Adherence to Inhaled Therapy and Its Impact on Chronic Obstructive Pulmonary Disease (COPD)

Magdalena Humenberger; Andreas Horner; Anna Labek; Bernhard Kaiser; Rupert Frechinger; Constanze Brock; Petra Lichtenberger; Bernd Lamprecht


BMC Pulm Med. 2018;18(163) 

In This Article


The primary outcome parameter of this retrospective analysis was to describe the characteristics of an Upper Austrian COPD cohort based upon degree of adherence to inhaled therapy and its association with spirometrically defined COPD stages. Moreover, we explored adherence as a risk factor for the poor outcome of exacerbation risk and we described further influencing factors on adherence.

Data of patients hospitalized for COPD exacerbations at the department for pulmonology in a tertiary care hospital in Upper Austria and discharged with a guideline conform inhaled therapy in 2012 were analyzed. The following observation period was 24 months. Patients who died within the first six months of the observation period were excluded due to the short observation period. However, patients who died afterwards but during the observation period, were included until death. Hence, the observation period was shorter in these patients and it was assumed that these individuals would have continued with the same adherence routine prior to their death.

Inclusion criteria were age > 40 years, COPD diagnosis (GOLD spirometry class I – IV) based on lung function testing (post-bronchodilator FEV1/FVC < 70%) and a prescribed permanent inhaled therapy. Inhaled therapy was prescribed according to the risk assessment (A – D) as proposed in the GOLD report 2011.[24]

All patients discharged in 2012 were screened looking for a diagnosis with ICD-10-Code 44.0–44.9 at time of discharge. COPD diagnosis and stage were verified by the most recent lung function performed in 2012. Based on lung function criteria, patients with partial post-bronchodilator reversibility were included. However, patients with complete reversibility (ΔFEV1 > 12%, or > 0.2 l) were excluded.

Adherence to inhaled therapy, based on the 24 months observation period, was defined according to the percentage of prescribed inhalers dispensed to the patient and classified as follows: Complete adherence (> 80%), partial adherence (50–80%) and low adherence (< 50%). 80% is a frequently used threshold for the differentiation of adherence (high or low).[23] We decided to further divide the participants according to adherence into three groups to show more precise results in the low adherence group (partial and low adherence).

Additionally, adherence was reported as mean medication possession ratio (MPR),[25] and categorized by sex, FEV1%pred, smoking status and inhaled therapy. The MPR was calculated using the ratio of personal adherence months to the whole observation period of each participant.

For a permanent inhaled therapy, one medical prescription per month for each device was assumed for complete adherence. Complete data therefore was provided by the Upper Austrian Health Insurance (OÖGKK).

Statistical analysis was performed using SAS 9.3. Figures and tables were created with Microsoft Excel 2016. The adherence category (complete, partial, low) was the underlying and central variable in all statistical analyses performed in this study. The association between adherence and most important covariates (age, sex, FEV1%pred, smoking status) is shown in a descriptive overview using means and proportions.

Nonparametric Chi-square test, Mann-Whitney U test and t-test were used to investigate differences between groups according to adherence category.

COPD control was assessed using the rate of severe exacerbations leading to hospitalization per year. Concerning exacerbations leading to hospitalization a binary univariate and multivariate logistic regression analysis was performed, based on odds ratios, to determine the influence of several factors (age, sex, FEV1%pred, smoking status, adherence) on exacerbation rates.

Results are mainly expressed as frequencies or as mean. Statistical significance was defined as p < 0.05 for all analyses in this study.