FDA Center for Drug Evaluation and Research

division of analgesic, anti-inflammatory, and ophthalmologic drug products, HFD-550,

9201 Corporate Blvd, Rockville MD 20850

Tel: (301) 827-2040



DATE: January 19, 2005

TO: Members and Consultants

Arthritis and Drug Safety and Risk Management Advisory Committees

FROM: Jonca Bull, M.D., Director

Office of Drug Evaluation V, Office of New Drugs

Paul Seligman, M.D., M.P.H., Director

Office of Pharmacoepidemiology and Statistical Science

RE: Overview of the February 16th, 17th, and 18th Arthritis and Drug Safety and Risk Management Advisory Committees Meeting to Discuss the Overall Benefit to Risk Considerations for COX-2 Selective Nonsteroidal Anti-Inflammatory Drugs And Related Agents.

Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most commonly used drug products, indicated for conditions ranging from the management of acute pain and primary dysmenorrhea, to the relief of the signs and symptoms of rheumatoid arthritis and osteoarthritis. The known toxicities associated with NSAIDs include gastrointestinal ulceration, hemorrhage and perforation, renal impairment, hepatic effects, fluid retention, and hypertension. Of the two known isoforms of the cyclooxygenase enzyme, COX-1 has been thought to be responsible for the maintenance of the mucosal integrity in the upper gastrointestinal tract and for maintaining normal platelet function, while COX-2 has been thought to be produced in response to stimuli and responsible for inflammation and pain. The development of COX-2 selective NSAIDs was expected to provide an improved safety profile over nonspecific NSAIDS, characterized by fewer clinically important gastrointestinal adverse events. Concerns about the potential for cardiovascular risk arose during pre-clinical testing when it was determined COX-2 is also produced in vascular endothelium, as well as other tissues. This concern resulted in the addition of cardiovascular safety outcome monitoring in clinical trials for rofecoxib (Vioxx), a highly selective COX-2 inhibitor.

The first clinical evidence of the potential for cardiovascular risk in humans of COX-2 selective drugs came in 2000, with the results of Merck’s Vioxx Gastro Intestinal Outcomes Research Trial (VIGOR). In this study, patients with rheumatoid arthritis who had received 50 mg of Vioxx (twice the recommended chronic dose) were found to have a reduced risk of severe gastrointestinal toxicity, but to be at greater risk for cardiovascular thromboembolic events than patients treated with naproxen. Analyses of multiple other large studies of Vioxx in Alzheimer’s disease and osteoarthritis did not show this signal at the recommended dose of 25 mg. These data were discussed at the February 2001 Arthritis Advisory Committee meeting. A subsequent labeling change agreed to in April 2002, incorporated these findings. In September 2004, we received data from a long-term, placebo-controlled trial that further elaborated the cardiovascular risk, this time at 25 mg of Vioxx per day.

The APPROVe (Adenomatous Polyp Prevention On Vioxx) study revealed that those patients who received Vioxx had twice the risk for cardiovascular thromboembolic events than those who received placebo, with risk emerging after 18 months of exposure. These data have resulted in a reconsideration of the benefit /risk balance not only of VIOXX, but of COX-2 selective NSAIDs in general.


In December 2004, preliminary data from one long-term placebo controlled trial of Celebrex to prevent adenomatous colon polyps (the APC trial) suggests there may be an increased risk of cardiovascular thromboembolic events from Celebrex. However, an identical study, the Prevention of Spontaneous Adenomatous Polyps (the preSAP trial), as well as earlier trials, did not demonstrate such risk.

Other COX-2 trials have revealed the occurrence of more cardiovascular thromboembolic events among coronary artery bypass graft (CABG) surgery patients who received parecoxib, a pro-drug of valdecoxib (Bextra), in the immediate postoperative period.

Many questions regarding cardiovascular risk have arisen as a result of these data, such as whether the observed cardiovascular risk is a class effect or is specific to particular drugs within the class. What is the mechanism of this effect? Hypotheses under consideration include a direct effect of loss of the influence of COX-2 on the endothelium, and an effect based on relatively small but sustained increases in blood pressure. Complicating the picture further is that prospective studies assessing cardiovascular risk relative to placebo from traditional, non-selective, NSAIDS have never been conducted. In December 2004, the National Institutes of Health (NIH) described an increased risk of cardiovascular events associated with the use of naproxen in comparison to Celebrex and placebo in an Alzheimer’s disease prevention study. The National Institutes of Health (NIH) has reported that these data contrast with other studies in which naproxen has appeared to have less risk than COX-2 selective NSAIDs (i.e., the VIGOR study). This raises the question as to whether there may be a similar risk for cardiovascular thromboembolic events associated with the nonselective NSAIDs.

Over the years that the COX-2 selective NSAIDS have been marketed, results from large epidemiology (case-control) studies have been mixed regarding the risk for cardiovascular events from these drugs. It is important to understand what we can learn from epidemiological studies, and how best to weigh their results in relation to controlled clinical trial data.

Based upon all of the information currently available, many in the public, the scientific community, and FDA have raised questions about whether there should there be continued marketing of COX-2 selective NSAIDs. What is your view? Is there a patient population for whom the risk is warranted, given the known potential for benefit? If COX-2 drugs continue to be marketed, how much and what kind of information is necessary in order to justify the marketing of a new COX-2 selective NSAID? What about the non-selective NSAIDs? What are the ethical and clinical trial design hurdles that must be overcome in order to study these questions? These are the types of issues that we will ask you to address at the Advisory Committee meeting.

Over the three days of the upcoming meeting, you will hear presentations covering a broad range of topics to better inform your discussion. We will begin Day 1 with an overview of the risks of gastrointestinal adverse events from NSAIDs, as this is where the history and hope for COX-2 selective NSAIDS began. This will be followed by a presentation on possible mechanisms for the cardiovascular risk of COX-2 selective NSAIDs. Cardiovascular risk data will then be presented from the clinical trials of rofecoxib (Vioxx), celecoxib (Celebrex), valdecoxib (Bextra), and naproxen. On Day 2, you will hear presentations of data regarding the development programs of two COX-2 selective NSAIDs that have been submitted to the FDA, but are not approved for marketing in the U.S., etoricoxib and lumiracoxib. Findings from epidemiology studies of naproxen and the COX-2 agents will be also discussed focusing on study design features, as well as the strengths and limitations of these data for assessing cardiovascular risks. On Day 3, you will hear perspectives on broad population based interpretations of the cardiovascular risk related to these drugs based on available data. We will ask you to help us consider how to take these into account in facing the challenges of drug development for future products in this area.

The accompanying background package contains information from multiple clinical trials and epidemiology studies, some of which will be presented in summary form at the Advisory Committee meeting. These illustrate the many challenges of discerning drug-related cardiovascular risks among patients with multiple other risk factors for myocardial infarction and other cardiovascular events.

Over the next few weeks we will be drafting questions for your consideration during the meeting in order to facilitate your preparation for the discussions. We look forward to your participation and value your expert contributions in weighing the extensive and sometimes conflicting evidence about the benefits and risks of selective and nonselective NSAIDs.