If asked to name the world’s most dangerous virus, many people would cite Ebola viruses, which cause gory, often fatal, hemorrhagic illness; or perhaps H5N1 avian influenza, which kills 60 percent of its victims. These highly lethal viruses present great danger to those they infect, justifying elaborate biosafety precautions taken to avoid exposure. However, their ability to spread from person to person is limited, so the danger they pose is thankfully local rather than global – limited to those directly exposed and perhaps a few close contacts. Indeed, the fact that hundreds of people have suffered infection from Ebola and from “wild-type” or natural H5N1 avian flu, yet these have not spread to become a global pandemic, is the best evidence that such viruses at present cannot sustain large-scale human-to-human transmission.
In the summer of 2011, laboratories in the Netherlands, the US, and Japan published reports of flu viruses related to naturally occurring (“wild-type”) H5N1 avian flu that had been deliberately modified to become transmissible through the air between ferrets, and thus potentially in humans, for whom ferrets are the best animal model. Some of these viruses differ by only 5 mutations from wild-type H5N1, but nonetheless seem to have overcome the one obstacle that limits the global impact of natural H5N1 and many other virulent, but low-transmission viruses.
The human lethality of wild-type H5N1 is perhaps 1000 times greater than that of the 2009 pandemic flu strain, 10 to 30 times that of the 1918 flu, six times that of SARS, and even greater than that of smallpox. If mammalian-transmissible H5N1 combined even a fraction of the lethality of their wild-type precursors with the ability to transmit efficiently between humans, then a laboratory accident leading to sustained transmission of these viruses could pose a threat greater than that of any other known infectious agent. Once it takes root in the human population, influenza is essentially impossible to stop with currently available approaches, partly because vaccine-manufacturing capacities are a fraction of global needs.
The exceptional risk posed by mammalian-transmissible H5N1 viruses justifies exceptional precautions to avoid their accidental release. Until January 2013, influenza researchers observed a self-imposed “pause” on research on these viruses. This moratorium was always meant to be temporary, so a critical question is what regulations on research with mammalian-transmissible H5N1 viruses should now take its place.
Most importantly, the number of laboratories and scientists handling these viruses should be limited. Smallpox, perhaps the closest rival in terms of the global threat it poses, is permitted in only two laboratories in the world. Every time these viruses are shipped or used in an experiment, there is a risk – very small in the best labs, but not zero – of a laboratory accident. SARS virus has three times been the source of laboratory infections (fortunately not leading to sustained spread); the last human case of smallpox was in a laboratory worker; and both the UK’s 2007 foot-and-mouth disease epidemic and the global spread of influenza A/H1N1 (which was part of seasonal influenza from 1977 to 2009) are attributed to laboratory accidents.
More precautions are needed
National authorities should permit and fund experiments only if the benefits they promise outweigh the small, but quantifiable, risk of a laboratory accident that could lead to extensive spread of the virus. In particular, experiments should not be permitted unless investigators can compellingly demonstrate that there is scientific knowledge with direct public health benefit that can be gained from these experiments, which could not be gained from experiments on less transmissible, or less lethal, influenza viruses. Even if the answer to this question is yes, there should be a presumption against increasing the number of labs possessing the viruses, to reduce the risk. The highest quality of laboratory worker training, skill and experience should be required for work on these viruses, and physical containment should be optimized to mitigate the risks of accidental exposure. Laboratory workers should not handle the virus unless they have received H5N1 vaccine (small vaccine stocks do exist) and shown an immune response.
Some of these requirements are in place in some laboratories in some countries, but efforts must be redoubled, and best practices enforced in any lab that deals with this especially dangerous virus – and perhaps with the small number of other viruses, such as SARS coronavirus, smallpox (on which research is already heavily restricted), which combine high virulence with high transmissibility, and have for that reason been called “potential pandemic pathogens.”
Rich countries where this research has started should take the lead in responsible regulation, but the technology to handle these viruses exists in all parts of the world, and a global consensus on applying a risk-benefit calculus before permitting experiments is needed, led by international organizations (notably WHO). While any reduction in the risk of accidents would be beneficial, coordination across agencies within national governments and across borders is needed to avoid “shopping” by investigators for sites with the loosest regulation and to establish an international ethic of caution surrounding these viruses. Because of how influenza viruses (and scientific norms) spread, decisions in Berlin have consequences in Boston, Bangkok, and Bamako.
Some work on mammalian-transmissible H5N1 viruses has tremendous value and should be done in the most careful laboratories; for example, the work published already helped to quell speculation that such viruses could never transmit among mammals, and has returned H5N1 to the attention of those who plan for pandemics. But now that we know this, much of the most important work to understand pathogenesis, treatment, and vaccination against H5N1 can be done on the far less transmissible, wild-type H5N1 viruses, or on transmissible forms of other, less lethal flu viruses.
These precautions would be stricter than those for any infectious agent at present, except smallpox. Proposing such restrictions does not come easily to a scientist, as replication of results and competition help drive the progress of science. But just as we forbid some scientifically valuable experiments to avoid cruelty to animals or risk to human research subjects, we should surely be willing to consider restrictions on experiments where, if something goes wrong, there is a massive threat to human lives.