Abstract
Background and Aim
Adherence to inhaled corticosteroids (ICS) is essential for asthma control, yet little is known about whether adherence differs across inflammatory phenotypes. Patients with type 2 (T2)-low biomarkers have reduced corticosteroid responsiveness, which may influence adherence. We aimed to evaluate adherence to ICS across inflammatory phenotypes and to examine whether associations between adherence and clinical risk factors differ by inflammatory phenotype.
Material and Methods
This retrospective cohort study included adult asthma patients managed at a respiratory outpatient clinic between May 2020 and May 2022. Inflammatory phenotype was defined using blood eosinophils (cut-off 0.15×109/L) and FeNO (cut-off 25 ppb). Patients were categorized into T2-low (two negative T2-biomarkers), possible T2-high (one positive T2-biomarker), and T2-high (two positive T2-biomarkers). Adherence was calculated as medication possession ratio (MPR), with >0.8 defining good adherence. Data was obtained from electronic health records and national registries. Logistic regression models adjusted for relevant covariates were performed.
Results
Of 960 phenotyped patients (165 T2-low, 538 possible T2-high, 257 T2-high), 484 were included in the adherence analysis. Adherence did not differ significantly between phenotypes, although patients with T2-low asthma showed a tendency towards lower adherence (0.66) than patients with possible T2-high and T2-high asthma (both 0.74). Across phenotypes, increasing age was associated with higher odds of good adherence (OR 1.25, p <0.001). Among male patients, those with T2-high (OR 0.13, p = 0.022) and possible T2-high (OR 0.14, p = 0.023) phenotypes had lower odds of good adherence than those with T2-low phenotype.
Conclusion
In patients with asthma, adherence to ICS did not differ significantly across inflammatory phenotypes, although patients with T2-low asthma tended to show lower adherence. Among males, adherence varied by phenotype, with T2-high and possible T2-high asthma associated with lower odds of good adherence.
Keywords: medication possession ratio, T2 inflammation, biomarkers, blood eosinophils, FeNO
Introduction
Asthma is a heterogeneous airway disease, with a large subgroup driven by type 2 (T2) inflammation, reflected by elevated fractional exhaled nitric oxide (FeNO) and blood eosinophils. Inhaled corticosteroids (ICS) constitutes the cornerstone of asthma treatment, with patients with T2-high asthma typically showing a good treatment response.1 In contrast, a subgroup of asthma patients presents with low T2-biomarker levels and may have a disease driven by other inflammatory pathways,2 whereas in some cases low biomarker levels may reflect ICS-induced suppression rather than a true T2-low phenotype. Consequently, distinguishing true T2-low disease from ICS-suppressed T2 inflammation remains a challenge when classifying inflammatory phenotypes. Nevertheless, a few studies have suggested that the T2-low inflammatory phenotype is associated with a reduced response to ICS and, in some cases, treatment-refractory asthma.3,4
Medication adherence describes the extent to which patients take their medication as prescribed.5 In asthma management, adherence to ICS is a key challenge, with studies showing that patients overall take less than 50% of the prescribed controller medication.6 Previous research has closely linked poor adherence to adverse outcomes, including loss of asthma control, exacerbations, and hospitalizations.7 Current asthma treatment guidelines emphasize the importance of assessing adherence, as reasons for poor adherence vary widely among patients.8 Known factors include younger age, milder disease,6,9 complex treatment regimens, inhaler technique difficulties,10,11 patient beliefs,12 mental health conditions such as anxiety and depression,13 and financial burden.14
Increased use of short-acting β2-agonists (SABA) is a marker of poor asthma control and is associated with hospitalizations, higher systemic corticosteroid (SCS) exposure, and mortality.15–17 Whether patients with T2-low asthma, who may have a reduced response to corticosteroid treatment, have higher SABA use remains unclear, as this relationship has not been well studied.
Risk factors such as smoking and obesity have been linked to low levels of T2-inflammation,18,19 reduced corticosteroid responsiveness,20,21 and poor outcomes.22 However, evidence regarding their impact on adherence are conflicting.23,24
Furthermore, evidence directly linking inflammatory phenotype with adherence is sparse.17,23 One study found no association between adherence and inflammatory biomarkers, including blood eosinophils, FeNO, and IgE, while a subsequent analysis of the same cohort found similar adherence across clinical asthma phenotypes.22,24 However, whether adherence differs across inflammatory phenotypes defined by combined T2 biomarkers, and whether reduced corticosteroid responsiveness in these phenotypes translates into lower adherence, remains unknown.
As maintenance-and-reliever therapy (MART) has become increasingly adopted in asthma management, interpretation of prescription-based adherence measures has also become more challenging, as dispensed ICS may reflect both maintenance and rescue use.8
Adherence may be influenced by patient’s perceived treatment benefit. Therefore, we hypothesized that ICS adherence would be lower among asthma phenotypes with limited corticosteroid responsiveness, such as T2-low asthma.
The primary objective of this retrospective cohort study was, therefore, to evaluate adherence to ICS across inflammatory phenotypes, with particular focus on differences between T2-low and T2-high asthma. Secondary objectives were to assess whether adherence differed according to risk factors associated with reduced corticosteroid responsiveness, namely obesity and smoking. We also examined whether demographic and clinical factors linked to disease burden were associated with poor adherence, and whether these associations varied across phenotypes.
Material and Methods
Study Design and Population
This retrospective cohort study used routinely collected clinical data. Patients were eligible for inclusion if they met the following criteria:
-
-
A registered diagnosis of asthma in a respiratory specialist outpatient care clinic according to the International Classification of Diseases (ICD-10: J45x or J46),
-
-
Age ≥18 years at index date, and
-
-
At least one visit to the respiratory outpatient clinic at Copenhagen University Hospital - Hvidovre between May 1st, 2020, and May 1st, 2022.
Patients with other chronic respiratory diseases, including chronic obstructive pulmonary disease (COPD) and interstitial lung disease (ILD), were excluded a priori. Patients who did not respond to the study invitation and therefore did not provide informed consent were excluded. Details of the cohort formation are presented in Figure 1.
Figure 1.
Flowchart of included patients in the study.
The index date was defined as the date of the patient’s first outpatient visit within this period.
Inflammatory phenotypes were pragmatically defined using blood eosinophils (cut-off 0.15 ×109/L) and FeNO (cut-off 25 ppb). The highest historically documented value was used for classification. The patients were categorized into three groups: T2-low (two negative biomarkers), possible T2-high (one positive biomarker), and T2-high (two positive biomarkers).
Data Sources and Variables
Baseline data, including BMI, smoking status (categorized as never, previous, or current), FeNO, Asthma Control Questionnaire (ACQ) scores,25 spirometry results, and prescribed ICS doses were obtained from electronic health records (Sundhedsplatformen, Epic Systems Inc., USA).
Test results from blood samples, including blood eosinophil counts, total Immunoglobulin E (IgE), and specific IgE for common aeroallergens, were retrieved from the laboratory database (LABKA), which since 2015 has contained nationwide data on blood samples analyzed.
Hospital contacts, administration of biologics, and comorbidities (relevant ICD-10 codes are listed in Supplementary Materials) were retrieved from the Danish National Patient Registry.
Prescription data for the year preceding each patient’s index date were extracted from the Danish National Prescription Registry. This included type of medication, strength, redeemed doses, and number of redeemed prescriptions. Data were collected for medications, including ICS, SABA, long-acting β2-agonists (LABA), long-acting muscarinic antagonists (LAMA), leukotriene receptor antagonists (LTRA), allergen immunotherapy (AIT), azithromycin and biologics. In addition, we extracted information on treatment with SCS as well as antibiotics prescribed for respiratory tract infections (Anatomical Therapeutic Chemical (ATC) codes are provided in Supplementary Materials). Anti-inflammatory reliever (AIR) was defined as low-dose ICS/formoterol prescribed as reliever therapy only, MART included low-dose ICS/formoterol only (doses listed in Supplementary Materials), ICS/LABA and ICS/LABA/LAMA included any combination and dosage of these agents.
Outcomes
The main outcome of interest was adherence to ICS treatment, expressed as medication possession ratio (MPR), calculated as the ratio between redeemed and prescribed doses during the year preceding each patient’s index date. MPR values were not capped at 1.0. Changes in prescribed maintenance treatment were accounted for using treatment initiation and discontinuation dates. Overlapping prescription redemptions were included.
An MPR ≥0.8 was considered indicative of good adherence, as this threshold is widely used in adherence studies, although it has not been specifically validated as the optimal threshold for asthma.26–28
Daily ICS dose exposure was reported as budesonide equivalent doses, according to a previously described method.29
SABA use was defined as redeemed ≥1 SABA inhaler, and annual SABA doses were recorded among users. An exacerbation was defined as treatment with systemic corticosteroids for ≥3 consecutive days and/or hospital contact due to asthma worsening. Annual SCS exposure was recorded among users. Antibiotic use was defined as redeemed prescriptions for antibiotics commonly used for respiratory tract infections (ATC codes are provided in Supplementary Materials), and annual antibiotic courses were recorded among users.
Statistical Methods
Baseline characteristics were summarized according to inflammatory phenotype. Continuous variables were presented as medians with interquartile ranges (IQR), and categorical variables as counts and percentages. Differences between groups were assessed using appropriate parametric or non-parametric tests depending on data distribution.
Associations between adherence and inflammatory phenotype were examined using logistic regression adjusted for sex, age, BMI, and smoking status. A second model additionally included indicators of disease burden and treatment: exacerbations, SABA use, ACQ score, FEV1, and MART use. Potential interactions with obesity, smoking, sex, and SABA use were examined in exploratory analyses. No adjustment was made for multiple testing. Results are reported as odds ratios (ORs) with corresponding 95% confidence intervals (CIs) and p-values.
Sensitivity analyses were performed with MPR capped at 1.0, using an alternative MPR threshold of ≥0.75 to define good adherence, and restricting the analysis to patients not receiving MART. All statistical analyses were carried out using R version 4.5.1.
Results
Cohort Characteristics
A total of 960 patients were included: 165 with T2-low, 538 with possible T2-high, and 257 with T2-high asthma. Baseline characteristics are presented in Table 1. Patients with T2-low asthma were younger, predominantly females, and had higher FEV1. Patients with T2-high asthma more frequently had chronic rhinosinusitis with nasal polyps (CRSwNP), whereas current smokers were less often classified as T2-high. Supplementary Table S1 summarizes blood eosinophils and FeNO, and Supplementary Table S2 presents data for the subgroup with additional specific IgE measurements.
Table 1.
Baseline Characteristics of the Study Population According to Inflammatory Phenotype
| n | Overall, n = 960 | T2-Low, n = 165 | Possible T2-High, n = 538 | T2-High, n = 257 | p-Value | |
|---|---|---|---|---|---|---|
| Female, n | 960 | 665 (69%) | 134 (81%) | 391 (73%) | 140 (54%) | <0.001 |
| Age, years | 960 | 49 (36, 65) | 41 (33, 57) | 49 (37, 64) | 56 (42, 69) | <0.001 |
| BMI, kg/m2 | 802 | 26.3 (23.1, 30.4) | 25.5 (22.4, 29.1) | 26.5 (23.4, 30.5) | 26.1 (23.0, 30.5) | 0.058 |
| BMI, n | 656 | 0.3 | ||||
| Normal weight (BMI <25 kg/m2) | 311 (47%) | 60 (56%) | 166 (45%) | 85 (47%) | ||
| Overweight (25≤ BMI <30 kg/m2) | 144 (22%) | 24 (22%) | 80 (22%) | 40 (22%) | ||
| Obese (BMI ≥30 kg/m2) | 201 (31%) | 24 (22%) | 121 (33%) | 56 (31%) | ||
| Smoking status, n | 959 | 0.002 | ||||
| Never | 500 (52%) | 99 (60%) | 255 (47%) | 146 (57%) | ||
| Previous | 348 (36%) | 47 (28%) | 208 (39%) | 93 (36%) | ||
| Current | 111 (12%) | 19 (12%) | 75 (14%) | 17 (6.6%) | ||
| PackYears, years | 304 | 15 (6, 30) | 20 (7, 34) | 15 (6, 30) | 15 (5, 28) | 0.4 |
| Comorbidities, n | 960 | |||||
| Rhinitis | 84 (8.8%) | 18 (11%) | 49 (9.1%) | 17 (6.6%) | 0.3 | |
| CRSwNP | 9 (0.9%) | <5 (NA) | <5 (NA) | 6 (2.3%) | 0.033 | |
| CRSsNP | 16 (1.7%) | <5 (NA) | <10 (NA) | 7 (2.7%) | 0.3 | |
| GERD | 43 (4.5%) | 9 (5.5%) | 26 (4.8%) | 8 (3.1%) | 0.4 | |
| Depression | 28 (2.9%) | 7 (4.2%) | 15 (2.8%) | 6 (2.3%) | 0.5 | |
| Anxiety | 42 (4.4%) | 9 (5.5%) | 22 (4.1%) | 11 (4.3%) | 0.8 | |
| ACQ | 602 | 0.86 (0.43, 1.57) | 0.71 (0.29, 1.43) | 0.86 (0.29, 1.71) | 0.86 (0.43, 1.57) | 0.2 |
| ACQ, n | 602 | 0.2 | ||||
| Controlled (ACQ <1.5) | 442 (73%) | 61 (81%) | 242 (72%) | 139 (73%) | ||
| Uncontrolled (ACQ ≥1.5) | 160 (27%) | 14 (19%) | 95 (28%) | 51 (27%) | ||
| FEV1% predicted | 673 | 91 (80, 102) | 96 (90, 108) | 90 (78, 102) | 90 (78, 99) | <0.001 |
Notes: Data are presented as median (IQR) for continuous variables and n (%) for categorical variables. Group comparisons were performed using the Kruskal–Wallis rank-sum test for continuous variables and Pearson’s Chi-squared or Fisher’s exact test for categorical variables. Percentages were anonymized (NA) for cells with <5 observations.
Abbreviations: ACQ, Asthma control questionnaire; BMI, Body mass index; CRSsNP, Chronic rhinosinusitis without nasal polyps; CRSwNP, Chronic rhinosinusitis with nasal polyps; FEV1, Forced expiratory volume 1 second; GERD, Gatroesophageal reflux disease; NA, Not applicable.
Medication redemption data are presented in Table 2. Median daily ICS doses were similar between groups, although patients with T2-high asthma tended to receive higher doses (p = 0.053). T2-low asthma had the highest use of MART (75% vs 55–63%, p < 0.001) and ICS/LABA overall (91% vs 79–84%, p = 0.010), whereas T2-high asthma had the highest use of ICS/LABA/LAMA (23% vs 11–16%, p = 0.006) and LTRA (12% vs 3.6–11%, p = 0.013). Fewer T2-low patients redeemed antibiotic courses (6.7% vs 16–19%, p = 0.002), and SABA (4.2% vs 10–11%, p = 0.037), and they had fewer exacerbations (p = 0.016), compared with possible T2-high and T2-high asthma.
Table 2.
Adherence and Medication Use Across Phenotypes
| n | Overall, n = 960 | T2-Low, n = 165 | Possible T2-High, n = 538 | T2-High, n = 257 | p-Value | |
|---|---|---|---|---|---|---|
| Adherence, MPR | 484 | 0.73 (0.37, 0.99) | 0.66 (0.22, 0.98) | 0.74 (0.38, 0.99) | 0.74 (0.41, 1.07) | 0.3 |
| Adherence (≥ 0.8) | 484 | 215 (44%) | 27 (40%) | 118 (44%) | 70 (47%) | 0.6 |
| ICS daily dose, BUD-eq µg/day | 457 | 395 (197, 658) | 395 (132, 548) | 395 (197, 631) | 395 (197, 736) | 0.053 |
| ICS | 890 | 243 (27%) | 30 (20%) | 134 (27%) | 79 (33%) | 0.019 |
| ICS/formoterol (AIR) | 484 | 23 (4.8%) | <5 (NA) | 11 (4.1%) | 12 (8.1%) | 0.022 |
| ICS/formoterol (MART) | 890 | 557 (63%) | 111 (75%) | 315 (63%) | 131 (55%) | <0.001 |
| ICS/LABA | 890 | 747 (84%) | 134 (91%) | 425 (84%) | 188 (79%) | 0.010 |
| ICS/LABA/LAMA | 890 | 153 (17%) | 16 (11%) | 82 (16%) | 55 (23%) | 0.006 |
| LTRA | 960 | 94 (9.8%) | 6 (3.6%) | 58 (11%) | 30 (12%) | 0.013 |
| Azithromycin | 960 | 19 (2.0%) | <5 (NA) | 13 (2.4%) | <5 (NA) | 0.5 |
| AIT | 960 | 14 (1.5%) | <5 (NA) | <10 (1.3%) | <5 (NA) | 0.8 |
| Biologic | 960 | <5 (NA) | <5 (NA) | <5 (NA) | <5 (NA) | 0.7 |
| Exacerbation (≥1*) | 960 | 51 (5.3%) | <5 (NA) | <30 (NA) | <25 (8.2%) | 0.016 |
| SABA (≥1 inhaler redeemed) | 960 | 91 (9.5%) | 7 (4.2%) | 55 (10%) | 29 (11%) | 0.037 |
| SABA, doses/year | 91 | 200 (180, 400) | 180 (60, 200) | 240 (200, 515) | 200 (200, 400) | 0.12 |
| Antibiotic (≥1 course redeemed) | 960 | 145 (15%) | 11 (6.7%) | 86 (16%) | 48 (19%) | 0.002 |
| Antibiotic, course/year | 145 | 1 (1, 2) | 2 (1, 3) | 1 (1, 2) | 1 (1, 2) | 0.6 |
Notes: Data are presented as median (IQR) and n (%). Kruskal–Wallis and Pearson’s Chi-squared tests were used where appropriate to assess group differences. Percentages were anonymized (NA) for cells with <5 observations. *Exacerbation defined as ≥3 consecutive days of SCS treatment and/or an asthma-related hospital contact.
Abbreviations: AIR: Anti-inflammatory reliever, AIT: Allergen immunotherapy, BUD-eq: Budesonide equivalent, ICS: Inhaled corticosteroids, LABA: Long-acting β2-agonist, LAMA: Long-acting muscarinic antagonist, LTRA: Leukotriene receptor antagonist, MART: Maintenance-and-reliever therapy, MPR: Medication possession ratio, NA: Not applicable, SABA: Short-acting β2-agonist.
Clinical characteristics were compared between patients included in the final analysis and those excluded due to missing biomarker or adherence data, where data were available. Included patients were younger, while patients excluded due to missing biomarker data had lower FEV1, higher ACQ scores, higher ICS exposure, and higher MPR (Supplementary Table S3).
Adherence
Adherence data was available for 484 patients, and distribution according to phenotype is shown in Figure 2, with no significant difference between groups. MPR in the overall cohort was 0.73 (Table 2), with similar levels demonstrated in the possible T2-high and T2-high groups: 0.74. In contrast, the T2-low group demonstrated a lower median adherence of 0.66, although this difference was not statistically significant. Among T2-low patients, 40% had good adherence, compared with 44% and 47% of possible T2-high and T2-high patients, respectively. When MPR was capped at 1.0, median MPR decreased to 0.66, but remained similar across inflammatory phenotypes. Using an alternative threshold of MPR ≥0.75 to define good adherence likewise did not substantially change the findings (Supplementary Tables S4 and S5).
Figure 2.
Adherence presented as MPR (medication possession ratio) across inflammatory phenotypes.
Abbreviation: MPR, Medical possession ratio.
In the multivariable analysis, inflammatory phenotype was not associated with good adherence (Table 3). Higher age was associated with increased odds of good adherence (OR 1.25, 95% CI 1.10–1.41). Neither smoking status nor BMI showed a significant association with adherence. Exacerbations were associated with higher odds of good adherence (OR 2.83, 95% CI 1.18–7.37), while other markers of disease burden or MART use were not significantly associated with adherence (Supplementary Table S6). In analyses restricted to patients not receiving MART, median MPR decreased to 0.66, with no significant between-group differences. No statistically significant association between inflammatory phenotype and good adherence was observed, although confidence intervals were wide (Supplementary Table S7).
Table 3.
Multivariable Logistic Regression Model for Good Adherence (MPR ≥0.8) According to Inflammatory Phenotype
| OR | 95% CI | p-Value | |
|---|---|---|---|
| T2 phenotype | |||
| T2-low | — | — | |
| Possible T2-high | 0.91 | 0.50, 1.66 | 0.8 |
| T2-high | 1.00 | 0.52, 1.93 | >0.9 |
| Age/10 years | 1.25 | 1.10, 1.41 | <0.001 |
| Sex | |||
| Male | — | — | |
| Female | 1.45 | 0.93, 2.26 | 0.10 |
| BMI/5 kg/m2 | 0.88 | 0.74, 1.04 | 0.13 |
| Smoking | |||
| Never | — | — | |
| Previous | 0.83 | 0.55, 1.25 | 0.4 |
| Current | 0.79 | 0.38, 1.61 | 0.5 |
Notes: Odds ratios (OR) with 95% confidence intervals (CI) are presented. Age: OR per 10 years increase, BMI: OR per 5 kg/m2 increase, MPR: medication possession ratio.
In exploratory interaction analyses, a significant interaction between inflammatory phenotype and sex was observed (Table 4). Among males, both T2-high and possible T2-high asthma were associated with lower odds of good adherence compared with T2-low asthma (OR 0.14, 95% CI 0.02–0.73 and OR 0.13, 95% CI 0.02–0.72, respectively). No such differences were observed among females. No significant interactions were observed between phenotype and smoking or obesity, while a significant interaction was observed between T2-high phenotype and SABA exposure (OR 9.63, 95% CI 1.33–199), although with a wide confidence interval (Supplementary Table S6D).
Table 4.
Multivariable Logistic Regression Model for Good Adherence (MPR ≥0.8) Including an Interaction Between Inflammatory Phenotype and Sex
| n | OR | 95% CI | p-Value | |
|---|---|---|---|---|
| T2 phenotype (male) | ||||
| Possible T2-high vs T2-low | 57 vs 9 | 0.14 | (0.02, 0.73) | 0.023 |
| T2-high vs T2-low | 60 vs 9 | 0.13 | (0.02, 0.72) | 0.022 |
| T2 phenotype (female) | ||||
| Possible T2 high vs T2-low | 181 vs 44 | 1.08 | (0.49, 2.41) | 0.9 |
| T2-high vs T2-low | 74 vs 44 | 1.20 | (0.49, 2.95) | 0.7 |
| T2 phenotype modified effects | ||||
| Possible T2-high * sex | 0.031 | |||
| T2-high * sex | 0.027 |
Notes: Odds ratios (OR) with 95% confidence intervals (CI) are presented. MPR: medical possession ratio.
Discussion
In this retrospective cohort study, we did not observe statistically significant differences in adherence to ICS across inflammatory phenotypes. However, patients with T2-low asthma showed a tendency towards lower adherence compared with both possible T2-high and T2-high asthma.
Evidence directly examining adherence across inflammatory phenotypes is limited. One study found no difference in adherence between patients with blood eosinophil counts ≤0.4 x109/L and >0.4 x109/L, but it used a high cut-off, a single biomarker, and self-reported/prescription data for assessment of adherence, limiting accuracy and validity.23 Another study reported better adherence among patients with allergy, defined by antihistamine redemption, but it did not include other T2-biomarkers relevant for defining inflammatory phenotypes.17
Overall adherence in this cohort (median 0.73) was higher than previously reported in the same outpatient clinic,30 although the majority of patients still remain below the commonly used threshold for good adherence of >0.8. This may partly reflect that the population managed in specialist care, with many patients treated at higher GINA steps, which has previously been associated with better adherence than in primary care.31 An increase in ICS prescriptions reported during the COVID-19 pandemic may also have influenced MPR estimates during the study period.32 However, this is unlikely to fully explain the pattern observed. Another possible explanation may be that 63% of the cohort received MART regimens, which may have contributed to the apparently high adherence estimates as doses administered as needed is indistinguishable from maintenance doses in registry data. When restricting the analysis to patients not receiving MART, MPR was generally lower. This was also observed in patients with T2-low asthma, in whom MART use was most prevalent. However, differences between inflammatory phenotypes remained non-significant, with considerable uncertainty due to the small sample size and wide confidence intervals. Thus, MART use may increase MPR without necessarily reflecting greater adherence to prescribed maintenance treatment. The outliers in Figure 2 may also reflect MART-related ICS use. Importantly, MART use was not equally distributed across phenotypes in our cohort. As higher ICS exposure may suppress T2 biomarkers such as blood eosinophils and FeNO, this may increase the risk of inflammatory phenotype misclassification. We sought to minimize this risk by using the highest historical T2-biomarker measurements for phenotype classification. Furthermore, the absence of a statistically significant association between inflammatory phenotype and adherence in the analysis restricted to non-MART patients supports the overall finding, although the smaller sample size limits the precision of this sensitivity analysis. Notably, the study period coincided with a shift in treatment guidelines towards recommending treatment with AIR and MART instead of rescue use of SABA.33,34 More knowledge of how adherence is best evaluated in the context of increasing MART use is needed, especially in registry-based studies.
Patients with T2-low asthma were clinically distinct from patients with T2-high asthma, and our findings suggest a lower disease burden among T2-low patients. These differences in severity across phenotypes align with previous studies showing that T2-high asthma is associated with more severe disease than T2-low asthma.35 It is therefore plausible that their slightly lower adherence may relate to fewer symptoms, supporting a theory previously suggested,17 rather than a perceived lack of treatment effect, although the underlying reasons for adherence cannot be determined from our data. In contrast, T2-high patients showed several indicators of higher disease burden, which may motivate more consistent medication use. Exacerbations were associated with higher adherence across phenotypes; although other indicators of disease burden were not. Previous studies have similarly reported increased adherence following an asthma exacerbation,36 raising the possibility that exacerbations may temporarily reinforce medication-taking behavior.
Interestingly, the association between adherence and phenotype appeared to differ by sex, although this exploratory interaction analysis was limited by small subgroup sizes and wide confidence intervals. Males with T2-high and possible T2-high phenotypes had lower adherence compared to males with T2-low phenotypes. Although previous studies examining sex differences in adherence have reported inconsistent results,37 a post-hoc analysis demonstrated sex-specific differences in treatment response, with females showing greater benefit from biomarker-guided therapy.38 Females are known to be overrepresented among patients with severe asthma,39 and we found that females were also more prevalent in the T2-low phenotype, which is consistent with previous findings.40 Taken together, these exploratory findings suggest that sex may influence patterns of adherence across phenotypes. However, the biological and behavioral mechanisms underlying these differences are unknown, and further studies are needed to clarify potential sex-specific differences across phenotypes.
In this study, we observed the lowest proportion of current smokers in the T2-high group, whereas obesity did not differ across phenotypes. Although, evidence on how smoking and BMI relate to adherence is inconsistent,23,24 one study interestingly reported relatively high adherence to ICS (>0.75) among individuals with obesity or smokers.22 In this study, however, we did not observe any significant differences in adherence across smoking status or BMI. Taken together, these findings suggest that the poorer asthma outcomes commonly reported in smokers and individuals with obesity are unlikely to be driven by reduced adherence. Paradoxically, the greater symptom burden in these subgroups may even promote higher adherence, yet with limited clinical benefit due to steroid insensitivity and treatment-refractory disease.20,21 Thus, smoking and obesity represent relevant treatable traits, but the influence on adherence remain unclear.
We did not observe increased SABA use in T2-low asthma, on the contrary, this group had the lowest proportion of SABA redemptions across phenotypes. This also likely reflects the higher symptom burden we found in T2-high asthma in this cohort, potentially driving both increased use of rescue and controller medication. Overall, SABA use was low in the present cohort, with few patients reaching ≥3 canisters/year associated with increased exacerbation risk.41 This may reflect heightened awareness of the risks associated with SABA overuse, updated treatment guidelines with increasing implementation of MART,33,34 and that these patients were managed in specialist care.
In our specialist cohort, T2-high patients were older than T2-low patients, which may reflect selection towards patients with more severe disease being followed in specialist care, whereas a larger proportion of T2-low patients may be managed in primary care. However, estimating the prevalence of asthma phenotypes in primary care is challenging due to infrequent measurement of T2 biomarkers.42 Older age was associated with better adherence across phenotypes, which aligns with earlier studies.17 This may partly reflect a greater symptom burden among older patients, although further research is needed to understand the motivations for ICS use across age groups.
This study has some limitations. The retrospective design relies on routinely collected clinical data, which introduces a risk of misclassification of phenotypes, as biomarker values may be missing, or influenced by ICS treatment or other modifying factors. The study period overlapped with the COVID-19 pandemic, which challenged routine outpatient visits, including access to spirometry and FeNO measurements. This likely contributed to missing data and a reduced sample size. Missing biomarker data may have introduced selection bias, as patients excluded due to missing biomarker data differed from included patients for several clinical characteristics. This may reflect differences in patterns of clinical follow-up and biomarker assessment during the study period, although the reasons for these differences cannot be determined from our data. Consequently, the direction and magnitude of potential selection bias remain uncertain. We initially planned to include specific IgE as a third T2-biomarker, but IgE measurements were available only for a small subset of patients, and those tested were predominantly positive. Restricting analyses to those with complete data for all three biomarkers would have substantially reduced the cohort and introduced further selection bias. Consequently, specific IgE was not included in the phenotyping, which may have limited the ability to capture the allergic component of T2-high asthma. This is relevant as allergy has been associated with better adherence to ICS in a previous study.17 Furthermore, MPR estimates were influenced by MART use, as registry data cannot distinguish maintenance from as-needed doses. The higher proportion of T2-low patients treated with a MART regimen, compared with the other inflammatory phenotypes, further complicates the interpretation of MPR. Although addressed by restricting the analysis to non-MART patients, the small sample size and wide confidence intervals limited the precision of this sensitivity analysis. Future studies will need to stratify by MART regimen when assessing adherence. Finally, some non-significant differences could reflect limited statistical power.
Conclusion
In conclusion, among patients with asthma managed in specialist care and with available biomarker data, adherence did not differ significantly across phenotypes, although adherence was numerically lower in patients with T2-low asthma. Across phenotypes, older age was associated with good adherence. Exploratory analysis suggested lower odds of good adherence among males with possible T2-high and T2-high asthma. Future studies should explore factors influencing adherence across phenotypes and sex, including perceived effect of ICS and treatment behavior, and consider adherence measures that distinguish between maintenance and rescue use in patients receiving MART.
Funding Statement
This study was supported by a research grant from Boehringer Ingelheim. Boehringer Ingelheim had no role in the design, analysis or interpretation of the results in this study. Boehringer Ingelheim was given the opportunity to review the manuscript for medical and scientific accuracy as it relates to Boehringer Ingelheim substances, as well as intellectual property considerations.
AI Statement
In the preparation of this manuscript, the authors used artificial intelligence tools (ChatGPT, version 5.2, OpenAI) solely for language editing and coding assistance. The authors have reviewed the terms of use of the AI tool and confirm its suitability for use in connection with this publication. The authors take full responsibility for the integrity of the entire content of the manuscript, including the accuracy of all references. The authors have reviewed and verified the originality and accuracy of the content, analyses, interpretations, and conclusions presented in the manuscript.
Abbreviations
ACQ, Asthma Control Questionnaire; AIR, Anti-inflammatory reliever; AIT, Allergen immunotherapy; BMI, Body mass index; BUD-eq, Budesonide equivalent; CI, Confidence interval; COPD, Chronic obstructive pulmonary disease; CRSsNP, Chronic rhinosinusitis without nasal polyps; CRSwNP, Chronic rhinosinusitis with nasal polyps; FeNO, Fractional exhaled nitric oxide; FEV1, Forced expiratory volume in 1 second; GERD, Gastroesophageal reflux disease; GINA, Global Initiative for Asthma; ICD-10, International Classification of Diseases, Tenth Revision; ICS, Inhaled corticosteroids; IgE, Immunoglobulin E; IQR, Interquartile range; LABA, Long-acting β2-agonist; LAMA, Long-acting muscarinic antagonist; LTRA, Leukotriene receptor antagonist; MART, Maintenance-and-reliever therapy; MPR, Medication possession ratio; NA, Not applicable; OR, Odds ratio; ppb, Parts per billion; SABA, Short-acting β2-agonist; SCS, Systemic corticosteroids; T2, Type 2.
Ethics Approval and Informed Consent
The study was conducted in accordance with the Declaration of Helsinki. Approvals were obtained from the hospital administration at Copenhagen University Hospital - Hvidovre (WZ: 22052754) and the Center for Health, Capital Region of Denmark (R-22024823). Informed consent was obtained from eligible patients in accordance with regulatory requirements. Furthermore, the study was registered in the Capital Region of Denmark’s legal platform (P-2022-268).
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Disclosure
CHM reports research grants from Skibsreder Per Henriksen, R. and Hustrus Foundation, Boehringer-Ingelheim DK, A.P. Møller Foundation and Danish Lung Association outside the submitted work. KEJH reports personal fees for lectures and educational events and/or consulting outside the current work from AstraZeneca, Chiesi, GSK and SanofiGenzyme, and unrestricted research grants paid to their institution from AstraZeneca, GSK, SanofiGenzyme and Novo Nordisk. HM reports personal fees for lectures and/or advisory board meetings from GSK, Chiesi, AstraZeneca, Norvatis, Sandoz, Teva, ALK-Abelló Nordic A/S outside the submitted work. CGW reports personal fees for lectures from Chiesi and GSK, a travel grant from Sanofi as well as consulting fees from GSK. CGW participated in voluntary activities with AstraZeneca and Sanofi and received no fees for these. NSG reports personal fees for lectures and educational events during the last three years from AstraZeneca, Boehringer-Ingelheim DK and Chiesi outside the submitted work. CSU reports personal fees for consulting or lectures from AstraZeneca, Berlin-Chemie Menarini, Boehringer Ingelheim, Chiesi, Covis Pharma, GlaxoSmithKline, Hikma Pharmaceuticals, IQVIA, Novartis, Novo Nordisk, Orion Pharma, Pfizer, Roche, Sanofi Genzyme, Takeda, TEVA and TFF Pharmaceuticals outside the submitted work. The authors report no other conflicts of interest in this work.
References
- 1.Smith AD, Cowan JO, Brassett KP. et al. Exhaled nitric oxide: a predictor of steroid response. Am J Respir Crit Care Med. 2005;172(4):453–12. doi: 10.1164/rccm.200411-1498OC [DOI] [PubMed] [Google Scholar]
- 2.Lindsley AW, Lugogo N, Reeh KA, Spahn J, Parnes JR. Asthma Biologics Across the T2 Spectrum of Inflammation in Severe Asthma: biomarkers and Mechanism of Action. J Asthma Allergy. 2025;18:33–57. doi: 10.2147/JAA.S496630 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pavord ID, Brightling CE, Woltmann G, Wardlaw AJ. Non-eosinophilic corticosteroid unresponsive asthma. Lancet. 1999;353(9171):2213–2214. doi: 10.1016/S0140-6736(99)01813-9 [DOI] [PubMed] [Google Scholar]
- 4.Berry M, Morgan A, Shaw DE, et al. Pathological features and inhaled corticosteroid response of eosinophilic and non-eosinophilic asthma. Thorax. 2007;62(12):1043–1049. doi: 10.1136/thx.2006.073429 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Vrijens B, De Geest S, Hughes DA, et al. A new taxonomy for describing and defining adherence to medications. Br J Clin Pharmacol. 2012;73(5):691–705. doi: 10.1111/J.1365-2125.2012.04167.X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Williams LK, Joseph CL, Peterson EL, et al. Patients with asthma who do not fill their inhaled corticosteroids: a study of primary nonadherence. J Allergy Clin Immunol. 2007;120(5):1153–1159. doi: 10.1016/j.jaci.2007.08.020 [DOI] [PubMed] [Google Scholar]
- 7.Murphy AC, Proeschal A, Brightling CE, et al. The relationship between clinical outcomes and medication adherence in difficult-to-control asthma. Thorax. 2012;67(8):751–753. doi: 10.1136/thoraxjnl-2011-201096 [DOI] [PubMed] [Google Scholar]
- 8.Global Initiative for Asthma. Global strategy for asthma management and prevention - updated 2025. 2025. Available from: https://ginasthma.org/. Accessed Sep 24, 2026.
- 9.Canonica GW, Domingo C, Lavoie KL, et al. Asthma patients’ and physicians’ perspectives on the burden and management of asthma: post-hoc analysis of APPaRENT 1 and 2 to assess predictors of treatment adherence. Respir Med. 2024;227:107637. doi: 10.1016/J.RMED.2024.107637 [DOI] [PubMed] [Google Scholar]
- 10.Oga T, Ito R, Mita C, et al. Analysis of adherence to ICS/LAMA/LABA in patients with asthma: a retrospective observational cohort study using medical claims data. Respir Investig. 2025;63(4):569–575. doi: 10.1016/J.RESINV.2025.04.017 [DOI] [PubMed] [Google Scholar]
- 11.Dierick BJH, Achterbosch M, Eikholt AA, et al. Electronic monitoring with a digital smart spacer to support personalized inhaler use education in patients with asthma: the randomized controlled OUTERSPACE trial. Respir Med. 2023;218:107376. doi: 10.1016/J.RMED.2023.107376 [DOI] [PubMed] [Google Scholar]
- 12.Adouni Lawani M, Zongo F, Breton MC, et al. Factors associated with adherence to asthma treatment with inhaled corticosteroids: a cross-sectional exploratory study. J Asthma. 2018;55(3):318–329. doi: 10.1080/02770903.2017.1326131 [DOI] [PubMed] [Google Scholar]
- 13.Alqarni AA, Aldhahir AM, Siraj RA, et al. Asthma medication adherence, control, and psychological symptoms: a cross-sectional study. BMC Pulm Med. 2024;24(1):1–10. doi: 10.1186/S12890-024-02995-X/FIGURES/4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hung CT, Erickson SR, Wu CH. Cost-related non-adherence to medications among adults with asthma in the USA, 2011–2022. Thorax. 2025;80(1):16–23. doi: 10.1136/THORAX-2024-221778 [DOI] [PubMed] [Google Scholar]
- 15.Stanford RH, Shah MB, D’Souza AO, Dhamane AD, Schatz M. Short-acting β-agonist use and its ability to predict future asthma-related outcomes. Ann Allergy Asthma Immunol. 2012;109(6):403–407. doi: 10.1016/J.ANAI.2012.08.014 [DOI] [PubMed] [Google Scholar]
- 16.Makhinova T, Barner JC, Richards KM, Rascati KL. Asthma controller medication adherence, risk of exacerbation, and use of rescue agents among Texas medicaid patients with persistent asthma. J Manag Care Spec Pharm. 2015;21(12):1124–1132. doi: 10.18553/jmcp.2015.21.12.1124 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Packham S, Ödling M, Bossios A, Konradsen JR, Stridsman C. Adherence to inhaled corticosteroid therapy and treatment escalation in the Swedish adult asthma population. Respir Med. 2024;231(Jan). doi: 10.1016/j.rmed.2024.107714 [DOI] [PubMed] [Google Scholar]
- 18.Shimoda T, Obase Y, Kishikawa R, Iwanaga T. Influence of cigarette smoking on airway inflammation and inhaled corticosteroid treatment in patients with asthma. Allergy Asthma Proc. 2016;37(4):e50–e58. doi: 10.2500/AAP.2016.37.3944 [DOI] [PubMed] [Google Scholar]
- 19.Sharma V, Ricketts HC, Steffensen F, Goodfellow A, Cowan DC. Obesity affects type 2 biomarker levels in asthma. J Asthma. 2023;60(2):385–392. doi: 10.1080/02770903.2022.2051548 [DOI] [PubMed] [Google Scholar]
- 20.Sutherland ER, Goleva E, Strand M, Beuther DA, Leung DYM. Body mass and glucocorticoid response in asthma. Am J Respir Crit Care Med. 2008;178(7):682–687. doi: 10.1164/RCCM.200801-076OC [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lazarus SC, Chinchilli VM, Rollings NJ, et al. Smoking Affects Response to Inhaled Corticosteroids or Leukotriene Receptor Antagonists in Asthma. Am J Respir Crit Care Med. 2007;175(8):783. doi: 10.1164/RCCM.200511-1746OC [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ilmarinen P, Vähätalo I, Tuomisto LE, Niemelä O, Kankaanranta H. Long-term adherence to inhaled corticosteroids in clinical phenotypes of adult-onset asthma. J Allergy Clin Immunol Pract. 2021;9(9):3503–3505.e3. doi: 10.1016/j.jaip.2021.04.057 [DOI] [PubMed] [Google Scholar]
- 23.Papi A, Ryan D, Soriano JB, et al. Relationship of Inhaled Corticosteroid Adherence to Asthma Exacerbations in Patients with Moderate-to-Severe Asthma. J Allergy Clin Immunol Pract. 2018;6(6):1989–1998.e3. doi: 10.1016/J.JAIP.2018.03.008 [DOI] [PubMed] [Google Scholar]
- 24.Vähätalo I, Kankaanranta H, Tuomisto LE, Niemelä O, Lehtimäki L, Ilmarinen P. Long-term adherence to inhaled corticosteroids and asthma control in adult-onset asthma. ERJ Open Res. 2021;7(1):00715–02020. doi: 10.1183/23120541.00715-2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Juniper EF, O’Byrne PM, Guyatt GH, Ferrie PJ, King DR. Development and validation of a questionnaire to measure asthma control. Eur Respir J. 1999;14(4):902–907. doi: 10.1034/j.1399-3003.1999.14d29.x [DOI] [PubMed] [Google Scholar]
- 26.Chongmelaxme B, Chaiyakunapruk N, Dilokthornsakul P. Association between adherence and severe asthma exacerbation: a systematic review and meta-analysis. J Am Pharm Assoc. 2020;60(5):669–685.e2. doi: 10.1016/j.japh.2020.02.010 [DOI] [PubMed] [Google Scholar]
- 27.Asamoah-Boaheng M, Bonsu KO, Farrell J, Oyet A, Midodzi WK. Measuring Medication Adherence in a Population-Based Asthma Administrative Pharmacy Database: a Systematic Review and Meta-Analysis. Clin Epidemiol. 2021;13:981–1010. doi: 10.2147/CLEP.S333534 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Asamoah-Boaheng M, Farrell J, Osei Bonsu K, Midodzi WK. Determining the optimal threshold for medication adherence in adult asthma patients: an analysis of British Columbia administrative health database in Canada. J Asthma. 2022;59(12):2449–2460. doi: 10.1080/02770903.2021.2014862 [DOI] [PubMed] [Google Scholar]
- 29.von Bülow A, Hansen S, Sandin P, et al. Use of High-Dose Inhaled Corticosteroids and Risk of Corticosteroid-Related Adverse Events in Asthma Findings From the NORDSTAR Cohort. J Allergy Clin Immunol Pract. 2025;13(7):1609–1619.e5. doi: 10.1016/J.JAIP.2025.01.023 [DOI] [PubMed] [Google Scholar]
- 30.Jensen FF, Håkansson KEJ, Overgaard Nielsen B, Weinreich UM, Ulrik CS. Self-reported vs objectively assessed adherence to inhaled corticosteroids in asthma. Asthma Res Pract. 2021;7(1):7. doi: 10.1186/S40733-021-00072-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Toh MR, Gxz N, Goel I, et al. Asthma prescribing trends, inhaler adherence and outcomes: a Real-World Data analysis of a multi-ethnic Asian Asthma population. Prim Care Respir Med. 2024;34(1):35. doi: 10.1038/s41533-024-00391-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Dhruve H, d’Ancona G, Holmes S, Dhariwal J, Nanzer AM, Jackson DJ. Prescribing Patterns and Treatment Adherence in Patients with Asthma During the COVID-19 Pandemic. J Allergy Clin Immunol Pract. 2021;10(1):100. doi: 10.1016/J.JAIP.2021.09.032 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Global Initiative for Asthma. Global strategy for asthma management and prevention - updated 2019. 2019. Available from: https://ginasthma.org/. Accessed Sep 24, 2026.
- 34.Global Initiative for Asthma. Global strategy for asthma management and prevention - updated 2020. 2020. Available from: https://ginasthma.org/. Accessed Sep 24, 2026.
- 35.Ricciardolo FLM, Sprio AE, Baroso A, et al. Characterization of T2-Low and T2-High Asthma Phenotypes in Real-Life. Biomedicines. 2021;9(11). doi: 10.3390/biomedicines9111684 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Staehr Holm F, Håkansson KEJ, Ulrik CS. Adherence with controller medication in adults with asthma–impact of hospital admission for acute exacerbation. J Asthma. 2022;59(9):1899–1907. doi: 10.1080/02770903.2021.1971702 [DOI] [PubMed] [Google Scholar]
- 37.Dima AL, Hernandez G, Cunillera O, Ferrer M, De Bruin M. Asthma inhaler adherence determinants in adults: systematic review of observational data. Eur Respir J. 2015;45(4):994–1018. doi: 10.1183/09031936.00172114 [DOI] [PubMed] [Google Scholar]
- 38.Eastwood MC, Busby J, Jackson DJ, et al. A Randomized Trial of a Composite T2-Biomarker Strategy Adjusting Corticosteroid Treatment in Severe Asthma: a Post Hoc Analysis by Sex. J Allergy Clin Immunol Pract. 2023;11(4):1233–1242.e5. doi: 10.1016/J.JAIP.2022.12.019 [DOI] [PubMed] [Google Scholar]
- 39.Jackson DJ, Busby J, Pfeffer PE, et al. Characterisation of patients with severe asthma in the UK Severe Asthma Registry in the biologic era. Thorax. 2021;76(3):220–227. doi: 10.1136/thoraxjnl-2020-215168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Heaney LG, Perez de Llano L, Al-Ahmad M, et al. Eosinophilic and noneosinophilic asthma: an expert consensus framework to characterize phenotypes in a global real-life severe asthma cohort. Chest. 2021;160(3):814–830. doi: 10.1016/j.chest.2021.04.013 [DOI] [PubMed] [Google Scholar]
- 41.Nwaru BI, Ekström M, Hasvold P, Wiklund F, Telg G, Janson C. Overuse of short-acting β2-agonists in asthma is associated with increased risk of exacerbation and mortality: a nationwide cohort study of the global SABINA programme. Eur Respir J. 2020;55(4). doi: 10.1183/13993003.01872-2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Kerkhof M, Tran TN, Allehebi R, et al. Asthma Phenotyping in Primary Care: applying the International Severe Asthma Registry Eosinophil Phenotype Algorithm Across All Asthma Severities. J Allergy Clin Immunol Pract. 2021;9(12):4353–4370. doi: 10.1016/j.jaip.2021.07.056 [DOI] [PubMed] [Google Scholar]