You are here
Should celecoxib (Celebrex) be officially declared as no riskier for the heart than other non-steroidal anti-inflammatory drugs (NSAIDs) naproxen and ibuprofen, and do those other NSAIDs interfere with aspirin for cardiovascular prevention?
Deliberations on Tuesday by a joint FDA advisory committee centered largely on the randomized PRECISION trial, with discussion wading deep into the trial's details, but with little substantial objections or concern voiced over its findings or conduct.
PRECISION randomized osteoarthritis or rheumatoid arthritis patients with a history of, or risk factors for, cardiovascular (CV) disease to treatment of their arthritis pain with celecoxib, ibuprofen, or naproxen. It found noninferiority of celecoxib for the endpoint of death, non-fatal myocardial infarction, or non-fatal stroke compared with both naproxen and ibuprofen.
Members of the Arthritis Advisory Committee and the Drug Safety and Risk Management Advisory Committee largely agreed with the industry presenters, who concluded that "celecoxib carries CV risk that is no worse than CV risk with ibuprofen or naproxen," although reviewers cautioned that this must be interpreted in context of the doses used in the trial, which were prescription strength and cannot be extrapolated to apply to intermittent or over-the-counter dosing.
Celebrex maker Pfizer requested an update to the product label to include the PRECISION data, which the panel will vote on Wednesday among other items.
The issue harkens back to data that began to emerge in the early 2000s on cardiovascular thromboembolic risk with the COX-2 selective NSAIDs and led to withdrawal of rofecoxib (Vioxx) and valdecoxib (Bextra) from the market in 2004.
Celecoxib stayed the course, though the FDA warned that this drug increased risk of serious adverse CV events compared with placebo, and required that non-selective NSAIDs carry a black box warning of potential increased risk of serious adverse CV events, too.
PRECISION started up in the wake of that debacle and realization that there was little cardiovascular safety data on the NSAIDs.
Primary investigator of PRECISION Steven Nissen, MD, of the Cleveland Clinic, was serving on an FDA advisory committee in 2001 when he and two other leading cardiologists raised a "cautionary flag" about the safety of rofecoxib.
Nissen noted that there had been widespread assumption that naproxen was the safest for the the heart among non-aspirin NSAIDs, while "after the withdrawal of rofecoxib, many observers assumed that all COX-2 inhibitors increased major adverse cardiovascular events."
However, PRECISION showed the importance of randomized trial data over theory, he noted. "The findings highlight differences in outcomes that appear related to multiple pharmacologic effects of these drugs, not necessarily their COX-1 vs COX-2 selectivity."
The other key question at the 2-day meeting is how the trial should influence labeling and contraindications for naproxen, ibuprofen, and aspirin.
PRECISION's population was enriched for CV disease, and about 45% were taking cardioprotective doses of aspirin.
All the NSAIDs work by inhibiting cyclooxygenase (COX). Aspirin has the shortest window, with maximal release within an hour for immediate release and 3.5 to 6 hours for enteric-coated formulations, during which it irreversibly binds with COX-1.
"Celecoxib is 350 times more selective for COX-2 than COX-1," as the FDA briefing document added, so even low-dose aspirin is still more than enough to irreversibly maximally inhibit platelet activation in its presence. It was not seen to have potential for interaction in pharmacodynamic studies; in PRECISION, there was no significant interaction of the primary results by stratification for low-dose aspirin use.
"We agree with the sponsor that studies in volunteers demonstrate that celecoxib 200 mg bid does not interfere with the antiplatelet activity of aspirin used at cardioprotective doses," an FDA reviewer said at the meeting.
Ibuprofen and naproxen are both reversible inhibitors of COX-1, along with COX-2, so they can interfere with aspirin's antiplatelet activity, as seen by the biomarker serum thromboxane.
Ibuprofen may be more popular by sales figures shown at the meeting. But as the FDA reviewer noted at the committee meeting, "the available data indicate that ibuprofen administration can attenuate the antiplatelet effects of aspirin. The timing of dosing of ibuprofen relative to aspirin and the aspirin formulation have major effects on the extent of interactions."
In the Kontakt randomized open-label trial of immediate-release aspirin and various naproxen dosing strategies, there was an inhibitory effect on the biomarker that was least when aspirin was given 30 minutes before naproxen but still evident when naproxen was given 8 hours prior to aspirin. The FDA reviewer called this "unequivocal evidence" of interaction, although with unclear clinical relevance for cardiovascular outcomes.
"The initial impression that naproxen carried lower cardiovascular risk was related to the fact that its estimated effect was driven by indirect comparisons that were largely dominated by its comparison with high doses of the COX-2 inhibitor with the most consistent CV toxicity, i.e. rofecoxib 50 mg daily," Milton Pressler, MD, of Pfizer, said in discussing the PRECISION trial his company sponsored.
Nissen, though, pointed to PRECISION's findings as more solid than the pharmacodynamic and meta-analyses that have suggested interaction.
The trial stratified for aspirin use at baseline, so it could test if the CNT meta-analysis that suggested in theory that ibuprofen interference would confer advantage to celecoxib.
But what they actually saw was no interaction and, if anything, trends in the opposite direction, Nissen noted. "The biomarker does not translate to clinical effect."
In a recently-published propensity-weighted analysis, the lowest major cardiovascular event risk was lowest with celecoxib, similar to main trial. But instead of seeing a widening of difference between celecoxib and the other NSAIDs in the presence of aspirin use, there was actually a narrowing. "This is the opposite of what would be expected if there was interference with aspirin," Nissen said. The same was true for GI outcomes and renal outcomes as well as the composite of all three types of outcomes.
"These data do not show evidence that these theoretical biomarker measurements of platelet function are actually translating to an observable clinical outcome," he said, adding, "I personally believe that a single large well-performed trial is more compelling than meta-analyses."
How the panel will weigh these findings as they deliberate in the second day of discussions remains to be seen. They will be asked is there is a clinically significant interaction aspirin with one or more of the other NSAIDs and whether over-the-counter ibuprofen and naproxen should carry warnings that they should not be taken by people using aspirin for cardioprotection.
While enteric-coated formulations account for about 60% to 70% of aspirin use, there isn't good data for interaction of the NSAIDs with it, the panel heard. That topic is sure to enter into the discussion as well.