By Kevin Davies
July 20, 2005 | This month, a group of influential editors from some of the world’s boutique medical journals begin enforcing a new policy that would bar drug companies from publishing clinical trial data unless those trials are registered in a suitable public database.
Calls for mandatory clinical trial registration intensified last year following New York State attorney general Eliot Spitzer’s lawsuit against GlaxoSmithKline (GSK) for suppressing negative clinical data on the antidepressant Paxil in children. GSK quickly settled for a reported $2.5 million and promise of full disclosure by the end of 2005.
The International Committee of Medical Journal Editors (ICMJE) — which includes the Journal of the American Medical Association, the Lancet, and the New England Journal of Medicine — issued a joint declaration promising to reject reports pertaining to new unregistered trials effective July 1 and to insist upon registration by this September for trials already under way.
This hard-line approach extends well beyond the dozen members of ICMJE. Other prominent journals, including the British Medical Journal and the open-access journal PLoS Medicine, are backing the ICMJE’s revised trial publication criteria, although there are differences of opinion on whether approved databases must be not-for-profit.
Despite assertions from drug companies that they would readily post trial data in registries and/or their own Web sites, shockingly, it seems some aren’t. The New England Journal editor-in-chief Jeff Drazen publicly rebuked GSK, Merck, and Pfizer for “making a mockery” of an open registry. A survey of entries in the NIH-based Clinicaltrials.gov database (see “Is That A Target On Her Back?” March 2005 Bio•IT World March 2005, page 1) revealed that the aforementioned pharmas frequently declined to identify the drug under investigation or reveal the formal title of the study. For example, GSK omitted key information in 90 percent of cases. As a recent New York Times editorial stated: “Any American gullible enough to believe that the drug industry can be trusted to report fully on what clinical trials it is sponsoring or what results were found must be sorely disappointed by recent developments.”
he importance of a transparent, comprehensive trial registry is unquestionable, even as the furor over the recall of Vioxx and other blockbuster drugs recedes. Annual prescription drug spending in the United States exceeds $160 billion, of which about 40 percent — $65 billion — is wasted on drugs that produce no benefit, or worse, harm the health of the patient. Last month, the Times reported that Johnson & Johnson persuaded the FDA to revise label warnings on its blockbuster heartburn drug Propulsid, before its eventual withdrawal in 2000 after links to dozens of deaths.
Complete, open access to trial data, while of limited everyday use to patients, would provide an invaluable resource to the medical community, particularly when trying to evaluate the full clinical profile of a new drug.
At a 2004 meeting in Geneva organized by the World Health Organization (WHO), participants established 20 key data points required to register each trial, including a unique study identifier. Given legitimate concerns about competitive information, one proposal is for sensitive data to be held in escrow for a short time. WHO also plans to build a Web portal to provide access to approved registries. These currently include Clinicaltrials.gov and London-based Current Controlled Trials, which provides free access to users but, at least for the time being, requires a registration fee (see “A Proliferation of Trial Registries,” May 2005 Bio•IT World, page 56).
Just as Congress is threatening to pass draconian laws to regulate steroid abuse (with a threatened two-year ban for first-time offenders), there is also talk of compulsory trial registry legislation, with offenders facing $10,000-per-day fines. The drug industry should recognize the merits of a transparent comprehensive registry of clinical trials rather than tempt Congressional interference. While the result could be an “unstructured deluge of information that few of us have time to process,” as Nature Medicine opined, the benefits — for physicians, scientists, editors, and indirectly, patients — are worth it.
Author's note: In May's First Base, I stated that the first virus genome was sequenced in 1978. My thanks to William Kuziel for pointing out that a Belgian group sequenced the RNA genome of the MS2 virus two years prior.