COX2 inhibitors and Graham's study
This is the response we sent to the Lancet concerning this study.
They may or
not accept to publish it, but they don't have online reactions. It may
BMJ readers, too. I hope this does not compromise the acceptability by the
Lancet, just in case. Please go to the Lancet and read Graham's full
Cardiovascular risks of rofecoxib and other NSAIDs
Patrick Blin, MD, PhD (1), Mathieu Molimard, MD, PhD (1), Annie
(1), Jacques Benichou, MD, PhD (2), Nicholas Moore, MD, PhD, FRCP (Edin)
(1) Department of Pharmacology, Université Victor Segalen Bordeaux 2,
Bordeaux, France; INSERM U657, Bordeaux, France; CHU de Bordeaux,
(2) INSERM U657, Rouen, France; Unité de Biostatistique, CHU de Rouen,
In their recent and excellent paper (1) Graham et al show an
adjusted odds ratio for the occurrence of myocardial infarction (MI) and
sudden cardiovascular death in current users of high-dose rofecoxib,
naproxen, and non-steroidal anti-inflammatory drugs (NSAIDs) other than
ibuprofen or celecoxib, and in recent (1 to 60 days) users of any NSAID,
compared to remote (>60 days) users of any NSAID (including rofecoxib
celecoxib). These results are rather surprising, because in most other
NSAIDs are generally cardioprotective (2) or have no effect on
mortality (3, 4, 5), or the risk of myocardial infarction increases after
withdrawal (6). However, these studies were comparative to no use rather
than remote use of NSAIDs. In Dr Graham’s study, the reference group is
group of remote users, ie, those patients who have not used the drugs
the last 60 days, and the risk of MI is increased in recent or current
use within the last 60 days. However, recent or current use may be
of regular or longer term use of NSAIDs which may select different
with more active disease, at higher risk of cardiovascular disease (7).
The main thrust of the paper, however, is the increased risk of
compared to celecoxib: rofecoxib is not different from the non-selective
NSAIDs, all of which are associated with a higher adjusted risk of MI than
celecoxib, a fact that the authors do not emphasize. So that the real
of this paper is a “protective” effect of celecoxib, rather than increased
with rofecoxib. Interestingly this only appears when the crude odds ratios
adjusted for cardiovascular risk (except for high-dose rofecoxib). The
elements in the cardiovascular risk adjustment are given in table 2 of the
paper. Some of these elements are indeed risk factors (previous history of
cardiovascular admission), others are described as indicative of risk, but
in fact use of cardiovascular drugs that are protective such as beta-
ACE inhibitors, lipid-lowering drugs, etc used in primary or secondary
prevention. Looking at table 2, there seems to be a higher use of these
protective drugs in the celecoxib group than in all the other groups.
the greater use of cardioprotective drugs explain the large change induced
this group by adjustment? The cardioprotective effect of celecoxib is
not generally admitted, and the recent Adenoma Prevention with Celecoxib
(APC) study, similar to the rofecoxib APPROVe study (8) also showed a dose
dependent increase in cardiovascular risk with celecoxib (9).
This does not explain the apparently higher risk associated with
high dose. However there were only 18 patients (10 cases and 8 controls)
with high-dose rofecoxib in the current use group, resulting in an
situation, where a single misclassified subject (in dose or group for
controls that were not clinically reviewed) might make these results non-
significant. Not that there could not be an increased risk with high-dose
rofecoxib, as suggested by other studies.
Finally the computation of an excess risk resulting in 88000 to
related to the use of rofecoxib extrapolated from data from clinical
ludicrous. The same type of extrapolation had been done for GI bleeding
when the COX2 inhibitors were put on the market (10-12). These
computation were wrong (13). The present one are probably wrong too. The
risks in continuous use clinical trials in selected patients or in
indications can hardly be applied indiscriminately to the intermittent use
in real life. Since they used data from clinical trials to compute excess
with rofecoxib, one can wonder why the authors did not also use the data
from the APC trial (9) to compute the excess risk with celecoxib. At the
the authors should have used data from their own study, which may be
construed as representative of real use and risks, as they did in a
version of the paper (14), where the computations gave “only” 27000
However, the data from their study indicate not only an increase risk with
rofecoxib rather than celecoxib, but also an increased risk with the non-
specific NSAIDs, which are much more widely used, so that the real message
should have been that the use of any NSAIDs (including rofecoxib) rather
celecoxib resulted in several hundred thousands or millions of excess
Without rejecting a possible cardiovascular risk with rofecoxib (for
there may be some strong arguments), obviously more studies are needed.
And risks or benefits found in studies in indications that have nothing to
with the common indications and usage patterns of NSAIDs (or for that
any other drug) should be extrapolated only with the greatest of caution.
Conflict of interest: the authors are involved in a 46000 patient
study of the
usage patterns and risks associated with COX2-selective inhibitors and
NSAIDs in France, the results of which should be available within a few
months. This study, financed by Merck & Co and Pfizer, Inc, is done at
behest of the French authorities under the control of an independent
scientific committee. The authors have no financial benefit from this
its results or any other involving these drugs. Some of the authors have
working on issues related to NSAIDs safety for years, and have formed a
number of opinions they may tend to favour, possibly, which could skew
1. Graham D, Campen D, Hui M, Spence M, Cheetham C, Levy G, et al.
of acute myocardial infarction and sudden cardiac death in patients
with cyclo-oxygenase-2 selective and non-selective non-steroidal anti-
inflammatory drugs: nested case-control study. Lancet 2005;365:Published
online Jan 25, 2005.
2. Kimmel SE, Berlin JA, Reilly M, Jaskowiak J, Kishel L, Chittams J, et
effects of nonselective non-aspirin non-steroidal anti-inflammatory
medications on the risk of nonfatal myocardial infarction and their
with aspirin. J Am Coll Cardiol 2004;43(6):985-90.
3. Garcia Rodriguez LA, Varas-Lorenzo C, Maguire A, Gonzalez-Perez A.
Nonsteroidal antiinflammatory drugs and the risk of myocardial infarction
the general population. Circulation 2004;109(24):3000-6.
4. Mamdani M, Rochon P, Juurlink DN, Anderson GM, Kopp A, Naglie G, et
al. Effect of selective cyclooxygenase 2 inhibitors and naproxen on short-
term risk of acute myocardial infarction in the elderly. Arch Intern Med
5. Ray WA, Stein CM, Hall K, Daugherty JR, Griffin MR. Non-steroidal anti-
inflammatory drugs and risk of serious coronary heart disease: an
observational cohort study. Lancet 2002;359(9301):118-23.
6. Fischer LM, Schlienger RG, Matter CM, Jick H, Meier CR. Discontinuation
of nonsteroidal anti-inflammatory drug therapy and risk of acute
infarction. Arch Intern Med 2004;164(22):2472-6.
7. Nissen SE, Tuzcu EM, Schoenhagen P, Crowe T, Sasiela WJ, Tsai J, et al.
Statin therapy, LDL cholesterol, C-reactive protein, and coronary artery
disease. N Engl J Med 2005;352(1):29-38.
8. FDA. Vioxx (Rofecoxib) information. http://www.fda.gov/cder/drug/
infopage/vioxx/default.htm (accessed Feb1, 2005). 2004.
9. FDA. Celebrex (celecoxib) information. http://www.fda.gov/cder/drug/
infopage/celebrex/default.htm (accessed Feb1, 2005). 2004.
10. Singh G. Recent considerations in nonsteroidal anti-inflammatory drug
gastropathy. Am J Med 1998;105(1B):31S-38S.
11. Tramer MR, Moore RA, Reynolds DJ, McQuay HJ. Quantitative estimation
of rare adverse events which follow a biological progression: a new model
applied to chronic NSAID use. Pain 2000;85(1-2):169-82.
12. Wolfe MM, Lichtenstein DR, Singh G. Gastrointestinal toxicity of
nonsteroidal antiinflammatory drugs. N Engl J Med 1999;340(24):1888-99.
13. Moore N. Comment on 'Quantitative estimation of rare adverse events
which follow a biological progression: a new model applied to chronic
use' Tramer et al., Pain 2000;85:169-182. Pain 2001;91(3):401-2.
14. Graham D. Risk of Acute Myocardial Infarction and Sudden Cardiac
Death in Patients Treated with COX-2 Selective and Non-Selective NSAIDs.
December 2, 2004.
we are presently doing a
46000 patient study of the
usage patterns of rofecoxib,
celecoxib, and NSAIDs, and
hope to have results by next
may, which we will of course
consider superior to all other
Competing interests: No competing interests