Classification of COPD patients and compliance to recommended treatment in Greece according to GOLD 2017 report: the RELICO study

Chronic obstructive pulmonary disease (COPD) is a multifactorial clinical condition, mainly related to smoking, characterized by chronic respiratory symptoms, structural pulmonary abnormalities, impaired lung function, or a combination of these findings, and is often accompanied by multiple, clinically significant comorbidities. According to current data on pathogenesis, COPD derives from various lifelong, dynamic, and cumulative interactions between genomic and environmental factors, that modulate the development, maintenance, and function of the lungs, through various biological mechanisms, including but not limited to airways’ and systemic inflammation [1]. It is this diversity in the etiopathogenesis, pathophysiology and clinical manifestations that makes early, accurate diagnosis and management of COPD particularly challenging. Possibly this is also the reason for the notable variance in the reported prevalence of COPD throughout the world [2], as researchers use different sampling schemes, diagnostic criteria, and measurement modalities. Various sources of data are currently available on the epidemiological features of COPD, either for USA [3], China [4], Europe [5] and worldwide [6].

In 2017 GOLD (Global Initiative for Chronic Obstructive Lung Disease) issued a new report for COPD diagnosis, management, and prevention. The goal of the GOLD assessment is to guide therapy based on the severity of airflow limitation as estimated by spirometry, the intensity of symptoms and the risk of future exacerbations or hospitalizations. Earlier GOLD guidelines [7] used a simple spirometric grading system to classify COPD patients and determine optimal treatment. In 2011, this system was updated to also include the symptomatic assessment and the risk of exacerbation of each patient [8]. The so-called “ABCD” tool combines the data from spirometry, the modified British Medical Research Council questionnaire (mMRC) or the COPD Assessment Test (CAT) and the history of patient’s exacerbations to classify patients into four groups. It has been noted though, that this classification had some limitations. Firstly, the ABCD assessment tool performed no better than the spirometric grades for mortality prediction or other important health outcomes in COPD. Secondly, group “D” outcomes were modified by two parameters: lung function and/or exacerbation history, which caused confusion. Under this perspective, in 2017, a revision of the ABCD assessment tool was proposed that separated spirometric grades from the “ABCD” groups. Therefore, grading according to spirometry as well as group classification according to patient-reported symptoms (CAT and mMRC assessment) and exacerbations were reported for each patient [9].

A previous study in Greece has demonstrated that the overall prevalence rate in the studied population (adults 35 years old and smokers of at least 100 cigarettes per lifetime) was 8.4%. It was found that 11.5% of the men and 4.8% of the women had spirometrically confirmed COPD [10]. Another two, more recent studies, with much narrower geographical coverage, estimated a COPD prevalence of 9% (mean patients’ age 52 years) in parts of northern Greece [11] and 18% (mean patients’ age 54 years) in Thessaly region [12]. However, a substantial number of COPD patients remain undiagnosed. In addition, there is little information regarding the treatment of COPD patients in the real world setting and the adherence to the GOLD recommendations. We have hypothesized that these changes in GOLD recommendations were not yet fully integrated in clinical practice in Greece. In any case, few data exist for the real-world treatment of COPD in Greece and even fewer for the adherence of physicians to treatment recommendations. The aim of the present study was to investigate (1) the everyday clinical practice on the COPD therapeutic management in Greece, (2) the severity of the disease in relation to the treatment provided and (3) the correlation of the data with GOLD 2017 recommendations.

Data acquisition 3740 subjects were initially screened and 125 of them excluded from the study, because they did not cover all inclusion criteria and/or at least one exclusion criterion was applicable. After recruitment, no discontinuation of subjects has been confirmed. Thus, datasets from 3615 COPD patients were collected. Various missing values have been retrospectively (at the stage of data processing) encountered (Additional file 1: Figure S1).

Demographics male to female ratio was 3:1 in our study sample (male participants 73.9%) with an age ranging from 40 to 95 years (mean and median age 69 years, with a SD of 10 years). Two thirds of our study sample belonged to the 60–79 years age group (67.7%), while the remaining third was roughly divided between the 40–59 years (18.5%) and the 80–99 years age group (13.8%). The mean (and median) body mass index (BMI) was higher than 27 (with a SD of 5.1) with almost three quarters of our study sample belonging to the overweight (25–29.9, 43%) or obese (30 + , 29.2%) BMI group. Only forty-eight subjects (1.3%) were underweighted to severely underweighted (Additional file 1: Figure S2). It should be noted that two subjects < 40 years of age were mistakenly classified as eligible study participants, and their data were included into the analysis. Nevertheless, they represent only 0.05% of the study sample, therefore not altering the results presented or the study inferences.

Vital signs The mean systolic/diastolic blood pressure (± SD) was 130/79 (12.5/9.4), the mean heart rate (± SD) 80 (9.6) and the mean respiratory rate (± SD) 18.2 (9.3). No differences were found in BMI, systolic/ diastolic blood pressure, heart, and respiratory rates between the 3 age groups (40–59, 60–79 and 80–99 years) among men and women (Additional file 1: Figure S2).

COPD-related medical history we calculated the time since initial COPD diagnosis and subtracted it from the current age of participants (i.e. age at recruitment) to estimate their age at first COPD diagnosis. The available data (N = 2552) indicate that the mean age (± SD) at COPD diagnosis was 63.8 years (10.2) and that 85% of patients were between 50–79 years of age at the time of diagnosis (Additional file 1: Figure S2). According to the subjects’ overall CAT score (Additional file 1: Table S1), their history of COPD exacerbations, either mild or serious enough to lead to hospitalization (Additional file 1: Table S2), participants were classified into one of four groups of the ABCD assessment tool. Based on the available data (N = 3520) more than half (59%) of the subjects were classified as belonging to group B, while almost all of the remaining cases were classified as belonging to groups A (18%) and D (21%). Only a tiny portion of the participants (2%) was categorized into group C (Fig. 1).

ABCD assessment tool in relation to age and years since diagnosis of patients significant differences exist between the mean age [F_3,2484 = 18.06, p < 0.001] or the time since initial COPD diagnosis of subjects (Additional file 1: Figure S3) [F_3,2484 = 19.88, p < 0.001] belonging to each of the four groups based on the ABCD assessment tool. Overall, both variables tend to slightly increase when moving to a higher group. From the available data (N = 2485) we conclude that subjects in group A are on average 2.5 years younger compared to those belonging to group B (p < 0.001) and more than 4 years younger compared to those belonging to group C (p = 0.008) or D (p < 0.001). Moreover, subjects in group B are almost 2 years younger compared to those in group D (p = 0.002). Similarly, subjects in group A were diagnosed with COPD more recently, at least 6 months, compared to those in group B (p = 0.052) and at least 2 years from those in groups C (p = 0.01) or D (p < 0.001). The same applies for subjects in group B compared to those of group D (p < 0.001) (Fig. 2). No associations were found between ABCD group allocation with gender or BMI (Additional file 1: Table S3).

ABCD assessment tool in relation to comorbid conditions of patients the group of the ABCD assessment tool that the patients belong to and their age group (< 59, 60–79, 80–99 years) relate to the burden of comorbid conditions of COPD patients, according to the available data (N = 3518). In particular, subjects belonging to group D were much more likely to suffer from a second comorbid condition (excluding COPD) compared to subjects belonging to groups A or B, who were predominantly diagnosed with just one comorbid condition [F(3, 21.6) = 15.3, p < 0.001]. Furthermore, subjects belonging to a higher age group gradually increased their likelihood of having one or two comorbid conditions (excluding COPD) [F(2, 39.9) = 28.4, p < 0.001] (Fig. 2). A list of the comorbid conditions of this study sample can be found in Additional file 1: Table S4.

ABCD assessment tool in relation to severity of airflow limitation the severity of airflow limitation was determined based on each patient's latest available spirometric results (Additional file 1: Figure S4) as mild (FEV₁ ≥ 80%pred), moderate (FEV₁ = 50–79%pred), severe (FEV₁ = 30–49%pred) and very severe (FEV₁ < 30% %pred). Most subjects, independent of their ABCD group allocation, were measured as suffering from moderate COPD. Moreover, the portion of subjects suffering from severe COPD increased as we move from group A to group D (Fig. 2).

ABCD assessment tool, COPD treatment practices and vaccination combined therapy with LAMA/LABA bronchodilators was the most frequently prescribed therapeutic option in our study sample (applied to 48% of the subjects), while other frequent options included combination of LABA, LAMA and ICS (28% of subjects) (Additional file 1: Figure S5). As a short acting bronchodilator, SABA was the most frequently used option (15.5% of subjects) (Additional file 1: Table S5). By combining the available data from the ABCD classification tool and the treatment choices of the corresponding subjects (N = 3520), it appears that only 18% of group A patients were treated according to the GOLD 2017 suggestions, while the corresponding percentage for group B and C patients was 59% and 58%, respectively. Treatment of group D patients showed the highest percentage of adherence to GOLD 2017 recommendations (82%) (Additional file 1: Figure 3). Finally, subjects were showing higher probability for being vaccinated against both, Influenza virus and S. pneumoniae, or at least one of them, if suffering from severe or very severe COPD, or belonging to groups B and D (Additional file 1: Table S6). More specifically, over 82% of severe to very severe COPD patients (or patients belonging to groups B and D) were vaccinated against at least one of the pathogens, while this value drops to below 75% for patients with mild COPD (or those belonging to groups A and C).

Compliance to treatments Perfect compliance for most drug categories (referring either to LAMA, LABA, ICS and their combinations, or to supporting treatments for symptomatic relief, such as short-acting bronchodilators or methylxanthines) was at the level of 65–75%, with the exception of antibiotics (this refers to prophylactic low dose azithromycin in a small number –110– of patients, approaching 85%), mucolytic agents and antioxidants (just above 50%). Poor compliance (missing doses with a frequency of more than twice weekly) was in most cases not exceeding 10%, and in the cases of the principal treatment options (LAMA, LABA, ICS and their combinations) not exceeding 6% (Additional file 1: Table S7, Figure S5).

This large COPD non-interventional, epidemiological study primarily focused on generating real-world data related to the course of COPD patients under typical treatment in the community [13, 14]. The short duration of the study and the enrollment of sequential patients from each participating physician and healthcare site aimed at avoiding the selection bias and include a representative population of Greek COPD patients. The study provided valuable information regarding the demographic and medical profile of COPD patients in Greece, their classification based on the revised ABCD assessment tool, and whether clinicians in Greece prescribe medications for the chronic treatment of the disease in agreement with the GOLD 2017 recommendations. It is worth mentioning that the main reason for designing this study around the GOLD classification and adherence to the 2017 revision is the lack of any national (Greek) guidelines published or proposed in COPD and the fact that most physicians of the country refer to the GOLD documents for the management of COPD.

The outcomes of this study indicate that only an exceedingly small fraction of COPD patients could be registered into group C of the ABCD assessment tool, proposed under the revised GOLD 2017 suggestions. Only a small minority of patients suffered from two or more annual exacerbations, or at least from one serious enough to lead to hospitalization, and at the same time claim that their daily routine is almost as good as normal (CAT score < 10). Moreover, this study shows that there was no strong relationship between the ABCD classification tool with COPD severity, as indicated by the spirometric measurements. Most patients, from all ABCD groups, suffer from COPD of moderate severity.

Moreover, the ABCD classification system corelates with the age of COPD patients, the duration of their disease, as well as the number of comorbid conditions they suffer from. Patients of Group C and D were on average older than those of Group B, who were on average older than those of Group A. Similarly, patients of Groups C and D suffered from COPD for a longer period compared to those of Group B, who have suffered from COPD for a longer period compared to those of Group A. Finally, Group D patients had significantly more chances of having three chronic conditions (including COPD) compared to Group A or B patients, who usually suffered from up to one more chronic condition (besides COPD).

Furthermore, the outcomes of this study indicate that, from a therapeutic point of view, there is an astonishingly low adherence to GOLD 2017 recommendations for clinicians in Greece, when it comes to Group A patients (just 18%), a moderate adherence when it comes to Group B or C patients (58–59%) and a good but not perfect adherence (exceeding 80%) when it comes to Group D patients. In patients belonging to Groups A–C, deviations from GOLD 2017 suggestions refer mainly to prescribing more drugs than necessary, or inappropriate drugs, rather than prescribing less drugs than supposed to. We should note, though, the fact that the ABCD tool is aimed at being used for the initial assessment and management of patients with COPD, and after receiving treatments, many patients experience a decrease in their symptoms and exacerbations frequency, thus if reassessed (like in the case of this study) with the ABCD tool, they may be grouped to less severe groups, only because their disease is well controlled by the treatments they receive. This could be an alternative explanation for this impressive mismatch between ABCD classification and applied treatment practices in Greece. If this is the case, COPD specialists need to consider more adequately de-escalation from initial pharmacological interventions. It should be noted that overtreatment of COPD patients in all GOLD groups, outside of evidence-based recommendations, is a common phenomenon observed both in primary care and in specialist settings [15].

The distribution of COPD patients into the different ABCD groups of the GOLD 2017 recommendations shows similarities as well as some differences, compared to other studies with similar recruitment criteria. For instance, only 2% of our study participants were distributed to Group C, which is the same portion also reported by Cui et al. [16] in China (in a sample of over 1200 patients), Tudoric et al. [17] from a sample of over 3300 patients from central and eastern Europe, and Duarte-de-Araújo et al. [18] in Portugal (in a sample of over 300 patients), although higher percentages have been also reported [19]. In the great majority of studies, just like in our work, group C is the smallest among the four groups of the ABCD system [16‐18, 20‐22], group B the biggest followed (in a few cases exceeded) by group D [16‐22]. The sociodemographic profile of our study sample is also remarkably close to that reported by other similar studies. Finally, other studies -similarly to ours- report the best adherence to GOLD 2017 treatment guidelines for Group D and increasing overprescribing practices as we move to previous groups [18‐21].

Two strong points of this study are the large number of COPD patients it contains and its nationwide extant, across the Greek territory. On the other hand, it is known that the GOLD report uses a fixed FEV1/FVC value (0.70) to define airway obstruction. This approach leads to a significant percentage of false-positive diagnoses of mild and moderate COPD. These patients, who are in good respiratory health may be incorrectly identified as having an abnormally low FEV1. This is a long-standing controversy with the use of the lower limit of normal for the FEV1/FVC ratio. Finally, as mentioned above, similar studies have been conducted in other parts of the world.

Furthermore, all patients included in the study were stable COPD patients, managed by specialist pulmonologists either in the outpatient clinics of the hospitals or in private practices. A part of the private practice visits may have been covered by the patients’ insurance and part on private expenses, whereas the vast majority of the hospital visits were covered by the patients' insurance. Since the study was not part of a referral program for COPD, we should not expect differences between the two populations. The analysis of all data was performed considering all the patients as a single population.

Based on the demographics and COPD-related medical history data of the study sample, we could infer that the most commonly encountered COPD patient involves an overweight or obese male individual aged > 50 years, usually with one or two other chronic comorbid conditions (most likely some type of cardiovascular disease), having a CAT score > 10 (thus belonging to group B or D according to the ABCD classification tool) and spirometric parameters indicative of a moderate stage of the disease, who also suffers from up to two, usually mild, exacerbations annually. Our data show a significant discrepancy between the complex ABCD system of the GOLD 2017 classification and real-life clinical practice that may reflect both the complexity of that classification and the need for better dissemination of treatment recommendations to practicing clinicians.