Comparing Costs and Healthcare Resource Utilization (HCRU) Using LAMA versus LABA/ICS at Treatment Initiation for COPD: Findings from CITRUS (Comparing the Incidence of Tiotropium and ICS/LABA in Real-World Use in South Korea) Study

Introduction

Chronic obstructive pulmonary disease (COPD) is a common, preventable, and treatable disease. The global prevalence of COPD is estimated to be 10.3%.¹ COPD was the third leading cause of death worldwide in 2019,² and the most common cause of death from non-communicable diseases (NCDs) in 2016.³ COPD also causes significant healthcare resource utilization and economic burden across countries, including South Korea.^4–6

According to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines, initial therapy should be chosen based on the severity of dyspnea and the risk of exacerbation. Inhaled bronchodilators such as long-acting muscarinic antagonists (LAMA), long-acting beta-2 agonists (LABA), or a combination of both are the critical components of COPD treatment. Although the American Thoracic Society (ATS) and GOLD 2023 update recently announced that LAMA/LABA combination therapy is more effective than LAMA or LABA monotherapy^1,7 in patients with dyspnea or exercise tolerance, LAMA monotherapy still remains a widely used and valuable component in COPD treatment,^8,9 especially in GOLD group A.

Unlike asthma, where inhaled corticosteroids (ICS) are an essential part of treatment, the GOLD 2023 guidelines do not encourage the use of LABA/ICS in COPD.^1,10 The use of ICS as a component of triple therapy is reserved for patients with recurrent exacerbations and high blood eosinophil count.^1,11 However, ICS is widely prescribed in real-world practice outside of the above recommendation.^12,13 A recent US study showed that LABA/ICS is used in 28.5% of patients with moderate or severe exacerbations, while only 5.7% of patients utilized triple therapy.¹⁴

The INSPIRE study demonstrated that LAMA monotherapy is similar to LABA/ICS in exacerbation prevention.¹⁵ Additionally, in a real-world study, treatment with LABA/ICS showed no exacerbation risk reduction over LAMA except in a group of patients with eosinophils above 4%.¹⁶ However, no real-world study compares LAMA with LABA/ICS treatment regarding medical costs and healthcare resource utilization (HCRU). This study compared all-cause medical costs and HCRU between LAMA and LABA/ICS during a three-year follow–up period using a health insurance claims database in primary and secondary care. We also explored COPD-related medical costs and HCRU rates between the treatment groups.

Materials and Methods

Study Design

This study was a retrospective, non-interventional cohort study using a national health insurance claims database from the National Health Insurance Service (NHIS) from South Korea.¹⁷ The NHIS data provides demographic characteristics and healthcare information of COPD patients, including comorbidities and medical treatment. The study period was from January 1, 2002 through April 30, 2016, consisting of the baseline period of 36 months and the patient enrollment period of 124 months (January 1, 2005 to April 30, 2015). The definition of index date was the date of the first prescription date of LAMA or fixed-dose combination (FDC) of LABA/ICS for COPD treatment during the enrollment period. The baseline period is defined as 36 months before the index date to assess patients’ baseline demographic information and comorbidities. Enrolled patients were followed for a minimum of 12 months and a maximum of 124 months from the index date (Figure 1A). The follow-up ended when the LAMA or LABA/ICS was discontinued, or the study period ended.

Figure 1 (A) Study scheme. (B) Study flow chart. Abbreviations: LAMA, long-acting muscarinic antagonist; LABA/ICS, long-acting beta-2 agonist plus inhaled corticosteroids; COPD, chronic obstructive pulmonary disease; MPR, medical possession rate; LTRA, leukotriene receptor antagonist; ILD, interstitial lung disease; IPF, idiopathic pulmonary fibrosis.

Study Population

We used the International Classification of Diseases, Tenth Revision (ICD-10) code to confirm COPD. Patients with ICD-10 code (J43.x-44.x, except J430), as the primary diagnosis or within the first four additional diagnoses, were considered to have COPD.^18,19 Then, we enrolled the patients for the analysis who met the following criteria: Patients aged ≥55 years as of the index date; Patients who were prescribed with LAMA monotherapy or LABA/ICS FDC for the first time during the enrollment period; Patients who were prescribed with LAMA monotherapy or LABA/ICS FDC at least twice within 12 months from the index date. The LAMAs included tiotropium bromide, aclidinium bromide, glycopyrronium bromide, and umeclidinium bromide. The FDCs of ICS and LABA were salmeterol and fluticasone, formoterol and budesonide, vilanterol and fluticasone furoate, and formoterol and fluticasone. We excluded the current asthma patients with active treatment at the index date but included patients with past asthma history who were recorded with ≥1 claim with ICD-10 codes for asthma in the previous 36 months before the index date. We also minimized the selection of asthma patients by excluding patients aged <55 years.

We also excluded the patients on any combinations other than LAMA or LABA/ICS at the index date or who received any inhaler treatments in the previous 12 months before the index date. Patients with their medical possession rate (MPR) ≤80% in the first 12 months were also excluded.

Study Outcome

The co-primary outcomes of this study were all-cause medical costs and all-cause HCRU over a three-year follow-up period. COPD-related medical costs and COPD-related HCRU were secondary outcomes.

Medical costs per-person-per-month (PPPM) were obtained overall and separately based on the type of service (inpatient visit, outpatient visit, or pharmacy visit). All-cause and COPD-related medical costs, consisting of services covered by the NHIS and co-payment, included the total medical, hospitalization, outpatient, and pharmacy costs during 12, 24, and 36 months of follow-up, were calculated. The medical costs were shown as the mean cost PPPM. All costs were converted to US dollars using the average 2019 exchange rate: 1 USD = 1166.51 KRW (South Korean Won).

All-cause and COPD-related HCRU included any services directly provided by the healthcare system during 12, 24, and 36 months of follow-up period from the index date, including inpatient visit, outpatient visit, and pharmacy visit. HCRU rate was presented as the number of any medical visits PPPM. For outpatient visits, HCRU was confined to outpatient visits with the ICD-10 code of COPD (J43.x–J44.x, except J430) with the prescription of COPD-related medication. For inpatient visits, analyses were confined to hospitalization with the ICD-10 codes of COPD or COPD-related diseases (pneumonia: J12.x–J17.x, pulmonary thromboembolism: I26, I26.0, and I26.9; dyspnea: R06.0; or acute respiratory distress syndrome: J80) with the prescription of COPD-related medication.^20–22

Propensity Score Matching and Statistical Analysis

We used the propensity score (PS) matching to reduce the potential confounding and to balance comparability between LAMA and LABA/ICS treatment groups, because the initial treatment is likely to be selected according to the demographics and baseline characteristics of the patients. PS was estimated using multiple logistic regression analysis based on age, sex, income quartiles, modified Charlson Comorbidity Index (CCI), asthma history, and COPD exacerbation history.

Patients’ baseline demographic and clinical characteristics were analyzed using descriptive statistics. Continuous and categorical variables were summarized using descriptive statistics. The comparison of descriptive statistics between study groups was evaluated by t-test for continuous variables and chi-squared test for categorical variables. HCRU rates and medical costs were also analyzed using descriptive statistics and presented as the number of inpatient hospitalizations, outpatient visits, outpatient pharmacy dispensations, and the mean cost PPPM. In addition, we conducted subgroup analyses according to sex, age group, and history of asthma or COPD exacerbation.

We compared the difference in medical costs and HCRU rates between the study groups by t-test for parametric and Wilcoxon rank-sum test for non-parametric variables. In addition, the medical costs and HCRU rates were calculated and compared across the study groups using a generalized linear model (GLM) with gamma distribution for costs and Poisson, negative binomial, or zero-inflated negative binomial distribution for HCRU rates based on the overdispersion parameter and the number of zeros. All statistical tests were two-sided with a significance level of 0.05. We used the SAS^® 9.4 software (SAS Institute, North Carolina, US) via SAS Enterprise Guide version 6.1 for the statistical analysis.

Ethics Approval

This study protocol followed the Declaration of Helsinki and was approved by the ethical committee of Konkuk University Medical Center (Institutional Review Board No.: KUMC2020-06-013). Informed consent was waived because only de-identified data were used for analytical purposes.

Results

Baseline Characteristics

Figure 1B shows the flow chart of the study. The total of 1,285,717 patients aged at least 40 years and diagnosed with COPD from January 1, 2002 to April 30, 2016 was screened. Among them, 112,955 patients were prescribed with either LAMA or LABA/ICS at least twice within 12 months from the index date. After the application of the aforementioned exclusion criteria and PS matching, 4888 COPD patients were included in this study: 2444 in the LAMA group and the rest 2444 in LABA/ICS group. Table 1 shows the demographic characteristics of each treatment group. The mean age of the enrolled patients was 69.7 years, and males comprised 75% of the cohort. About 10% of the patients were treated at clinics without inpatient units while 90% of the patients were treated in secondary care facilities. Twenty-five percent of the patients had a history of COPD exacerbation, and 48% of them had a history of asthma.

Table 1 Demographic and Clinical Characteristics in PS Matched Population

The follow-up period was significantly longer in the LAMA treatment group than in the LABA/ICS treatment group (812.3 days vs 681.1 days). However, other baseline characteristics did not differ significantly between treatment groups.

Medical Costs

For three years of follow-up, the adjusted all-cause medical cost in the LAMA and LABA/ICS treatment group was 403.08 USD and 474.50 USD PPPM, respectively (Figure 2, Tables 2 and 3). The cost ratio was 1.18 (95% confidence interval [CI]=1.10–1.26, p<0.0001). The COPD-related medical cost was also lower in LAMA compared to LABA/ICS treatment group (216.37 vs 267.32 USD PPPM, cost ratio 1.24, 95% CI=1.13–1.35, p<0.0001). These differences were maintained irrespective of the site of care (Figure 2). Regardless of age and history of asthma, the all-cause and COPD-related medical costs in the LAMA treatment group were lower than those in the LABA/ICS treatment group (Tables 2 and 3). For the patients without a history of COPD exacerbations, both costs were significantly lower in the LAMA treatment group. LAMA reduced the all-cause medical cost for patients with a high risk of exacerbation. However, the COPD-related medical cost was similar between the treatment groups for patients with a high risk of exacerbation (Tables 2 and 3).

Table 2 Adjusted All-Cause Medical Cost* Up to 3 Years in the Matched Population

Table 3 Adjusted COPD-Related Medical Cost* Up to 3 Years in the Matched Population

Figure 2 All-cause and COPD-related cost up to 3 years in the matched population. (a) Adjusted all-cause cost up to 3 years in the matched population. (b) Adjusted COPD-related cost up to 3 years in the matched population.

Healthcare Resource Utilization (HCRU)

Table 4 shows all-cause HCRU for three years of follow-up. The all-cause HCRU rate was not significantly different between the LAMA and LABA/ICS treatment groups (2.81 vs 2.73 visits PPPM, respectively, p=0.2875). In subgroup analyses, inpatient hospitalization and ER visit rates were significantly lower in the LAMA treatment group than in the LABA/ICS treatment group. LABA/ICS treatment was associated with higher all-cause HCRU than LAMA treatment for patients without asthma history (Table S1). COPD-related HCRU was lower in the LABA/ICS group compared to the LAMA group (0.60 vs 0.66 visits PPPM, respectively, p<0.0001), resulting from a significant difference in outpatient visits (Table 4).

Table 4 All-Cause and COPD-Related Healthcare Utilization* Up to 3 Years

Discussion

In this study, we compared the all-cause and COPD-related medical costs and healthcare resource utilization (HCRU) in COPD patients of initial treatment with either LAMA or LABA/ICS. We found that initiation with LAMA treatment for COPD costs an average of 403.08 USD PPPM for all-cause medical cost and 216.37 USD PPPM for COPD-related medical cost which were lower than those of patients initiating treatment with LABA/ICS. Most of the differences resulted from the inpatient hospitalization cost, which was lower in the LAMA treatment group. The cost ratio for all-cause and COPD-related medical costs with LABA/ICS treatment was 1.18 and 1.24, respectively (p<0.0001, Tables 2 and 3), compared to the LAMA treatment. This economic benefit was obtained by decreasing the risk of severe exacerbation, defined as worsening of symptoms requiring hospitalization.¹ As exacerbation is a significant contributor to the economic burden of COPD and most medical expenses of exacerbation arise from hospitalization,^23,24 this finding can be interpreted as LAMA can reduce severe exacerbations compared to LABA/ICS for treatment naïve COPD patients. However, the economic benefit of starting COPD treatment with LAMA disappeared in patients with a high risk of exacerbation (frequent exacerbator or severe exacerbator in the previous year).

Compared with the results regarding medical costs, the analyses of HCRU showed a different pattern. The all-cause HCRU rate was not significantly different between the LAMA and LABA/ICS treatment groups (2.81 vs 2.73 visits PPPM, respectively, p=0.2875). The COPD-related HCRU was higher in the LAMA group than the LABA/ICS group (0.60 vs 0.66 visits PPPM, respectively, p<0.0001). This finding was not consistent with that of a previous large, randomized study, which reported a similar rate of HCRU between the treatment groups.¹⁴ However, the main difference in HCRU was due to the difference in outpatient visits in our study. This pattern was preserved in patients without a history of asthma (Tables S1 and S2). As frequent outpatient visits can reduce the risk of COPD exacerbation by earlier and more prompt adjustment of inhaler therapy,²⁵ this finding of more outpatient visits in the LAMA treatment group can be another explanation for decreased severe exacerbations and reduced medical costs described above.

Our real-world findings of reduced medical costs in the LAMA compared to the LABA/ICS treatment group provided more rationale for using LAMA bronchodilators over LABA/ICS combination as an initial treatment for the patients with low risk of exacerbation, showing the reduced medical cost directly and the increased adherence indirectly, in addition to the delayed escalation to triple therapy.¹⁷ The findings of our study also provide another aspect, medical costs, to the recent changes in the GOLD 2023 update, in which the role of LABA/ICS in COPD is much diminished and the addition of ICS at treatment initiation is only suggested as a component of triple therapy for GOLD risk group E.¹ Furthermore, although other recent real-world comparative studies showed that LABA/ICS was as effective as LAMA or LAMA/LABA combination in preventing exacerbation, in the same studies LABA/ICS was associated with pneumonia more than LAMA or LAMA/LABA.^16,26 Therefore, our study adds financial insight to the recent recommendation in which the use of ICS should be limited to patients with a high risk of exacerbation, preferably as a component of triple therapy.¹

The recently released ATS guidelines for COPD recommend the LAMA/LABA combination for symptomatic patients over LAMA monotherapy. However, there was no consideration of medical cost in this recommendation.²⁷ Although several studies suggest superiority of LAMA/LABA compared to LAMA monotherapy in exacerbation prevention,^28–30 there are also reports showing adequacy of LAMA monotherapy regarding exacerbation.^31–34 As lower financial status and poverty are consistently associated with COPD,¹ initiation with LAMA can be an acceptable option in treating COPD patients with a low risk of exacerbation, especially in low-income countries.

The strength of our study was that we used propensity score matching to reduce the confounding effect. Also, the operational definition of COPD used in this study is well replicated^18–22 and validated in several previous studies,^12,35,36 suggesting that the results of our study reflected the actual real-world situation of COPD management well enough. Moreover, we enrolled treatment naïve patients, which contributes to the existing literature with a different type of study population in COPD assessing economic burden. However, this study has several limitations. This study was based on a health insurance claims database, so we could not assess lung function data, inhaler adherence or technique. Even though we used propensity score matching, some residual confounding from missing covariates, including baseline lung function measures and symptom scores, might influence the results. Also, although we showed that LAMA and LABA/ICS group did not differ in the history of exacerbations, other severity measures such as mMRC (Modified Medical Research Council) grade were not given. Furthermore, we did not investigate the influence of smoking on the inhaler medications such as ICS, which could have contributed to the utilization of medical resources in our patients. For inhaler adherence, we only enrolled patients with medical possession rate (MPR) over 80% in the first 12 months, and with such inclusion criterion we think that the issue of adherence was well controlled. Another limitation of our study is relatively long study period (January 1, 2002 through April 30, 2016), which could contain influences of advancements in COPD treatment and inhaler devices.

In summary, COPD patients initiating treatment with LAMA were associated with lower all-cause and COPD-related medical costs than those starting with LABA/ICS. There was no difference in all-cause HCRU but higher COPD-related HCRU in patients initiating treatment with LAMA, resulting from a significant difference in outpatient visits. LAMA monotherapy is a cost-efficient option for treatment initiation of symptomatic COPD.