You are on page 1of 11

ARTHRITIS & RHEUMATISM

Vol. 52, No. 4, April 2005, pp 12051215


DOI 10.1002/art.20985
2005, American College of Rheumatology

Evaluation of the Efficacy of Etoricoxib in


Ankylosing Spondylitis

Results of a Fifty-TwoWeek, Randomized, Controlled Study

Desiree van der Heijde,1 Herbert S. B. Baraf,2 Cesar Ramos-Remus,3 Andrei Calin,4
Arthur L. Weaver,5 Michael Schiff,6 Margaret James,7 Jan E. Markind,7 Alise S. Reicin,7
Agustin Melian,7 and Maxime Dougados8

Objective. To assess the efficacy, safety, and tol- tients global assessment of disease activity, and the
erability of etoricoxib, a cyclooxygenase 2 (COX-2) Bath Ankylosing Spondylitis Functional Index.
selective inhibitor, administered continuously over 52 Results. Of the 387 patients randomized to receive
weeks for the treatment of ankylosing spondylitis (AS). treatment, 301 (77.8%) completed part I and 284
Methods. This 2-part, multicenter, double-blind, (75.9%) completed part II. Compared with placebo over
parallel-group, 52-week study evaluated 2 doses of etori- 6 weeks, those receiving 90 mg etoricoxib, 120 mg
coxib (90 and 120 mg) compared with naproxen at 1,000 etoricoxib, and naproxen demonstrated significantly
mg. A 6-week, active-comparator and placebo- (P < 0.001) greater improvement in all primary end
controlled period (part I) was followed by a 46-week points; treatment effects (expressed as the difference in
active-comparatorcontrolled period (part II). The pri- least squares mean change versus placebo) were 2129
mary outcome measures (on 100-mm visual analog mm for spine pain, 1825 mm for disease activity, and
scales) were patients assessment of spine pain, pa- 1115 mm for function. Compared with patients receiv-
ing naproxen, significantly greater improvement in all
1
primary end points was demonstrated in the combined
Desiree van der Heijde, MD: University Hospital, Maas-
tricht, The Netherlands; 2Herbert S. B. Baraf, MD: Center for group receiving either 90 mg etoricoxib or 120 mg
Rheumatology and Bone Research, Wheaton, Maryland; 3Cesar etoricoxib over 6 weeks, in each individual etoricoxib
Ramos-Remus, MD: Centro Medico de Occidente, IMSS, Guadala- treatment group over 6 weeks, and in the combined
jara, Mexico; 4Andrei Calin, MD: Royal National Hospital for Rheu-
matic Diseases, Bath, UK; 5Arthur L. Weaver, MD: Arthritis Center of etoricoxib group over 1 year (all P < 0.05); results for
Nebraska, Lincoln; 6Michael Schiff, MD: Denver Arthritis Clinic, secondary and exploratory end points were generally
Denver, Colorado; 7Margaret James, MS, Jan E. Markind, PharmD, consistent with those from the primary analysis. Among
Alise S. Reicin, MD, Agustin Melian, MD: Merck & Co., Inc.,
Whitehouse Station, New Jersey; 8Maxime Dougados, MD: Rene all groups, there were no significant differences in the
Descartes University, Cochin Hospital, Paris, France. incidence of overall clinical, drug-related, or serious
Dr. van der Heijde has served as a consultant to and has adverse experiences (AEs) and discontinuations due to
received less than $10,000 as a speaker for Merck. Dr. Baraf has
received less than $10,000 as a speaker for Merck. Dr. Ramos-Remus AEs. Safety observations during part II were generally
has received more than $10,000 as a consultant and speaker for Merck. consistent with those in part I.
Dr. Weaver has received more than $10,000 as a consultant and Conclusion. Etoricoxib at doses of 90 mg and 120
speaker for Merck, Amgen, Wyeth, and TAP. Ms James, Ms Markind,
and Drs. Reicin and Melian have stock ownership in Merck. Dr. mg demonstrated superior efficacy compared with pla-
Dougados has received less than $10,000 as a consultant and speaker cebo over 6 weeks, and compared with naproxen over 1
for Merck. year. These study results demonstrate that etoricoxib is
Address correspondence and reprint requests to Desiree van
der Heijde, MD, Professor of Rheumatology, Department of Internal generally safe, well-tolerated, and efficacious for the
Medicine, Division of Rheumatology, University Hospital Maastricht, treatment of AS.
PO Box 5800, 6202 AZ Maastricht, P Debyelaan 25, 6229 HX
Maastricht, The Netherlands. E-mail: dhe@sint.azm.nl.
Submitted for publication July 2, 2004; accepted in revised Ankylosing spondylitis (AS) is a chronic inflam-
form January 11, 2005. matory autoimmune disease that is characterized by
1205
1206 VAN DER HEIJDE ET AL

pain, disability, and loss of mobility (17). Optimal ment of spine pain [on 100-mm visual analog scale (VAS)] and
treatment typically requires high doses of nonsteroidal an increase of 30% [minimum 12 mm] compared with the
pain rating at the screening visit) after the washout period for
antiinflammatory drugs (NSAIDs), which, over the long
prestudy NSAIDs.
term, carry a risk of toxicity in the gastrointestinal (GI) Patients with concurrent rheumatic disease (e.g., sys-
tract, including risk of perforations, ulcers, and bleeds temic lupus erythematosus, gout, Pagets disease of bone) that
(PUBs) (1,8). could confound the evaluation of efficacy were excluded. In
Etoricoxib is a cyclooxygenase 2 (COX-2) selec- addition, patients with acute peripheral articular disease (de-
fined by onset within 4 weeks prior to study of active [painful
tive inhibitor with demonstrated efficacy in numerous or swollen] peripheral arthritis) were also excluded. Patients
painful and/or inflammatory conditions (916). Similar with chronic peripheral arthritis were, however, eligible if spine
to several other COX-2 selective inhibitors, etoricoxib pain was the primary source of pain. Additional exclusion
has demonstrated efficacy and can be considered at least criteria included 1) use of corticosteroid therapy within 1
month prior to the screening visit, 2) use of analgesic medica-
as effective as nonselective NSAIDs, but has a superior
tions within 3 days of study entry and through week 6
GI safety and tolerability profile (1721). Recently, in a (acetaminophen use was permitted prior to study entry), and
12-week efficacy study involving patients with rheuma- 3) use of nonstudy NSAID or selective COX-2 inhibitor, with
toid arthritis (RA), 90 mg etoricoxib was found to be the exception of low-dose aspirin (100 mg once daily), which
superior to 1,000 mg naproxen. This finding suggested was allowed for cardiovascular prophylaxis.
Study design. This double-blind, parallel-group, 52-
the possibility that, in some patients, etoricoxib might week study evaluated patients with AS in 2 parts. Part I
confer greater efficacy than that provided by commonly consisted of a 6-week, active-comparator and placebo-
used NSAIDs in certain inflammatory conditions (13). controlled treatment period to evaluate the safety, tolerability,
The aim of the present study, in its entirety, was and efficacy of etoricoxib at 90 mg and 120 mg once daily for
the treatment of AS. All patients who successfully completed
to assess the efficacy, safety, and tolerability of etori- part I or who discontinued part I (due to lack of efficacy
coxib administered continuously over a 52-week period following at least 2 weeks of treatment during part I) were
for the treatment of AS. A 90 mg dose of etoricoxib (as given the opportunity to continue into part II, which was the
well as a dose of 120 mg) once daily was compared with double-blind, active-comparator, 46-week period.
1,000 mg naproxen (500 mg twice daily), which is a Study drugs. If the patient fulfilled the eligibility
criteria, the patient was randomly allocated to a treatment
commonly used NSAID for the treatment of AS (22). sequence using a computer-generated random-allocation
Prior studies have demonstrated the efficacy and high schedule. Based on the original randomization schedule, pa-
compliance associated with naproxen in the treatment of tients who received placebo during part I were reassigned
AS (2225). The 1,000 mg naproxen dose was selected (1:1:1) to etoricoxib 90 mg daily, etoricoxib 120 mg daily, or
naproxen 500 mg twice daily for part II. Patients who received
because treatment of AS typically requires high doses of etoricoxib or naproxen during the first 6 weeks continued to
NSAIDs (for optimal results) and it is the highest dose receive the same regimen throughout the 52-week study.
recommended for long-term use (26). During part I, patients received 3 bottles of study
medication at randomization and at weeks 2 and 4. Each bottle
contained active medication or matching placebo. Patients also
PATIENTS AND METHODS received acetaminophen (considered a rescue medication
for breakthrough AS pain). During part II, study medication
Patient population. Study subjects were outpatients was dispensed in blister cards, each containing active medica-
who fulfilled the modified New York criteria for AS (27). All tion or matching placebo, for 7 days.
patients were randomized to receive treatment after their Assessment criteria. Clinical assessments were per-
written informed consent was provided. Of the 44 international formed at screening, flare/randomization, and the week 2, 4, 6,
study sites participating in the trial, patients were screened and 8, 16, 26, 34, 43, and 52 visits, and if the patient discontinued
enrolled at 43 sites. For each study site, the protocol and treatment, an assessment was performed at the discontinuation
consent form were approved by an institutional review board visit. The primary end points were patients assessment of
or ethics review committee. Other criteria for inclusion were spine pain (100-mm VAS), patients global assessment of
1) minimum age of 18 years, 2) diagnosis of AS made 6 disease activity (100-mm VAS), and Bath Ankylosing Spondy-
months prior to study start, 3) history of positive therapeutic litis Functional Index (BASFI) (100-mm VAS) (28).
benefit with NSAIDs, 4) routine NSAID intake (defined as use Secondary end points were the following parameters.
of NSAIDs at least 25 of the previous 30 days) and at a Patients global assessment of response to therapy (PGART)
therapeutic dose level for 30 days prior to study enrollment, was performed on a 04-point Likert scale, where 0
5) use of approved nonstudy antirheumatic therapy at a stable excellent, 1 good, 2 fair, 3 poor, and 4 no response.
dose for required time periods (methotrexate for 3 months, Also recorded was the patients assessment of pain at night,
sulfasalazine for 3 months, other disease-modifying antirheu- which was rated on a 03-point scale, as follows: 0 not
matic drugs for 6 months), and 6) satisfaction of flare criteria bothered, 1 bothered a little, 2 bothered a lot, 3
(worsening of AS defined as 40 mm on the patients assess- bothered terribly. The Bath Ankylosing Spondylitis Disease
ETORICOXIB IN ANKYLOSING SPONDYLITIS 1207

Figure 1. Distribution of study patients in part I (randomization to week 6) and part II (week 6 to week
52). patients were eligible to enter part II after completion of part I (6 weeks of treatment) or after
discontinuation due to lack of efficacy after 2 weeks of treatment in part I. P 0.001 versus placebo.
the percentage combined across the 2 etoricoxib dose groups was significantly less than the percentage
in the naproxen group (P 0.01). patients were lost to followup, moved, or withdrew consent, or were
discontinued due to a protocol deviation or for other reasons, or the study site was terminated.

Activity Index is a composite of 6 questions regarding disease well, 1 well, 2 fair, 3 poor, 4 very poor. The
activity, and each response was recorded on a 100-mm VAS. investigators global assessment of response to therapy was
The investigators global assessment of disease activity was also rated on a 04-point Likert scale, as follows: 0 excellent,
performed on a 04-point Likert scale, as follows: 0 very 1 good, 2 fair, 3 poor, 4 no response. Use of
1208 VAN DER HEIJDE ET AL

acetaminophen (paracetamol) at 325 mg for pain relief was the combined etoricoxib treatment group against placebo. A
also assessed. Finally, the proportion of patients discontinuing similar approach was taken for comparisons with naproxen.
due to lack of efficacy (29) was recorded. The secondary end points, including an assessment of the onset
Exploratory end points included results of the modi- of effect during the first 6 days of dosing, were analyzed
fied Schober test, the occiput-to-wall test, chest expansion, and similarly, with the exception of discontinuations due to lack of
duration of morning stiffness (recorded on a 0100-mm VAS, efficacy, which were evaluated using Fishers exact test (32).
with the values between 0 and 100 corresponding to time During part II, the maintenance of treatment effect for
increments from 0 hours to 2 hours, respectively). Additional etoricoxib at 90 mg and 120 mg and naproxen at 1,000 mg over
post hoc analyses of the percentage of patients fulfilling the 52 weeks was examined via longitudinal plots of the mean
ASsessment in Ankylosing Spondylitis (ASAS) International changes from baseline among patients who were maintained
Working Group 20% response criteria (ASAS20) were per- on the same treatment regimen in parts I and II. For the
formed at week 6. The ASAS response criteria are defined as longitudinal plots, missing values were replaced with the last
a 20% (and at least 10-mm) improvement in at least 3 of 4 observed value, and patients who discontinued part I due to
prespecified end points (patients global assessment of disease lack of efficacy and continued into part II were included by
activity, patients assessment of spine pain, BASFI, and dura- carrying forward the last observed value in part I. In addition,
tion of morning stiffness), with no worsening of the fourth the time-weighted average changes from baseline were evalu-
domain to the same extent. The ASAS20 partial remission ated in a manner similar to that described for part I.
criteria (a score of 20 mm in each of the above-mentioned 4 Safety. Differences between treatment groups in the
domains) (30) were also assessed at week 6. proportions of patients with AEs or who exceeded the pre-
Spontaneously reported adverse experiences (AEs) defined limits of change were assessed by 95% confidence
were documented at all postbaseline visits (weeks 252). intervals (95% CIs) using Wilsons score method (32). For
Complete blood cell counts, serum chemistry panel results, prespecified clinical AEs and for predefined limits of change in
and findings on urinalysis were obtained at all visits. In laboratory parameters, each of the active treatments was
addition, vital signs, including blood pressure and heart rate, compared with placebo using Fishers exact test (32).
and weight were recorded at each visit. For part II, the same safety parameters as evaluated in
Sample size. For the primary efficacy hypothesis, it was part I were examined for clinically meaningful differences
predicted that sample sizes of 75 patients per group would between the 90 mg etoricoxib, 120 mg etoricoxib, and 1,000 mg
provide at least 83% power ( 0.05 by 2-tailed test) to detect naproxen treatment groups; no statistical testing was per-
the following differences in the mean (SD) change from formed.
baseline between the etoricoxib and placebo groups: a
between-group difference of 14.6 30 mm in the patients
assessment of spine pain, 15.7 31.3 mm in the patients RESULTS
global assessment of disease activity, and 13.2 22 mm in the
BASFI rating. In a previous study of NSAIDs in the treatment
Patients and study course. Of the 387 patients
of AS, a mean (SD) difference of 17 27 mm in comparison enrolled and randomized in the study, 301 (77.8%)
with the placebo group for the change from baseline in completed the first 6-week period. Of the 81 patients
patients assessment of spine pain was observed (31). who discontinued part I of the study due to lack of
Statistical analysis. Efficacy. The primary efficacy ana- efficacy after completing at least 2 weeks of treatment,
lyses were based on the modified intent-to-treat principle (i.e.,
inclusion of all patients in the analysis population for whom a 77 continued into part II. A total of 374 patients entered
baseline value and at least 1 postbaseline measurement were the final 46-week period of the study; 284 (75.9%)
available) and the per-protocol approach (part I only). completed the 1-year study (Figure 1).
In part I, the primary measure of treatment effect was Patient characteristics at the initiation of the trial
the time-weighted average of all measurements collected over
are summarized in Table 1. There were no clinically
the 6-week treatment period. The time-weighted average
changes from baseline for each efficacy end point were ana- meaningful differences among the treatment groups for
lyzed using an analysis of variance or covariance (ANOVA/ any of the evaluated characteristics. The most common
ANCOVA) model, with treatment and presence/absence of reason for discontinuation throughout the study was lack
chronic peripheral arthritis as the main effects and baseline of efficacy. Time to discontinuation for lack of efficacy
value as the covariate for end points measured at baseline. A
step-down procedure was used to control for multiplicity, as among all treatment groups is presented in Figure 2.
follows. The combined etoricoxib group was compared with During the first 6 weeks, a significantly (P
the placebo group using a t-test generated from the ANCOVA; 0.001) smaller percentage of patients discontinued treat-
if this was significant (i.e., 0.05 by 2-tailed test), then ment due to lack of efficacy in the 90 mg etoricoxib, 120
each individual dose was compared with placebo to assess mg etoricoxib, and naproxen groups compared with
which dose(s) contributed to the observed effect. If the com-
bined result was not significant, no further testing was per- placebo (Figure 1) and in both etoricoxib groups com-
formed. To claim the efficacy hypothesis, each primary end pared with naproxen (P 0.008 for the combined
point must show a significant difference for the comparison of etoricoxib group versus naproxen; P 0.014 for 90 mg
ETORICOXIB IN ANKYLOSING SPONDYLITIS 1209

Table 1. Baseline characteristics of the patients with ankylosing spondylitis at randomization*


Etoricoxib
Naproxen
Placebo 90 mg 120 mg 1,000 mg Total
(n 93) (n 103) (n 92) (n 99) (n 387)
Demographic
Female, % 20.4 26.2 21.7 20.2 22.2
Age, mean SD years 43.7 12.1 43.1 12.1 42.5 12.0 45.0 11.4 43.6 11.9
Age range, years 2371 2074 2078 1874 1878
Clinical
History of iritis, no. (%) 31 (33.3) 39 (37.9) 29 (31.5) 31 (31.3) 130 (33.6)
History of chronic peripheral arthritis, no. (%) 37 (39.8) 41 (39.8) 36 (39.1) 41 (41.4) 155 (40.1)
History of corticosteroid use, no. (%) 30 (32.3) 24 (23.3) 22 (23.9) 22 (22.2) 98 (25.3)
Concomitant DMARD use, no. (%) 18 (19.4) 27 (26.2) 18 (19.6) 23 (23.2) 86 (22.2)
Baseline efficacy values
Patients assessment of spine pain on VAS, 77.22 15.24 77.95 13.94 77.96 14.16 77.20 16.45 77.58 14.92
mean SD
Patients global assessment of disease activity 64.26 20.99 63.19 20.84 64.29 21.60 64.65 22.17 64.08 21.33
on VAS, mean SD
Bath Ankylosing Spondylitis Functional Index 54.12 26.99 56.89 22.48 55.23 25.07 54.09 23.23 55.11 24.37
on VAS, mean SD

* DMARD disease-modifying antirheumatic drug.


Fewer than 10% were taking concomitant corticosteroids.
Includes auranofin, azathioprine, cyclosporine, leflunomide, methotrexate, methotrexate sodium, and sulfasalazine.
0100 mm visual analog scale (VAS).

etoricoxib alone versus naproxen; P 0.068 for 120 mg percentage of discontinuations occurring in the
etoricoxib alone versus naproxen). No notable differ- naproxen group (Figure 1). The mean compliance rate
ences were observed among the treatment groups in over 52 weeks was 95% in all groups.
discontinuation rates due to clinical AEs or laboratory Assessment of efficacy. In comparison with pa-
AEs, or due to other reasons. During the final 46 weeks, tients receiving placebo over the 6-week treatment pe-
a pattern of discontinuations was observed similar to riod, those receiving 90 mg etoricoxib and 120 mg
that seen in part I, with the most common reason for etoricoxib demonstrated statistically significantly greater
discontinuation being lack of efficacy and the highest improvement (P 0.001) in all 3 primary end points
(Figures 3A, B, and C). Baseline mean values for the
primary end points were generally similar among the 4
groups. The greatest treatment effect among the active
treatment groups compared with the placebo group
occurred for pain (a difference in least squares mean
change of 2129 mm), followed by disease activity (a
difference in least squares mean change of 1825 mm),
followed by function (a difference in least squares mean
change of 1115 mm on the BASFI ratings). In combi-
nation, the 90 mg and 120 mg etoricoxib treatment
group demonstrated statistically significantly greater im-
provement in treatment responses compared with the
naproxen group for all 3 primary end points (P 0.05).
The individual etoricoxib 90 mg and 120 mg dose groups
demonstrated generally similar results, with either sta-
tistically significant or numerically greater improvement
Figure 2. Summary of patient discontinuations due to lack of efficacy
in comparison with naproxen. When comparing primary
over the 1-year period (parts I and II). Plot shows time until initial end points, there were no clinically meaningful differ-
discontinuation. R randomization (baseline). ences between the 90 mg etoricoxib group and the
1210 VAN DER HEIJDE ET AL

120 mg etoricoxib group. Results for the secondary and


exploratory efficacy end points were generally consistent
with those for the primary end points (Table 2).
The following ASAS20 response rates were
achieved in the placebo-controlled period: 64.7% in the
90 mg etoricoxib group, 64.8% in the 120 mg etoricoxib
group, 52.5% in the naproxen group, and 20.4% in the
placebo group (P 0.001, active treatment groups
versus the placebo group). The proportion of patients
achieving ASAS20 partial remission criteria was 14.7%,
17.6%, 9.1%, and 3.2% in the 90 mg etoricoxib, 120 mg
etoricoxib, naproxen, and placebo groups, respectively.
In addition, the majority of these patients (40 of 43
patients) reported a good to excellent response to
treatment.
The difference in pain relief was evaluated for
each of the 4 treatment groups on days 16 (at 4 hours
postdose) with the patients assessment of spine pain as
the end point. Statistically significant improvement (P
0.001) in pain relief was shown with all active treatments
relative to placebo on days 26. On day 1, statistically
significant improvement (P 0.02) in pain relief was
shown for the combined 90 mg and 120 mg etoricoxib
group compared with the placebo group, and for the 90
mg etoricoxib group compared with the placebo group.
Maintenance of treatment effects was demon-
strated during part II of the study. At the end of the
52-week treatment period, both etoricoxib dose groups
demonstrated significantly greater treatment effects
compared with the naproxen group for all 3 primary end
points. There was no loss of treatment effect over time
(Figures 3A, B, and C).
In an additional post hoc analysis, the number of
patients who had a good to excellent response on the
PGART based on the time-weighted average response
observed over 6 weeks of treatment was determined.
Good to excellent responses were observed by 71.6%,
69.2%, and 54.5% of patients in the 90 mg etoricoxib,
120 mg etoricoxib, and naproxen groups, respectively,
and by 25.0% in the placebo group. The proportion of
Figure 3. Primary end points for patients maintained on the same patients who had a good or excellent response was
treatment regimen over the 1-year period (parts I and II). For significantly greater (P 0.05) in the combined 90 mg
longitudinal plots, missing values were replaced with the last observed
and 120 mg etoricoxib group than in the naproxen
value, and patients who discontinued part I due to lack of efficacy and
continued into part II were included by carrying forward the last group, and was significantly greater (P 0.001) in the 90
observed value in part I. A, Patients assessment of spine pain (100-mm mg etoricoxib, 120 mg etoricoxib, and naproxen groups
visual analog scale [VAS]). B, Patients global assessment of disease relative to the placebo group.
activity (100-mm VAS). C, Bath Ankylosing Spondylitis Functional Assessment of tolerability. During the study,
Index (100-mm VAS). Horizontal broken line indicates the baseline.
etoricoxib was generally well tolerated (Table 3). During
Vertical broken line denotes the end of part I and the beginning of part
II. S screening visit; R randomization (baseline) visit (between the first 6 weeks, AEs were reported by 39.8%, 47.6%,
screening and randomization was considered the washout period for 48.9%, and 41.4% of patients in the placebo, 90 mg
prior ankylosing spondylitis therapy); LS least squares. etoricoxib, 120 mg etoricoxib, and naproxen groups,
Table 2. Primary, secondary, and exploratory efficacy end points over the 1-year study*
Etoricoxib

Combined 90 mg Naproxen
Placebo 90 mg 120 mg and 120 mg 1,000 mg
(n 93) (n 100) (n 90) (n 190) (n 97)
Primary end points
Patients assessment of spine pain on VAS
6 weeks 12.6 2.3 41.5 2.2 41.6 2.4 41.5 1.6 33.7 2.3
1 year NA 42.9 2.2 44.6 2.4 43.7 1.6 35.4 2.3
Patients global assessment of disease activity on VAS
ETORICOXIB IN ANKYLOSING SPONDYLITIS

6 weeks 3.4 2.2 27.9 2.1 26.6 2.2 27.3 1.5 20.9 2.1
1 year NA 29.5 2.2 30.1 2.3 29.8 1.6 22.6 2.2
Bath Ankylosing Spondylitis Functional Index on VAS
6 weeks 4.0 1.9 19.4 1.8 19.1 1.9 19.3 1.3 14.6 1.8
1 year NA 21.7 1.8 22.4 2.0 22.0 1.4 16.1 1.9
Secondary end points
Bath Ankylosing Disease Activity Index on VAS
6 weeks 6.4 1.9 28.6 1.8 28.5 1.9 28.6 1.3 23.6 1.8
1 year NA 30.4 1.9 31.6 2.0 31.0 1.4 24.5 1.9
Patients assessment of pain at night (03 scale)
6 weeks 0.18 0.06 0.87 0.06 0.88 0.06 0.87 0.04 0.63 0.06
1 year NA 0.85 0.06 0.89 0.06 0.87 0.04 0.68 0.06
Exploratory end points
Modified Schober test results, cm
6 weeks 0.0 0.11 0.53 0.11 0.71 0.11 0.62 0.08 0.55 0.11
1 year NA 0.56 0.11 0.70 0.11 0.62 0.08 0.60 0.11
Duration of morning stiffness, 100-mm VAS (minutes)#
6 weeks 4.6 2.3 (5.5) 25.2 2.2 (30.2) 23.6 2.4 (28.3) 24.4 1.6 (29.3) 20.9 2.3 (25.1)
1 year NA 28.8 2.3 (34.5) 29.0 2.4 (34.8) 28.9 1.6 (34.7) 22.4 2.3 (26.9)

* Values are the least squares mean SEM changes from baseline to 6 weeks and over 1 year, by analysis of time-weighted average response in a modified intent-to-treat approach.
VAS visual analog scale; NA not applicable.
P 0.001 versus placebo.
P 0.050 versus naproxen.
P 0.010 versus naproxen.
0.050 P 0.100 versus naproxen.
# Values were determined on the basis of question number 6 of the Bath Ankylosing Spondylitis Disease Activity Index questionnaire. Results are recorded both as VAS scores
and as the number of minutes on a 120-minute scale.
1211
1212 VAN DER HEIJDE ET AL

Table 3. Summary of safety data over the 1-year study*


Etoricoxib
Naproxen
Placebo 90 mg 120 mg 1,000 mg
(n 93) (n 103) (n 92) (n 99)
Placebo-controlled safety data (part I)
All AEs
One or more AEs 37 (39.8) 49 (47.6) 45 (48.9) 41 (41.4)
Any drug-related AE 16 (17.2) 30 (29.1) 22 (23.9) 22 (22.2)
Any serious AE 0 (0.0) 0 (0.0) 1 (1.1) 0 (0.0)
Discontinued due to an AE 0 (0.0) 2 (1.9) 0 (0.0) 1 (1.0)
Most common AEs
Headache 3 (3.2) 8 (7.8) 4 (4.3) 2 (2.0)
Diarrhea 4 (4.3) 7 (6.8) 4 (4.3) 1 (1.0)
Upper respiratory tract infection 3 (3.2) 1 (1.0) 6 (6.5) 2 (2.0)
Heartburn 1 (1.1) 6 (5.8) 3 (3.3) 4 (4.0)
Safety data over 1 year (parts I and II)
No. receiving same treatment over 1 year 92 83 78
All AEs
One or more AEs 76 (82.6) 68 (81.9) 52 (66.7)
Any drug-related AE 40 (43.5) 30 (36.1) 30 (38.5)
Any serious AE 7 (7.6) 6 (7.2) 6 (7.7)
Discontinued due to an AE 8 (8.7) 4 (4.8) 6 (7.7)
Most common AEs
Upper respiratory tract infection 10 (10.9) 16 (19.3) 9 (11.5)
Dyspepsia 14 (15.2) 3 (3.6) 2 (2.6)
Diarrhea 12 (13.0) 6 (7.2) 5 (6.4)
Headache 10 (10.9) 8 (9.6) 4 (5.1)
Pharyngitis 10 (10.9) 4 (4.8) 4 (5.1)
Heartburn 9 (9.8) 6 (7.2) 6 (7.7)
Abdominal pain 5 (5.4) 2 (2.4) 7 (9.0)
Asthenia/fatigue 8 (8.7) 4 (4.8) 2 (2.6)

* Values are the number (%) of patients. AE adverse experience.


Determined by the investigator to be possibly, probably, or definitely drug-related.
Incidence 5% in any treatment group.
Incidence 7.5% in any treatment group.

respectively. There were no significant differences in the cation discontinuation. There were 2 deaths in the
incidence of overall clinical AEs, drug-related clinical naproxen group: 1 patient died as a result of broncho-
AEs, laboratory AEs, drug-related laboratory AEs, se- pneumonia, peritonitis, and complications of a jejunal
rious AEs, or discontinuations due to AEs among all 4 perforation, and 1 patient died following traumatic
treatment groups. The most commonly reported clinical injuries from a motor vehicle accident.
AEs were headache (2.07.8%), diarrhea (1.06.8%), With regard to AEs of special interest (i.e., GI
heartburn (1.15.8%), and upper respiratory tract infec- nuisance symptoms, edema, and hypertension), there
tion (1.06.5%). were no statistically significant between-group differ-
The safety profile associated with an additional ences. Among the patients with digestive system or
46 weeks of treatment was generally consistent with the abdominal pain AEs across the 1-year continuous study,
observations made during the initial 6 weeks of the the incidence of discontinuations was similar in all 3
study. Although the absolute incidence of AEs was active treatment groups (4.3%, 1.2%, and 6.4% in the 90
higher in part II due to the longer time period, the mg etoricoxib, 120 mg etoricoxib, and naproxen groups,
pattern was qualitatively similar to that in part I, with no respectively). There were no patients in the 1-year
important differences between treatment groups in the continuous treatment groups who discontinued due to
incidences of AEs, drug-related AEs, or discontinua- an edema-related AE. In the 90 mg etoricoxib, 120 mg
tions due to AEs. During the final 46 weeks of the study, etoricoxib, and naproxen groups, there was a similar
there was 1 death in the 120 mg etoricoxib group; the incidence of edema-related AEs (2.2%, 2.4%, and 2.6%,
patient died of cardiac arrest 21 days after study medi- respectively) and a similar incidence of hypertension-
ETORICOXIB IN ANKYLOSING SPONDYLITIS 1213

related AEs (7.6%, 7.2%, and 6.4%, respectively). One response (based on the PGART) in each group was
patient, in the 120 mg etoricoxib group, discontinued the performed. The results of the post hoc analysis corrob-
study due to a hypertension-related AE. orated those of the primary end points; more patients
All cardiovascular thrombotic serious AEs (i.e., had a good to excellent response in the 90 mg and 120
myocardial infarction, ischemic heart disease, cerebro- mg etoricoxib combined treatment group in comparison
vascular accident, cardiac arrest) and significant GI AEs with the naproxen group (P 0.05), and more patients
that were reported in this study occurred during part II. had a good to excellent response in the active treatment
Four patients (3.2%), 1 patient (0.8%), and none in the groups than in the placebo group (P 0.001). The
90 mg etoricoxib, 120 mg etoricoxib, and naproxen clinical significance of the difference in efficacy results
groups, respectively, had investigator-reported cardio- was further corroborated by the discontinuation rates
vascular thrombotic serious AEs (all were confirmed by due to lack of efficacy in the etoricoxib and naproxen
the adjudication process). In addition, 3 patients (2.4%) groups, and by the ASAS20 results. More patients
receiving 120 mg etoricoxib and 4 patients (3.2%) re- discontinued due to lack of efficacy in the naproxen
ceiving naproxen experienced investigator-reported group than in either the 90 mg etoricoxib or 120 mg
PUBs, of which 4 were confirmed. etoricoxib treatment groups. Moreover, the ASAS20
response rate was higher among patients receiving etori-
coxib than among those receiving naproxen.
DISCUSSION
Efficacy with 90 mg etoricoxib compared with 120
This study demonstrated significantly superior mg was generally similar over 6 weeks and over 1 year
efficacy of 90 mg and 120 mg etoricoxib, once daily, and for all primary, secondary, and exploratory end points.
naproxen, 500 mg twice daily, compared with placebo for By day 2, both doses demonstrated superior efficacy
the treatment of AS. In addition, etoricoxib at 90 mg and compared with placebo (P 0.001). Based on these
120 mg and naproxen at 1,000 mg were shown to be results, we conclude that 90 mg etoricoxib provides full
generally safe and well tolerated over 6-week and 1-year efficacy in AS.
treatment periods. Because AS is an arthritic condition In this study, both the 90 mg and 120 mg doses of
for which optimal response often requires extremely etoricoxib were generally safe and well tolerated. The
high doses of NSAIDs, both 90 mg (the recommended overall incidence of AEs was generally similar among
dose for RA) and 120 mg (a dose higher than that the active treatment groups. There were too few GI
recommended for long-term use) of etoricoxib were experiences to make any meaningful assessments of GI
tested in this study. The comparator, naproxen, was safety. Additional safety information is also provided in
tested at 1,000 mg, which is the highest dose recom- the substantial database of etoricoxib phase II/III studies
mended for long-term use in the United States and the (17,18). Recently published safety data from these trials
maximum approved internationally (26). To increase the include data from a combined analysis of upper GI AEs
power to detect differences from placebo, combined across 10 phase II/III studies (including osteoarthritis,
analyses of the etoricoxib 90 mg and 120 mg dose groups RA, and noninflammatory chronic low-back pain); those
were performed, using a step-down approach that ana- results demonstrate a superior tolerability profile of
lyzed the dose combined and then looked for individual etoricoxib (60, 90, or 120 mg) in comparison with the
differences from placebo. The efficacy of both individual nonselective NSAIDs diclofenac, naproxen, or ibupro-
and combined etoricoxib 90 mg and 120 mg dose groups fen. In those analyses, treatment with etoricoxib reduced
was shown to be statistically superior to placebo. Like- the incidence of clinically important upper GI events by
wise, the efficacy of the combined 90 mg and 120 mg 50% compared with treatment with nonselective
etoricoxib group was shown to be significantly greater NSAIDs (18).
than that of naproxen at 500 mg twice daily for all In the present study, hypertension-related and
primary end points over 6 weeks. Moreover, each indi- edema-related AEs were relatively uncommon and the
vidual dose group was shown to have statistically signif- incidence of these AEs in the etoricoxib and naproxen
icantly superior efficacy (P 0.05) compared with groups was generally similar. In this study, there were
naproxen over 1 year. too few cardiovascular experiences to make any mean-
In order to confirm that the differences in the ingful assessments of cardiovascular safety. However,
etoricoxib groups compared with naproxen were clini- additional analyses of pooled data (across phase IIb/III
cally meaningful, even at 6 weeks, a post hoc analysis of etoricoxib clinical studies) from more than 6,700 pa-
the percentage of patients with a good to excellent tients have been performed. These analyses compared
1214 VAN DER HEIJDE ET AL

the incidence of serious cardiovascular thrombotic AEs, Treatment of active ankylosing spondylitis with infliximab: a
randomised controlled multicenter trial. Lancet 2002;359:118793.
which had been adjudicated by an external committee of 4. Braun J, Sieper J, Breban M, Zink A, Alten R, Golder W, et al.
specialists blinded to the patients treatment groups. Anti-tumour necrosis factor therapy for ankylosing spondylitis:
These analyses compared etoricoxib at doses 60 mg international experience. Ann Rheum Dis 2002;61 Suppl 3:5160.
5. Braun J, Sieper J. Building consensus of nomenclature and disease
with placebo, naproxen, and the non-naproxen NSAIDs classification for ankylosing spondylitis: results and discussion of a
ibuprofen and diclofenac. These analyses detected a questionnaire prepared for the International Workshop on New
decreased incidence of serious cardiovascular throm- Treatment Strategies in Ankylosing Spondylitis, Berlin, Germany,
1819 January 2002. Ann Rheum Dis 2002;61 Suppl 3:617.
botic AEs among those receiving naproxen compared 6. Dougados M. Diagnosis and monitoring of spondylarthropathy.
with those receiving etoricoxib, but no discernible dif- Compr Ther 1990;16:526.
ference in serious cardiovascular thrombotic AEs was 7. Dougados M, Dijkmans B, Khan M, Maksymowych W, van der
Linden S, Brandt J. Conventional treatments for ankylosing spon-
observed between patients receiving etoricoxib com- dylitis. Ann Rheum Dis 2002;61 Suppl 3:4050.
pared with those receiving non-naproxen NSAIDs or 8. Miceli-Richard C, Dougados M. NSAIDS in ankylosing spondyli-
placebo (33). Recent evidence resulting from long-term tis. Clin Exp Rheumatol 2002;20 Suppl 28:S656.
9. Leung AT, Malmstrom K, Gallacher AE, Sarembock B, Poor G,
studies in patients with spontaneous adenomatous pol- Beaulieu A, et al. Efficacy and tolerability profile of etoricoxib in
yposis suggests that long-term treatment (1 year) with patients with osteoarthritis: a randomized, double-blind, placebo
selective COX-2 inhibitors (i.e., celecoxib, rofecoxib) and active-comparator controlled 12-week efficacy trial. Curr Med
Res Opin 2002;18:4958.
may be associated with increased risk of cardiovascular 10. Zacher J, Feldman D, Gerli R, Scott D, Hou SM, Uebelhart D, et
thrombotic events relative to placebo (34,35). Additional al. A comparison of the therapeutic efficacy and tolerability of
large, long-term clinical safety outcome studies with etoricoxib and diclofenac in patients with osteoarthritis. Curr Med
Res Opin 2003;19:72536.
etoricoxib and other nonselective COX-2 inhibitors are 11. Birbara CA, Puopolo AD, Munoz DR, Sheldon EA, Mangione A,
presently ongoing. Results from these studies are not yet Bohidar NR, et al. Treatment of chronic low back pain with
available. etoricoxib, a new cyclo-oxygenase-2 selective inhibitor: improve-
ment in pain and disability: a randomized, placebo-controlled,
In summary, over the 6-week placebo-controlled 3-month trial. J Pain 2003;4:30715.
period, the efficacy of the etoricoxib 90 mg and 120 mg 12. Collantes E, Curtis SP, Lee KW, Casas N, McCarthy T, Melian A,
dose groups was superior to that of placebo, and over the et al. A multinational randomized, controlled, clinical trial of
etoricoxib in the treatment of rheumatoid arthritis. BMC Fam
1-year active-comparatorcontrolled period, the efficacy Pract 2002;3:10.
of the etoricoxib 90 mg and 120 mg dose groups was 13. Matsumoto AK, Melian A, Mandel DR, McIlwain HH, Borenstein
superior to that of naproxen. Nevertheless, naproxen D, Zhao PL, et al. A randomized, controlled, clinical trial of
etoricoxib in the treatment of rheumatoid arthritis. J Rheumatol
was also shown to be effective in this study, with 2002;29:162330.
significant differences from placebo noted for all pri- 14. Malmstrom K, Kotey P, Cichanowitz N, Daniels S, Desjardins PJ.
mary end points. The efficacy results for 90 mg etori- Analgesic efficacy of etoricoxib in primary dysmenorrhea: results
of a randomized, controlled trial. Gynecol Obstet Invest 2003;56:
coxib compared with 120 mg were generally similar. The 659.
results from this 1-year study thus suggest that etoricoxib 15. Malmstrom K, Kotey P, Coughlin H, Desjardins PJ. A random-
is efficacious and generally safe and well tolerated for ized, double-blind, parallel-group study comparing the analgesic
effect of etoricoxib to placebo, naproxen sodium, and acetamino-
the treatment of AS. phen with codeine using the dental impaction pain model. Clin J
Pain 2004;20:14755.
ACKNOWLEDGMENTS 16. Schumacher HR Jr, Boice JA, Daikh DI, Mukhopadhyay S,
Malmstrom K, Ng J, et al. Randomised double-blind trial of
The authors wish to thank Karen M. Giallella, BA, for etoricoxib and indomethacin in treatment of acute gouty arthritis.
her assistance with study monitoring, David A. Krupa, MS, for BMJ 2002;324:148892.
performing the post hoc statistical analyses, Calogera L. Mc- 17. Hunt RH, Harper S, Callegari P, Yu C, Quan H, Evans J, et al.
Cormick, BA, for editorial assistance, and Cheryl Russo for Complementary studies of the gastrointestinal safety of the cyclo-
manuscript preparation. oxygenase-2-selective inhibitor etoricoxib. Aliment Pharmacol
Ther 2003;17:20110.
18. Hunt RH, Harper S, Watson DJ, Yu C, Quan H, Lee M, et al. The
REFERENCES gastrointestinal safety of the COX-2 selective inhibitor etoricoxib
assessed by both endoscopy and analysis of upper gastrointestinal
1. Amor B, Dougados M, Khan MA. Management of refractory events. Am J Gastroenterol 2003;98:172533.
ankylosing spondylitis and related spondylarthropathies. Rheum 19. Curtis S, Bolognese J, Lee M, Watson DJ, Harper S. GI-related
Dis Clin North Am 1995;21:11728. treatment discontinuations with etoricoxib compared with nonse-
2. Braun J, Brandt J, Listing J, Rudwaleit M, Sieper J. Biologic lective cyclooxygenase inhibitors (NSAIDs) and placebo [abstract].
therapies in the spondyloarthritis: new opportunities, new chal- Ann Rheum Dis 2002;61 Suppl 1:187.
lenges. Curr Opin Rheumatol 2003;15:394407. 20. Curtis S, Bolognese J, Lee M, Watson DJ, Harper S. Usage of
3. Braun J, Brandt J, Listing J, Zink A, Alten R, Golder W, et al. concomitant gastroprotective agents with etoricoxib, nonselective
ETORICOXIB IN ANKYLOSING SPONDYLITIS 1215

cyclooxygenase inhibitors (NSAIDs), and placebo [abstract]. Ann 29. Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P,
Rheum Dis 2002;61 Suppl 1:207. Calin A. A new approach to defining disease status in ankylosing
21. Harper S, Bolognese J, Lee M, Watson DJ, Curtis S. Fewer spondylitis: the Bath Ankylosing Spondylitis Disease Activity
GI-related treatment discontinuations with etoricoxib compared Index. J Rheumatol 1994;21:228691.
with nonselective cyclooxygenase inhibitors (NSAIDs) [abstract]. 30. Anderson JJ, Baron G, van der Heijde D, Felson DT, Dougados
Arthritis Rheum 2001;44 Suppl 9:S317. M. Ankylosing Spondylitis Assessment Group preliminary defini-
22. Calin A, Elswood J. A prospective nationwide cross-sectional tion of short-term improvement in ankylosing spondylitis. Arthritis
study of NSAID usage in 1331 patients with ankylosing spondylitis. Rheum 2001;44:187686.
J Rheumatol 1990;17:8013. 31. Dougados M, Gueguen A, Nakache JP, Velicitat P, Veys EM,
23. Wasner C, Britton MC, Kraines RG, Kaye RL, Bobrove AM, Fries
Zeidler H, et al. Ankylosing spondylitis: what is the optimum
JF. Nonsteroidal anti-inflammatory agents in rhematoid arthritis
duration of a clinical study? A one year versus a 6 weeks
and ankylosing spondylitis. JAMA 1981;246:216872.
24. Burry HC, Siebers R. A comparison of flurbiprofen with naproxen non-steroidal anti-inflammatory drug trial. Rheumatology (Ox-
in ankylosing spondylitis. N Z Med J 1980;92:30911. ford) 1999;38:23544.
25. Hill HF, Hill AG. Ankylosing spondylitis: open long-term and 32. Chow SC, Liu JP. Design and analysis of clinical trials: concepts
double-blind crossover studies with naproxen. J Clin Pharmacol and methodologies. 2nd ed. New York: John Wiley & Sons; 2003.
1975;15:35562. 33. Curtis SP, Mukhopadhyay S, Ramey D, Reicin A. Cardiovascular
26. Sweetman SC, editor. Naproxen. In: Martindale: the complete safety summary associated with the etoricoxib development pro-
drug reference. 33rd ed. London: Pharmaceutical Press; 2002. p. gram [abstract]. Arthritis Rheum 2003;48 Suppl 9:S616.
602. 34. Bresalier RS, Sandler RS, Quan H, Bolognese JA, Oxenius B,
27. Van der Linden S, Valkenburg HA, Cats A. Evaluation of Horgan K, et al. Cardiovascular events associated with rofecoxib in
diagnostic criteria for ankylosing spondylitis: a proposal for mod- a colorectal adenoma chemoprevention trial. N Engl J Med 2005.
ification of the New York criteria. Arthritis Rheum 1984;27:3618. E-pub ahead of print.
28. Calin A, Garrett S, Whitelock H, Kennedy LG, Ohea J, Mal- 35. Solomon SD, McMurray JJ, Pfeffer MA, Wittes J, Fowler R, Finn
lorie-P, et al. A new approach to defining functional ability in P, et al. Cardiovascular risk associated with celecoxib in a clinical
ankylosing spondylitis: the development of the Bath Ankylosing trial for colorectal adenoma prevention. N Engl J Med 2005.
Spondylitis Functional Index. J Rheumatol 1994;21:22815. E-pub ahead of print.

You might also like