Rationale, design, and governance of Prospective Randomized ...

Rationale, design, and governance of Prospective
Randomized Evaluation of Celecoxib Integrated Safety
versus Ibuprofen Or Naproxen (PRECISION),
a cardiovascular end point trial of nonsteroidal
antiinflammatory agents in patients with arthritis
Matthew C. Becker, MD, a Thomas H. Wang, MD, a Lisa Wisniewski, RN, a Kathy Wolski, MPH, a Peter Libby, MD, b
Thomas F. Lüscher, MD, c Jeffrey S. Borer, MD, d Alice M. Mascette, MD, e M. Elaine Husni, MD, MPH, f
Daniel H. Solomon, MD, MPH, g David Y. Graham, MD, h Neville D. Yeomans, MD, i Henry Krum, MBBS, PhD, FRACP, j
Frank Ruschitzka, MD, c A. Michael Lincoff, MD, a and Steven E. Nissen, MD a for the PRECISION Investigators
Cleveland, OH; Boston, MA; Zurich, Switzerland; New York, NY; Bethesda, MD; Houston, TX; and Sydney and
Melbourne, Australia
Background Pain management in patients with osteoarthritis or rheumatoid arthritis often requires long-term use of
nonsteroidal antiinflammatory drugs (NSAIDs). However, the relative cardiovascular safety of these therapies remains uncertain.
Methods The Prospective Randomized Evaluation of Celecoxib Integrated Safety versus Ibuprofen Or Naproxen
(PRECISION) trial will evaluate the cardiovascular safety of celecoxib, ibuprofen, and naproxen. Approximately 20,000
patients with symptomatic osteoarthritis or rheumatoid arthritis at high risk for, or with, established cardiovascular disease
will be randomized in this double-blind, triple dummy, multinational, multicenter study. The primary end point is the composite
of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke. The trial will continue until 762 primary events
occur with at least 18 months follow-up. Noninferiority of any of the regimens will require a 97.5% upper CI of the hazard ratio
(HR) ≤1.33 and point estimate ≤1.12 for both intent-to-treat (ITT) and modified ITT populations.
Conclusion PRECISION, the first study of patients with high cardiovascular risk chronically treated with a cyclooxygenase-
2 selective inhibitor or nonselective NSAID, will define the relative cardiovascular safety profile of celecoxib, ibuprofen, and
naproxen and provide data to help guide NSAID use for pain management for this population. (Am Heart J 2009;157:606-12.)
The cardiovascular (CV) risks of nonsteroidal antiinflammatory
drugs (NSAIDs) have received intense scientific
and public attention after the withdrawal of
rofecoxib from the market in September of 2004. As
patients with chronic arthritis may continue to require
NSAID therapy to achieve an acceptable quality of life,
From the a Department of Cardiology, Cleveland Clinic Foundation, Cleveland, OH,
b Cardiovascular Division, Department of Medicine, Brigham and Women's Hospital,
Boston, MA, c Department of Cardiology, University Hospital of Zurich, Zurich, Switzerland,
d Division of Cardiovascular Medicine, State University of New York Downstate Health
Sciences Center, New York, NY, e National Heart, Lung, and Blood Institute, Bethesda, MD,
f Orthopedic and Rheumatologic Institute, Cleveland Clinic Foundation, Cleveland, OH,
g Department of Rheumatology, Brigham & Women's Hospital, Boston, MA, h Department of
Gastroenterology, Baylor College of Medicine, Houston, TX, i Dean of Medicine's Unit,
University of Western Sydney, Sydney, and j Department of Epidemiology & Preventive
Medicine, Monash University, Melbourne, Australia.
Trial Registration #: NCT00346216.
Submitted July 21, 2008; accepted December 12, 2008.
Reprint requests: Steven E. Nissen, MD, FACC, Department of Cardiovascular Cleveland
Clinic, Medicine, F15, 9500 Euclid Ave, Cleveland, OH 44195.
E-mails: nissens@ccf.org, beckerm2@ccf.org
0002-8703/$ - see front matter
© 2009, Mosby, Inc. All rights reserved.
doi:10.1016/j.ahj.2008.12.014
more robust data are needed to understand the riskbenefit
ratio of long-term NSAID use in this population.
Nonsteroidal antiinflammatory drugs differ with respect
to their selectivity in inhibiting cyclooxygenase (COX),
an enzyme that exists in 2 isoforms, COX-1 and COX-2.
Older NSAIDs are considered nonselective because these
drugs inhibit both COX isoforms at recommended
therapeutic doses. Nonsteroidal antiinflammatory drugs
such as celecoxib and rofecoxib are considered COX-2
selective because they do not substantially inhibit COX-1
at recommended dosages. Both rofecoxib and celecoxib
have been associated with an increase in CV risk in
placebo-controlled trials conducted in patients with
colonic polyps. 1-3 However, few data exist comparing the
relative or absolute CV risks of nonselective NSAIDs or
selective COX-2 inhibitors in the population in which
they are most commonly used—patients with rheumatoid
arthritis (RA) or osteoarthritis (OA).
The relative extent of COX-1 versus COX-2 inhibition
has potential implications for development of both
thrombotic CV events and adverse gastrointestinal (GI)
effects in patients treated with NSAIDs. Cyclooxygenase-1

American Heart Journal
Volume 157, Number 4
Becker et al 607
Figure 1
Trial randomization and design.
inhibition reduces platelet aggregation by inhibiting
formation of thromboxane A 2 , an endogenous promoter
of platelet aggregation. Cyclooxygenase-2 inhibition
reduces endothelial production of prostacyclin, a vasodilatory
eicosanoid that blocks platelet aggregation in
vitro. The COX-2 selective inhibitors were originally
developed based upon their potential to reduce GI
bleeding, 4-7 a well-established liability of COX-1 inhibition.
Thus, the balance between COX-1 and COX-2
inhibition has both advantages and potential liabilities in
determining the safety of all NSAIDs.
The Prospective Randomized Evaluation of Celecoxib
Integrated Safety versus Ibuprofen Or Naproxen (PRECI-
SION) trial is investigating the CV safety of celecoxib
relative to 2 other commonly used nonselective NSAIDs
in patients with arthritis who are at high CV risk.
Trial governance
PRECISION is governed by an executive committee
(EC) composed of CV, gastroenterology, and rheumatology
specialists. Because of potential concerns about
conflict of interest and the appropriateness of analyses in
the preapproval and postapproval studies of rofecoxib, 7-10
trial governance included specific measures to ensure
academic integrity within the framework of an industrysponsored
trial. Accordingly, members of the EC have
agreed to not accept honoraria, consulting fees, or other
compensation related to nonselective NSAIDs or selective
COX-2 inhibitors during the course of the trial except in
relation to preexisting obligations to data and safety
monitoring committees or events adjudication committees
for ongoing government or industry-sponsored trials.
To this end, full disclosure of commercial interactions is
required of EC members to the coordinating center and to
each other. Representatives of the trial sponsor may
attend EC meetings but cannot vote. Moreover, EC
meetings can include confidential sessions that exclude
the sponsor representatives. The Cleveland Clinic Coordinating
Center for Clinical Research (C5Research,
Cleveland, OH) will perform all data analyses using a
complete copy of the finalized database for all planned
publications or presentations. An independent data
monitoring committee (DMC) will provide guidance
regarding trial conduct and safety concerns. The PRECI-
SION trial is officially registered with www.clinicaltrials.
gov (NCT00346216) and is funded by Pfizer. In addition,
the authors are solely responsible for the design and
conduct of this study, all study analyses, the drafting and
editing of the paper, and its final contents.
Study design
PRECISION is a multicenter, multinational study that
uses a randomized, double-blind, triple-dummy, 3-arm
(celecoxib, ibuprofen, or naproxen) parallel group design
(Figure 1). Randomization is stratified according to the
primary diagnosis (either OA or RA), aspirin use, and
geographic region. Those patients with both OA and RA
are assigned to the RA stratum. At randomization, patients
will receive a lower dose of the active treatment,
celecoxib, 100 mg bid; ibuprofen, 600 mg tid; or
naproxen, 375 mg bid, in a 1:1:1 allocation ratio. At
subsequent visits, in patients with RA, investigators may
increase the dose to a maximum as follows: celecoxib to
200 mg bid, ibuprofen 800 mg tid, or naproxen 500 mg
bid. For patients with OA, ibuprofen and naproxen may
be increased, but no upward dose titration is permitted
for patients who receive celecoxib, consistent with Food
and Drug Administration-approved labeling restrictions

608 Becker et al
American Heart Journal
April 2009
Table I. Major inclusion criteria⁎
Age N55 y
Hypertension
Dyslipidemia (LDL N160 mg/dL or HDL b40 mg/dL in females and
b35 mg/dL in males or subjects currently undergoing lipid-lowering
therapy with statin drugs, fibrates, prescription ω 3-acid ethyl esters, or
prescription niacin [≥1,000 mg/d])
Family history of premature CV disease (MI, angina pectoris, heart failure,
cardiac death or coronary revascularization, stroke, carotid
endarterectomy, or other arterial surgery or angioplasty for
atherosclerotic vascular disease in a parent, grandparent, or sibling
with symptom onset or diagnosis before age 55 y for males and 65 y for
females)
Current smoker (cigarette smoking in the past 30 d)
LVH
Documented ankle brachial index b0.9
History of microalbuminuria, urine protein-creatinine ratio N2
LDL, Low-density lipoprotein; HDL, high-density lipoprotein; LVH, left ventricular
hypertrophy.
⁎ Patients deemed at high risk of CV disease must have 3 of the above.
allowing use of ≤200 mg/d (except in countries where
the approved dosage for OA is 400 mg daily). Esomeprazole
is provided to all patients for gastric protection
during the study. The trial size and duration are eventdriven;
estimated study duration is 48 to 60 months, but
all patients must be observed for a minimum of 18
months. The primary composite endpoint is the first
occurrence of CV death (including hemorrhagic death),
nonfatal myocardial infarction (MI), or nonfatal stroke.
Investigators are encouraged to provide optimal CV
preventive management to study subjects as mandated
by local guidelines. Patients receiving low-dose aspirin
(≤325 mg daily) at the time of randomization are
permitted to continue this therapy regardless of their CV
risk. Analysis of the effects of investigational therapies
will include patients with or without concomitant
aspirin therapy.
Study subjects
Inclusion criteria
PRECISION is enrolling patients ≥18 years of age with a
clinical diagnosis of OA or RA who, as determined by the
individual patient and physician, require daily treatment
with NSAIDs to maintain their quality of life. In patients
with RA, administration of disease-modifying antirheumatic
drugs or oral corticosteroids (≤20 mg prednisone
daily) is permitted. However, stability of disease-modifying
antirheumatic drugs or corticosteroid therapy is
required, defined as use of the same medications for at
least 3 months and stable dosage for at least 1 month.
A key inclusion requirement is presence of, or high risk
for, CV disease. Established disease is defined as ≥50%
occlusion of ≥1 coronary artery by angiography, ≥50%
occlusion of a carotid artery by angiography or ultrasound,
history of stable angina, symptomatic peripheral
arterial disease, or any of the following: prior MI, unstable
Table II. Major exclusion criteria
Unstable angina, MI, CVA, CABG b3 m from randomization
Planned coronary, cerebrovascular, or peripheral revascularization
Uncontrolled hypertension (SBP N140 mm Hg, DBP N90 mm Hg)
Uncontrolled arrhythmia b3 m of randomization
NYHA class III-IV heart failure or ejection fraction ≤35%
Acute joint trauma
Aspirin N325 mg daily
Oral corticosteroid, prednisone (or equivalent corticosteroid) N20 mg
daily
Warfarin
GI ulceration b60 d of randomization or perforation, obstruction, or bleed
b6 m of randomization
Inflammatory bowel disease, diverticulitis active b6 m of randomization
AST, ALT, or BUN N2× the upper limit of normal
Creatinine level N1.7 mg/dL in men, 1.5 mg/dL in women
Lithium therapy
Malignancy b5 y before randomization
Other known, active, significant GI, hepatic, renal, or coagulation
disorders
Allergy to study medications
CVA, Cerebrovascular accident; CABG, coronary artery bypass surgery; SBP, systolic
blood pressure; DBP, diastolic blood pressure; NYHA, New York Heart Association;
AST, aspartate transaminase; ALT, alanine transaminase; BUN, blood urea nitrogen.
angina, percutaneous coronary intervention, coronary
artery bypass graft surgery, transient ischemic attack,
ischemic stroke, prior carotid endarterectomy, or other
arterial surgery or angioplasty. These qualifying events
must have occurred ≥3 months before randomization. In
addition, diabetes mellitus is considered a “CV disease
equivalent.” Inclusion under the category of high risk for
CV disease requires subjects to have ≥3 atherosclerotic
risk criteria outlined in Table I.
Exclusion criteria
Subjects who otherwise meet inclusion criteria are
excluded if, at the time of screening, hypertension is
uncontrolled (blood pressure N140/90), severe (New York
Heart Association Functional class III or IV) heart failure,
or known ejection fraction ≤35% are present, atrial
fibrillation or other serious arrhythmia have occurred
within the past 3 months, treatment with N325 mg/d of
aspirin or warfarin anticoagulation is clinically required,
moderately severe liver or kidney disease exists, or a major
GI hemorrhage or high bleeding risk are present. Table II
describes the major exclusion criteria.
End points
The primary endpoint is the first occurrence of an
Antiplatelet Trialists Collaboration (APTC) end point,
including CV death (including hemorrhagic death),
nonfatal MI, or nonfatal stroke. Myocardial infarction is
defined according to the ACC/AHA guidelines. 11,12
Stroke is defined as an acute neurologic cerebral vascular
event with focal neurologic signs lasting N24 hours.
Secondary endpoints include the first occurrence of a
major adverse CV event defined as a composite of CV

American Heart Journal
Volume 157, Number 4
Becker et al 609
Table III. Secondary end points
First occurrence of MACE: composite of CV death (including hemorrhagic
death), nonfatal MI, nonfatal stroke, hospitalization for unstable angina,
revascularization, or hospitalization for TIA
Clinically significant gastrointestinal events (CSGIEs)
Patient assessment of arthritis pain (visual analog scale)
MACE, Major adverse cardiovascular events; TIA, transient ischemic attack.
death (including hemorrhagic death), nonfatal MI, nonfatal
stroke, hospitalization for unstable angina, revascularization,
or hospitalization for a transient ischemic
attack. In addition, clinically significant GI events and
arthritis pain are evaluated as secondary endpoints.
Clinically significant GI event is defined as symptomatic
gastric or duodenal ulcer; gastroduodenal, small bowel or
large bowel perforation, or hemorrhage; gastric outlet
obstruction, or acute GI hemorrhage of unknown origin.
Arthritis pain assessment is evaluated using a 100-mm
visual analog scale (Table III). 13
Other clinically significant renal or vascular events,
including initiation of dialysis, hospitalization for acute
renal failure, congestive heart failure, or uncontrolled
hypertension will also be evaluated. An independent
clinical events committee composed of multidisciplinary
specialists at C5 Research, blinded to study treatment
allocation, will review and adjudicate all CV, renal, and GI
events (including iron deficiency anemia of GI origin).
Rescue medications
Maintaining patients in their initially assigned randomized
treatment group requires particular attention in a
safety trial. Crossovers between randomized therapies
and loss of patient to follow-up represent confounders
that could potentially undermine interpretability of the
study. However, patients with arthritis often have
exacerbations of symptoms. To limit crossovers and
dropouts, the PRECISION trial allows use of rescue
medications including acetaminophen (up to 4 g daily),
opioids, tramadol, propoxyphene, intraarticular steroids
or hyaluronic acid, and other nonpharmacologic therapies.
During rescue therapy, the study drug will be
continued, if possible. Nonserious adverse events or poor
tolerability may also compromise maintenance of patients
in their originally randomized treatment group. Accordingly,
the study permits temporary discontinuation of
drug therapy, if required. Investigators are instructed to
restart trial medication as soon as possible after such
“drug holidays.”
Statistical methods
PRECISION will be completed when the protocolspecified
number of adjudicated primary end point
events has been reached, and all patients have been
followed for at least 18 months. The primary statistical
hypothesis is that none of the treatments is inferior to
either of the others (“noninferiority trial”). Three
pairwise comparisons (celecoxib vs naproxen, ibuprofen
vs naproxen, and celecoxib vs ibuprofen) will be
evaluated to assess noninferiority.
The noninferiority definitions used in this trial require
both that the upper limit of the 1-sided 97.5% CI for the
hazard ratio not exceed 1.33 and the point estimate of the
hazard ratio not exceed 1.12. If either the upper CI or the
point-estimate exceeds these limits, the noninferiority
criterion will not be met. On the basis of historical
examples, 14 we anticipate that approximately 40% of
those originally randomized will withdraw from their
assigned treatment during the trial but will continue to be
observed for adverse events. Therefore, the trial has
prespecified that noninferiority must be demonstrated in
the intent-to-treat (ITT) population using 2 different
censoring criteria. A standard ITT analysis will include the
first occurrence of any of the composite APTC end points
through 30 months (900 days) postrandomization,
regardless of whether study drug was discontinued.
Subjects who remain event-free will be censored at time
last known to be alive or 900 days, whichever occurs first.
A modified analysis will include the first occurrence of
any composite APTC end point up to and including
30 days after permanent study drug discontinuation, also
censored after 900 days of exposure. These more
stringent criteria—requiring that noninferiority be met
for both a standard and modified ITT analyses— allow
enhanced robustness in testing the primary noninferiority
hypothesis. Although both the ITT and modified ITT
analyses will be censored after 900 days, all outcome data
will be collected for patients through the end of the trial
to enable alternative analyses.
A Cox proportional hazards model, adjusting for
stratification factors (geographic region, diagnosis of OA/
RA, and aspirin use) will be used to calculate the hazard
ratio and CI for the primary analysis. In this trial, each
treatment is being compared to 2 alternative treatments
and therefore has 2 opportunities to be shown inferior.
From a clinical perspective, if one of the treatments is
found to be inferior to either of the comparators, the
treatment will be considered inferior. Because this trial is
assessing noninferiority across all comparisons adjustment
for multiple comparisons to control type I error is
unnecessary. Kaplan-Meier curves will be generated to
show the survival distribution for each treatment. Major
adverse CV events and clinically significant GI event
secondary end points will also be analyzed using Cox
proportional hazard models adjusted for study region,
use of low-dose aspirin, and diagnosis of RA. The change
from baseline in the patient's Global Assessment of
Arthritis, Patient's Assessment of Pain (visual analog
scale), and Health Assessment Questionnaire Disability
Index will be analyzed using analysis of covariance with
treatment and study region as factors and the baseline
score as covariates.

610 Becker et al
American Heart Journal
April 2009
For each pairwise comparison, a total of 508 APTC
events will be necessary to achieve 90% power, when
truth is no increase, for the upper 97.5% confidence
limit for the hazard ratio to exclude an increase of 33%.
Assuming an annual event rate of 2.0% for APTC end
points, approximately 20,000 patients must be enrolled
to generate the requisite 508 events in each pairwise
comparison. Missing data during the on-treatment
period will be imputed using Last Observation Carried
Forward methodology.
Data monitoring committee
monitoring guidelines
A data monitoring committee is responsible for safeguarding
the interests of trial participants, assessing the
safety and efficacy of the interventions during the trial,
and for monitoring the overall trial conduct. The data
monitoring committee will provide recommendations to
the EC about stopping or continuing the trial. The data
monitoring committee may also formulate recommendations
regarding the selection/recruitment/retention of
participants, improving protocol adherence, and the
procedures for data management and quality control. The
data monitoring committee will advise the clinical trial
leadership group, the EC. The EC will be responsible for
promptly reviewing the DMC recommendations, deciding
whether to continue or terminate the trial, and
determining whether amendments to the protocol or
changes in study conduct are required. The DMC
recommendations will be explicit and include specific
expectations from the EC.
A formal interim analysis meeting of the DMC will
occur every 4 to 6 months after trial commencement—
the primary purpose being to safeguard subjects against
harm related to study treatment. The monitoring guideline
for termination of a treatment group is met if it has an
excess of ≥47 primary composite events (CV death,
nonfatal MI, or nonfatal stroke) compared to either of the
other 2 treatment groups. An excess of ≥47 events would
be a strong indication of inferiority. 14 The total number
and percentage of primary composite events will be
presented to the DMC by treatment group. An excess of
≤28 events in one treatment group compared to another
would be an indication of noninferiority; however, the
trial would continue as planned and not be terminated
solely based on early evidence of noninferiority.
Discussion
The 2 most common arthritic disorders, OA and RA,
affect an estimated 46 million individuals in the United
States alone. 15 Nonsteroidal antiinflammatory drugs
effectively reduce pain and inflammation, thereby
enabling patients to function more normally. Given that
arthritis often affects relatively older patients, the
intersection of CV disease and arthritis is substantial.
Furthermore, some forms of arthritis, particularly RA,
are associated with an increased risk of CV morbidity
and mortality. 16,17 Recent controversy over the CV
safety of conventional NSAIDs and COX-2 inhibitors
has created considerable consternation and public
debate. Appropriately designed clinical trials have not
directly addressed the CV effects of conventional
nonselective NSAIDs primarily because these agents
were approved before concerns emerged regarding CV
risk of NSAIDs. Accordingly, there exist few data with
which to guide practitioners when tailoring an
analgesic regimen for the patient at moderate or high
risk for future coronary events.
Conducting a “safety trial” requires many special
considerations and defining a population in which
clinical equipoise exists is challenging. 18 Accordingly,
we chose to conduct this trial only in patients in whom
daily NSAID therapy was deemed essential to maintain
an acceptable quality of life. Because these patients
require NSAIDs, a study to compare the risks of available
agents, even if all tested drugs have risks, remains
ethically appropriate. Although this design will help
define the relative CV safety profile of the 3 active
comparators, the absence of a true placebo arm makes
determination of the risk associated with each of these
agents impossible. Given the paucity of randomized data
regarding the relative CV safety of the “legacy”
nonsteroidal agents, it remains unclear which regimen
conveys the least risk of CV harm. Therefore, although
some data suggest that celecoxib (at doses of ≥400 mg
daily) may be associated with an increased risk of CV
events as compared to placebo, it is unknown if this risk
is greater than that of the other nonsteroidal comparators
used in this trial and commonly prescribed worldwide.
This is the essential dilemma that the PRECISION
trial is designed to answer.
Previous trials from which inferences regarding the
CV safety of celecoxib were obtained included populations
that were generally younger, with fewer comorbidities,
and at lower risk for CV disease. 1,19
Examination of data from previous OA/RA trials 7,20
suggests that the PRECISION population will likely be
older, more often female, with a greater number of
comorbid illnesses, and therefore be at substantially
higher risk for CV events. We chose a population at high
risk for CV events to provide a sufficiently large number
of events to yield confidence intervals narrow enough to
answer the clinical question. As patients with high CV
risk frequently have arthritis requiring daily NSAIDs, this
population has high clinical relevance. Furthermore,
demonstration of noninferiority in a high-risk population
provides valuable inferences regarding the relative safety
of these agents in lower risk populations.
As for the choice of comparators, we chose 3 agents that
are commonly prescribed and represent a spectrum of

American Heart Journal
Volume 157, Number 4
Becker et al 611
COX-2 selectivity. Celecoxib, naproxen, and ibuprofen all
have potential advantages and disadvantages. Naproxen is
the least COX-2 selective and likely exhibits the greatest
potential for adverse GI effects but offers the potential
advantage of greater inhibition of thromboxane relative to
prostacyclin. 21 Ibuprofen exhibits less dominant COX-1
inhibition, providing a more balanced COX-1 versus COX-
2 inhibitory effect. Finally, celecoxib is the most COX-2
selective of the 3 agents, providing the greatest likelihood
of avoiding GI adverse effects. When taken concomitantly,
however, celecoxib may exacerbate aspirin-induced
gastric mucosal injury. Given the large percentage of
aspirin users expected in this trial and the known
gastrotoxic potential of the other 2 comparators, esomeprazole
is provided to all patients for gastric protection
during the study. Notably, evidence for an adverse CV
effect for celecoxib exists almost exclusively with a total
daily dose ≥400 mg, whereas the PRECISION trial limits
exposure to 200 mg daily in most patients. 30 There also
remain important unanswered questions regarding the
safety of naproxen and ibuprofen because these agents
may interfere with antiplatelet efficacy of aspirin. 22,23
Both observational and randomized controlled trials
have associated COX-2 inhibitors with an increased CV
risk. 1-3,7,24-31 However, with rare exception, these data
are limited by short-term follow-up, low CV event rates,
variable dosing regimens, and the lack of active controls.
None of the prior studies prespecified CV outcomes as
primary end points. Although data from the placebocontrolled
Adenoma Prevention with Celecoxib (APC)
study 19 suggest a dose-related increase in a composite of
CVevents, several lines of evidence support the CV safety
of celecoxib when used at the lower dosages selected for
this trial. The Celecoxib Long-Term Arthritis Safety Study
(CLASS) compared celecoxib, diclofenac, and ibuprofen
in patients with OA or RA and showed no significant
difference in CVevents. 6 A nested case-control study of 1.4
million NSAID-treated patients comparing those currently
exposed to COX-2 selective and nonselective NSAIDs to
patients with remote exposure to any NSAID found a dosedependent
CV risk associated with rofecoxib when
compared to celecoxib. 11 When compared to remote
NSAID use, celecoxib did not increase risk of CV events.
Furthermore, the analysis provided no evidence that
naproxen protects from serious CV events. 26
An independent systematic review and meta-analysis of
observational studies comparing the CV effects of COX-2
selective and nonselective NSAID use to remote or nonuse
also suggested a dose-related risk of rofecoxib use that was
most evident during the first month of treatment. 27
Celecoxib was not found to share this risk, Notably,
however, a nonselective NSAID—diclofenac—was found
to have the largest adverse CV risk. 28 Other studies have
suggested that rofecoxib or nonselective NSAIDs administration
have a greater risk of admission for heart failure
compared to patients on celecoxib. 28 Difference in
structure and pharmacologic effects may explain the
differences in CV risks for rofecoxib to celecoxib.
Unfortunately, most data comparing COX-2 inhibitors
to conventional NSAIDs are derived from prescriptionbased
databases, nested case-control series, or epidemiologic
studies. Observational studies are considered
“hypothesis generating” at best and there exist many
examples where such findings were not confirmed in
subsequent randomized trials. Therefore, only a largescale,
randomized clinical trial can establish the relative
risks of celecoxib and conventional NSAIDs. By addressing
this important public health issue, PRECISION will
help define the optimal analgesic strategy in the large
population of patients with chronic arthritic pain and
heightened CV risk.
Acknowledgements
We gratefully acknowledge Thomas Fleming, Jay
Brophy, Leslie Crofford, Debabrata Mukherjee, David
Peura, and Lynn Shemanski for their efforts while
serving as members of the DMC and those of William
Spalding and Kenneth Verburg for assistance with
manuscript preparation.
References
1. Arber N, Eagle CJ, Spicak J, et al. Celecoxib for the prevention
of colorectal adenomatous polyps. N Engl J Med 2006;355:
885-95.
2. Bertagnolli MM, Eagle CJ, Zauber AG, et al. Celecoxib for the
prevention of sporadic colorectal adenomas. N Engl J Med 2006;
355:873-84.
3. Bresalier RS, Sandler RS, Quan H, et al. Cardiovascular events
associated with rofecoxib in a colorectal adenoma chemoprevention
trial. N Engl J Med 2005;352:1092-102.
4. Schnitzer TJ, Burmester GR, Mysler E, et al. Comparison of lumiracoxib
with naproxen and ibuprofen in the Therapeutic Arthritis Research and
Gastrointestinal Event Trial (TARGET), reduction in ulcer complications:
randomised controlled trial. Lancet 2004;364:665-74.
5. Simon LS, Weaver AL, Graham DY, et al. Anti-inflammatory and
upper gastrointestinal effects of celecoxib in rheumatoid arthritis:
a randomized controlled trial. Jama 1999;282:1921-8.
6. Silverstein FE, Faich G, Goldstein JL, et al. Gastrointestinal toxicity
with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis
and rheumatoid arthritis: the CLASS study: a randomized
controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA
2000;284:1247-55.
7. Bombardier C, Laine L, Reicin A, et al. Comparison of upper
gastrointestinal toxicity of rofecoxib and naproxen in patients with
rheumatoid arthritis. VIGOR Study Group. N Engl J Med 2000;343
(21):1520-8, 2 p following 8.
8. Curfman GD, Morrissey S, Drazen JM. Expression of concern
reaffirmed. N Engl J Med 2006;354:1193.
9. Curfman GD, Morrissey S, Drazen JM. Expression of concern:
Bombardier et al. “Comparison of upper gastrointestinal toxicity of
rofecoxib and naproxen in patients with rheumatoid arthritis,” N Engl
J Med 2000;343:1520-8. N Engl J Med 2005;353:2813-4.
10. Psaty BM, Kronmal RA. Reporting mortality findings in trials of
rofecoxib for Alzheimer disease or cognitive impairment: a case study

612 Becker et al
American Heart Journal
April 2009
based on documents from rofecoxib litigation. Jama 2008;299:
1813-7.
11. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA 2002
guideline update for the management of patients with unstable
angina and non-ST-segment elevation myocardial infarction—
summary article: a report of the American College of Cardiology/
American Heart Association task force on practice guidelines
(Committee on the Management of Patients With Unstable Angina).
J Am Coll Cardiol 2002;40:1366-74.
12. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines
for the management of patients with ST-elevation myocardial
infarction; A report of the American College of Cardiology/
American Heart Association Task Force on Practice Guidelines
(Committee to Revise the 1999 Guidelines for the Management of
patients with acute myocardial infarction). J Am Coll Cardiol 2004;
44:E1-E211.
13. Scott J, Huskisson EC. Graphic representation of pain. Pain 1976;2:
175-84.
14. O'Brien PC, Fleming TR. A multiple testing procedure for clinical trials.
Biometrics 1979;35:549-56.
15. Hootman JM, Helmick CG. Projections of US prevalence of
arthritis and associated activity limitations. Arthritis Rheum 2006;
54:226-9.
16. Philbin EF, Ries MD, Groff GD, et al. Osteoarthritis as a determinant
of an adverse coronary heart disease risk profile. J Cardiovasc Risk
1996;3:529-33.
17. Maradit-Kremers H, Crowson CS, Nicola PJ, et al. Increased
unrecognized coronary heart disease and sudden deaths in
rheumatoid arthritis: a population-based cohort study. Arthritis
Rheum 2005;52:402-11.
18. Freedman B. Equipoise and the ethics of clinical research. N Engl J
Med 1987;317:141-5.
19. Solomon SD, McMurray JJ, Pfeffer MA, et al. Cardiovascular risk
associated with celecoxib in a clinical trial for colorectal adenoma
prevention. N Engl J Med 2005;352:1071-80.
20. Cannon CP, Curtis SP, FitzGerald GA, et al. Cardiovascular outcomes
with etoricoxib and diclofenac in patients with osteoarthritis and
rheumatoid arthritis in the Multinational Etoricoxib and Diclofenac
Arthritis Long-term (MEDAL) programme: a randomised comparison.
Lancet 2006;368:1771-81.
21. Salpeter SR, Gregor P, Ormiston TM, et al. Meta-analysis:
cardiovascular events associated with nonsteroidal anti-inflammatory
drugs. Am J Med 2006;119:552-9.
22. Patel TN, Goldberg KC. Use of aspirin and ibuprofen compared with
aspirin alone and the risk of myocardial infarction. Arch Intern Med
2004;164:852-6.
23. Kimmel SE, Berlin JA, Reilly M, et al. The effects of nonselective nonaspirin
non-steroidal anti-inflammatory medications on the risk of
nonfatal myocardial infarction and their interaction with aspirin. J Am
Coll Cardiol 2004;43:985-90.
24. McGettigan P, Henry D. Cardiovascular risk and inhibition of
cyclooxygenase: a systematic review of the observational studies of
selective and nonselective inhibitors of cyclooxygenase 2. JAMA
2006;296:1633-44.
25. Hippisley-Cox J, Coupland C. Risk of myocardial infarction in patients
taking cyclo-oxygenase-2 inhibitors or conventional non-steroidal
anti-inflammatory drugs: population based nested case-control
analysis. Bmj 2005;330:1366.
26. Graham DJ, Campen D, Hui R, et al. Risk of acute myocardial
infarction and sudden cardiac death in patients treated with cyclooxygenase
2 selective and non-selective non-steroidal antiinflammatory
drugs: nested case-control study. Lancet 2005;365:
475-81.
27. Solomon DH, Schneeweiss S, Glynn RJ, et al. Relationship between
selective cyclooxygenase-2 inhibitors and acute myocardial
infarction in older adults. Circulation 2004;109:2068-73.
28. Mamdani M, Juurlink DN, Lee DS, et al. Cyclo-oxygenase-2 inhibitors
versus non-selective non-steroidal anti-inflammatory drugs and
congestive heart failure outcomes in elderly patients:
a population-based cohort study. Lancet 2004;363:1751-6.
29. Solomon SD, Wittes J, Finn PV, et al. Cardiovascular risk of celecoxib
in 6 randomized placebo-controlled trials: the cross trial safety
analysis. Circulation 2008;117:2104-13.
30. Gislason GH, Jacobsen S, Rasmussen JN, et al. Risk of death or
reinfarction associated with the use of selective cyclooxygenase-2
inhibitors and nonselective nonsteroidal antiinflammatory drugs after
acute myocardial infarction. Circulation 2006;113:2906-13.
31. Mukherjee D, Nissen SE, Topol EJ. Risk of cardiovascular events
associated with selective COX-2 inhibitors. JAMA 2001;286:
954-9.