Adherence to inhalers and comorbidities in COPD patients. A cross-sectional primary care study from Greece

We used a convenience sampling approach to select COPD patients served by primary care and living in rural and semi-urban areas across Greece. The study was approved by the local medical ethics committee of the University Hospital of Crete Greece (protocol number 7985) and the patients gave written informed consent. Detailed methods have been previously published [26‐28]. Subjects were recruited by 53 facilities with participants ranging between 5 and 20 patients per area. The participation rate of subjects varied from 78 to 91%. In summary, 257 COPD patients, previously diagnosed by spirometry (e.g. a post-bronchodilator forced expiratory volume in 1 sec (FEV₁)/forced vital capacity (FVC) ≤0.7 in patients with symptoms and/or history of smoking) [29] during an assessment by chest physicians, were enrolled between 2015 and 2016. Through structured interviews, the general practitioners collected cross-sectional information including demographic characteristics, medical history, lifestyle, COPD Assessment Test (CAT [30]) and Modified Medical Research Council Dyspnea Scale (mMRC [31, 32]) scores, annual number of exacerbations and hospitalisations. COPD patients were classified with ABCD grading system according to GOLD 2019 guidelines [19].

Comorbidities were provided by the primary care physicians based on known patient medical history and/or chronic medication used. Findings from the COPDGene study showed that a comorbidity count performs well to quantify comorbidity in a diverse population with COPD, compared to weighted score or weighted score based upon statistical selection procedure [33], therefore we used a comorbidity count (sum of comorbidities; range from 0 to 9) and grouped our COPD patients into those with at least two (≥2) comorbidities or less.

Adherence to inhalers in COPD patients was assessed with the Test of Adherence to Inhalers (TAI; http://​www.​taitest.​com/​) [34]. It is a 12-item questionnaire which is designed to assess difficulties in following treatment with inhalers in people with asthma or COPD. It consists of 10 items addressing the patient, and 2 additional items which the practitioner can use to identify patients with low adherence. The 10-item TAI is used by healthcare professionals to assess only adherence and degree of adherence, and they may use the 12-item TAI to assess type of non-compliance. Scoring of each of the 10-item ranges from 1: worst compliance to 5: best compliance, providing a total sum score of between 10 (minimum) and 50 (maximum), which can determine the degree of adherence: good (50 points), intermediate (46 to 49 points) or poor (≤ 45 points). Items for healthcare professionals are scored with 1 or 2 points (poor or good knowledge of the regimen and/or inhalation technique) and may offer insights on patient’s type or pattern of non-compliance as follows: sporadic, deliberate or unconscious [34, 35]. A single patient may have more than one type or pattern of non-compliance and a patient with good adherence may be an unconscious non-complier.

Baseline characteristics are presented as count (n) and percentage (%), or mean and standard deviation (SD). If the distribution of continuous data was not normal, then median (interquartile range (IQR)) was used. Univariate analysis was performed using Chi-square or Fisher’s exact tests for categorical outcomes, and Student’s t-test or Mann-Whitney U test for continuous outcomes with normal or not distribution, respectively. Moreover, multivariate logistic regression models were developed to test associations of comorbidities and poor adherence to inhalers (intermediate and good adherence pooled as reference group) with COPD outcomes, adjusting for age, gender, smoking status and occupational status. All tests were two-tailed with significance level at 5%, and were performed in the IBM-SPSS Statistics software (version 23).

Literature reports good adherence levels ranging from 70 to 90% in clinical trials, but this deteriorates in clinical practice and PHC (i.e. 10–40%) [12, 14, 36]. The findings in this PHC population residing in rural/semi-urban areas of Greece (only 26% with good/intermediate adherence), agree with previously reported rates [14, 36]. This may reflect the inability of physicians to provide inhaler training and educate patients, or the patients’ inability to receive and comply to health information or just the insufficient or incorrect use of the devices [37]. However, it warrants attention the fact that nearly 70% of patients found to have deliberate non-compliance in the present Greek COPD population. This is in line with findings from another Greek rural population in 2014, where 46.4% of COPD and asthma patients had stopped their medications completely, decreased dosages or used similar medications that had in the past, during the austerity period [4], highlighting the impact of the global financial crisis on COPD patients’ adherence to therapy in rural areas. Furthermore, poor adherence to inhalers in our study could be explained by the fact that the majority of the patients were retired and a portion were unemployed, striving to cope with the ongoing financial crisis. In a Greek nationwide COPD study (Greek Obstructive Lung Disease Epidemiology and health ecoNomics: GOLDEN study) [16] which included both urban and semi-urban/rural areas, a slightly higher proportion of subjects reported good adherence, however less subjects were unemployed or retired in that study compared to ours. This observation supports the hypothesis that higher rates of poor adherence in rural areas might be attributed to poorer socioeconomic status, an association indirectly also seen in our study (e.g. occupational status vs. poor adherence).

We found that poor adherence to inhalers is associated with higher risk of exacerbation and worse health status. In a Greek COPD study, omission of medication was also associated with risk of exacerbations [16]. Furthermore, adherence to inhalers has been found to reduce disease severity and related exacerbations [38‐40]. In agreement with our findings is also the study by Montes et al. who they also used TAI questionnaire for assessing adherence to inhalers, and found that poor adherence was significantly associated with worse CAT score and higher risk for exacerbations [12]. These findings suggest the importance of adherence to inhalers which may have a direct impact on reducing exacerbations and improving symptoms.

As far as comorbidities in COPD patients are a concern, the findings of the current study are in line with EPIC [41] and EPOCA [42] studies that found at least two comorbidities in 39 and 42% of patients, respectively (42.5% in our study). However, in the Greek GOLDEN study 55% of COPD subjects had at least 2 comorbidities [16]. This high prevalence of comorbidities in COPD patients, projects the significance of managing this major health asset that influences overall health status and QoL.

Additionally, our finding of a significant burden of comorbid conditions in individuals with COPD supports previous investigations, with overweight/obesity, hypertension and diabetes being the most frequent [43‐45], potentially due to shared underlying risk factors and mechanisms [46]. The large number of overweight/obese COPD patients nearly doubled the proportion of subjects with ≥2 comorbidities, when BMI (≥25 Kg/m²) was included in the comorbidity sum score. Prevalence of overweight/obese in our study does not deviate from intercountry data [47]. However, we have not found an association between BMI and GOLD status. There are a few studies assessing the effect of BMI on COPD outcomes, suggesting that obesity, especially if mild-moderate, does not influence COPD symptoms [48, 49].

Moreover, studies suggest that a large proportion of COPD subjects reaching PHC but also secondary/tertiary care frequently have echocardiographical abnormalities proving underlying, sometimes undiagnosed, hypertension and heart disease [50‐52]. However, these echocardiographic abnormalities had limited impact on CAT score [52]. Similarly, we showed null associations between CAT and mMRC and heart disease and hypertension.

GERD is one of the most common gastrointestinal conditions in the general population and it is also listed by GOLD as a common comorbidity in COPD [19]. Evidence suggesting that GERD might worsen COPD outcomes remains controversial and the pathophysiological relationship between comorbid GERD and COPD remains unclear [53]. We did not find any association between GERD and COPD outcomes, however it is important to note that only GERD had a significant association with inhaler adherence in our study. Most data come from observational studies thus causal inference is difficult. However, the positive effect of GERD on inhaler adherence could be misleading with regards to which of the two is the cause and which is the outcome. There are studies proposing confounding by indication e.g., more symptoms due to GERD increase need for inhaler and therefore better adherence, while other studies suggest that respiratory medications are possible risk factors for GERD [53, 54], and others propose share risk factors for both diseases e.g. smoking [53]. More studies are needed to understand causal inference.

Data of the German COPD cohort COSYCONET were the first that revealed associations of COPD GOLD status with a number of comorbid conditions [25]. Even though we found an association of diabetes with GOLD status, this association did not remain significant after adjustment for occupational status, neither any of the associations of the rest of the comorbidities. This discrepancy and/or smaller signal may be explained by large difference in characteristics including socioeconomic status, study settings and prevalence of comorbidities with our Greek population, and also by the smaller sample size of this study.

According to cumulating evidence in the literature, management of COPD patients may benefit by raising PHC professionals awareness and attention for assessing patient adherence to inhalers as well as for screening and managing comorbidities regularly. The findings of the current study highlight that the presence of poor adherence to inhaled medication influences COPD outcomes. PHC physicians could become the lead coordinators of adherence-oriented interventions or COPD action plans, since they have the closest and continuous relation with COPD and chronic patients. They could help them comprehend the chronic nature of disease and effectiveness of treatment, importance of complete adherence to inhaled therapy and necessity of other supportive actions; including smoking cessation and exercise [36, 48]. Furthermore, treatment of comorbidities should become an important part of COPD management as it may influence inhaler adherence and overall disease status.

To the best of our knowledge, this is the first Greek study linking poor adherence and comorbidities to COPD outcomes in patients residing in rural and semi-urban settings in a real-life study. Additionally, assessment of type of adherence added value on translating patients’ compliance. The overall characteristics (e.g., male [5] and overweight/obese [47] overrepresentation), number of comorbidities and most common comorbidities of the present COPD population do not deviate much from those in a nationwide COPD study [16, 55]. On the other hand, several limitations should be discussed along with current findings. Primarily, the cross-sectional design may have inserted several biases, while it lacks ability to show causal associations. Patients are representative solely for PHC population of non-urban population, rather than those attending only secondary-tertiary care facilities. Unfortunately, in the Greek national healthcare system spirometry was not available in the rural primary care setting of this study, neither a linkage between information provided by specialists and GPs, therefore associations with outcomes could not be corrected for airflow obstruction at the time of assessment. In addition, in Greece patients can bypass the GP and visit a specialist or hospital without a referral, and since an electronic medical record linking all GPs with specialists and hospitals was not available at the time, misclassification of outcomes (e.g., exacerbations, hospitalizations) could exist.