Respiratory Medicine 129 (2017) 199e206

Contents lists available at ScienceDirect

Respiratory Medicine
journal homepage: www.elsevier.com/locate/rmed

Initiating or changing to a fixed-dose combination of Fluticasone

propionate/Formoterol over Fluticasone propionate/Salmeterol: A
real-life effectiveness and cost impact evaluation
Simon Wan Yau Ming a, John Haughney b, Iain Small c, Stephanie Wolfe d, John Hamill e,
Kevin Gruffydd-Jones f, Cathal Daly g, Joan B. Soriano h, Elizabeth Gardener a,
Derek Skinner i, Martina Stagno d'Alcontres a, David B. Price a, b, *
a
Observational and Pragmatic Research Institute, Singapore
b
Academic Primary Care, University of Aberdeen, Aberdeen, UK
c
Peterhead Health Centre, Aberdeen, UK
d
Primary Research Ltd, Norwich, UK
e
McMullans Pharmacy, Belfast, UK
f
University of Bath, Bath, UK
g
Elmham Surgery, Norfolk, UK
h n Hospital Universitario de la Princesa (IISP) Universidad Auto
Instituto de Investigacio noma de Madrid, Madrid, Spain
i
Optimum Patient Care, Cambridge, UK

a r t i c l e i n f o a b s t r a c t

Article history: Objective: Asthma has a substantial impact on quality of life and health care resources. The identification
Received 4 May 2017 of a more cost-effective, yet equally efficacious, treatment could positively influence the economic
Received in revised form burden of this disease. Fluticasone propionate/Formoterol (FP/FOR) may be as effective as Fluticasone
21 June 2017
Salmeterol (FP/SAL). We evaluated non-inferiority of asthma control in terms of the proportion of pa-
Accepted 22 June 2017
tients free from exacerbations, and conducted a cost impact analysis.
Available online 24 June 2017
Methods: This historical, matched cohort database study evaluated two treatment groups in the Opti-
mum Patient Care Research Database in the UK: 1) an FP/FOR cohort of patients initiating treatment with
Keywords:
Asthma
FP/FOR or changing from FP/SAL to FP/FOR and; 2) an FP/SAL cohort comprising patients initiating, or
Cost-effectiveness remaining on FP/SAL pMDI combination therapy. The main outcome evaluated non-inferiority of effec-
Fixed-dose combination inhalers tiveness (defined as prevention of severe exacerbations, lower limit of the 95% confidence interval (CI) of
Formoterol the mean difference between groups in patient proportions with no exacerbations is
3.5% or higher) in
GINA patients treated with FP/FOR versus FP/SAL.
Real-life Results: After matching 1:3, we studied a total of 2472 patients: 618 in the FP/FOR cohort (174 patients
initiated on FP/FOR and 444 patients changed to FP/FOR) and 1854 in the FP/SAL cohort (522 patients
initiated FP/SAL and 1332 continued FP/SAL). The percentage of patients prescribed FP/FOR met non-
inferiority as the adjusted mean difference in proportion of no severe exacerbations (95%CI) was 0.008
(
0.032, 0.047) between the two cohorts. No other significant differences were observed except acute
respiratory event rates, which were lower for patients prescribed FP/FOR (rate ratio [RR] 0.82, 95% CI 0.71,
0.94).
Conclusions: Changing to, or initiating FP/FOR combination therapy, is associated with a non-inferior
proportion of patients who are severe exacerbation-free at a lower average annual cost compared
with continuing or initiating treatment with FP/SAL.
© 2017 Elsevier Ltd. All rights reserved.

* Corresponding author. Academic Primary Care, Division of Applied Health Sci-
ences, University of Aberdeen, Polwarth Building, Foresterhill, Aberdeen AB25 2ZD,
UK.
E-mail address: dprice@opri.sg (D.B. Price).
1. Introduction
Asthma is a heterogeneous disease characterised by chronic
airway inflammation. The global prevalence of this disease has

http://dx.doi.org/10.1016/j.rmed.2017.06.016
0954-6111/© 2017 Elsevier Ltd. All rights reserved.
200 S. Wan Yau Ming et al. / Respiratory Medicine 129 (2017) 199e206
increased by 9.5%, from 334 million affected in 2005 to 358 million
people in 2015 [1,2]. Asthma has a substantial impact on quality of
life and health care resources, and its associated burden is high.
Treatments aimed at optimal disease control and prevention of
acute exacerbations are well established in national and interna-
tional guidelines [3]. However, international patient surveys have
found that asthma is often uncontrolled, underlining the need for
optimisation of asthma management [4,5].
Inhaled corticosteroids (ICS) have become a mainstay of asthma
therapies [6,7] with a short-acting b2-agonist (SABA) often pre-
scribed for quick symptom relief. However, many patients being
treated with ICS and SABA do not have fully controlled symptoms
[3]. The addition of a long-acting b2-agonist (LABA) to increase
asthma control as an alternative to upward titration of ICS doses is
well established [8].
The co-prescription of ICS with a separate LABA has become an
integral part of asthma treatment guidelines [3,9]. The develop-
ment and prescription of a fixed-dose combination (FDC) ICS/LABA
has been shown to improve patient compliance, consequently
reducing the risk of ICS therapy discontinuation, as well as reducing
the health care costs associated with separate drug inhalers, and
importantly, reducing respiratory-related deaths and life-
threatening episodes [6,7,9e12]. Many different FDC ICS/LABA
combination therapies are now available. An FDC ICS/LABA
composed of Fluticasone propionate (FP) and Formoterol fumarate
(FOR) has been developed for asthma treatment. FP/FOR combines
the anti-inflammatory and bronchodilating properties of two
effective compounds in a single inhaler. It has been suggested that
the bronchodilator effects of Formoterol reduce the need for rescue
medication and result in an increased number of episode-free days
compared with patients receiving an alternate LABA, Salmeterol
[13e15]. Its quick onset of action has a positive impact on patient
adherence [16]. Other FDC therapies, such as FP and Salmeterol (FP/
SAL), have also been found to be highly effective at maintaining
early and sustained improvements in asthma control [17,18]. There
is evidence that the safety profile is similar for either Formoterol or
Salmeterol combined with Fluticasone propionate (i.e. FP/FOR and
FP/SAL) [19].
All the above-mentioned factors are recognised by the Global
Initiative for Asthma (GINA) as being as vital to clinical outcomes
as the inhaled drug itself [3]. Asthma care in the United Kingdom
(UK) costs the National Health Service (NHS) over GBP 1 billion
per year [20]. The prescription of a more cost-effective, yet
equally efficacious, treatment has the potential to positively in-
fluence the economic burden of asthma care. Within the UK,
Seretide®, a formulation of FP/SAL, has the greatest volume of
prescribed units to treat patients with asthma [21]. A recent
alternative, Flutiform®, a formulation of FP/FOR, was launched in
the UK in September 2012 and currently accounts for only 4.6% of
overall ICS/LABA units prescribed for patients with asthma
compared with the 51.8% prescribed Seretide® [21]. Our previous
studies found that 88.4% of patients that were switched from FP/
SAL to FP/FOR remained on the new therapy, suggesting a high
retention rate for FP/FOR [22]. We also established that FP/FOR
provides an effective treatment option for patients with asthma
[23].
The objective of this study was to determine non-inferiority of
asthma control in patients prescribed FP/FOR versus those pre-
scribed FP/SAL, and to conduct a cost impact analysis of FP/FOR
compared with FP/SAL. Asthma control is defined as the proportion
of patients who are free from severe exacerbations during the study
period.
2. Methods
2.1. Study design
This was a historical, matched cohort study comparing patients
with asthma prescribed FP/FOR (the active comparator) with those
prescribed FP/SAL (the reference group). To represent real-life
treatment pathways, we initially identified patients according to
two treatment paths: an initiation path comprising those initiating
their first FDC ICS/LABA as either FP/FOR or FP/SAL on the index
date, and a change path, comprising individuals already on FP/SAL
who either received a repeat prescription for FP/SAL, or changed to
FP/FOR on the index date. Thus, the final FP/FOR cohort consisted of
patients initiating FP/FOR combination therapy and changing from
FP/SAL to FP/FOR, while the final FP/SAL cohort consisted of pa-
tients remaining on, or initiating FP/SAL combination therapy. The
index date was defined as the date when patients received their
first prescription of ICS as either FP/FOR or FP/SAL or changed to FP/
FOR.
The study was designed, implemented and reported in accor-
dance with the criteria of the European Network of Centres for
Pharmacoepidemiology and Pharmacovigilance (ENCePP number
ENCEPP/SDPP/12631) and followed the ENCePP code of conduct.
2.2. Data source
The study utilised data from the Optimum Patient Care Research
Database (OPCRD) [24]. The database has been approved by the
Trent Multicentre Research Ethics Committee for clinical research
use and, as of January 2017, contains anonymous, longitudinal data
for over 3 million patients from over 580 general practices in the
UK. The data includes information on diagnoses, prescriptions, in-
vestigations, hospital referrals, and admissions. Prescription costs
were taken from the National Health Service (NHS) Dictionary of
Medicines and Devices website [25], which contains up-to-date
prescription costs. Resource use costs were adapted from the Per-
sonal Social Services Research Unit (PSSRU) 2015 document [26].
2.3. Patient population
Patients eligible for the study were aged between 12 and 80
years, had a coded diagnosis compatible with asthma, and at least 2
years of continuous practice data comprising 1 baseline year and 1
outcome year. Patients were excluded from the analysis if their
records contained diagnostic codes for a chronic respiratory illness
other than asthma. Patients with more severe exacerbations who
were prescribed maintenance oral steroids were excluded. In
addition, patients from practices in the FP/FOR cohort where fewer
than 5 patients changed therapy were excluded, as the change
would more likely be precipitated by poor asthma control rather
than wholesale change for cost reasons. Inclusion and exclusion
criteria are specified in online supplementary material (Table S1).
2.4. Study outcomes
The primary outcome included patients with asthma experi-
encing no severe exacerbations in the outcome period and
compared the proportion of those who were prescribed FP/FOR
with those prescribed FP/SAL (non-inferiority analysis with a priori
limit for the difference in proportions set at 3.5%).
Secondary outcomes were assessed using previously described
composite measures [27e30]. Briefly, these included a comparison
 S. Wan Yau Ming et al. / Respiratory Medicine 129 (2017) 199e206
of: the rate of exacerbations; the proportion of patients with
frequent exacerbations (defined as 2 per year); risk domain
asthma control (defined by the absence of lower respiratory-
related hospital attendance or admission, emergency department
(ED) attendance, primary care physician consultation for lower
respiratory tract infections, or acute oral steroids); the acute res-
piratory event rate (defined as the prescription of antibiotics or oral
steroids in a lower respiratory primary care consultation or hospital
or ED attendance with a lower respiratory code); overall asthma
control (defined by the presence of risk domain asthma control and
>200 mg/day SABA use); SABA use; treatment stability; and the
number of lower respiratory-related hospitalisations.
Exploratory outcomes included comparison of ICS use,
controller/reliever ratio, and incidence of oral thrush in treatment
groups.
Cost impact outcomes included health care resource use, pre-
scriptions for respiratory-related conditions, and combined costs.
The total respiratory-related drug costs included prescriptions for
ICS, LABA, long-acting muscarinic antagonists (LAMA), leukotriene
receptor antagonists (LTRA), short-acting muscarinic antagonists
(SAMA), FDC ICS/LABA, theophylline, antibiotic and oral steroids
linked with a lower respiratory consultation. Lower respiratory-
related health care resources consisted of combined and dis-
aggregated primary care physician consultations, nurse asthma
reviews, ED attendance, and hospital admissions with a lower
respiratory-related diagnostic code.
2.5. Sample size
A sample size calculation was set a priori to determine our pri-
mary objective of evaluating non-inferiority of FP/FOR compared
with FP/SAL in preventing exacerbations, assessed in this study by
comparing the proportion of patients who were free of severe
asthma exacerbations during a one-year follow-up. The term
exacerbation is used based on the American Thoracic Society/Eu-
ropean Respiratory Society (ATS/ERS) Joint Taskforce definition: an
asthma-related hospital admission, emergency department atten-
dance, or an acute course of oral corticosteroids [31]. Preliminary
data from a historical, matched database study showed that the
difference in proportions of patients with no severe exacerbations
was 4.5% [32]. For the purposes of this study, a more stringent non-
inferiority limit of 3.5% was chosen for the upper limit of the 95%
confidence interval (CI) for the mean difference in primary out-
comes. With a minimum of 511 patients with asthma treated with
FP/FOR and 2044 treated with FP/SAL, this study provided 90%
power to reject the null hypothesis that FP/FOR and FP/SAL are non-
inferior in terms of the proportion of patients with no
exacerbations.
2.6. Statistical analysis
Statistical analysis was carried out using SPSS Statistics version
22 (IBM SPSS Statistics, Feltham, Middlesex, UK), and SAS version
9.3 (SAS Institute, Marlow, Buckinghamshire, UK).
Demographic and clinical characteristics, comorbidities, and
previous asthma exacerbations with p < 0.05 were considered as
potential confounders during modelling. In the event of significant
differences within the cohorts for certain variables, an assessment
was made for clinical relevance by a panel of medical experts and
the relevant variable was considered for matching criteria.
The final matching criteria were cohort path type (initiation or
change), age, gender, number of asthma exacerbations, SABA daily
dose (0e150 mg, 151e300 mg, 301e450 mg, 451e600 mg, >600 mg),
ICS daily dose (0e250 mg, 251e500 mg, >500 mg), smoking status,
and rhinitis diagnosis. Exact matching for categorical variables and
201
coarsened exact matching for numeric variables were used to
match patients using a ratio of 1:3 nearest neighbour matching
between treatment arms without replacement. We used a ratio of
1:3 to improve the power of the study with minimal loss of patients
in the intervention arm.
Conditional logistic regression was used for the primary, non-
inferiority analysis of the proportion of patients experiencing no
severe asthma exacerbations, and for the secondary outcomes of
the proportion of patients with frequent asthma exacerbations
(defined as 2 per year), risk domain asthma control, overall
asthma control and treatment stability. Conditional Poisson
regression was used to compare rate ratios of severe asthma ex-
acerbations, clinical asthma exacerbations and hospitalisations
between cohorts. Conditional ordinal regression was used for SABA
use. Minimal adjustment for residual confounding was made in
each of the fitted models. Adjustment for residual confounding was
based on non-collinear predictors of outcomes determined from
multivariable analyses on the unmatched sample and the balance
of cohorts after matching, together with clinical input. A p-value
lower than 0.05 was considered to be statistically significant.
3. Results
We identified a total of 671 eligible patients in the FP/FOR group
(203 initiated FP/FOR combination therapy; 468 changed from FP/
SAL to FP/FOR) and 67,054 in the FP/SAL group (25,719 initiated FP/
SAL combination therapy; 41,335 remained on FP/SAL) that met the
inclusion criteria (Figs. S1 and S2).
Matching resulted in 2472 unique patients: 618 in the FP/FOR
cohort who were matched 1:3 with 1854 in the FP/SAL cohort
(Table S2).
The mean age (SD) was 50 years (18) and 62% were female for
both matched groups. The percent predicted peak flow reading was
76% for both FP/FOR and FP/SAL. Each final cohort was composed of
55% non-smokers, 21% current smokers and 24% ex-smokers. Active
rhinitis, defined by the use of rhinitis drugs and a physician diag-
nosis, was 18% in both cohorts (Table 1).
At baseline, a higher percentage of patients in the FP/FOR cohort
were found to have better asthma control compared with the FP/
SAL group, 57% vs. 53%, p ¼ 0.005 (Table 1).
3.1. Main analysis: primary outcome e proportion of ‘no
exacerbations’
The matched combined FP/FOR cohort was found to be non-
inferior for FP/FOR compared with FP/SAL, in terms of the propor-
tion of patients with no severe exacerbations [0.008 (
0.032,
0.047)] using the non-inferiority boundary of
3.5% (Table 2).
3.2. Secondary outcomes
The acute respiratory event rate was significantly lower for the
combined FP/FOR cohort compared with the combined FP/SAL
cohort ([RR] 0.82, 95% CI 0.71e0.94). Other secondary outcomes
were not significantly different between the two cohorts (Table 3).
3.3. Exploratory outcomes
Prescription of FP/FOR compared to FP/SAL is associated with a
greater medication possession (p ¼ 0.017) and controller to reliever
ratio (p < 0.001) (Table S3). The number of associated lower res-
piratory consultations was lower for the FP/FOR group compared
with the FP/SAL group (p < 0.001). Other exploratory outcomes
were non-significant between cohorts.
202 S. Wan Yau Ming et al. / Respiratory Medicine 129 (2017) 199e206

Table 1
Baseline matched demographics and clinical characteristics.

Characteristic FP/FOR FP/SAL Total p value

(n ¼ 618) (n ¼ 1854) (n ¼ 2472)

Age, mean (SD)* 50.0 (18.3) 49.9 (18.1) 49.9 (18.1) n/a
Female sex, n (%)* 384 (62.1) 1152 (62.1) 1536 (62.1) n/a
Initiation or Change path*, n (%)
Initiation 174 (28.2) 522 (28.2) 696 (28.2) n/a
Change 444 (71.8) n/a
Remain n/a 1332 (71.8)
Smoking status*, n (%)
Current smokers 130 (21.0) 390 (21.0) 520 (21.0) n/a
Ex-smokers 149 (24.1) 447 (24.1) 596 (24.1)
Non-smokers 339 (54.9) 1017 (54.9) 1356 (54.9)
Percent predicted peak flow readings
N (% non-missing) 413 (66.8) 1266 (68.3) 1679 (67.9)
Mean (SD) 76.3 (19.2) 75.5 (19.0) 75.7 (19.0) 0.213
Index year, mean (SD) 2013 (0.5) 2008 (3.1) 2009 (3.4) <0.001
Exacerbations (ATS/ERS definitiona), median (IQR)* 0 (0e1) 0 (0e1) 0 (0e1) 0.353
0, n (%) 430 (70) 1290 (70) 1720 (70) n/a
1, n (%) 120 (19) 360 (19) 480 (19) n/a
>2, n (%) 68 (11) 204 (11) 272 (11) n/a
Acute respiratory eventsb, median (IQR) 0 (0e1) 0 (0e1) 0 (0e1) 0.004
Mean (SD) 0.7 (1.1) 0.8 (1.1) 0.8 (1.1)
0, n (%) 354 (57) 978 (53) 1332 (54) 0.013
1, n (%) 148 (24) 504 (27) 652 (26)
>2, n (%) 116 (19) 372 (20) 488 (20)
Acute oral steroid courses, median (IQR) 0 (0e1) 0 (0e1) 0 (0e1) 0.965
0, n (%) 431 (70) 1299 (70) 1730 (70) 0.090
1, n (%) 119 (19) 360 (19) 479 (19)
>2, n (%) 68 (11) 195 (11) 263 (11)
Risk domain asthma controlc
Controlled, n (%) 354 (57) 978 (53) 1332 (54) 0.005
Uncontrolled, n (%) 264 (43) 876 (47) 1140 (46)
Overall asthma controld
Controlled, n (%) 172 (28) 455 (25) 627 (25) 0.016
Uncontrolled, n (%) 446 (72) 1399 (76) 1845 (75)
Lower respiratory tract infection consultations resulting in script for antibiotics, median (IQR) 0 (0e1) 0 (0e1) 0 (0e1) 0.003
Asthma consultations, median (IQR) 1 (1e2) 1 (1e3) 1 (1e3) 0.002
Asthma-review consultations, mean (SD) 1.2 (0.9) 0.9 (1.0) 1.0 (1.0) <0.001
Primary care consultations 8.7 (7.0) 8.9 (7.5) 8.9 (7.4) 0.394
SABA scripts, mean (SD) 5.3 (4.5) 5.0 (4.7) 5.0 (4.6) 0.012
Rhinitis diagnosis and drugs*, n (%) 110 (17.8) 330 (17.8) 440 (17.8) n/a
Hospital inpatient admissions
0, n (%) 617 (99.8) 1838 (99.1) 2455 (99.3) 0.098
>1, n (%) 1 (0.2) 16 (0.9) 17 (0.7)
ED attendance
0, n (%) 617 (99.8) 1836 (99.0) 2453 (99.2) 0.072
>1, n (%) 1 (0.2) 18 (1.0) 19 (0.8)

SD ¼ standard deviation; IQR ¼ interquartile range; ATS/ERS ¼ American Thoracic Society/European Respiratory Society; SABA ¼ short-acting beta agonist; FP/FOR-
¼Fluticasone propionate/Formoterol; FP/SAL¼Fluticasone propionate/Salmeterol.
*Matching criteria.
a
Defined as an occurrence of an unscheduled hospital admission/ED Attendance/out of hours attendance for asthma OR use of acute courses of oral steroids OR primary care
physician consultations for lower respiratory tract infection.
b
Defined as the prescription of antibiotics or oral steroids in a lower respiratory primary care consultation or hospital or ED attendance with a lower respiratory code.
c
Defined by the absence of asthma related hospital attendance or admission, ED attendance, primary care physician consultation for lower respiratory tract infections or
acute oral steroids.
d
Defined by the presence of risk domain asthma control and no treatment change.

Table 2
Primary outcome results.

Matched cohort Treatment group Mean difference in proportion of ‘no exacerbations’ (FP/FOR-FP/SAL), % (95% CI) Non-inferiority met? (Lower 95% CI > -3.5%)

FP/FOR FP/SAL
(n ¼ 618) (n ¼ 1854)

Exacerbations (ATS/ERS definition)

No, n (%) 458 (74) 1372 (74) 0.008 (
0.032, 0.047)
3.5%: MET
Yes, n (%) 160 (26) 482 (26)
n/a

Abbreviations: ATS/ERS, American Thoracic Society/European Respiratory Society; FP/FOR, Fluticasone propionate/Formoterol; FP/SAL, Fluticasone propionate/Salmeterol; CI,
confidence interval.
 S. Wan Yau Ming et al. / Respiratory Medicine 129 (2017) 199e206 203

Table 3
Summary of secondary outcomes.

Matched cohort Treatment group RR/OR* (95% CI)

FP/FOR FP/SAL
(n ¼ 618) (n ¼ 1854)

Exacerbation rate (ATS/ERS definition) 0.97a (0.82e1.14)

Mean (SD) 0.4 (0.8) 0.4 (0.8)
Median (IQR) 0 (0, 1) 0 (0, 1)
Frequent Exacerbation (ATS/ERS Definition)c 0.99b (0.72e1.36)
2, n (%) 61 (10) 174 (9)
Risk domain asthma control 1.11c (0.91e1.37)
Controlled, n (%) 383 (62) 1115 (60)
Uncontrolled, n (%) 235 (38) 739 (40)
Acute respiratory event rate 0.82d (0.71e0.94)
Mean (SD) 0.6 (1.0) 0.7 (1.2)
Median (IQR) 0 (0e1) 0 (0e1)
Overall asthma control 1.21e (0.97e1.52)
Controlled, n (%) 180 (29) 463 (25)
Uncontrolled, n (%) 438 (71) 1391 (75)
SABA use (daily dose) 0.95f (0.78e1.16)
200 mcg, n (%) 245 (40) 695 (37)
>200 & 400 mcg, n (%) 167 (27) 522 (28)
>400 mcg, n (%) 206 (33) 637 (34)
Treatment stability 1.10g (0.89e1.35)
Yes, n (%) 342 (55) 990 (53)
No, n (%) 276 (45) 864 (47)
Hospitalisations 2.07h (0.83e5.16)
Yes, n (%) 8 (1) 14 (1)
No, n (%) 610 (99) 1840 (99)

*Comparison is FP/FOR versus FP/SAL.

Abbreviations: IQR, interquartile range; ATS/ERS, American thoracic society/European Respiratory Society; SABA, short-acting beta agonist; CI, confidence interval; FP/FOR,
Fluticasone propionate/Formoterol; FP/SAL, Fluticasone propionate/Salmeterol; SAMA, short-acting muscarinic antagonist; GERD, gastro-oesophageal reflux disease; LTRA,
leukotriene receptor antagonist; NSAID; nonsteroidal anti-inflammatory drug; ICS, inhaled corticosteroid; CCI, Charlson comorbidity index.
a
Rate ratio: adjusted for baseline ICS daily dose, baseline acute oral steroid courses (categorised), baseline SAMA use (categorised) and baseline LTRA use. Unadjusted result
rate ratio 0.97 (95% CI, 0.82e1.15).
b
Odds ratio: logistic regression adjusted for GERD diagnosis, baseline acute oral steroid courses (categorised), baseline SAMA use, gender, baseline ICS daily dose (cat-
egorised), NSAID use and baseline LTRA use. Unadjusted odds ratio 1.06 (0.79, 1.41).
c
Odds ratio: adjusted for baseline asthma exacerbations (clinical definition), ICS daily dose, cardiovascular diagnosis, smoking status, LTRA use, gender, rhinitis diagnosis
and paracetamol use. Unadjusted odds ratio 1.08 (0.90, 1.29).
d
Rate ratio: adjusted for gender, heart failure diagnosis, NSAIDS use, rhinitis diagnosis, baseline exacerbations (ATS definition), GINA control code, baseline LTRA use and
baseline SAMA use. Unadjusted rate ratio 0.88 (0.77e1.00).
e
Odds ratio: adjusted for baseline exacerbations (clinical definition), baseline SABA dose, cardiovascular diagnosis and CCI score (categorised). Unadjusted odds ratio 1.23
(1.04e1.47).
f
Odds ratio: adjusted for baseline antibiotics, baseline SABA dose, CCI score (categorised), age and smoking status. Unadjusted odds ratio 0.93 (0.82e1.05).
g
Odds ratio: adjusted for baseline exacerbations (clinical definition), rhinitis diagnosis, GINA control code, age, gender, smoking status, number of GP consultations
(categorised) and LTRA use. Unadjusted odds ratio 1.08 (0.90, 1.29).
h
Odds ratio: adjusted for index year and baseline number of antibiotics. Unadjusted odds ratio 1.72 (0.72, 4.15).

3.4. Health care costs
Lower respiratory-related drug costs (including prescriptions
for all SABA, SAMA, ICS, ICS/LABA, LABA, LAMA, LTRA, theophylline,
lower respiratory-related antibiotics or oral corticosteroid pre-
scriptions) were significantly reduced in the FP/FOR cohort
compared with the FP/SAL cohort (median of GBP 316 versus GBP
360 per annum p < 0.001) (Table 4, Fig. 1).
Lower respiratory-related health care resource costs were
significantly decreased for the FP/FOR cohort compared with the
FP/SAL cohort, although in terms of clinical relevance there was no
difference (Table 4, Fig. 1).
The total respiratory-related health care costs (the aggregate of
drug costs and resource costs) were significantly different: costs
were lower for the FP/FOR cohort (median of GBP 351 versus GBP
399 per annum, p < 0.001) (Table 4, Fig. 1).

4. Discussion
Table 4
Cost impact in GBP of prescribed medications, resource use and total cost. The matched combined FP/FOR asthma cohort was non-inferior
Characteristic FP/FOR FP/SAL Total p-valuea compared with the combined FP/SAL asthma cohort in terms of the
(n ¼ 618) (n ¼ 1854) (n ¼ 2472) proportion of patients with no severe exacerbations.
Asthma-related drug cost (GBP) Using exacerbation prevention (ATS/ERS definition) and sec-
Median (IQR) 316 (198e493) 360 (219e557) 352 (215e544) <0.001 ondary care as the measure of treatment effectiveness the observed
Asthma-related resource cost (GBP) similarity is not unexpected, given that the active preventative
Median (IQR) 14 (14e42) 14 (0e56) 14 (14e56) <0.001
medication was FP in both prescription cohorts.
All asthma-related medical costs (GBP)
Median (IQR) 351 (228e557) 399 (248e638) 385 (242e614) <0.001 The total number of acute respiratory events in the outcome
period was significantly different between the two cohorts. This
Abbreviations: FP/FOR, Fluticasone propionate/Formoterol; FP/SAL, Fluticasone
propionate/Salmeterol.
indicates that prescription of FP/FOR may be associated with fewer
a
p-value obtained from fitting a generalised linear model using Gamma distri- lower respiratory primary care consultations, where the patient
bution and log link, with standard errors adjusted for matching. presents with a condition requiring antibiotics or oral
204 S. Wan Yau Ming et al. / Respiratory Medicine 129 (2017) 199e206
Fig. 1. Median asthma-related costs in GBP of prescribed medications, resource use, and total cost. Abbreviations: FP/FOR, Fluticasone propionate/Formoterol; FP/SAL, Fluticasone
propionate/Salmeterol.
corticosteroids, as compared with FP/SAL. Alternative explanations
for the observed difference include individual prescribing prefer-
ences of primary care physicians, differences in the coding system
used, and the use of different codes to record events. The observed
reduction in the number of associated lower respiratory primary
care consultations in the outcome period for the FP/FOR cohort
strengthens this hypothesis.
Furthermore, the acute respiratory event rate was found to be
significantly lower for the combined FP/FOR cohort compared with
the combined FP/SAL cohort. Although superiority of this outcome
was not the aim of the study, one possible explanation for the
reduced acute respiratory event rate in the FP/FOR cohort may be
that patients do not present to their primary care physician as often
for respiratory conditions owing to improved symptom control
with Formoterol as with Salmeterol. Alternatively, the difference in
average index date between the cohorts (2013 for FP/FOR and 2008
for FP/SAL) may indicate an increased awareness of effective
asthma treatment and less reliance on antibiotic prescription in
patients prescribed FP/FOR owing to improved disease manage-
ment. The large difference in index year (2013 for FP/FOR compared
with 2008 for FP/SAL) reflects the later licensing of FP/FOR (from
September 2012). SABA daily dose was similar for FP/FOR and FP/
SAL patients, which may indicate comparable day-to-day symptom
control when the disease is controlled. Other differences between
the two cohorts, such as better risk domain asthma control in the
FP/FOR cohort, are also potentially related to the difference in index
year as a result of improved data recording practices or changes in
management guidelines. Although there were many statistically
significant differences in the baseline characteristics between the
groups, the magnitude of most was not clinically meaningful.
Our real-world study supports findings reported in a rando-
mised controlled trial (RCT) where FP/FOR was found to be com-
parable to FP/SAL in terms of improvements in lung function and
measures of asthma control [14]. The authors of this study, in
accordance with other researchers, found FP/FOR to have a more
rapid onset of action and suggest that if patients perceive the
benefits of FP/FOR more rapidly than FP/SAL, this could have a
positive impact on adherence [13,14]. This correlates with a study
by Bender et al., who found that patients expressed a wish for more
immediate symptom relief [33].
This study found that the cost impact of FP/FOR was lower than
that of FP/SAL and is driven by the associated lower total respira-
tory drug cost. The exploratory lower resource cost analysis also
indicates that there are fewer demands on health care resources
irrespective of drug costs. This was reiterated in a recent budget
analysis where FP/FOR was revealed to be the least costly option for
the NHS in comparison with FP/SAL [13]. Although a more in-depth
health economic analysis would provide a more robust comparison,
our cost impact analysis shows that the lower current price of FP/
FOR is associated with lower overall respiratory-related costs
compared with FP/SAL prescriptions.
Our data provide evidence that despite the use of ICS/LABA,
there is still a large proportion of patients (>70%), in both cohorts,
who are uncontrolled and >60% of patients using SABA rescue
therapy daily. This agrees with previous studies such as Asthma
Insights and Reality in Europe (AIRE) and the International Asthma
Patient Insight Research (INSPIRE) study that have shown that only
a small proportion of patients achieve guideline targets for asthma
control [34,35]. The INSPIRE study found that >70% of patients
prescribed ICS/LABA had uncontrolled asthma according to Asthma
Control Questionnaire data. Our findings also reflect those found in
this global study where 74% of patients used their SABA therapy
every day [35], further demonstrating poor asthma control. Mul-
tiple factors have been identified as contributors to the lack of
control including incorrect diagnosis, sub-optimal inhaler tech-
nique, poor compliance, smoking and co-morbidities [36], and
patient choice [37]. Underestimation of disease severity and hence
insufficient prescription or therapy-resistant disease is also a
prevalent problem. Our previous studies including Recognise
Asthma and Link to Symptoms and Experience (REALISE) [38,39]
and the EUrope and CANada (EUCAN) study [40] demonstrated
that patients with asthma reported features indicating uncon-
trolled disease though the majority believed to be controlled. This
implies that patients have low expectations of long-term asthma
management. Regardless of the cause, these data show that there is
a large proportion of patients globally that have suboptimally
controlled asthma.
Real-world evidence corroborates and reinforces the results
 S. Wan Yau Ming et al. / Respiratory Medicine 129 (2017) 199e206
from clinical trials. The use of a large database enabled us to analyse
data on real-life patients from a high-quality source, the OPCRD,
which has previously been used in respiratory research [24,41,42].
This allowed us to include a broad range of patients comparable to
those typically examined in primary care settings. In addition, we
studied outcomes over a full year both before and after the index
patient review/change to balance seasonal influences and other
transient confounders on outcome measures. Finally, the matching
identified a comparator group that had most key variables evenly
distributed at baseline.
As with other observational studies, there is a potential for se-
lection and physician bias and residual confounding. We addressed
this by performing a matched analysis, and residual confounding
was reduced by undertaking adjusted analyses. While the OPCRD is
a well validated and maintained database [24,41], we cannot rule
out the possibility of inaccurate or missing data. To achieve the
objectives, only patients with complete data for the study period
were included, precluding evaluation of asthma-related deaths and
other rare adverse effects with either inhaler. We relied on pre-
scription information from the OPCRD primary care database, thus
the data analysed allowed us to identify instances of when pre-
scriptions were written by general practitioners (GPs). It does not
include information on whether the prescription was filled and
therefore, we are unable to guarantee consumption of the medi-
cations in this study. Study findings are dependent on the avail-
ability and quality of the data in the OPCRD, which contains limited
information on hospitalisations. Lastly, there are differences in in-
dex date because FP/FOR was licensed later than FP/SAL in the UK.
This may have influenced study outcomes due to advances in
asthma treatment guidelines.
5. Conclusion
Changing to or initiating patients on FP/FOR rather than FP/SAL
for asthma treatment is associated with non-inferior clinical out-
comes. An exploratory analysis showed that FP/FOR is associated
with improved prescription possession as well as lower respiratory
related medication, health care resource use, and overall aggre-
gated cost in a real-life UK patient population.
Acknowledgements
With institutional support from NAPP Pharmaceutical Group
Ltd. All named authors meet the International Committee of
Medical Journal Editors (ICMJE) criteria for authorship for this
manuscript, take responsibility for the integrity of the work and
have given final approval to the version to be published. The au-
thors would like to thank Rosalind Bonomally for medical writing
support.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.rmed.2017.06.016.
Disclosures
SWYM is an employee of the Observational & Pragmatic
Research Institute (OPRI). Observational and Pragmatic Research
Institute Pte Ltd conducted this study, with institutional support
from Mundipharma and has conducted paid research in respiratory
disease on behalf of the following organizations: UK National
Health Service, British Lung Foundation, Aerocrine, AKL Research
and Development Ltd, AstraZeneca, Boehringer Ingelheim, Chiesi,
Meda, Mundipharma, Napp, Novartis, Pfizer, Respiratory
205
Effectiveness Group, Takeda, Teva Pharmaceuticals, Theravance,
and Zentiva.
JH has received reimbursements for attending symposia, fees for
speaking, organising educational events, funds for research or fees
for consulting from Almirall, AstraZeneca, Boehringer Ingelheim,
Chiesi, Cipla, GlaxoSmithKline, Mundipharma, Novartis, Pfizer and
Teva.
IS has received honoraria and support to attend international
conferences from Almirall, AZ, Boehringer Ingelheim, Chiesi, GSK,
NAPP, Novartis, and Orion.
SW reports being an advisory-board member for Merck Sharp &
Dohme.
JH has received honorarium for presentations on behalf of Pfizer,
NAPP, AstraZeneca, GSK, Teva, Almirall, Novartis and Boehringer
Ingelheim; sponsored to attend National and International con-
ferences including registration, flights, accommodation and food by
NAPP, Teva and Pfizer/Novartis.
KGJ has spoken on behalf of and acted as a consultant for
AstraZeneca, Boehringer Ingelheim, Chiesi, GlaxoSmithKline,
Mundipharma/NAPP.
CD has spoken at training events sponsored by NAPP, Teva and
AstraZeneca.CD has attended advisory boards sponsored by
AstraZeneca, Teva, Boehringer Ingelheim, Merck Sharp & Dohme,
Almirall and Actavis.
JBS participated in training, conferences, scientific committees
and/or advisory boards sponsored by: AirLiquide, Almirall, Astra-
Zeneca, Boehringer-Ingelheim, CHEST, Chiesi, ERS, GEBRO, Grifols,
GSK, Linde, Lipopharma, Mundipharma, Novartis, Pfizer, OPRI, Rovi,
SEPAR and Takeda.
EG is an employee of the Observational & Pragmatic Research
Institute (OPRI).
DS in an employee of Optimum Patient Care (OPC).
MSDA is an employee of the Observational & Pragmatic
Research Institute (OPRI).
DP has board membership with Aerocrine, Amgen, AstraZeneca,
Boehringer Ingelheim, Chiesi, Meda, Mundipharma, Napp, Novartis,
and Teva Pharmaceuticals; Chiesi, Meda, Mundipharma, Napp,
Novartis, and Teva Pharmaceuticals; consultancy with Almirall,
Amgen, AstraZeneca, Boehringer Ingelheim, Chiesi, GlaxoSmithK-
line, Meda, Mundipharma, Napp, Novartis, Pfizer, Teva Pharma-
ceuticals, and Theravance; grants and unrestricted funding for
investigator-initiated.
Studies (conducted through Observational and Pragmatic
Research Institute Pte Ltd) from UK National Health Service, British
Lung Foundation, Aerocrine, AKL Research and Development Ltd,
AstraZeneca, Boehringer Ingelheim, Chiesi, Meda, Mundipharma,
Napp, Novartis, Pfizer, Respiratory Effectiveness Group, Takeda,
Teva Pharmaceuticals, Zentiva, and Theravance; payment for lec-
tures/speaking engagements from Almirall, AstraZeneca, Boeh-
ringer Ingelheim, Chiesi, Cipla, GlaxoSmithKline, Kyorin, Meda,
Merck Mundipharma, Novartis, Pfizer, Skyepharma, Takeda, and
Teva Pharmaceuticals; payment for manuscript preparation from
Mundipharma and Teva Pharmaceuticals; payment for the devel-
opment of educational materials from Novartis and Mundipharma;
payment for travel/accommodation/meeting expenses from Aero-
crine, Boehringer Ingelheim, Mundipharma, Napp, Novartis, Teva
Pharmaceuticals, and AstraZeneca; funding for patient enrolment
or completion of research from Chiesi, Teva Pharmaceuticals, Zen-
tiva, and Novartis; stock/stock options from AKL Research and
Development Ltd which produces phytopharmaceuticals; owns 74%
of the social enterprise Optimum Patient Care Ltd, UK and 74% of
Observational and Pragmatic Research Institute Pte Ltd, Singapore;
and is peer reviewer for grant committees of the Medical Research
Council, Efficacy and Mechanism Evaluation Programme, and
Health Technology Assessment.
206 S. Wan Yau Ming et al. / Respiratory Medicine 129 (2017) 199e206

References [21] E. Farrington, et al., FP/FORM versus FP/SAL within clinical practice: an
updated budget impact analysis in asthma, Adv. Ther. 33 (5) (2016) 794e806.
[22] I.S. Daina Lim, Stephanie Wolfe, John Hamill, Kevin Gruffydd-Jones,
[1] F.D. Martinez, D. Vercelli, Asthma, Lancet 382 (9901) (2013) 1360e1372.
Cathal Daly, David Price, Real life effectiveness of changing fixed-dose com-
[2] G.B.D. Disease, I. Injury, C. Prevalence, Global, regional, and national incidence,
bination therapy from Seretide® metered dose inhaler (MDI) to Flutiform® in
prevalence, and years lived with disability for 310 diseases and injuries, 1990-
UK patients with asthma, in: Respiratory Effectiveness Group (REG) Summit
2015: a systematic analysis for the Global Burden of Disease Study 2015,
2014, England, London, 2014.
Lancet 388 (10053) (2016) 1545e1602.
[23] A.o.b.o.t.R.E. Group, Meeting abstracts from the Respiratory Effectiveness
[3] Global Initiative on Asthma (GINA), Global strategy for asthma management
Group 2015 Winter Summit - databases and registries around the world:
and prevention [cited 2017 20th April]; Available from: http://ginasthma.org/,
maximizing the yield, Pragmatic Observational Res. 2016 (6) (2015) 13e38.
2016.
[24] OPCRD, The Optimum patient care research database (OPCRD) [cited 2017
[4] L.P. Boulet, et al., Evaluation of asthma control by physicians and patients:
19th April]; Available from: http://optimumpatientcare.org/opcrd/, 2016.
comparison with current guidelines, Can. Respir. J. 9 (6) (2002) 417e423.
[25] NHS, Dictionary of Medicines and Devices (dmþd) [cited 2017 20th April];
[5] K.F. Rabe, et al., Worldwide severity and control of asthma in children and
Available from: https://apps.nhsbsa.nhs.uk/DMDBrowser/DMDBrowser.do,
adults: the global asthma insights and reality surveys, J. Allergy Clin. Immunol.
2015.
114 (1) (2004) 40e47.
[26] Unit, P.S.S.R, Unit costs of health and social care [cited 2016 13 December];
[6] P. Montuschi, Pharmacotherapy of patients with mild persistent asthma:
Available from: http://www.pssru.ac.uk/project-pages/unit-costs/2015/index.
strategies and unresolved issues, Front. Pharmacol. 2 (2011) 35.
php, 2015.
[7] D.D. Sin, et al., Pharmacological management to reduce exacerbations in
[27] N. Barnes, et al., Asthma control with extrafine-particle hydrofluoroalkane-
adults with asthma: a systematic review and meta-analysis, JAMA 292 (3)
beclometasone vs. large-particle chlorofluorocarbon-beclometasone: a real-
(2004) 367e376.
world observational study, Clin. Exp. Allergy 41 (11) (2011) 1521e1532.
[8] D. Price, et al., Is there a rationale and role for long-acting anticholinergic
[28] G. Colice, et al., Asthma outcomes and costs of therapy with extrafine beclo-
bronchodilators in asthma? NPJ Prim. Care Respir. Med. 24 (2014) 14023.
methasone and Fluticasone, J. Allergy Clin. Immunol. 132 (1) (2013) 45e54.
[9] (BTS), B.T.S, BTS/SIGN British guideline on the management of asthma [cited
[29] D. Price, et al., Prescribing practices and asthma control with
2017 12 April ]; Available from: https://www.brit-thoracic.org.uk/standards-
hydrofluoroalkane-beclomethasone and Fluticasone: a real-world observa-
of-care/guidelines/btssign-british-guideline-on-the-management-of-asthma/,
tional study, J. Allergy Clin. Immunol. 126 (3) (2010) 511e518 e1-10.
2016.
[30] E. Israel, et al., Increased dose of inhaled corticosteroid versus add-on long-
[10] D.S. Postma, H.A. Kerstjens, N.H. ten Hacken, Inhaled corticosteroids and long-
acting beta-agonist for step-up therapy in asthma, Ann. Am. Thorac. Soc. 12
acting beta-agonists in adult asthma: a winning combination in all? Naunyn
(6) (2015) 798e806.
Schmiedeb. Arch. Pharmacol. 378 (2) (2008) 203e215.
[31] K.F. Chung, et al., International ERS/ATS guidelines on definition, evaluation
[11] L. Rosenhall, et al., Budesonide/Formoterol in a single inhaler (Symbicort)
and treatment of severe asthma, Eur. Respir. J. 43 (2) (2014) 343e373.
reduces healthcare costs compared with separate inhalers in the treatment of
[32] D. Price, et al., Clinical effectiveness and cost impact in UK patients with
asthma over 12 months, Int. J. Clin. Pract. 57 (8) (2003) 662e667.
asthma switching from Fluticasone-propionate/Salmeterol pMDI to
[12] A.D. Bell, R.A. McIvor, The SMART study, Can. Fam. Physician 53 (4) (2007)
Fluticasone-propionate/Formoterol pMDI [Abstract 51], Prim. Care Respir.
687e688.
Med. (2015) 25.
[13] R. Aalbers, et al., Onset of bronchodilation with Fluticasone/Formoterol
[33] B.G. Bender, et al., Factors influencing patient decisions about the use of
combination versus Fluticasone/Salmeterol in an open-label, randomized
asthma controller medication, Ann. Allergy Asthma Immunol. 98 (4) (2007)
study, Adv. Ther. 29 (11) (2012) 958e969.
322e328.
[14] A. Bodzenta-Lukaszyk, et al., Fluticasone/Formoterol combination therapy is
[34] K.F. Rabe, et al., Clinical management of asthma in 1999: the asthma insights
as effective as Fluticasone/Salmeterol in the treatment of asthma, but has a
and reality in Europe (AIRE) study, Eur. Respir. J. 16 (5) (2000) 802e807.
more rapid onset of action: an open-label, randomized study, BMC Pulm. Med.
[35] M.R. Partridge, et al., Attitudes and actions of asthma patients on regular
11 (2011) 28.
maintenance therapy: the INSPIRE study, BMC Pulm. Med. 6 (2006) 13.
[15] J.J. Condemi, Comparison of the efficacy of Formoterol and Salmeterol in pa-
[36] J. Haughney, et al., Achieving asthma control in practice: understanding the
tients with reversible obstructive airway disease: a multicenter, randomized,
reasons for poor control, Respir. Med. 102 (12) (2008) 1681e1693.
open-label trial, Clin. Ther. 23 (9) (2001) 1529e1541.
[37] J. Haughney, et al., Features of asthma management: quantifying the patient
[16] A. Ullman, et al., Onset of action and duration of effect of Formoterol and
perspective, BMC Pulm. Med. 7 (2007) 16.
Salmeterol compared to salbutamol in isolated Guinea pig trachea with or
[38] D. Price, M. Fletcher, T. van der Molen, Asthma control and management in
without epithelium, Allergy 47 (4 Pt 2) (1992) 384e387.
8,000 European patients: the REcognise asthma and LInk to symptoms and
[17] R. Aalbers, et al., Adjustable maintenance dosing with budesonide/Formoterol
experience (REALISE) survey, NPJ Prim. Care Respir. Med. 24 (2014) 14009.
compared with fixed-dose Salmeterol/Fluticasone in moderate to severe
[39] D. Price, et al., Time for a new language for asthma control: results from
asthma, Curr. Med. Res. Opin. 20 (2) (2004) 225e240.
REALISE Asia, J. Asthma Allergy 8 (2015) 93e103.
[18] G. Shapiro, et al., Combined Salmeterol 50 microg and Fluticasone propionate
[40] J. Sastre, et al., Insights, attitudes, and perceptions about asthma and its
250 microg in the diskus device for the treatment of asthma, Am. J. Respir.
treatment: a multinational survey of patients from Europe and Canada, World
Crit. Care Med. 161 (2 Pt 1) (2000) 527e534.
Allergy Organ J. 9 (2016) 13.
[19] A. Bodzenta-Lukaszyk, et al., Efficacy and safety profile of Fluticasone/For-
[41] CPRD, The clinical practice research Datalink (CPRD) [cited 2017 19th April],
moterol combination therapy compared to its individual components
https://www.cprd.com/home/, 16 Dec 2015.
administered concurrently in asthma: a randomised controlled trial, Curr.
[42] R.C. Jones, et al., Opportunities to diagnose chronic obstructive pulmonary
Med. Res. Opin. 29 (5) (2013) 579e588.
disease in routine care in the UK: a retrospective study of a clinical cohort,
[20] NHS, An outcomes strategy for COPD and asthma: NHS companion document
Lancet Respir. Med. 2 (4) (2014) 267e276.
[cited 2016 13 Dec 2016]; Available from: https://www.gov.uk/government/
uploads/system/uploads/attachment_data/file/216531/dh_134001.pdf, 2012.