Treatment of Chronic Low Back Pain With Etoricoxib, a New

Cyclo-oxygenase-2 Selective Inhibitor: Improvement in Pain and

DisabilityA Randomized, Placebo-Controlled, 3-Month Trial
Charles A. Birbara,* Anthony D. Puopolo, David R. Munoz, Eric A. Sheldon,
Antoinette Mangione, Norman R. Bohidar, and Gregory P. Geba for the Etoricoxib
Protocol 042 Study Group

Abstract: We evaluated etoricoxib, a novel COX-2specific inhibitor, in 319 patients with chronic
low back pain (LBP) in this double-blind, placebo-controlled trial. Patients were randomized to a 60
mg dose (n 103) or 90 mg dose (n 107) of etoricoxib, or placebo (n 109), daily for 12 weeks. The
primary endpoint was low back pain intensity scale (Visual Analog Scale of 0- to 100-mm) time-
weighted average change from baseline over 4 weeks. Other endpoints included evaluation over 3
months of low back pain intensity scale, Roland-Morris Disability Questionnaire (RMDQ), low back
pain bothersomeness scale, patient- and investigator-global assessments, Patient Health Survey (MOS
SF-12), rescue acetaminophen use, and discontinuation due to lack of efficacy. Etoricoxib provided
significant improvement from baseline versus placebo in pain intensity (4 weeks: 12.9 mm and 10.3
mm for 60-mg and 90-mg doses, P < .001 for each; 12 weeks: 10.5 mm and 7.5 mm for 60-mg and
90-mg doses, P .001 and .018, respectively). Etoricoxib at either dose led to significant improvement
in other endpoints, including RMDQ scores, bothersomeness scores and global assessments. Etori-
coxib given once daily provided significant relief of symptoms, and disability associated with chronic
LBP that was observed 1 week after initiating therapy, was maximal at 4 weeks, and was maintained
over 3 months.
2003 by the American Pain Society
Key words: Etoricoxib, low back pain, chronic, disability, treatment.

L
ow back pain (LBP) is a major cause of disability and
lost work time, second only to upper respiratory
tract infections, and represents the most costly of all
occupational health problems.3,11,35 The condition af-
flicts up to 90% of adults at some point in their lives35
and often requires expensive diagnostic procedures and
hospitalizations.11 The costs associated with low back
pain are estimated at 20 to 50 billion dollars annually in
the United States, with chronic LBP accounting for 75%
to 90% of that total.22 In industrialized societies, such as
the United States, this condition is believed to have
evolved as a consequence of the modern lifestyle, which
includes physical inactivity and being overweight, as well
as the extension of life span, associated deconditioning,
spinal malalignment, disc herniation and degenerative
Received August 23, 2002; Revised February 21, 2003; Accepted March 8,
2003.
From the *University of Massachusetts School of Medicine, Worcester,
MA, Milford Emergency Associates, Milford, MA, Internal Medicine
Northwest, Tacoma, WA, Miami Research Institute, Miami, FL, Albert
Einstein Medical Center, Philadelphia, PA, and Merck and Co, Inc, West
Point, PA.
Supported by Merck & Co, Inc, West Point, PA.
Address reprint requests to Gregory P. Geba, MD, MPH, Clinical Develop-
ment, Merck and Co, Inc, PO Box 4, HM-202, West Point, PA 19486-0004.
E-mail: gregory_geba@merck.com
2003 by the American Pain Society
1526-5900/2003 $30.00 0
doi:10.1016/S1526-5900(03)00633-3
bone changes.11 Despite knowledge of the factors con-
tributing to low back pain, the cause cannot be precisely
determined in individual cases. The problem of diagnosis
is compounded by sensitive but nonspecific imaging
studies.11 This can lead to misdiagnosis, unnecessary sur-
gery, and disability, which contribute to litigation and
absenteeism, with associated significant costs (reviewed
in 22) and a negative societal impact.
Low back pain is most commonly treated with analge-
sic medications. Opioid and nonopioid analgesics, and
especially nonsteroidal anti-inflammatory drugs
(NSAIDS) are employed. Persistent symptoms, despite an-
algesics, have traditionally led physicians to prescribe
physical rehabilitation. A more recent approach has
been to stress a return to normal activities, because nei-
ther bed rest, facet injections, transcutaneous nerve
stimulation, nor traction have been shown to improve
pain and disability in randomized clinical trials.11 Other
historical approaches include acupuncture, which has
not proven to be effective; spinal manipulation, which
may provide some, though modest, benefit; and mas-
sage therapy, which may have a role in the treatment of
acute back strain.24
The effectiveness of these medical treatments for
chronic back pain has not been confirmed.11 Although
most acute episodes of back pain are self-limited, ap-

The Journal of Pain, Vol 4, No 6 (August), 2003: pp 307-315 307

308
proximately 5% to 10% of patients develop symptoms
lasting 3 or more months.3 Many of these patients are
treated with NSAIDs.17,29 Despite their widespread use,
dual COX inhibiting NSAIDs have not previously been
definitively proven to be effective in the treatment of
LBP.17,29 In a recent review of 51 randomized, controlled
trials for acute and chronic low back pain, NSAIDs ap-
peared to be somewhat effective only for short-term
global improvement in patients with acute low back
pain,29 but this effect waned after more than a few
weeks. Patient tolerability of chronic NSAID therapy,
driven mainly by gastrointestinal symptoms, may have
been an obstacle in maintaining favorable out-
comes.17,29,36
Treatment with conventional NSAIDs, which inhibit cy-
clo-oxygenases 1 and 2 (COX-1/COX-2), can be compli-
cated by gastrointestinal mucosal injury.36 This is
thought to be due to NSAID-induced inhibition of
COX-1, the enzyme responsible for the generation of
gastroprotective prostanoids.14,15 The COX-2 inhibitors
were developed to provide the analgesic and anti-in-
flammatory effects of NSAIDs, without causing gastroin-
testinal adverse events mediated by COX-1 inhibi-
tion.6,23,28,30 The COX-2 inhibitors have been associated
with improved gastrointestinal tolerability compared to
dual-COX-inhibiting NSAIDs.7,32,33 COX-2 also mediates
generation of prostanoids responsible for pain, inflam-
mation, and fever,20,21 which are blocked by COX-2 in-
hibitors.
Herein we report the results of a randomized, placebo-
controlled, parallel-group, 12-week, double-blind, piv-
otal phase III study designed to evaluate the efficacy,
safety and tolerability of etoricoxib, a novel COX-2 inhib-
itor, in the treatment of chronic LBP. To guide dose se-
lection, we used previously available data showing the
minimal dose of etoricoxib necessary to produce maxi-
mal effectiveness in 2 other chronic musculoskeletal pain
syndromes (osteoarthritis and rheumatoid arthritis) ob-
tained from studies ranging from 6 weeks to 1 year (re-
viewed in 9 and 18). In those studies, efficacy was maxi-
mal by the second to fourth week and was maintained
over the duration of the study. We therefore tested both
the 60 and 90 mg doses in this study and evaluated the
efficacy of etoricoxib at various points over a trial of
12-weeks duration to determine onset of efficacy and
durability of effect, and to assess the impact of etoricoxib
on improving the disability often associated with this
chronic pain condition. This approach to the evaluation
of chronic musculoskeletal pain is consistent with previ-
ously published guidance.1,2,10
Materials and Methods
Study Population
This study enrolled outpatients between 18 and 75
years of age who had low back pain for at least 3 months,
were regular users, over at least the past 30 days, of
either an NSAID or acetaminophen for treatment of their
low back pain, and were otherwise in good health. Pa-
tients were eligible if they were assessed as Class 1 (pain
Etoricoxib in Chronic Low Back Pain
without radiation to an extremity and without neurolog-
ical signs) or Class 2 (pain with radiation to an extremity,
but not below the knee and without neurological signs)
on the Quebec Task Force system on Spinal Disorders.1
Muscle relaxants, physical therapy, and chiropractic or
alternative therapy (such as acupuncture) were permit-
ted, if their use was stable for the month preceding the
screening visit and was expected to remain stable for the
duration of the study. Patients with low back pain due to
secondary causes (eg, malignancy, inflammatory disease,
osteoporosis, fibromyalgia, ochronosis, vertebral frac-
ture, infection, juvenile scoliosis, or congenital malfor-
mation) were excluded. Also excluded were patients
who had surgery for low back pain within 6 months of
screening, had symptomatic depression that would inter-
fere with completion of the study questionnaires,
abused drugs or alcohol within the past 5 years, used
opioids more than 4 days in the month before screening,
or had corticosteroid injections within 3 months before
screening. Low dose aspirin (81 mg/day) was continued
during the trial if used for cardiovascular prophylaxis
prior to enrolling in the study.
Study Design
This randomized, placebo-controlled, parallel-group,
double-blind study was conducted at 46 centers in the
United States from March to November 2000. The study
was conducted in accordance with the principles of good
clinical practice. The protocol was approved by the insti-
tutional review board at each study site. At the screening
visit, patients provided written informed consent, under-
went a general physical examination and laboratory
safety assessment, and had the following assessments:
low back pain intensity scale, low back pain bothersome-
ness scale, Patient Global Assessment of Disease Status
(PGADS), Roland-Morris Disability Questionnaire
(RMDQ),27 Hospital Anxiety and Depression Scale
(HADS),37 and the Patient Health Survey (MOS Short
Form [SF]-12).31 After the screening visit, patients discon-
tinued prestudy analgesics for a stipulated washout pe-
riod of 4 to 15 days, depending upon the half-life of the
previous medication used to treat LBP. At the completion
of washout, patients returned to the study site for re-
evaluation. Patients were randomized if they met the
following flare criteria: (1) a score of 40 mm on the low
back pain intensity scale; (2) an increase of 10 mm on the
low back pain intensity scale compared to the screening
visit; and (3) a worsening of PGADS by 1 point (on a 0 4
point scale) compared to the screening visit.
Within each study center, patients were randomly al-
located to the treatment groups using a computer-gen-
erated random allocation schedule. Eligible patients
were randomized to a 90-mg dose of etoricoxib, a 60-mg
dose of etoricoxib, or placebo (1:1:1). All treatments
were double-blinded and double-dummy. Data-analysis
personnel were also blinded during the screening pro-
cess while reviewing and making corrections to the data
before the data file was locked for final statistical anal-
ysis and reporting. No patients in this study were un-
blinded prematurely.
ORIGINAL REPORT/ Birbara et al
The study drug was taken once daily, every morning,
for a total of 12 weeks. Acetaminophen (up to 1950 mg
daily) was provided as rescue analgesia for use through-
out the study. Patients were instructed not to take rescue
analgesia within the 12 hours before each efficacy visit.
Patients returned at 1, 2, 4, 8, and 12 weeks postrandom-
ization for efficacy and safety assessments. All random-
ized patients underwent safety assessments at each
study visit, including measurement of vital signs, record-
ing of adverse experiences (AEs), laboratory safety as-
sessment (at 2, 4, 8, and 12 weeks, or at discontinuation,
and 2 weeks post-therapy), and physical examination (at
4 and 12 weeks, or at discontinuation). Investigators
evaluated all AEs according to their intensity and rela-
tionship to study drug while still blinded to the treat-
ment.
Statistical Analysis
The primary statistical analysis of efficacy employed a
modified intent-to-treat approach, which included anal-
ysis of all patients who received at least one dose of
therapy and provided data from 1 efficacy evaluation
(97% for primary endpoint). Only observed data (at
scheduled, unscheduled, and discontinuation visits) were
included in each patients average response; no data
were carried forward or imputed in the computation. A
per-protocol analysis, which excluded patients with a
prespecified set of significant protocol violations, also
was conducted for the primary efficacy endpoint as a
supportive analysis. There were no substantial differ-
ences in results obtained using the intent-to-treat or per-
protocol approaches.
The prespecified hypotheses were that once-daily
etoricoxib would demonstrate superior improvement in
low back pain compared with placebo as assessed by the
low back pain intensity scale time-weighted average
over 4 weeks (primary) and 3 months (secondary), and
that etoricoxib would be safe and well tolerated over 3
months. Other objectives included evaluation of effi-
cacy provided by etoricoxib in terms of the RMDQ, the
low back pain bothersomeness scale, patient- and in-
vestigator-global assessment of response to therapy,
patient health survey (MOS SF-12), PGADS, use of res-
cue acetaminophen and discontinuations due to lack
of efficacy.
Data from a study of corticosteroid injections for LBP
and from the phase IIb trials demonstrating the effect
of etoricoxib in the treatment of osteoarthritis were
used to calculate the detectable difference for the pri-
mary hypothesis. The parallel-group design with 100
patients per group provided 95% power to detect (
0.05, 2-tailed) a 12-mm difference between etoricoxib
and placebo with respect to the average change from
baseline over 4 weeks on the low back pain intensity
scale.
Treatment effect was measured by the time-weighted
average response by taking the time between adjacent
observations divided by the time from the flare/random-
ization visit to the last observation and using it as the
weight for the computation of the average treatment
309
response. The time-weighted average response of a pa-
tient who discontinued the study after only one on-treat-
ment visit was based on only the response at that visit.
Analysis of covariance (ANCOVA) was used to model
treatment response as a function of treatment, study
center, and baseline covariate. Treatment-by-study-cen-
ter and treatment-by-baseline-covariate interactions
were evaluated prior to running the final model. Time-
weighted average responses (least mean squares) for
each treatment group were calculated and differences
between groups compared using ANCOVA with treat-
ment, study center and baseline value of the dependent
value as the covariate. Between group differences were
assessed in the context of the final statistical model. A
step-down, closed testing procedure where the 90-mg
dose was first compared to placebo, followed by the
60-mg dose versus placebo, was employed to assess the
dose-response relationship. For prespecified discontinu-
ations, prespecified AEs, or prespecified laboratory val-
ues outside of predefined limits of change, Fishers exact
test was used for pairwise comparisons of proportions of
patients between treatment groups.
Results
The study enrolled 319 subjects. Baseline characteris-
tics were generally similar in all groups (Table 1).The
mean (SD) age was 52 ( 13) years and 61% were female.
Nearly 85% of all patients were prior NSAID users. Pa-
tients had LBP for a mean duration of 11 ( 10) years and
had a mean pretreatment pain intensity (after flare) of
77 ( 13) mm on the low back pain intensity scale 100
mm VAS. Approximately 13% of patients were classified
as having moderate to severe anxiety, but only 4% of
patients met classification of depression based on the
HADS questionnaire. Fig 1 summarizes the randomiza-
tion and patient disposition.
Primary EndpointLow Back Pain
Intensity Scale
The etoricoxib 60-mg and 90-mg groups both pro-
vided statistically significant improvement in low back
pain intensity (0- to 100-mm VAS) over the first 4
weeks (primary hypothesis; 60 mg vs placebo: 12.94
mm [95% CI: 19.03, 6.86]; 90 mg vs placebo: 10.29
[16.26, 4.31); P values .001 vs placebo for both
doses). Over 12 weeks (secondary hypothesis) results
were similar (Table 2).Treatment responses to etori-
coxib at 60 mg or 90 mg over 4 and 12 weeks were not
significantly different (P .394 and P .359, respec-
tively; Fig 2).
Additional Endpoints
All other endpoints were prespecified in the protocol.
Both the 60-mg and 90-mg doses of etoricoxib provided
significant improvements compared to placebo over 4
and 12 weeks in terms of RMDQ (Fig 3), the low back pain
bothersomeness scale (Fig 4), PGADS (Fig 5), and patient-
and investigator-global assessment of response to ther-
apy (Table 2). The treatment effects provided by the 2
doses of etoricoxib were not significantly different from
310 Etoricoxib in Chronic Low Back Pain

Table 1. Baseline Patient Characteristics by Treatment Group

ETORICOXIB

STATISTIC PLACEBO 60 MG 90 MG TOTAL

Age (years)
No. 107 101 106 314*
Mean (SD) 51.0 (13.7) 52.3 (13.3) 52.2 (12.4) 51.8 (13.1)
Range 23 to 75 21 to 75 20 to 75 20 to 75
Gender, [n (%)]
Female 59 (55.1) 64 (63.4) 67 (63.2) 190 (60.5)
Male 48 (44.9) 37 (36.6) 39 (36.8) 124 (39.5)
Race, [n (%)]
White 94 (87.9) 88 (87.1) 96 (90.6) 278 (88.5)
Black 8 (7.5) 6 (5.9) 5 (4.7) 19 (6.1)
Hispanic 5 (4.7) 4 (4.0) 5 (4.7) 14 (4.5)
Asian 0 (0.0) 1 (1.0) 0 (0.0) 1 (0.3)
Native American 0 (0.0) 1 (1.0) 0 (0.0) 1 (0.3)
Indian 0 (0.0) 1 (1.0) 0 (0.0) 1 (0.3)
Prior analgesic use, [n (%)]
Non-NSAID 15 (14.3) 21 (20.8) 12 (11.3) 48 (15.4)
NSAID 90 (85.7) 80 (79.2) 94 (88.7) 264 (84.6)
Duration of chronic low back pain
No. 107 101 106 314
Mean (SD) 10.7 (10.0) 12.1 (10.5) 10.8 (10.8) 11.2 (10.4)
Range 1 to 44 1 to 49 1 to 46 1 to 49
Low back pain intensity scale (100-mm VAS)
No. 107 101 105 313
Mean (SD) 76.9 (12.8) 76.9 (12.7) 77.8 (13.6) 77.2 (13.0)
Range 41 to 100 47 to 100 40 to 100 40 to 100
Low back pain bothersomeness scale (0 to 4-point Likert)
No. 107 101 105 313
Mean (SD) 3.0 (0.7) 3.0 (0.6) 3.1 (0.7) 3.0 (0.7)
Range 1 to 4 2 to 4 2 to 4 1 to 4
Patient global assessment of disease status (0 to 4-point Likert)
No. 106 101 106 313
Mean (SD) 2.9 (0.7) 2.9 (0.8) 3.0 (0.8) 2.9 (0.7)
Range 1 to 4 1 to 4 1 to 4 1 to 4
Roland-Morris Disability Questionnaire (0 to 24-point scale)
No. 107 101 106 314
Mean (SD) 14.1 (4.9) 15.4 (5.0) 14.7 (5.0) 14.7 (5.0)
Range 2 to 24 1 to 24 1 to 24 1 to 24
SF-12 Physical Component Summary Scale
No. 104 97 101 302
Mean (SD) 32.6 (8.6) 30.1 (7.4) 32.1 (8.4) 31.6 (8.2)
Range 15 to 54 15 to 55 15 to 55 15 to 55
SF-12 Mental Component Summary Scale
No. 104 97 101 302
Mean (SD) 49.1 (11.2) 47.8 (10.5) 48.3 (10.7) 48.4 (10.8)
Range 19 to 68 25 to 70 24 to 66 19 to 70

*All patients from 1 study center (2, placebo; 2, etoricoxib 60 mg; 1, etoricoxib 90 mg) were excluded from the ITT and PP analyses due to possible data integrity
issues.

each other. Incremental improvements in physical and
mental components of the exploratory MOS SF-12 end-
point were significantly different between etoricoxib
and placebo for the physical component only for the
90-mg dose (P .039).
All patients were dispensed acetaminophen at base-
line for use as pain rescue medication during the treat-
ment period (treatment weeks 1 through 12). Very few
patients required acetaminophen as rescue medication
during the course of the study (average use 1.2 tablets/
day). Differences between groups were not statistically
significant. A significantly smaller proportion of patients
assigned to 60 mg or 90 mg of etoricoxib discontinued due
to lack of efficacy compared to those on placebo (5.9%,
11.3% and 25.2%, respectively; P values: .001, and .012
for 60-mg and 90-mg doses vs placebo, respectively).
ORIGINAL REPORT/ Birbara et al 311

Figure 1. Subject disposition. This figure presents the flow of patients through the study from screening to completion of the study.
Patients who did not complete the study are indicated. *The main reasons patients did not qualify for randomization were the
following: did not meet protocol flare criteria, patients not in good health on the basis of medical history, physical exam, and routine
laboratory tests, and patients were not taking either an NSAID or acetaminophen on a regular basis (24 of the previous 30 days) for
the treatment of low back pain. Patient lost to follow-up, withdrew consent, protocol deviation, or moved. Patients from 1 study
center (2, placebo; 2, etoricoxib 60 mg; 1 etoricoxib 90 mg) were discontinued due to possible data integrity issues.

Safety and Tolerability etoricoxib groups compared to placebo. Etoricoxib at a

Clinical adverse events were reported by 51 (46.8%), 60
(58.3%), and 56 (52.3%) patients in the placebo, 60-mg
etoricoxib, and 90-mg etoricoxib treatment groups, re-
spectively, with no statistically significant differences be-
tween active treatments and placebo. The most com-
monly reported adverse events were headache (5.5%,
11.7%, and 5.6%), nausea (2.8%, 5.8%, 7.5%), diarrhea
(1.8%, 3.9%, 8.4%) and respiratory tract infections
(6.4%, 5.8%, 2.8%) in the placebo, etoricoxib 60 mg and
etoricoxib 90 mg groups, respectively. There were no
significant differences in serious adverse events leading
to discontinuation among groups. The incidence of one
or more clinical adverse events, serious adverse events,
all discontinuations due to adverse events, incidence of
discontinuation due to specific adverse events including
gastrointestinal, edema-related, hypertension-related
events and adverse events of congestive heart failure or
pulmonary edema were not significantly different in the
dose of 90 mg led to a higher incidence of adverse
events, judged by the investigator to be drug-related,
compared to placebo (27.1% vs 11.9%; P .006), but 60
mg was not significantly different from placebo (20.7%
vs 11.9% P .133). There were no significant differences
in drug-related adverse events leading to discontinua-
tion (events [%]: 5 [4.6%], 8 [7.8%] and 10 [9.3%] in the
placebo, etoricoxib 60 mg and 90 mg groups, respec-
tively; P .29 for either active vs placebo).
Discussion
In this 12-week placebo-controlled trial, etoricoxib, 60
or 90 mg once daily, provided significant improvement
across several important efficacy measures in the treat-
ment of chronic LBP. Etoricoxib was effective in terms of
reducing low back pain intensity scores, the primary end-
point, decreasing disability (Roland-Morris), ameliorat-
ing interference in life activities (bothersomeness
312 Etoricoxib in Chronic Low Back Pain

Table 2. Efficacy Endpoints (Modified Intent-to-Treat Approach)

60 MG VS. PLACEBO 90 MG VS. PLACEBO
DIFFERENCE IN LS DIFFERENCE IN LS
ENDPOINT MEAN* (95% CI) MEAN* (95% CI)

Low back pain intensity scale over 4 weeks 12.94 10.29

(100 mm VAS; 0 no pain to 100 extreme pain) (19.03, 6.86) (16.28, 4.31)
Low back pain intensity scale over 12 weeks 10.45 7.50
(100 mm VAS; 0 no pain to 100 Extreme pain) (16.77, 4.14) (13.71, 1.28)
Low back pain bothersomeness scale over 12 weeks 0.38** 0.33**
(4 point Likert; 0 not at all to 4 extremely) (0.62,0.14) (0.57,0.09)
Patient global assessment of disease status over 12 weeks 0.29 0.31
(4 point Likert; 0 very well to 4 very poor) (0.54,0.04) (0.55,0.07)
Patient global assessment of response to therapy over 12 weeks 0.51 0.44**
(4 point Likert; 0 excellent to 4 none) (0.80,0.21) (0.74,0.15)
Roland-Morris Disability Questionnaire over 12 weeks 2.42 2.06**
(0 to 24 point Scale) (3.87,0.98) (3.46,0.65)
SF-12 Physical Component Summary Scale over 12 weeks 1.65 2.58
(0.86,4.16) (0.13,5.02)
SF-12 Mental Component Summary Scale over 12 weeks 1.68 0.92
(0.51,3.87) (3.08,1.23)
Investigator global assessment of response to therapy over 12 0.51 0.43**
weeks (4 point Likert; 0 excellent to 4 none) (0.81,0.21) (0.73,0.14)
Rescue acetaminophen use (tablets per day) over 12 weeks 0.25 0.06
(0.50,0.01) (0.31, 0.19)
Discontinuations due to lack of efficacy 19.3 13.9
(28.8,9.6) (24.1, 3.5)

*LS Least-squares.

Low back pain Intensity was measured by a question to each patient: During the past week, how much low back pain did you have?

Negative values reflect benefit of etoricoxib compared to placebo.

Positive values reflect benefit of etoricoxib compared to placebo.

P .001.

P .05.
**P .01.

scores), and led to significant positive response to ther-
apy as judged by global assessment of both the patient
and the physician.
The 2 doses employed in this study gave comparable
results. Although on some endpoints (MOF SF-12, and
rate of discontinuation due to lack of efficacy) the 90-mg
dose of etoricoxib appeared to provide greater efficacy
than the 60-mg dose, these differences were not statis-
tically significant. This finding is not surprising because
the 60-mg dose of etoricoxib led to improvements com-
pared to on-treatment baseline scores recorded at
screening prior to randomization to study drug for each
of the endpoints tested. It is possible that, for individual
patients, 90 mg might provide additional benefit, but
this cannot be concluded from the results of this study.
The magnitude of the treatment response observed
for the primary endpoint (low back pain intensity score)
has been shown to be clinically meaningful.2 The robust
results obtained for the secondary endpoints provide
strong support of the primary endpoint data. The im-
provements in pain and disability shown in this trial con-
trast with previous trials,17 which evaluated the efficacy
of NSAIDs in low back pain. Koes et al17 reported that
although 5 of 10 previous placebo-controlled studies in
low back pain suggested benefits of NSAIDs, when given
at doses commonly employed in the treatment of osteo-
arthritis and rheumatoid arthritis compared to placebo
in this conditionthis was true for subgroups only in 2 of
these, and in 2 other trials the differences did not reach
statistical significance. Patients with chronic low back
pain were evaluated in only 1 of these trials, which was
deemed to be of low methodologic quality. Generally,
the methodological shortcomings in these previous trials
included lack of statistical power to detect treatment
effects due to small sample sizes, and study of heteroge-
neous patient populations that may have confounded
the results. The use of the Quebec Task Force classifica-
tion was recommended by an expert panel12 to correct
for this deficiency and therefore was used in this study.
A retrospective meta-analysis of pooled data from pla-
cebo-controlled trials of sufficient methodological qual-
ity was used to obtain an estimate of the effect of NSAIDs
in low back pain.17 This analysis suggested that NSAIDs
provided greater efficacy compared to placebo in terms
of the prespecified endpoint defining success of the trial,
which varied according to the study, especially demon-
strable early on after initiating therapy (OR 0.53; 95%
CI: 0.32, 0.89 using a fixed-effect model). However, the
meta-analysis was dominated by studies of short dura-
tion. Durability of effect beyond 2 weeks could not be
ORIGINAL REPORT/ Birbara et al
Figure 2. Low back pain intensity scale (0- to 100-mm VAS).
Least-squares mean change (mm) from baseline (flare/random-
ization visit) over the 12-week treatment period (all-patients
treated approach). Screening (S) to baseline (R) is the washout
period from previous analgesic. Data points for each treatment
group were shifted along the x-axis to maximize legibility at
each time point. (Abbreviation: SE, standard error.) *P .001
time weighted average versus placebo. P .001 time weighted
average versus placebo. P .018 time weighted average
versus placebo.
assessed because only 1 inconclusive study provided data
after more than 2 weeks of treatment. Curiously, despite
the anti-inflammatory effects of NSAIDs, these agents
were not associated with improved outcomes in the sub-
group of patients with signs of radiculopathy due to root
compression, which is often associated with inflamma-
tion and potentially amenable to anti-inflammatory
therapy. However, failure of monotherapy with NSAIDs
in this scenario may not be entirely surprising given that
radiculopathic pain may also involve ischemia and neu-
ropathic components (reviewed in 34) and thus may re-
quire an approach that targets several mechanisms op-
erant in such complex pain.
This is the first placebo-controlled study to evaluate
the efficacy and durability of benefit of a highly selective
COX-2 inhibitor in the treatment of patients with chronic
low back pain over the course of 3 months. Because the
majority of patients in this trial had a long history of back
symptoms (average 11 years), consistent with previous
reports describing the characteristics of patients with this
chronic pain syndrome,11 testing for maintenance of ef-
fect was of particular importance. The results indicated
that the effects of etoricoxib appeared evident after 1
week of therapy, were confirmed by 2 weeks, reached a
plateau at 4 weeks, and were maintained at the 4-week
level over the remainder of the 12 weeks of the trial.
The RMDQ, which measures LBP-related function and
correlates with improvements in work status,12,27 is an
endpoint ideally suited for clinical trials. This question-
naire employs 24 sentences to ascertain the degree of
symptom-limited function in activities of daily living as a
binary response (affirmative or negative). Mean differ-
ences from placebo for the RMDQ at 12 weeks were
2.42 and 2.06 units for the etoricoxib 60-mg and
90-mg groups, respectively, which are consistent with a
313
Figure 3. Roland-Morris Disability Questionnaire scores (0- to
24-point scale). Least-squares mean change (points) from base-
line (flare/randomization visit) over the 12-week treatment pe-
riod (modified intent-to-treat approach). Screening (S) to base-
line (R) is the washout period from previous analgesic. Data
points for each treatment group were shifted along the x-axis to
maximize legibility at each time point. (Abbreviation: SE, stan-
dard error.) *P .001 time weighted average versus placebo. P
.001 time weighted average versus placebo. P .004 time
weighted average versus placebo.
magnitude of effect previously reported to be clinically
meaningful.2
Additional validated assessment tools used to demon-
strate efficacy focused on other clinically relevant end-
points, consistent with recommendations by an interna-
tional panel addressing methodology in evaluating low
back pain efficacy in clinical research.10 Etoricoxib pro-
vided statistically superior efficacy compared to placebo
in all but 1 of the remaining endpoints measured (8 of 9
overall), and improvement from baseline in all endpoints
measured. This contrasts with the lack of consistency of
results in prior NSAID back pain trials. This difference
may be explained by having adequately powered this
study, selecting appropriate patients with confirmed an-
algesic-responsive pain, and applying the Quebec Task
Force criteria as suggested by an expert panel.29 Another
possible explanation may lie in the compliance chal-
lenges of NSAIDs, which in some studies resulted in sig-
nificant dropout due to gastrointestinal adverse ef-
fects.11
For 1 of the secondary endpoints, testing effects on
quality of life using the MOS SF-12 (mental and physical
components), subjects assigned to etoricoxib did not ex-
perience a change from baseline that was statistically
significantly greater than placebo. The MOS SF-12 was
chosen over the longer survey (MOS SF-36) in consider-
ation of the numerous additional efficacy assessments
the patients were asked to complete, and because of
data suggesting its validity as a surrogate for the SF-36 in
other pain models.16,31 However, despite clear improve-
ments in all other parameters, the MOS SF-12 was not
sensitive enough in this study of chronic low back pain to
discriminate from placebo. Previous published findings
have questioned the validity of the shorter question-
naire in studies with less than 200 patients per treatment
314
Figure 4. Low back pain bothersomeness scale (0- to 4-point
Likert scale). Time weighted average change (points) from base-
line (flare/randomization visit) over the 12-week treatment pe-
riod (modified intent-to-treat approach). Screening (S) to base-
line (R) is the washout period from previous analgesic. Data
points for each treatment group were shifted along the x-axis to
maximize legibility at each time point. (Abbreviation: SE, stan-
dard error.) *P .001 time weighted average versus placebo. P
.002 time weighted average versus placebo. P .007 time
weighted average versus placebo.
arm and placed in doubt its appropriateness as a tool to
assess efficacy in low back pain13, 25; the results of this
study lend further support for this view.
The safety profile suggests that etoricoxib was well
tolerated as a treatment for chronic LBP. Gastrointestinal
adverse events can be associated with use of dual COX-1
and COX-2 inhibiting NSAIDs.19 COX-2 selective inhibi-
tors were developed based on the concept that these
specific agents would have improved gastrointestinal
tolerability relative to conventional NSAIDs.26,32 In this
study, at both doses, the rate of gastrointestinal adverse
events with etoricoxib was not significantly different
from placebo. The favorable tolerability profile observed
in this study may have contributed to the outcomes ob-
served in terms of maintenance of pain relief, positive
effects on disability scores and favorable global assess-
ment of response to therapy over this 3-month trial.
The risk of certain renal and vascular events needs to
be considered when treating patients with NSAIDs.8,19
Similar precautions have been recommended when con-
sidering treatment with the available COX-2 inhibi-
tors.4,5 The incidence of renal and vascular events in the
present study was not significantly different between
the 60-mg and 90-mg etoricoxib doses and placebo.
However, the required sample size was based on pre-
dicted efficacy differences and not on estimates of po-
tential differences in adverse event profile. Thus, addi-
tional studies might be pursued to further determine
whether any advantages exist, for an individual patient,
or subgroups of patients, for 1 of the doses employed.
In conclusion, in this 12-week treatment trial, etori-
coxib significantly reduced pain scores, lessened disabil-
ity, reduced impact of back pain on sense of well being
Etoricoxib in Chronic Low Back Pain
Figure 5. Patients global assessment of disease status (0- to
4-point Likert scale). Least-squares mean change (points) from
baseline (flare/randomization visit) over the 12-week treatment
period (modified intent-to-treat approach). Screening (S) to
baseline (R) is the washout period from previous analgesic. Data
points for each treatment group were shifted along the x-axis to
maximize legibility at each time point. (Abbreviation: SE, stan-
dard error.) *P .001 time weighted average versus placebo. P
.012 time weighted average versus placebo. P .013 time
weighted average versus placebo. P .022 time weighted av-
erage versus placebo.
and was well tolerated. Both doses of etoricoxib were
efficacious. Onset of efficacy was observed as soon as 1
week after initiating therapy (first efficacy assessment),
was maximal by 4 weeks and was maintained over the
remainder of the 3-month trial. The data from this study,
obtained across multiple clinical endpoints, confirm the ef-
ficacy of etoricoxib in reducing pain and disability and sup-
port its use as an effective and generally well tolerated
option in the management of chronic low back pain.
Acknowledgments
We thank Drs J. Rush, T. Dobbins, R. Petruschke, W.
Straus, and J. Yates for their review of this manuscript
and for valuable comments. We gratefully acknowledge
the contribution of A. Polis who provided statistical as-
sistance, M. Dixon and C. Skalky who provided manage-
rial support of this study, K. OBrien, L. Mucciola, and A.
Pyeron for their expertise in monitoring this trial, and L.
Mucciola and R. Petruschke for their assistance with
manuscript preparation.
Investigators: The etoricoxib protocol 042 study group
included the following investigators: M. Adamczk, R. Ad-
elizzi, C. Birbara, D. Bong, D. Britt, D. Brune, B. Caciolo, R.
Cavanaugh, T. Coats, C. Codding, G. Collins, L. Colman, E.
DeNoia, D. Evanko, C. Fisher, Jr, F. Fried, E. Gillie, J.
Green, R. Halley, D. Herrington, G. Hill, R. Jackson, R.
Jones, R. Karr, T. Lefton, D. MacPeek, J. Malachowski, A.
Mangione, D. Mehta, A. Mollen, D. Munoz, T. Murphy, N.
Nayak, T. OBarr, D. Orchard, D. Petrone, K. Pryhuber, A.
Puopolo, G. Pyke, W. M. Ryan, L. Schiffman, B. Schwartz,
E. Sheldon, N. L. Smith, R. Swezey, L. Warrick, M. Weitz,
and J. Zuzsga.
ORIGINAL REPORT/ Birbara et al
References
1. Atlas SJ, Deyo RA, Patrick DL, Convery K, Keller RD, Singer
DE: The Quebec task force classification for spinal disorders
and the severity, treatment, and outcomes of sciatic and
lumbar spinal stenosis. Spine 21:2885-92, 1996
2. Bombardier C, Hayden J, Beaton DE: Minimally clinical
important difference: low back pain: outcome measures.
J Rheumatol 28:431-8, 2000
3. Borenstein DG: Chronic low back pain. Rheum Dis Clin
North Am 22:439-56, 1996
4. Boyd IW, Mathew TH, Thomas MC: COX-2 inhibitors and
renal failure: the triple whammy revisited. Med J Aust 173:
274, 2000
5. Brater DC, Harris C, Redfern JS, Gertz B: Renal effects of
COX-2 selective inhibitors. Am J Nephrol 21:1-15, 2001
6. Buttgereit F, Burmester GR, Simon LS: Gastrointestinal
toxic side effects on nonsteroidal anti-inflammatory drugs
and cyclooxygenase-2-specific inhibitors. Am J Med 19(suppl
3A):13S-19S, 2001 110
7. Cannon GW, Breedveld FC: Efficacy of cyclooxygenase-
2-specific inhibitors. Am J Med 19(suppl 3A):6S-12S, 2001 110
8. Clive DM, Stoff JS: Renal syndromes associated with non-
steroidal anti-inflammatory drugs. N Engl J Med 310:563-72,
1984
9. Cochrane DJ, Jarvis B, Keating GM: Etoricoxib. Drugs 62:
2637-51, 2002
10. Deyo RA, Battie M, Beurskens AJ, Bombardier C, Croft P,
Koes B, Malmivaara A, Roland M, Von Korff M, Waldell G:
Outcome measures for low back pain research: a proposal
for standardized use. Spine 23:2003-2013, 1998
11. Deyo RA, Weinstein JN: Low back pain. N Engl J Med
344:363-370, 2001
12. Dionne CE, Von Korff M, Koepsell TD, Deyo RA, Barlow
WE, Checkoway H: A comparison of pain, functional limita-
tions, and work status indices as outcome measures in back
pain research. Spine 24:2339-2345, 1999
13. Gandek B, Ware JE, Aaronson NK, Alonso J, Apolone C,
Bjorner J, Ballinger M, Fukuhara S, Kaasa S, Leplege A, Sul-
livan M: Cross-validation of item selection and scoring for
the SF-12 Health Survey in nine countries: results from the
IQOLA Project. International Quality of Life Assessment.
J Clin Epidemiol 51:1171-1178, 1998
14. Hawkey CJ: Nonsteroidal anti-inflammatory drug gas-
tropathy. Gastroenterology 119:521-535, 2000
15. Hernandez-Diaz S, Rodriguez LA: Association between
nonsteroidal anti-inflammatory drugs and upper gastroin-
testinal tract bleeding/perforation: an overview of epidemi-
ologic studies published in the 1990s. Arch Intern Med 160:
2093-2099, 2000
16. Jenkinson C, Layte R: Development and testing of the UK
SF-12 (short form health survey). J Health Serv Res Policy
2:14-18, 1997
17. Koes BW, Scholten RJPM, Mens JMA, Bouter LM: Efficacy
of non-steroidal anti-inflammatory drugs for low back pain:
a systematic review of randomized clinical trials. Ann Rheum
Dis 56:214-223, 1997
18. Matsumoto AK, Melian A, Mandel DR, McIlwain HH, Bo-
renstein D, Zhao PL, Lines CR, Gertz BJ, Curtis S: A random-
ized, controlled, clinical trial of etoricoxib in the treatment
of rheumatoid arthritis. J Rheumatol 29:1623-1630, 2002
315
19. McCarthy DM: Comparative toxicity of non-steroidal an-
ti-inflammatory drugs. Am J Med 107:37S-46S, 1999
20. Meade EA, Smith WL, DeWitt DL: Differential inhibition
of prostaglandin endoperoxidase synthase (cyclooxygenase)
isozymes by aspirin and other nonsteroidal anti-inflamma-
tory drugs. J Biol Chem 268:6610-6614, 1993
21. Mitchel JA, Akarasereenont P, Thiemermann C, Flower
RJ, Vane JR: Selectivity of nonsteroidal anti-inflammatory
drugs as inhibitors of constitutive and inducible cyclooxy-
genase. Proc Natl Acad Sci USA 90:11693-11697, 1994
22. Nachemson AL: Newest knowledge of low back pain: a
critical look. Clin Orthop 279:8-20, 1992
23. Noble SL, King DS, Olutade JI: Cyclooxygenase-2 enzyme in-
hibitors: place in therapy. Am Fam Physician 61:3669-3676, 2000
24. Philadelphia Panel evidence-based clinical practice
guidelines on selected rehabilitation interventions for low
back pain. Phys Ther 81.1641-1674, 2001.
25. Riddle DL, Lee KT, Stratford PW: Use of SF-36 and SF-12
health status measures: a quantitative comparison for
groups versus individual patients. Med Care 39:867-878,
2001
26. Riendeau D, Percival MD, Brideau C, Charleson S, Dube
D, Ethier D, Falgueyret JP, Friesen RW, Gordon R, Greig G,
Guay P, Mancini J, Ouellet M, Wong E, Xu L, Boyce S, Visco D,
Girard Y, Prasit P, Zamboni R, Rodger IW, Gresses M, Ford-
Hutchinson AW, Young RN, Chan CC: Etoricoxib (MK-0663):
preclinical profile and comparison with other agents that
selectively inhibit cyclooxygenase-2. J Pharmacol Exp Ther
296:558-566, 2001
27. Roland M, Morris R: A study of the natural history of back
pain: part 1: development of a reliable and sensitive mea-
sure of disability in low-back pain. Spine 8:141-144, 1983
28. Simon LS: Role and regulation of cyclooxygenase-2 dur-
ing inflammation. Am J Med 106:37S-42S, 1999
29. van Tulder MW, Scholten RJPM, Koes BW, Deyo RA: Non-
steroidal anti-inflammatory drugs for low back pain: a sys-
tematic review within the framework of the Cochrane col-
laboration back review group. Spine 25:2501-2513, 2000
30. Verburg KM, Maziasz TJ, Weiner E, Loose L, Geis GS,
Isakson PC: Cox-2-specific inhibitors: definition of a new
therapeutic concept. Am J Ther 8:49-64, 2001
31. Ware JE Jr, Kosinski M, Keller SD: A 12-Item short-form
health survey: construction of scales and preliminary tests of
reliability and validity. Med Care 34:220-233, 1996
32. Watson DJ, Harper SE, Zhao PJ, Quan H, Bolognese JA,
Simon TJ: Gastrointestinal tolerability of the selective cyclo-
oxygenase-2 (Cox-2) inhibitor rofecoxib compared with
non-selective Cox-1 and Cox-2 inhibitors in osteoarthritis.
Arch Int Med 160:2998-3003, 2000
33. Weaver AL: Rofecoxib: clinical pharmacology and clinical
experience. Clin Ther 23:1323-1338, 2001
34. Winkelstein BA, Weinstein JN, DeLeo JA: The role of me-
chanical deformation in lumbar radiculopathy: an in vivo
model. Spine 27:27-33, 2002
35. Wipf JE, Deyo RA: Low back pain. Med Clin North Am
79:231-246, 1995
36. Wolfe MM, Lichtenstein DR, Singh G: Gastrointestinal
tolerability of nonsteroidal anti-inflammatory drugs. N Engl
J Med 340:1888-1899, 1999
37. Zigmond A, Snaith RP: The hospital anxiety and depres-
sion scale. Acta Psychiatr Scand 67:361-370, 1983