COX-2 Class Labeling From Vioxx/Celebrex Data Proposed By FDA Consultant

Class labeling for COX-2 inhibitors incorporating data from Pharmacia's Celebrex and Merck's Vioxx large-scale safety studies may be useful in explaining differential safety results in gastrointestinal and cardiovascular adverse events, a consultant to FDA's Arthritis Advisory Committee suggested Feb. 8.

Is there a "way of working all of this up in a class labeling for the COX-2 [inhibitors] that would somehow summarize the information gleaned from both very large studies?" voting committee consultant Allan Sampson, PhD, University of California-San Diego, asked.

COX-2 class labeling could be used in place of or in addition to the nonsteroidal anti-inflammatory drug class warnings of increased upper gastrointestinal toxicity found on Celebrex and Vioxx labeling.

While Merck's 8,000-patient Vioxx Gastrointestinal Outcomes Research Study showed rofecoxib to have an advantage in GI safety over the traditional NSAID naproxen, the arthritis committee suggested that any benefit could be offset by a lack of cardiovascular protection (¹ ).

Pharmacia was able to show a GI advantage over the NSAID ibuprofen only in a secondary comparison of systematic and complicated ulcers, and showed no benefit when compared to the NSAID diclofenac, according to the 8,000-patient Celecoxib Long-term Arthritis Safety Study.

Celecoxib was not statistically different from comparators in cardiac events, although there were trends toward a greater incidence of anginal events and myocardial infarctions in the celecoxib group (² (Also see "Celebrex, Vioxx GI Safety Differences May Be Indistinguishable - FDA Cmte." - Pink Sheet, 12 Feb, 2001.)).

There is "a lot of information in [the] CLASS and VIGOR studies and...a wealth of opportunity for people that would like to do some meta-analytical work combining these two studies to try to tease out a stronger effect or...a stronger inference," Sampson suggested.

"I think the absence of cardioprotective effect for both COX-2 inhibitors should be emphasized in product literature....There is nothing I heard from either yesterday or today which suggests that either agent has a cardioprotective effect as do the non-[COX-2]-selective agents," voting committee consultant Steven Nissen, MD, The Cleveland Clinic Foundation said.

"I think we need to investigate whether there is an excess of cardiovascular events to longer-term exposure in both of these agents in comparison to placebo," he suggested.

In analyst conferences following the meeting, Pharmacia pointed to committee recommendations for cardiovascular warnings in Merck's Vioxx labeling, while Merck maintained that the committee's conclusions about a lack of cardioprotective effect extend to both COX-2 inhibitors (³ (Also see "Pharmacia, Merck Explain FDA Cmte. Review Of COX-2 Safety To Investors" - Pink Sheet, 19 Feb, 2001.)).

The committee discussed whether sponsors should focus on specific organ safety, such as gastrointestinal or cardiovascular, or if overall "global" safety should be established first.

"The benchmark probably should be overall safety," Nissen maintained. A drug should be shown to "produce less serious effects [overall] than another drug," or else sponsors can choose the "right" endpoint and comparator, and "get it to show almost anything."

"There is [a] potential to make a serious mistake when you focus all your attention in this early development on this target organ concept," Nissen pointed out. "So in preapproval, I really do think we don't want to lose the FDA's focus on overall safety."

Committee consultant David Wofsy, MD, University of California-San Francisco, contended that "in this particular instance there was a reason why organ-specific toxicity was the right thing to look at first" and then to "go beyond that and look broadly at other things."

Wofsy stressed, however, that "if a drug does not have an efficacy advantage and it is being put forward primarily because of its safety advantage [like Celebrex and Vioxx], a particularly thorough safety evaluation needs to happen in whatever sequence before a final decision is made."

While Vioxx and Celebrex labeling will focus on GI and cardiovascular effects, "the question is to what degree other organ systems impact these studies and to what degree should we be monitoring other organ systems," Acting Committee Chair Nigel Harris, MD, Morehouse School of Medicine, noted. "I think that these are really muddy waters and I think that maybe it is a separate point that the FDA does need to think through some of these issues with regards to safety trials," Harris recommended.

FDA medical officer Lawrence Goldkind, MD, noted that an approach "where specifically [an] organ safety claim [is] considered...would be a marked change from the past in terms of what is asked for preapproval to have a large safety database, particularly a comparative safety database."

Nissen questioned whether the "net benefit" in GI toxicity "here exceeds the net harm" in cardiovascular events, as seen in the Vioxx database. "It actually would be a lot easier for me to advocate a COX-2 inhibitor for a young patient without cardiovascular risk because I can see where the benefits would be outweighing the risk," he said.

"When you consider that an atherosclerotic event is the cause of death in about 50% of the American population, you are talking about the potential for an awful lot of morbidity and mortality as you treat those patients with agents that may increase the risk of that endpoint," he added.

The committee debated whether a GI safety advantage could have been generalized over all NSAIDs if Celebrex had shown a conclusive advantage over the comparators in Pharmacia's studies.

"Based upon the cumulative experience of the studies, it appeared that diclofenac and ibuprofen fall on the lower end of the [toxicity] spectrum," Committee guest expert Byron Cryor, MD, University of Texas Southwestern Medical Center, observed. "If you are showing a difference between the ones that fall on the lowest end, you would expect that you find clearly a difference with the ones that were more toxic."

Wofsy contended, however, that with 10 different available NSAIDs to use as comparators, it is hard to predict the performance of celecoxib or rofecoxib based on the positive results from one or two NSAIDs. "There are 10 NSAIDs out there. If we did them all...would we find that you were better than nine and not better than one, or would we find that you were better than one and not better than nine?" he asked.

Some committee members suggested FDA choose one standard NSAID comparator to avoid the question of whether the results could be generalized across the class. FDA should pick "the gold standard for NSAIDs and make sure that everybody uses the same one from now on," voting committee consultant Michael Wolfe, MD, Boston University, recommended.

Committee Chair Harris also urged FDA to choose comparators for future COX-2 comparisons. "In the future, when one is designing studies like this, what advice should we give them in terms of comparative drugs?...Which drug is best representative of the non-steroidals? Is it one, is it two [or] is it three?"

FDA Office of Drug Evaluation V Director Robert Delap, MD/PhD, put the ball back in the committee's court. "I hate to say that in reality, we'll probably come back to you and ask you what you think is the drug you should compare to so that we can tell our sponsors, and have some public discussion of that."

Finding an appropriate comparator, however, will be difficult, Nissen indicated. Is it realistic to ask sponsors to conduct "studies for each of the dozen or so potential comparators before we agree there is some benefit?" he asked. "It is not going to happen [and] it is not reasonable to make it happen."

Determining whether either celecoxib or rofecoxib was shown to have a clinically meaningful safety advantage depends on how one looks at a number of different variables, Cryor pointed out.

In the case of Celebrex, "our answer's going to be different if we make the comparison with ibuprofen versus diclofenac," he reiterated. Other factors include the endpoint measured, concomitant aspirin use, the study time and risk group characteristics.

Celebrex showed a statistically significant improve-ment in symptomatic/complicated ulcers over ibuprofen in patients not taking aspirin, but the positive effects were not replicated in aspirin users. Merck's VIGOR study did not enroll patients on aspirin.

"How we define clinically meaningful safety advantage also has to be considered with respect to time," Cryor continued. While in the first 90 days, there were no differences in safety between ibuprofen and celecoxib, some separation was seen at 12 months, he noted.

Safety results also depended on different risk group characteristics, Cryor added. While celecoxib showed no safety advantage in patients older than 75, in those with a history of upper GI bleeds, celecoxib did show an advantage over traditional NSAIDs.