Dynamics of inhaled corticosteroid use are associated with asthma attacks

Abstract

Inhaled corticosteroids (ICS) suppress eosinophilic airway inflammation in asthma, but patients may not adhere to prescribed use. Mean adherence-averaging total doses taken over prescribed-fails to capture many aspects of adherence. Patients with difficult-to-treat asthma underwent electronic monitoring of ICS, with data collected over 50 days. These were used to calculate entropy (H) a measure of irregular inhaler use over this period, defined in terms of transitional probabilities between different levels of adherence, further partitioned into increasing (H_inc) or decreasing (H_dec) adherence. Mean adherence, time between actuations (Gap_max), and cumulative time- and dose-based variability (area-under-the-curve) were measured. Associations between adherence metrics and 6-month asthma status and attacks were assessed. Only H and H_dec were associated with poor baseline status and 6-month outcomes: H and H_dec correlated negatively with baseline quality of life (H:Spearman r_S = - 0·330, p = 0·019, H_dec:r_S = - 0·385, p = 0·006) and symptom control (H:r_S = - 0·288, p = 0·041, H_dec: r_S = - 0·351, p = 0·012). H was associated with subsequent asthma attacks requiring hospitalisation (Wilcoxon Z-statistic = - 2.34, p = 0·019), and H_dec with subsequent asthma attacks of other severities. Significant associations were maintained in multivariable analyses, except when adjusted for blood eosinophils. Entropy analysis may provide insight into adherence behavior, and guide assessment and improvement of adherence in uncontrolled asthma.

Figure 1

Assessment protocol timeline, visits and clinical measures. EMD electronic monitoring device, ACT asthma control test, AQLQ asthma quality of life. (Microsoft Powerpoint, version 2101, https://office.live.com/start/powerpoint.aspx).

Figure 2

Calculation of entropy. Panel (a) shows a perfectly-adherent time series (green) in the background comprising 100% of prescribed puffs for 100% of the time, and an example patient time series (blue) overlaid atop the perfectly-adherent series, with instances of under- and over-adherence, both mapped to the state-based series. Panel (b) displays the corresponding transitional probability matrix, while panel (c) allows us to visualise the same matrix (and the “disorder”) in a 3-dimensional graph. The entropy of the transitional probability matrix is then calculated as ${\sum }_{i,j=0}^{N-1}{P}_{i,j}(-\ln (P_{i,j}))$. (Python Software Foundation, version 3·6 http://www.python.org).

Figure 3

Sample adherence time series from 4 different patients over the study period. All patients had ‘good adherence’ as defined by mean adherence PT_mean (note not capped at 100% for the purpose of demonstrating variability), however with different increasing and decreasing entropy (H_inc and H_dec) measures, which better reflect the variability in patient inhaler adherence behaviour. Panel (a) demonstrates a patient who took their inhaled controller on average 84% of prescribed doses with calculated entropy 2.24 and increasing entropy 1.61. Panel (b) demonstrates a second patient who took their controller 90% of prescribed doses, with calculated entropy of 2.32 and increasing entropy of 0.92. Panel (c) demonstrates a third patient who took their controller 102.5% of prescribe doses, with calculated entropy of 1.85 and decreasing entropy of 0.95. Panel (d) demonstrates a fourth patient who took their controller 102.5% of prescribed doses, with calculated entropy of 1.86 and decreasing entropy of 0.59. (Python Software Foundation, version 3·6 http://www.python.org).

Figure 4

Consort diagram demonstrating flow of participants through the study. EMD electronic monitoring device. ^†Due to EMD detachment. (Microsoft Powerpoint, version 2101, https://office.live.com/start/powerpoint.aspx).

Figure 5

Entropy metrics (over day 0–50) predict asthma outcomes (over days 0–180). Panel (a): Entropy (H) and attacks requiring hospitalisation. Panels (b–d): Decreasing entropy (H_dec) and attacks requiring general practitioner (GP) visit, oral corticosteroids or hospitalisation respectively. The boxes depict the 25th, 50th, and 75th percentiles while the whiskers depict the minimum and maximum values in the data. The individual data points are also shown as dots. (R version 3·3 https://www.R-project.org/).