This was a retrospective, population-based, observational study aimed at comparing the effectiveness of maintenance treatment with LABA/LAMA and LABA/ICS in patients with COPD in Catalonia (Spain). Data were obtained from a large administrative healthcare claims database and outcomes were measured during the year after initiation of either LABA/LAMA or LABA/ICS fixed combinations for maintenance therapy of COPD. The study was approved by the Research and Ethics Committee of the IDIAP Jordi Gol Institute of Research In Primary Care (Barcelona, Spain). Since anonymised data were collected retrospectively, no informed consent was considered necessary.

The data was obtained from SIDIAP database (Information System for the Development of Research in Primary Care) which contains anonymized computerized primary care medical records from 5.8 million people in Catalonia, representing more than 80% of the total population.15 The SIDIAP database provides information from medical records (demographic data, past medical history, visits to the general practitioner (GP), clinical variables, prescriptions, immunizations, and referrals), laboratory results and medication dispensed. It is also linked to other databases in Catalonia, such as the hospital discharge summaries database and the mortality registry. The SIDIAP database has been validated and extensively used for other studies on COPD.16,17

The study inclusion criteria were: (1) adults of both sexes older than 40 years; (2) at least one diagnosis of COPD (International Classification of Disease – 10th Edition [ICD-10] codes J41, J42, J43 and J44); (3) spirometry with forced expiratory volume in 1 second (FEV1)/forced vital capacity (FVC)<0.7; (4) in smokers or former smokers; (5) at least one dispensation for LABA/LAMA (fixed form) or a LABA/ICS (fixed form) between 1 January 2013 and 31 December 2019, with the first prescription defined as the index date; (6) a minimum follow-up period of 3 months after initiation of treatment; and (7) at least 12 month of available medical record data prior to the index date (“pre-index”).

The exclusion criteria were: (1) patients with any evidence of LABA/LAMA or LABA/ICS or LABA/LAMA/ICS prescriptions in free or fixed form during the pre-index period; (2) diagnosis of asthma in the year prior to the index date; (3) diagnosis of lung cancer, interstitial lung disease, or lung transplant identified at any time prior to the index date; and (4) a COPD exacerbation 30 days before the index date.

Patients were classified according to the Global Initiative for Chronic Obstructive lung Disease (GOLD) spirometric categories based on FEV1(% predicted) as GOLD I–IV.4 Patients were also classified according to clinical phenotypes.18 Those with two or more exacerbations during the year before the index date were classified as having a frequent exacerbator phenotype, while patients with only one exacerbation were classified as infrequent exacerbators and the remaining COPD patients were considered non-exacerbators. Patients with a concomitant diagnosis of asthma were included in the asthma COPD overlap (ACO) phenotype.19,20

All analyses were conducted for the total population as well as in subgroups of patients at high or low risk of exacerbations based on (1) previous history of exacerbations in the year preceding cohort entry (low exacerbation risk: 0 hospitalisations and 0–1 outpatient exacerbations; high exacerbation risk: ≥1 hospitalization or ≥2 outpatient exacerbations); and (2) blood eosinophil count (baseline eosinophils<300cells/μL or ≥300cells/μL) as identified based on the laboratory result value that was closest but prior to the index date (within 6 months). Finally, according to both variables, patients were classified into four subgroups of high/low risk of exacerbation with high/low blood eosinophil counts.

Exposure measures were based on pharmacy dispensation of LABA/LAMA and LABA/ICS over up to 1 year of follow-up. Exposure was based on the current use of LABA/LAMA or LABA/ICS as defined by the dispensation date plus the day's supply recorded at the time of dispensation.

A gap of 15 days was allowed between dispensations to allow for delays in obtaining medication refills and for continued use beyond the days supplied on which medication had been missed due to imperfect adherence.

Patients were followed from the index date until the earliest of the following: first occurrence of a COPD exacerbation, hospitalization for community-acquired pneumonia, escalation to triple therapy, switch in treatment, discontinuation of COPD treatment, the end of the study period, or (for the main analyses) 1 year after cohort entry. The treatment segment ended at the earliest of the following events: (1) 15 days after the end of the observed day's supply for the medication received on the index date without a subsequent dispensing of COPD medication (i.e., discontinuation); (2) initiation of triple therapy; or (3) any other change in the use of study medication, including a change to a different combination therapy, from a fixed-form to a free-form combination therapy, or from combination therapy to monotherapy.

The primary outcome was to compare the risk of first COPD exacerbation after initiating treatment with LABA/LAMA versus LABA/ICS. Exacerbation episodes were identified using outpatient visits with a diagnostic code indicative of a respiratory exacerbation, or receipt of corticosteroids and/or antibiotics used for treating exacerbations, in accordance with previously published algorithms.16,17 To avoid misclassification and over-estimation of exacerbations, consecutive episodes with less than 21 days between prescriptions or GP visits were considered as a single event.

Secondary outcomes included the risk of hospitalization for community-acquired pneumonia, risk of pharmacy dispensation indicating escalation to triple therapy, and risk of any of the following events occurring: COPD exacerbation, hospitalization for pneumonia or escalation to triple therapy. Pneumonia was defined as hospitalization with one of the following diagnostic codes: 481.x–486.x, 487.0, 507.x, 507.0, 507.1, 507.8, 510.0, 510.9, 511.0, 513.0, 514.x, 517.1, 519.8, 530.84 (as per the International Classification of Diseases, version 9, Clinical Modification: ICD-9-CM; used up to October 2015) and J10.0, J11.0, J12-J18, J22, J69, J85.0, J85.1 and J86 (ICD-10; used from October 2015 onwards) (pneumonia without hospitalization was not captured). These definitions of hospitalization data have been used in several previous studies in patients with COPD.21 Escalation to triple therapy was defined as any combination of pharmacy dispensations resulting in the overlapping use of a LAMA, LABA, and ICS through any fixed or free combination for at least 1 day.

Descriptive analysis was performed with baseline sociodemographic, clinical characteristics and baseline treatments of patients included. Categorical variables were described using absolute frequencies and corresponding percentages. Continuous variables following a normal distribution were described using the mean and standard deviation (SD), while those that did not follow a normal distribution were described using the median and interquartile range.

Time in months between the start of the study treatment and the appearance of the first event (exacerbation, escalation to triple therapy, or any event) was described using the restricted mean survival time (the average time-to-event in a restricted time period) and using Kaplan–Meier analysis over 1 year of follow-up.

The risk of outcomes (exacerbation, escalation to triple therapy, or any event) in patients receiving LABA/ICS versus LABA/LAMA was described using hazard ratios (HRs) with a 95% confidence interval (CI) and p values. Cox proportional hazard models were performed as an as-treated analysis to derive HRs.

Two study sub-cohorts (LABA/LAMA and LABA/ICS) were built based on 1:1 matching through the greedy-matching algorithm (without replacement) and on the propensity score.22 A logistic regression model was used to construct a propensity score that predicts the patient's probability of initiating LABA/LAMA or LABA/ICS treatment. The following variables, included in the model, were collected on the index date: age distribution, sex, time elapsed from COPD diagnosis to the start of treatment, ACO, previous exacerbations and previous treatment billed to compare the evolution at one year of follow-up between cohorts.

All analyses were performed for the whole COPD population and the matched cohort and were also performed according to risk subgroups (high or low risk of exacerbation based on previous exacerbation history and blood eosinophil levels).

All statistical analyses were performed using the statistical software package (SPSS version 20.0, IBM, Chicago, IL, USA).

The SIDIAP database included 236,162 patients older than 40 years with a diagnosis of COPD; of these, 92,132 (39%) initiated either LABA/LAMA or LABA/ICS during the study period. After excluding those without smoking habits or spirometric confirmation of COPD, the final population of our study was composed of 14,046 confirmed COPD patients; 11,329 (80.6%) initiated LABA/ICS and 2717 (19.4%) LABA/LAMA (Fig. 1).

Fig. 1.

Patients’ disposition in the study and formation of the study groups.

(0.31MB).

Patients initiating LABA/ICS were significantly older (69.2 years (SD: 10.4) versus 67.9 (SD: 9.7): p<0.001), were less frequently active smokers (35.5% versus 43.5%; p<0.001), had a higher prevalence of a previous diagnosis of asthma (13.5% versus 4.9%; p<0.001) and had a lower percentage of presenting at least one exacerbation the previous year (70.7% versus 73.3%; p<0.001) while those who had exacerbations presented more episodes (mean number of exacerbations 1.86 (SD: 1.06) versus 1.75 (SD: 0.95); p<0.001). A full description of the characteristics of the study population is shown in Table 1.

After matching for the most important variables related to the risk of exacerbations, the matched population consisted of 1650 patients per treatment arm (Table 1).

During the one-year follow-up, 68.5% of patients in the global population experienced an exacerbation (69.6% in the LABA/ICS group and 64.4% in the LABA/LAMA group). The mean time to the first exacerbation was 6.03 months (95%CI: 5.94–6.12) for LABA/ICS and 6.4 months (95%CI: 6.21–6.59) for LABA/LAMA; p<0.001. The time to escalation to triple therapy and to the combination of all events was also significantly prolonged in LABA/LAMA compared to LABA/ICS patients (Table 2, Fig. 2a–c). Differences were only non-significant for hospitalization due to the extremely small number of events (29 (0.21%) for LABA/ICS and 12 (0.08%) for LABA/LAMA).

Fig. 2.

Kaplan–Meier curves for cumulative incidence of the first event. a–c in the COPD confirmed population: (a) time to the first moderate/severe exacerbation; (b) time to the escalation to triple therapy; (c) time to the combined event. d–f in the matched cohort: (d) time to the first moderate/severe exacerbation; (e) time to the escalation to triple therapy; (f) time to the combined event.

(0.35MB).

These results were replicated in the matched cohort, in whom time to the first exacerbation, time to escalation to triple therapy and time to the combined event were significantly prolonged in patients initiating LABA/LAMA compared to those initiating LABA/ICS (p<0.001 for all three comparisons, Table 2 and Fig. 2d–f). The number of hospitalisations for pneumonia was very small, and the differences were not significant (6 (0.4%) for LABA/ICS and 10 (0.6%) for LABA/LAMA).

The results in terms to time to the next exacerbation, time to escalation to triple therapy or both combined were very similar among the different subgroups of patients classified according to the frequency of previous exacerbations and blood eosinophil levels. Due to the small numbers of hospitalization for pneumonia by subgroups (between 2 and 7 in the whole population and between 0 and 3 in the matched cohort) no statistical comparisons were conducted for this variable. In all cases, the time to the next event was numerically superior for LABA/LAMA versus LABA/ICS and the differences were statistically significant for the time to triple therapy and the time to the next event in all four subgroups with low and high risk of exacerbations and high and low blood eosinophil counts. The time to the next exacerbation was significantly longer for LABA/LAMA users in patients with infrequent exacerbations and high blood eosinophils (7.2 months, 95%CI: 6.6–7.6, versus 6.8, 95%CI: 6.5–7.0; p<0.05) and in patients with frequent exacerbations and low blood eosinophil counts (5.2, 95%CI: 4.8–5.6, versus 4.5, 95%CI: 4.3–4.7; p<0.01) (Table 3).

Similarly, for patients in the matched cohorts, the time to the next event was always numerically longer in the LABA/LAMA compared to the LABA/ICS group and differences were highly significant for time to escalation to triple therapy and combined outcome for the four subgroups (p<0.001 for all comparisons) and significant with p<0.05 for the time to the next exacerbation in the subgroup of patients with infrequent exacerbations and low blood eosinophils (Table 3).

The risk of exacerbations was higher for LABA/ICS users (adjusted HR [aHR]: 1.12, 95%CI: 1.07–1.18). Similarly, the risk of escalation to triple therapy (aHR: 3.26, 95%CI: 2.92–3.63) and the risk of combined outcome (aHR: 1.48, 95%CI: 1.41–1.56) were also significantly higher for patients treated with LABA/ICS (Fig. 3a).

Fig. 3.

Risk of exacerbation, escalation to triple therapy, or any event in patients receiving LABA/ICS versus LABA/LAMA for the whole COPD population and the matched cohort. (a) The global population; (b) patients at low risk of exacerbations; (c) patients at high risk of exacerbations; (d) patients with low blood eosinophil counts; (e) patients with high blood eosinophil counts. Footnote: Hazard ratios were derived using Cox proportional hazard models. aHR: adjusted hazard ratio; CI: confidence interval; COPD: chronic obstructive pulmonary disease; ICS: inhaled corticosteroid; LABA: long-acting beta-2 agonist; LAMA: long-acting antimuscarinic agent.

(0.25MB).

Similar results were observed for patients stratified by the frequency of baseline exacerbations and for patients with eosinophil levels lower or higher than 300cells/μL. Only in the subgroup of patients with more than 300eosinophils/μL were no significant differences in the risk of exacerbations observed between the two treatment groups (Fig. 3b–e).

The analysis in the matched cohort showed similar results. The aHR estimates varied from 1.18 to 3.39 for risk of exacerbation, escalation to triple therapy or the combined outcome, and the results were non-significant only for the risk of exacerbation in patients with more than 300eosinophils/μL (Fig. 3a–e).