Introduction

Chronic obstructive pulmonary disease (COPD) is one of the top three causes of death worldwide.1 There is currently no cure for COPD, so treatment can help slow the progression of the condition and control the symptoms.2 COPD is a chronic lung disease affecting men and women worldwide, characterized by persistent respiratory symptoms and airflow limitation due to airway and/or alveolar abnormalities.1,3 Lung functional abnormalities lead to a limitation of the airflow, which could destroy parts of the mucus, blocking the airways and inflaming and swelling the airway lining.3 Common symptoms of COPD include breathlessness, difficulty breathing, chronic cough, dyspnea, cough or sputum production, and fatigue.1,3 These symptoms commonly lead to a reduction of quality of life due to the difficulties in developing normal activities, occasioned by breathlessness and acute worsening of respiratory symptoms.3 COPD is a leading cause of morbidity and mortality and induces a substantial and increasing social and economic burden1 with a considerable prevalence of 11.8% of adults over 40 years in 2019 and with a mortality of 26.8 deaths per 100,000 inhabitants in 2020.4,5

As a chronic and progressive disease, COPD is associated with significant concomitant chronic pathologies in most patients, which increase associated morbidity and mortality, creating a higher economic impact.1 COPD represents one of the pathologies with the most important economic burden for the Spanish healthcare system, reaching 3000 million euros (€, 2011) per year.6 The annual cost of COPD to the healthcare system is €3238 (~91%) per patient, with indirect costs amounting to €300 (~9%) per patient, bringing the total societal cost of the disease to €3538 per patient per year. Among the direct healthcare costs, hospitalizations caused by exacerbations account for the most significant proportion of expenditure, around 70–80%7.Mortality increases with the frequency of severe exacerbations, that could generate a 4.1 times greater mortality risk compared to patients with stable COPD, particularly if these require a hospital admission.8

The major objectives of COPD treatment are the reduction of disease symptoms, the reduction of the frequency and severity of exacerbations, and the improvement of the quality of life and patient survival.9 Long-acting inhaled bronchodilator medications are recommended as initial maintenance therapy for many patients with COPD. These medications include long-acting muscarinic antagonists (LAMA) and long-acting β2-agonists (LABA). Combinations of long-acting bronchodilator agents (LAMA/LABA) and inhaled corticosteroids (ICS) combined with LABA (ICS/LABA) are also used as initial or follow-up therapy in patients with more severe symptoms or at risk of COPD exacerbations.10 For patients who remain symptomatic despite dual therapy (a combination of ICS/LABA or LAMA/LABA) and are at high risk of exacerbations, the GOLD and the Spanish COPD guidelines recommend a triple combination of ICS/LAMA/LABA.9,11

ETHOS trial (Efficacy and Safety of Triple Therapy in Obstructive Lung Disease) was a randomized, double-blind, multi-centred, parallel-group study to assess the efficacy and safety of the ICS/LAMA/LABA triple-fixed therapy budesonide/glycopyrrolate/formoterol fumarate (BUD/GLY/FOR) relative to GLY/FOR (LAMA/LABA) and BUD/FOR (ICS/LABA) in over 52 weeks in patients with moderate to very severe COPD.12 BUD/GLY/FOR are well-established pharmacological agents in the treatment of COPD, combined in a single metered-dose inhaler (MDI) device. BUD/GLY/FOR showed benefits over dual therapy with a LAMA/LABA or ICS/LABA combination concerning the annual rate of moderate or severe COPD exacerbations, symptoms, and health-related quality of life.12

The prevalence and burden of COPD are projected to increase over the coming decades due to continued exposure to COPD risk factors, disease characteristics, and aging of the population,1 so having comparative evidence regarding the economic impacts of these therapies will be desirable. The present analysis aimed at assessing the cost-effectiveness of triple fixed-dose therapy BUD/GLY/FOR compared with dual therapies LAMA/LABA and ICS/LABA in patients with moderate to severe COPD, from the Spanish National Healthcare System (NHS) perspective.

Methods Model Overview

Given that COPD is a chronic progressive disease, a semi-Markov model was developed to estimate the costs and outcomes of moderate to severe COPD patients receiving three alternatives: triple fixed-dose therapy BUD/GLY/FOR; dual PT003 GLY/FOR (LAMA/LABA) MDI therapy; or dual PT009 BUD/FOR (ICS/LABA) MDI therapy. A lifetime horizon was used with a cycle length of 1-month following the natural history of the disease. The baseline characteristics of the modelled cohort were obtained from the ETHOS trial13 (Table 1).

Table 1 Clinical Inputs Used in the Model

The model (Figure 1) comprised ten health states: nine forced expiratory volume in 1 second (FEV1)-related health states and a death state. FEV1 is a spirometry measure for assessing COPD and other lung diseases.1 The FEV1-related health states were divided by three levels of severity, defined according to the GOLD 2017 guidelines:14 moderate (GOLD 2): patients have post-bronchodilator FEV1 ≥50% and <80% of predicted; severe (GOLD 3): patients have post-bronchodilator FEV1 ≥30% and <50% of predicted; very severe (GOLD 4): patients have post-bronchodilator FEV1 <30% of predicted. Each moderate, severe, and very severe FEV1 health state was duplicated three times to separately capture patients who had experienced no, moderate, and severe exacerbations. Exacerbations health states were defined according to ETHOS trial.12 A moderate exacerbation was characterized as an exacerbation requiring systemic corticosteroids or antibiotics for at least 3 days, and a severe exacerbation as requiring hospitalization or resulting in death.12

Figure 1 Model structure.

Abbreviations: Exac: exacerbation; FEV1: forced expiratory volume in one second.

Dashed black arrows indicate the transition between more severe FEV1 health states. Although not directly shown on this diagram, health state transitions between the moderate and very severe FEV1 health states are permitted.

The risk of experiencing an exacerbation (or subsequent exacerbation) and the risk of death were differentiated between patients who had not experienced an event and those who had had a prior event. Differentiating FEV1 health states by exacerbations permitted different risks, costs, and utility values to be applied, incorporating ‘tunnel states’. These tunnel states were sub-states within each FEV1 health state that permitted the increased resource use and utility decrease associated with moderate and severe exacerbations. Simulated patients entered each health state at the start of an exacerbation episode. A simulated patient only occupied the tunnel state for one model cycle; after that, the patient transitioned to either the post-exacerbation health state or the death state. A simulated cohort of patients entered the model in the no exacerbation health states (moderate, severe, and very severe FEV1) according to the FEV1 severity of patients at baseline in the ETHOS trial (Table 1).13 The distribution of patients in the no exacerbation health states at baseline was assumed to be equivalent for all treatments.

Patients in the cohort could not transition back to the no exacerbation health states once they experienced an exacerbation (moderate or severe). Furthermore, since COPD is considered a progressive disease, patients could not experience an improvement in lung function and transition to a less severe FEV1 health state. Death was the absorbing state to which patients could transition from all states.

All inputs were validated by an expert panel of three Spanish healthcare professionals (a pneumologist, a general practitioner, and a hospital pharmacist) to ensure the study perspective fits the current Spanish clinical practice.

Efficacy Lung Function (FEV1) and Exacerbations

The transition probabilities between health states were calculated by reviewing the effect of each treatment over the simulated cohort of patients (Table 1). Based on ETHOS trial15 data (Tables S1 and S2) at baseline, the rates of moderate and severe exacerbations were analysed for the overall patient population stratified by lung function (FEV1) severity. A single rate ratio was applied for each type of exacerbation. It was assumed that there was no increased probability of presenting exacerbations after a previous exacerbation or no long-term decrease in the effect of treatment on lung function and exacerbations for each of the treatments. All 52-week data were transformed into monthly probabilities to reflect the model cycle length.

Treatment Discontinuation

The treatment discontinuation risk (monthly probability) (Table 1) for the therapies was derived from the ETHOS trial.13 The reasons for premature treatment discontinuation were mainly adverse events, patients’ withdrawal from the trial, or lack of efficacy. The risk of treatment discontinuation was applied for the lifetime horizon.

Mortality

The mortality rates for females and males in the general population were weighted by the proportion of females and males in the ETHOS trial overall population (Table 1).13 The ETHOS trial collected data on patient mortality during the 52-week follow-up of the study.15 These data were used for the first year (monthly probability was adjusted to align with the cycle duration) (Table 1). From the second year on and beyond, COPD-related mortality was presented as the increased risk of death due to COPD compared to the risk of death for the general population at the same age (Table 1). Relative risk of death by lung function (FEV1) health state without exacerbation history were derived from the literature (moderate FEV1: 1.40; severe FEV1: 2.60 and very severe FEV1: 2.60)16 and were adjusted according to the experts’ recommendations (moderate FEV1: 1.40; severe FEV1: 2.55 and very severe FEV1: 2.65). The general population all-cause mortality rates were sourced from the life tables of the Spanish National Statistics Institute.17 It was also assumed that the risk of death by FEV1 health state post-exacerbation was equivalent to the risk of death without an exacerbation history. Severe exacerbation-related deaths were modelled separately, only patients who experienced a severe exacerbation were at risk of severe exacerbation-related deaths. This risk was applied to these patients when an exacerbation occurred. A probability of 12% was considered and reflected the estimated mortality 90 days after hospitalization for a severe exacerbation (Table 1).18,19 This input was set to 0% in the first year to avoid double-counting since overall mortality rates from the trial were used. Patients who experienced a moderate exacerbation had a risk of death depending on the FEV1 health state when the exacerbation occurred.

Utility Values

Utility values for each lung function (FEV1) severity state and utility decrements for exacerbations and AEs were included. The utility values ascribed to all FEV1 health states (exacerbations and no exacerbations) were derived from the EuroQoL 5-dimension 5-level (EQ-5D-5L) questionnaire collected in the ETHOS trial (Table 1).13 The loss of utility (dis-utility) associated with the exacerbation was applied as QALY losses per event in line with the roflumilast NICE assessment18 and the publication by Samyskin et al20 (Table 1). The QALY loss was assumed to occur during one cycle only, beyond which the patient’s quality of life returned to the level associated with their lung function status.

In addition, disutilities associated with AEs were ascribed at the beginning of the model when the entire cohort was at risk of treatment-related AEs to ensure that the full impact of AEs was captured. The AEs-related disutility was assumed to be captured in the underlying EQ-5D-5L analysis of the ETHOS trial data.13 Therefore, no additional AEs-related utility decrements were considered to avoid double-counting in the base-case analysis.

Costs

Costs included treatment-specific costs (drug acquisition and subsequent treatment, rescue medication, and adverse events) and disease management costs: by lung function (emergency room visit, specialist visit, primary care visit, nursing visit, telephone calls to physician, and telephone nursing consultation) and by exacerbation (hospitalization in the intensive care unit, hospitalization in the coronary care unit, hospitalization in the general ward, treatment with oral corticosteroids, ambulance transport, and primary care visit). All costs were estimated in euros 2021.

Drug acquisition costs were derived from an official database in Spain21 (Table 2) (plus the rebate of 7.5% according to Spanish regulation RDL 8/2010, as appropriate22). All LAMA/LABA therapies had the same prices. However, ICS/LABA treatment costs were calculated as a price average.21 Upon discontinuation of the initial drug, only drug acquisition costs were considered. They were calculated as weighted average costs based on market share determined by the experts. Patients receiving BUD/GLY/FOR assumed that 95% of cases continued to receive triple therapy treatment while receiving the additional subsequent treatment roflumilast.23 The remaining 5% of patients de-escalated to double therapy (4% LAMA/LABA and 1% ICS/LABA). Patients that received LAMA/LABA treatment moved onto triple therapy in 90% of cases and received an ICS/LABA therapy as a subsequent treatment in 10%. Patients initially receiving ICS/LABA treatment moved onto triple therapy treatment in 100% of cases.

Table 2 Costs Inputs Used in the Model

The number of inhalations per administration and the number of administrations per day by treatment were derived from the product summary of product characteristics (SmPC) for each treatment.24 The efficacy of subsequent treatment was assumed equal to BUD/GLY/FOR triple therapy, and subsequent treatment cost to be the lowest of all triple therapies in the Botplus database21 (Table 2). This assumption favoured the dual therapies opting for the most conservative approach. The rescue medication-related cost was included as the average number of rescue medication inhalations required per month per treatment and by FEV1 state (Table 2). Rescue medication costs were calculated using the cost of salbutamol21 and the mean daily number of rescue inhaler uses observed in ETHOS trial.13 No rescue medication costs were applied after treatment discontinuation. The experts decided to include in the model those adverse events reported with a frequency greater than 3% and the cardiovascular events, both based on the ETHOS trial.13 Therefore, nasopharyngitis, upper respiratory tract infection, confirmed pneumonia, bronchitis, and confirmed major adverse cardiovascular events were included. Unit costs were obtained from the Spanish healthcare costs database25 (Table 2). More details about resource use and costs for AEs are shown in Table S3. The costs associated with AEs were ascribed in the first cycle only.

Lung function and exacerbation management costs, such as resource use (annual frequency), were derived from the ETHOS trial13 and were validated according to the experts’ recommendations. In addition, unit costs were obtained from the Spanish healthcare costs database25 (Table 2). More details about resource use and costs for lung function and exacerbation management are shown in Table S4. Disease management costs were ascribed separately for patients who discontinued and were not included as a subsequent treatment cost.

Model Outputs

The outputs were measured as total costs, the number of exacerbations, life years (LYs), and quality-adjusted life-years (QALYs). The cost-effectiveness was expressed as incremental cost effectiveness ratio (ICER) per exacerbation avoided and per life year (LY) gained, and per quality-adjusted life-year (QALY) gained (ICUR). Cost-effectiveness was assessed using a willingness-to-pay (WTP) threshold of €25,000/QALY.26

An annual discount rate of 3% was applied to costs and benefits, following Spanish guidelines.27

Sensitivity Analysis

One-way (OWSA), scenario, and probabilistic sensitivity analyses (PSA) were conducted to explore the parameter uncertainty of the model inputs. OWSA was estimated using 95% CI from ETHOS trial or ±20% the value defined for the base case. The result of the OWSA was plotted on a tornado diagram with the 10 most sensitive parameters presented.

Two scenario analysis was performed based on experts’ opinions. In the first one, the characteristics of the population (average age of patients and rates for females and males) were modified. Although patients from Spanish centers participated in the ETHOS trial, including data from the Spanish population was considered necessary, taking into account the perspective of our analysis; Therefore, instead of the data from the ETHOS trial (64.7 and 40.3% female, respectively), data from an epidemiological study on COPD (60 and 52.6% female, respectively) in the Spanish population were used.28

As the ETHOS trial lasted only 52 weeks, a second scenario analysis with a 1-year time horizon was performed, calculating the annual costs and showing the differences between treatments on an annual basis.

The PSA was performed using a Monte Carlo simulation with 10,000 iterations. The gamma distribution was used for cost and resource use. The beta distribution was used for probabilities and utilities. Finally, a log-normal distribution was applied to the relative treatment effects (relative risk/rate ratio).

Results Base-Case Analysis

Over a lifetime horizon, the rate of exacerbations was lower with BUD/GLY/FOR (12.80) than LAMA/LABA (13.36) and ICS/LABA (13.23). In addition, BUD/GLY/FOR had more LYs and QALYs (10.32 and 7.55, respectively) compared to LAMA/LABA (10.14 and 7.41, respectively) and ICS/LABA (10.06 and 7.32, respectively). The total costs were higher for the BUD/GLY/FOR (€, 16,520.45) vs LAMA/LABA (€, 16,044.47) and vs ICS/LABA (€, 15,489.48). Drug initial costs were lower in the BUF/GLY/FOR (€, 1,893.19) than LAMA/LABA (€, 2,814.86) and similar to ICS/LABA (€, 1,978.51) (Table 3). Subsequent treatment was the only cost higher for BUD/GLY/FOR. The estimated incremental cost per exacerbation avoided for BUD/GLY/FOR was €850.95 vs LAMA/LABA and €2,422.26 vs ICS/LABA (Table 3). The incremental cost per LY gained (ICER) for BUD/GLY/FOR vs LAMA/LABA and vs ICS/LABA was €2,733.38 and €4,111.15, and the ICUR was €3,461.19 and €4,545.24, both respectively.

Table 3 Results of the Base-Case Analysis

Sensitivity Analyses

The OWSA showed that base-case results were robust to changes in input parameters. The most sensitive parameter was the subsequent treatment costs for BUD/GLY/FOR and LAMA/LABA or ICS/LABA (Figures 2 and Figure 3); even so, BUD/GLY/FOR continued to be cost-effective.

Figure 2 OWSA BUD/GLY/FOR vs LAMA/LABA.

Abbreviations: BUD/GLY/FOR: budesonide/glycopyrrolate/formoterol fumarate. FEV1: forced expiratory volume in one second; ICS: inhaled corticosteroid; ICUR: incremental cost-utility ratio; LABA: long-acting beta adrenoreceptor agonist; LAMA: long-acting muscarinic antagonist. RR: Relative Risks.

Figure 3 OWSA BUD/GLY/FOR vs ICS/LABA.

Abbreviations: BUD/GLY/FOR: budesonide/glycopyrrolate/formoterol fumarate. FEV1: forced expiratory volume in one second; ICS: inhaled corticosteroid; ICUR: incremental cost-utility ratio; LABA: long-acting beta adrenoreceptor agonist; LAMA: long-acting muscarinic antagonist. RR: Relative Risks.

In the model developed, the QALY difference observed between compared treatments resulted from different utility values for each lung function (FEV1) severity state and utility decrements for exacerbations and AEs. Therefore, a lower utility value of severe FEV1 health status would lead to fewer QALYs and higher ICUR. However, contrary to expectations, with ICS/LABA, a lower utility value of the very severe health state FEV1 caused QALYs to be higher. This is because the probability of progressing to a very severe FEV1 state in ICS/LABA treatment was faster than in BUD/GLY/FOR. Therefore, the number of individuals accumulated in very severe FEV1 state is higher and QALYs are higher too in ICS/LABA treatment.

In the first alternative scenario, the characteristics of the population (ETHOS clinical trial population in base-case) were changed for the characteristics of Spanish population. In this case, the ICUR of BUD/GLY/FOR versus LAMA/LABA was €1,707.76, and versus ICS/LABA was €3,935.68, remaining lower than the WTP threshold (25,000€/QALY), which also supports the results’ robustness. The second alternative scenario (1-year time horizon) showed that LAMA/LABA and ICS/LABA were dominated by BUD/GLY/FOR so, BUD/GLY/FOR is both clinically superior and cost saving.

The PSA showed that 91.32% vs LAMA/LABA and 99.29% vs ICS/LABA of simulations were in the northeast quadrant, indicating that BUD/GLY/FOR was cost-effectiveness vs dual therapies by a willingness to pay (WTP) threshold of €25,000 per QALY gained (Figures 4 and Figure 5).

Figure 4 PSA BUD/GLY/FOR vs LAMA/LABA.

Abbreviations: ICER: incremental cost-effectiveness ratio; QALYs: quality-adjusted life-year.

Figure 5 PSA BUD/GLY/FOR vs ICS/LABA.

Abbreviations: ICER: incremental cost-effectiveness ratio; QALYs: quality-adjusted life-year.

The cost-effectiveness acceptability curve showed that BUD/GLY/FOR (98.8%) has a higher probability of being cost-effectiveness than LAMA/LABA (0.6%) and ICS/LABA (0.6%) at a WTP threshold of €25,000/QALY (Figure 6).

Figure 6 Cost-effectiveness acceptability curve.

Abbreviations: BUD/GLY/FOR: budesonide/glycopyrrolate/formoterol fumarate.ICS: inhaled corticosteroid; LABA: long-acting beta adrenoreceptor agonist; LAMA: long-acting muscarinic antagonist.

Discussion

COPD is a highly prevalent chronic disease that produces a great economic impact and is projected to increase over the coming decades, so strategies to address this disease are necessary.1 The results of this cost-effectiveness analysis based on a Markov model showed that triple therapy BUD/GLY/FOR is dominant with a 1-year time horizon and cost-effective treatment option over lifetime horizon for patients with moderate to severe COPD compared to dual therapies LAMA/LABA and ICS/LABA from the Spanish NHS perspective. BUD/GLY/FOR was associated with a lower total exacerbation per patient (12.80) than LAMA/LABA (13.36) and ICS/LABA (13.23) and more LYs (10.32 vs 10.14 and 10.06, respectively) and QALYs (7.55 vs 7.41 and 7.32, respectively) per patient gained. The analysis resulted in an ICER of €2,733.38 and €4111.15 per LY gained, and an ICUR of €3,461.19 and €4,545.24 per QALY gained, for BUD/GLY/FOR vs LAMA/LABA and ICS/LABA respectively, both below the threshold of €25,000/QALY. Furthermore, the results were robust to various sensitivity analyses and showed that variation in critical parameters did not impact the results in a great deal.

Others cost-effectiveness analyses of different single-inhaler triple therapy in patients with COPD and risk of exacerbations versus dual therapies were identified.29–31 All of them showed triple therapy was cost-effective vs dual therapies with an ICUR of £ 6418 (£ 2018, UK),29 £ 4104 (£ 2018, UK),29 £ 1098 (£ 2015/2016, UK)28 for a lifetime horizon in line with our analysis, including one from the Spanish perspective with an ICUR of € 642 (€ 2019) over a 3-year time horizon.31 Results of different modelling approaches for the same decision problem showed similar outcomes as our study, strengthening the robustness of our results. Furthermore, these findings suggest that the data on cost-effectiveness in the UK and Spain may extend to other countries.

This study has some limitations, common to all cost-effectiveness analyses using clinical trial data and Markov models, mainly due to assumptions such as QALY loss was assumed to occur during one cycle only. Although validated by clinical experts, these assumptions introduce uncertainty into the findings. Firstly, the time horizon was for a lifetime; however, the ETHOS trial duration was 52 weeks, so it was assumed that there was no long-term decrease in the treatment effect on lung function and exacerbations for each of the treatments and the efficacy of subsequent treatment was equalled to BUD/GLY/FOR triple therapy. Alternatively, several sensitivity analyses were conducted to ascertain the impact of long-term efficacy assumptions and confirmed that the results were not impacted. Specifically, in the scenario where the time horizon was one year, the results favoured BUD/GLY/FOR therapy. Secondly, the approach does not fully capture the link between exacerbations, symptoms, lung function and the exact exacerbation history to the extent that a patient-level simulation model would. A Markov model is unable to ascribe probabilities of future health state transitions according to the nature or timing of past events.32 However, overcoming this was possible due to “tunnel states”, allowing the analysis to capture exacerbation management costs and the utility decrement and increased risk of mortality associated with the exacerbation events experienced in an individual cycle. In addition, the model underestimates the mortality risk for patients with multiple exacerbations. In this regard, the experts indicated that the probability of presenting an exacerbation after the “post-exacerbation” health state does not depend only on the previous exacerbation, as there are other issues to consider (eg, smoking, corticosteroid treatments, etc.). Therefore, the experts assumed that the best approach was not to include it in any of the comparators. Thirdly, subsequent treatment discontinuation was not considered. Patients who went from double to triple therapy stayed there until the end, and those who were on BUF/GLU/FOR went on to BUF/GLU/FOR plus roflumilast (95%) until the end of their lives. Therefore, those who received dual therapy would never receive BUF/GLU/FOR plus roflumilast, from which they could benefit from so that this model can be considered conservative. Nevertheless, BUD/GLU/FOR remained a cost-effective treatment option considering a number of uncertainties across all sensitivity and scenario analyses, suggesting that the results were robust when adapting to the reality of COPD management in clinical practice and that the developed model is reliable.

In addition to the robustness of the sensitivity analysis results, a strength of this study is that the proposed economic model structure was appropriate for the decision problem. The model considered the inclusion of different severity levels of exacerbation which allowed to include costs and HRQoL decrements as captured. In addition, this allowed different risk of disease progression for patients who have experienced an exacerbation after initiation of treatment.

Moreover, all inputs were validated according to the experts’ recommendations.

Our study demonstrated a beneficial effect of BUD/GLY/FOR vs dual therapies on the annual rate of exacerbations, symptoms, and health-related quality of life following guidelines’ recommendations9 and, consequently, demonstrating to be cost-effective in terms of LYs and QALYs.

In conclusion, the results of this cost-effectiveness analysis based on a Markov model showed that treatment with the fixed-dose combination of BUD/GLY/FOR was cost-effective compared with dual therapies. These results may help inform future decision-making processes in the Spanish NHS.

Funding

The study was funded by Astrazeneca Farmacéutica Spain S.A.

Disclosure

Dr Juan Antonio Trigueros reports grants from GlaxoSmithKlein, Menarini Lab, and Chiesi, during the conduct of the study. Dr Adolfo Baloira reports personal fees for advice and conferences from Astra Zeneca, outside the submitted work. Mrs Susana Aceituno, Mrs Ana Calvo and Drs Miriam Prades report their employer, Outcomes’10, has received fees from AstraZeneca Farmacéutica Spain, S.A. for its contribution to the project coordination as well as to the drafting of this manuscript. Carolina Touron, Anisia Martínez, and Covadonga Torres are employees of AstraZeneca Farmacéutica Spain, S.A. The authors report no other conflicts of interest in this work.

References

1. Global Initiative for Chronic Obstructive Lung Disease. Report; 2022. Available from: https://goldcopd.org/2022-gold-reports-2/. Accessed June1, 2022.

2. NHS. Chronic obstructive pulmonary disease (COPD). Available from: https://www.nhs.uk/conditions/chronic-obstructive-pulmonary-disease-copd/treatment/. Accessed June1, 2022.

3. World Health Organization. Available from: https://www.who.int/news-room/fact-sheets/detail/chronic-obstructive-pulmonary-disease-(copd). Accessed July1, 2022.

4. Soriano JB, Kendrick PJ, Paulson KRet al. Prevalence and attributable health burden of chronic respiratory diseases, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet Respir Med. 2020;8(6):585–596.

5. National Center for Epidemiology. Carlos III Institute of Health. Availabre from: http://raziel.cne.isciii.es/. Accessed November1, 2021.

6. Gómez Sáenz JT, Quintano Jiménez JA, Hidalgo Requena A, et al. Enfermedad pulmonar obstructiva crónica: morbimortalidad e impacto sanitario [Chronic obstructive pulmonary disease: morbimortality and healthcare burden]. Semergen. 2014;40(4):198–204. Spanish.

7. Izquierdo JL. The burden of COPD in Spain: results from the confronting COPD survey. Respir Med. 2003;97(Suppl C):S61–S69. doi:10.1016/S0954-6111(03)80026-4

8. Soler-Cataluña JJ, Martínez-García MA, Román Sánchez P, Salcedo E, Navarro M, Ochando R. Severe acute exacerbations and mortality in patients with chronic obstructive pulmonary disease. Thorax. 2005;60(11):925–931.

9. Miravitlles M, Calle M, Molina J, et al. Actualización 2021 de la Guía Española de la EPOC (GesEPOC). Tratamiento farmacológico de la EPOC estable [Spanish COPD guidelines (GesEPOC) 2021: Updated pharmacological treatment of stable COPD]. Arch Bronconeumol. 2021;58:69–81.

10. Skolnik NS, Nguyen TS, Shrestha A, Ray R, Corbridge TC, Brunton SA. Current evidence for COPD management with dual long-acting muscarinic antagonist/long-acting β2-agonist bronchodilators. Postgrad Med. 2020;132(2):198–205. doi:10.1080/00325481.2019.1702834

11. The GLOBAL STRATEGY FOR DIAGNOSIS, MANAGEMENT AND PREVENTION of COPD (updated 2021), the pocket guide (updated 2021) and the complete list of references examined by the committee. Available from: https://goldcopd.org/. Accessed June1, 2022.

12. Rabe KF, Martinez FJ, Ferguson GT, et al. Triple inhaled therapy at two glucocorticoid doses in moderate-to-very-severe COPD. N Engl J Med. 2020;383(1):35–48. doi:10.1056/NEJMoa1916046

13. Study to assess the efficacy and safety of PT010 relative to PT003 and PT009 in subjects with moderate to very severe COPD (ETHOS). Available from: https://clinicaltrials.gov/ct2/show/NCT02465567. Accessed November1, 2021.

14. Global Initiative for Chronic Obstructive Lung Disease. Report; 2017. Available from: https://goldcopd.org/wp-content/uploads/2017/02/wms-GOLD-2017-FINAL.pdf. Accessed June1, 2022.

15. Data on file.

16. Shavelle RM, Paculdo DR, Kush SJ, Mannino DM, Strauss DJ. Life expectancy and years of life lost in chronic obstructive pulmonary disease: findings from the NHANES III Follow-up Study. Int J Chron Obstruct Pulmon Dis. 2009;4:137–148. doi:10.2147/COPD.S5237

17. Spanish National Statistics Institute. Life tables;2020. Available from: https://www.ine.es/. Accessed November1, 2021.

18. National Institute for Health and Care Excellence. Roflumilast for treating chronic obstructive pulmonary disease 2017. Available from: https://www.nice.org.uk/guidance/ta461. Accessed November1, 2021.

19. National COPD Audit Programme. COPD: who cares when it matters most? – outcomes report; 2014. Available from: https://www.rcplondon.ac.uk/projects/outputs/copd-who-cares-when-it-matters-most-outcomes-report-2014. Accessed June1, 2022.

20. Samyshkin Y, Kotchie RW, Mörk A-C, Briggs AH, Bateman ED. Cost-effectiveness of roflumilast as an add-on treatment to long-acting bronchodilators in the treatment of COPD associated with chronic bronchitis in the United Kingdom. Eur J Health Econ. 2014;15(1):69–82.

21. Botplus. Online database of drug prices in Spain. Available from: https://botplusweb.portalfarma.com/botplus.aspx. Accessed November1, 2021.

22. Deductions for medications in Spain. Available from: https://www.sanidad.gob.es/profesionales/farmacia/notasInfor.htm. Accessed November1, 2021.

23. Rabe KF. Update on roflumilast, a phosphodiesterase 4 inhibitor for the treatment of chronic obstructive pulmonary disease. Br J Pharmacol. 2011;163(1):53–67.

24. AEMPS. CIMA. Available from: https://cima.aemps.es/cima/publico/home.html. Accessed November1, 2021.

25. Gisbert R, Brosa M Spanish health costs and cost-effectiveness ratios database: eSalud. Barcelona: Oblikue Consulting, S.L.; 2007. Available from: http://www.oblikue.com/bddcostes/. Accessed November1, 2021.

26. Sacristán JA, Oliva J, Campillo-Artero C, et al. ¿Qué es una intervención sanitaria eficiente en España en 2020? [What is an efficient health intervention in Spain in 2020?] Gaceta Sanitaria. 2020;34(2):189–193 .

27. López-Bastida J, Oliva J, Antoñanzas F, et al. Spanish recommendations on economic evaluation of health technologies. Eur J Health Econ. 2010;11(5):513–520. doi:10.1007/s10198-010-0244-4

28. Soriano JB, Alfageme I, Miravitlles M, et al. Prevalence and Determinants of COPD in Spain: EPISCAN II. Arch Bronconeumol. 2021;57(1):61–69.

29. Fenwick E, Martin A, Schroeder M, et al. Cost-effectiveness analysis of a single-inhaler triple therapy for COPD in the UK. ERJ Open Res. 2021;7:00480–2020. doi:10.1183/23120541.00480-2020

30. Martin A, Shah D, Ndirangu K, et al. Is single-inhaler triple therapy for COPD cost-effective in the UK? The IMPACT trial. ERJ Open Res. 2022;8(1):00333–2021.

31. Schroeder M, Benjamin N, Atienza L, et al. Cost-effectiveness analysis of a once-daily single-inhaler triple therapy for patients with chronic obstructive pulmonary disease (COPD) using the FULFIL trial: a Spanish perspective. Int J Chron Obstruct Pulmon Dis. 2020;15:1621–1632. doi:10.2147/COPD.S240556

32. Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation. Oxford University Press; 2006.