You are on page 1of 52

m rmists

101 Lucas Valley Road, Suite 210
Tel: (415) 479-8628 l
l San Rafael, California
Fax: (415) 479-8608 l

December 7,200O

Janet Woodcock*.
Director, Center for Drug Evaluation and Research
Food an&Drug Administration
560 &hers Lane
l&c?ckville, MD 20857

Dear Dr. Woodcock:

We hereby petition the FDA to mandate patient medicine guides (MedGuides) for all
prescription drug Cox-II and NSAID preparations. This Petition is submitted pursuant to:
a) 2 1 CFR, Section 10.30; and b) Sections 355 (e) and 3 14.150 of the Federal Food, Drug
and Cosmetic Act.

It has been shown that NSAIDs induce upper GI bleeding which occurs in 5% to 10% of
the population who are taking this medication with an estimated 16,500 NSAID-related
deaths per year.

The cost of the NSAID-related complications exceeds two billion dollars per year.

As a class, NSAIDs are the most widely used drugs in the United States, with more than
seventy million prescriptions written annually. .

It has been reported by UK’s Committee on Safety of Medicines that over eleven deaths
have been associated with the use of Rofecoxib (Vioxx) during the drug’s first year on the
: .: market in the UK.
. I am sending along a copy of the November, 2000 “Worst Pills/Best Pills News” in which
it states adverse GI reactions from upper GI perforations, ulcerations and bleeding occur
especially in the over-65 years of age.

I have written the enclosed letter to Mr. Raymond Gilmartin, President/CEO, Merck
outlining P&I’s concerns.
We respectfully request FDA initiate these MedGuides on all NSAIDs and Cox-II drugs.

The longer the delay in educating patients, the larger the toll of preventable GI bleeds and
other serious damage to the public.


Nothing requested in this Petition will have an impact on the environment.


We certify that, to the best of our knowledge and belief, this Petition includes all
ews on which this Petition relies and that includes representative data
own to the petitioners which are unfavorable to the Petition.

Pharmacists Planning Service, Inc. (PPSI)
p+tarmisb pkx7nir-gservize,it-c
101 Lucas Valley Road, Suite 210 l San Rafael, California 94903
Tel: (4 15) 479-8628lFax: (4 15) 479-8608 le-mail:

December 7,200O

Raymond V. Gilmartin, President/CEO
Merck and Company, Inc.
One Merck Drive, POB 100
Whitehouse Station, New Jersey 08889-4044

Dear Mr. Gilmartin:

I am writing this letter to you regarding concerns of Pharmacists Planning Service, Inc.
(PPSI), a 501 C (3) non-profit public health, consumer, pharmacy education organization,
regarding the Committee on Safety of Medicines’ (the United Kingdom’s (UK) equivalent
to our US Food and Drug Administration (FDA)), most recent report in the September,
2000, issue of*“Current Problems in Pharmacovigilance”.

Our major concerns are regarding Merck’s prescription drug, Vioxx (Rofecoxib), which is
being used for the treatment of both osteoarthritis and acute pain in adults. To summarize
the concerns fi-om June, 1999 and up to July, 2000, a total of 1,120 reports of suspected
adverse reactions have been received by the Committee on Safety of Medicines.
-.. _. -
Adverse GI reactions accounted for almost half of all reports of these. Most were nausea,
upset stomach, diarrhea and abdominal pain. More serious were 68 reports or 12% of
patients reporting of upper GI perforations, ulcerations and bleeding (PUBS). In patients
with PIJBs, five died.

In this same report 177 reports of suspected cardiovascular adverse reactions with
swelling, high blood pressure and palpitations were recorded.

There were 15 rk&orts of heart failure, of aggravated heart failure. Of these 15 reports,

There were 9 reports ofheart attacks-THREE FATAL.

Various psychiatric reactions were reported with Rofecoxib (Vioxx) use--depression,
confusion, hallucinations. -The majority of the patients were reported to have recovered
after Rofecoxib was stopped. Other adverse drug reactions were reported with this drug
including hives, bronchospasms and worsening of asthma along with 65 cases reported of
kidney failure, 12 reports of abnormal liver function and serious skin rashes. Altogether
there were 1 I reports of death associated during the first year of marketing this drug in the
United Kingdom.

PPSI’s major concern is that in Merck’s fuII page ads to both healthcare professionals and
consumers in the direct-to-consumer advertising (DTCA) Merck is saying “Once daily

In the AMA “Archives of Interna Medicine”, June 24,200O article by Ric Day, M.D., et
al. entitled “A Randomized Trial of the Efficacy and Tolerability of the Cox-II Inhibitor
Rofecoxib vs. Ibuprofen in Patients with Osteoporosis” it states in the conclusion
“Rofecoxib was well tolerated and provided clinical efficacy comparable with a high dose
of the NSAID Ibuprofen”.

Since it has been reported that the mortality of patients hospitalized with NSAID-induced
upper GI bleeding is 5% to 1O%, with an estimated 16,500 NSAID-related deaths per year
with the cost of NSAID-related GI complications exceeding two million per year - how
safe is your Rofecoxib product in light of the UK study?

In the “Dear Doctor” letter from Merck signed by Gail A. Ryan, Professional Services,
which quotes a paper from “Arthritis and Rheumatism”, Volume 43; Number 5, May,
2000, printed by the American College of Rheumatology, by Grant W. Canon, M.D., et
al. entitled “Rofecoxib, a Specific Inhibitor of Cyclooxygenase 2, with Clinical Ef&zacy
Comparable with That of Diclofenac Sodium”, it states on Page 983 “The difference in
incidents of GI adverse events (comparing Rofecoxib with Diclofenac Sodium) were

PPSJ respectfuily requests:

1. Merck put out a patient package consumer information medicine guide at its earliest
.. convenience for Vioxx for each prescription dispensed by healthcare professionals.

2. -In light of the UK report that an education and awareness campaign be instituted
_: . ASAP to educate heaithcare professionals about the increased reports of death
i associated with the. use of Rofecoxib. ._

3. That in all full page direct-to-consumer ads (DTCA) the word “safety” be d&ted
and adequate warnings be given to consumers IN BOLD LETTERS on the adverse
risk reactions which can result in this issue.
4. That Rofecoxib is widely assumed (without definitive evidence) to be a much
safer NSAID for the gastrointestinal (GI) tract then the other drugs in its
class because it works by a new mechanism of action. One of PPSI’s concerns is
that since this drug works in a new way, it may also cause harm in a new way.
We would like this reported in a fair and balanced manner.

Finally, I would like to ask for more data and information on the safety and efficacy of this
product in lieu of the reported deaths in the UK.

How many deaths have been reported in the United States? If 11 deaths have been
reported in England in one year, is it safe to say that since England has a population of 60
million, there would be at least 55 deaths reported in the US from use of this drug?

Thank youjo? addressing PPSI’s concerns.
,/” ,

Worst Pills, Best Pills News

Update from the United Kingdom on Adverse Drug
Reactions Reported for the Arthritis Drug Rofecoxib

T he Committee on Safety of
Medicines, the United
Kingdom‘s (U.K.) equivalent
to our Food and Drug Administration
of Medicines. It was estimated that
557,100 prescriptions had been
dispensed for the drug up to the end
of May 2000.
Various psychiatric reactions were
reported with rofecoxib use. These
included depression (28 reports), I ’
con&s&n (14 reports) and--I__ hallucina-
(FDA), summarized the adverse Adverse GI reactions accounted for / tions (11
__+-..e -I-.reports).
_ The majority of ,
reactions they had received for the almost half (554) of all reports. Of patients were reported
i’ recove~~~-a~~~~~~~~~~~~~i~~as to have /
new osteoarthritis drug rofecoxib these, most (84 percent) were nausea, i.~, -,____
.- .
(VIOXX) during its first year of upsetstomach, diarrhea and stopped.
marketing in the September 2000 nal pa& More serious were 68 A
--Adverse drug reactions that are
issue of Cu~ent Problems itz (TiT~~er~~~~~~perfora tions, known tozc=-vith --.--~other NSAIDs \
___.-...-,- _,-.--
Pharmacovigilarzce. ulceration and bleeding. Eri!oxs, were_,_.also
.-.... with
- .“.”_.__._,,_ ,.__. rofecoxib.
We reviewed rofecoxib in the July ulcer%%T&leedini go by the These mcluded._“.__ hives (35 reports),
1999 issue of Worsf Pills, Best Pills News acrolwatients with brotihospasm
.._.. . ....+.-.. or worse%
____ -..-1-11 of
and recommended that it not be used PUBS, 44 (65 percent) recovered, but asthma,.. .(25...) repoaskgkixy failure (16
before July 2004 because it did not five died. More than two thirds (69 reports), &normal ,- _,,.,..,liver fun&i.on (12
offer any documented advantage in percent)
--x-e of the patients experiencing ’ reports), _,_-,..._,
and serious szshes (3
-. ,,,,,...,_,,.,
effectiveness or safety over the more PUBswere over 65years of-age. In six reports).
than 20 nonsteroidal anti-inflamma- cases, aspirin or another NSAID was Altogether there have been 11 i.
tory drugs (NSAIDS) currently on the also implicated, and another 10 /’ reports oix&&Et%%he
-- +
market in the U.S. Rofecoxib --- is widely patient; were taking aspirin at the uiexaecoxib
----- during the drug’s t
asstied (without definitive evidence) __ same time as rofecoxib. first year on the
I _“_,l--.------ I! --
to be.,..a-*.
a --...,--.
much------ safer NSAID for the i There were 177 reports of sus-
__.__ ,,._I-,.-- (GI) tract than the
_,._ pecked cardiovascular adverse
What You Can Do
---_drug ...- in its class because it reactions. Most of these were of
works by a--c
--_..-. new mechanism of action. , swelling (101 reports), high blood You should -_c--wait at least UntilJuly
One of our concerns is that a drug that pressure (31) and palpitations (19). 2004 to take ___._. rofecoxibzs
._.,.^ compel-
works in a new way may also cause There ling research_-_.-. ‘_is $%eTime ..-- FDA
I were 15 reports of heart failure, I-- ___._,..._...
harm in a new way. or aggravated heart failure. Of these hat allows a
- _ ,. . change,+-,iezdrug’s ’
Rofecoxib was first marketed in 15 reports, three patients died. prqduct .,.labeling
Britain i$&ne- 1999 md up through
._;- ?here were also nine g-f \” package i%&%” stating that this drug
II&-ZQQQ,a..totalsf 1;120 repoi@ .of 1..heart attack, thre.ef2tzIn .mo& of ‘is safer ory-----
-&._-.-.. more -IL_ef=e than the -
--. adverse reactions had been &e&ias, the p&er$had risk numerous other- older-NSAIDs
_,,_-“-^-.--.-“_.-. ...____ now
received factors for cardiovascultir disease. on thp mad-+
.-_-- by-.’the Committee on Safety .. -’

; TER.>-.I
i i

. ’
Your Changing Needs.

By calling the Merck National Service Center,
you can receive information on any topics
relevant to Merck, including, for example:

Research and Development
Educational Resources andservices ’ ^.
Medical TO&S

Pricing and Qontracts‘
Sales and Marketing
Packaging and Distribution
Policies and Procedures 1
Industry Health Issues
Merck Divisions and Subsidiaries

For product and serSice inforrbation, call the
Merck National Service Qnter:.
8:00 a.m. to 7:OOp.m. ET, Monday through Fridqy.
For product orders and direct account
inquiries only, call l-800-MERCK RX

Copyright 1997 by Merck &Co., 1~.
All rights reserved.
Remove adhesive label
I? -:’

June 26,200O

A Randomized Trial of tti ,

R. ,Day, Luza, 0. Castaneda,
Merck & Co., Inc.
U.S. Human Health
West Point, PA 19486-0004

Dear Doctor:
i Thank you for your interest in this reprint, “A Randomized Trial of the Efficacy and Tolerability
of the COX-2 Inhibitor Rofecoxib vs Ibuprofen in Patients With Osteoarthritis,” by Ric Day et al,
published in Archives of InternalMedicine, Volume 160, June 2000. We are pleased to provide
this article to you as requested.

This study, a randomized, double-blind trial in 809 adults in whom the knee or hip was the
primary source of pain, compared the clinical efficacy and tolerability of VIOXX’ (rofecoxib)
with that of ibuprofen.

Patients were randomized to one of four treatment groups: placebo; 12.5mg VIOXX once daily;
25-mg VIOXX once daily; or SOO-mgibuprofen three times a day. Clinical efficacy and safety
were monitored during a six-week treatment period.

VIOXX is indicated for relief of the signs and symptoms of osteoarthritis, the managementof
acute pain in adults (see CLINICAL STUDIES), and the treatment of primary dysmenorrhea.

VIOXX is contraindicated in patients with known hypersensitivity to rofecoxib or any other
component of VIOXX. VIOXX should not be given to patients who have experienced asthma,
urticaria, or other allergic-type reactions after taking aspirin or other nonsteroidal anti-
‘inflammatory drugs (NSAIDs). Severe,rarely fatal, anaphyl.actic-like reactions to NSAIDs have
been reported in such patients.

VIOXX is not a sulfonamide; therefore, VIOXX has no sulfonamide contraindication.

Serious gastrointestinal toxicity can occur with or without warning symptoms with NSAIDs.

Before prescribing VIOXX, please read the accompanying complete Prescribing Information.
Thank you for your interest in this information about VIOXX.

Very truly yours,

Gail A. Ryan
Professional Services

Enclosure: Prescribing Information for VIOXX
VIOXX is a registered trademark of Merck & Co., Inc.

: .$
1 *

,j A Randomized Trial of the Efficacy and Tolerability’
.,-g of the COX-2 Inhibitor Rofecoxib vs Ibuprofen
in Patients With Osteoarthritis
Ric Day, MD; Rriggs Morrison, MD; Armando Luza, MD; Oswald0 Castaneda, MD; Albert0 Strusberg, MD;
Menachem Nahir, MD; Knut Bjom Helgetveit, MD; Barbara Kress, RN; Brian Daniels, MD;James Bolognese;
Dave Krupa; Beth Seidenberg, MD; Elliot Ehrich, MD;for the Rofecoxib/Ibuprofen Comparator Study Group

Background: Nonsteroidal anti-inflammatory drugs Results: Both doses of rofecoxib demonstrated efficacy
(NSAIDS) inhibit both cyclooxygenase-1 (COX-1) and clinically comparable with ibuprofen as assessed by 3
cyclooxygenase-2 (COX-2). It is not known whether a primary end points (pain walking on a flat surface
specific inhibitor of COX-2 will provide efficacy in os- [Western Ontario and McMaster Universities Osteoar-
teoarthritis (OA) comparable with NSAIDs. Therefore, thritis Index], patient global assessment of response to
we compared the efficacy and safety of the rofecoxib, therapy, and investigator global assessment of disease
which specifically inhibits COX-2, with those of the status) according to predefined comparability criteria.
NSAID ibuprofen in patients with OA. Both rofecoxib doses and the ibuprofen dose provided
significantly (IY.001) greater efficacy than placebo on
Objective: To compare the clinical efficacy and toler- al1 primary end points. Results from secondary end
ability of rofecoxib (12.5 and 25 mg once daily) with ibu- points were consistent with those of the primary end
profen (800 mg 3 times daily). points. All treatments were well tolerated; the overall
incidence rates of clinical adverse experiences were not
Methods: A randomized, double-blind trial of 809 adults significantly different (P>.OS) among the treatment
with OA was conducted. Patients with OA in whom the groups.
knee or hip was, the primary source of pain were ran- _ .
- - dornized to 1 of 4 treatment groups on demonstration Conclusion: Rofecoxib was well tolerated and pro-
of disease activity: placebo; rofecoxib, 12,5 or 25 mg once vided clinical efficacy comparable with a highdose of the

daily; or ibuprofen, 800 mg 3 times daily. Clinical effi- ~- NSAID ibuprofen.
cacy and safety were monitored during a 6-week treat-
ment period. Arch Intern Med. 2000;160:1781-1787

ONSTEROIDAL anti-inflam-

mation and other stressors.15-20These dis-
From St Vincent’s Hospital, matory drugs (NSAIDS) tinct expression patterns have led to the
Darlinghurst, Australia are commonly used to proposal that prostaglandins produced by
(Dr Day); Merck Research treat the pain and inflam- COX-1 are largely responsible for physi-
Laboratories, Rahwfly, NJ mation caused bv/ a vari- ologic functions,2’ while COX-2-derived
(Drs Morrison, Daniels, ety of clinical disorders, including osteoar- prostaglandins mediate pathophysi-
Seidenberg, and Ehrich,
thritis (OA). The clinical effects of these ologic and inflammatory processes.*’
MS Kress, and Mssrs Bolognese
and Krupa); Clinica San Pablo,
drugs result primarily from the inhibition In vitro and ex vivo assayshave shown
Surca Lima, Peru (Dr Lupa); of the enzyme cyclooxygenase (COX), the that NSAIDs nonspecifically inhibit both the
Clinica Anglo-Americana, first step in the conversion of arachidonic COX-1 and COX-2 isoforms.21-25As pros-
Lima, Peru (Dr Castaneda); acid to prostaglandins.’ taglandins are involved in the mainte-
Department o/Rheumatology, Two COX isoforms (COX-1 and nance of gastrointestinal (GI) tract muco-
Rambam Medical Center, COX-2) have been identified and charac- sal integrity, the well-recognized toxic effects
Haifa, Israel (Dr Nahir); and terized.2-5 Cyclooxygenase-1 is constitu- of NSAIDs on the GI tract26have been pro-
Martina Hansens Hospital, tively active throughout the body6,7and is posed to result largely from inhibition of
Baerum, Norway CO%-1 activity. 21,27
only slightly upregulated in some cells in The therapeutic effects
(Dr Helgetveit). Dr Strusberg
is in private practice in
response to hormones or growth fac- of NSAIDs may be attributable to COX-2 in-
Cordoba, Argentina. See the tors.x,9 In contrast, under basal condi- hibition.2’~28~29 Therefore, agents that spe-
acknowledgments for a list of tions, COX-2 expression is restricted to the cifically inhibit COX-2 are being evaluated
the other members of the brain,lOJi reproductive tract,‘* kidney,13 to determine whether they have efficacy
Rofecoxibllbuprofen and pancreatic islet cells,14 but it is mark- equal to NSAIDs with an improved GI tract
Comparator Study Group. edly upregulated in response to inflam- safety profile.

METHODS Acetaminophen users: Patients who used acetamino-
phen instead of NSAIDs for the treatment of OA were ran-
domized if at both the screening and randomization visits
The primary objective of this randomized, placebo- they met all 3 of the following criteria: (I) they reported a
controlled clinical trial was to compare the clinical effi- minimum of 40 mm on the pain VAS (question 1 of the
cacy of rofecoxib (12.5 and 25 mg once daily) with ibu- WOMAC), (2) they reported a minimum of 40 mm on a
profen (800 mg 3 times daily). All subjects gave written separate VAS evaluating the patient’s global assessment of
informed consent. The study protocol was approved by the disease status, and (3) the investigator rated the global as-
institutional review boards or ethical review committees sessment of disease status as fair, poor, or very poor.
for all 49 investigative sites in 26 countries. The diagnosis of OA was based upon clinical and ra-
diographic evidence of OA (joint space narrowing and os-
STUDY DESIGN teophytes for knee and joint space narrowing for hip). Other
entry criteria included age 40 years or older, American Rheu-
On confirmation of eligibility, patients were randomized to matism Association (ARA; Steinbrocker system) func-
1 of 4 treatment groups by a computer-generated allocation tional class I, II, or II134;symptomatic for at least 6 months;
schedule: placebo, rofecoxib, 12.5 or 25 mg once daily, or the knee or hip the primary source of pain or disability;
ibuprofen, 800 mg 3 times daily. The primary purpose of this and, for women, postmenopausal or demonstrably non-
study was to compare the efficacy of rofecoxib with that of gravid. Patients were excluded if they had significant re-
ibuprofen; a smaller placebo group was included to confirm nal impairment (estimated creatinine clearance ‘0.50 ml/s
that rofecoxib and ibuprofen had efficacy greater than that [530 mUmin or serum creatinine level >177 pmoVL
of the placebo. Thus, the allocation was 1:4:4:4 for placebo, [>2.0 mg/dLl), clinically significant abnormal results of
both doses of rofecoxib, and ibuprofen. The masked alloca- physical examination or laboratory screening, a positive fe-
tion schedule was generated by an individual not otherwise cal occult blood test result, malabsorption, class III/IV an-
involved with the study and kept concealed from all study gina or congestive heart failure, uncontrolled hyperten-
participants. The allocation schedule was unblinded once all sion, stroke or transient ischemic attack within 2 years, active
data had been entered, reviewed, and certified. Medication hepatic disease, a history of recent neoplastic disease, or
was provided in blister packages; study blinding was main- an allergy to acetaminophen or NSAIDs. Patients were
tained by using a matching placebo for each study medica- excluded if they required aspirin at any dose, corticoste-
tion. Patients took 3 tablets each morning and 1 tablet at both roids, warfarin sodium., or ticlopidine hydrochloride.
midday and evening. Patients returned approximately every
2 weeks for 3 visits to assess both efficacy and safety. Pa- EFFICACY ASSESSMENTS
tients were provided open-label acetaminophen for osteoar-
thritic pain not adequately controlled by the study medica- To obtain a comprehensive assessment of the effect of
tion; the maximum daily dose of acetaminophen allowed was rofecoxib on the multiple clinical manifestations of OA, a
2600,mg, and the amount used was recorded. Patients re- variety of efficacy end points were included in the study
turned 7’to 10 days after their last dose of study medication that were derived from the assessments made by both the
for posttherapy safety assessment. patient and the investigator.
At each visit, the patient completed the WOMAC33 and
ENTRY CRITERIA a global assessment of overall disease status (loo-mm VAS,
ranging from “very well” to “very poor”). At treatment vis-
The study included 2 groups of patients with OA. NSAID its, the patient also rated the overall response of his or her
users: These patients discontinued their prior NSAID therapy OA to study medication on a O-to-4 Likert scale (%one”
on confirmation of eligibility. Following a washout period to “excellent”). The physician rated (1) overall assess-
(longer than 5 plasma half-lives of prior NSAID use), pa- ments of disease status (O-to-4 Likert scale of “very poor”
tients’ painwalking on a flat surface was assessedusing ques- to “very well”), (2) overall response of the patient’s OA to
tion 1 of the Western Ontario and McMaster Universities study medication, and (3) study joint tenderness. Exami-
Osteoarthritis Index (WOMAC 3.0),33 a patient-reported nation of study knee or hip joint for tenderness was per-
loo-mm visual analog scale (VAS). Patients were random- formed with the patient in the supine position. Tenderness
ized to the study if they reported a minimum of 40 mm and was defined as pain in response to passive motion or pres-
an increase of 15 mm on the VAS compared with the value sure; the hip was internally and externally rotated, and the
at the screening visit (ie, before discontinuation of NSAIDs), knee was moved through the full available range to detect
and if the investigator’s global assessment of disease status any end range pain and palpated around the medial and
worsened by at least 1 point on a O-to-4 Likert scale com- lateral joint lines while the knee was in the neutral posi-
pared with the screening visit. tion. Pain on palpation (knee only) or during passive I;%nge
: _ ._
I :-. _.__ ‘. __. --


Rofecoxib, 4-[4-(methylsulfonyl)phenyl]-3-phenyl- to 1000 mg. 3o Rofecoxib is, therefore, a specific inhibi-
2(5H)-furanone, inhibits human COX-2 with a-greater tor of the COX-2 isoform in humans.
than 800-fold degree of selectivity relative to COX-1 in Clinical evidence to support the hypothesis that
. :
an in vitro assay with Chinese hamster ovary cell lines rofecoxib has an improved GI tract safety profile
expressing COX-1 or COX-2.30 Using ex vivo human compared with NSAIDs would consist of data demon-
whole blood assays, rofecoxib showed dose-related in- strating that rofecoxib provihes improved GI tract
_ J hibition of COX-2 activit)i but no significant inhibition safety compared with an NSAID at doses that provide
x-i -.
of COX-1 activity with single oral doses ranging from 5 comparable clinical efficacy. In 2 large multicenter

of motion (hip and knee) was graded according to the fol- as a function of the categorical predictors treatment, study
lowing scale: 0, no pain; 1, patient states there is pain; 2, center, and history of ulcer or upper Gl tract bleeding, and a
patient states there is pain and winces; and 3, patient states continuous covariate, the baseline measurement. Mean pa-
there is pain, winces, and withdraws. tient change from baseline and SEs resulting from the
Other measurements of efficacy included amount of ANCOVA were used to compute 95% confidence intervals
rescue acetaminophen consumed and discontinuation from (CIs) for the between-treatment difference in mean re-
the study because of lack of efficacy of the study medica- sponse, tests to compare mean response with active treat-
tion. The primary end points were pain walking on a flat ments against the placebo, and posterior probabilities (based
surface, patient response to therapy, and investigator global on Bayesian analyses with noninformative prior distribu-
assessment of disease status. Secondary end points in- tions35) that the true mean differences in response to the ac-
eluded the WOMAC subscales (pain, stiffness, and disabil- tive treatments were within the predefined clinical compa-
ity), patient global assessment of disease status, investiga- rability bounds. All statistical tests for difference were 2-tailed
tor assessment of response to therapy, patients discontinued with P= .05; Ps.05 was considered statistically significant.
from the study because of lack of efficacy, acetaminophen The primary hypothesis of this study was that rofe-
use, and study joint tenderness. coxib would provide clinical efficacy comparable with
ibuprofen as assessed by 3 primary end points: pain walk-
TOLERABILITY ASSESSMENTS ing on a flat surface, patient response to therapy, and
investigator global assessment of disease status. The fol-
Spontaneously reported adverse experiences and vital signs lowing conditions had to be satisfied to conclude that the
were monitored at every visit. Laboratory investigations, treatments were clinically comparable: in any 2 of the 3
including hematology (complete blood cell count with primary end points, the 95% CIs of mean differences
differential), chemistry (electrolyte, urea nitrogen, creati- between treatment groups had to be within predefined
nine, total protein, albumin, calcium, alanine aminotrans- comparability bounds (+lO mm on a loo-mm VAS and
ferase, aspartate aminotransferase, alkaline phosphatase, and +O.S on a Likert scale), and all of the 3 posterior prob-
total bilirubin levels), and urinalysis (protein, glucose, pH, abilities were required to be 0.950 or lower. These clini-
red blood cells, and white blood cells, with microscopic cal comparability bounds are more conservative than
examination if there were any abnormal results) were per- those proposed by a consensus panel of rheumatolo-
formed at screening, randomization, 4 and 6 weeks of therapy. gists36 and were derived from results of previous OA tri-
and the posttherapy visit. For all adverse experiences, the in- als with rofecoxib. _
vestigator recorded the intensity, the relation to test drug Separate analyses were performed to evaluate effects
(“definitely not” and “probably not” related were scored as on treatment differences of subgroup factors, including race,
not drug-related adverse experiences; “possibly,” “prob- age, sex, study joint (knee vs hip), and prior OA medica-
ably,” and “definitely” related were scored as drug-related tion use (NSAID vs acetaminophen). These were assessed
adverse experiences), the outcome, and any action taken. individually by adding each subgroup factor and its inter-
action with treatment to the ANCOVA model for each of
STATISTICAL ANALYSIS the 3 primary end points.
Safety was assessed by comparing incidence rates of
‘The primary measure for each efficacy end point (except adverse experiences and exceeding predefined limits of
discontinuation because of lack of efficacy) was the mean change in laboratory and vital sign variables between the
response (change from baseline) over all observation times treatment groups. These between-group comparisons were
in the 6-week treatment period. All data collected from dis- calculated using the Fisher exact test; a step-down ap-
continuation and unscheduled visits were included in this proach (25 mg first, and if significant, followed by 12.5 mg)
analysis; no missing values were imputed. For each end was used for the comparisons of rofecoxib doses vs pla-
point, a patient had to have a baseline measurement and cebo.
at least 1 measurement during the 6-week treatment pe- This study had greater than 99% power to demon-
riod for the mean change from baseline to be computed. strate comparable efficacy (according to the criteria cited)
Only 14 of the 809 randomized patients were excluded from between rofecoxib and ibuprofen if their true difference is
the analysis for one or more of the primary end points be- 0. Power calculations were based on observed treatment
cause of missing baseline or on-treatment data. Eighty- effects in other placebo-controlled studies with rofecoxib.
five percent of the randomized patients had a measure- Since this study was designed using variability data from a
ment recorded for all 3 primary end points at all of the pilot study (data on file, Merck Research Laboratories), pro-
planned observation times. vision was made for larger variability. If the underlying SDS
For each end point, analysis of covariance (ANCOVA) were 25% larger than those observed in the pilot study, the
was used to model patient mean change from baseline power was approximately 94%.

endoscopy studies31,32 performed in patients with OA, and physician assessments of efficacy in the treatment
rofecoxib, 25 mg taken once daily, caused fewer endo- of OA.
scopically detected gastroduodenal ulcers than ibupro-
fen, 2400 mg (800 mg 3 times daily). Therefore, we
undertook a prospective randomized study with the
hypothesis that rofecoxib, 25 mg, would provide com- Between April 30 and November 7,1997,1023 patients
parable clinical efficacy with ibuprofen, 800 mg 3 with OA were screened and 809 were enrolled in the study
times daily. The main outcome measures were patient (Figure 1). Patients randomly assigned to the 4 treat-

: -
:; 1783
. .q
@ZOO0 American Medical Association. Ail rights resrrved.
214 PatientsExcluded

*ARA indicates American Rheumatism Association,, NSAIDs, nonsteroidal
anti-in flammatoty drugs.
ment groups had similar sociodemographic and clinical
characteristics, including baseline values for efficacy end
points (Table 1, Figure 1, and Figure 2; additional data rability bounds was greater than 0.950. The effect of ro-
not shown), fecoxib, 25 mg, was significantly superior to that of ibu-
Of the 809 patients, 709 (88%) completed the study; profen (Pc.05) for 2 of the 3 primary end points (patient
the overall discontinuation rate was comparable among response to therapy [P= ,005) and investigator global as-
treatment groups (Figure 3; additional data not shown). sessment of disease status [P= .OOSl).
There was a significantly higher discontinuation rate be- Analyses were performed to determine if the treat-
cause of clinical adverse experiences in the ibuprofen ment effects observed were consistent across various sub-
group compared with the placebo group (P<.O5), whereas groups of patients using treatment-by-subgroup analy-
both rofecoxib groups were not significantly different from ses for the 3 primary en.d points. Treatment effects were
placebo. There were significantly fewer discontinua- consistently observed whether the patients had knee vs
tions because of lack of efficacy (Ps.009) in the active hip as the primary study joint and whether they were an
therapy groups compared with the placebo group (Fig- aceiaminophen user vs an NSAID user at study entry. No
ure 3). The number of patients who withdrew for other statistically significant interactions were observed be-
reasons was similar between all groups (P> .OS). tween treatment and study center, sex, race?age, or ARA
functional class at study entry.
.‘:-._‘ ..- EFFICACY
Figure 2 presents the mean change from baseline dur-
ing the 6-week treatment period for the 3 primary end The incidence of any clinical adverse event was not sig-
points (pain walking on a flat surface, patient response nificantly different among the treatment groups (41.9%
to therapy, and investigator global assessment of dis- in the placebo group vs 50.8% [rofecoxib, 12.5 mgl, 53.3%
ease status) and the secondary end point of the physical [rofecoxib, 25 mgl, and 51.8% [ibuprofen] in the active
function subscale of the WOMAC; data for all primary groups). Drug-related clinical adverse events were more
and secondary end points are shown in Table 2. For all common in all active therapy groups compared with pla-
4 end points, the treatment effect was similar among all cebo (10.8% vs 27.5%, 31.0%, and 30.5%, respectively;
active groups and was superior to the placebo group. Maxi- P= .003). Clinical adverse events that led to discontinu-
mum treatment effects were seen (Figure 2) by the first ation from the study were most common in the ibupro-
evaluation (2 weeks) and were sustained throughout the fen group (1.4% vs 4.1%, 3.7%, and 8:4%, respectively;
6-week treatment period. The treatment effect of rofe- P=.O3 vs placebo for ibuprofen only); this was mostly
coxib was consistently seen for all primary and second- accounted for by adverse experiences related to the GI
ary end points (Figure 2 and Table 2); for each end point tract. Two symptomatic gastric ulcers were observed in
the effect was similar among all active groups, and all ac- the study; both were in the ibuprofen treatment group.
tive groups were superior to the placebo group. The most common clinical adverse experiences were epi-
_ The clinical efficacy of both rofecoxib doses was cdm- gastric discomfort (0% vs 5.7%, 5.8%, and 8.0%, respec-
parable with that of ibuprofen during 6 weeks of treat- tively) , diarrhea (4.1% vs 4.5%, 5.0%, and 5.2%, respec-
ment using the prespecified comparability criteria (see tively), and nausea (1.4% vs 2.9%, 6.6%, and 3.6%,
the “Methods” section). Forall 3 primary end points, the respectively). The incidence of any laboratory adverse
95% CIs for the difference of mean response between each event was not significantly different among the treat-
of the treatment pairs (rofecoxib, 25 mg, and ibuprofen; ment groups (4.1% in the placebo group vs 10.7% [ro-
rofecoxib, 12.5 mg, and ibuprofen; and rofecoxib, 25- fecoxib, 12.5 mgl, 7.6% [rofecoxib, 25 mg], and 13.4%
12.5 mg, and ibuprofen) were within the predefined com- [ibuprofen] in th e active groups). The mean changes in
parability bounds, and the posterior probability that the body weight and blood pressure were similar in all treat-
true mean difference was within the predefined compa- ment groups. Adverse experiences of edema or hyper-
Treatment Groups
A Placebo 0 Rofecoxib,12.5 mg 0 Rofecoxib,25 mg

72.5 + 14.6 (Baseline,Meanf SD)

8.3 (Baseline,Meani SD)

B -1.6- i
.0 -25-
zB -1.8- ,T’
I 6
2 -2.0 - -30 - I
S R 2 4 6 S R 2 4 6
Week Week

Figure 2. Treatment effects over time for the 3 primary clinical efficacy end points. S indicates screening visit; R, randomization visit; and WOMAC,Western
Ontario and McMaster Universities Osteoarthritis Index. Error bars indicate 84% confidence intervals. All active treatments were superior to placebo (P-UJOI).

cerning the role of inhibiting COX-1 vs COX-2 in terms
/Discontinued Becabs; of of the efficacy and safety of this widely prescribed class of
drugs. Previous work has demonstrated that specific in-
hibitors of COX-2 are efficacious in the treatment of OA,37
but left open the question of how that efficacy compares
with NSAIDs, which inhibit both COX-1 and COX-2. In _
this report, we demonstrated that the efficacy of rofe-
coxib, which specifically inhibits COX-2, was comparable
with that of a high dose of the NSAID ibuprofen, Impor-
tantly, we characterized the effect of rofecoxib on a vari-
ety of the clinical manifestations of OA, and in all cases,
we found the efficacy of rofecoxib to be comparable with
Placebo Rofecoxib,12.5 mg Rofecoxib,25 mg- Ibuprofen,2400 mg ibuprofen. These results were obtained in a large, diverse
Treatment_Group . population of patients from 26 different countries; and the
results were consistent across race, age, sex, study joint,
- >‘.l~-..-.‘II-.~ Figure 3. Patients d&continued from the study because of lack dfefficacy or
adverse experience. Thi?iafe of di.&ontinuation for Tackof efficacy was
and prior OA medication use (NSAID vs acetaminophen).
-2~:-.: i - :.. _
_=.-..;&- ..< . ., greater in-theplacebo group compared with-egch of the treatment groups _ -. Our data also demonstrated that rofeco-xib,-12.5 and25 mg;
. (Pc.05). The rate of discontinuation for an adverse event wasgreatir only in -. provided comparable clinical efficacy. Basedupon these and ( “.
the ibuprofen group compared with the placebo group (P<.O5).. _ other data, it is recommended that 12.5 mg be used as the _ .
initial dose of rofecoxib for the treatment of OA.38
. -tensidn were reported at similar rates in all treatment The NSAIDs are associated with a number of toxic ef-
groups: fects, the most important ofwhich are related to the GI tract
and the kidney. To firmly establish an improved safety pro-
file of rofecoxib in contrast to NSAIDs, it is important that
the safety profiles be compared using doses that provide
The discovery of 2 isoforms of COX, the target enzyme in- equivalent efficacy. This study rigorously demonstrated that
hibited by NSAIDs, has led to a number of questions con- both once-daily doses (12.5 mg and 25 mg) of rofecoxib
*A// values are least squares means (95% confidence intervals) exceptfor the end point“patientsdiscontinued because of lack of efficacy ”
t Western Ontario and McMaster Universities Osteoarthritis Index IWOMAC), a visual analog scale of 0 to 100 mm.
$PsOO9 compared with placebo.
3 Likert scale (O-4).
llSca/e of 0 to 3.

(which specifically inhibits COX-2) provided comparable In summarv. we have demonstrated that snecific in-
clinical efficacy with ibuprofen, 800 mg 3 times daily (a hibition of COX-i provides efficacy in the tre’atment of
dual COX-I and COX-2 inhibitor). Therefore, it is appro- OA that is comparable with that of high doses of the
priate to compare the safety and tolerability of rofecoxib, NSAID ibuprofen. The safety of rofecoxib, 12.5 and 25
12.5 and 25 mg, with those of ibuprofen, 2400 mg. mg once daily, was not significantly different from pla-
It is important to note that no adverse events unique cebo and ibuprofen, 800 mg 3 times daily.
to the specific inhibition of COX-2 were apparent in the
rofecoxib treatment groups as assessedin this 6-week con- Acceptedfor publication December 8, 1999.
trolled clinical trial. All active treatments were gener- This work wasfunded by grants-from Merck 6 Co Inc,
ally well tolerated. Adverse experiences potentially at- West Point, Pa.
tributable to renal effects of COX inhibition, such as Presentedas an abstract at the European LeagueAgainst
edema, hypertension, weight gain, and changes in blood Rheumatism 14th Congressin Glasgow, Scotland,June 6-12,
pressure, were comparable in all treatment groups, in- 1999.
cluding the placebo group. Other membersofthe Rofecoxibllbuprofen Comparator
An important clinical adverse effect of NSAIDs is their Study Group are asfollows: A. Alves de Matos, MD, Lisbon,
propensity to lead to serious upper GI tract events, such Portugal; Andre Beaulier, MD, Sainte-Foy, Quebec;JamieBeier,
as perforations, gastric and duodenal ulcerations, arid up- MD, Esbjerg,Denmark; William Bensen,MD, Hamilton, On-
per GI tract bleeding. In other clinical studies,31,32as as- tario; Marco Broggini, MD, Varese,Italy; Ricardo Castro, MD,
sessed by endoscopy, the incidence of abnormalities of San Jose, Costa Rica; Leg Ceder, MD, Lesingborg, Sweden;
the GI mucosae associated with rofecoxib, 25 and 50 mg, Bernard Combe, MD, Montpellier, France; Peter 7. Dawes,
was substantially less than that associated with ibupro- MD, Stoke-on-Trent,England; Marfin deAm&-Martins, MD,
fen, 800 mg 3 times daily. Our study was not intended Sao Paulo, Brazil; H. Eckhardt, MD, Vogtareuth, Germany;
to assess endoscopically diagnosed ulcerations and was Carlos Fuentealba,MD, Santiago, Chile; D. Geissler,MD, Kla-
not large enough or long enough to compare the inci- genjurt, Austria; Mel Games, MD, Lisbon, Portugal; Nigel
dence of serious upper GI tract events. However, in an Gilchrist, MD, Christchurch, New Zealand; Fransisco
overview analysis of all clinical trials3’ performed with Gutierrez, li/rD, Santiago, Chile; Kurt Haas, MD, Stockholm,
rofecoxib, including our study, the incidence of serious Sweden; Gabriel Herrero-Beaumont, MD, Madrid, Spain;
upper GI tract events was found to be significantly less Boulos Haraoui, MD, Montreal, Quebec;Bruce Kirkham, MD,
with rofecoxib compared with NSAIDs. Kogarah, Australia; Peter Kots, MD, Vienna, Austria; Michel
LaRoche, MD, Toulouse Cedex, France; l%ns lvar Lund, 17. Lyons-GiordanoB, PrattaMA,GalbraithW, DavisGL,Arner EC,Interleukin-I dif-
MD, Esbjerg, Denmark; Terry Macedo, MD, Auckland, New ferentially modulateschondrocyteexpressionof cycle-oxygenase-2and phos-
pholipaseAZ. Exp Cell Res. 1993;206:58-62.
Zeuland; Christian Marcelli, MD, Caen, France; Roberto
18. Mitchell JA,AkarasereenontP. Thiemermann6, Flower RJ,VaneJR. Selectivity
Marcolongo, MD, Siena, Italy; Lyn March, MD, St Leonards, of nonsteroidalanti-inflammatorydrugs as inhibitors of constitutiveand induc-
Australia; Erich Mur, MD, Innsbruck, Austria; Pasquale ible cycle-oxygenase.froc NatlAcad Sci U S A. 1994;90:11693-11697.
Or&e, MD, Nuples, Italy; G. Passero,MD, Rome,Italy; Karel 19. BerenbaumF, JacquyesC, ThomasG,Corvol MT, BereziatG, MasliahJ. Syner-
Puvelka, MD, Prague, Czech Republic;Jean Pierre Pelletier, gistic effect of interleukin-1p and tumor necrosisfactor 01or PGEZproduction
by articularchondrocytesdoesnot involve PLAZstimulation.ExpCellRes.1996;
MD, Montreal, Quebec;F. Raeman, MD, Merksem, Belgium; 222:379-384.
Erik Rosal,MD, Guatemala City, Guatemalq PeterRyan, MD, 20. VadasP, StefanskiE,Wloch M, Grouix B, van den BoschH, KennedyB. Secre-
Victoria, Australia; Manfred Shattenkirchner, MD, Munich, tory non-pancreaticphospholipaseA2 andcyclooxygenase-2expressionby tra-
Germany; F. Singer, MD, Laab in Walde, Austria; Malcolm cheobronchialsmooth musclecells. EurJ Biochem.1996;235:557-563.
Smith, MD, Daws Park, Australia; Hyman Tannenbaum, MD, 21. MeadeEA,SmithWL,deWittDL.Differentialinhibitionof prostaglandinendoper-
oxide synthase(cyclooxygenase)isozymesby aspirin and other non-steroidal
Montreal, Quebec;Glen Thomson, MD, Winnipeg, Manitoba; anti-inflammatorydrugs. J Biol Chem.1993;268:6610-6614.
Jesus Tornero, MD, Gualadajara, Spain; Desiree Van der 22. O’NeillGP,ManciniJA, KargmanS, et al. Dverexpressionof humanprostaglan-
Heijde, MD, Maastricht, Holland; Sol Villegas de Morales, din G/Hsynthase-1and -2 by recombinantvacciniavirus: inhibition by nonste-
MD, Caracas, Vaezuela; and P. Villeger, MD, St Gallen, Swit- roidal anti-inflammatorydrugs and biosynthesisof 15-hydroxyeicosatetraenoic
zerland. acid. Mel Pharmacol.1994;45:245-254.
23. Laneuville0, Breuer DK, Dewitt DL, Hla T, FunkCD, Smith WL. Differential in-
We thank Magaly Garcia Woolard, Francis D’Anzi, hibition of human prostaglandinendo-peroxideH synthases-I and -2 by non-
and]ovanni Gonzalezjor their help in conducting this study. steroidal anti-inflammatorydrugs. JPharmacol Exp Ther.1994;271:927-934,
Reprints: Briggs W. Morrison, Merck 6 Co Inc, 126 24. KargmanS, WonQE, Greig GM, et al. Mechanismof selective inhibition of hu-
E Lincoln Ave, RY 32-641, Rahway, NJ 07065. man prostaglandinG/H synthase-1and -2 in intact cells. BiochemPharmacol.
25 Berg J, ChristophT, Widerna M, BodenteichA. lsoenzymespecific cyclooxy-
genaseinhibitors: a whole cell assay system usingthe humanerythroleukemic
cell line HELandthe humanmonocyticcell line monomat 6. JPharmacol Toxi-
co/ Methods.1997;37:179-186.
VaneJR. Inhibitionof prostaglandinsynthesisas a mechanismof actionof aspirin-
like drugs. Naf New Biol. 1971;231:232-235. 26 GabrielSE,JaakkimainenL, BombardierC.Riskfor seriousgastrointestinalcom-
Sirios J, Simmons DL, Richards J. Hormonal regulation of messengerribo- plications related to use of nonsteroidal anti-inflammatory drugs: a meta-
nucleic acid encoding a novel isoform of prostaglandinendoperoxideH syn- analysis.Ann intern Med. 1991;15:787-796.
thasein rat preovulatoryfollicles: inductionin vivo andin vitro. dBio/ Chem.1992; 27 BakhleYS, Botting RM. Cyclooxygenase-2and its regulationin inflammation.
267:11586-l1592. Mediators Inflamm.1996;5:305-323.
3 KujubuDA,FletcherBS,VarnumBC,Lim RW,HerschmanHR.TISIO,A phorbol 28 MasferrerJL, Zweifel BS,ManningPT,et al. Selectiveinhibitionof induciblecy-
ester tumor promoter-induciblemRNAfrom Swiss 3T3 cells, encodesa novel clooxygenase2in vivo isanti-inflammatory andnon-ulcerogenic.froc NatlAcad
prostaglandin synthaseicyclooxygenasehomologue.J Viol Cbem.1991;266: Sci U S A. 1994;91:3228-3232.
12866-l2872. 29 ChanCC,BoyceS, BrideauC, et al. Pharmacologyof a selectivecyclooxygen-
4 Xie WL, ChipmanJG, RobertsonDL, EriksonRL, SimmonsDL. Expressionof a ase-2inhibitor, L745.337:a novel nonsteroidalanti-inflammatoryagentwith an
mitogen-responsivegeneencodingprostaglandinsynthaseis regulatedby mRNA ulcerogenicsparing effect in rat and nonhumanprimatestomach.J Pharmacol
splicing. Proc Nat/Acad Sci U SA. 1991;88:2692-2696. Exp Ther.1995;273:1531-1537.
5 O’BanionMK,WinnVD,YoungDA.cDNAcloningand!unctionalactivity of a glu- 30 EhrichEW,DallobA, de LepeleireI, et al. Characterizationof rofecoxib as a cy-
cocorticoid-regulatedinflammatonjcyclooxygenase.Pruc Nat/AcadSci US A. clooxygenase-2isoform inhibitor and demonstrationof analgesiain the dental
1992;89:4888-4892. pain model.C/inPharmacolTher.1999;65:336-347.
6 SeibertK, MasferrerJL. Roleof induciblecyclooxygenase(COX-2)in inflamma- 31 Lai& L, HawkeyC, HarperS. et al. Effectof the COX-2specific inhibitor (C-2SI)
tion. Receptor.1994;4:17-23. rofecoxibon ulcerformation:a double-blindcomparisonwith ibuprofenand pla-
7 KargmanS, CharlesonS, Cartwright M, et al. Characterizationof prostaglandin cebo [abstract]. Gastroenferology.1999;116;A232.
G/H synthase1 and 2 in rat, dog, monkey,and humangastrointestinaltracts. 32. Hawkey C, Laine L, Mortensen E, et al. Treatmentof osteoarthritis with rofe-
Gastroenterology.1996;i 11:445-454. coxib,a cyclooxygenase-2(COX-2)specificinhibitor,was associatedwith a lower
8. Dewitt D. Prostaglandinendoperoxidesynthase:regulationof enzymeexpres- incidenceof gastroduodenalulcerscomparedto ibuprofenand was comparable
sion. BiacbimBiopbysActa. 1991;1083:121-134. to placebotreatment[abstract]. In: Proceedingsof the 14th Congressof the Eu-
9. Wu KK,SandujaR, Tsai AL, FerhanogluB, Loose-MitchellDS.Aspirin inhibits ropeanLeagueAgainstRheumatism;14thCongress;June6-12,1999; Glasgow,
interleukin-l-inducedprostaglandinH synthaseexpressionin culturedendothe- Scotland.
lial cells. Proc Nat/Acad Sci US A. 1991;88:2384-2387. 33. BellamyN, BuchananWW,GoldsmithCH,CampbellJ, Stitt LW.Validationstudy
10. YamagataK, AndreassonKI, KaufmannWE,BarnesCA,Worley PF.Expression of WOMAC:a health status instrumentfor measuringclinically important pa-
of a mitogen-induciblecyclooxygenasein brain neurons:regulationby synaptic tient relevantoutcomesto antirheumaticdrug therapy in patientswith osteoar-
activity and glucocorticoids.Neuron.1993;11:371-386. thritis of the hip or knee.JRheumatol. 1988:15:1833-1840.
11. KaufmanWE,Worley PF,PeggJ, BremerM, lsakson P. COX-2.a synaptically 34. Arnett FC,EdworthySM, Bloch DA,et al. The AmericanRheumatismAssocia-
inducedenzyme,is expressedby excitatory neuronsat postsynapticsites in rat tion 1987 RevisedCriteriafor the Classificationof RheumatoidArthritis. Atibri-
cerebral cotiex.f’roc Nat/-AcadSciUSA. 1996;93:2317-232i. tis Rheum.1988;31:313-324. -
12. Lim H, PariaBC,DasSK,et al. Multiplefemalereproductivefailures in cyclooxy- 35. RoddaBE, Davis RL. Determiningthe probability of an important difference in
genase2-deficient mice. Cell. 1997;91:197-208. bioavailability.Clin PharmacolTher.1980;28:247-252.
13. Kbmhoff M, Grijne H-J, Klein T, Seyberlh HW, Nijsing RM. Localizationof cy- 36. BellamyN, CaretteS, FordPM, etal. Osteoarthritisantirheumaticdrug trials, Ill:
clooxygenase-1and-2 in adultandfetal humankidney:implicationfor renalfunc- settingthe deltafor clinical trials, results of a consensusdevelopment(Delphi)
tion. Am J fhysiol. 1997;272:F460+468. exercise.J Rheumalol.1992;19:451-457.
14 RobertsonEP.Dominanceof cyclooxygenase-2in the regulationof pancreatic 37 EhrichE,SchnitzerT, Kivitz A, et al, for the Rofecoxib OsteoarthritisPilot Study
islet prostaglandinsynthesis.Diabetes.1998;47:1379-1383. Group.Effectof specific COX-2inhibition in osteoarthritis of the knee:a 6 week
15 AkaraserennontP, Mitchell JA, AppletonI, ThiemermannC, VaneJR. Involve- doubleblind, placebocontrolled pilot study. J Rheumatol.1999;11:2438-2447.
mentof tyrosine kinasein the inductionof cycle-oxygenaseand nitric oxide syn- 38 Rofecoxibprescribinginformation.West Point, Pa:Merck & Co Inc.
thase by endotoxin in cultured cells. Br J fharmacol. 1994;113:1522-1528. 39 LangmanM. JensenD, HarperS, et al. Lower incidenceof clinically evidentup-
16 Blanc0A, HabibA, Levy-ToledanoS, Maciouf J. Involvementof tyrosine kinases per-G1perforations,ulcersandbleedsin patientstreatedwith rofecoxib vs. non-
inthe inductionof cycle-oxygenase-2in humanendothelialcells.BochemJ. 1995; specific. cyclooxygenase inhibitors [abstract]. Gasfroenferology. 1999;116:
312:419-423. A232.

^_ 1787

02000 American Medical Association. All rights reserved.
Fiud Retention and Edema: Fkdd retention and edema have been observed in
Some s,,en,staln VIOD? rofecoxibtsbletss”d orslsusQensk$ (WADVERSE
REACflONS). VIOX! should I e usedvdtb cati?n and should be mtrodwd dtbe
!owez recommended dose in paUenb wkb fluid retentfw hypertension. or heart

Pree&ing Asthma:FX6ent? with esthma~?y have aspirin-sensitive asthma. The
use of aspidn in pabenh wim aspidn-sensdlve estbms has been $ssoclsted with
were bronchosQ&sm. which can be fatal. Since cross-reacbW$ !odudlng brqn- Wng into eccoonttbe importance of the drug to me mother.
Ns: VlOxX b widmindicated in patients witb know hype,- chospasm, behneefl es irin and other NSAlDs has been reported I$ suck aspinn- &Wic use safety and effec6Veness In pediatric QsdentSbelow tbe age of ta
s.ensdive patients. VI0 & should npt be edmYslered to Pabents wltb 6x5 form of yea-s have not been evaluated.
ssoirin sensitivitv and should be osed with ceuflon in os(ients with preexting Garbtic Us&me patients who rewed v10)3( in OA dinicsltrials. 1,455 were
65 years Of sge or older (mis included.460 who were 75 yean or older). No sob-
stanhal diiredw in safety and effeCb*?&s were obssm@ @een the sub-
jectsand younger subjects. G@ersensitrvi@j of some ddermdtwdoals cannot k
ruledout Dosageadjushnentmtbe eideily isnotnecessary: how&w tberapvw@
VIOXX should be inmsted etdw lowest recommended dose. In 1 oftbese StUdles
a6 week. double-blind. rendwnized dlniwl hid). VIOXX 12.5 mg or25 mg once
L,.rb was administered to 174 OA patients 83 years of age. The safety rohie I”
this eMe$ QOPUlatiMlMS similar to thst of younger Qabe”tStrestxd wnR VIOXX.
AOYERSER,E&TtqNs: OA np mximatety 3,@ ~etients with OA we b’eeted
With? SQPmMmstslyI,#! paants recemd IOn for 6 months or longer,
end Pmximstely BM) psbents for 1 year or lax~e[ Th,efdiowing aregaph lists
al, a M evwts. regardless of cau.sal~y accurnng I” St least !aIn of pettentS
receiviog VIOY in 9 controlled shidles oi6 weeks’ to 6 months.’ dumhon con-
ducted 1” Pabents wth OA et Me MemQeuGdiy recommended doses (12.5 mg
and 25 mg). which indoded a placebo and/or QOsr6veControf group.
patients should be instmcted to seek immediate emergency help in tie
NSAlOs. .QQQ-SSl case of an anaphyledoid reaction (see w’“““‘*“‘
end in about 2%t% In late pregnancy. VIOXX should be i woided because it may Cause Qremetore
increasing tie likelihood plasma of the ductus adedosus.
Labors,o,y 7esis. Because wws GI trect ulcersdons and bleeding can aCCor
withoot warning symptoms, physicians should monitor for signs or Symptoms of
01 hh.d,“”

The snsml eddy pmfile of VIOXX 50 m g.d. I” OA clinid tnsls of up to 6
monttx ! 476pst1erds) was similartothat,ofV POF St the recommended OAdweS
of12.5mgand25mg,g,d.,enceptforahlgherincide~~ofGIsymQtwns (abdam-
inalpsin, epi sstric pain. heartburn. naoses. and vomdmg), lowerextremlfyedema
(6.3%). and a yfzwtension (8.2%).
toyotohwdis~ In me OA ebidii~, me following Spontaneous adverse evti owned in
9.0% 10 1.9% of QabmS Wilted Wltb VfOn, ,egWdteSsof CaUSSfdy:
io eldert” or deb

~ lams, i-nger &reuo n Of NsAlD therapy. smoting. atoholism, older age. and p&r
oemnl b&h ma,“!

xmsnrxln bteoreonsncv. vIO)(x Sboold beavoided because %may wse Pm
mabJrecbsurs OfioeructrEj a*noS”S The Idlowing serbus adverse events have been reQo!ted rarely (estimated
ITIWONS: General: MO& cannot be ex ected to substitote for cortico- 41%) in patws tsMop VIOxX,,re ardless of caosalii. Cases rewted onhi I”
stem!& or toheat ~dicosteroti insufficiency. I!propt dis~ntinoetion ot,cottico- the posbnsrkebng expsnence are !n9lcdted m Italics.
steroids may lead 10exacerbation ofcorticostero~d-res onwe Less. PahentS,on S.stiWwbr webm!awlar accident. cong&x heati failure, deep veno+
QWlOngedcoticoste@d $ers~y $oold have tbelr,4 erepy iapered slowly If a mmmbosis. myowdiil infarcbon. putmona,! embolis$ inslent lwhemlc
decision IS made to disconbnoe oWcos&roids. atIe&. onsteble angina Gl:cholecysbbs, colihs, colonic makgnant nSop.ssm, duo-
The Qharmaealagic activ’w of VlOxX in redogg jnflsmmation. a@Qos+ibb den~pMomtio4, duodend ylcer, esophagea ok-+ gee@ perforabp?, g=@
fever, may diminish the utility of tb$se disgpostm SI ns m dete$ng mftiaus okzc GI bleeding, hepsfia’s, mtesdnal obstrocdon, #md!ce, pancreabbs. Hems
complicadons of presumed ooninfecbous, wmful conB,itlons. end L.wn hatic; ag!awfocytc$3 leokwada, lymphoma. inrombwtow~.
,rnrn”r?&sfa m: slla~*ld rea*orl, .mgioedeme. Nelwus L$stam e.vPbC
H~gatic Ef@rS: Bordertine elevations of 1 or mop live, tests me ocw in up to PsychMnc: Mnhsion, hallucinations. Sxin and Skin Ap~endsW:
15 /. of pabents Wing NSAIDs, and rotable elevebons of ALT or Ai T (-3 or more wars tin reams, irwlodin Gfeverw-Johnson syndrome. U‘W@i3f Svsfl:
times tbe wper limit of normal have hew repolted in aQQmximete$ 1% of acute renal Wkr.?, breast ms9lgnant neoplasm. infersbaal nepphnts, prostsbc
patients in clinical t*ls wdb AWDS. These la~orstwy sbn~rmskties may genie ininice p obl doses upto 30 mWXgpt+) mslignwd neoplasm, omliiiss~s. woisen~ng chmnic rwnltaifom.
mgmss. may remain uM.swd, or mey be tmoyt with co*meng @spy. and2-,ok,fi,e ma” exwsureatl moand 5 mod
Aare mses of severe hepstic rexbo~s, !mclod!ng pond!ce and fats fulmmsnt In t-year mntrolled dinical Wls end in extension Studies for UP to 86 FQ
(-@ Q&h treated Wim vton f0, t Y&X or longer), me adue~Q!ne~W
hep.stiUs,liver newsis. and bepsttc fs&re Some wdh fsM outcome) have been prohk was qosKsb”@ similar to mat observed in st!&s of shodr doraboa
reported with NSAlDs. In controlled dintcal ir!als of VIOXX. tie incidence of bQ,-
deriine elevations of liver tests .st doses of 12.5 end 25 mg daily was compa-
rable m the incidence observed w6J ibuprofen .s? “!I lower than that ob,seNed with
didofensc. I” Q,sWbC-W”,,o&d tdS.ts. a Q,o,W”Steb,0.5% of Qs”e”tS takl”g
rofemxib (12 5 mg or25 mg q.d.) and 0.140 ofpst~ents Wing Qlscebo had notable
aIevam of ALT or AX
A patient wiU3 symptoms and/or signs wggesdng Ii*, sfunction, or i? Thesdveneexperience profile ~nme ensigesb.studwns ecardiHmilar to
whom an abnonal bye, t& has occurred, should be montiore2’ caretul?, for v those mQOlkd in the OA studies. The follovdng~addidiooal she ,pSe
denceoftbedeveto e~ofamareseverehepatlc,reacha~whiieantheraW~~ which ooc~lm.l sl an inc,dence of X2% of pabents treateo ti $?$
VloxX Use of VI0 ET. IS not recommended in Pabents wlh mode@ or severe obswed h me postde”tel pain surgery StodW pastdental emacbon *sd*s
hepatic ~msofficiewy. If diniti signs and symptoms coMstent w~tb bver dlseese (dly =ket).
dwel orilsystem~cmsnirestabonsOccor(eg..wsinophilis. resh.etc.l.VIOXX In 110 patim treated Hith VlOxX (m-age age -65 rs) in mq posloml~
shaoi$ bsdiswtinued. pedic empry pain study, the most eommonb repoited E em ex~erms were
mstipahm, few,, and nausea
WSARE AND AlJhNNISlRAllOk VIOXX is sdminblered orally. Ttx Iwrest dose
o,VWXshauld be woghtforeech Q&nt
OA: The recommended stati” dose of VlOxX is 12.5 mg once de@. Some
eden6 may rsmb additional nefit by incrwsi~ me dose to 25 mg on:e de&
MaMgemenfofAcote Pain a@,‘Dysmen~tiee:~ ream-
mendedintbsl dose otVlOM~s50 m once daiix Sobse oent @sesshould be 50
mg once d.4 as need@. “se ofVl0 lx for more Ulsn s days in management Of
psm be not ‘a, n sb,d@d.
Ora\ Sqemhm~VIOxX Od Su ension 12.5 m@ mL or 25 mg.5 mL may be
subeM for VlOxX Tablets 12.“p mg or 25 mg, respechvely, m soy of the above
indic&ns. Sheke before using.
Formore dettWinfonatim, coosubyowMenkrand~~fUfl
Pm”g ,mmm

Nuning hfolhen;Rofe+oxib is excreted in me milk of tactetkw ratsstc?manba-
time similar to those m Plesme. There wds w inmease in Pop mortetdy and I

Merck& Co.,Inc.

Dear Doctor:

Thank you for your interest in this reprint, “Rofecoxib, a Specific Inhibitor of Cyclooxygenase 2, with
Specific Efficacy Comparable with that of Diclofenac Sodium: Results of a One-Year, Randomized,
Clinical Trial in Patients with Osteoarthritis of the Knee and Hip,” by Grant W. Cannon et al, published
in Arthritis & Rheumatism, Volume 43, Number 5, May 2000. We are pleased to provide this article to
you as requested.

This study, a randomized, double-blind, active-comparator-controlled trial in 784 adults, compared the
clinical efficacy of VIOXX’ (rofecoxib) with that of diclofenac. The study also evaluated the safety and
tolerability of VIOXX.

Patients were randomized to one of three treatment groups: 12.5mg VIOXX once daily, 25mg VIOXX
once daily, or 50-mg diclofenac three times a day. Clinical efficacy and safety were evaluated over a one-
year continuous treatment period. Patients could receive additional arthritis medication during the last six

VIOXX is indicated for relief of the signs and symptoms of osteoarthritis, the management of acute pain
in adults (see CLINICAL STUDIES), and the treatment of primary dysmenorrhea.

VIOXX is contraindicated in patients with known hypersensitivity to rofecoxib or any other component of
VIOXX. VIOXX should not be given to patients who have experienced asthma, urticaria, or other allergic-
type reactions after taking aspirin or other nonsteroidal anti-inflammatory drugs (NSAIDs). Severe, rarely
fatal, anaphylactic-like reactions to NSAIDs have been reported in such patients.

VIOXX is not a sulfonamide; therefore, VIOXX has no sulfonamide contraindication.

Serious gastrointestinal toxicity can occur with or without warning symptoms with NSAIDs.

Before prescribing VIOXX, please read the accompanying complete Prescribing Information. Thank you
for your interest in this information about VIOXX. - .,..

Very truly yours,

Gail A. Ryan
Professional Services

Enclosure: Prescribing Information for VIOXX
VIOXX is a registered trademark of Merck @ Co., Inc.

R-VIO-2609( I)-OSOO
Vol. 43 No. 5 May 2000
Copyright 0 2000 by American College of Rheumatology
Published by Lippincott Williams & Wilkins Printed in U.S.A.
-.-- _.____. ------ ---__-___


Results of a One-Year, Randomized, Clinical Trial in Patients with
Osteoarthritis of the Knee and Hip


Objective. To compare the clinical efficacy of cording to predefined statistical criteria, to that of 150
rofecoxib, a specific inhibitor of cyclooxygenase 2 (COX- mg of diclofenac per day in this l-year study. Specific
2), with that of diclofenac in patients with osteoarthritis inhibition of COX-2 provided therapeutic efficacy in OA.
(OA) and to evaluate the safety and tolerability of
rofecoxib. Nonsteroidal antiinflammatory drugs (NSAIDS)
Methods. We performed a randomized, double- are widely used in the treatment of osteoarthritis (OA)
blind, active comparator-controlled trial in 784 adults (1,2). Although NSAIDs effectively control mild-to-
with OA of the knee or hip. Patients were randomized to moderate joint pain associated with OA, their use is
1 of 3 treatment groups: 12.5 mg of rofecoxib once daily, accompanied by the risk of significant gastrointestinal
25 mg of rofecoxib once daily, and 50 mg of diclofenac 3 (GI) toxicity, including GT perforation, ulceration, and
times daily. Clinical efficacy and safety were evaluated bleeding (PUB) (3-5).
over a l-year continuous treatment period. NSAIDs act by inhibiting the synthesis of prosta-
Results. Rofecoxib at dosages of 12.5 and 25 mg glandins by the enzyme cyclooxygenase (COX) (6,7).
demonstrated efficacy that was clinically comparable to Two COX isoforms are now recognized. COX-1, which
that of diclofenac, as assessed by all 3 primary end is constitutively expressed, sustains the routine physio-
points according to predefined comparability criteria. logic function of prostaglandins, including gastric muco-
-Results from secondary end points were consistent with sal protection; COX-2 is induced chiefly in response to
those of the primary end points. There were small pathologic processes, including pain and inflammation
statistical differences favoring diclofenac for 2 of the (5-8). Prostaglandins synthesized by the inducible
end points. All treatments were well tolerated. COX-2 isoform mediate acute inflammatory responses
Conclusion. Rofecoxib was well tolerated and in animal models (9). .
provided efficacy that was clinically comparable, ac- In vitro and ex vivo assays have shown that
NSAIDs are non-isoform specific, inhibiting both the
Supported by.Mc_rck Research Laboratories, Raliway. NJ. COX-1 and COX-2 isoforms (10-16). Since prostaglan-
Grant W: Cannon. MD: Department of Veterans Affairs dins are involved in the maintenance of GI mucosal
Medical Center, and Kiniversity .of Utah, Salt Lake City: Jacques R.
Caldw~ll. MD: Gainesville, Florida;, Peter Hoc, MD: Baltimore, integrity and since only the COX-1 isoform is present in
Maryland; Barry McLean, MD: Birmingham;‘Alabama; Beth Seiden- the normal GI mucosa, the GI toxicity of NSAIDs has
berg. MD, James Boiogncsc, MStat, Elliot Ehrich. MD, Suarabh been proposed to result largely from inhibition of
. Mukhopadhyay, PhD. Brian Daniels. MD: Merck Sr Company. Inc..
Rahway. New Jersey. COX-1 activity (12,17-20). The therapeutic effects of
Address reprint requests to Brian Daniel% MD. Merck & NSAIDs may be primarily attributable to COX-2 inhi-
Company, Inc., 126 East.Linculn Avenue. RY 32-637, Rahway, NJ bition (9,21,22). Therefore, agents that specifically in-
Submitted for publication July 16. 1999; accepted in revised hibit COX-2 were developed and evaluated because of
form January 5, 2000. their potential to provide clinical efficacy comparable to

that of NSAIDs with a reduced risk of GI toxicity source of pain or disability. Patients were in functional class I,
(23-25). II. or III according to the Steinbrocker criteria (29). The study
Rofecoxib (VIOXF; Merck. Rahway, NJ) is a included 2 groups of OA patients, based on the treatment they
received for OA at the time of enrollment: those who took
specific inhibitor of COX-2 in humans. IJsing ex vivo NSAIDs and those who took acetaminophen.
human whole blood assays, rofccoxih showed dose- The NSAID group was assessedat the screening visit,
related inhibition of COX-2 activity (26). The degree of and patients who satisfied entry criteria discontinued their
COX-2 inhibition was similar to that of NSAIDs. At NSAID therapy. Following a washout period, patients’ pain
doses of 15-40 times the proposed clinical dose, rofe- when walking was assessed on a patient-reported loo-mm
visual analog scale (VAS). Patients were randomized into the
coxib had no dose-dependent inhibition of COX-1 (27). study if they had at least moderate pain when walking (40 mm)
This report describes the results of a large, ran- and a minimum increase in pain when walking (15 mm)
domized, clinical trial comparing rofecoxib, 12.5 and 25 compared with the level at screening. In addition. the physi-
mg once daily, with diclofenac sodium, SO mg 3 times cian’s assessmentof disease status had to be worse compared
daily, in the treatment of patients with knee and hip OA. with the screening level.
The acetaminophen group was randomized if at both
In this study, rofecoxib provided efficacy in OA that,
the screening and randomization visits (no acetaminophen
according to predefined statistical criteria, was clinically allowed within 12 hours of assessments),the patients reported
comparable to a high dose of the NSAID diclofenac. In at least moderate pain when walking (40 mm). In addition, the
a study using serial endoscopy for the presence of ulcers patient’s and physician’s assessmentsof disease status had to
in OA patients, rofecoxib demonstrated a GI safety be fair, poor. or very poor.
Women were postmenopausal or demonstrably non-
profile equivalent to that of placebo and significantly gravid. Patients were excluded if they had significant lenal
better than that of ibuprofen (28). Our findings, together impairment. clinically significant abnormalities on physical or
with those reported by Laine et al (28), show that in the laboratory examinations at the screening visit, positive results
treatment of OA, rofecoxib is as effective as diclofenac. on fecal occult blood testing, class III/IV angina or uncon-
and has the potential to improve the GI safety profile. trolled congestive heart failure, uncontrolled hypertension. a
stroke or transient ischemic attack withifi 2 years of study,
active hepatic disease,a history of recent neoplastic disease.or
PATIENTS AND METHODS an allergy to acetaminophen or NSAIDs. Patients were ex-
cluded if the? iequired aspirin at any dose, corticnsteroids,
All patients gave written informed consent before warfarin, or tlclopidine.
screening and enrollment in the study. The study protocol and Patients with a historv of gastroduodenal ulcer or GI
procedures were approved by the institutional review boards bleeding were allowed to participate.
for all investigative sites. The investigators who participated in Efficacy measurements and end points. Well-validated
the Rofecoxib Protocol 035 Study Group are listed in Appen- measurementsof efficacy were obtained at screening, random-
dix A. ization, and following 2, 4, 8, 12, 26, 39, and 52 weeks of
Study design. Patients were screened (screening visit) treatment. At each of these visits, the patients completed the
to ensure study eligibility. tipon confirmation of eligibility (see Western Ontario and McMaster Universities Osteoarthritis
entry criteria), patients were randomized (randomization visit) Index (WOMAC) (30), and both the patients and the physi-
by a computer-generated allocation schedule to 1 of 3 treat- cians completed an assessmentof disease status. Patients and
ment groups: rofecoxib 12.5 mg once daily, rofecoxib 25 mg physicians completed an assessment of response to therapy
once daily, OTdiclofenac SOmg 3 times daily (150 mg/day). following 2, 4, 8, 12, and 26 weeks of treatment.
Study blinding was maintained by using a matching placebo for There were 3 primary end points for this study: pain
each study medication. Patients took 3 tablets each morning when walking (100~mm VAS, which is question 1 of the
and 1 tablet at both midday and evening. Patients were WOMAC), patient’s assessment of response to therapy (S-
provided open-label acetaminophen (maximum dosage of 2.6 point scale,where 0 = none and 4 = excellent), and physician’s
gmidap) that could be taken for OA pain that was not assessmentof disease status (5-point scale, where 0 = very
adequately controlled by the study medication. poor and 4 = very well). All 3 end points were used to
Patients returned to the study center following 2, 4, 8, determine clinical and statistical comparability, as described in
12. 19, 26, 33. 39, 45, and 52 weeks of therapy to assessboth the statistical section (see below).
efficacy and safety. Patients who did not enter a voluntary Other end points were patient’s assessmentof disease
extension at the end of the l-year treatment period returned status (loo-mm VAS, where 0 = very well and 100 = very
7-10 days after their last dose of study medication for post- poor), physician’s assessmentof response to therapy (S-point
therapy safety assessments. scale, where.0 = none and 4 = excellent), WOMAC subscales
Entry criteria. Patients were a minimum of 40 years of Pain, Stiffness, and Functional Ability (100-mm VAS),
old and had both clinical and radiographic eviderice of OA. study-joint tenderness (O-3 scale, where 0 = no pain and 3 =
Patients with OA of- the knee or hip we’re eligihle for study. patient states that there is pain; winces and withdraws), and
Radiographic criteria for OA of the knee were joint space amount of rescue acetaminophen consumed (number of
narrowing and the presence-of osteophytes: the radiographic 325mg tablets).
criterion for 0.4 of the hip was joint space narrowing. The Safety assessments. Spontaneously reported adverse
study joint (either the knee or the hip) had to be the primary experiences were recorded throughout the study. Vital signs

Table 1. Baseline characteristics of the patients, according to treatment group*
12.5 mg 25 mg 150 mg Total
Characteristic (n = 259) (n = 257) (n = 26X) (n = 784)
Female sex, no. (%) 169 (65.3) 175 (68.1) 185 (69.0) 529 (675)
Race. no. (c/o)
White ’ ’ 236 (91.1) 229 (89.1) 237 (88.4) 702 (89.5)
African American 19 (7.3) 23 (8.9) 23 (8.6) 65 (8.3)
Other 4 (1.5) 5 (1.9) 8 (3.0) I7 (2.2)
Age, mean 2 SD years 62.8 _f 10.2 62.8 2 10.3 62.5 _f 10.1 63.6 -t 10.2
Weight, mean 2 SD kg 92.4 t 22.2 87.9 -c 19.6 88.0 z 21.0 89.4 i- 21.0
Duration of OA. mean 2 SD years 11.1 ir 8.9 11.5 t 8.7 11.4 t 9.4 8.7 _c 9.0
Functional class. no. (‘6)
Class I 31 (12.0) 39 (15.2) 38 (14.2) 109 (13.9)
Class II 173 (66.8) 176 (68.5) 168 (62.7) 5 I7 (65.‘))
Class III 54 (20.8) 42 (16.3) 62 (23.1) 1.58(20.2)
Study joint, no. (‘h)
Hip 61 (23.6) 68 (26.5) 61 (22.8) 190 (24.2)
Knee 198 (76.4) 189 (73.5) 207 (77.2) 594 (75.8)
Previous OA medication (%)
NSAIDs 240 (92.7) 238 (92.6) 242 (90.3) 720 (91.8)
Acctaminophcn 19 (7.3) 19 (7.4) 26 (9.7) 64 (8.2)
Primary outcome nicasure?
Pain when walking (WOMAC). 75.9 2 15.0 77.5 t 14.7 75.8 + IS.4 76.4 2 15.0
O-l OO-mm VAS
Physician’s asscssmcnt of disease 2.9 t_ 0.7 2.9 Ic_0.7 3.0 i 0.7 2.‘) -i- 0.7
status. O-4 Likert scale

* Duration of osteoarthritis (OA) was dctcrmined by patient report. Functional class was dctcrmincd
according to the Stcinhrocker criteria. NSAIDs = nonstcroidal antiinflammatory drugs: WOMAC =
Wcstcrn Ontario and McMastcr Universities Osteoarthritis Index; VAS = visual analog scale.
t The third primary outcome mcasurc, patient’s assessmentof response to therapy, was not examined past
week 26 and dots not have basclinc (randomization) values hccausc it asscsscd the patient’s rcsponsc to

were monitored at every visit. Laboratory investigations. in- daily. As a comparability trial, specific predefined criteria were
cluding hematology, blood chemistry. and a urinalysis. were established. Clinical comparability was declared if the follow-
performed at all visits. For all clinical adverse experiences. the ing criteria were met: for all 3 primary end points. the 95%
investigator recorded the intensity. the relation to test drug. confidence intervals (95% Cls) of the difference in the mean
the outcome, and any action taken. The investigator also treatment response between 2 treatments were within ? 10 mm
determined if a laboratory adverse event was study-drug on a lOO-mm VAS and +O.S on a Likert scale. These clinical
related. comparability bounds are more conservative than those pro-
Statistical analysis. This study tested the hypothesis posed by a consensus of academic rheumatologists and em-
that rofecoxib, 12.5 mg and 2.5 mg once daily, would have ployed in a study comparing meloxicam with diclofenac
clinical efficacy comparable to that of diclofenac. 50 my 3 times (31.32). This study had >99% power to demonstrate compa-

Table 2. Numbers of patients who cntcred, completed. and discontinued the study. according to
treatment group
Study status 12.5 mg- 25 mg I50 mg Total
Entered the study, no. of patients 255, 257 268 784
Completed the study. no. (c/) 161 (62.2) 142 (55.3) 145 (54.1) 44x (57.1)
Discontinued the study, no. (% ) 98 (37.8) 1 IS (44.7) 123 (45.9) 336 (42.Y)
Clinical adverse experience 37 (14.3) -32 (12.5) 41 (15.3) 1 IO (14.0)
Laboratory adverse cxpcricnce 1 (0.4) 2 (0.8) 14 (5.2) 17 (2.2)
Lack of efficacy 36 (13.9) 56 (21.8) 43 (16.0) 135 (17.2)
Deviation from protocol IO (3.9) 12 (4.7) 11 (4.1) 33 (4.2)
Patient withdrew consent 9 (3.5) 9 (3.5) I I (4.1) 2’) (3.7)
Other” 5 (1.9) 4 (1.6) 3 (1.1) 12 (1.5)
* Includes patients who moved and patients who were lost to followup.

rable efficacy (according to the criteria cited) bctwcen 25 mg of
rofecoxih and diclofenac if their true difference is 0.
For the determination of comparability. the .3 primary -5
end points were analyzed as the averaged response over the -10
52-week treatment period (first 26 weeks only for patient’s -15
assessment of response to therapy). All data collected from
discontinuation and unscheduled visits were included in this
analysis; no missing values were imputed. The comparability -25
analysis was also performed on data from the first 12 weeks -30
and the first 26 weeks. -35
The responses of primary and secondary end points
were analyzed using an analysis of covariance model. with -40
treatment, study center, and history of ulcer or upper GI
bleeding as main effects. and basclinc as the covariate. For end
points without basclinc mcasurcmcnts (i.c., patient’s/ SR24 8 12 26 39 52
physician’s assessment of response to therapy), the baseline
value of a relevant variahlc (i.e.. patient’siphysician’s asscss-
ment of disease status) was used as the covariatc in the model.

% -1.5 -
Between November 1996 and April 1997, 1,128 5
patients were screened, and 784 (69.5%) were enrolled 3 -2.0 -
into the study. Patients not randomized were excluded s
$ -2.5
for a variety of reasons, including failure to meet OA
diagnostic criteria (13X?), abnormalities found on -3.0 -
screening physical or laboratory examinations (12.1%), -3.5 1 8 ’ j 1
failure to satisfy randomization OA activity criteria SR24 8 12 26
(1.3%), and their reconsideration of participation in the c 0.0
study (4.2%). All treatment groups had similar haselinc _ -0.2
characteristics and primary efficacy outcome measures G -0.4
at enrollment (Table 1). All randomized patients with s
!$, -0.6
OA of the knee and 96% of those with OA of the hip .z -0.8
fulfilled the American College of Rheumatology classi-
& -1.0
fication criteria for OA of those regions (33,34).
is -1.2
A total of 44X of the 784 patients (57. I %) com- 6~ -1.4
pleted 1 year of study therapy (Table 2); the overall
discontinuation incidence was similar among treatment .iJ -1.6
.grpups. There were no statistically significant differences
in the incidence of discontinuation because of lack of

SR24 8 12 26 39 52
efficacy of the study therapy or clinical adverse experi- Week
ence among the treatment groups. The increased discon- a 12.5 mg Rofecoxib V 25 mg Rofecoxib 0 150 mg Diclofenac
tinuation rate because of laboratory adverse experiences
Figure 1. Treatment response over time for the 3 primary clinical
in the diclofenac group was due to elevations in serum efficacy end points: A, pain when walking (baseline mean 76.4 mm on
transaminases, as discussed later in the Results (see a 100-mm visual analog scale). B, patient’s assessment of responsk to
below). therapy (not assessed at baseline visit), and C, physician’s assessment
Efficacy. Figure I presents the response ‘over of disease status (baseline mean 2.9.on a 5-point scale, where 0 = very
.-. .. ‘..’ time fo’r all-3 piimary end points. Both pain when poor and 4 = very well). Scales in A ind C were normalized to the
randomization mean; the scale in B was inverted for consistency with
.-.,: . - waiking’and physiditin’s-assessment Df disease status, the
other end points. On all graphs, decreasing values indicate improve-
. 2 end poiuts evaluated at baseline (randomization), ment. S = screening visit; R = randomization visit; 95% CI = 95%
demonstrated significant improvements from baseline confidence interval.
for all treatment groups. The mean response for the
.._. primary end point of patient’s assessment of response to have baSeline values because it assessed the patients’
therapy was similar among all treatment groups. This response to therapy.
end point was not examined past week 26 and does not For all primary end points, treatment responses

Table 3. Comparability analysis for the 3 primary end points over I
year or 6 months nf trcatmcnt. as indicated*
Comparing Comparing
2.5 mg nf 12.5 mg of
rofccoxih with rofccoxih with
150 mg of 150 mg of
Primary end point diclofenac diclofcnac

Pain when walking (WOMAC I .98 ( - I .6h. 5.62) 1.81 (- I .xs. 5.44)
question I). O-IOf)-mm
Patient’s assessment of 0.19 (n.ns, 0.33) 0.24 (0. IO. 0.38)
response to therapy. O-4
Likert scalet
Physirian’s assessment of 0.17 (0.05. 0.29) 0.13 (O.Ol.ft.25)
disease status, O-4 I.ikcrt

* Comparability was defined as the difference in the lcasl squares
mean(95’6 confidence interval). and must hc within 2 IO mm on the
visual analogscale (VAS) and ~0.5 units on the Likert scale. Positive
mean diffcrcnccs favor diclofcnac over rofccoxih. WOMAC = West-
ern Ontario and McMastcr Univcrsitics Osteoarthritis Index.
f Data for this end poinl were not collected after week 26: therefore.
the analysis concerns the first 26 weeks of treatment.

were seen within 2 weeks (first time point measured) for
all treatment groups. The treatment responses were
sustained throughout the entire year of treatment (or 26 0.0
weeks for patient’s assessment of response to therapy) at
a generally consistent level.
The primary hypothesis of this study was that -0.4
rofecoxib would provide comparable clinical efficacy to
that of diclofenac in the treatment of OA. To test this
hypothesis, comparability criteria for the 3 primary end -0.8
points were prespecified (as discussed in Patients and
Methods). The difference in the mean treatment re-
sponse between 2 treatments is calculated as treatment -1.2
A minus treatment B, and this difference has an associ-
-1.4 1 ’ ’ ’
ated 9.5% CI. The 2 treatments would be considered SR24 8 12 26
clinically comparable if the 95% CI of the difference Week
does not extend beyond the predefined bound of ~10 m 12.5 mg Rofecoxib v 25 mg Rofecoxib 0 150 mg Diclofenac
mm on the VAS and 20.5 on the Likert scale.
Figure 2. Treatmenr responses over time for the 3 secondary end
Table 3 presents the comparability data. For all 3 points: A, Physical Function subscale of the Western Ontario and
primary end points, the 95% Cls for the difference in the McMaster Universities Osteoarthritis Index (WOMAC) (haseline
mean treatment response for each of the treatment pairs mean 69.6 on a 100-mm visual analog scale [VAS]). B. Stiffness
(25 mg of rofecoxib and diclofenac; 22.5 mg of rofecoxib subscale of the WOMAC (baseline mean 72.9 on a l(N)-mm VAS). and
and diclofenac) were within the predefined comparabil- C, study-joint tcndcrness (baseline mean 2.0 on a O-3 scaE:\iihfre 0 =
no pain and 3 = patient states that there is pain; winces and
ity bound. Thus, both 12.5 and 25 mg of rofecoxib withdraws). Study-joint tenderness was not assessed after week 26.
demonstrated clinical efficacy comparable to that of 150 Scales were normaljz~d to the randomization mean. On all graphs.
mg diclofenac over 1 year of continued treatment. The decreasing values indicate imprtrvemcnt. S = screening visit: R =
same conclusion was reached when comparability was randomization visil: 95% CI T 9S”r confidence interval.
analyzed for the first 12 weeks or the first 26 weeks of
There were small differences favoring diclofenac statistical significance: patient’s assessment of response
compared with rofecoxib for 2 end points that reached to therapy and physician’s assessment of disease status.

Table 4. Summary of won&u-y efficacy end points over the 12-month treatment period”

Diclofenac. 150 mg
Efficacy end pirinl 12.4 mg (n = 259) 25 mg (n = 257) (n = 268)
Pain suhxcalc (WOMAC). (I-IOO-mm VAS -26.7 (-29.5. -23.9) -27.3 (-30.1. -24.42) -29.6 (-32.4, -26.8)
Physical Function suhscalc (WOMAC). (I-IOO-mm VAS -23.4 (-26.4, -20.4) -23.8 (-26.8, -20.7) -25.8 (-28.9, -22.R)
Stiffness subscale (WOMAC). 0-IOO-mm VAS -24.5 (-27.7, -21.3) -25.2 (-28.4, -22.0) -27.7 (-30.9, -24.5)
Patient’s assessment of disease status, O-IOO-mm VAS -28.5 (-31.7, -25.3) -27.1 (-30.3, -23.9) -31.5 (-34.7, -28.3)
Physician’s asscssmcnt of rcsponsc to therapy. O-4 Likert scale? -2.5 (-2.66, -2.43) -2.5 (-2.61, -2.39) -2.8 (--2.9. -2.6)
Study-joint tendcrncss. O-3 Likcrt scale -1.1 (-1.2. -1.0) -1.2(-1.X -1.1) -1.1 (-1.2. -1.0)
Acctaminophcn USC(for rcscuc). tahlctsiday 0.8 (0.7,0.9) 0.X (0.7, 0.9) 0.7 (

* Values are the least squares mean (c)S? confidcncc intctval). rcprcscnting the mean change from the time of randomization. Negative values
indicate improvcmcnt in the end point compared with randomization (visit 2). except for acctaminophen use. See Table 1 for definitions,
‘? Data for this end point wcrc not collected after week 26: thercforc, the analysis concerns the first 26 weeks of treatment.

However, these differences and their respective 95% CIs tively). Patients discontinued because of cardiovascular
were within the comparability bounds predefined for events at a similar incidence (2.3%, 3.1% and 3.7% in
this study. the 12.5mg rofecoxib, 25-mg rofecoxib, and diclofenac
OA is a complex disease with multiple clinical groups, respectively). The most frequent individual ad-
manifestations. Secondary end points were collected to verse experiences resulting in discontinuation were diar-
assess the response to treatment in a variety of domains. rhea, dyspepsia, epigastric discomfort, and myocardial
The secondary end points of the WOMAC Stiffness infarction (Table 5). No single adverse experience ac-
subscale, WOMAC Physical Function subscale, and counted for discontinuation in >2 patients per treat-
study-joint tenderness (Figure 2) demonstrated signifi- ment group.
cant changes from baseline in all treatment groups. A total of 2, 2, and 3 patients in the 12.5mg
Results for additional secondary end points were con- rofecoxib, 25mg rofecoxib, and diclofenac groups, re-
sistent with the finding of clinical comparability among spectively, experienced a symptomatic gastric or duode-
treatment groups (Table 4). nal ulcer. There were no episodes of GI bleeding in this
The consistency of the treatment effects of rofe- study.
coxib and diclofenac among patients of various sub- The incidence of drug-related lower extremity
groups was compared. Treatment-by-factor analysis for
the 3 primary end points showed that there was no
Table 5. Summary of advcrsc experiences”
statistically significant interaction with treatment for -
various subgroups. including location of the study joint Rofecoxib
(knee or hip), previous OA medication (NSAID or 12.5 mg 25 my I50 mg
acetaminophen), age, and sex. (n = 259) (n = 257) (n = 268)
Safety. All safety data are reported for the entire
Any clinical adverse event 86.9 84.0 86.2
l-year treatment period. The incidence of each clinical Any drug-related clinical 30.9 30.4 32.5
adverse event and drug-related (as assessed by the advcrsc eventi
investigator) adverse event was similar among the treat- Most frequent adverse went
Upper respiratory infection 23.9 25.7 17.9
ment groups. The most frequent adverse events were Sinusitis 8.9 1.4 7.1
upper respiratory infection and sinusitis: the most fre- Most frcqucnt GI adverse
quent GI adverse events were nausea, diarrhea. and Nausea 7.4 9.7
heartburn (Table 5). The differences in incidence of GI Diarrhea 6.9 12.1 10.4
adverse. events were not statistically significant. Heartburn 5.4 5.1 3.0
Any laboratory adverse event 14.4 18.4 27.4
The incidence of patients who discontinued the Discontinuation due to
study because of clinical adverse events was similar adverse event
among the 3 treatment groups (Table 2 j. The majority of Diarrhea 0.4 0.8 0.4
Dyspepsia 0.4 0.8 0.0
discontinuations were because of adverse experiences Epigastric discomfort 0.8 0.0 0.7
related to the G[ or cardiovascular systems. Patients Myocardial infarction 0.4 0.4 0.7
discontinued because of -GI symptoms at a similar
* Values are percentages of patients. GI = gastrointestinal.
incidence (4.6%. 3.1%, and 3.7% in the 12.5mg rofe- + Determined by the investigator to be possibly. probably, or definitely
coxib, 2%mg rofecoxib. and diclofenac groups, respec- medication related.

Table 6. Blood prcssurc and swum crcatininc lcvcls over time*
Variahlc. trcatmcnt group Screening visit Week 12 Week 2h Week 52
Systolic blood prcssurc (mm F{g)
Rofccoxih 12.5 13h.7 _f 16.5
Rofccoxih 25 mg 136.2 i 15.6 134.4 t 17.3 136.8 I 17.4
mg 135.(, i 1.4 136.9 2 16.7 133.9 i 14.9 136.8 2 17.0
Diclofcnac 151)
mg 136.4 t 16.3 134.9 2 14.7 134.3 t- 15.4 133.7 t 16.3
Diastolic hlood
Rofccoxib 12.5 prcssurc (mm Hg)
Rofccoxih 25 mg X0.8 -e 7.x 79.9 i- 8.2 80.1 t- 8.2 80.2 _t 8.0
Diclofcnac 150mg 80.4 +- 8.4 81 .I) -+ 9.4 79.5 t 9.0 x1.5 +- 9.6
Swum crcatininc (m&d]) 81.0 i- x.7 79.9 F 9.3 79.1 -c 9.7 79.x t 8.9
Rofccoxih 12.5 mg- 1.14 t 0.2 I.14 2 0.2 1.11 -c 0.2
Rofccoxih 25 1.1s 2 0.2
mg 1.13 + 0.2 1.16 2 0.2 I.16 t 0.2 I.13 2 0.2
Diclofcnac 150 mg I. 13 _f 0.2 I.15 2 0.2 1.15 r 0.2 1.10 2r 0.2
* Values arc the mean t SD.

edema (reported by the patient) over the year of treat- peripheral arterial occlusions, was numerically lower in
ment was similar among the treatment groups (3.9%, the rofecoxib groups (1.5%. 2.3%, and 3.4% in the 12.5-mg
1.9%, and 3.4% in the 125mg rofecoxib, 25-mg rofe-
coxib, and diclofenac groups, respectively). The clinical
significance of these events was minor, since over the
entire i-year duration of the study, only a single patient
( 12.5mg rofecoxib group) discontinued therapy because
of lower extremity edema and most cases resolved with
continuation of treatment. There were 4 episodes of
congestive heart failure: 1 in the 12.5-mg rofecoxib
group and 3 in the diclofenac group. There were no
meaningful changes in blqod pressure or serum creati-
nine levels over the year of treatment among the 3
groups (Table 6). No patient had a clinical episode of
acute renal failure.
There were more laboratory adverse events in the
diclofenac group compared with the rofecoxib groups,
largely due to a greater incidence of increased serum J
SA2 4 8 12 19 26 33 39 45 52
aminotransferase levels. The diclofenac group had pro-
nounred mean changes in alanine and aspartate amino-
transferase levels compared with the rofecoxib groups
(Figure 3). These elevations caused 11 diclofenac pa-
tients (4.1%) to discontinue therapy, compared with
none of the patients in the rofecoxib groups.
Because of its action in inhibiting the function of
platelets. prolonged therapy with low-dose aspirin re-
duces the risk of thromboembolic cardiovascular events
(35). Although a similar epidemiologic case has not been
made for NSAIDs, there has been a theoretical concern
that specific inhibition of COX-2, which does not effect
ptatelet function, may not protect the cardiovascular
-30 ‘a 8 ’ 8 8
system to the same extent as NSAIDs. In this l-year $2 4 8 12 19 26 33 39 45 52
study that included patients with cardiovascular risk Week
factors (hypertension in 45%, angina in 3%, hypercho- H 12.5 mg Rofecoxib v 25 mg Rofecoxib 0 150 mg Diclofenac
lesterolemia in lf%, and diabetes in 7%). the incidence
of thromboembolic cardiovascular events, such as myo- Figure 3. Aminotransfcrasc values over time, cxpresscd as the geometric
mean pcrccntagc of change from hascline. A, Alanine aminotransfcrase:
cardial infarction, stroke, transient ischemic attack, and B, aspartatc aminotransferasc. See Figure I for definitions.

rofecoxib, 2%mg rofecoxib, and diclofenac groups. respec- NSAID widely accepted HS efficacious in the treatment
tively). of OA.
All studies are sub,jcct to dropouts, which affect
the interpretation of the data. The 43% incidence of
discontinuation for this l-year study was less than that of
The discovery of 2 isoforms of COX. the target a long-term study of OA comparing naproxen with
enzyme inhibited by NSAIDs. poses a number of ques- acetaminophen (39). When adjusted for duration of
tions concerning the role of COX-I and COX-2 in the exposure, it is lower than the 30-345: discontinuation
efficacy and safety of this widely prescribed class of rate for l2-week OA studies without placebo controls
drugs. Previous studies have demonstrated that specific (40,41). The overall incidence of discontinuation was
COX-2 inhibitors are efficacious in the treatment of OA. similar among all treatment groups. The effects of
but did not answer the question as to how that efficacy dropouts on the results were minimized by employing an
compares with the efficacy of NSAIDs (36). intention-to-treat analysis and by not imputing values for
In this study. the efficacy of rofecoxib was com- patients who discontinued.
parable to that of a high dose of the NSAID diclofenac. Overall. during I year of treatment. all treat-
Predefined clinical comparability criteria were used for ments were generally well tolerated. It is important to
the primary analysis because the scientific question of note that no adverse event unique to the specific inhi-
interest was whether rofecoxib had clinical efficacy bition of COX-2 was observed in the rofecoxih treat-
comparable to that of an NSAID in the treatment of ment groups. The overall incidences of adverse events
OA. Small statistical differences favoring diclofenac and discontinuations for clinical adverse events were
compared with 25 mg of rofecoxib were seen for 2 of the similar among the treatment groups.
end points: ~0.20 Likert units for patient’s assessment The hypothesis of improved GI safety and toler-
of response to therapy and physician’s assessment- of ahility for inhibitors that are specific for (‘OX-2 cannot
be answered with the results of a single trial. Early
disease status. These differences and the associated 9SV;
published results for rofecoxih and celecoxih demon-
CI were well within the clinical comparability bounds
strate an improved GI safety profile compared with
prespecified for this study. These bounds are more
NSAIDs (42,43). An endoscopic study of OA patients
conservative than those recommended by a panel of
confirmed the significantly improved GI safety profile of
expert rheumatologists (31). In addition, 0.2 Likert units
rofecoxih in comparison to standard NSAID therapy
is markedly smaller than the recently determined mini-
(28). In that study, patients who received rofecoxib (25
mal perceptible clinical improvement of 0.5 units for
and 50 mg) had significantly fewer endoscopic gastrodu-
these end points in OA studies (37). Thus, both the
odenal ulcers compared with those who received ibupro-
12.5mg and the 2%mg doses of rofecoxih are clinically
fen (2.4 gm) over 6 months of treatment (9.6%. 14.7%.
comparable by the strict, prespecified comparability and 45.8% for the 2S-mg rofecoxih, SO-mg rofecoxih, and
criteria to diclofenac for all 3 primary end points. ibuprofen groups, respectively). In addition, the inci-
In addition, the effects of rofecoxib on a variety dence of endoscopic gastroduodenal ulcers in the rofe-
of the clinical manifestations of OA. as assessed by coxih group was equivalent to placebo for the 12-week
secondary end points, were comparable to those of placebo treatment period.
diclofenac for all end points. These results were oh- In the present l-year study, there were fewer
tained in a representative population of OA patients, symptomatic ulcers in the combined rofecoxib groups
and the results were consistent across study joint, age. (0.8%) compared with the diclofenac group (1.2%). The
and sex. numhers of patients in the study were too small to
There are several potential limitations f?~ this support a conclusion of a decrease in the incidence of
study. No placebo group was included in the study PUB events. A combined analysis of all OA clinical
because of the inability to maintain patients who have studies has been performed and demonstrated a statis-
painful OA symptoms oti a regimen of placebo for 1 tically important decrease in PUBS for rofecoxib-treated
year. The treatment responses in this study were mark- patients compared with NSAID-treated patients (44).
edly similar to those of rofecoxib groups in other Thus, based both on the endoscopy data and the analysis
placebo-controlled trials that clearly demonstrated sig- of clinical PUB events, rofecoxih appears to have a
nificant..djfferences compared with place-ho (36,38). In meaningful improvement in GI safety compared with
addition, the respOnses seen with rofecoxib were com- NSAIDs.
parable-to those seen with high doses of diclofenac, an Treatment with rofecoxib for 1 year did not have

an effect on serum aminotransferase levels. In contrast, 5. Richardson C. Emery P. The clinical implications of inhibition of
the inducible form of cycle-oxygcnasc. Drug Saf 1996; 15:249-60.
the elevations in serum aminotransferase levels in the 6. Vane JR. Batting RM. New insights into the mode of action of
diclofenac group were consistent with the published anti-inflammatory drugs. Inflamm Rcs 1995:44:1-10.
experience (45). 7. DC Brum-Fcrnandcs AJ. New pcrspectivcs for nonstcroidal anti-
inflammatory therapy. J Rhcumatol 1997;24:246-8.
The most common renal effects of NSAIDs at- 8. Donnelly MT. Hawkcy CJ. COX-II inhibitors: a new generation of
tributable to the inhibition of COX are a reduction in safer NSAIDs? Aliment Pharmacol Ther 1997; 11:227-36.
glomerular filtration rate (GFR) and reductions in the 9. Masferrer JL, Zweifel BS, Manning PT, Hauscr SD, Leahy KM,
Smith WG. ct al. Selective inhibition of inducible cyclooxygenase
excretion of sodium, with the potential for fluid reten- 2 in vivo is antiinflammatory and nonulccrogenic. Proc Nat! Acad
tion and edema. The intrarenal distribution and regula- Sci U S A 1994;91:3228-32.
IO. O’Ncill GP. Mancini JA, Kargman S, Ycrgey J. Kwan MY,
tion of renal COX-2 by sodium intake suggests a role for Falgucyrct JP, et al. Overexprcssion of human prostaglandin G/H
this enzyme in renal physiology and in the renal effects synthase-1 and -2 by recombinant vaccinia virus: inhibition by
of NSAIDs (46,4’7). It has been previously shown that nonstcroidal anti-inflammatory drugs and biosynthesis of 15-
hydroxycicosatctracnoic acid. Mel Pharmacol 1994:45:245-54.
the acute (24-48 hours postdose) sodium-retaining ef- 1 I. Laneuvillc 0, Breucr DK. Dewitt DL, Hla T. Funk CD. Smith WL.
fect of 50-mg rofecoxib is comparable to that of the Differential inhibition of human orostaglandin cndo-ocroxidc H
synthases-I and -2 by nonstcroidal anti-inflammatory drugs.
NSAID indomethacin (48). This effect resolves over the J Pharmacol Exo Thcr 1994;271:927-34.
14 days of treatment with rofecoxib, in contrast to the 12. Mcadc EA. Smith WL. dcWitt DL. Differential inhibition of
persistence of this effect with indomethacin. In addition, prostaglandin cndopcroxidc synthase (cyclooxygcnasc) isozymes
by aspirin and other non-stcroidal anti-inflammatory drugs. J Biol
rofecoxib did not significantly affect the GFR (48). In Chem 1993;268:6610-4.
this l-year study, the renal effects of rofecoxib were 13. Kargman S, Charlcson S. Cartwright M, Frank J, Ricndcau D,
similar to those of diclofenac, as assessed by spontane- Mancini J. et al. Characterization of urostaelandin
. ., G/H svnthasc
. I
and 2 in rat, dog. monkey, and human gastrointestinal tracts.
ous reports of lower extremity edema. Most of these Gastrocntcrology 1996:111:445-54.
events resolved while continuing study therapy, and few 14. Berg J. Christoph T. Widerna M, Bodentcich A. Isoenzyme
patients discontinued treatment because of these events. specific cyclooxygcnase inhibitors: a whole cell assay system using
the human erythrolcukemic cell line HEL and the human mono-
There -were no significant effects on the mean diastolic cytic cell lint mono mat 6. J Pharmacol Toxicol Methods 1997;37:
or systolic blood pressure or on serum creatinine levels. 179-86.
15. Wong E, DeLuca C. Boily C, Charlcson S, Cromlish W. Denis D,
- In summary, the specific inhibition of COX-2 ct al. Characterization of autocrine inducible prostaglandin H
with rofecoxib at a dosage of 12.5 mg and 25 mg once synthase-2 (PGHS-2) in human osteosarcoma cells. Inflamm Rcs
daily provided comparable clinical efficacy to that of 1997;46:5 l-9.
16. Miralpcix M, Camacho M, Lopez-Belmontc J. Canalias F, Bcleta
diclofenac 50 mg 3 times daily in the treatment of OA of J. Palacios JM. ct al. Selective induction of cycle-oxygenase-2
the knee and hip. Rofecoxib was generally well tolerated. activity in the pcrmancnt human cndothelial ccl1 lint HUV-EC-C:.
biochemical and pharmacological characterization. Br J Pharma-
col 1997;121:171-80.
ACKNOWLEDGMENTS 17. Langman MJS. Wcil J. Wainwright P, Lawson DH. Rawlins MD,
Logan RFA. et al. Risks of hlecding peptic ulcer associated with
We thank the investigators who participated in the individual non-steroidal anti-inflammatory drugs. Lancet 1994;
Rofecoxib Protocol 035 Study Group. We also thank Mary 343: 1075-8.
Carson, John Benedetto, Kathryn Lunga, James Halsey, Jo- 18. Fries JF. Assessing and understanding patient risk. Stand J Rhcu-
matol Suppl 1992:92:21-4.
vanni Gonzalez, and Jill-Ann Bardi Warren for their essential
19. Gabriel SE. Jaakkimainen L. Bomhardier C. Risk for serious
contributions to the project. gastrointestinal complications related to USCof nonsteroidal anti-
inflammatory drugs: a meta-analysis. Ann Intern Med 1991;llS:
REFERENCES 20. Bakhle YS. Botting RM. Cyclooxygenasc-2 and its regulation in
inflammation. Med Inflamm 1996;5:305-23.
1. Lane NE. Pain management in osteoarthritis: the role of COX-2 21. Seibcrt K, Masferrcr JL. Role of inducible cvclooxvgenase
inhibitors. J Rhcumatol 1997;24 Suppl 49:20-4. (COX-2) in inflammation. Receptor 1994;4: 17-23. . ’-
2. Brandt KD. Osteoarthritis. In: Isselbacher KJ. Braunwald E, 22. Chan CC, Bovce S. Bridcau C. Ford-Hutchinson AW. Gordon R,
Wilson J, Martin J, Fauci A, Kasper D, editors. Harrison’s Guay D. et al. Pharmacology of a selective cyclooxygenase-2
principles of internal medicine. 12th ed. New York: McGraw-Hill; inhibitor. L745.337-a novel nonstcroidal anti-inflammatory agent
1991. p. 1475-9. with an ulccrogcnic sparing effect in rat and nonhuman primate
3. Fries JF. NSAID gastropathy: the second most deadly rheumatic stomach. J Pharmacol Exp Ther 1995;274:1531-7.
disease? Epidemiology and risk appraisal. J Rheumatol 1991;18 23. Vane JR, Botting RM. New insights into the mode of action of
Suppl 28:6-10. anti-inflammatory drugs. Inflamm Rcs 1995:44:1-IO.
4. Garcia Rodriguez L,A, Cattaruzzi C, Troncon MG, Agostinis L. 24. Griswold DE, Adams JL. Constitutivc cyclooxygenase (COX-1)
Risk of hospitalization for upper gastrointestinal tract bleeding and inducible cyclooxygenasc (COX-2): rationale for selective
associated with ketorolac. other nonsteroidal anti-inflammatory inhibition and progress to date. Med Res Rev 1996;16:181-206.
drugs, calcium antagonists, and other antihypertensive drugs. Arch 25. Prasit P. Riendeau D. Selective cyclooxygcnase 2 inhibitors. Annu
Intern Med 1998;15&:33-9. Rep Med Chem 1997;30: 179-88.

26. Brideau C, Kargman S. Liu S, Dalloh AL. Ehrich EW. Rodger IW, Schwartz BD, ct al. Preliminary study of the safety and efficacy of
et al. A human whole blood assay for clinical evaluation of SC-58635, a novel cyclooxygenase 2 inhibitor: efficacy and safety in
biochemical efficacy of cyclooxygenase inhibitors. Inflamm Res two placebo-controlled trials in osteoarthritis and rheumatoid
1996;45:68-74. arthritis, and studies of gastrointestinal and platelet effects. Ar-
27. Ehrich E. Dalloh A. van Hecken A. DcprC M, DcLcpelcirc I, thritis Rhcum 1998;41:1591-602.
Bridcau C, et al. Demonstration of selective COX-2 inhibition by 43 Lanza FL. Rack MF. Simon TJ. Quan H, Bolognese JA, Hoover
MK-966 in humans [abstract]. Arthritis Rheum 1996:39 Suppl ME. et al. Specific inhibition of cyclooxygcnase-2 (COX-2) with
9:S81. MK-0966 is associated with less gastroduodcnal damage than
28. Lame L, Harper S, Simon T. Bath R. Johanson J. Schwartz H. et either aspirin or ibuprofen. Aliment Pharmacol Ther 1999;13:
al. A randomized trial comparing the effects of rofecoxib. a 76 1-7.
cvclooxveenase 2-soecific inhihitor. with that of ihuorafen on the 44. L,angman M. Jensen D, Watson DJ, Harper S, Zhao P. Bolognese
gastrodu~denal mucosa of patients with ostcnarthri&. Gastrocn- J, et al. Adverse upper gastrointestinal effects of rofccoxih com-
terology 1999:117:776-83. pared with NSAIDs. JAMA 1999;282: 1929-33.
29. Steinbrocker 0, Traeger CH, Batterman RC. Therapeutic criteria 45. Purcell P. Henry D, Melville G. Diciofenac hepatitis. Gut 1991;
in rheumatoid arthritis. JAMA 1949;140:659-62. 32:1381-5.
30. Bellamy N. Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. 46. Guan Y, Chang M, Cho W, Zhang Y, Redha R. Davis L, et al.
Validation study of WOMAC: a health status instrument for
Cloning, expression, and regulation of rabbit cyclooxygenasc-2 in
measuring clinically important paticnl relevant outcomes to anti-
renal medullary interstitial cells. Am J Physiol 1997;273:Fl8-26,
rheumatic drug therapy in patients with osteoarthritis of the hip or
47. Yang T. Singh I, Pham A, Sun D. Smart A. Schnermann J. et al.
knee. J Rheumatol 1988;15:1833-40.
Differential response of renal cortical and inner mcdullary
31 Bellamy N. Carette S, Ford PM, Kean WF. IeRiche NGH, Lussier
A. et al. Osteoarthritis antirheumatic drug trials. IIJ. Setting the cyclooxygenasc-2 (COX-2) expression to dietary salt [abstract].
delta for clinical triahs-results of a consensus development J Am Sot Ncphrol 1997;8:S621.
(Delphi) exercise. J Rheumatol l992:19:451-7. 48. Catella-Lawson F. McAdam B. Morrison BW, Kapour S. Kuobu
32. Hawkey C, Kahan A. Steinbruck K, Alegre C. Baumclou E. D, Antes L, et al. Effects of specific inhibition of cyclooxygenase-2
Begaud B. et al. Gastrointestinal tolerability of mcloxicam com- on sodium balance. hemodynamics, and vasnactivc eidosanoids.
pared to diclofenac in osteoarthritis patients. Br J Rheumatol J Pharmacol Exp Ther 1999:289:735-41.
33. Altman R, Asch E, Bloch D. Bole G, Borenstein D. Brandt K, et
al. Development of criteria for the classification and reporting of APPENDIX
osteoarthritis: classification of osteoarthritis of the knee. Arthritis STUDYGROUP
Rheum 1986;29: 1039-49.
34. Altman R. Alarcon G, Appelrouth D, Bloch D. Borenstein D. Investigators of the Rofecoxih Protocol 035 Study Group are
Brandt K, et al. The American College of Rheumatology criteria as follows: T. Adamson. MD (San Diego. CA). J. Aneelo. DO (New
for the classification and reporting of osteoarthritis of the hip. Orleans, LA). K. Bahrt,‘MD (South PlaYinfield: NJ). A: Bobrove: MD
Arthritis Rheum 1991:34:505-14. (Palo Alto, CA). B. Bockow, MD (Seattle, WA). A. M. Brabham, MD
35. The Steering Committee of the Physicians’ Health Study Research (Columbia, SC), D. Britt, MD (Longmont. CO). A. Brodsky, MD
Group. Preliminary report: findings from the aspirin component of (Dallas, TX). J. Caldwell, MD (Gainesville, FL). G. Cannon, MD (Salt
the ongoing physicians’ health study. N Engl J Med 1988318: Lake City, UT). C. K. Champion. MD (Bloomfield Hills, Ml). M.
262-4. Cohen, MD (Jacksonville, FL), J. Conte. MD (Barrington. RI). W.
36. Ehrich E, Schnitzer T, Kivitz A. Weaver A, Wolfe F, Morrison B. Cryan, MD (Long Beach, CA), M. Doyle, MD (Ferndale. MI), M.
et al. MK-966, a highly selective cox-2 inhibitor. was effective in Drehobl, MD (San Diego, CA), V. Elinoff, MD (Endwell. NY), H. W.
the treatment of osteoarthritis of the knee and hip in a h-week Emori, MD (Medford, OR), J. Ervin. MD (Kansas City. MO), M.
placebo controlled study [abstract]. Arthritis Rhcum 1997;40 Ettinger, MD (Stuart, FL), G. Fiocco, MD (Lacrosse, WI), C. Fisher,
Suppl 9:S85. MD (Newport News, VA), R. Fleischmann, MD (Dallas, TX), W.
37. Ehrieh E, Davies G, Watson W, Bolognese J. Seidcnberg B, Gruhn, MD (Charlotte, NC), J. Hatford, MD (Cincinnati, OH), M.
Bellamy N. Minimal perceptible clinical improvement in osteoar- Heller, MD (Peabody, MA), D. Henry, MD (Salt Lake City, UT), P.
thritis: WOMAC and global assessments. influence of aee. gender Holt, MD (Baltimore, MD), R. Hynd, MD (Oklahoma City, OK), D.
and geographic regioi[abstract]. Arthritis Rheum 199f;42-Suppl James. DO (Tulsa, OK), S. Kafka, MD (Rochester, NY), J. Kaine, MD
9S143. (Sarasota, FL), J. Kay, MD (Burlington, MA). J. Kearney. MD (Camp
38. Saag K, Fisher C, McKay J, Ehrich E. Zhao P-L, Bolognese J, et Hill, PA), L. Kirby, MD (Sun City. AZ). D. Koster, MD (Albuquerque,
al, for the MK-0966 Phase III Protocol 033 Study Group. MK- NM), R. Levin, MD (Palm Harbor, FL). J. Markenson, MD (New
0966, a specific cox-2 inhibitor, has clinical efficacy comparable to York, NY), B. McLean, MD [Birmingham. AL), F. G. McMahon, MD
ibuprofen in the treatment of knee and hip osteoarthritis in a (New Orleans, LA). A. Nafziger, MD (Cooperstown, NY), K. Nies.
6-week controlled clinical trial [abstract]. Arthritis Rheum 1998;41 MD (Torrance, CA), J. Norton. MD (Colorado Springs. CO), H.
Suppl 9:S196. Offenberg. MD (Daytona Beach, FL), C. Oshrain, MD (Rochester,
39. Williams HJ, Ward JR, Egger MJ, Neuner R, Brooks RH, Clegg NY), R. Z. Paster. MD (Oregon, WI), H. Resnick, MD (Lake Jackson,
DO; et al. Comparison sf naproxen and acetaminophen in a TX), P. Ripley, MD (South Yarmouth, MA), J. Robbins, MD (Sacra-
two-year study of treatment of osteoarthritis of the knee. Arthritis mento, CA), L. Robison. MD (Memphis, TN), N. Roche. MD
Rheum 1993;36:1196-206. (Fremont, CA). D. Ross, MD (Mobile, AL), M. Russell. MD (Tallas-
40. Ward DE, Veys EM, Bowdler JM. Roma J. Comparison of see. AL), A. Safdi, MD (Cincinnati, OH), M. Schiff, MD (Denver,
aceclofenac with diclofenac in the treatment of osteoarthritis. Clin CO). T. Schnitzer, MD, PhD (Chicago, IL), K. Solinsky, MD (West ,.:
Rheumatol 1995;14:656-62. Babylon, NY), J. C. Stringer, MD (Manlius, NY), P. Toth, MD
41. Bellamy N, Buchanan WW, Chalmers A. Ford PM, Kean WF. (Indianapolis, IN). R. Trapp. MD (Springfield, IL). 0. Troum, MD
cunnar RG, et al. A multicenter study of tcnoxicam and diclofe- (Santa Monica, CA), B. Walsh, DO (Tucson. AZ), A. Weaver. MD
nac in ,patients with osteoarthritis of the knee. J Rheumatol (Lincoln, NE), C. Wiesenhutter. MD (Coeur d’Alene, ID). C. Wise.
1993;20:999-1004. MD (Richmond, VA), S. Wolfe, MD (Springboro, OH), and D.
42. Simon IS, Lanza FL, Lipsky PE, Hubbard RC, Talwalker S, Young, MD (Carmichael, CA).
I . . . :


. .._
GI STUDIES WITH VlOXri” (rofecoxib)

., ,_~
anti recelvtm asoirin were not included in~these’tiu

-’ ~:~~&er?f.o!l
higher risk for developing a GI b!eed than

,. _-

_,, ./. .- .;
. ,. ..,~,‘.; it,.., ”

,, ., : . . ., ,_
.., _- .L.. 1 :

/ i , .. _
m Placebo (N=l58)
HVIOXX 25 mg (N=186)
n VIOXX 50 mg (N=i78)
n Ibuprofen 2400 mg (N=l67)
tKO.001 “5 ibuprofen 2400 mg

6 Weeks 12 Weeks 24 Weeks
Night Pain Relief Morning Stiffness Relief
Baseline Mean = 65 mm Baseline Mean = 74 mm
H Placebo n VIOXX12.5 mg q.d. H VIOXX 25 mg q.d. n Ibuprofen 800 mg t.i.d.
I .,1 ,,,..>l,..->
; ./ II .;; .... ‘,&f..” ‘++&&;:.
‘-I( ., @2~~~,:*$$“$~$*&~. ii’ -li;<+i,g+s^o,
- ‘““‘$&qg$$$$f
P<O.OOl for VIOXX vs placebo; P=NS‘for ViOXX vs Ibuprofen. ;,‘,,-.“i.
I *,._ l,i:,l_,.lX ,_ . i’ ,.,.... :,$&g@*&$
* Randomized, double-blind, placebo- and ~~~~~~~~~dr~~~~~~~~.‘~~~~liel-group study to assess the effi&y @J&$
dosing (12.5 mg q.d., 25 mg q.d.) vs ibuprofen (600?$ ti.d.) in paints with OA of the hip or knee. \“../
,i ‘XIL,/ : ..‘-0”
L. >ps5,‘,
The Western Ontario and McMaster_UniGersities (WOMAC) questionnaire was used, which is based on patient’s assessme
obtained from a 24question survey comprised of three subscales: Pain (5 questions), Stiffness (2 questions), and Physic.
(17 questions). ., ‘_ , (.( <‘_i,.,,1_,.__ 3,:.i., ,“>O
_,. ‘_ ..~.,,g;+,
The six-week measurements of “Pain at Night” (WOMAC Question 3) and “Morning Stiffness” (WOMAC Question 6) vZe$

. -



Tablets represented at actual size; 12.5:mg tablet shown.’

., <..,y ;

Baseline Mean = 76 mm

i v-) I:., I; “i’..I &~,~+;~~~<~“~~ :.- ..d;+~i”..-” ‘W”“:
’ The Western Ontario and M&laster Universities (WOMAC) questionnaire was used, which is based on patient’s,
assessment of pain, obtained from a 24-question survey comprised of three subscales; Pain (5 questions)! ..
Stiffness (2 questions), and Physical Function (17 questions), _l_, .,. .:,
~~rr;i j ,,,,,,;;,:~~~~~~:~*~ +&~;..~i*~~ i
fi&<;*;@g>;::J :>.g~?~~*
The primary endpoint Pain Walking on a Fiat Surface (WOMAC Question 1 from the Pain subsiale~
measured by a 0-loo-mm VAS, in which 0 mm = No Pain and 100 mm = &reme Pafi.?:
,\ ,<,,.<.“‘;y;: “; ,~~~~~~~~~~~~~~~~
-- ; ‘.+‘,.<?.~b, 1,-k.,‘..!?Q$@yp
..i ,,.,,d,?
l Once-daily VIOXX 12.5 mg and 25 mg were ~1,+:;~,-;;~;~;~~,
comparable to diclofenac 150 mg (50 mg t.i.d.). .’
l Patients could receive additional arthritis
medications during the last six months
of the study. ,-. .
l VIOXX is not a-substitute for aspirin for
cardiovascular prophylaxis because it
-does not affect platelet function.
.’ Before prescribing VIOXX, please read
the complete
.. Prescribing Information


rotecoxlb) .
550 ing

n woxx 50 mg (N=l10) W Naproxen Sodium 550 mg (N=55) Placebo W53)
~a ..,.

I---L- --d’,
r of Datients 80 vears

1. Schwartz JI et al. Cyclooxygen,
by rofecoxib reverses naturally oc
humans. C/in Phbr$,a?6rTfyy.“,
1999;65(6):653-660.* ‘. ,,:.,--,‘;
2. Laine L et al.“A ‘rar&i;;i;kZ trr
the effect of rofecoxib, a ‘cy&oXjgen_. _ -r--,
inhibitor, with that of ibuorofen in the .-

-. -,-.- -.-..-r.- VII *LyYL-‘L tI”III J

Serwces, WPl-27, Merck & Co., Inc.,‘WeNesii%int,‘PA ”
19486. Please specify information package
DA-VlO5(2). ,. , ,. . ..I.
.. :I&.., .,.,$.;-.-‘:~
‘> %&
4. Cannon CW,et al.kofe?ox$, a specrfic rnhibrtor of’%.” “.”
cyclooxygenase 2, with clinical efficacy comparable
with that of diclofenac sodium: results of a-one-year,
randomized, clinical t;hl^iii.~~eii~~~~~~~~e~~~hritis
of the knee and hip. Arthritis Rheum. 2000; ‘. * .’
43(5):978-987. . ’ “;“; 1’: ‘;5Ez>;!z;:,F&~,Q;- ; i:,,
5. Morrison BW et al: Ahalge%‘effXa~ o?& ”.
cyclooxygenase;2-sdecifrc inhibitor rofecnxih in

Product Class OA Dose Regimen* WAC’

Celebrex NSAID 200 mg 200-mg capsule q.d. $2.62
(celecoxib) or
100-mg capsule b.i.d. $2.38
Arthrotec NSAID/ Ar-throtec 50 50-mg tablet t.i.d. $3.66
(diclofenac sodium prostaglandin (50 mg didofenac
and misoprostol) analog sodium/200 pg

Daypro NSAID 1200 mg Two 600-mg caplets q.d. $2.56 ‘“J ’ ”

Reiafen NSAID 1000 mg Two 500-mg tablets q.d. $2.04
(nabumetone) 1500 mg Two 750-mg tablets q.d. $2.40
2000 mg Two 750-mg tablets + $3.42
one 500-mg tablet q.d.
. . /

i . “_

Voltaren-XR NSAID 100 mg 100-mg tablet q.d. $2.60
(diclofenac sodium)
VIQXX@ VIOXX@ (rofecoxib

metaboiite is reversible
tablets and oral suspension1

in humans to a limited extent (~5%).
VIOXX@ (rofecoxib tabletsand oral suspension)

(rofecoxib tablets and oral suspension) OA clinical studies, once daily treatment in the morning with
These metabolites are inactive as COX-1 or 60X-2 inhitiitors. VIOXX 12.5 and 25mg was associated with a significant
Cytochrome P450 plays a minor role in metabolism of rofe- reduction in joint stiffness upon first awakening in the morn-
DESCRIPTION coxib. Inhibition of CYP 3A activity by administration of keto- ing. At doses of 12.5 and 25 mg, the effectiveness of VIOXX
VIOXX’ (rofecoxib) is described chemically as 4-[4-(methyl- conazole 400 mg daily does not affect rofecoxib disposition. was shown to be comparable to ibuprofen 800 mg TID and
sulfonyl)phenyll-3-phenyl-2(5M-furanone. It has the follow- However, induction of general hepatic metabolic activity by diclofenac 50 mg TID for treatment of the signs and symp-
ing chemical structure: administration of the non-specific inducer rifampin 600 mg toms of OA. The ibuprofen studies were &week studies; the
daily produces a 50% decrease in rofecoxib plasma concentra- diclofenac studies were 12.month studies in which patients
tions. (Alsosee5rugI”teractions.~ could receive additional arthritis medication during the last6
Rofecoxib is eliminated predominanrly by hepatic metabo-
lism with little (cl%) unchanged drug recovered in the urine. Analgesia, including Dysmenorrhea
Following a single radiolabeled dose of 125 mg, approxi-
In acute analgesic models of post-operative dental pain,
mately72%ofthedosewasexcreted into the urineas metabo-
poet-orthopedic surgical pain, and primary dysmenorrhea,
lites and 14% in the feces as unchanged drug.
VIOXX relieved pain that was rated by patients as moderate to
The plasma clearance after 12.5. and doses was severe.Theanalgesiceffect(includingonsetofactionjofasin-
approximately 141 and 120 mUmin, respectively. Higher
gle 5Omg dose of VIOXX was generally similar to 550 mg of
plasma clearance was observed at doses below the therapeu-
naproxen sodium or 400 mg of ibuprofen. In single-dose
tic range. suggesting the presence of a saturable route of post-operative dental pain studies, the onset of analgesia with
metabolism (i.e., “on-linear elimination). The effective
asingle50-mgdoseofVlOXXoccurredwithin45minutes. Ina
half-life (based on steady-state levels) was approximately multiple-dose study of post-orthopedic surgical pain in which
Rofecoxib is a white to off-white to light yellow powder. It is 17 hours.
patients received VIOXX or placebo for up to 5 days. 50 mg of
sparingly soluble in acetone, slightly soluble in methanol and VIOXX once daily was effective in reducing pain. In this study,
isopropyl acefate, very slightly soluble in ethanol, prac&ally patients on VIOXX consumed a significantly smaller amount
insoluble in odanol! and insoluble in water. The empirical for- of additional analgesic medication than patients treated with
mula for rofecoxib IS C1+lIIOIS, and the molecular weight is The pharmacokinetics of rofecoxib are comparable in men
end women. placebo (1.5 versus 2.5 doses per day of additional analgesic
314.36. medication forVlOXXand placebo, respectively).
Each tablet OfVlOXXfor oral administration contains either Geriatric
12.5 mg. 25 mg, or 50 mg of rofecoxib and the following inac- After a single dose of 25 mg VIOXX in elderly subjects (over
tive ingredients: croscar~ellose sodium, hydroxypropyl cel- 65 years old) a 34% increase in AUC was observed as com-
lulose, lactose, magnesium stearate. microcrystalline pared to the young subjects. Dosage adjustment in the elderly Special Studies
cellulose, and yellow ferric oxide. The 50 mg tablets also con- is not nacessary; however, therapy with VlOXXshould be initi- Upper Endoscopy in Patients with Osteoarthritis
tain red ferric oxide. atedatthelowest recommended dose. Two identical (U.S. and Multinational) endoscopy studies in
Each 5 mL of the oral suspension contains either 12.5 or a total of 1516 patients were conducted to compare the per-
25 mg of rofecoxib and rhe following inactive ingredients:cit- Pediafric centage of patients who developed endoscopically detectable
ric acid (monahydrate), sodium citrate (dihydratel, sorbitol VIOXX has not been investigated in patients below 18 years gastroduodenalulcerswith V10XX25 mg dailyor mg daily,
solution, strawberry flavor, xanthan gum, end purified water. of age. ibuprofen 2400 mg daily, or placebo. Entn/ criteria for these
Added as preservatives are sodium methylparaben 0.13% and Race studies permitted enrollment of patients with active Helico-
sodium propylparaban 0.02%. Meta-analysis of pharmacokinetic studies has suggested a batter py/ori infection. baseline gastroduodenal erosions,
slightly (lo-15%) higher AUC of rofecoxib in Blacks and His- prior history of a” upper gastrointestinal perforation. ulcer, or
CLINICAL PHARMACOLOGY pa&&as compare&to Caucasians. No dosage adjustment is bleed (PUB). and/or age 265 years. However, patients receiv-
necessary on the basis of race. ing aspirin (including low-doseaspirinforcardiovascuiar pro-
Mechanism of Action phylaxis) were not enrolled in these studies. Patients who
VIOXX isa nonsteroidal anti-inflammatory drug that exhib- Hepatic insufficiency
were 50 years of age and older with osteoarthritis and who
its anti-inflammatory. analgesic, and antipyretic activities in A pharmacokinefic study in mild (Child-Pugh score 56) had no ulcers at baseline were evaluated by endoscopy after
animal models.ThemechanismofacIionofVlOXXisbelieved hepatic insufficiency patients indicated that rofecoxib AUC weeks 6. 12. and 24 of treatment. The placebo-treatment
ro be due to inhibition of prostaglandin synthesis, via inhibi- was similar between these patients and healthy subjects.
groupwasdixontinued atweek 16 by design.
tion of cyclooxygenase-2 (COX-2). At therapeutic concentra- Limited data in patients with moderate (Child-Pugh score 7-9)
Treatment with VIOXX25 mgdailyor50 mgdailywas asso-
tionsin humans,VlOXXdoesnotinhibiithecyclooxygenase-1 hepatic insufficiency suggest a trend towards higher AUC ciated with a significantly lower percentage of patients with
(COX-1) isoenzyme. (about 69%) of rofecoxib in these patients, but more data are
endoscopic gastroduodenal ulcers than treatment with ibu-
needed to evaluate pharmacokinetics in these patients. profen 2400 mgdaily. However, thestudiescannot rule out et
Pharmacokinetics Patients with severe hepatic insufficiency have not been
Absorption least some increase in the rate of endoscopic gastroduodenal
studied. ulcerswhencomparingVlOXXtoplacebo.SeeFigures land2
The mean oral bioavailability of VIOXX at therapeutically
recommended doses of 12.5,25, and 50 mg is approximately Renal Jnsufficisncy andtheaccompanyingtablesfortheresunsofthesestudies.
33%. The area under the curve (AUCI and peak plasma level In a study (N=6) of patients with end stage renal disease
(&,,I following a single 25-mg dose were 3286 ($8431 ng-hri undergoing dialysis, peak rofecoxib plasma levels and AlJC
mL and 207 (till) n@mL, respectively. Both C,,, and AUC declined 18% and 9%, respectively, when dialysis occurred
ara roughly dose proportional across the clinical doss range. four hours after dosing. When dialysis occurred 48 hours after
Figure I
At doses greater than 50 mg, there is a iess than proportional dosing, the elimination profile of rofecoxib was unchanged.
increase in C,,, and AUC, which is thought to be due to the COMPARISON TO IBUPROFEN
While renal insufficiency does not influence the pharmacoki-
low solubility ofthe drug in aqueous media. The plasma con- netics of rofecoxib, use of VIOXX in advanced renal disease is Life-Table Cumulative Incidence Rate of Gastroduodenal
centration-time profile exhibited multiple peaks. The median not recommended at present because no safety information is Ulcers 2 3mm” (Intention-to-Treat)
time to maximal concentration IT,,&, as assessed in nine available regarding the use of VIOXX in these patients.
pharmacokinetic studies, is 2 to 3 hours. Individual Tmaxvalues
in these studies ranged between 2 to 3 hours. This may not Drug Interactions (Also see PRECAUTIONS, Drug
reflect rate of absorption as T,, may occur es a secondary Interactions.) U.S. Study
peak in some individuals. With multiple dosing, steady-state
conditions are reached by Day 4. The AlJc0.2~~ and C,, at Genl?ral
steady state after multiple doses of 25 mg rofecoxib was In human studies the potential for rofecoxib to inhibit or
4018(t1140) ng-hr/mL and 327 (+104) nghL, respectively. induce CYP 3A4 activity was investigated in studies using the
Theaccumulation factor based on geometric meanswas 1.67. intravenous erythromycin breath fest and the oral midazolam
VIOXX Tablets 12.5 mg and 25 mg are biaequivalent to test. No significant difference in erythromycin demethylation
VIOXX Oral Suspension 12.5 mg/5mL and 25 mg/5 mL, was observed with rofecowib (75 mg daily) compared to pla-
respectively. cebo, indicating no induction of hepatic CYP3A4. A 30%
reduction of the AUC of midazolam was observed with rofe-
Food and Antacid Effects coxib (25mg daily). This reduction is mosl likely due fo
Food had no significant effect on either the peak plasma increased fira pass metabolism through induction of intesti-
concentration (C,,,) or extent of absorption fAUC) of rofe- nal CYP 3A4 by rofecoxib. In vitro studies in rat hepatocytes
coxibwhenVlOXXtabletsweretakenwirha highfatmeal.The also suggest that rofecoxib migh? be a mild inducer for
time to peak plasma concentration (T,,,). however, was CYP 3Ab. Time by Treatment
delayed by1 to2 hours.Thefoodeffectonthesuspensionfor- Drug interaction studies with rofecoxib have identified
_- niulation has not beenstudied. VIOXX tabletscan be adminis- potentially significant interactions with rifampin, methotrex-
tered without regard to timing of meals. ate and warfarin. Patients receiving fhese agents with VIOXX
There Mies a 13% and 8%.decrease in AUC when VIOXX was should be appropriately monitored. Drug interaction studies
ad&+isfe?&l~ with calcium carbonate antacid and magne- do not support the potential for clinically important interac-
s&@~tiinu~...antaoid to elderly subjects, respectively, tions between antacids arxirn$tidine with rofecoxib. Similar
-” Therk v&an $proliinBte 20% decrease in C,,j of rofeeoxib to ‘experience ,witti other nonsteroiddl?anti-inflammatow
sith Biiherantacid. drugs (NSAiDs), studies wilfi iofecoXib suggdst the potential
_ for interaction
the pharmacokinetics
with ACE inhibitors. The effects of rofecoxib on
and/or pharmacodynamics of ketocon-
Rofecoxib is approximately 87% bound to human plasma azole, prednisonelprednisolone. oral contraceptives. and
protein over the rangeof concentrationsof0.05 to25 mcg/mL. digoxin have been studied in viva and clinically important
The apparent volume of distributmn at steady state IVdsrj is interactions havenotbeen found.
approximately 91 Lfollowing a 12.5mg dose and 86 L follow-
Rofecoxib has been shown tqcross the placenta in rats and Osteoarthritis lOA)
rabbits, and theblood-brain barrier in rats. VIOXX has dem&rated significantrddu&nin joint pain
comparedtoplacebo. VIOXX wasevaiuatedforthetreatment TABLE1
Metabolism . gf &e signs ‘and symptoms of OA of the knee and hip in
Endoscopic GanrDd”adenai”lcen
“.S srrdu
at 12we&
Metabolism of rofecoxib is primarily mediated through placebo- and active-controlled clinical trials of 6 ro 86 weeks
: ‘reduction b~cytosolicenzymes.Theprincipal metabolicprod- duration that enrolled approximately 3900 patients. In
ucts are the cis-dihydro and trans-dihydro derivatives of rofe- patients with OA, treatment with VIOXX 12.5 mg and 25 mg
coxib, whichaccountfornearly56%ofrecovered radioactivity once daily resulted in improvement in patient and physician
in the urine. An additional 8.8% of the dose was recovered es global assessmentsand in the WOMAC (Western Ontarioand
the glucuronide of the hydroxy derivative, a product of oxida- McMaster Universities) osteoarthritis questionnaire. includ-
tive metabolism. The biotransformation of rofecoxib and this ing pain, stiffness,andfunctionalmeasuresofOA. Insixstud- VIOXX25 mg 7,188 4.1% 0.4, 10.16,1.0!3
7--Registered mdemark of MERCK & CO., Inc.. Whitehouse Station, . ies of pain accompanying OA flare, VIOXX provided a VIoM50mg WI78 7.3% il.74
New Jersev. USA stgnificant reduction in pain at the first determination (after
lbuprafen 421167 27 7% 2.79
COPYRIGHT 0 MERCK&CO., Inc., ,998 one week in one study, after two weeks in the remaining five
All rights resewed. studies); this continued for the duration of the studies. In all *byYetable a”aly5is
VIOXXB (rofecoxibtablets and oral suspension) VIOXP (rofecoxibtablets and oral suspensionl

matory drugs (NSAIDs). Minor upper gastrointestinal prob- incidence observed with ibuprofen and lower than that
Figure 2 lems, such as dyspepsia, are common~a~~ may also occur at observed with diclofenac. In placebo-controlledtrials, approx-
COMPARISON TO IBUPROFEN any time during NSAID therapy. Therefore, physicians and imately 0.5% of patients taking rofecoxib (12.5 or 25 mg QD)
patientsshould remain alert for ulcerationand bleeding, even and 0.1% of patients taking placebo had notable elevations of
Life-Table Cumulative lncidance Rate of Gastmduodenal in the absence of previous GI tract symptoms. Patients should ALT or AST.
Ulcers 2 3mm” IIntention-to-Treat) be informed about rhe signs and/or symptoms of serious Gl A patient with symptoms and/or signs suggesting liver dys-
toxicity and the steps to take if they occur. The utility of peri- function, or in whom a” abnormal liver test has occurred,
odic laboratov monitoring has not been demonstrated, nor should be monitored carefully for evidence of the develop-
hasit beenadequatelyassesaed.Onlyo”einfivepatie”tswho ment of a more severe hepatic reaction while on therapy with
Multinational Study develop a serious upper GI adverse event on NSAID therapy is VIOXX. Use of VIOXX is not recommended in patients with
50- symptomatic. k has been demonstrated that upper GI ulcers, moderate orsevere hepaticinsufficiency(see Pharmacokinet-
gross bleeding or perforation, caused by NSAIDs. appear to its, Specie/Populetians). If clinical signs and symptoms con-
d 40- occur in approximately 1% of patients treated for 3-6 months. sistent with liver disease develop, or if systemic
and in about l-4% of patients treated for one year. These
manifestations occur (e.g., eosinophilia, rash, etc.), VIOXX
trends continue thus, increasingthe likelihood of developing a
.xf ^ 3J- serious GI event at some time during the course of therapy.
should bediscontinued.
c E However,even short-termtherapy is notwithout risk.
Xl- It is unclear, at the present time, how the above rates apply
.P Renal Effects
P to VIOXX (see CLINICAL STUDIES, Special Studies, Upper Long-term administration of NSAlDs has resulted in renal
z 10 - Endoscopy in Patients with Osteoarthritis). Among 3357 papillary necrosis and other renal iniurv. Renal toxicity has
patients who received VIOXX in controlled clinical trials of alsobeen seen in patients in whom renalprostaglandins-have
i; 0
om 6-weeks to one-year duration (most were enrolled in aiIompensatory role in the maintenance of renal perfusion. In
awdt 12-Week”’ 24-VVeek six-month or longer studies) at a daily dose of 12.5 mg to these patients, administration of a nonsteroidal anti-inflam-
Time by Treabnenf 50 m@, a total of 4 patients experienced a serious upper GI matory drug maycausea dose-dependent reduction in pros-
event, using protocol-derived criteria. Two patients experi- tartlandin formation and, secondarilv. in renal blood flow.
enced en upper GI bleed within three months (at day 62 and which may precipitate overt renal de&mpansation. Patients
81, respectively) (0.06%). One additional patient experienced at greatest risk of this reaction are those with imoaired renal
an obstruction within six months (Day 130) and the remaining
function, heartfailure,liverdysfunction. thosetak/ngdiuretics
patient developed a” upper GI bleed within 12 months (Day
and ACE inhibitors, and the elderly. Discontinuation of NSAID
322) (0.12%). Approximately 23% of these 3357 patients were
in studiesthat requiredthemtobefreeof ulcersatstudyentry. therapy is usually followed by recovery to the pretreatment
r. gc 0.001 versus ibuprqfen 2400 mg state. Clinical trials with VIOXX at daily doses of 12.5 end
esults of analyses using a 2 5mm gastroduodenal ulcer endpoint It is unclear if this study population is representative of the
general population. Prospective, long-term studies required 25 ma have shown renal effects le.a.. hvoertension. edema)
were oonsisant.
“‘*The primary endpoint was the cumulative incidence of to compare the incidence of serious, clinically significant sirnil& to those observed with co&pa&or NSAI&: these
gastroduodenal u!cer at 12 weeks upper GI adverse events in patientstakingV\OXXvscompara- occur with an increased frequency with chronic use of VlOXX
tar NSAID products havenot been performed. ;;;c;j”s above the 12.5 to 25 mg range. (See ADVERSE REAC-
TABLE2 NSAlDs should be prescribed with extreme caution in
patientswith a prior historyofulcer diseaseorgastrointestinal Caution should ix used when initiating treatment with
bleeding. Moat spontaneous reports of fatal GI events are in VIOXX in patients with considerable dehydration. It is advis-
elderly or debilitated patients and therefore special care able to rehydrate patients first and then starl therapy with
should be taken in treating this population. To minimize the VIOXX. Caution is also recommended in patients with
potential risk for an adverse 01 event, the low& effective pm-existing kid”eydisease(seeWARNINGS,Adva”cedRe”al
dose should bs usad for the shortest possible duration. For Disease).
VIOXX25 rnQ 1.M high risk patients, alternate therapies that do not involve
9/187 5.3% 10.36.3111)
NSAlDs shou.ld be considered.
VIOXX6Qmg tLw2 8.8% 1.73 la.65.rsl) Studies haveshownthatpatientswithapriorhistoryofpep- Anemia is sometimes seen in patients receivina VIOXX. In
Ibuprofen 49na7 232% 5.72 tic ulcer disease and/orgastrointestinal bleeding and who use placebo-con~rolledtrials,therewkre nosignificanthifferenws
NSAIDs, haveagreaterthan lo-fold higher riskfor developing observed between VIOXX and placebo in clinical reports of
‘by lie table an&%is a GI bleed than patients with neither of these risk factors. In anemia. Patients on long-term treatment with VIOXX should
addition to a past history of ulcer disease. pharmacoepidemi-
ological studies have identified several other co-therapies or have their hemoglobin or hematocrit checked if they exhibit
The c&relation between findings of endoscopic studies,
and the relative incidence of clinically serious upper GI events co-morbid conditions that may increase the risk for GI bleed- any signs or symptoms of anemia or blood loss. VIOXX does
that may be observed with different products, has not been ing such as: treatment with oral corticosteroida, treatment not generally affect platelet counts, prothrombin time (PT), or
fully eslabliphed. Seriousclinically significant upper GI bleed-. with anticoagulants, longerduratian of NSAID therapy. smok- partial thromboplastin time IPTTI, end does not inhibit platelet
ing has been dbsetied in patients receiving VIOXX in ‘don-- .. ing, alcoholism, older age. and poor general health status. aggregation at indicated dosages (see CLINICAL STUDIES,
>m!led trials; albeit infrequently (see WARNINGS, Specialstudies, Platelet&
Gastroinresfinal Hill Effects -Risk of Gl Ulceration, Bleeding, .,
Anaphylactoid Reactions
end PerforationI. Prospective, long-term studies required to Fluid Reten$on and Edeme
As with NSAlDs in general, anaphylactoid reactions have
eonware the incidence of serious, clinicallv sionificent uooer Fluid retention and edema have been observed in some
occurred in patients without know” prior exposure to VIOXX.
Gladverseeventsin patientstakin~VIOXX~e&uscompa~~tor patients taking VIOXX (see ADVERSE REACTIONS). VIOXX
In post-marketing experience, rare cases of anaphylactoid
NSAID products have not been performed.
reactions and angioedema have been reported in patients should be used with caution. and should be introduced at the
receiving VIOXX. VIOXX should not be given to patients with lowest recommended dose in patients with fluid retention,
Assessment of Fecal Occult Blood Loss in Heelthy Subjects the aspirin triad. This symptom complex typically occurs in hypertension, or heart failure.
Occuftfecal blood loss associated with VIOXX 25 mg daily, asthmatic patients who experience rhinitis with or without
VIOXX 50 mg daily, ibuprofen 2400 mg per day, and placebo nasal polyps, orwho exhibit severe, potentially fatal broncho-
was evaluated in a study utilizing S’Cr-tagged red blood curls spasm aftertaking aspirin or other NSAlDs (seeCONTRAINDI- Preexisting Asthma
in 67 healthy males. After 4 weeks of treatment with VIOXX CATIONS and PRECAUTIONS, Preexisdng Asthma). Patients with asthma may have aspirin-sensitive asthma.
25 mg daily or VIOXX 50 mg daily, the increase in the amount Emergency help should be sought in cases where an anaphy- The use of aspirin in patients with aspirin-sensitive asthma
of fecal blood loss was not statistically significant compared lactoid reaction occurs. has been associated with severe bronchospasm which can be
with placebo-treated subjects. In contrast, ibuprofen 2400 mg fatal. Since cross reactivity, including bronchospasm.
per day produced a statistically significant increase in fecal Advanced Renal Disease between aspirin and other nonsteroidal anti-inflammatory
blood loss as compared with placebo-treated subjec$ and No safety information is available regarding the use of drugs has bee” reported in such aspirinsensitive patients,
VIOXX-treated subjects. The ~linica,llelevance ofJh$ ftnding VIOXX in patients with advanced kidney disease. Therefore, VIOXX should not be administered to patients with this form
is unknown. : treatmentwith VIOXX is not recommended in these patients. Of aspirin sensitivity and should be used with caution in
,@ If VIOXX therapy must be initiated, close monitoring of the patients with preexisting asthma.
Platelets patient’s kidney function is advisable (see PRECAUTIONS,
Multiple doses of VIOXX 12.5,25. and up to 375 mg admin- Renal Effects).
istered daily up to ‘12 days had rro effect an bleeding time rela- Information for Patients
Jive to placebo. Similarly, bleeding time was not altered in a Pregnancy _ VIOXX can cause discomfort and, rarely, more serious side
single dose studywith 500 or 1000 mg of VIOXX. Therewas no I” late pregnancy VIOXX should be avoided because it may effects, such as gastrointestinal bleading, which may result in
inhibition of ax viva arachidonic acid- or collagen-induced cause premature closure of the ductus arteriosus. hospitalization and eve” fatal outcomes. Although serious GI
platelet aggregation with 12.525, and 50 mg of VIOXX. tract ulcerations end bleeding can occur without warning
_.. _ PRECAUTIONS symptoms, patients should be alert for the signs and symp-
toms of ulcerations and bleeding, and should ask for medical
advice when observing any indicative signs or symptoms.
VIOXX is indicate& VIOXX cannot be expected to substitute for corticosteroids Patients should be apprised of the importance of this fol-
Forreliefofthe signsandsymptomsofosteoarthritis. ortotreat corticostemid insufficiency.Abrupt discontinuation low-up (see WARNINGS, Gestrointestinal(G/)Eff~s- Riskof
For the management of acute pain in adults (see CLINICAL of CorticosCeroids may lead to exacerbation of corticoste-
GI Ulceration, Needing and Perforation).
STUDIES). raid-responsive illness. Patients on prolowed corticosteroid
Forthe treatment of primary dysmenorrhea. theraovshould have their theraov taoered slowlv if a decision

is made to discontinue corticosterokfs. Patients should promptly report signs or symptoms of gas-
The pharmacological activity of VIOXX in reducing inflam- trointestinal ulceration or bleeding, skin rash, unexplained
CONTRAINDICATIONS _- mation, and possibly fever, may diminish the utility of these weight gain, or edema to their phvsicians.
VIOXX is Eon&indicated in patients with known hypersen- diagnostic signs in detecting infectious complications of pre- Patientsshould beinformed ofihewarningsignsandsymp
sitivityto rofecoxib oranyothercomponent of VIOXX. sumed noninfectious, painfulconditions. tams of hepatotoxicity (e.g., nausea, fatigue, lethargy, pruri-
VIOAX should not be give” to patients who have erperi- tus, jaundice, right upper quadrant tenderness, and”flu-like’
enced asthma, urticaria. or allergic-type reactions after taking nepetic Effects symptoms). If these occur, patients should be instructed to
aspirin or other NSAIDs. Sever&, rarely fatal, anaphylactic-like Borderlineelevationsofoneormore livertests mayoccurin stoptherapyandseekimmediatemedical therapy.
reactions to NSAlDs have been reported in such patients (see upto 15%ofpatientstaki”g NSAIDs,and notableelevationsof Patients should also be instructed to seek immediate emer-
WARNINGS, Anaphylactoid Reactions and PRECAUTIONS, ALTorAST~approximatelythreeormoretimss~eupperlimif gency help in the case of an anaphylactoid reaction (see
Preexisting Asthma). of normal) have bee” reported in approximately 1% of WARNINGS).
patienrsinclinical trialswith NSAlDs.Theselaboratoryabnor- In late pregnancy VIOXX shbuld be avoided because it may
WARNINGS malities may progress, may remain unchanged, or may be cause prematureclosure oftheductus arteriosus.
Gastrointestinal (GO Effects -Risk of Gf Ulceration, Bleeding, transient with continuing therapy. Rare cases of severe
and Perforation hepatic reactions, including jaundice and fatal fulminant hep-
Serious gastrointestinal toxicity such as bleeding, ulcer- atitis, liver necrosis and hepatic failure (some with fatal out- Laboratory Tests
ation, and perforation of the stomach, small intestine or large come) have been reported with NSAIDs. In controlled clinical Because serious GI tract ulcerationsand bleeding can occur
intestine, can occur at any time, with or without warning trials of VIOXX, the incidence of borderline elevations of liver without warning symptoms, physicians should monitor for
symptoms, in patients treated with nonsteroidal anti-inflam- tests at doses of 12.5 and 25 mg daily was comparable to the signs orsymptoms of GI bleeding.
VIOXX@ (rofecoxib tablets and oral suspension) VlOXX@ (rofecoxibtabletsand oral suspension) VIOXX@ (rofecoxibtabletsandoral suspension)

coxib had no effect on fertility in female rats at doses up to
ACEinhibitors: Reports suggest that NSAlDs may diminish 30 mg/kg (appro%imately 19-and ‘I-fold human exposure at 25
theantihypertensiveeffeCtofAngiotensinConvertingEnzym8 and 56 nig daily based on AUI&~I. Placebo VIOXX Ibuprofen Diclofanac
(ACE) inhibitors. In patients with mild to moderate hyparten- 1250r25mo 2WVma 110nlo
sion. administration of 25 mg daily of VIOXX with the ACE
inhibitor banazepril, 10 to 46 mg for 4 weeks, was associated
with an average increase in mean arterial pressure of about Pregnancy
3 mm Hg compared to ACE inhibitor alone. This interaction Teratogenic effects: Pregnancy Category C.
shoutd be given consideration in patients taking VIOXX con- 4.1 3.4 4.6 5n
Rofecoxibwas not teratogenic in rats at doses up to 50 mg/
comitantlywith ACEinhibitors. kg/day (approximately 28. and IO-fold human exposure at 25 1.0 2.2 2.0 26
and 50 mg daily based on AU&&. There was a slight, 2.2 3.0 2.7 3.4
Aspirin: Concomitant administration of low-dose aspirin non-statistically significant increase in theoverall incidence of 3.1 2.9 1.5 3.2
with VIOXX may result in an increased rate of Gl ulceration or vertebral malformations onlv in the rabbit at doses of 50 ma/
1.1 3.7 3.8 3.4
other complications, compared to use of VlOXX alone. At kg/day (approximately l- or&fold human exposure at25 anb
steady state, VIOXX 60 mg once daily had no effect on the 50 mg daily based on AU&&. There are no studies in preg- 7.8 8.5 5.8 8.2
anti-platelet activity of low-dose (81 mg once daily) aspirin, as nant women. VIOXX should be used during pregnancy only if
assessed by ex viva platelet aggregation and serum TX& the potential benefit iustifiesthe ootential risk to the fetus. 1.3 3.5 50 1.6
generation in clotting blood. VIOXX is not a substitute forbspi- Nbnteratogenic egects: Rofecoxib produced peri-implanta-
rinforcardiovascularprophylaxis. tion and post-implantation losses and reduced embryo/fetal
survival in rats and rabbits at oral doses 210 and 2775 mg/kg/ 6.8 8.5 7.1 18.8
Cimetidine: Co-administration with high doses of cimeti- day, respectively(approxim%tely9-and3-fold [rats1 andl-and 27 3.5 4.7 4.0
dine /800 mg twice daily] increased the C,,, of rofecoxib by <l-fold [rabbits! human exposure based on theAUCc.2rat 25 2.8 3.8 9.2 5.4
21%, the AUC,,.lmh, by 23% and the tm by 15%. These small and 50 mg daily). These changes are expected with inhibition
3.6 4.2 5.2 4.6
changes are not clinically significant and no dose adjuament ofprostaglandinsynthesisand are not the result of permanent
is necessary. alteration of female reproducdve function. There was an 29 5.2 7.1 7.4
increase in the incidance of postnatal pup mortality in rats at
Oigoxin: Rofecoxib 75 mg once daily for 11 days does not Z5 mg/kg/day (approximately 5. and 2.foid human exposure 20 27 1.8 2.4
alter the plasma concentration profile or renal elimination of at 25 and 50 mg daily based on AUC&). In studies in preg-
digoxin afterasingle0.5 mgoraldose. nant rats administered single doses of rofecoxib, there w%s %
treatment-related decrease in the diameter ofthe ductus arte- 19 25 1.4 2.8
Furosemide: Clinical studies. as well as wst-marketina riosus at all doses used (3-300 ms/kg: 3 mg/kg is approxi-
observations, have shown that NSArDs can rkduce the natri mately 2- and <l-fold human exposure at 25 or 50 mg daily tIesdacha 7.5 4.7 6.1 n-0
ureticeffectoffurosemideand thiazidesin someo%tients.This based onAUCo.z,J. Aswith other drugs known to inhibit pros-
response has been attributed to inhibition of renal prostaglan- tagtandin synthesis, useofVlOXXduringthethirdtrimesterof
din synthesis. pregnancy should be avoided.

Ketoconezole: Ketoconazole 400 mg daily did not have any
clinically important effect on the pharmacokinetics of rofe- Labor and delivery
corib. Rofecoxib produced no evidence of significantly delayed Thegeneraisafetyprofile ofV10XX50 mg QDin OAclinical
labor or parturition in females at doses 15mg/kg in rats trials of up to 6 months (476 atients) was similar to that of
Lithium:NSAIDs have oroduced an elevation of olasma lith- (approximately lo- and J-fold human exposure as measured VIOXX at the recommended 8 A doses of 12.5snd 25 mg QD,
ium levels and a reduction in renal lithium clear&e. Thus, except for a higher inctdence of gastromtestmal symptoms
bytheAUC%.24%t25and50 mgl.TheeffectsofVlOXXo”l%bor Iabdominal pam, epigastric pain, heartburn, nausea and
when VlOXX and lithium are administered concurrently, sub- and delivery in pregnant women are unknown. vomiting). lower extremity edema (6.3%) and hypertension
jectsshould beobaervedcarefullyforsignsof lithium toxicity. Merck & Co.,~l& maintains a registry to monitor the preg- (8.2%).
nancyoutcomesofwomenexposed toVlDXXwhilepregnant. In the OA studies, the following spontaneous adverse
Mefhotiex.+te:VlOXX75 mg administeredonce daily for 10 Healthcare providers are encouraged to report any prenatal events occurred .in >O.l% to l.% of patients treated with
days increased plasma concentrations by 23% as measured exposure to VIOXX by calling the Pregnancy Registry at (800) VlOXXregardlessofcausality:
by AUCD-Z,+,r in patients receiving methotrexate 7.5 to 15 mg/ 9868889.
week for rheumatoid arthritis. An equivalent magnitude of Bodyasa Who/e:abdominal distension! abdominaltender-
reduction in methotrexate renal clearance was observed. At ness. abscess, chest pain, cnilh, contuston, cyst, dlaphrag-
matlc hernia. fever, fluid retention, flushing, fungal infection,
24 hours postdose,
with methotrexate
a similar proportion
alone (94%) Zlnd subsequently
of patients treated
treated with
.. infection, laceration, pain, pelvic pain. peripheral edema,
Nursing mothers -postoperative pain, syncope, trauma. upper extremity
‘methbirexate co-administeied riitR of iof&oxib (66%) edema, viral syndrome.
Rofecoxib is excret&d in the milk of lactating rats at concen-
had methotrexate plasma concentrations below the measur-
trations similar to these in plasma. There-was an increase in
able limit (5 nglmL). The effects of the recommended doses Cardiovascular System: angina pectoris, atrial fibrillation,
for osteoarthritis (12.5 and 25 mg) of ViOXX on plasma meth- -pup mortality and a decrease in pup body weight following bradycardia, hematoma, irregular heart beat, p%[pitation,
otrexate levels are unknown. Standard monitoring of metho- exposure of popsto milk from dams administered VIOXX dur- rematureventricular contraction, tachycardia, venous insuf-
trexate-related toxicity should be continued if ViOXX and ing lactation. The dose tested represents an approximate 18- &iency.
methotrexateare administered concomitantly. and B-fold human exposure at 25 and 50mg based on
AU&+ It is not known whether this drug is excreted in Digestive System: acid refiux, aphthous stomatitis, consti-
human milk. Because many drugs are excreted in human milk pation, dental caries, dental pain, digestive gas symptoms.
Oral Contraceptives: Aofecoxib did not have any clinically dry mouth, duodenal disorder, dysgeusia, esophagltis, flatu-
important affect on the pharmacokinetics of ethinyl estradiol and because of the potential for serious adverse reactions in
nursing infants from VIOXX. a decision should be made lence, gastric disorder gastritis, gastroenteritis. hematoche-
and norethindrone. zia, hemorrhoids, infectious gastroenteritis. oral infection,
whethertodiscontinuenursingortodiscontinuethedrug,t%k- oral lesion,oral ulcer,vomiting.
ingintoaccountthe importanceofthedrugtothemother.
Prednisone/prednisolone: Rofecoxib did not have any clini-
cally important effect on the pharmacokinetics of predniw- Eyes, Ears, Nose, and Throat: allergic rhinitis. blurred
loneorprednisone. vision, cerumen Impaction, conjunctivitis, dry throat,
.^ epistaxis, laryngitis, nasal congestion. nasal secretion, oph-
Pediatric Use thalmic mjection, otic pain, obtis. otitis medta, pharyngitis,
fiifampin: Cmadministration of VIOXX with rifampin Safety and effectiveness in pediatric patients below the age tinnitus,tondllitis.
600 mg daily, a potent inducer of hepatic metabolism, pro- of 18yearshavenot beenevaluated.
duced an approximate 50% decrease in rofecoxib plasma con- immune System: allergy, hypdrsensitivity, insect bite reac-
L centrations. Therefore, a starting daily dose of 25 mg of tion.
VIOXX should be considered for the treatmem of osteoarthri-
tis when VIOXX is co-administered with potent inducers of Metabolism and Nutrition: appetite change, hypercholes-
hepatic metabolism. Geriatc(c Use - terolemia, weightgain.
OfthepatientswhoreceivedVlOXX in osteoarthritisclinical
trials. 1455 were 65 veals of aqe or older (this included 460 Musco/oa~e/era/Sysrem:anklesprain. arm pain,,arthralgia,
We&win: Anticoagulant activity should be monitored, par- who were 75 years br older). ?Jo substantial differences in back strain, bursitis, cartila e trauma, .oint swelling, muscu-
ticularlyinthafirstfewdaysafteriniti%tingorchangingVlOXX safety and effectiveness were observed between these sub- lar cramp, muscular disor 9 er, muscu Iar weakness, muscu-
therapv in oatients receivina warfarin or similar aoants. since jects and younger subjects. Greater sensitivity of some older loskeletal pain. muaculoskeletal stiffness, myalgia.
the& t&&ts are at %n in&s%d risk of bleeding complica- individualscannot be ruled out. Dosage adjustment in the eld- osteoarthritis, tendinitis, traumatic arthropathy, wrist frac-
tions. In &gle and multiple dose studies in healthy subjects ture.
erly is not necessary; however, therapy with VIOXX should be
receiving-,both warfarin and rofecoxib, prothrombin time initiated at the lowest recommended dose.
(measured as INR) was increased by approximately 8% to Nervous. System: hypesthesia, insomnia, median nerve
In one of these studies (a six-week, double-blind, random- “europathy, migr%ine,muscul%rspasm,paresthesia.sci%tica,
11%. In post-marketing experience, b&ding eve& have ized clinical tria11, VlOXX 12.5 or 25 mg once daily was admin- somnolence,vertigo.
been reported, predominantly in the elderly, in association iStered to 174 osteoarthritis patients 280 years of age. The
with increases in prothrombin time in patients receiving
safety profile in this elderly population was similar to that of
VlOXX concurrently with warfarin.
younger patients treated wtih VIOXX.
RespiratorySysrem:%shm%, cough, dyspnea, pneumonia,
Carcinogenesis, Mutagenesis, Impairment of Ferrility pulmonan/congestion,respiratory infection.
Rofecoxib was not carcinogenic in micegiven oral doses up
to 30 mg/kg (male) and 60 mgikg (female) (approximately 5. Skin andSkinAppendages:%brasion, alopecia, atopic der-
and 2.fold the human exposure at 25 and 50 mg daily based ADVERSE REACTIONS
matitis, basal cell carcinoma, blister, cellulitis,cont%ct derma-
on AU4.P4)andinm%le%nd femaleratsgivenoral doses upto Osteoarthritis titis, herpes simplex, herpes zoSter, nail unit disorder,
8 mg/kg (approximately 6- and 2-fold the human exposure at Approximately 3600 patients with osteoarthritis were perspiratton, pruritus, rash, skin erythema, urticaria, xerosis.
25 and 50 mg daily based on AUCO& for two years. treated with VIOXX: approximately 1400 patients received
Rofecoxib was not mutagenic in an Ames test or in a V-73 VIOXXforG monthsor longer and approximately 800 patients Urogenital System: breast mass, cystitis, dysuria, meno-
mammaliancellmutagenesisassay, norclastogenicinachro- for one year or longer. The following table of adverse experi- paus%! symptoms, menstrual. disorder, nocturia, urinary
mosome aberration assay in Chinese hameter ovary (CHO) ewes lists all adverse events, regardiess of causality. occur- retentton,vaginitis.
cells, inan in vitroandan in vivoalk%lineelution%ss%y,or inan ring in at least 2% of patients receiving VIOXX in nine The followmg serious adverse events have been reported
rarely (estinwtadcO.l%1 in atients taking VIOXX. regardless
in vivochromosomal aberration test in mouse bone marrow. controlled studiesof&weektoCi-monthdurationmnductedin of causality. Cases reporte cr only I” the post-marketing expe-
Rofecoxib did not impair male fertility in rats at oral doses patients with OA at the therapeutically recommended doses rience are Indicated in italics.
up to 100 mg/kg (approximately 20- and 7-fold human expo- (12.5 and 25 mg), which included a placebo and/or positive
sure at 25 and 50 mg daily based on the AU&&) and rofe- control groop. Cardiovascular: cerebrovascular accident, congestive
VIOXXa (rofecoxibtablets and oral suspension) VIOXXQ lrofecoxib tablets and oral suspansicn) VIOXP (rofecoxib tablets and oral suspension)

heartfailure,deepvenausthrombosis. myocardial infarction, OVERDOSAGE No. 3611 -Tablets VIOXX. 25 mg, are yellow, round, tablets
pulmonary embolism, transient ischemic attack, unstable engraved MRK 110 on one side and VIOM on the other. They
No overdosesofVlOXXwere reported during clinical trials. are supplied as follows:
angina. AdministrationofsingledosesofVIO~XldOO mgto6healthy NDC0006-01 lo-31 unit ofuse bottlesof
volunteersand multipledosesof250 mgfdayfor14daysto75 NDC00066110-28unitdosepackagesof 100
Gastrointestinal; cholecystitis, colitis. colonic malignant healthyvolunteersdtd notresult inserioustoxicity. NDC0006-0110-68 bottles of 100
neoplasm, duodenalperforation, duodenal ulcer, esophageal In the event of overdose, it is reasonable to employ the NDC0006.0110.82 bottles of 1000
o/ca~ gastric wforaiion, gastric ulcer, gastrointestinal bleed- usual supportive measures, e.g., remove unabsorbed mate-
ing,tntestina Pobstruction,pancreatitis. NDC 0006-O 11 O-80 botttes of 8000.
rial from the gastrointestinal tract, employ clinical monitor-
in and institute supportive therapy, if required.
R ofecoxib is not removed by hemodialysis; it is not known No. 3818 -Tablets VtOXX, 50 mg, are orange, round, tab-
Hemicandlymphatic:IvmDhoma. lets engraved MRK 114 on one side and VtOXX on the other.
whether rofecoxib is removed by peritoneal dialysis.
Immune System: anaph ylactoid reaction, angioedema. NDC0006-0114-31 unitof use bottlesof
DOSAGE AND ADMINISTRATION NDC0006-0114-28 unit dose packages of 100
Nervous System: aseptic meningitis. VIOXX is administered orally. The lowest dose of VIOXX NDC 0006-0114-68 bottles of 100
should be sought for each patient. NDCOO06-0114-74 bottlesof 500
NOC 0006-01 l&81 bottles of 4000.
Psychiatric: hallucinations.
No. 3784 - Oral Suspension VIOX~, 12.5 mg/5 mL is an
Urogenital System: acute renal failer+ breast malignant The recommended starting dose of VIOXX is 12.5 mg once opaque. white to faint yellow suspension with a strawberry
neoplasm, interstitial nephritis, prostatlc malignant neo- daily. Some patients may receive additional benefit by flavor that is easily resuspended upon shaking.
plasm, urolithiasis, worseningchronicrenalfailore. increasing the dose to 25 mg cmce daily. The maximum rec- NDC 0006-3784-64 unit of use bottles containing 150 mL
ommended dailydoseis 25 mg. (12.5 me/!3 mL).
In l-year controlled clinical trials and in extension studies
for up fc~ 86 weeks (approximately 800 patients treated with Management of Acute Pain and Treatment of Primary
VIOXX for one ear or longer), the adverse experience profile Dysmenorrhea No. 3785 - Oral Suspension VIOXX. 25 mg/6 mL, is an
wasqualitative ystmtlartothat observed in studiesofshorter The recommended initial dose of VIOXX is 50 mg cmce opaque. white to faint yellow suspension with a strawberry
duration. ’ . daily. Subsequent doses should be 50mg once daily as flavor that is easily resuspended upon shaking.
needed. Use of VIOXX for more than 5 days in management NDC 0006-3765-64 unit of use bottles containing 150 mL
of p$n has not been studied(seeCLINICALSTUDIES,Ana/ge- (25 mg15 mL).
Analgesia, including primary dysmenorrhea sia, mcluding dysmenorrhea).
A roximately one thousand patients were treated with VIOXX tablets may be taken with or without food. storage
VI&S m analgesia studies. All patients in post-dental sur- VIOXX Tabablats:
gerypainstudiesreceived onlya singledose ofstudymedica- store at 25’C (77’F). excursions ermitted to
tion. Patients in primary dysmenorrhea studies may have VIOXX Oral Suspension 12.5 mg/5 mL or 25 mg/5 mL may 15.3O*C (59-86°F). [See USP Controlled Room ? emperature.1
taken u to 3 daily doses of VIOXX, and those ‘in the
post-art R opedlc surgery pain study were prescribed
doses of VIOXX.
5 daily
be substituted for VIOXX Tablets 12.5 or25 mg, respectively,
in any of the above indications. Shake before using.
vl~ta~eoral Suspeoslon:
25°C 177°F) excursions permitted to
15-30-C (59~~6”FL[See USP’Controtled RoomTemperature.
Theadverse experience profile in the analgesiastudies was HOW SUPPLIED
wanerally similar to those reported in the osteolfihritis stud- Rx only
bs. The’ fotlowin additional adverse experience. which No. 3810 - Tablets VIOXX. 12.5 mg, are cream/off-white,
occurred at an incl I ence of at least 2% of patients treated with round. shallow cup tablets engraved MRK 74 on one side and
VIOXX, was observed in the post-dental pain surgery studies: VIOXX on the other. They are supplied as follows: Discby:
post-dental extraction alveolitis(drysocket). - NDC 0006-0074-31 unit of use bottles of 30 0 MERCK&CO., INC., Whitehouse Station, NJ 06869, USA
In 110 patients treated with VIOXX (average age approxi- NDC0006-0074-28unitdose packagesof 100
mately 65 yearsl in the post-orthopedic sur<ery$ain study, NDC 0006-0074-68 bottles of 100
the most common1 reported adverse experiences were con- NDC 0006-0074-82 bottles of 1000 Issued May 2000
stipation,fever, an J nausea. NIX 0006-0074-80 bottles of 8000. Printed in USA 9183805