Introduction
Chronic obstructive pulmonary disease (COPD) is a heterogeneous disease with varying
phenotypes, i. e. clinical and pathophysiological differences in symptoms, airflow limitation,
frequency of exacerbations and comorbidities [1 ]. According to current guidelines, COPD is associated with a permanent airflow limitation
and an increased inflammatory response to inhalative noxae in the airways. The mainstays
of COPD treatment are long-acting bronchodilators and, depending on the exacerbation
risk, inhaled corticosteroids (ICS) [2 ]
[3 ]. Even though the European Medicines Agency concluded that the benefits of ICS medicines
in treating COPD, i. e. reduction of exacerbations, continue to outweigh their risks
[4 ], the concern of adverse effects, notably pneumonia [5 ] but also skin bruising, candidiasis, cataracts, osteoporosis and diabetes [6 ], is a constant matter of debate. The key challenge for physicians is to identify
patients with the best benefit/risk profile in clinical practice. Therefore, the identification
of a simple biomarker associated with a beneficial treatment response to ICS in COPD
is gaining more interest. Chronic airway inflammation in COPD is usually neutrophilic
but an elevated blood eosinophil count has been proposed as a potential biomarker
of ICS responsiveness in COPD [7 ]
[8 ]. Within the past five years, numerous post-hoc analyses have consistently associated
higher peripheral blood eosinophil counts with increased responsiveness of patients
with COPD to ICS [9 ]
[10 ]
[11 ] or a worsening of exacerbation frequency after ICS withdrawal [12 ]. This relationship has been supported by results of more recently published secondary
analyses, in which blood eosinophil count was shown to predict exacerbation risk and
clinical response to ICS [13 ]. Furthermore, in an extended analysis of the IMPACT study [14 ], which investigated inhaled triple therapy compared with fixed-dose combination
dual bronchodilators (long-acting beta-2 agonists [LABA]/long-acting muscarinic antagonists
[LAMA] and ICS/LABA), the baseline blood eosinophil count was linked to an ICS-associated
exacerbation reduction following a linear pattern. In the IMPACT study, the value
of blood eosinophil count in the management of COPD was prospectively confirmed [15 ]. This accumulating evidence on the role of blood eosinophils as a biomarker prompted
the Global Initiative for Chronic Obstructive Lung Disease (GOLD) committee to refine
the existing treatment algorithm by incorporating eosinophil counts into treatment
decisions. The 2020 revision of the GOLD report states that in exacerbating patients,
ICS and bronchodilator combination therapy could be considered as a first-line option
for patients with eosinophil counts of ≥ 300 cells/µl [2 ]. With blood eosinophil count as a biomarker to be utilized for phenotyping and predicting
treatment responses in COPD, there is a need to better understand when, why and how
frequently such blood tests and other measures are performed in daily clinical practice.
The aim of this non-interventional study was to evaluate the attitudes of private
German respiratory specialists towards the use of selected measures, including blood
eosinophil counts, and the execution of these measures in routine clinical care of
patients with COPD. We collected information on private German respiratory specialists’
opinions on the use of diagnostic tests at initial diagnosis and during follow-up
using a doctor’s questionnaire (DQ). A retrospective evaluation of patient medical
records was used to examine the actual execution of these tests within routine medical
care in the respective private practices.
Methods
This German multicentre non-interventional study consisted of a cross-sectional questionnaire
and a 12-month retrospective patient record study. The study was conducted from April
2018 to October 2018 (https://clinicaltrials.gov , NCT03465332) and was approved by the ethics committee of the Bavarian Medical Association
in Germany.
The study consisted of a pilot study and two main parts. In the pilot study, respiratory
specialists who were not participating in the survey tested the DQ and, based on the
results, the DQ was revised (a translated version of the DQ details can be found in
the supplement). During the first part of the main study, 27 private respiratory specialists
were enrolled and data on their perspectives on the diagnosis and treatment of patients
with COPD were collected via the DQ. The participating doctors were all respiratory
specialists working in private practices, to whom patients in Germany were either
referred or had direct access to.
Eligibility criteria for the participating doctors included more than ten years of
experience in respiratory medicine, supervision of at least 500 COPD patients (at
least 100 patients over the last 12 months), a focus on COPD, board certification
as a respiratory physician, and informed consent to participate in this study and
also share files of consenting patients (while complying with data protection rules).
In the second part, 251 patients with COPD were selected from participating doctors
and retrospective medical data were collected from patient records. Each site was
given a pre-defined minimum and maximum number of patients and a predefined recruitment
window. Every consecutive patient visiting the site within the recruitment window,
who met inclusion criteria and consented to participate in the study, was included.
Inclusion criteria for patients comprised written informed consent, being ≥ 40 years
of age, a current or former smoker with > 10 pack years, ≥ 12 months since initial
COPD diagnosis, no concurrent asthma diagnosis and ≥ 12 months of documented disease
history at participating study centres. Exclusion criteria for patients were pregnancy
and breastfeeding (in the last 12 months), and patients currently participating in
any interventional study and/or patients with severe comorbidities interfering with
COPD therapy. In accordance with the study design, the physicians transferred existing
retrospective data from selected patients who may have received any COPD medication
into an electronic case report form (eCRF).
The primary endpoints were the assessment of the selection of the diagnostic tests,
including blood eosinophil counts commonly used by private respiratory specialists
in routine clinical care in Germany, and the reasons for the choice of these tests.
Secondary endpoints included the investigation of any relationship between the blood
eosinophil test selection (based on the DQ) and drug selection (based on information
from the medical record review). Furthermore, comprehensive retrospective data were
transferred from the patient records into the eCRF on the use of diagnostic measures,
including lung function and imaging, patient-reported outcome (PRO) questionnaires
(COPD assessment test [CAT], modified Medical Research Council [mMRC] dyspnea scale),
COPD maintenance and other treatments in the previous 12 months, history of exacerbations,
hospitalisation due to COPD and selected concomitant diseases. A COPD exacerbation
was defined as a sustained worsening of respiratory symptoms that required treatment
with systemic corticosteroids, antibiotics, or hospital admission, or any combination
thereof.
Endpoints, including data from DQs and patient data, were analysed descriptively.
If needed for certain explorative analyses, statistical tests were applied, appropriate
to the level of measurement, e. g. the paired-samples t-test for continuous data or
Fisher’s exact test for categorical data. All tests performed are descriptive and
have no confirmatory character; therefore, no correction for multiple tests was applied.
Results
Doctors’ perceptions of common diagnostic measure usage
Twenty-seven centres participated in the study by completing the DQ and recruiting
patients. At all sites, participating doctors were private respiratory specialists,
who had treated at least 100 COPD patients in the previous 12 months; 19 sites (70.4 %)
reported having treated over 1000 patients in the previous 12 months.
According to the DQ, assessment of lung function performed by body plethysmography
was the most commonly used diagnostic measure by all respiratory specialists (100 %)
([Fig. 1 ]). It was performed for diagnosis in 24 centres (88.9 %), regularly in 26 centres
(96.3 %) and “when needed” in 5 centres (18.5 %) ([Table 1 ]). All physicians considered that body plethysmography was useful as a diagnostic
tool for COPD, as well as for monitoring of long-term therapy. The benefit for their
patients was considered “high” by 92.6 % of the physicians ([Table 1 ]). Regarding spirometry, 88.9 % of physicians stated that they usually use this method
([Fig. 1 ]); either regularly (70.4 %) or for diagnosis (59.3 %). The most frequently named
advantage of this measure was monitoring of short-term therapy, which was mentioned
by 88.9 % of the physicians.
Fig. 1 Diagnostic measures – results of the doctor’s questionnaire (DQ). *Multiple responses
were possible; AADT = alpha-1 antitrypsin deficiency; CT = computed tomography; IgE = immunoglobulin
E; N = 27 (number of respiratory specialists)
Table 1
Doctors’ reasons for the use of diagnostic measures in routine clinical practice – results
of the DQ
Diagnostic measures
BP
Spirometry
Chest X-ray
CAT
mMRC
Eosinophil count
Respiratory specialists, n (%)
Time/frequency of execution
[* ]
For diagnosis
24 (88.9)
16 (59.3)
16 (59.3)
12 (44.4)
7 (25.9)
16 (59.3)
Regularly
26 (96.3)
19 (70.4)
14 (51.9)
17 (63.0)
6 (22.2)
3 (11.1)
When needed
5 (18.5)
8 (29.6)
13 (48.1)
7 (25.9)
5 (18.5)
10 (37.0)
Advantages of this measure
[* ]
Diagnostic for…
27 (100)
20 (74.1)
22 (81.5)
15 (55.6)
15 (55.6)
18 (66.7)
13 (48.1)
8 (29.6)
16 (59.3)
4 (14.8)
4 (14.8)
15 (55.6)
Monitoring of…
24 (88.9)
24 (88.9)
7 (25.9)
17 (63.0)
12 (44.0)
4 (14.8)
27 (100)
23 (85.2)
13 (48.1)
20 (74.1)
15 (55.6)
8 (29.6)
Recommended by guidelines
9 (33.3)
8 (29.6)
3 (11.1)
5 (18.5)
5 (18.5)
3 (11.1)
Benefit for the patient
None
0
0
2 (7.4)
2 (7.4)
4 (14.8)
4 (14.8)
Low
0
2 (7.4)
1 (3.7)
3 (11.1)
6 (22.2)
4 (14.8)
Medium
2 (7.4)
10 (37.0)
11 (40.7)
17 (63.0)
11 (40.7)
14 (51.9)
High
25 (92.6)
15 (55.6)
13 (48.1)
5 (18.5)
6 (22.2)
5 (18.5)
* Multiple responses were possible; BP = body plethysmography; CAT = COPD assessment
test; COPD = chronic obstructive pulmonary disease; DQ: doctor’s questionnaire; mMRC = modified
Medical Research Council (dyspnea assessment test); N = 27 (number of respiratory
specialists)
Chest X-ray was the second most common diagnostic measure as stated in the DQ (96.3 %;
[Fig. 1 ]) and was mostly considered as a test for diagnosis (81.5 %); over half of the physicians
also used this for detection of concomitant diseases (59.3 %; [Table 1 ]). Other frequently used measures were symptom and health-related quality of life
assessment by CAT, the determination of alpha-1-antitrypsin deficiency (AADT), and
measurement of the differential blood count, each considered by 85.2 % of the physicians
([Fig. 1 ]). Physicians reported that the CAT was completed regularly in 63.0 % of cases and
was mainly considered useful for monitoring the effects of long-term therapy (74.1 %),
but also for monitoring of short-term therapy (63.0 %) and a useful diagnostic for
COPD (55.6 %) ([Table 1 ]). The eosinophil count was a blood test that was reported as “routinely used” by
81.5 % of the respiratory specialists ([Fig. 1 ]).
Further analysis of doctors’ perceptions of blood eosinophil tests revealed that more
than half of the respiratory specialists (59.3 %) reported that blood eosinophil counts
were analysed for diagnosis ([Table 1 ]). Only three physicians (11.1 %) indicated regular analysis of blood eosinophil
counts. Two-thirds of respondents regarded blood eosinophils as a useful diagnostic
tool for COPD and 55.6 % felt it was an important diagnostic tool for concomitant
diseases, followed by 29.6 % of respondents who found it to be useful for monitoring
of long-term therapy. The benefit of blood eosinophil counts for their patients was
evaluated as medium (51.9 %) ([Table 1 ]).
The physicians were asked to add “other” assessment tests applied in their routine
usage. In total, 36 additional tests were reported by 19 out of 27 physicians (70.4 %),
with the most frequently mentioned tests being blood gas analyses, other lung function
tests (e. g. diffusion capacity) and bronchodilator reversibility tests (data not
shown).
In addition to the evaluation of the commonly used measures, physicians were asked
which tests they would classify as relevant biomarkers. A majority of 24 out of 27
physicians (88.9 %), considered blood eosinophil count to be a relevant biomarker
for COPD. High-sensitive C-reactive protein was considered relevant by 13 (48.1 %)
and total IgE determination by 10 physicians (37.0 %) (data not shown).
Retrospective analysis of patient data
Patient demographics and disease characteristics
The respiratory specialists recruited 251 patients, of whom two were excluded from
analysis; one patient due to incomplete data and one patient due to a protocol deviation.
The evaluable patient population therefore comprised 249 patients.
Detailed patient characteristics including comorbidities and exacerbations during
the retrospective period are displayed in ([Table 2 ]). According to analysis of the medical records, a higher proportion of male than
female patients were included in the study (57.8 % vs. 42.2 %, respectively). The
mean age was 66.8 years, with 42.6 % of the patients being < 65 years. More than half
of the patients (64.3 %) had stopped smoking, while 35.7 % were current smokers. About
half of the patients (50.6 %) were categorised as having GOLD II disease (moderate
severity with a percentage predicted forced expiratory volume in 1 s [FEV1 ] value between 50 % and 79 %) and 31.7 % of patients were classified in high risk
GOLD groups C and D.
Table 2
Patient demographics and disease characteristics
Baseline
Patients
Sex, n (%)
144 (57.8)
105 (42.2)
66.8 (8.6)
Age groups, n (%)
106 (42.6)
143 (57.4)
6.4 (4.5)
Time groups, n (%)
27 (10.8)
222 (89.2)
Smoking status at baseline, n (%)
160 (64.3)
89 (35.7)
Pack years (years), mean (SD)
33.9 (18.0)
35.7 (15.5)
Allergic comorbidities, n (%)
233 (93.6)
Concomitant diseases (system organ classes), n (%)
133 (53.4)
79 (31.7)
69 (27.7)
40 (16.1)
39 (15.7)
32 (12.9)
23 (9.2)
GOLD 2017[* ] assessment of severity of obstruction, n (%)
15 (6.0)
126 (50.6)
79 (31.7)
29 (11.6)
GOLD 2017[* ] risk classes, A/B/C/D classification, n (%)
34 (13.7)
136 (54.6)
42 (16.9)
37 (14.9)
Retrospective 12-month period
Exacerbations per patient, mean (SD)
0.5 (0.9)
Number of exacerbations, n (%)
171 (68.7)
56 (22.5)
22 (8.8)
COPD = chronic obstructive pulmonary disease; SD = standard deviation
* Global Initiative for Chronic Obstructive Lung disease (GOLD) 2017; risk classes
based on COPD assessment test and exacerbations; N = 249 (number of patients)
Within the retrospective 12-month period, the largest proportion of all documented
COPD drugs was the combination of LABA and LAMA (18.6 %), followed by LAMA (16.8 %)
and short-acting beta-2 antagonists (SABA) (16.5 %). ICS, given alone or combined
with LABA and/or LAMA, represented 21.0 % of the documented drugs (data not shown).
Diagnostic measures – patient documentation
According to the retrospective data derived from patient records, body plethysmography
was performed in 72.7 % of all patients and spirometry in 30.9 % of patients ([Fig. 2 ]). During the 12 months prior to inclusion in the study, the CAT was used for more
than half of the patients (61.8 %), followed by chest X-ray in 40.6 % of patients
and the mMRC dyspnea scale in 22.1 % of patients. Blood eosinophil counts were performed
in 18 patients (7.2 %).
Fig. 2 Diagnostic measures – patient documentation. *Multiple responses were possible; body
plethysmography always included spirometry parameter; AADT = alpha-1 antitrypsin deficiency;
CT = computed tomography; IgE = immunoglobulin E; N = 249 (number of patients)
The mean percentage of blood eosinophils in all 18 patients was 1.5 % (± 1.3 %). The
proportion of patients receiving ICS was slightly higher among patients with blood
eosinophil results than in patients without, especially when single ICS and/or orally
administered prednisolone was given (22.2 % vs. 11.3 % and 22.2 % vs. 10.0 %, respectively)
(data not shown).
Diagnostic measures – subgroup analysis
The two different parts of the study (“DQ” and “patient data” [i. e. information from
the medical record review]) were compared with each other. In some diagnostic measures
(for example, the CAT and the mMRC dyspnea scale), the physicians’ opinions of their
routine use statistically significantly correlated with the retrospective data from
the patient records ([Table 3 ]). In contrast, the physicians’ statements on other routinely performed diagnostic
measures (including blood eosinophil count) were not reflected by their patients’
documentation.
Table 3
Comparison of diagnostic measures between DQ and patient documentation based on patient
population
Diagnostic measures
Patients recruited by physicians who reported use of the measure
Patients in whom the measures were performed
Physicians who reported routine use of the measure
Physicians who reported not to routinely use the measure
N = 249 (%)
% of patients
% of patients
P-value[* ]
Body plethysmography
249 (100)
72.7
–
– [** ]
Chest X-ray
242 (97.2)
41.7
0
0.0436
Spirometry
211 (84.7)
30.3
34.2
0.7035
COPD Assessment Test
199 (79.9)
67.8
38.0
0.0002
AATD screening
210 (84.3)
1.0
7.7
0.0285
Differential blood count
203 (81.5)
6.9
6.7
0.3171
Blood eosinophils
192 (77.1)
8.3
3.5
0.3801
Total IgE
199 (79.9)
5.5
0
0.1274
Specific IgE
197 (79.1)
4.6
0
0.2107
Medical examination of the skin
185 (74.3)
3.2
4.7
0.6981
CT-scan of thorax
183 (73.5)
6.6
4.5
0.7652
mMRC (dyspnea assessment test)
107 (43.0)
50.5
0.7
< 0.0001
AATD = alpha-1 antitrypsin deficiency; CT = computed tomography; DQ = doctor’s questionnaire;
IgE = immunoglobulin E
* Fisher’s exact test
** could not be calculated for this specific comparison
Discussion
In accordance with current GOLD recommendations [2 ] our results confirm that lung function analysis plays a crucial role in diagnosing
and monitoring COPD, since all physicians stated routine use of body plethysmography.
Lung imaging by chest X-ray and PRO-questionnaires such as the CAT and (less frequently)
the mMRC dyspnea scale were also commonly used measures, according to the DQ and the
documented patient data. Our analysis revealed that blood eosinophil count gained
importance due to increased physician awareness that blood eosinophils might qualify
as a useful biomarker in COPD, since 88.9 % of the physicians regarded them as such
and 81.5 % reported blood eosinophils to be a commonly used measure. However, according
to the retrospective analysis of patient documentation, blood eosinophil counts were
determined in only 18 out of 249 patients (7.2 %) in the observational phase from
19 April 2017 to 30 August 2018. This test was mainly performed for diagnosis and
not on a regular basis. More than two-thirds (66.7 %) of all physicians considered
the blood eosinophil count to be an initial diagnostic tool for COPD and 11.1 % reported
its use as a regular assessment.
The study aimed to evaluate the availability, relevance and execution of diagnostic
measures in real-life settings, which was accomplished by a cross-sectional DQ and
the retrospective evaluation of patient data. One limitation of the study was the
retrospective nature of the patient data analysis. Most importantly, at the time of
data collection by the DQ, the role of eosinophils was much less pronounced than today.
The international GOLD recommendations for using eosinophil counts in the treatment
algorithm were published after completion of our study. As yet, the evidence base
for using eosinophils in the COPD treatment algorithm is weak and the implementation
of eosinophils as a biomarker in German recommendations for the treatment of COPD
[3 ] is still ongoing. Therefore, it would be useful to repeat the study with the aim
of collecting information about the changing perceived importance of blood eosinophils
over time. On the contrary, the retrospective design can also be regarded as a strength
because the actions taken by participating physicians were not influenced by the study
design. In addition, a minimum set of inclusion and exclusion criteria allowed the
enrolment of a representative set of practices and patients throughout Germany, with
a broad range of COPD stages.
Although the number of analysed patients was relatively low, the patient characteristics
in our study reflect that of a typical real-life COPD population. The patients recruited
were very similar to the German DACCORD population, in which 49.2 % of patients were
categorized into the GOLD II stage at baseline (50.6 % in our study) and approximately
three-quarters were free from exacerbations in the previous year (68.7 % in our study)
[16 ]. The GOLD 2017 distribution of patients in our study was similar to that observed
in the international COPD gene cohort and the European real-world population (Adelphi
Real World Respiratory Disease Specific Programme) with ~30 % of patients being categorized
to high risk GOLD classes C and D [17 ]
[18 ]. No comorbidities were recorded for only 16.1 % of all patients. Vascular disorders
(e. g. hypertension) and cardiac diseases were the most common concomitant diseases,
as reported in other real-world studies [16 ]
[17 ]
[18 ].
Concerning lung function analysis, respiratory specialists stated regular use of body
plethysmography (100 % usage) and spirometry (88.9 %) via the DQ. The retrospective
analysis revealed that body plethysmography was performed in 72.7 % of patients and
spirometry in 30.9 %. As shown by the study eCRF, the body plethysmography panel mandatorily
requested spirometry values. Thus, all patients underwent spirometry as recommended
by both the GOLD and German guidelines. According to GOLD recommendations and the
German guidelines, body plethysmography is not mandatorily recommended for diagnosis
of COPD, but analysis of the complete set of lung volumes, including total lung capacity
and residual volume, is important and may be necessary in some cases for further differential
diagnosis [2 ]
[3 ]. Body plethysmography is available to respiratory physicians as an advanced diagnostic
tool. It is mandatorily included in the reimbursement system of German secondary respiratory
care. Therefore, it might be used more frequently than spirometry alone in German
private respiratory practices, which has been reported before in other studies [19 ]
[20 ].
In concordance with the current national and international COPD guidelines, chest
X-ray was chosen for exclusion of differential diagnoses by 81.5 % of all respiratory
specialists and for the determination of concomitant diseases by 59.3 %. In the patient
documentation, chest X-ray was performed in 40.6 % of patients, which was assumed
to be related to the fact that the first diagnosis of COPD for all patients was prior
to the observation period of one year.
Other analyses that were not performed on a regular basis included alpha-1 anti-trypsin
deficiency (AATD), medical examination of the skin, total immunoglobulin E (IgE) and
specific IgE. While the latter three are used to rule out allergic disorders, the
regular screening of COPD patients for AATD is recommended by national and international
guidelines [2 ]
[3 ]. However, the disconnect between the recommendations and clinical practice has been
published before [21 ].
Regarding the assessment of symptoms, the CAT was documented retrospectively in 61.8 %
of patients and the mMRC dyspnea scale in 22.1 %. Subgroup analyses, which aimed to
examine whether the measures indicated in the DQ correlated to the patient data documented
at the corresponding sites, revealed a significant association for both the CAT and
mMRC dyspnea scale. Thus, the physicians’ assessments of their routine usage did correlate
with the documented patient data. It should be noted that the CAT, which was documented
in a large proportion of patients, was stated in the DQ to be completed regularly.
The use of the CAT is recommended by current guidelines and is a well-proven, quick,
and reliable tool for the assessment of health-related quality of life [22 ].
In our study, the majority of physicians (88.9 %) stated that blood eosinophils are
a relevant biomarker of COPD, as indicated in the DQ. According to the retrospective
analysis, the blood eosinophil count was routinely performed in 7.2 % of patients.
In the DQ, most physicians reportedly performed the blood eosinophil count during
initial diagnosis, which all patients underwent before inclusion in the study. This
could explain the low utilisation of the blood eosinophil test. Nevertheless, a single
analysis of blood eosinophil count at initial diagnosis would not be sufficient to
guide the appropriate pharmacotherapy (e. g. to assign treatment with corticosteroids
or not). Varying stability of blood eosinophils over time was shown in the German
multicentre COSYCONET study [23 ], therefore regular analysis for all patients would be required to achieve concordant
analyses. On the other hand, very recent analysis of the IMPACT trial revealed that
two blood eosinophil count measurements do not appear to provide additional information
to predict ICS treatment response in COPD versus one measurement [24 ].
The subgroup of 18 patients who underwent blood eosinophil testing in our study received
more corticosteroids during the observational period (data not shown), thus the physicians
showed a tendency towards drug selection based on blood eosinophil count. However,
due to the low number of patients in that group, conclusions on the correlation of
corticosteroid use and blood eosinophil count should be interpreted with caution.
The results of our study cannot be fully representative for the following two limitations:
Firstly, the data collection time is referring to the years 2017 – 2018 and may likely
not be representative for the actual situation. Secondly, even though we succeeded
in including physicians with high expertise in the area, the doctor’s questionnaire
was completed exclusively by office-based, private respiratory specialists. Therefore,
it is unclear to what extent the findings apply to routine clinical practice in German
hospitals.
Conclusion
Our results confirm that lung function, imaging and the use of PRO questionnaires
play a crucial role in diagnosing and monitoring COPD. Our analyses may suggest that
that German respiratory physicians are aware of the role of blood eosinophil count
in the future management of COPD, however, during 2017 – 2018 eosinophil count analysis
was not yet implemented in German respiratory secondary care.
Disclosures
Writing support was provided by Dr. Angelika Schedel, Freelance Medical Writer located
in Nuremberg, Germany. This support included the development of the first and subsequent
drafts of the manuscript, under the guidance of the authors, and coordination of author
comments and approval, and was funded by GlaxoSmithKline.
Trademarks are the property of their respective owners.
