Chicago, IL - The COX-2 inhibitor etoricoxib (Arcoxia, Merck), showed almost identical rates of thrombotic cardiovascular events as diclofenac in patients with arthritis in the large-scale MEDAL trial, but rates of other side effects differed between the two agents [1].

The study, which was presented today at the American Heart Association (AHA) 2006 Scientific Sessions and published online simultaneously in the Lancet, showed similar rates of effectiveness against arthritis symptoms with the two drugs. While the two agents also showed similar rates of lower GI events and complicated upper-GI events, etoricoxib was associated with small increases in congestive heart failure and discontinuations due to edema or hypertension, while diclofenac showed higher rates of uncomplicated upper-GI events and discontinuations due to gastrointestinal and hepatic adverse events.

Presenting the study at the AHA meeting, Dr Christopher Cannon (Brigham and Women's Hospital, Boston) said he did not want to make recommendations on how arthritis patients should be treated—it was up to arthritis experts and guidelines committees to do that—but "from our data it would seem likely that different drugs could be more appropriate for different patients." He concluded that this study did provide one piece of the puzzle. "We addressed the question: Does increasing COX-2 selectivity increase cardiac events? And we found the clear answer to this is: No, it doesn't."

Criticism over diclofenac as comparator

But the trial has received a fair amount of criticism, particularly on the decision to use diclofenac as the comparator agent. Dr Steven Nissen (Cleveland Clinic, OH), who was cochair of the late-breaking clinical-trial session at which the study was presented, commented to heartwire: "In the recent JAMA meta-analysis, diclofenac showed the highest rates of adverse cardiovascular outcomes—slightly higher even than Vioxx. Being noninferior to diclofenac is not reassuring."

In an editorial accompanying the Lancet paper [2], Drs Luis Alberto García Rodríguez (Centro Español de Investigación Farmacoepidemiológica, Madrid, Spain) and Paola Patrignani (G d'Annunzio University, School of Medicine, Chieti, Italy) also voice concern about diclofenac as the control arm. They explain that diclofenac also has some selectivity for COX-2, and it would have been preferable to have used a less-COX-2-selective nonsteroidal anti-inflammatory drug (NSAID) as the comparator, such as naproxen or ibuprofen. The editorialists say that both naproxen and ibuprofen are commonly used in the general population, where they are associated with a more favorable cardiovascular profile than that of diclofenac, and for this reason the application of the MEDAL results to general practice is limited.

The editorialists suggest that the rate of cardiovascular events associated with no treatment in the MEDAL trial may be estimated by comparing the intention-to-treat and per-protocol analyses, and by doing this they calculate that four extra cardiovascular events per 1000 patients treated for a year were associated with the use of etoricoxib or diclofenac, a number that they say is in line with a recent opinion from the European Medicines Agency.

In an interview with heartwire, García Rodríguez said he was not reassured by this new study. "For me, the increases in heart failure, hypertension, and edema shown with etoricoxib are worrying, and the fact that etoricoxib did not reduce complicated peptic ulcer makes the only safety benefit of the COX-2 drugs compared with traditional NSAIDs fade away. With the available information, it is difficult to find somebody for whom the risk/benefit assessment for etoricoxib would make it the NSAID of choice among the vast armamentarium."

Califf more accepting

Dr Robert Califf (Duke University), the designated discussant of the trial at the AHA's late-breaking clinical-trial session, was kinder about the MEDAL trial and the prospects for etoricoxib. He said the trial provided "a relatively definitive result for the direct comparison done." But he pointed out that there were many questions left unanswered, including the possibility that both agents may increase cardiovascular risk more than no agent or a different agent. "The single comparison in this trial leaves open the issue of whether naproxen or a different NSAID would have shown a better result. But from a regulatory view, these data are reassuring, and armed with this information on etoricoxib and diclofenac and the difference in price between these two agents, people can start to make decisions about who should be taking which one," Califf concluded.

The results

In the MEDAL study, 34 701 patients with arthritis were randomly assigned to etoricoxib (60 mg or 90 mg daily) or diclofenac (150 mg daily) for an average of 18 months. The primary end point—the occurrence of thrombotic cardiovascular events—was similar between the two groups.

MEDAL: Rate of thrombotic cardiovascular events (per 100 patient-years)

End point	Etoricoxib	Diclofenac	HR (95% CI)
Thrombotic CV events	1.24	1.30	0.95 (0.81-1.11)

Upper-GI events (perforation, bleeding, obstruction, ulcer) were more common with diclofenac, but there was no difference in complicated upper-GI events.

MEDAL: Rate of upper GI events (per 100 patient-years)

Outcome	Etoricoxib	Diclofenac	HR (95% CI)
Upper-GI events	0.67	0.97	0.69 (0.57-0.83)
Complicated upper-GI events	0.30	0.32	Not significant

Etoricoxib was associated with small increases in the incidence of heart failure (absolute increase of 0.1%-0.4% depending on dose) and discontinuations due to edema (absolute increase 0.1%-1.1%) and hypertension (absolute increase 0.4%-1.6%), while diclofenac showed higher rates of discontinuations because of hepatic adverse events (absolute increase 1.2%-4.7%).

Cannon defends diclofenac

At an AHA press conference, Cannon defended the use of diclofenac as the comparator agent. He pointed out that diclofenac was the most widely prescribed NSAID in the world, and, unlike ibuprofen and naproxen, it does not interfere with the antiplatelet effects of low-dose aspirin. He noted that the assertion that diclofenac has more of an increased risk of cardiovascular events than other NSAIDs had come from a summary of observational studies recently published in JAMA, but it was well known that observational data were unreliable. "I would prefer to rely on randomized data, and when all the randomized data comparing COX-2 selective drugs and NSAIDs are combined, all the agents apart from one look similar in terms of cardiovascular risk. The one that stands out is high-dose naproxen, which may have an antiplatelet effect."

Then asked why naproxen was not also included in the MEDAL program, Cannon replied: "There are two main questions in the whole NSAID/COX-2 cardiovascular-risk field that need to be addressed. Does increasing COX-2 selectivity increase cardiac risk? Is naproxen protective in terms of cardiovascular risk? One study cannot address all questions. With this trial in more 34 000 patients, we have shown a clear answer to the first question: increasing COX-2 selectivity does not increase cardiovascular risk per se. The naproxen question is still unanswered and needs to be addressed in future studies."

Another MEDAL investigator who has been heavily involved in the COX-2 field, Dr Garrett Fitzgerald (University of Pennsylvania, Philadelphia) made the point that the use of diclofenac as the comparator in this study provided new and useful information. "A comparison with naproxen, for example, in nonaspirin users would almost certainly give the results of VIGOR. Indeed, the overview of phase 2/3 studies that made this comparison had previously shown divergence of thrombotic events, so why reinvent the wheel?" he commented to heartwire.

Is etoricoxib approvable now?

Etoricoxib, although already available in many other countries, has not yet been approved by the FDA, which has been waiting for the MEDAL program. Whether the drug should now gain FDA approval is now the subject of much discussion.

Nissen says that because diclofenac was an inappropriate comparator, etoricoxib is not approvable based on these data. Cannon disagrees: "Randomized data suggest all NSAIDs and COX-2 inhibitors have similar cardiovascular risks apart from naproxen. But there needs to be more than one agent on the market. We have shown that etoricoxib is just as safe or safer, in some respects, compared with diclofenac, which is used by 15 to 20 million people worldwide. If the FDA can't approve etoricoxib now, it should remove from the market all the other 33 NSAIDs and COX-2 inhibitors apart from naproxen. Just because one drug is thought to be better than others, does this mean to say the others should not be available? For example, tPA opens arteries better than streptokinase, but does this mean streptokinase should not be sold any more?"

Fitzgerald's recommendations for treatment decisions

Fitzgerald summarized for heartwire how he would advise the use of NSAIDs and COX-2 inhibitors at the current time based on the randomized NSAID comparisons and placebo-controlled trials:

• Patients at risk of CV disease should head for the COX-1 end of the spectrum.
• Patients who are GI intolerant of naproxen should try combining it with a proton pump inhibitor and, if this doesn't work, should try ibuprofen.
• Patients at low CV risk who have GI intolerance of the above should consider drugs at the COX-2 end but be mindful that it is mechanistically plausible that CV-risk transformation might occur with continued dosing.
• For patients who are GI intolerant of traditional NSAIDs and at CV risk—no small number of people—there are very few data. However, depending on the balance of risk, their options are: naproxen or ibuprofen with clopidogrel; or a COX-2 with low-dose aspirin, recognizing that this may undermine the relative GI benefit.