The CONCEPT Trial: A 1-Year, Multicenter, Randomized,
Double-Blind, Double-Dummy Comparison of a Stable
Dosing Regimen of Salmeterol/Fluticasone Propionate
with an Adjustable Maintenance Dosing Regimen of
Formoterol/Budesonide in Adults with Persistent Asthma
J. Marl< FitzGerald, FRCPCI; Louis-Philippe Boulet, FRCPC2; and
Richard M.A. Follows, BSc 3
1Respiratory Division, Vancouver General Hospital, Vancouver, Canada; 2The Laval University
Cardiothoracic Institute, Laval Hospital, Quebec City, Canada; and 3GlaxoSmithKline Research
Development Limited, Greenflord, Middlesex, United Kingdom
ABSTRACT
Background: A patient-driven, adjustable mainte-
nance dosing (AMD) approach to asthma therapy, in
which the dose is adjusted by patients according to
the severity of their symptoms, has recently been com-
pared with fixed-dose therapy in open-label studies.
Objective: This study used a double-blind, double-
dummy design to compare the efficacy of 2 treatment
approaches: stable dosing of salmeterol/fluticasone pro-
pionate (SAL/FP) and AMD of formoterol/budesonide
(FOR/BUD).
Methods: This was a 1-year, muhicenter, random-
ized, double-blind, double-dummy study in adult patients
with symptomatic asthma that was not controlled by
therapy with 200 to 500 pg/d inhaled corticosteroid
(ICS) plus a long-acting beta2-agonist , or with >500 to
1000 iag/d ICS alone. Patients were randomized to
receive 1 inhalation of SAL/FP 50/250 pg BID or 2
inhalations of FOR/BUD 6/200 pg BID, both delivered
via dry powder inhaler devices. After 4 weeks of stable
dosing in both groups, eligible patients continued the
study for an additional 48 weeks, receiving either
a stable dose of SAL/FP or AMD of FOR/BUD. Ac-
cording to the AMD treatment plan, patients initially
halved their dose and subsequently stepped up or down
as indicated by the presence or absence of nocturnal
awakenings due to asthma, frequency of rescue medica-
tion use, and changes in morning peak expiratory flow
(PEF). The primary end point was the percentage of
symptom-free days. Other parameters included daily
asthma symptom scores, morning PEF, percentage of
days free of rescue medication use, daily rescue medica-
tion use, percentage of nighttime awakenings due to
asthma, percentage of weeks with well-controlled
asthma, and number of exacerbations requiring oral
corticosteroids or emergency department (ED) visits/
hospitalizations. Tolerability was assessed in terms of
adverse events spontaneously reported or elicited at
clinic visits.
Results: The intent-to-treat population comprised
688 patients (344 per treatment arm) with a mean age
of 45 years and a mean baseline forced expiratory vol-
ume in 1 second 81% of the predicted normal value.
After 4 weeks' stable dosing, 581 patients (295 SAL/FP,
286 FOR/BUD) continued beyond visit 3 into the
remaining 48-week treatment period. Over weeks 1
through 52, patients receiving stable dosing of SAL/FP
had a significantly greater percentage of symptom-free
days compared with those receiving AMD of
FOR/BUD (median, 58.8% vs 52.1%, respectively; P =
0.034). The incidence of asthma exacerbations requir-
ing oral steroids or an ED visit/hospitalization was
47% lower with SAL/FP compared with FOR/BUD
(adjusted annual mean rate, 0.18 vs 0.33; P = 0.008).
During weeks 5 through 52, patients in the FOR/BUD
AMD group used a mean of 1.8 inhalations/d (equiva-
lent to BUD 360 l~g/d), and 235 (82.2%) patients
stepped down to 1 inhalation/d. Mean (SD) daily ICS
Accepted[orpublicationMarch1,2005.
Express track online publication March 23, 2005.
doi:l 0.1016/j.clinthera.2005.03.006
0149-2918/05/$19.00
Printed in the USA. Reproduction in whole or part is not permitted.
Copyright © 2005 Exeerpta Medica, inc.
exposure over 52 weeks was 463 (81) lag FP and 480
(238) lag BUD in the respective treatment arms.
Conclusions: In this adult population with persistent
asthma, stable dosing of SAL/FP 50/250 lag BID result-
ed in significantly greater increases in symptom-free
days, days free of rescue medication, and morning PEF,
as well as almost halving the exacerbation rate, com-
pared with AMD of FOR/BUD 6/200 lag. The results
suggest that there is a minimum daily amount of main-
tenance therapy necessary to prevent exacerbations in
adults with persistent asthma. (Clin Ther. 2005;27:393-
406) Copyright © 2005 Excerpta Medica, Inc.
Key words: adjustable maintenance dosing, asthma
control, salmeterol/fluticasone propionate, formoterol/
budesonide, exacerbations.
INTRODUCTION
The combination of a long-acting beta2-agonist (LABA)
with an inhaled corticosteroid (ICS) has been shown
to be highly effective in the treatment of asthma. 1-4
Both national s,6 and international 7 asthma treatment
guidelines recommend a combination of these 2 medi-
cation classes as the preferred option for the manage-
ment of all but the mildest asthma. Two LABA/ICS
combinations are currently available containing either
salmeterol (SAL) and fluticasone propionate (FP) or
formoterol (FOR) and budesonide (BUD). SAL/FP* is
available both as a dry powder inhaler and as a
metered-dose inhaler, whereas FOR/BUDt is available
only as a dry powder inhaler.
Current asthma management guidelines state that
the aim of treatment is to achieve and maintain asth-
ma control without undue adverse effects from the
therapies used. They recommend that patients take a
stable dose of preventive medication, with periodic
reviews by the physician to assess whether treatment
adjustments are required. The Global Initiative for
Asthma (GINA) has developed specific criteria relat-
ing to symptoms, exacerbations, emergency depart-
ment (ED) visits, use of rescue medication, limita-
tion of activity, and lung function that should be met
over a sustained period if the disease is to be consid-
ered controlled. 7
*Trademark: Seretide ®, Advair ®, Diskus®, Advair Diskus®,
Accuhaler (GlaxoSmithKline, Ware, United Kingdom).
TM
?Trademark: Symbicort ® Turbuhaler ® (AstraZeneca, S6dertalje,
Sweden).
A patient-driven, adjustable maintenance dosing
(AMD) treatment approach has recently been evaluat-
ed in open-label studies comparing AMD of FOR/BUD
with fixed-dose treatment. 8-1° With AMD, the dose of
FOR/BUD was adjusted by patients according to the
severity of their symptoms. After an initial period of
2 inhalations BID, the dose of FOR/BUD was stepped
down to 1 inhalation BID if symptoms were controlled
or was temporarily stepped up to 4 inhalations BID for
7 or 14 days if asthma worsened. These studies report-
ed a reduction in asthma exacerbations with the AMD
approach compared with fixed-dose treatment. The
AMD treatment approach is recommended by the
manufacturer of FOR/BUD as a means of minimizing
the overall amount of medication required. This entails
using a maintenance dose of 1 or 2 inhalations/d during
periods with no symptoms and increasing the dose only
temporarily (up to 8 inhalations/d) during periods of
worsening. Although the original studies of AMD used
a maintenance dose of 1 or 2 inhalations BID, the inter-
national data sheet for the FOR/BUD Turbuhaler states
that "when control has been achieved, the dose should
be titrated to the lowest effective dose, which could
include Symbicort Turbuhaler given once daily. ''n
The efficacy of SAL/FP and FOR/BUD given in stable
dosing regimens has been documented in numerous stud-
ies. 12,13 Two studies have directly compared the efficacy
of stable doses of SAL/FP 50/250 lag BID and FOR/BUD
12/400 l~g BID-l°'14 Both studies reported similar im-
provements in their primary end points (odds ratio for
achieving a well-controlled asthma week 1° and mean
rate of all exacerbations [mild, moderate, and severe] 14)
with both treatments. The study by Aalbers et aP ° also
included a 6-month, open-label comparison of a stable
dose of SAL/FP with AMD of FOR/BUD. Again, similar
improvements were seen in the primary end point with
the 2 treatments. However, based on a search of the lit-
erature (OVID, MEDLINE, Cochrane Database), there
have been no published randomized, double-blind,
controlled studies comparing stable dosing and AMD
of these 2 combinations. Therefore, the aim of the
present trial--the CONtrol CEntred Patient Treat-
ment (CONCEPT) study--was to compare the efficacy
of stable dosing of SAL/FP 50/250 lag BID via Diskus
with AMD of FOR/BUD 6/200 lag via Turbuhaler in the
management of patients with persistent asthma. It was
not a comparison of 2 drugs used at equipotent doses
(eg, according to the GINA criteria, "medium" doses
of FP and BUD are 250-500 pg/d and 400-800 lag/d,

respectively7), but of 2 treatment strategies using
products appropriate for the population studied.
PATIENTS AND METHODS
Inclusion and Exclusion Criteria
Male or female outpatients aged >18 and <70 years
with a documented clinical history of asthma (con-
firmed by the medical record) and a forced expiratory
volume in i second (FEV1) between 60% and 90% of
the predicted normal value were eligible for enroll-
ment in the study. All patients had used either an ICS
at a dose equivalent to 200 to 500 pg/d beclometha-
sone dipropionate (BDP) combined with a LABA, or an
ICS alone at a dose equivalent to >500 to 1000 pg/d
BDP for >12 weeks before enrollment.
Exclusion criteria included lower respiratory tract in-
fection or use of systemic corticosteroids within 1 month
of study entry, a >10 pack-year smoking history, changes
to regular asthma therapy within 12 weeks of study entry,
and any significant disorder that in the investigator's opin-
ion, might put the patient at risk or influence the study
outcomes. The following medications were prohibited
during the study: inhaled cromones, leukotriene modi-
fiers, beta2-agonists (except salbutamol provided as res-
cue medication), xanthines, and inhaled anticholinergics.
Study Design
This was a randomized, double-blind, double-
dummy, parallel-group study carried out at 91 centers
in 15 countries. It consisted of 3 periods: a 2-week
run-in period, a 52-week double-blind treatment peri-
od, and a 2-week follow-up period (Figure 1). The
SAL/FP 50/250 pg x 1 BID +
Dlacebo Turbuhaler x 2 BID
200-500 gg ICS + LABA
or >500-1000 pg/d ICS
(BDP equivalent)
Randomization
FOR/BUD 6/200 pgx 2 BID +
I p acebo Dis <usx 1 BID
Run-in Weeks 1 4
Visit: 1 2 3
Week: 2 0 4
Figure 1. Study design. ICS = inhaled corticosteroid; LABA = long-acting beta2-agonist; BDP = beclomethasone dipropi-
onate; SAL = salmeterol; FP = fluticasone propionate; Turbuhaler = trademark of AstraZeneca, S6dert~Ije,
Sweden; Diskus = trademark of GlaxoSmithKline, Ware, United Kingdom; AMD = adjustable maintenance dos-
ing (visit 3 to 4, I inhalation of Turbuhaler BID, increased to 4 inhalations BID as required; From visit 4 onward,
I inhalation once daily or BID, increased to 4 inhalations BID as required); FOR = Formoterol; BUD = budesonide.
al.
double-blind treatment period was split into 2 phases:
weeks 1 through 4, during which the dose delivered
from both devices was kept constant, and weeks 5
through 52, during which the dose delivered via Diskus
remained unchanged but the dose delivered via
Turbuhaler was initially reduced to 1 inhalation BID
and subsequently adjusted according to the physician-
guided, self-managed AMD plan (Table I).
The study was conducted in accordance with the
Declaration of Helsinki and good clinical practice
guidelines. The study protocol, patient information
form, and informed consent form were approved by the
local ethics committees. All patients provided written
informed consent before beginning the study. Patients
who withdrew were encouraged to return to the clinic
for a checkup and to return all study-related materials.
Run-in Period
During the 2-week open-label run-in phase, patients
continued to take their current asthma medication,
with as-needed salbutamol as rescue medication. To be
eligible for randomization, patients had to have a total
daily symptom score of >2 (on a scale from 0 to 5,
described in the Efficacy Measures section) on >4 of the
last 7 evaluable days of the run-in period. In addition,
they had to demonstrate the ability to use a peak flow
meter and to correctly record values on a diary card.
Double-Blind Treatment Period
Eligible patients were randomized to 1 of 2 treat-
ment groups: SAL/FP 50/250 lag or FOR/BUD 6/200 pg
(equivalent to a 4.5/160-lag emitted dose). During the
SAL/FP Diskus 50/250 I~g x 1 BID + placebo Turbuhaler AMD ........... I
FOR/BUD Turbuhaler 6/200 pg A M D + placebo Diskus x I BID L ....... I
Weeks 5-52 Follow-up
4 S 6 7
16 28 40 52 54
395
 Table I. Schedule for adjustable maintenance dosing of active treatment or placebo administered via Turbuhaler*
during weeks 5 through 52.

Adjustment Criteria

Step-up: From 1 or 2 inhalations/d to Two consecutive days or nights with:

4 inhalations BID (judged by the patient)
Step-down: From 4 inhalations BID to
1 inhalation BID after 7 or 14 days o£
step-up treatment (judged by the
investigator)
Rescue medication used >3 times during the day
OR
Nighttime awakening due to asthma
OR
Morning PEF <85% of the mean of the last 7 days before visit 3
Last 2 consecutive days or nights with:
No rescue medication use
AND
No nighttime awakening due to asthma
AND
Morning PEF >_85% of the mean of the last 7 days before visit 3

PEF = peak expiratory flow.

*Turbuhaler is a trademark ofAstraZeneca, S6dert~lje, Sweden.

first 4 weeks of treatment (weeks 1--4), patients in the
stable-dose SAL/FP group received 1 inhalation of
SAL/FP 50/250 pg BID via Diskus and 2 inhalations
of placebo BID via Turbuhaler; patients in the AMD
FOR/BUD group received 2 inhalations of FOR/BUD
6/200 pg BID via Turbuhaler and 1 inhalation of place-
bo BID via Diskus. Salbutamol was taken as needed as
rescue medication for the duration of the study.
After the first 4 weeks, patients were eligible to con-
tinue the study beyond visit 3 (weeks 5-52) only if, dur-
ing the 7 days before visit 3, they had no nighttime
awakenings due to asthma and had not used salbuta-
tool for symptom relief on >2 days, as assessed by the
investigator based on the patients' diary cards. Patients
who did not meet these criteria were withdrawn.
Patients meeting the criteria were instructed to reduce
the Turbuhaler dose from 2 inhalations BID (active
drug or placebo) to 1 inhalation BID. At subsequent
clinic visits (visit 4 [week 16] onward), patients who
continued to meet the criteria were instructed to further
reduce the Turbuhaler dose to 1 inhalation/d in the
evening; this was in accordance with the international
data sheet for the FOR/BUD Turbuhaler 11 but at vari-
ance with the approach recommended in previous stud-
ies of AMD. 8-1° If the criteria were not met at later vis-
its, patients were told to revert to 1 inhalation BID.
Patients continued to take 1 inhalation BID via Diskus
(active drug or placebo) throughout the 52-week double-
blind treatment period.
Patients received oral and written instructions regard-
ing the AMD self-management plan and the actions to
be taken in the event of worsening asthma between clin-
ic visits, defined as 2 consecutive days with >3 occasions
of rescue medication use during the day, nighttime
awakenings due to asthma, or a morning peak expirato-
ry flow (PEF) <85% of the mean of the last 7 days of
weeks 1 through 4 (Table I). The value corresponding to
85% of the mean PEF from the last 7 days of weeks 1
through 4 was recorded on the front of the diary card to
alert the patient to the need to increase the number of
inhalations from the Turbuhaler device if the morning
PEF fell below this value. Whenever asthma control
became insufficient, as defined in Table I, patients were
to adjust the Turbuhaler dose to 4 inhalations BID for 7
or 14 days, without the need to consult their physician.
After 7 or 14 days, a step-down to 1 inhalation BID
was carried out in consultation with the investigator by
clinic visit or telephone. Any patient whose symptoms
persisted after stepping up the Turbuhaler dose to 4 in-
halations BID and who did not meet the criteria for a
step-down of the Turbuhaler dose after 14 days was to
contact the investigator, who instructed the patient to
step down the Turbuhaler dose to 1 inhalation BID and
implemented a short course of oral corticosteroids.
Follow-up Period
All patients were followed for 2 weeks after the last
clinic visit or the withdrawal visit in the case of pre-
mature discontinuation. Follow-up was conducted
either by clinic visit or telephone.
Randomization and Blinding
A centralized randomization schedule was generat-
ed using in-house software. Patients were registered
and randomized to treatment through an automated
system for central drug allocation that employed a tele-
phone interactive voice response system (IVRS). As a
patient was enrolled, the study site contacted the IVRS
to register the patient with the system. At the random-
ization stage, the study site again dialed into the IVRS,
which assigned the patient a randomization number.
The IVRS was also used to allocate treatment packs to
the patients and to record information on their status
in the study. To maintain double-dummy blinding,
patients received both Diskus and Turbuhaler devices
to use throughout the treatment period.
Efficacy Measures
Each morning, patients recorded their asthma
symptom score for the previous 24 hours, the number
of nighttime awakenings due to asthma, the number
of occasions (irrespective of the number of inhalations
taken) of salbutamol use (rescue only, not prophylactic
use) during the previous 24 hours, the number of Turbu-
haler inhalations taken in the previous 24 hours (weeks
5-52), and PEE
Daily asthma symptom scores were recorded on a
standard 6-point rating scale (0 = no symptoms, 1 =
symptoms for 1 short period, 2 = symptoms for ___2
short periods, 3 = symptoms for most of the 24-hour
period that did not affect normal activities, 4 = symp-
toms for most of the 24-hour period that did affect
normal activities, 5 = symptoms so severe that normal
activities could not be performed). The definition of
"short" in this context was subjective. A symptom-
free day (the primary end point) was defined as a 24-
hour period with a symptom score of 0.
PEF was measured using a mini-Wright peak flow
meter (Clement Clark, Harlow, United Kingdom)
before administration of any study or rescue medica-
tion. The investigator instructed all patients in the cor-
rect use of the peak flow meter and checked their abil-
ity to measure PEF correctly at each clinic visit. A
patient-instruction leaflet was provided with each
peak flow meter. Patients were asked to record only
the highest of 3 recordings. FEV 1 was assessed by
spirometry at each clinic visit.
A well-controlled asthma week was defined as a
week with no nighttime awakenings due to asthma, no
exacerbations, no ED visits, and no treatment-related
adverse events, plus >2 of the following: an asthma
symptom score >1 for <2 days; <2 days of rescue medi-
cation use (maximum of 4 occasions/wk); and a morn-
ing PEF >80% of the predicted normal value on each
day. Predicted normal values for FEV 1 and PEF were
calculated using standard reference values. 15
An asthma exacerbation was defined as a worsen-
ing of asthma requiring hospital treatment or treat-
ment with oral corticosteroids, either in the opinion of
the investigator or based on a morning PEF <70% of
the mean of the last 7 days in weeks 1 through 4 for
>2 consecutive days. A value representing 70% of the
mean PEF recorded during the last 7 days of weeks 1
through 4 was recorded on the front of the diary card,
and patients were instructed to contact the investiga-
tor as soon as possible if their morning PEF fell below
this value. Individual courses of oral corticosteroids
were classified as representing separate exacerbations
only if they were administered >1 week apart.
Compliance
Compliance with both devices was calculated based
on the number of doses remaining in each inhaler device
dispensed and returned by patients. Patients were
deemed compliant if the actual number of doses taken
was _+30% of the expected number of inhalations.
Tolerability
Tolerability was assessed by the recording of ad-
verse events at clinic visits. Spontaneously reported
and/or observed adverse events and patients' respons-
es to a standard open-ended question were recorded
by the blinded investigator, who assessed the relation-
ship between treatment and each adverse event.
Statistical Methods
The study was designed to detect a difference in
the primary end point, the percentage of symptom-free
days, between stable dosing of SAL/FP and AMD of
FOR/BUD. It was anticipated that a sample size of
347 patients per group would be sufficient to detect
such a difference based on a Mann-Whitney U test
with a 5% 2-sided significance level and 90% power.
Randomized patients who took >1 dose of study
medication and had _1 postrandomization diary
assessment were included in the intent-to-treat (ITT)
397
population, the primary population for all efficacy country grouping, and age. The model's goodness of fit
analyses. A secondary analysis of the primary end to the data was confirmed by a QQ plot 17 of the stan-
point was performed in the per-protocol (PP) popula- dardized deviance residuals. The mean daily asthma
tion, defined as all patients in the ITT population with symptom score and mean morning PEF for each patient
no major protocol violations that might influence the over the 52-week treatment period were compared
treatment effect. The safety analyses were based on all between treatments using analysis of covariance with
patients who received >1 dose of study drug. adjustment for baseline, sex, age, and country group-
The percentage of symptom-flee days was derived for ing. Exposure to oral corticosteroid treatment (in days)
each patient over the 52 weeks of double-blind treat- was summarized across treatments. In addition to the
ment. This end point was compared between treatments planned summary, oral corticosteroid exposure was
using the van Elteren extension 16 to the Wilcoxon rank later analyzed using the van Elteren extension 16 to the
sum test, stratified by country grouping. The percentage Wilcoxon rank sum test, stratified by country grouping.
of days free of rescue medication, percentage of night- There was no imputation of missing data in the
time awakenings due to asthma, and mean daily rescue analysis of any end point, including missing data
medication use (number of occasions) were analyzed in resulting from patient withdrawals. In the calculation
the same way. These end points were also summarized of percentages, the number of days or weeks with
over weeks 5 through 52. In addition to these planned nonmissing values was used as the denominator.
summaries, the percentage of symptom-free days and
percentage of days free of rescue medication over weeks RES U LTS
5 through 52 were analyzed later. Patient Population
The exacerbation rate over the 52-week treatment Nine hundred five patients were enrolled in the
period was analyzed using a maximum likelihood- study. The first patient was recruited on November
based analysis, assuming the negative binomial distri- 26, 2002, and the last patient completed the study on
bution, with time on treatment as an offset variable. July 28, 2004. The disposition of patients is shown in
The model included adjustment for treatment, sex, Figure 2. After the run-in period, 706 patients were

I Recruited
(N = 905) Withdrew before randomization (n - 199):
4 Adverse event
15 Consent withdrawn
8 Lost to follow-up
3 Protocol violation

Excluded from I-iq- due to I

~ - Randomized
(n = 706)
167 Did not meet entry criteria
2 Other
no postrandomizadon '1 /
efficacy data and/or tool< no 41
study reed cat on (n = 18)
+
Iqq- population
(n = 688)

SAL/FP FOR/BUD J
(n = 344) (n = 344)

Withdrew (n - 80): Withdrew (n = 93):

49 Did not meet visit 3 step-down criteria 52 Did not meet visit 3 step-down criteria
6 Adverse event 11 Adverse event
11 Consent withdrawn 6 Consent withdrawn
4 Lost to follow-up 2 Lost to follow-up
4 Protocol violation ~ 7 Protocol violation
0 Did not meet eligibility criteria
3 Lack of efficacy
3 Other
[ 2 Did not meet eligibility criteria
4 Lack of'efficacy
9 Other

Completed Completed ]
(n = 264) (n ~ 251)

Figure 2. Patient disposition, intent-to-treat (I-I-F) p o p u l a t i o n . SAL = salmeterol; FP = fluticasone propionate;

FOR = formoterol; BUD = budesonide.
_.... ...........................................................................................................................................................................................................................
randomized to double-blind treatment; 688 were in-
cluded in the ITT population (344 patients in each treat-
ment group). Eighteen patients were excluded from
the ITT population because of the lack of >_1 post-
randomization diary-card assessment and/or failure
to take study medication on any occasion. Data from
702 patients were included in the safety analysis. The
4 randomized patients excluded from the safety analy-
sis had no evidence of having taken any study medica-
tion, as indicated by the investigator on the case-
report form.
Three hundred forty-nine patients had >_1 deviation
from the protocol: 180 in the SAL/FP group and 169
in the FOWBUD group. The most common protocol
deviations (34%-35% in each treatment group) were
related to compliance, with patients taking an incor-
rect number of inhalations from either the Turbuhaler
and/or Diskus (most commonly from the Turbuhaler).
The proportion of patients who were compliant with
each device was similar in the 2 treatment arms: with
the Diskus, 80.8% of the SAL/FP group and 82.6% of
the FOR/BUD group were compliant; with the Turbu-
haler, 66.9% of the SAL/FP group and 68.3% of the
FOR/BUD group were compliant. After 4 weeks of
stable dosing, 295 of 344 (85.8%) patients in the
SAL/FP group and 286 of 344 (83.1%) patients in
the FOR/BUD group continued beyond visit 3 into the
5 through 52-week treatment period.
Overall, the 2 treatment groups were well balanced
with regard to demographic and baseline characteris-
tics (Table II). Enrolled patients had suboptimal asth-
ma control with their existing therapy; 42.2%
(2901688) of the ITT population were receiving LABA
therapy at study entry. Patients in both treatment
groups had similar percentages of symptom-free days
during the last 7 days of the run-in period (mean, 5%;
median, 0%). The results for the diary-card parame-
ters during treatment, which are summarized in Table
III, indicate that both groups had improvements from
baseline.
Symptom-Free Days
The percentage of symptom-free days over the 52-
week treatment period was significantly higher with

Table II. Baseline characteristics of'the intent-to-treat population.

SAL/FP FOR/BUD
Characteristic (n = 344) (n = 344)

Sex, no. (%)

Female 204 (59) 216 (63)
Male 140 (41) 128 (37)
Age, mean (SD), y 46 (14) 44 (14)
Body weight, mean (SD), kg 76 (16) 76 (17)
Asthma duration >10 years, no. (%) 197 (57) 200 (58)
ICS dose at entry, mean (SD), lag 509 (275) 515 (303)
Concurrent long-acting beta2-agonist use at entry, no. (%) 150 (44) 140 (41)
FEV~, mean (SD), L 2.53 (0.80) 2.52 (0.70)
FEV1, mean (SD), % predicted 82 (21) 81 (13)
Median % symptom-free days (25th, 75th percentile) ~ 0 (0, 0) 0 (0, 0)
Daily asthma symptom score, mean (SD) ~ 1.9 (0.6) 1.9 (0.5)
Median % days free of rescue medication (25th, 75th percentile) ~ 14.3 (0, 57.1 ) 14.3 (0, 71.4)
Median rescue medication use, no. ofoccasions/d (25th, 75th percentile) ~t 1.0 (0.4, 2.0) 1.1 (0.4, 2.0)
Median % nighttime awakenings due to asthma (25th, 75th percentile)* 14.3 (0, 50.0) 0 (0, 42.9)
Morning PEF, mean (SD), L/min* 357 (103) 362 (100)

SAL = salmeterol; FP = fluticasone propionate; FOR = formoterol; BUD = budesonide; ICS = inhaled corticosteroid; FEV1
forced expiratory volume in 1 second; PEF = peak expiratory flow.
*Measured over the 7 days before randomization. PEF was measured before study drug was taken.
Only occasions of-rescue medication use were recorded, not the actual number ofinhalations taken.

399
 Table III. Summary of diary-card parameters over the treatment period (weeks 1-52, unless otherwise specified).*

Parameter SAL/FP FOR/BUD

Intent-to-treat population
% Symptom-free days, median
Weeks 1-52 (344, 344) 58.8 (1.5, 90.6) 52.1 (0, 83.5)
Weeks 5-52 (305,299) 73.8 (2.7, 95.5) 64.9 (0.9, 89.5)
% Days free of rescue medication, median
Weeks 1-52 (344, 344) 90.5 (66.5, 98.3) 85.6 (58.5, 96.7)
Weeks 5-52 (305,299) 94.5 (80.6, 99.4) 90.7 (69.3, 98.5)
Daily occasions o f rescue medication use, median (344, 344) 0.11 (0.02, 0.43) 0.18 (0.04, 0.59)
Daily symptom score, mean (344, 344) 0.8(0.8) 0.9(0.8)
% Nighttime awakenings due to asthma, median
Weeks 1-52 (343,344) 1.1 (0, 6.3) 1.4 (0, 6.3)
Weeks 5-52 (304, 299) 0.3 (0, 2.8) 0.3 (0, 3.6)
Morning PEF, mean (343,344) 395 (104) 390 (100)
% Well-controlled asthma weeks, median (344, 344) 82.7 (40.0, 96.2) 71.2 (30.4, 92.3)
Per-protocol population
% Symptom-free days, median (164, 175) 68.5 (2.4, 94.8) 51.8 (0, 84.3)

SAL = salmeterol; FP = fluticasone propionate; FOR = formoterol; BUD = budesonide; PEF = peak expiratory flow.
*Numbers in parentheses after each parameter in column 1 are the numbers of patients from each treatment group who con-
tributed data to the relevant assessment. The patient numbers for weeks 5 through 52 include those with week-5 diary-card
data (due to a late visit 3) who did not continue beyond visit 3. For median values, figures in parentheses are the 25th and
75th percentiles; for the mean values, the figures in parentheses are the SDs.

stable dosing of SAL/FP compared with AMD of

FOR/BUD (median, 58.8% vs 52.1%, respectively; P = 100-

I
0.034). Figure 3 uses a box-and-whisker plot to 90-
show the median values and summarize the variabil- 80-
ity of the data in both treatment groups. The frequen-
70-
cy distribution of symptom-free days is shown in
Figure 4. 60-
Over weeks 1 through 4, when both groups were ~ 0-
receiving stable dosing, the percentage of symptom-
40-
free days was similar between the 2 treatments (medi-
an, 25.0% in both groups). However, during weeks 5 & 3o-
through 52, the percentage of symptom-free days was N 20-
significantly higher in the SAL/FP group compared ]0-
with the BUD/FOR group (median, 73.8% vs 64.9%,
O-
respectively; P = 0.030) (Figure 5). This difference over I I I

the 52-week treatment period would equate with an Salmeterol/ Formoterol/Budesonide

average of 24 additional symptom-flee days per year
in the stable-dose SAL/FP group compared with the
AMD FOR/BUD group, whereas the difference over
weeks 5 to 52 would equate with an average of 32
additional symptom-free days per year.
Analysis of the PP population over the 52-week
treatment period confirmed the statistically signifi-
cant difference between stable dosing of SAL/FP and
Huticasone Propionate
Figure 3. Box-and-whisker plots o f the percentage of
symptom-free days over weeks 1 through
52. The line across the box is the median;
the top and bottom portions of the box
represent the 2Sth and 7Sth percentiles,
respectively; and the whiskers indicate the
maximum and minimum values.
 • Salmeterol/fluticasone propionate

sot
40
[] Formoterol/budesonide

E
30
t~m
q..
o

I I ] I
0 0-G25 >25-G50 >50-G75 >75-100

% of Symptom-Free Days

Figure 4. F r e q u e n c y d i s t r i b u t i o n o f t h e p e r c e n t a g e o f s y m p t o m - f r e e d a y s o v e r w e e k s 1 t h r o u g h 52.

• Salmeterol/fluticasone propionate
[] Formoterol/budesonide
80-
73.84

70-
64.9

58.8*
60-
I 52.1

m 50-

E 40-

E
~_ 30-
0 25.0 25.0

20-

10-

I- I i
(n = 344) (n = 344) (n = 344) (n = 343) (n = 305) (n = 299)

Weeks 1-52 Weeks 1-4 Weeks 5-52

Figure 5. P e r c e n t a g e o f s y m p t o m 4 r e e days over weeks 1 through 52, 1 t h r o u g h 4, a n d 5 t h r o u g h 52. V a l u e s are

medians. *P = 0.034, tp = 0.030.
AMD of FOR/BUD (median, 68.5% vs 51.8%, re-
spectively; P = 0.009), indicating that the difference
between the 2 treatments was maintained, even in a
population that had adhered closely to the AMD
treatment plan for FOR/BUD.
Asthma Exacerbations
Thirty-nine (11.3%) patients who received stable
dosing of SAL/FP experienced an asthma exacerba-
tion, compared with 61 (17.7%) patients who
received AMD of FOR/BUD (Figure 6). Overall, there

A
70-

60-

O
50-

40-

30-
h
r"
~o
20-
O..

10-

0 -- I I 1 I I I I 1 I I I
0" 5 10 15 20 25 30 35 40 45 50 55

Time on SAL/FP (wk)

70-

60-

.O
50-

40-
,5
30-
c
_ ----_
20- m
o..

10-

0 I I I l I I I I I I I
0 5 10 15 20 25 30 35 40 45 50 55
Time on FOR/BUD (wk)

Figure 6. Incidence o f exacerbations over time in individual patients receiving (A) stable dosing o f salmeterol/
fluticasone propionate 50/250 pg (SAL/FP) BID and (B) adjustable maintenance dosing o f f o r m o t e r o l /
budesonide 6/200 pg (FOR/BUD). Each line represents 1 exacerbation in an individual patient. The
beginning and end of each line correspond to the beginning and end o f the exacerbation, as designated
by the investigator. More than 1 exacerbation in the same patient is indicated by separate lines in the
same row.
were 48% fewer exacerbations with stable dosing of Asthma Control
SAL/FP than with AMD of FOR/BUD (50 vs 96 exac- The median value for the percentage of well-
erbations, respectively). Two exacerbations resulted in controlled asthma weeks over the 52-week treatment
hospitalization in the SAL/FP group and 3 resulted in period was 82.7% with stable dosing of SAL/FP and
hospitalization in the FOR/BUD group. Three other 71.2% with AMD of FOR/BUD (descriptive statistics).
exacerbations required an ED visit in the FOWBUD The number of patients with sustained asthma control
group. Forty-eight of the exacerbations in the SAL/FP (ie, >7 of 8 weeks with well-controlled asthma) was con-
group were judged by the investigator to require oral sistently higher in the SAL/FP group compared with
corticosteroid treatment, compared with 90 in the the FOR/BUD group over the 3 treatment periods
FOR/BUD group. The adjusted annual mean exacer- assessed (weeks 5-12, 59.3% vs 55.9%, respectively;
bation rate was 0.18 for SAL/FP and 0.33 for FOR/ weeks 17-24, 68.2% vs 55.0%; weeks 45-52, 70.6% vs
BUD (adjusted treatment rate ratio, 0.53; 95% CI, 62.7%).
0.34-0.85; P = 0.008), representing a 47% reduction
in rate. Adjustment of the Turbuhaler Dose
The distribution of exposure to oral corticosteroids In total, 295 of 344 (85.8%) patients in the group
(in days) was significantly different between treat- receiving stable dosing of SAL/FP and 286 of 344
ment groups (P = 0.026). The total number of days of (83.1%) patients in the group receiving AMD of
exposure was 46% lower (301 days) in the SAL/FP FOR/BUD entered the period from weeks 5 through
group compared with the FOR/BUD group (559 52 and stepped down their Turbuhaler dose (active
days). drug or placebo) to 1 inhalation BID. Between weeks
16 and 52, 253 of 295 (85.8%) patients in the SAL/FP
Other Diary-Card Parameters group and 235 of 286 (82.2%) patients in the FOR/
The percentage of days free of rescue medication BUD group stepped down to 1 inhalation/d at some
over weeks 1 through 52 was significantly higher in time.
the SAL/FP group compared with the FOR/BUD During weeks 5 through 52, 71 of 295 (24.1%)
group (median, 90.5% vs 85.6%, respectively; P = patients in the SAL/FP group and 119 of 286 (41.6%)
0.008). Over weeks 5 through 52, the median value patients in the FOR/BUD group adjusted the Turbu-
for the percentage of days free of rescue medication haler dose to 4 inhalations BID at least once. The
was 94.5% in the SAL/FP group and 90.7% in the mean (SD) time during which patients used 4 inhala-
FOR/BUD group (P = 0.008). Daily rescue medica- tions BID was 5.1 (13.6) days in the SAL/FP group
tion use was significantly lower in the SAL/FP group and 10.9 (18.6) days in the FOR/BUD group.
compared with the FOR/BUD group (median, 0.1 vs Over the 52-week treatment period, the mean (SD)
0.2 occasion/d; P = 0.006) over the 52-week treat- daily ICS exposure in the SAL/FP group was 463
ment period. (81) I~g FP; in the FOR/BUD group, the mean daily ICS
Mean and adjusted mean daily symptom scores exposure was 480 (238) gg BUD. Diary-card data in-
over the 52-week treatment period were 0.8 for dicated that during weeks 5 through 52, patients in the
SAL/FP and 0.9 for FOWBUD (adjusted mean differ- FOR/BUD group used a mean of 1.8 (0.6) inhalations/d
ence, 0.1; 95% CI, 0.0-0.2; P = NS). The median per- (equivalent to BUD 360 pg/d) from the Turbuhaler.
centage of nighttime awakenings due to asthma over
the 52-week treatment period was 1.1% in the Adverse Events
SAL/FP group and 1.4% in the FOR/BUD group (P = The incidence of adverse events was similar in both
NS). During weeks 5 through 52, the median number treatment groups, with a low incidence (<3%) of phar-
of nighttime awakenings due to asthma was 0.3% in macologically predictable adverse events. One hundred
both treatment groups. sixty-nine of 348 (48.6%) patients in the SAL/FP group
Mean morning PEF over the 52-week treatment reported a total of 429 events, and 185 of 354 (52.3%)
period was significantly higher in the SAL/FP group patients in the FOR/BUD group reported 537 events.
compared with the FOR/BUD group (adjusted mean, Drug-related adverse events occurred in 22 (6.3%)
400.1 vs 390.6 L/rain, respectively; adjusted mean dif- patients in the SAL/FP group and 21 (5.9%) patients in
ference, 9.5 L/min; 95% CI, 2.7-16.3; P = 0.006). the FOR/BUD group.

i !i iai0 ii iiiiiiiii iiiiiiiiiiiiii

~ i i i i ~ i i ~ ! ~ i i ~
iiiii iiiiiiiiiiiiiiiiii!iiiiiiiiii
!/~ •g•••G•G•••••••••:•••~:•<••7•? ~
,iiiiiiiii
i!iiiiiii
il iiiiiiiiiiiiiii:iiiiiiiiiii!iiiiiiiii
iiiiiiiiii! i iiiili!ili!;iii;iiii
!!i i iiiiii,iiiii!¸!ii!iiiii i i :
iiiiii!!iJil!i!iiii,iiiiiiiiiiiiii ......
,~:i~:~i~;~i .....
40a
 Twenty-three serious adverse events occurred in
18 patients (9 in each group [2.6 SAL/FP, 2.5% FOR/
BUD]). Two serious adverse events leading to an asth-
ma exacerbation resulted in hospitalization in the
SAL/FP group (1 case of pneumonia/lung infiltration,
1 case of anaphylactic reaction). In the FOR/BUD
group, 3 patients were hospitalized in association with
asthma exacerbations.
DISCUSSION
This study is the first 1-year, double-blind, double-
dummy study to compare the efficacy of stable dosing
of SAL/FP 50/250 lag BID (mean [SD] daily ICS expo-
sure, 463 [81] pg FP) with AMD of FOR/BUD 6/200 lag
(mean daily ICS exposure, 480 [238] lag BUD) in the
management of adults with persistent asthma. In this
population, stable dosing of SAL/FP produced signifi-
cantly greater improvements in the percentage of
symptom-flee days (P = 0.034), percentage of days
free of rescue medication (P = 0.008), and morning
PEF (P = 0.006). More important, despite the fact that
patients receiving AMD of FOR/BUD 6/200 lag could
increase the number of inhalations in response to emerg-
ing symptoms, stable dosing of SAL/FP 50/250 pg BID
almost halved the rate of exacerbations requiring either
oral corticosteroids or ED visits/hospitalizations.
Although significantly higher percentages of
symptom-free days and days free of rescue medication
were achieved in the stable-dose SAL/FP group, the
overall percentages of symptom-free and rescue-free
days in the AMD FOR/BUD group were also high.
Even though the AMD approach in the FOR/BUD
group led to a reduction in the amount of maintenance
treatment used, improvements from baseline were seen
in all diary-card parameters, suggesting that when
taken regularly, low amounts of maintenance treatment
can be effective in improving symptoms and lung func-
tion, and in decreasing rescue medication use. However,
a higher daily amount of maintenance treatment may
afford greater protection against the triggers that can
result in the type of exacerbations requiring interven-
tion with oral corticosteroids or hospital care, a finding
consistent with the results of an earlier study) 8
On the initiation of effective maintenance treat-
ment, improvements in some parameters can occur
rapidly; however, changes in bronchial hyperrespon-
siveness take many months of treatment, 19 suggesting
that short-term improvements in symptoms are not
necessarily temporally concordant with changes in air-
way inflammation. 2° The same is likely to be true of
airway remodeling, which probably occurs over a
longer period. 21 It has been shown that increases in air-
way inflammation can occur in the absence of worsen-
ing symptoms, 22 indicating that changes in airway
inflammation may pass unnoticed by the patient
("silent" inflammation). The findings of 3 studies that
have included measures of airway inflammation as a
guide to treatment indicate that approaches taking air-
way inflammation into account are more effective than
plans that are predominantly symptom based, 23-26 sug-
gesting that the long-term control of airway inflamma-
tion should be the focus of treatment, rather than
treatment minimization to the lowest effective amount
that controls symptoms alone. However, because
quantitative measures of airway hyperresponsiveness
using either challenge with agents such as metha-
choline or cell counts from induced sputum are not
routinely performed in the general practice setting,
alternative strategies are required to ensure that the
amount of maintenance treatment is appropriately
adjusted to the level of disease.
A study of ICS treatment found that a combination
of education, regular follow-up, and an action plan
were effective in improving asthma control and air-
way hyperresponsiveness, even though adherence to
the symptom-based management plan was only 52%. 27
The improvements were accompanied by an increase
in the mean daily amount of ICS used and a decrease
in the proportion of patients using beta2-agonists.
Two recent studies that examined doubling the
amount of ICS at the first sign of an exacerbation
failed to show a benefit with this strategy, 2s,29 suggest-
ing that once control of airway inflammation has been
lost, it is not readily or rapidly regained.
Combination products containing SAL/FP and
FOR/BUD have been directly compared in 2 previous
studies) °,~4 However, the present study is the first
reported trial to use a double-blind, double-dummy de-
sign to compare stable dosing of SAL/FP with AMD of
FOR/BUD. Although the use of a double-blind, double-
dummy design increased the complexity of the study,
in that all patients were required to use 2 types of
inhaler and follow both treatment approaches, the
blinding minimized the influence of the bias associated
with knowing which treatment was active. This was
particularly important in a study in which patients and
investigators made decisions about treatment adjust-
ment. The percentage of patients who were compliant
with each device was similar between treatment arms,
suggesting that patients were not aware of which
device contained the active treatment and which the
placebo. The present study was also of longer duration
than the other 2 studies (1 year, compared with 614 and
71° months). A 52-week treatment period was chosen
to provide more opportunities for patients to use AMD
and a longer period over which to evaluate the effec-
tiveness of this treatment approach.
In the present study, patients could enter weeks 5
through 52 of treatment only if they met the stability cri-
teria for reducing the number of inhalations from the
Turbuhaler device and were then stepped down from
2 inhalations BID to 1 inhalation BID. If the patient met
the stability criteria for a step-down at subsequent clin-
ic visits, the number of inhalations from the Turbuhaler
could be further reduced to 1 inhalation in the evening.
By ensuring that patients' asthma was stable and then
stepping down the Turbuhaler dose, this study tested
the concept of using the minimum effective dose of
FOR/BUD. This approach led to a reduction in the
amount of maintenance treatment used in the FOR/
BUD arm; however, unlike patients in the stable-dose
SAL/FP group, patients in the AMD FOR/BUD group
increased their maintenance treatment for up to 2 weeks
at the first signs of asthma instability. Nonetheless,
patients in the FOR/BUD group had more asthma exac-
erbations than patients in the SAL/FP group, suggesting
that a higher amount of daily maintenance treatment
may be needed overall to reduce the risk of losing
asthma control in patients with persistent asthma.
Instructing patients to increase the number of inhala-
tions from the Turbuhaler from 1 or 2 inhalations/d to
8 inhalations/d did not always abort the exacerbation,
supporting the view that once control is lost, even
prompt intervention with high doses of maintenance
treatment may not always be sufficient to regain it.
These results also suggest that as in other studies using
AMD of FOR/BUD, 8-1° there is a minimum daily
amount of maintenance therapy necessary to prevent
exacerbations in adults with persistent asthma.
Rather than minimizing the amount of asthma
treatment by responding to daily changes in symp-
toms, an alternative approach may be to attempt to
eliminate disease variability as much as possible by
maintaining a level of treatment that keeps patients
symptom free and exacerbation free for as long as
possible. In addition, the latter strategy may better
address the underlying airway inflammation.
CONCLUSIONS
In this study in adult patients with persistent asthma,
stable dosing of SAL/FP 50/250 tag BID was more
effective than AMD of FOR/BUD 6/200 pg. Use of the
AMD strategy resulted in a reduction in the amount
of maintenance treatment used in the AMD arm com-
pared with the stable-dose arm. The results suggest
that in this patient population, a minimum daily
amount of maintenance treatment was necessary and
that if this dose was not maintained, an increase in
exacerbations was likely.
ACKNOWLEDGM ENTS
This study was funded by GlaxoSmithKline Research
& Development Limited, Greenford, United King-
dom. The authors ,thank Antonio Belgrave, Sally
Edmundson, Jo Barr, Jenny Huang, and Karen Hardy
of GlaxoSmithKline for their support of the study and
the CONCEPT clinical investigators who recruited
and cared for the patients in the study.
Dr. FitzGerald is a recipient of a Canadian institute
for Health Research BC Lung Investigator Award and
also a Michael Smith Foundation for Health Research
Distinguished Scholar Award.
REFERENCES
1. Shrewsbury S, Pyke S, Britton M. Meta-analysis of in-
creased dose of inhaled steroid or addition of salmeterol
in symptomatic asthma (MIASMA). BMJ. 2000;320:1368
1373.
2. StoloffS, Poinsett-HoJmes K, Dorinsky PM. Combination
therapy with inhaled long-acting J32-agonists and inhaled
corticosteroids: A paradigm shift in asthma management.
Pharmacotherapy. 2002;22:212-226.
3. Garcia-Marcos L, Schuster A, Cobos Barroso N. Inhaled
corticosteroid plus long-acting J~2-agonists as a com-
bined therapy in asthma. Expert Opin Pharrnacother. 2003;
4:23-39.
4. Ram FSF, Cotes CJ, Ducharme FM. Long-acting beta2-
agonists versus anti-leukotrienes as add-on therapy to
inhaled corticosteroids for chronic asthma. CochraneData-
fiase of Systematic Reviews. 2005; Issue 1. Availableat: http://
www.cochrane.org/coch rane/revabstr/ab003137.htm.
AccessedJanuary 23, 2005.
5. British Thoracic Society; Scottish Intercollegiate
Guidelines Network. British guideline on the management
of asthma. Thorax. 2003;58(SuppI I):il-i94.
6. National Asthma Education and Prevention Program.
Expert Panel Report: Guidelines [or the Diagnosis and Man-
agement ofiAsthma--Update on Selected Topics2002. NIH Pub-
lication 02-5075. Available at: http://www.nhlbi.nih.
 gov/guidelines/asth ma/execsumm.
pd£ Accessed November 7, 2004.
7. Global Initiative for Asthma. Global
Strategy for Asthma Management and
Prevention. NHLBI/WHO Workshop
Report. Bethesda, Md: National
Institutes of Health; 2002. NIH
Publication 02-3659.
8. St~illberg B, Olsson P, Jorgensen LA,
et al. Budesonide/formoterol adjust-
able maintenance dosing reduces
asthma exacerbations versus fixed
dosing. Int Y Clin Pract. 2003;57:
656-661.
9. FitzGeraldJM, Sears MR, Boulet LP,
et al, for the Canadian Investigators.
Adjustable maintenance dosing
with budesonide/formoterol reduces
asthma exacerbations compared
with traditional fixed dosing: A five-
month multicentre Canadian study.
Can RespirJ. 2003;10:427-434.
10. Aalbers R, Backer V, Kava T-I-, et al.
Adjustable maintenance dosing
with budesonide/formoterol com-
pared with fixed-dose salmeterol/
fluticasone in moderate to severe
asthma. Curt Med Res Opin. 2004;
20:225-240.
11. Symbicort prescribing information.
Available at: http://hcp.symbicort.com/
article/11435.aspx. Accessed Octo-
ber 11, 2004.
12. McGavin JK, Goa KL, Jarvis B.
Inhaled budesonide/formoterol com-
bination. Drugs. 2001 ;61:71-78.
13. Marl<ham A, Jarvis B. Inhaled
sal meterol/fluticasone propionate
combination: A review of its use in
persistent asthma. Drugs. 2000;60:
1207-1233.
14. Dahl R, Chuchalin A, Lindberg A, et
al. EXCEL: Regular maintenance ther-
apy with salmeterol/fluticasone pro-
pionate combination (SFC) reduces
exacerbations more effectively than
with formoterol/budesonide combi-
nation (FBC). Eur RespirJ. 2004;
24(Suppl 48):309S. Abstract.
15. Quanjer PH, Tammeling GJ, CotesJE,
et al. Lung volumes and forced ven-
tilatory flows. Report Working Party
Standardization of Lung Function
Tests, European Community for Steel
and Coal. Official statement of the
European Respiratory Society. Eur
RespirJ Suppl. 1993;16:5-40.
16, Lehman EL. Nonparametrics: Statistical
Methods Basedon Ranks. San Francisco,
Calif." Holden-Day Inc; 1975.
17. Atkinson AC. Two graphical displays
for outlying and influential observa-
tions in regression. Biometrika. 1981 ;
68:13-20.
18. PauwelsRA, L6fdahl CG, Postma DS,
et al, for the Formoterol and Cord-
costeroids Establishing Therapy
(FACET) International Study Group.
Effect of inhaled formoterol and
budesonide on exacerbations of asth-
ma [published correction appears in
N Engld Med. 1998;338:139]. N EnglJ
Med. 1997;337:1405-1411.
19. Woolcock AJ. What are the impor-
tant questions in the treatment of
asthma? Clin Exp Allergy Rev. 2001 ;1 :
62 64.
20. Reddel HI<, Jenkins CR, Marks GB,
et al. Optimal asthma control, start-
ing with high doses of budesonide
[published correction appears in Eur
Respird. 2000;16:579]. Eur RespirJ.
2000;16:226-235.
21. Ward C, Pals M, Bish R, et al.
Airway inflammation, basement
membrane thickening and bronchial
hyperresponsiveness in asthma.
Thorax. 2002;57:309-316.
22. de KluijverJ, Evertse CE, Schrumpf
JA, et al. Asymptomatic worsening
of airway inflammation during low-
dose allergen exposure in asthma:
Protection by inhaled steroids. Amy
Respir Crit Care Med. 2002;166:294-
300.
Address c o r r e s p o n d e n c e to: J. M a r k FitzGerald, F R C P C , Director, Center
f o r Clinical E p i d e m i o l o g y a n d E v a l u a t i o n , R e s p i r a t o r y Division,
Vancouver General Hospital, 828 West 10th Avenue, Vancouver, BC V 5 7
3J5, C a n a d a . E-maih markf@interchange.ubc.ca
23. SontJK, Willems LN, Bel EH, et al,
for the AMPUL Study Group.
Clinical control and histopatho[ogic
outcome of asthma when using air-
way hyperresponsiveness as an addi-
tional guide to Iong~term treatment.
Am J Respir Crit Care Med. 1999;159:
1043-1051.
24. Green RH, Brightling CE, McKenna
S, et al. Asthma exacerbations and
sputum eosinophil counts: A ran-
domised controlled trial. Lancet.
2002;360:1715-1721.
25. Jayaram L, Hussack P, Efthimiadis
A, et al. The LOMA study: Effect on
asthma exacerbations in the first
year. Amy Respir Crit Care Med. 2003;
167:A976. Abstract.
26. Pizzichini MM,Jayaram L, Pizzichini
E, et al. Does sputum cell counts
alter asthma exacerbations? The
LOMA study. Am J Respir Crit Care
Med. 2004;169:A366. Abstract.
27. Turner MO, Taylor D, Bennett R,
FitzGerald JM. A randomized trial
comparing peak expiratory flow and
symptom self-management plans
for patients with asthma attending a
primary care clinic. Am J Respir Crit
Care Med. 1998;157:540-546.
28. FitzGerald JM, Becket A, Sears MR,
et at, for the Canadian Asthma
Exacerbation Study Group. Doubling
the dose of budesonide versus main-
tenance treatment in asthma exacer-
bations. Thorax. 2004;59:550-556.
29. Harrison TW, Oborne J, Newton S,
Tattersfield AE. Doubling the dose
of inhaled corticosteroid to prevent
asthma exacerbations: Randomised
controlled trial. Lancet. 2004;363:
271-275.