Department of Homeland Security Bioterrorism Risk Assessment: A Call for Change (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

5 Risk Assessment for Unknown and Engineered Biothreat Agents How do we avoid becoming beguiled by the risks we have already experienced, and distracted from those that our enemy might be planning in the future? —Department of Homeland Security Secretary Michael Chertoff at Homeland Security Policy Institute, March 16, 2005 BIOLOGICAL THREAT RISK ASSESSMENTS NEED TO uncertain (IOM and NRC, 2006). Agents can be modified for INCLUDE UNKNOWN AND ENGINEERED AGENTS new properties in a variety of ways. Discoveries in this area in the past few years have included the following: Most of this report deals with assessing bioterrorism risk and prioritizing risks associated with known biologi- • Poxviruses with an IL-4 gene insert that can cause cal agents. However necessary the focus on risk associated severe disease in immunized or genetically resistant with known biological agents is, the committee strongly animals (Jackson et al., 2001). IL-4 (interleukin-4) believes that it is not sufficient. The Department of Home- is a mammalian protein that serves as one of several land Security’s (DHS’s) Biological Threat Risk Assessment important regulators of immune response. Ectromelia (BTRA) of 2006 only considers threats already known and (mousepox) virus is similar in lethality and contagious- at least partially characterized. However, the biological ness to smallpox in humans and is closely related to the threat spectrum is dynamic (Petro et al., 2003; IOM and smallpox virus. As with human smallpox, mice can be NRC, 2006) and therefore requires a proactive approach. protected by the same vaccine that is used to protect Some agents on the Category A list (such as several of humans against smallpox. However, when immunized the hemorrhagic fevers) of the Centers for Disease Control mice are infected with the IL-4 modified mousepox and Prevention (CDC) were discovered only within the virus, the effects are severe and similar to those in past few decades, and there are undoubtedly many more unimmunized mice. In addition, strains of mice that pathogens still undiscovered in nature (IOM, 1992, 2003; normally would be genetically resistant are not resistant Morse, 1991, 1995). Some of them may be similar to agents to the modified virus, but become sick in the same way already known. Others, such as Nipah virus infection and that more susceptible mice do. Fortunately, for reasons severe acute respiratory syndrome (SARS) (discovered in not clearly understood, these engineered strains do not 2003), may have completely unexpected characteristics. In transmit well to others. However, one can anticipate addition, previously unknown pathogens will continue to be that a future technically adept adversary could solve discovered or to evolve from nature. Some may be adopted this problem. by adversaries as “bioweapons of convenience,” just as the • Anthrax modified with a gene insert from a nonpatho- current biothreat agents were all zoonotic diseases (animal genic relative that can defeat a live anthrax vaccine diseases that can be transmitted to humans) adopted by older (Pomerantsev et al., 1997). Inserting a particular gene bioweapons programs because their biological or physical from a relatively harmless anthrax relative, Bacillus characteristics made them suitable. cereus, made anthrax able to infect and kill animals So far, with respect to biothreat agents, nature has been (hamsters) that had been immunized with the standard the greatest source of novelty, but the rapid advances in live vaccine used in Russia for human protection. This molecular biology and biotechnology and the increasing live vaccine, known as STI, is generally considered understanding of pathogenesis (the mechanisms by which highly effective, and (although comparative data are these organisms cause disease) cannot be ignored. These ad- lacking) is widely thought to be equivalent in efficacy vances suggest that the future will be even more complex and to the protective antigen (PA) protein-based vaccine used in the United States and United Kingdom. Earlier,   For the CDC list, see www.bt.cdc.gov/agent/agentlist-category.asp#a. at the 1995 International Anthrax Meeting in Salisbury, Accessed February 25, 2008. United Kingdom, the same Russian group had reported 42 

RISK ASSESSMENT FOR UNKNOWN AND ENGINEERED BIOTHREAT AGENTS 43 developing multi-drug-resistant anthrax. Although a such technological advances. In the meantime, conventional vaccine strain was used for this experiment, it could threats are likely to predominate. Nevertheless, if the history just as easily have been done with virulent anthrax. of PCR and other scientific advances is any indication, the • Reconstruction of viable 1918 pandemic influenza virus use of biotechnology to engineer novel threats will come in (Tumpey et al., 2005). The virus responsible for the time. It has been suggested that engineering “advanced bio- most notorious influenza pandemic in recorded history weapons” is a natural extension of advancing biotechnology. (with an estimated 50 million human deaths world- In the words of the authors of a recent publication on this wide) was recently reconstructed from several different subject (Petro et al., 2003, p. 161): sources using molecular techniques. This tour-de-force Advances in biological research likely will permit develop- of molecular biology (by Jeffery Taubenberger and col- ment of a new class of advanced biological warfare (ABW) leagues) made it possible to study the 1918 pandemic agents engineered to elicit novel effects. . . . Such new agents virus for the first time. The virus could be grown and and delivery systems would provide a variety of new use op- tested in several animal species, in which it caused tions, expanding the BW paradigm. Although ABW agents severe disease. The purpose of the work was construc- will not replace threats posed by traditional biological agents tive, to better understand how the 1918 pandemic virus such as Bacillus anthracis (anthrax) and Variola (smallpox), caused such serious disease. Even more recently, it they will necessitate novel approaches to counterprolifera- was shown that two specific amino acid changes in the tion, detection, medical countermeasures, and attribution. hemagglutinin (HA) surface protein of the H5N1 avian In consideration of these possibilities, the White House influenza virus would enable it to bind to human, rather recently released Homeland Security Presidential Directive than avian, tissues, a necessary first step in being able 18 (HSPD-18): Medical Countermeasures Against Weapons to readily infect humans (Yamada et al., 2006). of Mass Destruction (The White House, 2007) as a follow- up to the original biodefense strategy embodied in HSPD-10 The motives for all the work cited here are ostensibly (The White House, 2004). HSPD-10 is the document that, benign: to better understand these dangerous pathogens. But among other tasks, instituted the regular threat assessment it is easy to imagine how the same techniques could be ap- that constitutes the main thrust of this committee’s work. plied to other uses. At present, conducting this work requires Setting out the outlines of the U.S. biodefense strategy, specialized laboratory expertise at the postgraduate level or HSPD-10 states that “[t]he essential pillars of our national above, and the influenza genetic system is currently beyond biodefense program are: Threat Awareness, Prevention the technical capabilities of all but a few experts. However, and Protection, Surveillance and Detection, and Response advances in biotechnology will make all of these techniques and Recovery.” HSPD-18 takes this strategy a step farther, more accessible in the future (IOM and NRC, 2006). The mandating that “[o]ur Nation will use a two‑tiered approach powerful molecular technique for selectively copying de- for development and acquisition of medical countermea- oxyribonucleic acid (DNA) known as the polymerase chain sures, which will balance the immediate need to provide a reaction (PCR) was so esoteric in the early to mid-1990s that capability to mitigate the most catastrophic of the current performing it required painstaking technique by experienced CBRN (chemical, biological, radiological, and nuclear) scientists. PCR has now become so widely used and routine threats with long-term requirements to develop more flex- that it is commonplace in high school science projects and ible, broader spectrum countermeasures to address future is even taught to schoolchildren visiting museum exhibi- threats.” The biodefense tasks are divided into “Tier I: tions. As another example, the complete chemical synthesis Focused Development of Agent-Specific Medical Coun- of the poliovirus genome (a small ribonucleic acid [RNA] termeasures” for current and anticipated biological threats virus) required several years of work by experts, including and “Tier II: Development of a Flexible Capability for New overcoming a number of technical difficulties (Cello et al., Medical Countermeasures” (The White House, 2007). The 2002). Since then, the George Church Laboratory at Harvard latter specifically recognizes the diversity of possible future University has devised microchips that could be used to syn- biological threats, both natural and engineered, and the need thesize even larger genomes with far less effort (Tian et al., for broad-spectrum solutions. 2004), and other large-scale rapid DNA synthesis methods The BTRA of 2006 does not lend itself readily to the rapid are at the advanced development stage. There has also been assessment of new threats. Cybersecurity presents similar academic interest in “synthetic biology,” a kind of engineer- contrasts of comprehensiveness versus flexibility. Buckshaw ing using biological component parts to make entities with et al. (2005) developed a quantitative risk model based on desired functions (Bio FAB Group et al., 2006). It is clear the adversary’s attack preferences instead of the adversary’s that future possibilities will be limited more by imagination probabilities of attack. This has certain advantages (e.g., than by technical obstacles. Buckshaw at al. [2005, p. 24] note, “Adversary attack prefer- Very few individuals today are capable of using these ences are easier to measure and help develop the mitigation techniques, and it is likely to be some time before other than strategy. We need to consider all attacks since a capable state-sponsored terrorists will be able to take advantage of and adaptive threat will constantly change their actions in 

44 DEPARTMENT OF HOMELAND SECURITY BIOTERRORISM RISK ASSESSMENT response to our assurance activities.”). However, Buckshaw three concepts of information characteristics, information et al. (2005, p. 36) also note the same drawback of needing states, and security countermeasures. DHS could seek, large amounts of reasonably accurate data: “Data are the with its customers, multiple classification schemes that are benefit and bane. . . . If time is spent to get the data required most useful to each of its customers’ end-user communi- of a quantitative risk model, then one can produce insightful ties. Rather than requiring specific numbers (such as R0, or and clear recommendations for the decision maker. Because infection rate), identifying combinations of key variables as the data requirements are so large, we recommend that [this] qualitative categories would also help model vulnerabilities be used only on critical information systems.” and prioritize concerns that have not yet been foreseen by existing analyses. Although analyzing newly emerging threats may seem a INCLUDING UNKNOWN AND ENGINEERED daunting task, it actually appears to be quite feasible. While AGENTS IS CHALLENGING BUT POSSIBLE the number of all possible combinations of characteristics is There are several possible ways to deal with the unpre- enormous, it would not be necessary to deal with such vast dictable and dynamic future for both natural and engineered numbers of combinations in practice, because the analysis agents: would be limited to a number of key characteristics at rela- tively broad qualitative levels. The high, medium, and low • Concentrate on known agents, and develop new risk threat categories appear meaningful to some users and might assessments as each new threat is identified (e.g., from be sufficient. intelligence). While this avoids speculation about fu- Several efforts have been made to define the weapon or ture possibilities and possibly unnecessary work, it has terrorist potential of microbial agents. For example, Casa- several potential weaknesses. Risk assessment models devall and Pirofski (2004) recently attempted to identify the such as the BTRA of 2006 require large amounts of characteristics that might contribute to the weapon potential specific information about the agent and its properties. of an agent. Such efforts have repeatedly identified several Even with such well-known agents as Bacillus anthra- of the same characteristics—for example: cis (anthrax) or Yersinia pestis (plague), the critical data are approximate at best, with uncertainties that have • Ease of acquisition, proven elusive to quantify (indeed, some data may • Transmissibility, vary by strain of organism and conditions of assay). • Mode of spread (person to person or by direct exposure Obtaining these data for newly recognized or newly only, or both), engineered agents is likely to be even more difficult, • Case fatality rate, and to exact a significant time lag. • Ease of dissemination, • Attempt to identify every potential future threat. The • Frequency of serious sequelae (e.g., blindness or neu- committee believes that both the complexity of nature rological disease) in survivors, and and unforeseen advances in biotechnology will make • Availability and efficacy of countermeasures (vac- this task infeasible and may lead to a false sense of se- cine or other prophylactic measures) and therapeutic curity by leaving the United States unprotected against measures. newly engineered pathogens. There are too many theo- retical possibilities and, barring reliable intelligence All of these criteria are included in the BTRA of 2006. information, prioritization is likely to be exceedingly However, a true quantitative estimate is virtually impossible difficult. for newly recognized or poorly characterized agents. The • Consider “more-generic” categorization of agents and committee has prepared a rapid assessment tool that can be risks into groups by various properties, identifying the applied to any newly recognized agent for which there are most critical variables. This would provide a general only very limited specific data, and it suggests that DHS framework that could be used to classify newly identi- consider development of such a tool. Such a rapid assessment fied threats as they appear on the basis of even limited tool could use attributes similar to those listed above, with information. In addition, this approach may suggest the agent qualitatively categorized as being a low, medium, mitigation strategies that already apply to existing or high threat with respect to each attribute; these categories pathogens. The committee favors this open-ended ap- could be assigned numerical scores (e.g., 1, 2, and 3) and proach for new and emerging pathogens. these scores used as a signature to compare with known pathogens. Although such qualitative assessments cannot What are some possible standard situations or classifica- replace detailed risk assessment, the use of such a rapid as- tions to use in an effort to anticipate future threats? Consider sessment tool can aid DHS in focusing on areas where further the analogy presented by grouping threats in information expansion of possibilities and more evaluation are needed. assurance analyses. McCumber (1991) introduces a qualita-   T.Cox, Cox Associates. 2002. “What’s Wrong with Risk Matrices.” tive model for information security that incorporates the Unpublished. 

RISK ASSESSMENT FOR UNKNOWN AND ENGINEERED BIOTHREAT AGENTS 45 TABLE 5.1  An Illustrative Approach to Rapid Assessment, With Some Examples Attribute Anthrax Brucella Ebola Salmonella Smallpox 1.  ase of acquisition or synthesis E 3 3 1 3 1 2.  nvironmental stability E 3 1 1 2 2 3.  ransmissibility person to person T 1 1 1 2 3 4.  ase fatality rate (untreated) C 3 1 3 1 3 5.  ase of dissemination (estimated) E 2 2 1 3 2 6.  requency of serious sequelae (e.g., blindness F 2 2 2 1 3 or neurological disease) in survivors 7.  ack or unavailability of useful L 1 3 3 2 2 countermeasures or treatmenta 8.  eed for immediacy in diagnosis and treatment N 3 1 3 1 3 to ensure patient survival    Total 18 14 15 15 19 NOTE: 1 = Low threat, 2 = Medium threat, 3 = High threat. aHigh threat = Countermeasures/therapeutics nonexistent or not readily available. The committee’s rapid assessment tool is an adaptation to form the basis of a more proactive strategy to face future of the Multi-Attribute Risk Analysis (MARA) step 1 used threats. in the development of the DHS risk assessment process as described in the DHS report Bioterrorism Risk Assessment Recommendation: The BTRA should be broad enough to (DHS, 2006, Ch. 6, pp. 1-28). In MARA, there are 28 attri- encompass a variety of bioterrorism threats while allow- butes, each scored 0 through 4 by a panel of experts. Aggre- ing for changing situations and new information. DHS gating these categories to broadly reflect several key charac- should develop a strategy for the rapid assessment of teristics, such as those listed above, could form the basis for a newly recognized and poorly characterized threats. rapid assessment tool. Although results would, of course, be approximate, this rapid assessment would help DHS and its REFERENCES partner agencies determine whether a newly identified agent is a high priority for additional consideration. Bio FAB Group, D. Baker, G. Church, J. Collins, D. Endy, J. Jacobson, J. Keasling, P. Modrich, C. Smolke, and R. Weiss. 2006. “Engineering As a hypothetical illustration of this sort of rapid assess- Life: Building a Fab for Biology.” Scientific American 294(6):44-51. ment, a template and some worked examples are shown in Buckshaw, D.L., G.S. Parnell, W.L. Unkenholz, D.L. Parks, J.M. Wallner, Table 5.1. and O.S. Saydjari. 2005. “Mission Oriented Risk and Design Analy- While engineered agents are by definition novel, an engi- sis of Critical Information Systems.” Military Operations Research neered agent will likely be designed for a specific function. 10(2):19-38. Casadevall, A., and L.A. Pirofski. 2004. “The Weapon Potential of a Mi- The committee therefore anticipates that the evaluation of crobe.” Trends in Microbiology 12(6):259-263. such agents would be similar to the evaluation of novel Cello, J., A.V. Paul, and E. Wimmer. 2002. “Chemical Synthesis of Polio- natural agents. virus cDNA: Generation of Infectious Virus in the Absence of Natural Additionally, rigorous sensitivity analyses applied to Template.” Science 297(5583):1016-1018. the BTRA (as recommended by the committee elsewhere DHS (Department of Homeland Security). 2006. Bioterrorism Risk Assess- ment. Biological Threat Characterization Center of the National Biode- in this report) can help identify the key characteristics for fense Analysis and Countermeasures Center. Fort Detrick, Md. a rapid assessment and should be done regularly and on a IOM (Institute of Medicine). 1992. “Emerging Infections.” In Microbial variety of parameters. Since many of the parameters input to Threats to Health in the United States, J. Lederberg, R.E. Shope, and the BTRA result from the elicitation of expert opinion, the S.C. Oaks, Jr. (eds.). Washington, D.C.: National Academy Press. threats that emerge may only reinforce general expert con- IOM. 2003. Microbial Threats to Health: Emergence, Detection, and Re- sponse. Mark S. Smolinski, Margaret A. Hamburg, and Joshua Leder- sensus. Alternative consequence analyses of different routes berg (eds.). Washington, D.C.: The National Academies Press. of administration (such as large-scale food contamination) IOM and NRC (National Research Council). 2006. Globalization, Bio­ should be rigorously tested to ensure that these results are security, and the Future of the Life Sciences. Committee on Advances robust. It is also possible that even improvised suboptimal in Technology and the Prevention of Their Application to Next Genera- routes used by an adversary may cause significant morbid- tion Biowarfare Threats. Washington, D.C.: The National Academies Press. ity and mortality, or mental health consequences. This issue Jackson, R.J., A.J. Ramsay, C.D. Christensen, S. Beaton, D.F. Hall, and is further examined in Chapter 6, “Improving Bioterrorism A.I. Ramshaw. 2001. “Expression of Mouse Interleukin-4 by a Recom- Consequence Assessment.” binant Ectromelia Virus Suppresses Cytolytic Lymphocyte Responses It is said that generals are adept at fighting the last war. and Overcomes Genetic Resistance to Mousepox.” Journal of Virology The committee recommends expanding present approaches 75(3):1205-1210. 

46 DEPARTMENT OF HOMELAND SECURITY BIOTERRORISM RISK ASSESSMENT McCumber, J. 1991. “Information Systems Security: A Comprehensive Tumpey, T.M., C.F. Basler, P.V. Aguilar, H. Zeng, A. Solorzano, D.E. Model.” In Proceedings: 14th National Computer Security Conference. Swayne, N.J. Cox, J.M. Katz, J.K. Taubenberger, P. Palese, and A. Baltimore, Md.: National Institute of Standards and Technology. Garcia-Sastre. 2005. “Characterization of the Reconstructed 1918 Span- Morse, S.S. 1991. “Emerging Viruses: Defining the Rules for Viral Traffic.” ish Influenza Pandemic Virus.” Science 310(5745):77-80. Perspectives in Biology and Medicine 34(3):387-409. The White House. 2004. Homeland Security Presidential Directive 10 Morse, S.S. 1995. “Factors in the Emergence of Infectious Diseases.” [HSPD-10]: Biodefense for the 21st Century. Available at www.fas. Emerging Infectious Diseases 1(1):7-15. org/irp/offdocs/nspd/hspd-10.html. Accessed January 16, 2008. Petro, J.B., T.R. Plasse, and J.A. McNulty. 2003. “Biotechnology: Impact The White House. 2007. Homeland Security Presidential Directive 18 on Biological Warfare and Biodefense.” Biosecurity and Bioterrorism [HSPD-18]: Medical Countermeasures Against Weapons of Mass 1(3):161-168. Destruction. Available at www.fas.org/irp/offdocs/nspd/hspd-18.html. Pomerantsev, A.P., N.A. Staritsin, Yu.V. Mockov, and L.I. Marinin. 1997. Accessed January 16, 2008. “Expression of Cereolysine AB Genes in Bacillus Anthracis Vaccine Yamada, S., Y. Suzuki, T. Suzuki, M.Q. Le, C.A. Nidom, Y. Sakai-Tagawa, Strain Ensures Protection Against Experimental Hemolytic Anthrax M. Yukito, Y. Muramoto, M. Ito, M. Kiso, T. Horimoto, K. Shinya, Infection.” Vaccine 15(17-18):1846-1850. T. Sawada, M. Kiso, T. Usui, T. Murata, Y. Lin, A. Hay, L.F. Haire, Tian, J., H. Gong, N. Sheng, X. Zhou, E. Gulari, X. Gao, and G. Church. D.J. Stevens, R.J. Russell, S.J. Gamblin, J.J. Skehel, and Y. Kawaoda. 2004. “Accurate Multiplex Gene Synthesis from Programmable DNA 2006. “Haemagglutinin Mutations Responsible for the Binding of H5N1 Microchips.” Nature 432(7020):1050-1054. Influenza A Viruses to Human-type Receptors.” Nature 444(7117):378- 382.