COX-2 Inhibitor Use After Vioxx: Careful Balance or End of the Rope?

November 1, 2004
A. Mark Fendrick, MD Co-Editor in Chief

Volume 10, Issue 11 Pt 1

The American Journal of ManagedCare

This issue of examines interventions aimed at improvingdrug safety.1-4 The need for a rigorous researchagenda in pharmaceutical safety is highlighted byMerck's decision to voluntarily withdraw rofecoxib(Vioxx) on September 30, 2004. The cyclo-oxygenase-2(COX-2) selective inhibitor was pulled from the marketafter it was revealed that subjects receiving rofecoxibhad roughly twice the rate of myocardial infarction andstroke when compared to the rate in placebo controls.5

these innovative drugs were never intended toprovide pain relief superior to that of traditional nonsteroidalanti-inflammatory drugs (NSAIDs). Rather,the COX-2 selective inhibitors were meant to reducethe risk of gastrointestinal adverse events, commonwith NSAIDs, while affording patients symptomaticrelief comparable to that of the NSAIDs.

This widely publicized recall prompted tremendousmedia attention, much of which focused on financialmarkets and decision making by the US Food and DrugAdministration (FDA). Although tens of millions ofAmericans were using rofecoxib and the other 2 FDA-approvedCOX-2 selective inhibitors (celecoxib [Celebrex]and valdecoxib [Bextra]) when the withdrawal wasannounced, media reports provided little information toassist patients and their clinicians regarding therapeuticoptions. Largely lost in the firestorm over the cardiovascularsafety of COX-2 selective agents was the factthat Thus, the useof the available treatment alternatives in the "post-Vioxx" era must balance competing risks and benefits.

The New England Journal of Medicine

"Perspective" articles in the October 21, 2004, issueof called forextensive reexamination of the safety of the COX-2selective agents, but also of how the entire class of drugshas been approved and marketed.6,7 The following day,the European Medicines Agency announced that itwould reexamine all aspects of the cardiovascular safetyof the COX-2 inhibitors to assess whether changesneed to be made to existing labeling and whether additionalstudies are needed. In the meantime, millions ofpatients looked for counsel as to whether they shouldcontinue taking their COX-2 inhibitors.

Given the undesirable cardiovascular findings forrofecoxib, the lack of long-term, placebo-controlled cardiovascularsafety studies for either celecoxib or valdecoxibraises concern regarding continued use of theseother COX-2 selective agents. Fitzgerald emphasizedthis point, stating "the burden of proof now rests withthose who claim that this is a problem for rofecoxibalone and does not extend to other coxibs. We mustremember that the absence of evidence is not the evidenceof absence."7

COX-2 selective agents were developed and prescribedwith the best intentions, to relieve patients' pain and disability.Nevertheless, clinicians must always be awarethat treatment decision making has always been, andlikely always will be, about balancing potential benefitsagainst potential risks. With the alarming revelation thatmany COX-2 users may be at increased risk for myocardialinfarction or stroke, this tightrope balancing act is inthe spotlight.

Knowing what we now know, clinicians should beginby assessing each patient's risk factors for cardiovascularevents and consider initiating aspirin prophylaxisbefore prescribing pain relief. User-friendly tools areeasily accessed from the Internet to assess cardiovascularrisk and indications for aspirin.8 For patients with anincreased risk of myocardial infarction and stroke, noneof the COX-2 selective agents may be the ideal paintreatment for 2 important reasons. First, long-termcardiovascular safety data for the available COX-2selective agents in at-risk patients are currently lacking.6,7 Second, the gastrointestinal risk-reducing effectof COX-2 agents is diminished, if not eliminated, withconcomitant aspirin use, an underappreciated fact.9-11One national survey reported that more than 50% ofchronic COX-2 users over the age of 55 took regularaspirin therapy,12 potentially negating the safety advantageof this class of drugs in a majority of users.

For patients who do not have cardiovascular risk factorsthat require aspirin prophylaxis and who are at lowor no risk for gastrointestinal complications (Table,upper left quadrant), monotherapy with one of the traditionalnonselective NSAIDs offers a sound approach toanti-inflammatory therapy. For patients who do notrequire aspirin prophylaxis and are at risk for NSAID-inducedgastrointestinal complications (upper rightquadrant), a suitable initial approach would be to prescribeeither a COX-2 selective inhibitor or a traditionalNSAID and a proton pump inhibitor (PPI). In arandomized, controlled study of patients at high risk forgastrointestinal complications, recurrent gastrointestinalbleeding rates were not different for those taking atraditional NSAID and a PPI compared to a COX-2agent alone.13 While no statistical differences were seenbetween the 2 treatment arms, the high rates of eventsin both groups (approximately 5% at 6 months) drawsattention to the important clinical point that a "safer"treatment option does not mean completely safe.

For NSAID users with cardiovascular risk whorequire aspirin prophylaxis, COX-2 selective agentsshould be avoided for the reasons noted above. Since itis well established that multiple NSAID use (ie, aspirinand any NSAID) raises the likelihood of gastrointestinalcomplications,14 prophylaxis against the development ofgastrointestinal adverse events may be warranted.Recommending that COX-2 users at risk for cardiovascularevents simply switch to traditional NSAIDs ignoresthe profound clinical and economic effects of NSAID-relatedgastrointestinal adverse events, the reason theCOX-2 agents were developed. For individuals with nogastrointestinal risk factors (lower left quadrant), a traditionalnonselective NSAID should be prescribed witha PPI if the absolute risk of a gastrointestinal complicationwarrants gastroprotection. For the large number ofindividuals at risk for cardiovascular and gastrointestinaladverse events (lower right quadrant), if a traditionalNSAID must be used (in lieu of non-NSAID therapy),the additional measure of prescribing a gastroprotectiveagent is strongly recommended.

In light of the Vioxx withdrawal and the subsequentreassessments of COX-2 agents underway to definitivelyestablish cardiovascular safety, clinicians should balancethe benefits and risks of available agents in termsof pain relief and cardiovascular and gastrointestinalsafety. Until the burden of safety is met, we should considertaking our patients "back to the future" and useolder, well-studied agents to safely provide symptomaticrelief.

—A. Mark Fendrick, MD

Co-Editor in Chief

Am J Manag Care.

1. Sebaldt RJ, Petrie A, Goldsmith CH, Marentette MA. Appropriateness ofNSAID and coxib prescribing for patients with osteoarthritis by primary care physiciansin Ontario: Results from the CANOAR study. 2004;10:742-750.

Am J Manag Care.

2. Solberg LI, Hurley JS, Roberts MH, et al. Measuring patient safety in ambulatorycare: Potential for identifying medical group drug–drug interaction rates usingclaims data. 2004;10:753-759.

Am J Manag Care.

3. Fick DM, Maclean, JR, Rodriguez NA, et al. A randomized study to decreasethe use of potentially inappropriate medications among community-dwelling elderlyadults in a southeastern managed care organization. 2004;10:761-768.

Am JManag Care.

4. Parente ST, Kim SS, Finch MD, et al. Identifying patterns of controlled substanceutilization requiring evaluation using administrative claims data. 2004;10:783-790.

5. Merck announces voluntary worldwide withdrawal of VIOXX® [press release].Whitehouse Station, NJ. September 30, 2004. Available at: Accessed October 28, 2004.

N Engl JMed.

6. Topol EJ. Failing the public health – rofecoxib, Merck, and the FDA. 2004 Oct 21;351(17):1707-1709.

N Engl J Med.

7. Fitzgerald GA. Coxibs and cardiovascular disease. 2004 Oct21;351(17):1709-1711.

8. National Cholesterol Education Program. Risk Assessment Tool for Estimating10-year Risk of Developing Hard CHD. Available at: Accessed Oct. 14, 2004.


9. Silverstein FE, Faich G, Goldstein JL, et al. Gastrointestinal toxicity with celecoxibvs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoidarthritis. The CLASS study: A randomized controlled trial. 2000;284:1247-1255.


10. Schnitzer TJ, Burmester GR, Mysler, et al. Comparison of lumiracoxib withnaproxen and ibuprofen in the Therapeutic Arthritis Research and GastrointestinalEvent Trial (TARGET), reduction in ulcer complications: randomised controlledtrial. 2004 Aug 21;364(9435):665-674.


11. Laine L, Maller ES, Yu C, Quan H, Simon T. Ulcer formation with low-doseenteric-coated aspirin and the effect of COX-2 selective inhibition: a double-blindtrial. 2004 Aug;127(2):395-402

Arch InternMed.

12. Cox ER, Frisse M, Behm A, Fairman KA. Over-the-counter pain reliever andaspirin use within a sample of long-term cyclooxygenase 2 users. 2004 Jun 14;164(11):1243-1246.

N Engl J Med.

13. Chan FK, Hung LC, Suen BY, et al. Celecoxib versus diclofenac andomeprazole in reducing the risk of recurrent ulcer bleeding in patients with arthritis. 2002 Dec 26;347(26):2104-2110.


14. Fendrick AM, Garabedian-Ruffalo SM. A clinician's guide to the selection ofNSAID therapy. 2002;27:579-582.