July 21, 2004
Results of Drug Trials Can Mystify Doctors Through Omission
pediatrician who has read medical journals lately might have seen articles advocating the use of the antifungal drug Diflucan for the treatment of ringworm, a childhood skin infection.
But reading the extensive medical information enclosed in the drug's package, the same doctor would not learn that federal drug regulators had rejected Diflucan for ringworm use. Reviewing test data provided by the maker, Pfizer, the F.D.A. concluded that Diflucan was not very effective for ringworm - and that a child could risk liver damage if given high doses of the drug.
It is a quirk of the F.D.A.'s labeling rules. For a drug's government-approved uses, the label gives doctors important data from clinical tests involving those treatments while spelling out the drug's risks and providing directions for administering it to patients.
But labels are often allowed to remain silent about a test when the F.D.A. turns down a company's application for approval of a new use or a new patient group - as in the case of Diflucan and ringworm.
Pfizer and the F.D.A. both say that the Diflucan label conforms to the agency's rules and accurately reflects the drug's approved uses for adults and children, as well as its risks.
The labeling issue is another facet of an expanding debate over the incomplete disclosure and publication of the results of clinical drug trials. And it is one of growing importance to a number of physicians.
The doctors worry that ambiguous or incomplete drug labels may lead them to prescribe drugs for treatments that F.D.A.-reviewed tests have shown to be ineffective or potentially risky. Printed material accompanying many drugs says the drug's safety and effectiveness "has not been established'' in children - when some have been tested for pediatric use, with equivocal results.
Until recently, several popular antidepressants carried the "has not been established'' language despite tests in children and adolescents that raised questions about their efficacy. Regulators and manufacturers have since issued safety warnings about those drugs amid concerns that, in a sharply limited number of cases, they may cause depressed youngsters to consider suicide.
"If it has been tested, we want that to be reflected" on the label, said Dr. Richard Gorman, a pediatrician who is chairman of the committee on drugs of the American Academy of Pediatrics.
In recent weeks, the academy and other medical groups have met with lawmakers to discuss possible legislation requiring companies to list tests of drugs and medical devices in a public database. And yesterday, Representative Henry A. Waxman, Democrat of California, sent a letter to Health and Human Services Secretary Tommy G. Thompson, urging him to release data from all pediatric drug trials that have been conducted under federal laws that encourage such tests.
Dr. Shirley Murphy, director of the F.D.A.'s division of pediatric drug development, said that several members of an agency advisory committee that met in May had expressed concern that some drug labels omitted any reference to clinical drug trials that were negative or equivocal. The group included Dr. Gorman.
"They noticed that some labels have lots of information and other labels didn't have information," Dr. Murphy said.
The issue of test disclosure on labels is not limited to pediatric trials. With few exceptions, federal law bars the F.D.A. from acknowledging that clinical tests were conducted, if the agency receives the test data as part of a drug maker's effort to have an existing product approved for a new use or category of patients and the F.D.A. turns down the request.
Drug companies may publicly disclose such failures through a press release or a filing with the Securities and Exchange Commission, though such information would not necessarily find its way onto a label. And if a company chooses not to publicize a test failure, a doctor typically would not learn from a drug's label that such trials ever took place, a high-ranking F.D.A. official, Dr. Robert Temple, said.
"If we don't agree" with the company, said Dr. Temple, the F.D.A.'s associate director for medical policy, "the doctor would not know about that."
Dr. Temple said there were exceptions, like a case when the trial of a drug shows a new or increased safety risk beyond those previously indicated on the label. Other exceptions include pediatric drug trials conducted under a federal law passed in 2002, the Best Pharmaceuticals for Children Act.
As with an earlier law in 1997, the 2002 measure was passed to promote the testing of drugs in children because pharmaceutical companies had customarily not performed such trials. Instead, the practice was to seek approval for uses in adults, then leave it to pediatricians to prescribe the drugs on an off-label basis - as doctors are allowed to do, although drug makers cannot actively market drugs for such nonapproved uses.
To encourage tests in children and adolescents, who may react differently than adults do to some drugs, the laws gave drug makers a powerful financial incentive - an additional six months of patent protection from generic competition when the medications were tested for pediatric use, even if the trials showed the drugs to be of dubious value.
Dr. Gorman said he and others who helped draft the 1997 legislation quickly realized that they had failed to include a way to force a company to disclose all results from medications tested under the law. And so the 2002 law required that the F.D.A. release summary reports of such drug tests.
Currently, reports on about 30 drugs received by the agency since mid-2002 are on the Web at www.fda.gov/cder/pediatric. There, for example, a browser would find the summary report on Pfizer's pediatric tests of Diflucan. But Dr. Gorman says he now believes, as a result of the antidepressant controversy, that the 2002 bill was not explicit enough.
"We didn't go the next step - about what would go on the label," he said.
All involved agree that the laws have increased the amount of labeling information about the use of many drugs in children, like the age groups where they are effective or how the drugs should be administered. And in some cases, new warning information reflects safety issues that arose in tests, like those on packaging of some cortisone-based treatments for skin infections.
Of the 98 drugs that qualified for testing under the 1997 and 2002 laws, the labels on 76 have been changed, F.D.A. officials said, while to date the labels on 22 drugs have not.
The information gap between test findings and label information can occur even when F.D.A. officials are aware that doctors are increasingly using a medication off label for a particular problem.
In the case of Diflucan, according to a test summary, it was the agency that asked Pfizer to run the ringworm trials. Regulators were concerned, according to that report, that pediatricians were turning to Diflucan because of dissatisfaction with an older medication and because published studies, some sponsored by Pfizer, reported that Diflucan was superior.
After reviewing two Pfizer-run pediatric studies, the F.D.A. concluded earlier this year that Diflucan did not work any better than the older, generic drug, griseofluvin, which also poses toxicity risks. In the agency's report, regulators also raised concerns that children could face liver problems and other risks if doctors, seeking to treat ringworm more effectively, started using higher doses of Diflucan, which is known generically as fluconazole.
"As some adverse events may be dose-dependent," the F.D.A. document said, "it is expected that higher doses of fluconazole may approach dosages where the safety of pediatric patients may be at greater risk."
Dr. Murphy, at the F.D.A., said that the ringworm tests had not been noted on Diflucan's label because it did not win approval as a ringworm treatment.
A Pfizer spokeswoman, Mariann Caprino, said the company was discussing the drug's labeling language with F.D.A. officials. She also said the tests did not raise any safety concerns beyond those already mentioned in the label. The F.D.A. also turned down an effort last year by another drug maker, Allergan, to get its antibacterial eye ointment, Ocuflox, approved for use in infants because trial data suggested the drug was not particularly effective in newborns. But the label continues to say that the drug's "safety and effectiveness in infants below the age of one has not been established."
There is little question that some drugs fail in pediatric trials because the tests are poorly run or not properly designed for children. F.D.A. reviewers raised the possibility of a design flaw in their report about Ocuflox.
And while the F.D.A. review of pediatric test results for the popular migraine treatments Imitrex and Zomig found that the drugs worked no better than placebos in children, pediatric headache experts involved in those trials cited flaws in the test design. Dr. Gorman said it was important for all doctors, not just those who treat children, to know that trials with negative or equivocal outcomes have been conducted when considering the benefits and risks of a medication before prescribing it.
Dr. Temple said that, speaking personally, he also believed that drug companies and medical device producers should voluntarily disclose test results - regardless of what F.D.A. rules may require. Otherwise, they face the risk of seeming to bury information.
"If I was consulting companies, I would say 'Let it out,' " Dr. Temple said. "Now, it looks like something secret has been uncovered."