Characteristics, treatment patterns and burden of illness in US patients with asthma newly initiating multiple-inhaler triple therapy
•,,,,.
...
Abstract
Introduction For patients with asthma who remain symptomatic on medium-dose inhaled corticosteroid/long-acting β2-agonist, add-on long-acting muscarinic antagonist is a treatment option, which can be administered as multiple-inhaler triple therapy (MITT). A high proportion of patients (61.5%–88.2%) discontinue MITT use within 1 year postinitiation; however, which patients discontinue and their treatment patterns at initiation are unknown. This study aimed to understand the demographic, clinical and treatment-related characteristics of patients with asthma who newly initiated MITT, by discontinuation status.
Methods This retrospective cohort study used administrative data from IBM Truven MarketScan Commercial Claims and Encounters Database with Medicare supplement between 1 January 2016 and 31 December 2019. Adult patients with asthma who initiated MITT between 1 January 2017 and 31 March 2019 were included and were classified based on their discontinuation status. ‘Continuous users’ had continuous use of MITT and ‘discontinuers’ discontinued treatment within the 6-month period postinitiation. Demographics and clinical characteristics, asthma treatment use prior to MITT initiation (12-month baseline period), mode of MITT initiation and complexity of regimen were described.
Results Of 4132 patients (mean age: 49.0 years, 67.9% female), 78.0% (n=3224) were discontinuers; 22.0% (n=908) were continuous users. Demographic and other clinical and treatment-related characteristics during baseline were broadly similar between cohorts. A significantly higher proportion of continuous users versus discontinuers had ≥6 dispensed claims for short-acting β2-agonist canisters (16.0% vs 12.5%; p=0.006) during baseline and initiated a once-daily MITT regimen (35.2% vs 26.2%; p<0.001). Fewer continuous MITT users used a mix of once-daily and twice-daily regimens than those who discontinued MITT (64.3% vs 72.3%; p<0.001).
Conclusions Most patients with asthma discontinued MITT within 6 months. Results indicate that patients with a history of uncontrolled, symptomatic asthma and those using less complex triple therapy regimens at initiation are less likely to discontinue MITT than patients with controlled asthma and those using a complex MITT regimen.
What is already known on this topic
Observational studies in the USA, Japan and England have shown a high proportion of patients with asthma (61.5%–88.2%) discontinue multiple-inhaler triple therapy (MITT) within 1 year postinitiation. However, further studies are needed to better understand the characteristics and treatment patterns of patients who discontinue MITT compared with continuous users.
What this study adds
This retrospective cohort study confirmed high discontinuation of MITT (78.0%) in patients with asthma. Results suggest that patients with poorer asthma control before MITT initiation may be more likely to continue using the treatment, and those who initiate MITT using less complex regimens may be less likely to discontinue MITT than patients using complex regimens.
How this study might affect research, practice or policy
Results of this study indicate that patients’ history of asthma control and the complexity of triple therapy regimen at MITT initiation may be linked to the discontinuation of MITT.
Introduction
Asthma affects approximately 300 million people worldwide, contributing to a substantial clinical and economic burden.1 2 For adult patients with asthma who remain poorly controlled on at least medium-dose inhaled corticosteroid/long-acting β2-agonist (ICS/LABA) therapy, the Global Initiative for Asthma report recommends the addition of a long-acting muscarinic antagonist (LAMA) as a supplementary controller to ICS/LABA, administered as either multiple-inhaler triple therapy (MITT) or single-inhaler triple therapy (SITT).2
In observational studies in the USA, Japan and England, 61.5%–88.2% of patients discontinued MITT within 1 year postinitiation.3–5 However, further studies are needed to understand which patients with asthma are discontinuing MITT and their treatment patterns. The present study aimed to understand which patients discontinued MITT by describing demographics, clinical and treatment-related characteristics of patients with asthma who were new users of MITT by discontinuation status.
Methods
Study design
This retrospective cohort study used administrative data from the IBM Truven MarketScan Commercial Claims and Encounters Database with the Medicare supplement. The study period ranged from 1 January 2016 to 31 December 2019 and the index date was the date of first initiation of MITT between 1 January 2017 and 31 March 2019. The 12-month period before index was defined as the baseline period, and the 6-month period after index was the follow-up period (figure 1). Patients were required to have at least 9 months of follow-up data to enable classification of discontinuation status at 6 months after MITT initiation, accounting for the maximal allowable gap before discontinuation criteria were met.
Study design. aAll patients were required to contribute an additional 3 months of data (ie, ≥9 months of data) to allow for classification of discontinuation status at 6 months postinitiation, based on the longest allowable gap between episodes of MITT prescribing before the criteria for discontinuation were met. MITT, multiple-inhaler triple therapy.
Objective
The objective of this study was to describe the demographic, clinical and treatment-related characteristics of patients with asthma at the time of their first MITT initiation, by discontinuation status. Patients were characterised as either ‘continuous users’ (no discontinuation of MITT use within 6 months postinitiation) or ‘discontinuers’ (any discontinuation of MITT within 6 months postinitiation, with or without re-initiation, which included discontinuation of MITT to begin SITT).
Eligibility criteria
Patients were included in the study if they were ≥18 years of age on the earliest date of their first medical claim for asthma or first pharmacy claim for an asthma maintenance therapy, had ≥2 medical claims with an asthma code recorded ≥30 days apart and ≥1 pharmacy claim for an asthma maintenance therapy (alone or in combination: ICS, LABA, LAMA, xanthines, leukotriene receptor antagonists, chromones and biologics), during the study period. Patients were also required to have initiated MITT between 1 January 2017 and 31 March 2019, have ≥12 months of continuous enrolment and full health insurance coverage preindex date, and have ≥9 months of continuous enrolment and full health insurance coverage postindex date. Diagnosis codes for asthma and respiratory conditions were defined by the International Classification of Diseases, Tenth Revision (online supplemental table 1).6
Patients were excluded from the study if they had any MITT use during the baseline period, or any SITT use during the baseline period (including the index date). Patients with ≥1 medical claim of any of the following respiratory conditions at any point during the study period were also excluded: active respiratory tuberculosis, chronic obstructive pulmonary disease (COPD), cystic fibrosis or lung cancer.
MITT use and discontinuation
MITT initiation was identified based on overlapping dispensed prescriptions for all three MITT components. The discontinuation of MITT use was identified as a greater than 45-day gap in prescribing MITT components.7 The MITT algorithm was defined as any period of overlap (≥1 day) between ICS/LABA and LAMA dispensed on the same or different days. The recorded days of supply were used to calculate the prescription duration with no stockpiling or stretch periods applied.
Statistical analyses
Statistical analyses were intended to be descriptive in nature. However, a post hoc χ2 test was conducted to determine whether there were statistically significant differences in short-acting β2-agonist (SABA) use (proportion of patients with ≥6 dispensed claims for SABA canisters; cut-off based on Nwaru et al8) and MITT dosing regimen (proportion of patients initiating each regimen) between patient cohorts. These outcomes were selected for statistical analysis post hoc as the between-group differences in these outcomes were particularly large, so quantification of the differences was of interest.
A post hoc Cox proportional hazards model was used to determine whether MITT regimen complexity and other factors were independently associated with time to MITT discontinuation (outcome). Before building the model, the proportional hazards assumption was tested by examining a Kaplan-Meier curve (survival function vs survival time) by categories of MITT dosing scheduling complexity. After fitting the model, the proportional hazards assumptions were examined using a Schoenfeld residuals plot.
Survival analysis was used to determine whether MITT regimen complexity was independently associated with time to MITT discontinuation within 6 months of initiation. The categorical, independent variable was defined at index, and was classified as ‘once-daily’ (once-daily ICS/LABA plus once-daily LAMA [least complex dosing regimen]), ‘twice-daily’ (twice-daily ICS/LABA plus twice-daily LAMA) or ‘mix of once-daily and twice-daily’ (once-daily ICS/LABA plus twice-daily LAMA or twice-daily ICS/LABA plus once-daily LAMA [most complex dosing regimen, with discordant dosing regimens; reference category]).
Missing values for variables were not imputed, rather a complete case analysis was performed for univariate and multivariate models; this assumes data were missing at random. For univariate models, the analysis was conducted among all patients with non-missing data for each specific covariate. For the final model, only those patients with complete data for the included covariates were included. Univariate associations between proposed covariates and time to MITT discontinuation were conducted. Unadjusted and adjusted HRs with 95% CIs (and p values) were provided for all variables included in univariate and multivariable models, respectively.
Statistical analysis was conducted using SAS. Statistical power for the Cox regression was estimated using NCSS PASS Software (V.19.0.1).
Patient and public involvement
No patients or public were involved in this study, and no direct subject contact or primary collection of individual human patient data occurred.
Results
A total of 4132 new MITT users were included in the study. Of these patients, 78.0% (n=3224) discontinued MITT within 6 months after initiation (discontinuers), while 22.0% (n=908) had continuous MITT use (continuous users) for the full 6-month period after initiation.
The mean (SD) age among patients at the time of their MITT initiation was 49.0 (12.1) years and most patients were female (67.9%) (table 1). The most common comorbidities among patients were atopy (60.5%) and general cerebrovascular disorders (44.1%). Demographic characteristics of patients who were discontinuers were broadly similar to those who were continuous users, and the two cohorts also had a broadly similar comorbidity burden (table 1). A high proportion of patients had asthma symptoms, including breathlessness (44.1%) and coughing (51.7%) (table 1). A higher proportion of discontinuers had cough symptoms than continuous users (53.0% vs 47.4%), whereas a higher proportion of continuous users had breathlessness than discontinuers (47.9% vs 43.1%) (table 1).
Table 1
|
Patient demographics at the time of MITT initiation and clinical characteristics during baseline
During the baseline period, 45.1% of patients experienced one or more corticosteroid-defined exacerbations, 8.4% of patients had emergency room-defined exacerbations and 4.9% had hospitalisation-defined exacerbations (table 1). There were no differences in exacerbation history at baseline between discontinuers and continuous users (table 1).
Asthma treatment patterns and healthcare resource utilisation (HCRU) among patients were assessed during the baseline period. The proportion of patients using individual maintenance therapies was similar between cohorts, except for a few numerical differences. The most common maintenance medication was ICS/LABA (86.7% vs 80.2% in continuous users and discontinuers, respectively) (table 2). A lower proportion of continuous users had claims for LAMA than those who discontinued (22.0% vs 28.4%) (table 2). During baseline, the mean (SD) dispensed claims for SABA canisters was 3.0 (3.2) among all patients and was similar in continuous users and discontinuers (3.0 [3.3] vs 3.0 [3.1], respectively). However, a statistically higher (p=0.006) proportion of continuous MITT users (16.0%, n=145/908) had ≥6 dispensed claims for SABA canisters compared with MITT discontinuers (12.5%, n=402/3224) (table 2). All-cause and asthma-related HCRU were similar during baseline in patients who were continuous users of MITT and those who were discontinuers (table 2).
Table 2
|
Mode of MITT initiation, asthma maintenance and reliever therapy use, and HCRU during baseline
The regimen for MITT treatment components was assessed at the time of MITT initiation. Most MITT initiators (70.6%) used a mix of once-daily and twice-daily ICS/LABA and LAMA (the most complex regimen), with 28.2% using both once-daily ICS/LABA and once-daily LAMA (the least complex regimen), and 1.3% using both ICS/LABA and LAMA as a twice-daily regimen. A significantly higher (p<0.001) proportion of continuous users (35.2%, n=320/908) initiated a once-daily regimen of MITT than discontinuers (26.2%, n=843/3224), while more discontinuers (72.3%, n=2332/3224) used a mix of once-daily and twice-daily regimens than continuous users (64.3%, n=584/908) (online supplemental figure 1). A large increase in MITT discontinuation (decrease in survival probability) was observed around 30 days after initiation, with a subsequent smaller increase at around 90 days. The probability of MITT discontinuation was highest among those on a twice-daily regimen, followed by a mix of once-daily and twice-daily and once-daily regimens (figure 2). Post hoc analysis showed that compared with a mix of once-daily and twice-daily ICS/LABA and LAMA, patients on once-daily regimens were less likely to discontinue MITT (HR 0.8 [95% CI 0.8, 0.9]; p<0.001), whereas patients on twice-daily regimens were more likely to discontinue MITT (HR 1.3 [95% CI 1.0, 1.8]; p=0.0421). Further details on the Cox proportional hazards regression analysis and univariate and multivariate analyses for the association between MITT regimen and time to MITT discontinuation can be found in online supplemental tables 2–4.
Kaplan-Meier survival curve for time to discontinuation in patients initiating MITT, stratified by MITT regimen complexity. MITT, multiple-inhaler triple therapy.
Most patients initiated MITT following previous asthma maintenance therapy. The highest proportion of MITT initiators stepped up from ICS/LABA (52.4%), followed by direct MITT initiators or those who stepped up from other maintenance therapy (28.5%) and those who stepped up from LAMA (19.1%) (table 2).
Discussion
Among patients who initiated MITT for the first time, nearly 80% discontinued the treatment within 6 months. Previous studies in the UK, the USA and Japan reported similarly high levels of discontinuation (ranging from 61.5% to 88.0%) in the year after MITT initiation.3–5 Generally, similar trends were observed in demographic and clinical characteristics during baseline between continuous users and those who discontinued MITT use within 6 months, with a few differences. Notably, a higher proportion of continuous users had ≥6 dispensed claims for SABA canisters in the year prior to MITT initiation compared with discontinuers. Given that high levels of SABA use can be viewed as a proxy for asthma control, this suggests that patients with poorer asthma control prior to MITT initiation may be more likely to use the therapy long-term.2 8 9 In previous studies, the majority of patients with asthma used SABA as rescue medication before MITT initiation (54.8%–82.7%), with a study in Japan showing that the proportion of patients with SABA use decreased post-MITT initiation versus baseline (47.1% vs 54.8%).3 4 In addition, in the current study, a significantly greater proportion of continuous users were dispensed a simple MITT regimen at index (once-daily ICS/LABA plus once-daily LAMA), with more discontinuers using complex MITT regimens (a combination of once-daily and twice-daily ICS/LABA plus LAMA). In an observational study of patients with COPD in the USA, complex regimens (three inhalers including ICS+LABA+LAMA) have been associated with poorer adherence than simple regimens (two inhalers including ICS/LABA+LAMA or ICS+LABA/LAMA).10 Indeed, in the current study, once-daily regimens were associated with a lower probability of discontinuing MITT compared with complex MITT regimens. Possible causes of discontinuation of MITT may include the inconvenience of using multiple inhalers and burdensome regimens, inadequate insurance coverage, out-of-pocket expenses, switching to biologics treatment, etc.2 Though this is not the focus of this analysis, it is worth noting that future studies could quantify the above to shed light on the causes of MITT discontinuation.
This study used a real-world database to characterise the demographic, clinical and treatment-related characteristics of patients with asthma in the USA. Unlike in previous studies of MITT where patient characteristics were described for the full cohort of initiators,3–5 we stratified our analysis to better understand how characteristics of patients with continuous MITT use over 6 months differed from those who discontinued during that time. When interpreting the findings from this study, a few limitations should be considered. The MarketScan database provided a large sample of patients covered by private US insurance plans and some patients using Medicare; therefore, the results may not be entirely generalisable to the US population. However, the results likely better reflect a real-world patient population compared with prospective cohort or randomised controlled trials with more selective eligibility criteria than the current study. Administrative claims data may be vulnerable to coding inaccuracies and the presence of a diagnosis code may not indicate the presence of disease, which may lead to misclassification bias. To minimise these limitations, code lists were developed with clinical coding experts to appropriately identify an asthma diagnosis. MITT components may also be used to treat COPD, and therefore the algorithms for MITT may have included patients without recorded evidence of COPD who were not necessarily taking MITT for their asthma. As with other claims database studies, pharmacy claims data indicate the date on which the medication was dispensed. However, they do not guarantee that the medication was taken exactly as prescribed or on what exact days patients are taking their medications, leading to a risk of misclassifying MITT users and non-users. To mitigate this, the performance of several existing algorithms was tested and assessed in exploratory analyses, to identify a single robust algorithm for the main analyses.7 As there were no external standards or validated algorithms against which to compare, the assessment was subjective, and each metric was recalculated following adaptations to the original, existing algorithms. To explore the findings from this study, further observational studies, randomised controlled trials or prospective studies should be conducted using other secondary datasets.
Conclusion
Discontinuation of MITT is high in patients with asthma. Findings from this observational study indicate that patients with a history of uncontrolled, symptomatic asthma and those who initiate MITT using less complex regimens may be less likely to discontinue MITT than patients with controlled asthma and those using complex MITT regimens.
