• 4/19/2005
  • Cori Vanchieri
  • Journal of the National Cancer Institute, Vol. 97, No. 8, 552-553, April 20, 2005

Several cyclooxygenase 2 (COX-2) inhibitors have taken a beating from recent evidence that they can cause deaths from heart attack and stroke, but some cancer researchers say that the drugs are too promising as possible chemoprevention agents to abandon them completely.

Rofecoxib (Vioxx) was withdrawn from the market in September when a colon cancer chemoprevention trial revealed an increased risk of heart attacks and strokes among long-term users (see News, Vol. 96, No. 23, p. 1734, “Vioxx Withdrawal Alarms Cancer Prevention Researchers”). Soon after, valdecoxib (Bextra) was shown to increase heart attacks in people who recently had coronary artery bypass surgery.

Bad news for celecoxib (Celebrex), the COX-2 inhibitor used most frequently in prevention trials, came in December: It also increased the risk of heart attacks in a cancer prevention trial. Drug administration was stopped on dozens of cancer prevention trials, other studies were halted, and a U.S. Food and Drug Administration hearing was planned.

Two FDA panels—the Arthritis Drug Advisory Committee and the Drug Safety and Risk Management Advisory Committee—met jointly in mid-February. Overall, they viewed the cardiovascular risk as a class effect among the group of COX-2 inhibitors. But they voted to keep on the market the two drugs that are still being sold—celecoxib and valdecoxib—with new black box warnings and limits on direct-to-consumer marketing. They also voted narrowly to leave the door open for Merck to return rofecoxib to the market. The FDA is expected to decide soon whether to follow the committees’ recommendations.

The decision, if finalized, means that the cancer community can consider studying COX-2 inhibitors again. The National Cancer Institute is planning how to speed reporting of efficacy data and get more safety data. It is also reviewing existing trial designs with institutional review boards (IRBs) to see which studies can move forward. COX-2 researchers are now hedging their bets on the drugs, determined to focus their efforts on the most dangerous precancerous lesions and test lower doses of celecoxib in combination with other drugs.

“We still think that [nonsteroidal anti-inflammatory drugs (NSAIDs)] and COX-2 selective inhibitors may be very useful in both treating cancer and preventing cancer’s development in many organs,” said Ernest Hawk, M.D., Ph.D., director of NCI’s Office of Centers, Training, and Resources. “We now know of a risk for cardiovascular events that we didn’t know of previously. But even with that, they still may be useful in some cohorts.”

Does it Work in Cancer?

“Getting the efficacy data is absolutely critical,” Hawk added. Results from the Adenoma Prevention with Celecoxib (APC) trial, a study of more than 2,000 people who have had adenomas removed, are due this year. The APC was stopped in December after the trial data indicated an increased cardiovascular risk with celecoxib, but by that point, most of the participants had been taking the drug or placebo for close to the full 3-year period. The last colonoscopies should be finished by March. The original plan was to publish the efficacy data in fall 2005; Hawk said he intends to accelerate that timeline.

To help answer the lingering questions about celecoxib’s effect on cardiovascular health, the National Institutes of Health is also planning a meta-analysis across five large, placebo-controlled long-term celecoxib trials plus a sixth study in a high-risk group—people with diabetic retinopathy. NIH hopes to get this done before the end of 2005. Hawk said he also hopes to do an extension study of the APC.

With most of these agents, there appears to be a dose response for the cardiovascular effects. “Dose is an issue in most minds,” Hawk said. “It’s unknown precisely what dose would be safe, if there is one.”

At the FDA hearing, Hawk presented the data from the APC trial, in which participants were randomly assigned to take either 200 mg of celecoxib twice a day, 400 mg of the drug twice a day, or placebo for 3 years. A dose-dependent risk of death from cardiovascular causes, myocardial infarction, stroke, or heart failure emerged: the risk was 2.3-fold among those taking the lower dose (7.8 events per 1,000 patient-years) and 3.4-fold for people taking the higher dose (11.4 events per 1,000 patient-years, compared with 3.4 events per 1,000 patient-years among those taking placebo).

The APC results were published online February 15 in the New England Journal of Medicine, along with the results of the APPROVe trial, which tested daily use of high-dose rofecoxib for polyp prevention. Rofecoxib also increased cardiovascular events; there were 1.50 events per 100 patient-years in the rofecoxib group and 0.78 events per 100 patient-years in the placebo group.

Which Cohorts Are Worth Pursuing?

Raymond N. DuBois, M.D., Ph.D., director of the Vanderbilt-Ingram Cancer Center in Nashville and a longtime COX-2 investigator, admits that the upswell of concern has dampened COX-2 research in cancer. “It makes it much harder to think about it for widespread use, because of those side effects.”

So he and others are retooling priorities for this class of drugs, focusing on people most at risk for disease. Andrew J. Dannenberg, M.D., director of cancer prevention at Weill Medical College of Cornell University in New York, plans to begin a study of celecoxib in patients with aneuploid leukoplakia, a precursor to oral cancer. “These patients have an extraordinarily high risk of death within 5 years of diagnosis,” he said. “In this cohort, I think it is entirely reasonable and the science is sound to explore the potential utility of COX-2 inhibitors in altering the natural history of that disease to delay development of oral cancer.”

He said he remains excited about the prevention possibilities in cancers of the colon, oral cavity, breast, and bladder. “Bladder cancer is one of the best, though underappreciated, opportunities,” he added. Preclinical data indicate that celecoxib can prevent recurrence of superficial bladder cancer, which recurs frequently. “Preventing or delaying a recurrence of a frank malignancy would represent a significant step forward.” He also published a retrospective study in 2004 indicating that frequent aspirin use is associated with a decreased risk for estrogen receptor–positive breast cancer.
DuBois plans to study low-dose celecoxib—200 mg once per day—with another agent, perhaps an epidermal growth factor receptor tyrosine kinase inhibitor such as erlotinib (Tarceva). In mouse models of colon cancer, the two together were “very effective. The mice haven’t developed any polyps in 5 or 6 months,” he said. “We’ve never seen anything that effective in preclinical studies.”

Vanderbilt has several phase I and phase II studies of celecoxib in treatment and prevention that are all still on hold, Dubois said. Their IRBs are reviewing the risks and benefits now. “It’s a matter of sorting out where NIH is willing to go and where the drug companies are willing to go,” he said.

Hawk said NCI is deciding trial by trial which studies should be restarted. In the high-risk familial adenomatous polyposis (FAP) cohorts, NCI staff is meeting with IRBs to reinstitute the trials combining celecoxib with eflornithine and more cardiovascular monitoring, he said. All trials will have modified informed consent forms. In the situations where cancer risk is very low and competing strategies are effective—for example, in actinic keratosis—NCI is not planning to restart studies. “In many other settings, however, we are planning to reinstitute the trials as designed. We may adjust the dose in some cases, or the exclusion criteria,” he said. “It makes sense, for example, to make sure no one with uncontrolled hypertension or a strong family history” is enrolled.

Researchers are also trying to tease out a heart disease risk factor that could single out people who should stay away from the drugs. An analysis of the APC data did not reveal one. They had baseline data on age, gender, diabetes, smoking, and use of aspirin and lipid-lowering agents. Some of the factors—smoking and age—appeared to affect risk, but none made a statistically significant difference. Overall, the APC population was at elevated risk for heart disease. Their average age was close to 60 and about 45% had high blood pressure. About one-quarter were taking lipid-lowering agents.

If hypertension proves to be an important explanation for why COX-2 inhibitors are associated with increased cardiovascular disease risk, then more effective treatment for hypertension may be the answer, said Dannenberg, who attended the FDA hearing as a consultant to Pfizer.

NSAIDs and COX-2 Inhibitors

Laboratory studies continue to differentiate the various types of COX-2 inhibitors. Dannenberg published a study in the March 1 issue of Clinical Cancer Research that he says distinguishes celecoxib from rofecoxib. In prostate cancer cells that do not express COX-2, celecoxib stifled replication of the cancer cells; rofecoxib did not. He reported that celecoxib worked by lowering levels of cyclin D1, a protein essential to cell proliferation. “These results support the notion of a unique action by celecoxib that is independent of COX-2, and that’s different from Vioxx,” Dannenberg said.

Although the advisory committees decided that the cardiovascular risk is a class effect for the COX-2 inhibitors, the question they couldn’t answer is “how large is the class?” Do the traditional NSAIDs, such as naproxen (Aleve and Naprosyn) and ibuprofen (Motrin and Advil), belong in the group as well? There are hints that they do, but there are no published data yet. The Alzheimer’s Disease Anti-Inflammatory Prevention Trial (ADAPT), a study testing the effect of naproxen, celecoxib, and placebo on Alzheimer’s disease prevention, may be able to answer that question, at least for naproxen. That trial was also halted in December, but cardiovascular results are still being tabulated.

Dannenberg said an important point went underreported from the FDA hearing. “The advisory committees clearly expressed great concern about traditional NSAIDs having similar cardiovascular risk,” he said, plus they carry an increased risk of gastrointestinal toxicity. He predicted that in the next year, with intensified scrutiny, the risk of traditional NSAIDs will be validated. He has a paper under review of a retrospective study showing that smokers who took NSAIDs long term had a reduced risk of oral cavity cancer but they also had an increased risk of cardiovascular-related death. “That message needs to be shared as well.”