The COX-2 Story and Cancer

This article has been archived.
Please use for reference only.

Julia Draznin Maltzman, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: January 2, 2005


In the past few months three big selling pain relievers have come under immense public scrutiny; rofecoxib (Vioxx, Merck), celecoxib (Celebrex, Pfizer), valdicoxib (Bextra, Prizer). It all began September 30 th when Merck and Co voluntarily withdrew its acute pain medication drug, Vioxx. The announcement caused Merck stock prices to drop and the company lost more than a quarter of its market value in a single day. The trial that led to this shocking announcement was a cancer prevention trial called the APPROVe trial.

Mechanism of action

All non-steroidal anti-inflammatory drugs (NSAID) act by inhibiting the cyclooxigenase enzyme, known as COX for short. The older NSAIDs, however, inhibit all COX enzymes, the most important of which are COX-1 and COX-2. As these medications gained prominence and use, scientists realized that bleeding stomach ulcers, the major side effect seen with NSAIDs, were a direct result of the COX-1 enzyme inhibition. This led to the development of a newer class of drugs that inhibit only the COX-2 enzyme. Vioxx, Celebrex, and Bextra belong to this new group of COX-2 inhibitors. COX-2 inhibitors still decrease pain and inflammation effectively, but with a much lower incidence of bleeding ulcers and GI toxicities. No doubt that with the introduction of the COX-2 inhibitors there has been a drastic reduction in the number of hospitalizations and deaths due to NSAID induced ulcers and GI bleeds, however, these drugs are not without their own ill effects.

The active Cyclooxigenase 2 enzyme stimulates the production of prostacyclins. Prostacyclins are fatty acids that perform a number of hormone-like tasks in the body including promotion of inflammation, dilatation of blood vessels, and inhibition of platelet aggregation. By using these medications and blocking the COX-2 enzyme, prostacyclins production is stopped, inflammation is decreased, blood vessels are not dilated, and platelet aggregation is promoted. For people with arthritis, the decrease in inflammation is a positive effect, however, the promotion of platelet aggregation seems to have become a bigger issue than originally envisioned.

The cancer connection

The connection between cancer and inflammation is not new. More recently, however, multiple studies have documented overexpression of the COX-2 enzyme in several cancers including breast, colon, lung, prostate, and bladder. Based on these preclinical observations, scientists have reasoned that disrupting the COX-2 enzyme with an inhibitor may improve treatment and even prevention efforts. Last year more than a dozen studies examining Vioxx and Celebrex in people at high risk for lung, breast, prostate, colon, and bladder cancers were on the way. In addition, COX-2 inhibitors were also being studied along with standard treatments in patients who already had cancer. One particularly tantalizing study observed that that COX-2 inhibitors virtually eliminated many colon polyps in people who are predisposed to get them. Colon polyps lead to colon cancer and are removed when found on colonoscopic examination. Based on these observations the Adenomatous Polyp Prevention on Vioxx trial (APPROVe) was opened. This multi-institution effort randomly assigned healthy individuals who have already had polyps removed from their colons to take either Vioxx or placebo with the aim to understand if the COX-2 inhibitor prevented subsequent polyp growth. The answer to this question is not yet known and scientists hope to release preliminary information next spring.

Ironically, the same trial that hoped to show an extended application and efficacy for Vioxx, led to its demise. The participants taking the COX2 inhibitor had more heart attacks and strokes than those taking the placebo. A similar trial looking at polyp formation comparing Celebrex and placebo showed a similar side effect profile.

The APPROVe trial began in 2000 and enrolled over 2,500 patients. There were 25 confirmed cases of cardiovascular events in the patients taking placebo as opposed to the 45 cases in the group of patients on Vioxx. This calculates to a relative risk of 1.96 of a cardiovascular event on the drug with a significant p value (p=0.007). This risk became apparent only after 18 months of treatment. It is important to note that, to date, there is no difference in the overall mortality for both groups. On the other hand, about three quarters of patients completed three years of therapy and study participation before the study was stopped prematurely. The polyp data is currently being analyzed and will certainly yield very important information.

The big concern

The largest concern that has emerged is whether or not this cardiovascular risk is a class effect or is it limited to a few isolated COX-2 inhibitors? With the latest information on Celebrex, the uneasy suspicion is that it may in fact be a class effect. Merck’s shocking news in September prompted the NIH to scrutinize all trials using the COX-2 inhibitors. In mid December another cancer prevention clinical study was shut down because of the increased cardiovascular risk posed by Celebrex.

A five-year Celebrex trial by the National Cancer Institute was looking at polyp prevention in high-risk individuals. The study participants were divided into three groups. Two groups received Celebrex either 200mg twice daily or 400mg twice daily and the third group received placebo. A data safety monitoring committee looked at the interim results and determined that patients taking the higher dose of Celebrex had 3.4 times the risk of having a cardiovascular event. Individuals taking the lower dose more than doubled their risk of an event over placebo. In total, there were six cardiac events in the placebo group, 15 in the lower dose Celebrex group, and 20 events in the higher dose group, within 33 months of follow up. Unlike Merck, Pfizer did not withdraw its drug off the market as it cited other trials that contradicted these observations. Similarly, the FDA has not (to date) required Pfizer to withdraw its drug from the market.

Implications for cancer studies

The effect of the withdrawal of COX-2 inhibitors on cancer patients is not positive. Since this trial showed an increase in cardiovascular events in the preventive setting, the jury is still out as to the benefit of this class of medications in the treatment of cancer. As with all medications, the balance between the benefits and possible risks must be considered prior to initiation of therapy.

Depending on the severity of cardiovascular risk, some high-risk cancer patients may find that their risk of getting cancer outweighs the risk posed by COX-2 inhibitors. There may, in fact, be a narrow population of individuals for whom these drugs are appropriate if not necessary.

The recent warnings we have been hearing about the COX-2 inhibitor class are very real and should be heeded; however, these observations are not as definitive as a randomized controlled clinical trial. The rigor that we use to prove drug efficacy was not applied in this case when risk was observed. To some extent this is correct as it would be unethical and a waste of patient resources to conduct a prospective randomized clinical trial to evaluate the cardiovascular risk for all COX-2 inhibitors. However, these data should be interpreted carefully. It is too early to make strong conclusions with certainty and a blanket withdrawal of all COX-2 inhibitors may prove to be more counter-productive in the long run.

Finally, I will end by a reminder to: “never say never”. In the 1960’s, the sleeping aid Thalidomide was taken off the market due to the severe birth defects it caused when taken by pregnant women. Today, it is a drug with great potential in the treatment of many malignancies. Many patients with multiple myeloma have derived a great benefit – including prolongation of survival -- from using a drug that was banned 40 years ago.