)688 (5.0)489 (4.7)306 (5.0)308 (4.1)229 (4.7) LABA< 0.010.020.0109,599 (69.9)9,627 (70.1)8,115 (78.0)4,780 (78.5)5,599 (74.1)3,715 (76.1)1-21,423 (10.4)1,406 (10.2)893 (8.6)486 (8.0)663 (8.8)409 (8.4)≥ 32,712 (19.8)2,701 (19.7)1,399 (13.4)822 (13.5)1,263 (16.7)757 (15.5) LAMA< 0.010.030.01011,262 (82.0)11,273 (82.1)8,429 (81.0)4,635 (76.1)6,041 (80.3)3,687 (75.5)1-2933 (6.8)929 (6.8)775 (7.5)540 (8.9)560 (7.4)419 (8.6) ≥ 31,539 (11.2)1,532 (11.2)1,203 (11.5)913 (15.0)924 (12.3)775 (15.9) ICS0.020.010.01011,958 (87.1)12,034 (87.6)9,181 (88.2)5,347 (87.8)6,708 (89.1)4,342 (89.0)1-21,114 (8.1)1,092 (8.0)798 (7.7)485 (8.0)542 (7.2)356 (7.3)≥ 3662 (4.8)608 (4.4)428 (4.1)256 (4.2)275 (3.7)183 (3.8) Xanthines0.010.010.0102,885 (21.0)2,851 (20.8)2,485 (23.9)1,563 (25.7)1,780 (23.7)1,222 (25.0)1-66,245 (45.5)6,219 (45.3)4,770 (45.8)2,657 (43.6)3,465 (46.1)2,171 (44.5)≥ 74,604 (33.5)4,664 (34.0)3,152 (30.3)1,868 (30.7)2,280 (30.3)1,488 (30.5) Oral glucocorticoids0.010.010.0105,779 (42.1)5,836 (42.5)4,331 (41.6)2,547 (41.8)3,330 (44.3)2,143 (43.9)1-66,209 (45.2)6,173 (45.0)4,778 (45.9)2,734 (44.9)3,336 (44.3)2,149 (44.0)≥ 71,746 (12.7)1,725 (12.6)1,298 (12.5)807 (13.3)859 (11.4)589 (12.1) Oral antibiotics0.010.020.0103,579 (26.1)3,607 (26.3)2,637 (25.3)1,473 (24.2)1,972 (26.2)1,243 (25.5)1-68,222 (59.9)8,231 (59.9)6,293 (60.5)3,708 (60.9)4,523 (60.1)2,927 (60.0)≥ 71,933 (14.1)1,896 (13.8)1,477 (14.2)907 (14.9)1,030 (13.7)711 (14.6)Concurrent ICS exposure at LABA/LAMA FDC initiation (previously initiated)532 (3.9)489 (3.6)0.02434 (4.2)269 (4.4)0.01285 (3.8)187 (3.8)0.01Use of other medications in the past y prior to the index date α-blockers1,352 (9.8)1,401 (10.2)0.011,020 (9.8)619 (10.2)< 0.01733 (9.7)491 (10.1)< 0.01 β-blockers4,201 (30.6)4,222 (30.7)< 0.013,434 (33.0)2,007 (33.0)< 0.012,387 (31.7)1,554 (31.8)0.01 ACEIs/ARBs5,692 (41.4)5,672 (41.3)< 0.014,433 (42.6)2,620 (43.0)< 0.013,130 (41.6)2,086 (42.7)0.01 Antiplatelets/anticoagulant5,844 (42.6)5,851 (42.6)< 0.014,618 (44.4)2,752 (45.2)< 0.013,236 (43