The unpublished studies, in the hands of the journals Science and Nature, reveal how Dutch and American research teams managed to mutate the H5N1 avian flu virus to the point where it became highly transmissible in ferrets, which are considered the best predictor of how flu viruses might behave in humans.
The U.S. government asked the journals to publish only brief reports of the results of the work on the advice of its biosecurity advisers, the National Science Advisory Board for Biosecurity. The biosecurity board looks at so-called dual use research: science that is done for legitimate reasons, but which could be used for bioterrorism.
The request is non-binding; the U.S. government does not have the power to block the publications.
In exchange for asking the journals to limit the scope of what they reveal, the U.S. government said it would work with international partners to try to devise a system whereby parties that need to know the full details of the research — other flu scientists and public health officials, for instance — could be given access to the material.
But some skeptics questioned whether, in the age of Wikileaks, it would be possible to keep the information under wraps. A leading U.S. government scientist said the best they can do is try.
"In the era of the Internet and being able to spread information, there's no guarantee whatsoever that this will remain confidential. But you can just try as best you can," said Dr. Anthony Fauci, scientific director of the National Institute of Allergy and Infectious Diseases, which funded the studies.
"Is this really going to be able to be maintained? I don't know. It's probably not, in a totally strict confidential way."
The journals and researchers revealed they are working on removing the guts of the studies to comply with the request, but most stopped short of making firm commitments not to share the work more broadly.
Dr. Bruce Alberts, editor-in-chief of Science, said the journal's editors are deliberating what to do. Alberts said Science's response would depend on what steps the U.S. government takes to put the promised sharing system in place.
And one of the research groups based at Erasmus Medical Centre in Rotterdam, the Netherlands, warned it will not wait months for U.S. government officials to work out that system and figure out how to decide who has a legitimate need to see the material.
"If the U.S. government or WHO," — the World Health Organization —"are going to take two months to figure out a strategy to disseminate the results, then we will probably have already disseminated the results to those that need it, especially our very close collaborators, the people who have distributed the (original) virus to us," said Dutch virologist Ron Fouchier, who is the senior author of the study expected to be published in Science.
While the U.S. plans to include international partners in the discussions, it's not yet clear whether the WHO would have or would want a role in this endeavour.
Quick action may not be within reach. Just getting a decision from the National Science Advisory Board on Biosecurity on whether the papers should be published took more than two months. The group held multiple conference calls, some with as many as 60 or 70 people arguing on the line.
Acting chair Paul Keim, an anthrax expert, said the discussions were unlike any he's taken part in before. But he said in the end, the panel reached broad consensus that the studies should not be published in full.
Michael Osterholm, who is both an expert in biosecurity issues and in flu, took part in those discussions. He said the critical thing is not getting the studies out quickly, but getting them out safely.
"Time is important here, but we need to exercise caution," said Osterholm, who heads the Center for Infectious Diseases Research and Policy at the University of Minnesota.
"And I think the urgency of the situation is not in my mind one that says 'We have to tell the world everything about this today.' I think the urgency of the situation would be if we tell the wrong people about this and they do something with it that could cause really truly catastrophic damage in the future."
Fouchier has already submitted an abbreviated version of his paper to the journal, along with an editorial —co-written with Erasmus' head of virology, Ab Osterhaus — criticizing the panel's decision.
Fouchier developed an H5N1 virus that spread easily among ferrets with only "a handful" of mutations. He said in an interview Tuesday that all the mutations he used have been found in H5N1 viruses found in the wild, in birds. In some cases, some of the mutations have been seen in combination with each other, he said.
Osterhaus said he and Fouchier know of flu scientists elsewhere who are working with viruses that have some of these mutations, and they feel morally obligated to let them know about the potential danger.
Like Alberts, Nature Editor-in-Chief Dr. Philip Campbell did not indicate whether the journal would honour the request to withhold the key portions of the paper submitted to it by virologist Yoshihiro Kawaoka of the University of Wisconsin-Madison.
Kawaoka was not immediately available for comment, but his university issued a brief statement on his behalf saying the team would comply with the biosecurity panel's recommendation.
While the debate over whether to recommend that the studies not be put in the public domain has taken place largely in secret, news that the discussion was underway has spilled out into the broader scientific world.
Some called the decision to recommend restriction of the information a mistake.
"This is a bad day for virology, and for science in general," microbiologist Vincent Racaniello of Columbia University Medical Center wrote in his popular blog, Virology Blog.
"The decision by the NSABB sets a precedent for censoring future experimental results whose wide dissemination would benefit, not harm, humanity."
But others hailed the move, with some going so far as to insist the work should never have been done in the first place.
Biosecurity expert Dr. D.A. Henderson of the Center for Biosecurity at the University of Pittsburgh Medical Center said he thought the panel's recommendation was the right approach to talk, though he acknowledge it would be tough to keep the information secure.
"I think there's no perfect answer," said Henderson, who co-wrote a recent statement from his centre saying the risks of this research outweighed any possible benefit that might flow from it.
Fauci disagreed, however, defending the work. He said there is a legitimate need to know whether H5N1 has the capacity to become a virus that can spread easily from person to person. "And we need to know what signal characteristics might alert us to that," he said.
The H5N1 virus currently only infects humans rarely. Since the virus exploded through poultry flocks in Asian in 2003, 573 people have been known to have been infected and 336 of those people have died.