Now that world attention finally has been focused on the potentially human-contagious H5N1 Asian bird-flu virus, the international research community should take steps to deal with three other potential pandemic pathogens. Two are among history's nightmares: smallpox, which killed or maimed millions of people over centuries as it ravaged the world, and the resurrected 1918 pandemic flu virus, which may have killed 50 million people worldwide over just two years. The third is severe acute respiratory syndrome, or SARS, a newcomer that proved how lethal it could be in one natural outbreak during 2002 and 2003, when it killed 9.6 percent of those it infected, a fatality rate almost four times higher than the 1918 flu's.
These potential pandemic pathogens, or PPPs, pose a danger that goes well beyond the potential of other hazardous microbes that have made the news. The PPPs are all extremely deadly, highly contagious (or potentially highly contagious) in humans, and not currently present in human populations, meaning it would be a disaster to reintroduce them into the population. Of the three, smallpox, researched at only two facilities in the world by international agreement, poses the smallest threat of laboratory escape. At Russia's Vector and the U.S. Centers for Disease Control, lab workers have been vaccinated to prevent infections to themselves that might spread to others outside the lab. In stark contrast to the strict controls on smallpox research, however, SARS, the 1918 flu virus, and potentially human-contagious H5N1 bird flu are studied in laboratories throughout the world, using less than the highest biocontainment, known as Biosafety Level 4, or BSL-4, and there is no approved and stockpiled vaccine for any of them.
A quick search of PubMed, the National Library of Medicine database of medical research, identifies 30 labs that are working with live SARS virus and at least 10 using live 1918 flu virus. Counting the two labs in the Netherlands and United States that created Asian bird flu that is contagious in ferrets (which served as a model for human contagiousness), there are at least 42 facilities engaged in researching live PPPs. The PubMed search was anything but comprehensive, and the actual number is likely higher.
Simple mathematical analysis gives real reason for concern about the handling of these dangerous viruses. Consider the probability for escape from a single lab in a single year to be 0.003 (i.e., 0.3 percent), an estimate that is conservative in light of a variety of government risk assessments for biolabs and actual experience at laboratories studying dangerous pathogens. Calculating from this probability, it would take 536 years for there to be an 80 percent chance of at least one escape from a single lab. But with 42 labs carrying out live PPP research, this basic 0.3 percent probability translates to an 80 percent likelihood of escape from at least one of the 42 labs every 12.8 years, a time interval smaller than those that have separated influenza pandemics in the 20th century. This level of risk is clearly unacceptable. (A detailed analysis, additional arguments, documentation, and mathematical justification for these conclusions can be found in the research report written by one of us, "Sharpening Our Intuition on Man-made Pandemics.")
Awful as a pandemic brought on by the escape of a variant H5N1 virus might be, it is SARS that now presents the greatest risk. The worry is less about recurrence of a natural SARS outbreak than of yet another escape from a laboratory researching it to help protect against a natural outbreak. SARS already has escaped from laboratories three times since 2003, and one escape resulted in several secondary infections and one death.
What is the likelihood that the virus's escape could lead to a pandemic? Too high, given the lessons taught by the natural SARS outbreak a decade ago. During that outbreak, one woman infected in Hong Kong flew to Toronto, a city with outstanding public health capabilities. The woman initiated infections in 438 people in Canada, and 44 of them died. What if the next infected person flies to a crowded city in a poor nation, where surveillance and quarantine capabilities are minimal? Or to a war zone where there may be no public health infrastructure worthy of the name? In such settings, an all but unstoppable pandemic could be seeded before SARS were even identified. While the numbers of victims and fatalities from the 2002-2003 SARS outbreak are large -- 8,098 probable cases and 774 deaths -- they are nowhere near the very large numbers defining a pandemic.
The most convincing evidence that a lab escape could cause a pandemic is a comparison of the features of SARS and the 1918 pandemic flu. According to Appendix C pages 31 and 58 in the Final Supplementary Risk Assessment for the Boston University National Emerging Infectious Diseases Laboratories, in the case of each virus, one victim can infect an average of two to three others, the key measure of contagiousness. The 1918 flu killed about 2.5 percent of those infected. SARS, on the other hand, had a fatality rate of 9.6 percent. In other words, SARS appears to have as much pandemic potential as the 1918 flu, but with a fatality potential that is four times greater.
How devastating might such a pandemic be? Although precise numbers are hard to come by, in a conservative estimate, 15 percent of the world's population could be infected by a pandemic flu virus. Using that percentage and a world population of 7 billion, about 1 billion people would be infected and 100 million would die as a result. If extreme public health measures -- including quarantine -- were successfully undertaken, the toll might be reduced radically -- perhaps by as much as 90 percent -- but that still would mean 100 million victims and 12 million deaths.
As noted, about 30 labs now are working with live SARS virus worldwide. The probability of escape from at least one laboratory is high; the probability of an escape that leads to a major outbreak or pandemic is, on the other hand, likely low. Would one in 10 escapes lead to a major outbreak or pandemic? One in a hundred? One in a thousand? No one knows. But for any of these probabilities, the likelihood-weighted number of victims and deaths would be intolerably high. In a world of many and varied risks, following the "precautionary principle" strictly and universally can stand in the way of accomplishing worthwhile goals. But the precautionary principle should apply to research on potential pandemic pathogens because of the absolutely intolerable possible consequences of their escape.
Research on live SARS should be curtailed -- perhaps even discontinued -- until biological containment measures beyond BSL-3 and even BSL-4 can be put in place. Simply moving all SARS research to BSL-4 facilities will not substantially reduce the risk; there have already been three escapes from BSL-4 containment since 1990: a Marburg virus laboratory-acquired infection at the Vector facility in the Soviet Union in 1990, a foot and mouth disease virus escape from the Pirbright facility in England, and a SARS virus laboratory-acquired infection from a BSL-4-rated biosafety cabinet in a Taiwan laboratory.
The research community must move toward requiring an additional level of biosafety -- call it BSL-4-plus -- that adds special protections for laboratory work with potential pandemic pathogens and includes measures in two major categories.
To improve laboratory biosafety, the research community should agree on measures that:
- Train full-time technical staff who are dedicated to working with highly dangerous pathogens. These staffers would carry out experiments directed by scientists who would never need to be present in the BSL-4+ laboratory. With modern audio-video technology, research scientists can remotely monitor lab work as if they were present.
- Require lab staffers to follow up extended work shifts with periods of quarantine before they leave the biocontainment area. Such procedures would assure that no potential pandemic pathogen escapes from a BSL-4+ lab through a laboratory-acquired infection; anyone accidentally infected would show symptoms while still in quarantine.
- Mandate that potential pandemic pathogen experiments be carried out primarily in BSL-4+ labs (although BSL-3+ with the special protections like post-shift quarantine might suffice for some work).
- Locate the two or three such high-security labs the world needs in remote locations, where an aerosol escape or other containment failure would pose the least risk of infecting an outside community.
To place responsibility with the international community where it belongs and to provide maximum transparency, international policy makers should:
- Institute a global system for deciding whether a given potential pandemic pathogen experiment should be done at all and for monitoring the experiments that are approved.
- Require that international inspectors have access to facilities at any time, on short notice.
- Make experimental details from these labs available to inspectors and to authorized international bodies, to maximize the transparency of lab activities and to help dispel concerns about clandestine development of biological weapons. Appropriate ground rules can keep confidential information gained during the inspections from falling into the wrong hands.
As it stands, researchers often study potential pandemic pathogens -- including the extremely dangerous SARS virus -- in BSL-3 laboratories, ignoring National Institutes of Health guidelines, which clearly state, "Biosafety Level 4 is required for work with dangerous and exotic agents that pose a high individual risk of aerosol-transmitted laboratory infections and life-threatening disease that is frequently fatal, for which there are no vaccines or treatments…"
The World Health Organization should expand its focus beyond H5N1 bird flu and convene a meeting to take up concerns about research on live PPPs. The participants should be balanced among virologists, public health scientists, biosafety and biosecurity experts, and biological-weapons and arms-control experts. The goal should be nothing less than the elimination of the possibility that scientific research might cause a pandemic.