|
|
| |||||||||||||||||||||||
When is evidence from a series of similar cases strong enough to justify going directly to the public, possibly engendering grave concern among those using a drug or device? Are small, uncontrolled studies such as this one valid in our era of large and rigorously controlled case–control and cohort studies? The history of our involvement with the study by the Mayo Clinic and Meritcare illustrates our approach to this issue.
At the end of February 1997, we received a first manuscript by Connolly and her colleagues in which they described five cases of valvular heart disease in patients taking fen–phen. Information from pathological examination of the affected valves was available for only two patients. However, the authors indicated that they were aware of 11 other possible cases that were not as well documented. They believed that their few cases might be only the tip of a large iceberg. Given that the use of these drugs and valvular disease are both prevalent, the association in so few cases could have been a chance occurrence. After review, we thought there was still substantial uncertainty about the causal connection between fen–phen and valvular disease, but we offered to publish a description of their preliminary findings in our Correspondence section. While the manuscript was under review, however, the authors were accruing more cases.
In May, we received a revised manuscript from the two institutions. By then they had accumulated 24 cases of valvular disease in patients who had been taking fen–phen. Echocardiographic studies were available for all the patients, and severe, highly unusual valvular abnormalities were confirmed in the five patients who required valve replacement or repair. We believed that the possibility of a cause-and-effect relation between fen–phen and the cardiac lesions was strengthened by the new data, and outside reviewers agreed. After the authors responded to suggestions for revising the manuscript, we accepted the paper for publication on June 25. We then waived the Ingelfinger Rule, which prohibits a prepublication press conference, and gave the authors permission to report the findings publicly and release the manuscript. Members of the press were informed that a news conference would be held on July 8.
Why make such a fuss? Why didn't we wait the seven weeks until the final paper, with all the relevant detail, was published? In an editorial accompanying the published paper, Dr. Gregory Curfman, a deputy editor, explained that we waived the Ingelfinger Rule and allowed the investigators to release their data early because we believed that the findings might have immediate implications for the health of people using the drug combination.2 This decision was in accord with our long-standing policies.3 Subsequent observations, some of them detailed in the Correspondence section of this issue, add to the strength of the hypothesis that the link between fen–phen and valvular heart disease is causal. Nonetheless, more definitive data from case–control studies are required to confirm (or disprove) the apparent association. In the meantime, fenfluramine has been removed from the market at the request of the Food and Drug Administration, after further data suggested an incidence of valvular disease as high as 30 percent among users of the drug. Dexfenfluramine, which has been associated with pulmonary hypertension and valvular heart disease, has also been removed, but phentermine is still on the market.
Some of the physicians who criticized us for permitting the Mayo Clinic press conference had been caught off guard by the news reports. When their patients called them to ask whether they should continue taking fen–phen, the physicians were unable to answer, because they could not assess the evidence for themselves. The existence of an electronic medium (the World Wide Web) that allows us to transmit medical news instantly to physicians all over the world suggested a solution to this problem. In this case, Connolly and her coauthors offered to put the preliminary, unedited manuscript on their web site. We agreed, and it was installed on the Mayo Clinic web site at the time of the press conference, but not all interested physicians found it.
In the future, when we waive the Ingelfinger Rule for a study that we and the authors agree has critical public health implications, we will coordinate the press release with installation of the partially edited manuscript in full text on the Journal's web site (www.nejm.org). We will mark it clearly as incomplete, and when the final manuscript is available we will replace the incomplete version with it. Not all physicians have access to the Journal's web site, but the numbers of those who do are increasing rapidly, and we expect that most will soon be able to obtain information by this method.
Most of the studies that we and other journals publish have only an indirect and cumulative effect on medical practice. Physicians rarely change their prescribing practices on the basis of a single study, nor should they. But in the few instances in which a study suggests an immediate and important change in practice, we will continue to try to convey information to doctors and patients as quickly as possible, having made reasonable efforts to ensure scientific validity. The line between jumping the gun and waiting too long is a difficult one to tread, and we are aware that we may be criticized for premature release of alarming findings, just as we have been criticized in the past for withholding important information. In the case of fen–phen, reducing patients' exposure to these drugs by seven weeks may have reduced morbidity and even saved some lives. We undoubtedly will permit the early release of unpublished studies again.
Jerome P. Kassirer, M.D.
Marcia Angell, M.D.
References
| |||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
