Clinical Trials on Pharmaceuticals

Over the past few years, questions have been raised about how much access physicians and the general public should have to the results from clinical trials for pharmaceuticals. Currently, drugmakers are for the most part required to share newly discovered information about risks and side effects only with the Food & Drug Administration.

The American Medical Association's (AMA) Council on Scientific Affairs published a report on the publication of clinical trials in June. The report concluded that clinical trial results that show a positive benefit of a treatment are far more likely to be published--and go to press more quickly--than those indicating no benefit.

AMA passed a resolution that month calling on the Department of Health & Human Services to create a comprehensive centralized registry of clinical trial outcomes so that scientists, investigators, and clinicians could easily find information. The registry "would ensure that all trials with negative or positive results are made publicly available." Each trial, the resolution says, would be provided a unique identification number and placed on an electronic database of results from registered trials. This database would include trials of drugs, medical devices, and other treatments.

On Sept. 9, the International Committee of Medical Journal Editors announced that its 11 member journals will require as a condition for publication that clinical trials be listed in a clinical trials registry. When the results of the trials are known, they must be included in the registry. This policy applies to any clinical trial starting enrollment after July 1, 2005. "Honest reporting begins with revealing the existence of all clinical studies, even those that reflect unfavorably on a research sponsor's product," the statement says.

This summer, Merck and Eli Lilly separately announced that they would create public registries for their drug-testing data. And on Aug. 27, GlaxoSmithKline agreed to put all of its clinical trial results on its website within 10 months of completion. This decision was made as part of a lawsuit settlement with New York Attorney General Eliot L. Spitzer. He found that Glaxo had conducted at least five clinical trials on the use of the drug Paxil in depressed teenagers but published only one of them, which shows mixed results on efficacy. Data from the other studies, which suggest a possible increase in suicidal behavior with the use of Paxil, were not publicly disseminated.

Also in August, Sens. Tim Johnson (D-S.D.) and Christopher J. Dodd (D-Conn.) made their own call for a national registry of clinical drug trials. They wrote to Elias A. Zerhouni, director of the National Institutes of Health, and Acting FDA Commissioner Lester M. Crawford asking what additional authorities these agencies needed to create a national registry that would include results of all clinical drug trials, both positive and negative.

"Consumers and medical professionals ought to have complete information available when they consider a given drug," Johnson says. "Doctors and patients shouldn't have to play Russian roulette when it comes to the medicines they prescribe or receive," Dodd says.

"Doctors and patients shouldn't have to play Russian roulette when it comes to the medicines they prescribe or receive."

ALTHOUGH CONCERN about clinical trial results is an issue for all medicines, recent observations about antidepressants brought the issue to the fore.

Antidepressant medicines called selective serotonin reuptake inhibitors (SSRIs) and other antidepressants have been prescribed increasingly for depressed children and adolescents over the past decade. During this period, doctors have observed that some of these drugs seem to raise the risk of suicides and suicide attempts in young people, as well as the risk of aggressive behavior, shortly after the drug regimen is begun. A number of these physicians believe that if all the drugs' clinical trial data on youths had been publicly available, some tragedies would have been avoided. Several lawyers have filed class-action lawsuits against the makers of Paxil and other antidepressants on behalf of parents of young people who committed suicide or tried to commit suicide while taking these drugs.

Last February, FDA hosted an advisory committee meeting on the use of antidepressant drugs in children and adolescents and on what can be learned from the clinical trials conducted thus far. At the meeting, Russell Katz, director of FDA's Division of Neuropharmacological Drug Products, laid out the history of how the agency first became concerned about pediatric use of antidepressants.

When FDA reviewed the controlled trial data for Paxil, it became aware that Glaxo had categorized some events that could have represented suicidal behavior or suicidal thinking with descriptions that seemed "inappropriate," Katz said. FDA asked Glaxo to clarify its presentation of the data, and the company's response raised concern that Paxil may increase suicidal behavior, he said. As a result, in June 2003, the agency issued a notice recommending that the drug not be used to treat pediatric patients with depression.

In a further step, FDA asked the makers of other antidepressant drugs to search their databases for instances in which the drugs may have contributed to suicidal behavior in youth. Based on the responses received, the agency then advised doctors to use antidepressants with "caution" in pediatric patients, Katz said.

FDA also decided that the data on antidepressant drugs submitted by the manufacturers "may not have been collected or reported to us in a form that would permit us to adequately evaluate the potential relationship between these drugs and suicidal behavior," Katz explained. The agency then requested that all the relevant data gathered during the trials be submitted in a standardized format so that the data could be looked at collectively.

The agency asked a panel of experts at Columbia University to examine all the instances of potential suicidal behavior observed in the clinical trials from a number of antidepressant drugs and to use their expertise to classify these behaviors into discrete categories depending on seriousness.

AFTERWARD, two FDA scientists reanalyzed the behaviors as classified by the Columbia University panel. One analysis, by Tarek A. Hammad, a safety reviewer in FDA's Division of Neuropharmacological Drug Products, found that children and adolescents on SSRIs and other antidepressants are 1.8 times more likely to experience suicidal thoughts and behavior than those who are not taking the drugs. Andrew Mosholder, an epidemiologist in FDA's Division of Drug Risk Evaluation, using slightly different methods, concluded that youths who take antidepressants have 1.9 times the risk of experiencing suicidal behavior. In both analyses, Prozac was the only drug that seemed to decrease suicidal behavior. It is the only one FDA had approved for pediatric use. Doctors were prescribing the other antidepressants to youths in what is called off-label use.

Daniel Carpenter, a professor of government at Harvard University who is leading an extensive research effort on FDA issues, says it is not surprising that there is disagreement over the effects of antidepressants. It is especially difficult to interpret the effects of psychotropic drugs, he says. "It seems that many of these drugs are effective for many patients," he explains, "but the concern is just how effective.

"One of the problems is that people who go on antidepressants also receive some psychotherapy," he continues. So when they improve, it is hard to tell how much of the improvement should be attributed to the drug and how much to the therapy, he explains. Another problem is the placebo effect, which can be very strong in people who are being treated for depression, he says. "The belief in the promise of a treatment can itself have a curing power."

One reason FDA is devoting so much time to sorting out the effects of antidepressants in youths is that depression in adolescents is prevalent in the U.S. and in much of the world. Between 4 and 8% of U.S. adolescents have major depressive disorder, said Cynthia R. Pfeffer, professor of psychiatry at Weill Medical College of Cornell University, at the FDA meeting held in February.

After homicide and accidents, suicide is the leading cause of death among teenagers, killing about 1,600 in the 16 to 19 age group each year, says David Shaffer, professor of child psychiatry at Columbia University's College of Physicians & Surgeons. In 2002, according to the Centers for Disease Control & Prevention, an estimated 124,409 visits to U.S. hospital emergency rooms were made after attempted suicides or other self-harm incidents among young people 10 to 24 years old.

THE GOOD NEWS is that, even though the youth suicide rate started to increase in the late 1950s and tripled by 1990, it declined somewhat after 1994. Antidepressant drugs were first used to treat depressed teens in the late 1980s. One theory is that antidepressants are responsible for the decline in the suicide rate, but no one has hard evidence to support this theory, Shaffer explained.

Although at first blush it might seem purely advantageous to make all clinical trial data publicly available, experts find both benefits and risks in taking this step. They caution that a registry needs to be created carefully or it will be of little benefit to most users.

Peter Lurie, deputy director of Public Citizen's Health Advisory Group, says the results of Phase II, III, and IV clinical trials should be put on the registry. In other words, all but Phase I trials, which involve small patient groups and are used to measure toxicity, should be included. Any new trials for drugs that already have been approved and marketed, as well as medications that never made it through the approval process, should be on the registry, he says.

Currently, Lurie says, there is no way to find any information about the clinical trials conducted on most drugs that FDA did not eventually approve. The only exceptions are the unapproved drugs that FDA discussed in advisory committee meetings, he says. For these, clinical trial information is posted on the Internet, usually 24 hours prior to the meeting.

If the results for unapproved drugs were included in a registry, other companies would be less likely to test those same substances, thus avoiding some blind alleys, Lurie says. Consequently, a registry could potentially lower the overall cost of drug development, he says. "What you really want is a registry where the data are entered the moment it is [finalized]."

However, it is important to enter data in a form that can be interpreted, Lurie says. For example, the first level of data consists of questionnaires, sometimes thousands in one study, where yes or no answers are circled. There is a secondary level in which the answers are entered into spreadsheets. There is another level in which all the data, including all the different variables, are summarized. Then, there is a final level of analysis that would end up in a medical journal. In setting up a registry, FDA would need to establish "a relatively standardized form of collecting data with the right degree of rawness," he explains. "What you want is something that provides detail on the order of what appears in a medical journal, or perhaps somewhat more," he says.

Having a registry of clinical trials could be beneficial in several ways, Carpenter says. Giving physicians access to more information would allow them to make better decisions, he says. By enabling physicians to more easily choose the least expensive drug that will help the patient, a registry "would be consistent with recent movements toward cost-effectiveness in medicine and evidence-based medicine," he observes.

"Giving the layperson access to clinical trial results at the very earliest stages could in some cases do more harm than good."

BUT CARPENTER has one worry. "Giving the layperson access to clinical trial results at the very earliest stages could in some cases do more harm than good," he says. Often patients ask, "Is this medical procedure or treatment associated with death?" he explains. But deaths are common among patients who are taking drugs for heart disease or Parkinson's disease, for example. "If two out of 10,000 patients die from a heart attack during a clinical drug trial, the doctor knows that this death rate compares well with the adverse event readings for other drugs," Carpenter says. But the layperson might be afraid to take the new drug because of the associated fatalities. For this reason, it is important to put information in the proper context in a clinical trials registry, he notes.

There is a general movement toward making the results of clinical trials "somewhat more transparent," says Alan Goldhammer, associate vice president for regulatory affairs at the Pharmaceutical Research & Manufacturers Association (PhRMA). He stresses that data such as adverse drug reactions are always reported to FDA. In regard to SSRIs, he explains, "the data have always been in FDA's hands. It's just that the agency was struggling in trying to interpret them."

PhRMA is committed "to the timely communication of meaningful results of controlled clinical trials of marketed products or of investigational products that are approved for marketing, regardless of the outcome of the trials," he says. "If a product is not marketed, we don't see the utility of publishing those results because doctors can't prescribe the drug.

"Where PhRMA has a great interest in trying to improve transparency is the trials that for whatever reason did not get into the drug label," Goldhammer explains. "These could be trials that were only mildly positive and trials that could be confusing." However, he continues, it would be inappropriate to publicize the results of clinical trials if the data were invalid or didn't answer the questions for which the trial was designed.

Making clinical trial results publicly available would, in the long run, be very beneficial, says Robert Temple, acting director of FDA's Office of Drug Evaluation. "As a scientist, I feel that if someone does a trial in humans, the public ought to be aware of the results," he says. These data could be put to many useful purposes. They would allow people to pool data, to do meta-analyses, and to see if they can find things that weren't apparent in the initial look at the study, he says. "Those are useful processes that people learn from. So as a medical specialist, I would like to see trial data available if it were done properly and well," he explains.

However, FDA does not have the authority to require that all clinical trial results for drugs, even for approved drugs, be put in a public database, Temple says. Legislation would be required to make this mandatory. But even without legislation, there is a general voluntary movement toward making trial results public, he notes. The moves toward greater transparency are already happening, he explains, and are likely to continue no matter what Congress does.