C ontinual news on the outbreak status of the H5N1 subtype of influenza A virus in Southeast Asia points to one of the greatest challenges facing 21st-century society: the prediction and management of disasters. Hundreds of thousands die from influenza annually, with widespread and often devastating pandemics occurring episodically. The last flu pandemic occurred in 1968. Are we better able to mitigate the effect of a new pandemic than we were 37 years ago? Advances in science, vaccine strategies, and antiviral drugs provide this potential, but whether these can be applied in the short term in an effective global policy is not guaranteed. The continual threat of influenza A viruses such as avian H5N1 lies in their basic biology. The virus is easily transmitted and can be highly virulent. It is present at high frequencies in reservoirs of wild birds that can infect domestic animals, including horses, pigs, and poultry. Add to that the genetic plasticity of the virus, with high rates of mutation and a ready capacity for reassortment that allows it to combine with other strains to produce new and sometimes highly pathogenic variants. Most worrisome, the reassortment of human and bird strains could result in a pandemic virus that is transmissible among humans. A highly pathogenic avian H5N1 virus first appeared in Hong Kong in 1997. A mass cull of chickens alleviated the problem locally, but H5N1 viruses continued to circulate in birds in Asia, most recently among migratory species that could theoretically carry the virus for long distances. In 2003, H5N1 re-emerged in humans, causing almost 60 deaths in Asia to date, although there is no convincing evidence that the virus has evolved human-to-human transmission. ![Figure][1] CREDIT: CHRISTOPHER FURLONG/GETTY IMAGES How might we prevent and manage a future influenza pandemic? The most obvious requirement is a rapid and expansive influenza surveillance and response network. A permanent global task force has been proposed to perform this role,[*][2] and we strongly endorse this idea. However, such a task force will only be successful if national governments release data promptly and adhere to control measures should a pandemic arise. Such surveillance activity also needs to include humans, domestic animals, and wild birds. Second, we must develop further, effective intervention strategies to reduce transmission and disease. The development of vaccines against H5N1 strains, and ultimately against all subtypes, is a clear priority. Recent preliminary tests of a potential vaccine are encouraging. But using traditional approaches, fewer than 500 million people could currently be vaccinated with a two-dose monovalent pandemic influenza vaccine. New vaccine methodologies are in reach, but international agreements on production, intellectual property, distribution, and administration need to be aggressively pursued. Antiviral drug stockpiles are equally limited. Thus, because rapid global distribution networks of vaccines and antiviral agents have yet to be established, it is essential that logistical simulations be conducted to determine their possible limitations. Third, we need epidemiological models, to explore the spread and impact of potential influenza pandemics in the face of realistic control measures and how to manage pandemics when they arise. Two recent reports suggest that antiviral-based containment policies could be an effective strategy,[†][3] although this is contingent on a rapid, coordinated response to the emergence of a pandemic. We must also develop strategies to reduce the probability of pandemics. This will require a multitude of basic scientific information, including the probability and mechanism of reassortment; a measure of the exposure rates of influenza viruses at the human/animal interface; and, most critically, an understanding of how avian viruses evolve to develop sustained transmission networks in humans. It is therefore essential to conduct a global surveillance of genetic diversity in avian influenza viruses, sequencing complete genomes from these and mammalian strains to explore the polygenic nature of host adaptation. We also need to determine the extent of clonal genetic variation within individual hosts, because consensus sequences invariably hide strains with varying phenotypic properties. These data would also provide perspective on the evolutionary dynamics of viral pathogens at different spatial scales. Although a unified political effort is essential to avert or mitigate a major influenza pandemic, it must proceed in parallel with advances in basic science. The potential for avian H5N1 to cause a global human pandemic is presently uncertain because it cannot be predicted with current data. However, if an H5N1 pandemic does not emerge in the near term, the political will to continue the global preparations necessary for a future pandemic may falter. We cannot afford such a misstep. [1]: pending:yes [2]: #fn-1 [3]: #fn-2