The Food and Drug Administration should rethink its rejection of the Declaration of Helsinki.
Later this year, the US Food and Drug Administration (FDA) will adopt new standards for human clinical trials conducted without its advance sign-off in foreign countries. The rules will govern whether data from such trials can be used in applications to market the drug in question in the United States. Although these new standards specify how to run such trials to meet US requirements, they are worryingly silent on key issues relating to human rights, in contrast with the rules currently in effect. As a result, they could open the way to some ethically fraught decisions.
Take the case of the drug Surfaxin, a synthetic, inhaled version of a lung protein the absence of which is a leading cause of death in premature infants. Back in 2001, the drug’s manufacturer, Discovery Labs of Warrington, Pennsylvania, was looking for a suitable location in Latin America to run a trial on the therapy. But rather than compare its product to one of the several effective drugs already available, Discovery Labs was proposing to administer a placebo to the 325 infants in the control group.
The trial was redesigned only after the FDA — and unfavorable media attention — reminded Discovery Labs that a placebo-controlled trial of this type would be deemed unethical in the United States, and other developed countries, because effective treatments were available. As a result, the control group received alternative active treatments.
The FDA estimates that annually it receives data from around 575 foreign drug trials conducted without its knowledge, more and more of which come from trials run in the developing world. Currently, these trials must comply with the Declaration of Helsinki (or with local country laws, whichever offer the most protection) if sponsors want to use the data to win US marketing approval. The declaration, adopted in 1964, and revised several times since, is today endorsed by medical associations from 85 countries. It is widely considered to be the bedrock of protection for research subjects. Its 1989 revision, which the FDA uses as its present standard, states that any patient in any trial “should be assured of the best proven diagnostic and therapeutic method”.
“The FDA risks sending a message that ethical considerations are
expendable when research subjects live half a world away.”
Yet the FDA announced last month that it will shelve the declaration. Starting in October, the FDA intends to adopt a new standard it calls Good Clinical Practice (GCP), which is modelled on a 1996 document developed by drug regulators and pharmaceutical industry representatives from the United States, the European Union and Japan. Although GCP deals with subject protection, it is in essence a manual on how to conduct rigorous clinical trials, not a human-rights document. For instance, whereas Helsinki explicitly discourages the use of placebos for serious conditions where proven therapies exist, GCP is silent on this issue. So under the GCP guidelines, the FDA could accept data from Surfaxin placebo trials of the future.
The FDA argues that it should not be bound by Helsinki because the declaration is devised by a group it does not control, and is subject to periodic revisions that could confuse trial sponsors or contradict US law. But it is tempting to conclude that the FDA is dropping Helsinki not because it is changeable, but because the agency disagrees with the way it has been changing — in particular with its constraints on the use of placebos. (The US agency is more favourably disposed to placebo use than, say, its European counterparts.)
It makes sense for the FDA to adopt the GCP standard, giving foreign-based researchers guidelines that should help them generate the best data. But if the FDA jettisons Helsinki, the critical underpinning for such efforts, it risks sending a message that ethical considerations are expendable when research subjects live half a world away.
Sources: Nature, International Weekly Journal of Science Editorial
Nature 453, 427-428 (22 May 2008) Published online 21 May 2008
