A tale of two coxibs

Theresa Hegmann, MPAS, PA-C

Ms. Hegmann is Clinical Assistant Professor and Director of Curriculum and Evaluation, Physician Assistant Program, Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City. The author has indicated no relationships to disclose relating to the content of this article.

What’s up with the FDA lately? The regulatory agency seems to have been in the news repeatedly as one high-profile drug after another gets removed from the market or has a new “black box” warning added to its labeling. Remember troglitizone (Rezulin), pulled from the market in March 2000, after its liver safety had been debated for 3 years? Troglitazone generated more than $2 billion in sales for its manufacturer and was associated with at least 90 cases of liver failure in diabetic patients before it was withdrawn.¹ Or what about cerivastatin (Baycol), withdrawn by its maker in August 2001 because of increased rates of rhabdomyolysis? In the course of court proceedings, it surfaced that the company likely was aware of the increased risk of rhabdomyolysis as early as 4 months after cerivastatin was first approved in June 1997.^2,3 More recently, the very popular cyclooxygenase2 (COX2) inhibitor medications have been in the news. Rofecoxib (Vioxx) was pulled from the market in September 2004, and then in December 2004 we heard that the National Cancer Institute had decided to stop their “Adenoma Prevention with Celecoxib” trial prematurely because there was evidence of increased risk of MI, cardiovascular (CV) death, and stroke in patients taking celecoxib (Celebrex).

A marketing success story

The COX2 subclass of NSAIDs was supposed to be safer than older, nonspecific NSAIDS—or that is what we have been led to believe since 1998, when the first of these medications, celecoxib, was approved by the FDA. These drugs were marketed as having a lower risk for GI side effects such as ulceration, so clinicians have been prescribing them to many high-risk elderly patients with osteoarthritis. By the time Merck pulled rofecoxib from the market, more than 80 million patients had taken it, more than 10 million prescriptions per month were being written for it in the United States alone, and annual sales had reached $2.5 billion—an almost unimaginable figure for a single pharmaceutical.⁴

Given the huge number of patients taking COX2 inhibitors, even a slight excess of serious adverse drug reactions (ADRs) could represent a major public health issue. Using numbers from the colon polyp prevention trial (the Adenomatous Polyp Prevention on Vioxx, or APPROVe, study) that led to the withdrawal of rofecoxib, the observed excess of 16 MIs or strokes per 1,000 patients would translate into an additional 160,000 patients with MI or stroke for every 10 million patients taking rofecoxib.⁴ Unfortunately, this may underestimate the actual risk. The trial involved a lowrisk population without any known CV disease, whereas COX2 inhibitors are often preferentially prescribed to older, higher-risk patients in whom CV disease is common.

Credible warnings ignored

Why has it taken 6 or 7 years of prescribing COX2 inhibitors to our highest risk patients before we learn that these medications may have serious CV toxicity? Evidence of toxicity was available in 2001, when Mukherjee and colleagues compared the risk of thrombotic CV events in patients taking rofecoxib and naproxen in a major randomized trial involving 8,076 people (the Vioxx Gastrointestinal Outcomes Research, or VIGOR, trial) and calculated a relative risk of 2.38 for the rofecoxib group (CI 1.39-4.00, P=.002).⁶

Although this evidence could not be considered conclusive, it definitely should have raised a red flag. In fact, the authors concluded, “. . . we believe it is mandatory to conduct a trial specifically assessing CV risk and benefit of these agents. Until then, we urge caution in prescribing these agents to patients at risk for CV morbidity.” Unfortunately, no further trials were undertaken to evaluate the CV effects of rofecoxib or celecoxib, and it was an incidental finding in a trial with a completely different purpose—to evaluate the use of rofecoxib in the prevention of colon polyps—that finally led to the withdrawal of rofecoxib from the market.

Physiologic mechanisms

Can a reasonable physiologic explanation account for the observed increase in CV toxicity with rofecoxib and possibly other COX-2 inhibitors? This question is critical in determining whether these ADRs are likely to be a class effect.

COX-2 medications were developed based on the discovery in the early 1990s of two distinct isoforms of cyclooxygenase, both of which are involved in the production of prostaglandins. Initially, COX-1 was believed to be the most important source of prostaglandins involved in protection of the stomach lining and in the effects on platelets and endothelium, while COX-2 was considered the main culprit in causing inflammation and pain. Thus, it made sense at the time to design a drug targeted specifically to inhibit COX-2: A designer drug might be expected to block inflammation without as much risk for stomach ulceration.⁷

We now know that COX-2, not COX-1, is actually the dominant source of prostaglandin I2 in the endothelium responsible for vasodilatation and inhibition of platelet aggregation. On the other hand, thromboxane A2, a major COX-1 product in platelets, leads to platelet aggregation and vasoconstriction. Thus, selective inhibition of COX-2 might inhibit prostaglandin I2 formation while at the same time failing to suppress thromboxane A2. This is a potentially deadly combination—essentially an “antiaspirin effect”—that could account mechanistically for some of the observed side effects of COX-2 inhibitors, including elevated BP and an increased tendency to thrombotic events such as MI or stroke.⁷

With a physiologic mechanism to account for the CV toxicity of COX-2 inhibitor medications, it must be assumed that the CV risks that led to the withdrawal of rofecoxib may be a class effect. Because neither celecoxib nor valdecoxib (Bextra) has shown convincing evidence of decreased GI side effects, and because COX-2 inhibitors as a class have never been shown to be more effective than the nonselective NSAIDs for controlling pain or inflammation, it is difficult to argue for preferentially prescribing COX-2 inhibitors to any group of patients, even those at low risk for CV disease.

Problems at the FDA?

As providers, we rely on the FDA to decide whether drugs are safe and effective for their intended use before they go on the market and, in general, that part of the system works reasonably well. We must remember that regulatory approval for clinical use does not (and could never) guarantee the safety of a medication. The drug approval system must balance the societal need for access to promising new medications with the ever-present risk of missing rare but dangerous side effects.

Under the current system, adverse reactions occurring in 1% or more of patients are usually obvious before a drug goes to market.⁸ On the other hand, the existing system is not good for characterizing rare adverse reactions that may not show up until many thousands or hundreds of thousands of patients have been exposed. To gather this sort of information, our country relies heavily on a passive surveillance system called MEDWATCH, in which prescribers are encouraged to voluntarily report ADRs. In contrast to our highly structured and well-funded system for premarketing evaluation of pharmaceuticals, the postmarketing surveillance system is outdated, haphazard, and plagued by vast underreporting and built-in biases.

In the current system, the FDA does not have a clear mandate, nor sufficient funding, to actively investigate suspected ADRs. Instead, it relies on pharmaceutical manufacturers to collect, review, and analyze information regarding suspected ADRs that show up after a medication goes to market. In such a system, we should not be surprised that it took years for rofecoxib to be withdrawn from the market. The real problem with the system is not that a postmarketing ADR showed up but rather that it has taken so long for the FDA to investigate it and act on the issue. During that time, mass marketing of a medication with an unproven safety profile magnified the potential for harm.

Many of the experts who have reviewed the history of the development and marketing of the COX-2 inhibitors have come to essentially the same conclusion—that the postmarketing surveillance system for medications in this country is antiquated and fraught with conflicts of interest.^2,3,9 At the very least, legislation is needed to authorize the FDA to require pharmaceutical companies to conduct postmarketing trials as necessary to ensure drug safety. An even better option is the establishment of an independent agency to monitor the safety of medications, with the long-term goal of establishing an active, prospective, and systematic approach to postmarketing drug surveillance in the United States.

REFERENCES

1. Gale AM. Lessons from the glitazones: a story of drug development. Lancet. 2001; 357:1870-1875.

2. Psaty BM, Furberg CD, Ray WA, Weiss NS. Potential for conflict of interest in the evaluation of suspected adverse drug reactions. JAMA. 2004;292:2622-2631.

3. Fontanarosa PB, Rennie D, DeAngelis CD. Postmarketing surveillance—lack of vigilance, lack of trust. JAMA. 2004;292:2647-2650.

4. Topol EJ. Failing the public health—rofecoxib, Merck, and the FDA. N Engl J Med. 2004;351:1707-1709.

5. Solomon DH, Schneeweiss S, Glynn RJ, et al. Relationship between selective cyclooxygenase2 inhibitors and acute myocardial infarction in older adults. Circulation. 2004;109:2068-2073.

6. Mukherjee D, Nissen SE, Topol EJ. Risk of cardiovascular events associated with selective COX2 inhibitors. JAMA. 2001;286:954-959.

7. Fitzgerald GA. Coxibs and cardiovascular disease. N Engl J Med. 2004;351:1709-1711.

8. Meadows M. Why drugs get pulled off the market. FDA Consumer. 2002;36(1): 11-17.

9. Topol EJ. Arthritis medicines and cardiovascular events—“house of coxibs.” JAMA. 2005;293:366-368.