Chronic Obstructive Pulmonary Disease (COPD) is characterized by persistent airflow limitation and is a major cause of mortality and morbidity worldwide.1 The aim of treatment of COPD is to reduce symptoms, decrease the frequency and severity of exacerbations and improve exercise tolerance and quality of life.2,3 Pharmacological management is based on long-acting bronchodilators (LABD), either a long-acting muscarinic antagonist (LAMA) or a long-acting β2-agonist (LABA) or both, and inhaled corticosteroids (ICS).2,3 Current recommendations suggest to start with one LABD and step up to dual or triple therapy (TT) (ICS/LABA/LAMA), as necessary.2 Therefore; TT is recommended only for patients with frequent exacerbations whose symptoms are not adequately controlled with either LABA/LAMA or ICS/LABA combination.2,4

Despite these recommendations, TT is used extensively in current clinical practice, as has been reported in several studies around the world.5–9 Since TT is becoming increasingly important in clinical practice and over-prescription of inhaled medication is associated with increased costs and possible risk of side effects,10 it's important to understand the current situation of COPD patients with TT in real life and to evaluate the treatment pathways leading to TT. While some studies have been conducted to understand patient characteristics and factors related to TT use,7,11,12 and the efficacy and tolerability of TT have been evaluated in several clinical trials,13–15 there is limited “real world” information regarding pathways before and after initiation of TT.

The objectives of this study were to characterize COPD patients who initiated TT and to describe treatment pathways before and after TT initiation in a population-based sample of patients with COPD in Primary Care.

During the study period, a total of 197,189 patients with recorded diagnosis of COPD were identified; of whom 34,018 (17.2%) initiated TT and constituted the study population. Among these patients, 8.6% initiated TT prior to COPD diagnosis, 5.3% at COPD diagnosis and 86.1% after COPD diagnosis (Fig. 1).

Fig. 1.

Flow-chart diagram of the study population.

(0.24MB).

Pathways following triple therapy initiation

A total of 21,261 (62.5%) patients continued receiving TT throughout follow-up. However, 11,666 patients (34.3%) stepped down from TT and 1091 patients (3.2%) discontinued TT at any time during follow-up. Distribution of pathways following TT initiation was similar between subgroups (Fig. 2).

Fig. 2.

Distribution of pathways following triple therapy in overall and by subgroups of interest.

(0.1MB).

Time to step down or discontinuation

Over 5 years of follow-up, 11,666 (34.3%) patients stepped down from TT. Of them, 7710 (66.1%) did so within one year of TT initiation, and 9984 (85.6%) and 11,059 (94.8%) after two and three years, respectively. Patterns of step down were similar across subgroups (data not shown). The restricted mean time to step down following TT initiation was 38.9 months (95%CI, 51.3–57.9) and was similar across GOLD subgroups. However, among phenotypes, frequent exacerbators stepped down around 3 months earlier (36.2 months) than non-exacerbators (39.6 months) (Table 3). On the other hand, 1091 (3.2%) patients presented complete discontinuation with a restricted mean time to discontinuation of 58.4 months (95%CI, NA), similar across subgroups (Table 3).

Table 3.

Time to first event following Triple Therapy in the Incident Triple Therapy Cohort, Overall and by Subgroup of Interest.

	GOLD A/B (N=23,867)	GOLD C/D (N=10,151)	ACO (N=19,23)	Frequent exacerbators (N=56,34)	Non-Frequent Exacerbators (N=99,31)	Non-exacerbators (N=18,453)
Persistent N (%)	14,700 (61.6)	65,61 (64.6)	11,96 (62.2)	35,40 (62.8)	61,39 (61.8)	11,582 (62.8)
Restricted mean in months; (95% lower-upper limit)	26.3 (26–26.6)	22.8 (22.4–23.2)	25.2 (24.3–26.2)	20.6 (20.1–21.1)	23.9 (23.5–24.3)	27.3 (26.9–27.6)
Step-down N (%)	8323 (34.9)	3343 (32.9)	678 (35.2)	1959 (34.7)	3464 (34.8)	6243 (33.8)
Restricted mean in months; (95% lower-upper limit)	38.6 (37.7–39.4)	38.4 (37.3–39.4)	38.1 (36.5– 39.6)	36.2 (35.4–37.02)	37.8 (37.1–38.4)	39.6 (39.04–40.2)
Discontinuation N (%)	844 (3.5)	247 (2.4)	49 (2.5)	135 (2.4)	328 (3.3)	628 (3.4)
Restricted mean in months; (95% lower-upper limit)	57.8 (57.2–58.4)	58.4 (57.7–59.1)	58.7 (57.6–59.7)	58.4 (57.6–59.2)	57.9 (57.2–58.6)	57.9 (56.01–59.9)

ACO, asthma COPD overlap; FEV1, forced expiratory volume in 1s; GOLD, global initiative for obstructive chronic pulmonary disease.

Kaplan Meier analysis showed that patients starting TT had a 50% probability of step down at 5 years of follow-up. Higher proportions of step downs were observed in frequent exacerbators and patients with ACO; however, observed proportions were similar across GOLD subgroups (Fig. 3). Probability of TT discontinuation within 5 years was about 6%, with higher proportion in milder patients (GOLD A/B, non-exacerbators and non-frequent exacerbators) (Fig. 4).

Fig. 3.

A. Kaplan Meier curves to show rates of step-down following triple therapy at 5 years post index date. (A) by phenotype; (B) by GOLD groups AB/CD.

(0.17MB).

Fig. 4.

Kaplan–Meier curves to show rates of discontinuation following triple therapy at 5 years post index date. (A) by phenotype; (B) by GOLD groups AB/CD.

(0.17MB).

The current study provides a description of the clinical profile of patients with COPD who initiated treatment with TT and treatment pathways prior and following TT initiation in real-life clinical practice. We observed that patients who initiate TT often receive ICS containing regimens as prior treatment and without having experienced exacerbations in the year prior, suggesting that less severe patients are receiving TT, in contrast to treatment recommendations. Treatment with TT tends to remain persistent irrespective of the severity of the disease and the level of risk. Step down following TT is more likely in severe patients, especially during the first year.

Prescription of TT for COPD patients is common in current clinical practice7,20–22 and it's increasing in recent years. A recent study performed in UK using real world data showed an increase of incident TT prescriptions from less than 1% in 2001 to 26% in 2016, and an increase in prevalent TT prescription from 1% to 41% in the same period.9 A similar increase was observed in Spain, where a previous epidemiological study by Barrecheguren et al.8 showed that 12% of newly diagnosed COPD patients initiated TT during 2006–2012, which increased to 17% of incident TT during 2011–2015 in our study. Similarly, our study found 18% of patients eligible to be in the study to be receiving TT. Although this percentage is still lower than that observed in other international studies,6,7,21,22 findings support a consistent trend of increasing TT prescription in Primary Care.

Guidelines published during the study period recommended the use of TT in patients with severe disease, frequent exacerbations and persistent symptoms who were not adequately controlled with LABA/LAMA or ICS/LABA combination.2 These recommendations are supported by accumulated evidence from clinical trials showing the benefits of TT compared with dual or monotherapy in more severe patients with exacerbations.13,14 However, real life data indicate that TT prescription patterns often differ from the current recommendations and there has been a reported overuse of TT in different levels of health care, from Primary Care to hospital-based respiratory clinics.4,5,7,20–22 Interestingly, the majority of patients who initiated TT in our study population were classified as GOLD A/B, in whom TT is not recommended. This finding is not unique to Spain, as previous works performed in the UK,21 the US,7 central and Eastern Europe,23 Latin America24 and Japan25 have also shown the over-prescription of TT in patients who are not frequent exacerbators and belong to the GOLD groups A or B. These findings suggest that primary care physicians may rely more on symptoms than exacerbations for their decisions to step up pharmacologic therapy. This hypothesis is supported by several studies that reported progression to TT with no prior exacerbations history,6,7,11 including a recent study by Landis et al.,26 which found that dyspnea was the main driver for therapy changes in COPD patients in Primary Care. However, there is some evidence for the use of TT irrespective of reported exacerbation history. For example, subgroup analyses from the KRONOS trial suggest a benefit of TT over dual therapy even in patients with no exacerbation in the past year.27

The Spanish COPD guidelines3 at the time of the study recommended the use of ICS in patients identified as ACO and/or frequent exacerbators. It is unlikely that the high use of TT was due to a high prevalence of ACO, because only 5.6% of patients were identified as ACO and the prevalence of high blood eosinophil counts was not high enough to justify the use of TT. It is of note, however, that the mean value of blood eosinophil concentrations for our population were relatively high at 247cell/μl. This is similar to a previous study in Spain reporting a mean value of 254cell/μl in a population of 57,000 unselected patients with COPD at all levels of severity.28 Recently, Kolsum et al.29 obtained a mean eosinophil count of 210cell/μl in patients with COPD, which was significantly higher than counts observed in smokers without COPD and health controls.

Our results highlight prescription of ICS-combined regimens and a later escalation to TT, irrespective of exacerbation frequency.6,11,20,26,30–33 These findings may indicate however, that such use may be warranted for some patients and may be reflective of underreporting of exacerbations or control of exacerbations with ICS therapy. However, if used when not appropriate, prior therapy with ICS-containing regimens could result in limited clinical benefits. Consequently, GPs may decide to escalate treatment to TT instead of changing to a non ICS-containing regimen to control disease.

Interestingly, around 20% of patients who initiated TT had no history of previous COPD medication. A similar percentage has been observed in other studies8,34 and indicates again a deviation from treatment recommendations.2,3

We observed that the probability of de-escalation following TT within 5 years is around 50%, but particularly in patients with more severe disease (frequent exacerbators and patients with ACOS) and during the first year. This unexpected result may be due to lack of efficacy despite receiving TT,26 or on the contrary, due to an effective disease stability achieved after TT.35–37 In addition, we cannot rule out that more frequent contacts of severe patients with the healthcare providers may increase the probability of treatment changes, including de-escalation from TT. Unfortunately, we do not have enough clinical information to address the reason for de-escalation; long term follow up and complete sequencing of treatments after TT initiation may be warranted to best investigate reasons for change in treatment patterns.

In a number of cases, step down consisted of discontinuation of a LABD with persistence of the ICS component. This is consistent with other studies38 but contrasts clinical trials demonstrating that ICS can be safely discontinued in some patients with stable COPD, particularly in non-exacerbators with low blood eosinophil counts.39 We observed 6% probability of complete discontinuation of TT within 5 years, especially in less severe patients,40 which is not consistent with current recommendations that suggest LABD treatment in symptomatic COPD.2,3

The current study has some limitations; as in most real word data studies, caution should be taken when interpreting results relating to choice of therapy as the full context of the decision is not recorded. Full categorization of COPD patients was not possible, as symptoms data, including CAT scores and mMRC are not available in SIDIAP and information on lung function is limited; therefore, GOLD categorisation was only possible in two groups A/B or C/D. Although we used a validated algorithm to identify exacerbations, the risk of underreporting still exists and underreporting of exacerbations by patients is also is well documented in the literature. In addition, exacerbations leading to hospitalization or those treated directly by respiratory specialists were not available; however, these episodes are less frequent in and the majority of these episodes can be captured based on prescriptions following the exacerbation episode. Last, exacerbations were only considered over a 12-month period, longer-term monitoring could be useful to further understand exacerbation frequency and treatment patterns. In any case, if underreporting exists, it should affect equally to all subgroups and should not affect significantly the conclusions of the study. Since new evidence on efficacy and safety of inhaled medication is available and guidelines have been modified, it would be relevant to repeat the study to see if there has been any change in the trends observed.

In conclusion, our results suggest that there is extensive use of TT in primary care, and that the use of TT is not consistent with guideline recommendations. Most patients initiating treatment with TT continue on the same treatment over time regardless of severity of disease. Step down was most frequent in severe patients, while discontinuation was most common among patients with less severe disease. The findings suggest that further research into the drivers of TT prescription in clinical practice would be beneficial.

DL, ND, AB conceived and designed the study. MMo and IS coordinated statistical analyses. MM and MMo wrote manuscript. All authors contributed to study design and analysis, drafting, reviewing, and critically revising the paper, gave final approval of the version to be published, and agreed to be accountable for all aspects of the work.

Marc Miravitlles has received speaker fees from AstraZeneca, Boehringer Ingelheim, Chiesi, Cipla, Menarini, Rovi, Bial, Sandoz, Zambon, CSL Behring, Grifols and Novartis, consulting fees from AstraZeneca, Boehringer Ingelheim, Chiesi, GlaxoSmithKline, Bial, Gebro Pharma, Kamada, CSL Behring, Laboratorios Esteve, Ferrer, Mereo Biopharma, Verona Pharma, TEVA, pH Pharma, Novartis, Sanofi and Grifols and research grants from GlaxoSmithKline and Grifols. Miriam Barrecheguren has received speaker fees from Grifols, Menarini, CSL Behring, GSK and consulting fees from GSK and Novartis. D. Lambrelli, N. Dhalwani and A. Booth are employees of Evidera. The authors confirm there are no other conflicts of interest to report.