Thomas V. Inglesby, MD, is the Chief Executive Officer and Director; Anita Cicero, JD, is Chief Operating Officer and Deputy Director; and D. A. Henderson, MD, MPH, is a Distinguished Scholar, all at the Center for Biosecurity of UPMC, Baltimore, Maryland. Drs. Inglesby and Henderson are Coeditors-in-Chief of Biosecurity and Bioterrorism: Biodefense Strategy, Practice, and Science.
Over the past 8 years, H5N1 avian influenza has sickened 571 people, killing 59% of them. To give some perspective, the fatality rate of the virus that caused the 1918 Great Pandemic was 2%, and that pandemic killed on the order of 50 million people. Like all influenza strains, H5N1 is constantly evolving in nature. But thankfully, this deadly virus does not now spread readily through the air from person to person. If it evolves to become as transmissible as normal flu and results in a pandemic, it could cause billions of illnesses and deaths around the world—the proportion of the global population affected by past pandemics.
Scientists recently have announced that they genetically modified H5N1 in the laboratory and that this mutated strain spread through the air between ferrets that were physically separated from each other. This is ominous news. Since ferret influenza virus infection closely mirrors human infection and is similarly transmissible, these scientists appear to have created a bird flu strain with characteristics that indicate it would be readily transmissible by air between humans. In fact, the lead scientist on one of the experiments explicitly stated this.
The question is this: Should we purposefully engineer avian flu strains to become highly transmissible in humans? In our view, no. We believe the benefits of this work do not outweigh the risks. Here’s why.
There are no guarantees that such a deadly strain of avian flu would not escape accidentally from the laboratory. This particular experiment was performed by internationally respected scientists in biosafety conditions considered top of the line. They seem to have taken the expected and necessary precautions. The risk of a person accidentally becoming infected and starting an outbreak with this new strain is low. But it is not zero.
The safety record of most labs working at high biocontainment levels is outstanding, and the historic number of accidents is very low. In almost all situations, even if a laboratory worker comes in contact with a dangerous pathogen and becomes sick, the risk of extensive wide community spread is negligible. This is because very few dangerous pathogens are as highly transmissible as influenza is. An accidental escape of an influenza strain from a lab in 1977 proves the possibility: That accident led to widespread flu epidemics. Given the potential global consequences of an accident with the newly modified strain of avian flu, we are playing with fire.
We are not opposed to research in high-containment labs using dangerous pathogens, including H5N1. Over the past decade, the Center for Biosecurity has publicly argued for the importance of such research to develop diagnostics, medicines, and vaccines for the most threatening infectious diseases. But research and development for those purposes does not require engineering lethal viruses to make them more transmissible between humans. There is no virus disease today other than influenza that has exhibited the potential attributes of global transmission such as has been documented with flu over the past 2 centuries and more. Thus, it occupies a unique position in the pantheon of microbes.
Some who defend the value of this recent experiment have argued that it is important to determine whether the present H5N1 strain could acquire genetic characteristics of a pandemic strain and, if so, what the particular characteristics might be. The thinking is that currently circulating strains of the H5N1 virus could then be screened for these characteristics, which, if found, might serve as an early warning. It is a speculative hope but not worth the potential risk.
Others have argued that the benefit of this experiment and its findings will be to motivate more work on H5N1 flu vaccines. Scientists have already developed bird flu vaccines, and research and development work continues. Should there be more money and effort devoted to improving our vaccine and medical defenses against H5N1 and other flu strains? Absolutely. But should we undertake experiments that engineer a transmissible strain of deadly flu in hopes that the frightening results invigorate R&D for vaccine development?
Still others have argued that this experiment may help directly in the vaccine development process. That argument is that now that we have this new transmissible strain, we should use it to test whether our current vaccines are effective against it. But it is unclear that vaccine protection against this engineered strain would correlate with vaccine protection against an H5N1 pandemic strain that evolves naturally in the world. Again, we would argue that the risk of accidental release outweighs the potential and uncertain benefits of engineering a pandemic strain.
An eminent scientific journal is considering whether to publish this work. A critical tenet of the advancement of science is the publication of new research in a form that allows other scientists to reproduce the work. This principle should be followed in almost all conceivable circumstances. But in this circumstance, it shouldn’t.
Publishing the methods for transforming the H5N1 virus into a highly transmissible strain would show other scientists around the world how to do it in their own labs. One concern is the possibility that the strain would be recreated for malevolent purposes. Even disregarding this risk (which we shouldn’t), scientific publication would encourage others that this is a research initiative worthy of additional exploration.
There are already checks in the research review process that include scientists’ consideration of these issues at project conception, deliberation at the time of funding decisions, institutional biosafety committee approvals, and the scientific publication review process. In principle, these elements should provide the necessary checks and balances. In this case, it will be important to understand the reasoning and decision making that led to the execution of this work. In the future, when an experimental plan calls for engineering a lethal virus into a highly transmissible one, then the rationale for funding and approval should be made explicit publicly.
Whether this experiment is published or not, it is a reminder of the power of biology and its potential. We need new approaches for the rapid development of large quantities of medicines or vaccines to protect us against new emerging viruses. But engineering highly transmissible strains of avian flu is not the way to get us there.