If a clinical trial is halted halfway through, can it still provide useful data? That's a question facing an Alzheimer's disease prevention study, in which thousands of volunteers received either the painkillers Celebrex or naproxen, or a placebo. The trial, stopped at the end of 2004, suggests that naproxen is associated with a higher risk of heart and stroke problems than Celebrex is--a finding that contradicts many other published studies. Critics say the results cannot be trusted, in part because the trial ended so early.
When Vioxx was pulled from the market for safety reasons in the fall of 2004, it threw a world of clinical trials into turmoil. Vioxx belongs to a class of drugs known as COX-2 inhibitors, and although approved to treat pain from arthritis and other conditions, the drugs were also considered potential preventers of cancer and Alzheimer's disease because of their anti-inflammatory properties. Indeed, the trial that forced Vioxx out of circulation was designed to test the drug's ability to inhibit colon cancer. With Vioxx gone, researchers studying related drugs needed to weigh whether the possible risks were worth the potential benefits.
One of the largest prevention studies at the time was the Alzheimer's Disease Anti-inflammatory Prevention Trial (ADAPT), which was half-complete by late 2004. The trial had enrolled 2528 participants 70 years of age and older and divvied them into three groups: One got Celebrex, a COX-2 inhibitor; one got naproxen, an anti-inflammatory that's not COX-2 specific; and one got a placebo. After Vioxx was pulled, the trial's leaders and the National Institutes of Health, which was funding it, made the controversial decision to halt it, even though the trial's data safety monitoring board had recently recommended it continue. "It was almost a political pressure in this hysteria over COX-2," says Steven Abramson, a rheumatologist at New York University Hospital for Joint Diseases, who was not involved in the study.
Now, nearly 2 years later, the ADAPT team has published its safety data. (There's no word yet on the drugs' possible effectiveness against Alzheimer's.) Reporting today in PLOS Clinical Trials, the group writes that 37 out of 1070 volunteers in the placebo group and 28 out of 717 volunteers in the Celebrex group suffered heart or stroke-related problems, a similar percentage. In the naproxen group, however, the number was 40 out of 713. That works out to about 5.6% of those taking Celebrex or placebo, versus 8.3% of those taking naproxen, who would be predicted to have complications over a 3-year time period.
The results contradict a number of studies that show that naproxen has fewer, if any, of the cardiovascular risks of Celebrex. In fact, two studies published earlier this year in the Journal of the American Medical Association and the British Medical Journal, which examined hundreds of studies of anti-inflammatory drugs, found no increased risk of naproxen.
First author Barbara Martin, an epidemiologist at Johns Hopkins University in Baltimore, Maryland, admits the results are not conclusive. "I think the jury's still out" on whether traditional anti-inflammatories such as naproxen are risky, she says. One issue, says Martin, is that the cardiac events reported in the trial were not assessed by outsiders, as is common. Instead, those running the trial took the words of volunteers at face value. There are more problems with the study, says Steven Nissen, a cardiologist at the Cleveland Clinic in Ohio, who believes that naproxen is not harmful. "If you take a snapshot of a clinical trial at any given point in time, you can see almost anything," he says.
Still, others say that no matter what the trial's reliability, the safety of naproxen and other traditional anti-inflammatories are not assured. "Naproxen should not be given a free pass," says Abramson.
Related sites
)
)
)
)
)
)
)
)
)
