The COX-2 inhibitors—An update
Article Outline
The Multinational Etoricoxib and Diclofenac Arthritis Long-term (MEDAL) Program1 presented in this issue of the Journal is an effort to clarify the role of nonsteroidal antiinflammatory drugs (NSAIDs) in patient care. Knowledge about the balance between benefit (pain control) and harm (increased risk of serious cardiovascular events) is not well established for any of the marketed selective and nonselective NSAIDs.
The objective of this commentary is to review briefly the outcome of the US Food and Drug Administration (FDA) hearing in February 2005 and the subsequent actions taken, comment on the new safety data on celecoxib, and finally discuss the strengths and weaknesses of the MEDAL Program.
FDA hearing and actions taken
The FDA convened a Joint Arthritis and Drug Safety and Risk Management Advisory Committee in February 2005 to review the data on the COX-2 selective inhibitors. After 3 days of deliberations, the Joint Committee reached the following conclusions or made the following recommendations.2
Class effect
All FDA-approved COX-2 selective inhibitors—celecoxib, rofecoxib, and valdecoxib—increase the risk of serious cardiovascular events, primarily acute myocardial infarction, although celecoxib appeared to be less harmful than the others. New data on celecoxib discussed below question whether that is indeed the case.
Remain on the market
The committee supported leaving celecoxib on the market conditional on adding a Black Box warning, limiting direct-to-consumer advertising (DTCA) and introducing a treatment guide. The vote regarding valdecoxib was split. The FDA agreed with those opposed to letting it remain on the market, and Pfizer subsequently withdrew the drug.
Black Box warning
It was recommended that a Black Box warning be added to the labeling for celecoxib. The label was changed on July 29, 2005,3 and a softly worded Box was added, which describes in nonspecific terms what celecoxib may do. The tone is not discouraging, and the Box will hardly help patients make educated decisions about whether to take the drug.
DTCA
The excessive use of COX-2 selective inhibitors was a concern. A ban on DTCA was not deemed possible because it violates the right to commercial free speech. However, the regulations for DTCA require that Black Box warnings be included in any advertisement. This likely dampens the promotion value, and to date, there has been no resumption of DTCA for celecoxib.
Medication guide
One of the most important recommendations was the development of a medication guide. The objective of this tool is to inform and remind patients about the potential risks of a specific medication, and it is to be presented to patients each time they fill a prescription. The hope of the committee was that a properly worded guide might discourage some patients from using high doses of celecoxib. Unfortunately, as with the Black Box warnings, the guide requirement was later expanded by FDA to include all NSAIDs, and it does not specifically warn against the COX-2 selective inhibitors. The fact that these selective agents are associated with higher risk of serious cardiovascular events is not emphasized.4, 5, 6
Need for comparative trials
The shortage of informative outcomes from trials was obvious at the hearing. One explanation is the practical and ethical difficulty of conducting placebo-controlled trials of pain medications, especially long-term trials. Another is the manufacturers' apparent reluctance to conduct well-designed active control trials. A neutral or unimpressive outcome of a trial comparing an expensive new drug with an inexpensive generic drug could undermine future marketing efforts.
The cardiovascular risk associated with traditional NSAIDs is largely unknown, but some may increase the risk of myocardial infarction. If a COX-2 selective inhibitor is noninferior (equal) to a traditional NSAID in an active-control trial, we will not know whether both are harmful (and how harmful) or neutral. The committee recognized the dilemma and concluded by recommending naproxen as the “preferred” comparator among the traditional NSAIDs. After all, naproxen was associated with a lower risk of myocardial infarction in Vioxx Gastrointestinal Outcomes Research (VIGOR) trial,4 Therapeutic Arthritis Research and Gastrointestinal Event Trial,5 and Successive Celecoxib Efficacy and Safety Study-1 (SUCCESS-1) trial.6
New data on celecoxib
A recently published systematic review of 6 randomized double-blind trials comparing celecoxib to placebo, diclofenac, ibuprofen, or paracetamol (each with a minimum duration of 6 weeks) showed that the odds ratio for myocardial infarction was 1.88 (95% confidence interval [CI], 1.16-3.08).7 For the 4 placebo-controlled comparisons, the odds ratio was 2.26 (95% CI, 1.0-5.1). A particular strength of this review is its focus on longer term trials and the exclusion of a large number of uninformative, small, short-term trials in primarily low-risk subjects.
This negative report was supported by another recent clinical trial of 13,274 patients with osteoarthritis. Compared to diclofenac 50 mg bid or naproxen 500 mg bid, celecoxib 100 mg bid increased the risk of myocardial infarction (odds ratio, 5.0; 95% CI, 0.6-33).6
These new data demonstrate that the risk of myocardial infarction with celecoxib is similar in magnitude to the reported 2.24-fold risk with rofecoxib.8 The 33% increase in celecoxib use between the first and fourth quarter of 2005 raises safety concerns.
The MEDAL Program
The MEDAL Program8 has strengths and weaknesses. Its major strength is the sample size of 34,701 enrollees, with approximately 40,000 person-years of exposure and an anticipated 635 primary events. In contrast to most of the earlier trials, the cardiovascular events will be independently classified using prespecified diagnostic criteria. Only a modest proportion of study participants, approximately 11%, have a history of atherosclerotic cardiovascular disease. The program will provide additional information about health outcomes associated with long-term use of etoricoxib (not available in the United States) and diclofenac.
The program has several limitations. The selection of diclofenac as a comparator was made before the committee recommendation to use naproxen. It would clearly be an advantage to use the same drug (in a specified dose) as a comparator in all active control trials of NSAIDs. Among the available NSAIDs, diclofenac has more relative COX-2 activity than ibuprofen or naproxen.9 Hence, the comparison may not be very informative for most “traditional” NSAIDs.
The selection of a combined arterial and venous disease outcome is another limitation. Because the COX-2 selective inhibitors have no documented effect of increasing the risk of venous disease events, their inclusion in the primary outcome has 2 disadvantages. First, it will add noise to the comparison and dilute any differential treatment effect on the arterial disease events, and, second, it will dilute power by halting the program before reaching 635 arterial disease events.
The primary reliance on a per protocol analysis rather than an intention-to-treat analysis is also debatable. The latter underestimates differences, but the former introduces biases. There is no ideal solution, but the latter would be preferred because interpreting biased findings is not possible without knowing their directions and magnitude. The conclusion in the CONSORT statement10 that' there is greater confidence in results when the conclusions are consistent' is an endorsement of presenting both analytic approaches. Moreover, there is no scientific support at all for the proposed modified intention-to-treat approach.
Establishing an upper boundary for noninferiority is also a challenge. If 2 treatments are so similar that the upper boundary of the 95% CI excludes group differences of 10% to 15% or more, one would agree that the 2 are similar enough to be considered interchangeable. An upper boundary of 30% is perhaps large. Note that we get very excited when a new drug is reported to improve mortality or morbidity by 20%.
Conclusions
The FDA only partially adopted the recommendations from the FDA hearing. The emerging safety data on celecoxib are disturbing. The MEDAL Program will address parts of the major outstanding safety questions regarding the selective and nonselective NSAIDs.
References
- Cannon CP, Curtis SP, Bolognese JA, et al. Clinical trial design and patient demographics of the Multinational Etoricoxib and Diclofenac Arthritis Long-term (MEDAL) Study Program: cardiovascular outcomes with Etoricoxib vs. Diclofenac in patients with osteoarthritis and rheumatoid arthritis. Am Heart J [In press].
- Decision Memo—Analysis and Recommendations for Agency Action—COX-2 Selective and Non-selective NSAIDs. (Issued 4/6/2005, posted 4/15/05) http://www.fda.gov/cder/drug/infopage/COX2/NSAIDdecisionMemo.pdf.
- Celebrex label with Black Box warning. http://www.fda.gov/cder/foi/label/2005/020998s018,019lbl.pdf.
- Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. VIGOR Study Group. N Engl J Med. 2000;343:1520–1528
- Comparison of lumiracoxib with naproxen and ibuprofen in the Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET), cardiovascular outcomes: randomized controlled trial. Lancet. 2004;364:675–684
- Celecoxib versus naproxen and diclofenac in osteoarthritis patients: SUCCESS-I Study. Am J Med. 2006;119:255–266
- Risk of cardiovascular events and celecoxib: a systematic review and meta-analysis. J R Soc Med. 2006;99:132–140
- Risk of cardiovascular events and rofecoxib: cumulative meta-analysis. Lancet. 2004;364:2021–2029
- . Cyclooxygenases: new forms, new inhibitors, and lessons from the clinic. FASEB J. 2004;18:790–804
- Reporting of noninferiority and equivalence randomized trials. An extension of the CONSORT statement. JAMA. 2006;295:1152–1160
PII: S0002-8703(06)00470-4
doi:10.1016/j.ahj.2006.05.022
© 2006 Mosby, Inc. All rights reserved.
