NEWS - Valdecoxib (Bextra) meta-analysis showing CV risk is inappropriate, say experts

[Guest guest]

Valdecoxib (Bextra) meta-analysis showing CV risk is " inappropriate, "

say experts

Rheumawire

Nov 12, 2004

Zosia Chustecka

The meta-analysis suggesting that valdecoxib (Bextra, Pfizer) increases

cardiovascular (CV) risk that hit the headlines is " inappropriate, " say

coxib experts, and the CV signal seen in trials involving high-risk

patients undergoing coronary artery bypass graft surgery (CABG) is not

relevant to chronic use of lower doses for osteoarthritis (OA) and

rheumatoid arthritis (RA).

Presented by Dr Garret FitzGerald (University of Pennsylvania,

Philadelphia) during an invited lecture at the American Heart

Association meeting earlier this week, the meta-analysis involved a

total of 7500 patients, of whom about 2000 were post-CABG surgery. It

showed that valdecoxib was associated with twice as many myocardial

infarctions (MIs) or stroke events than placebo (relative risk 2.19).

FitzGerald pointed out that this was higher than the 1.96 relative risk

shown with rofecoxib (Vioxx, Merck) in the APPROVE trial, which led to

the drug being withdrawn, commenting: " This is a time bomb waiting to go

off. "

Dr Lee Simon (Beth Israel Deaconess Hospital, Boston, MA) says this

meta-analysis is inappropriate and its conclusions problematic. He also

points out that these data are not new. " What FitzGerald did was take a

meta-analysis that has already been published [1], and he subtracted out

a certain number of patients that did not meet his convenienceit's not

clear exactly which patients, but that published meta-analysis has

around 8000 patients, while the analysis FitzGerald was talking about

had around 5000 without the CABG patientshe took some out, and I don't

think it's possible to do such a thing, " he tells rheumawire. " Also, he

didn't have access to the original data set, and furthermore, he added

in the 2 CABG trials, which I don't think is appropriate. " These 2

trials in CABG patients (involving about 2000 patients) reflect an

entirely different patient population and a different way of delivering

the drug, he points out: the patients were given intravenous parecoxib

(the prodrug) immediately after surgery, followed by oral valdecoxib at

a much higher dose than is used in arthritis. " More important, I don't

know where he got these data fromthese data are not publicly known, only

one of these trials has been published, and the 2 trials themselves had

important differences in both the duration of therapy and the dose of

medications used. "

So is this inappropriate because it's comparing apples and oranges?

" It's even worse than that, " says Simon. " There are certain rules about

performing meta-analyses, because they are always fraught with problems,

and so you want to be sure that the trials are comparable enough to be

combined. . . . I don't think that this meta-analysis, based on

selection of trials, would have met the most commonly used criteria to

allow trials to be combined. "

" Therefore, I would say that the conclusions are problematic, because I

don't know how to interpret them. I don't know that he had the original

data sources, and under those circumstances I wonder about really what

these data are telling me, " Simon comments. " And to suggest that this

is a new meta-analysis is misleading. "

Simon advises rheumatologists reading reports on the latest news about

the coxibswhich are flowing thick and fast in the wake of the rofecoxib

withdrawalto be " very critical of reports in the public press. " He urges

all physicians to " appraise very carefully what the report is based on

and then make every attempt to get hold of the literature that it

appears to be referring to, to make sure that we are basing our

decisions on evidence and not on hype. "

Dr Vibeke Strand (Stanford University, CA) agrees with these criticisms

of the FitzGerald analysis and adds several points from a perspective of

closer involvement, as the second author of the published meta-analysis.

" The data were taken apart and used in pieces, " she tells rheumawire.

Their published meta-analysis was restricted to trials of the approved

indication of chronic use of valdecoxib for osteoarthritis and

rheumatoid arthritis, with a total of more than 8000 patients and

various nonselective nonsteroidal anti-inflammatory drugs (NSAIDs) used

as comparators. " We saw no cardiovascular signal, " she says, even though

these patients had a slightly higher background rate of cardiovascular

disease (42%-50%) than has been seen in other studies and a low use of

aspirin for cardioprotection (12%-15%).

Strand points out that a CV signal with valdecoxib may be evident only

in 2 trials carried out in very high-risk surgical patients undergoing

CABG, in whom valdecoxib was used at a much higher dose (40 mg and 80

mg) than is used for arthritis and in conjunction with the intravenous

formulation of its prodrug. In the first of these studies, most of the

MIs that were seen occurred during surgery and before the first dose of

these drugs, so " it would be hard to say that the drug was implicated, "

while some of the MIs in the second studyin which patients not only

underwent CABG but were also put on a cardiopulmonary pump for 12 to 24

hoursoccurred more than 5 drug half-lives after the last dose. " These

findings are difficult to interpret, " she comments. " There does seem to

be a signal, " she says, but she argues that observations in these very

high-risk CABG patient populations may not be relevant to the use of

10-mg and 20-mg daily doses in OA and RA, and misleading reports may

further " scare patients away. " Strand also points out that the results

of the second CABG study have yet to be published, precluding detailed

peer review of their appropriate use in meta-analyses.

However, FitzGerald says that the findings in the CABG trials are like

" the canary in the cold mine, " as the likelihood of detecting a signal

of increased cardiovascular risk with COX inhibitors is expected to be

linked to the degree to which the hemostatic system is activated. " If

the thrombosis risk is elevated, as it is in the CABG patients, we are

much more likely to see an effect if it is there. The clustering of

these events in CABG patients is exactly the sort of signal we should be

concerned about. "

Several of the rheumatologists contacted by rheumawire for their

reactions to the news of the valdecoxib meta-analysis commented that

several factors need to be borne in mind when considering prescribing a

selective COX inhibitor.

Strand emphasized the GI advantages of the coxibs over traditional

NSAIDS and says they are " absolutely worth using for GI benefit. " COX-2

inhibitors offer numerical GI benefit even with coadministration of

low-dose aspirinseen in both the CLASS (with celecoxib) and TARGET (with

lumiracoxib) studiesand 2 recent studies have shown less small-bowel

mucosal injury with a COX-2 inhibitor compared with a nonselective NSAID

combined with a proton pump inhibitor.

Strand adds that part of the cardiovascular issue with the coxibs may be

related to their effects on hypertension and edema, which are similar to

those of nonselective NSAIDs. There are dose-related increases in

hypertension and edema with rofecoxib, but not with celecoxib, and there

appears to be a dose-related increase with valdecoxib at 40 mg and 80 mg

but not at the lower doses of 10 mg and 20 mg, she comments. In

addition, she points out that recent epidemiological studies suggest

there may also be a cardiovascular signal with some of the traditional

nonselective NSAIDsin 1 study with ibuprofen and in another with

indomethacin, diclofenac, and naproxen. However, interpretation of these

data are challenging, since trials with NSAIDs did not enroll patient

populations of the same size and length of treatment as the studies that

have been carried out with coxibsVIGOR and APPROVE with rofecoxib, CLASS

with celecoxib, TARGET with lumiracoxib, and EDGE with etoricoxib.

Dr Crofford (University of Kentucky, Lexington) comments that

it's " very important for physicians to consider an entire grid of

information when prescribing NSAIDs and COX-2 inhibitors. It appears

that drugs with higher COX-2 selectivity (rofecoxib,

valdecoxib/parecoxib) are most likely to cause cardiovascular thrombotic

events, although those with less selectivity (celecoxib) are not yet

completely exonerated. The risk is higher in those with preexisting CV

disease or risk factors. Though aspirin may eliminate the risk, that is

not yet proved. Furthermore, aspirin use appears to eliminate the GI

benefit of the COX-2 inhibitors. Finally, there are other strategies to

reduce the risk of upper-GI toxicity, the most common (though not the

only) GI toxicity associated with NSAIDs. For all these reasons, I would

recommend using COX-2 inhibitors with caution. "

Dr Matteson (Mayo Clinic, Rochester, MN) says: " This new

meta-analysis heightens concerns about the potential for all COX-2 drugs

to carry with them an increased risk of cardiovascular complications.

The underlying question of risk can best be addressed in properly

controlled prospective trials designed to answer the question, but the

concerns raised by the study cannot be ignored. For patients at high

risk of GI bleeding, the COX-2 drugs are useful in the management of

inflammation and pain where no other alternatives are available. It

would seem prudent to limit exposure to them to as brief a period as

possible. "

Dr Topol (Cleveland Clinic, OH), editor-in-chief of our sibling

website, theheart.org, who raised the alarm over the increased

cardiovascular risk of coxibs several years ago, comments: " I agree that

the data on valdecoxib in CABG patients look quite concerning, but this

is the first time any coxib has been tested in a true high-risk

population. " He adds that he " will await peer review of this work. "

Source

White WB et al. " Effects of the cyclooxygenase-2

specific inhibitor valdecoxib versus nonsteroidal antiinflammatory

agents and placebo on cardiovascular thrombotic events in patients with

arthritis. " Am J Ther 2004; 11:244-250.

I'll tell you where to go!

Mayo Clinic in Rochester

http://www.mayoclinic.org/rochester

s Hopkins Medicine

http://www.hopkinsmedicine.org