A Framework for Assessing the Health Hazard Posed by Bioaerosols (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.

2 Exploring the Complexity of Health Risk In developing a standard framework for evaluating the detection of biological aerosols, it is necessary to consider the biological and physical characteristics of the samples under study. Biological warfare agents (BWAs) are currently grouped into three categories: bacteria, viruses, and toxins. 2.1 BIOLOGICAL WARFARE AGENTS 2.1.1 Bacteria Bacteria are single-celled microorganisms that are capable of replication independent of other living cells. Some can be grown in a simple liquid medium containing amino acids, salts, and other basic substances or on the surface of agar medium in a petri dish. The ease with which large quantities of pathogenic bacteria can be prepared in the laboratory contributes to their threat as biological weapons. Bacteria are often characterized by morphology (e.g., rod, sphere, or spiral), motility, and nutritional requirements (e.g., oxygen consuming). Bacteria can be pathogenic, parasitic, symbiotic, or free living, though combinations of these broad categories also exist (e.g., pathogenic, parasitic bacteria). Examples of bacterial pathogens and their associated diseases are: • Bacillus anthracis—anthrax • Yersinia pestis—plague • Francisella tularensis—Tularemia There are several approaches to quantifying bacteria, including flow cytometry, fluorescent labeling, and culture (i.e., colony counting). The ability to form colonies is one commonly used indication of viability for bacteria and their spores.1 While there are other indications of viability, colony counting is the accepted “gold standard”; the resultant unit of measure is the colony forming unit (CFU). A limitation of the CFU is that several bacteria in close proximity on the plate will form a single colony (i.e., one CFU does not necessarily equal 1 For the purposes of this report, the ability of bacteria to form cultures on artificial media will be considered a measure of viability. The committee acknowledges that this is more a direct measure of culturability rather than viability, but finds the uncertainty introduced by this issue falls into the broader category of uncertainty in threat assessment. See Box 3.1 for more information regarding uncertainty in bioaerosol detection. 15 

16 one viable bacterium). An additional limitation is that some bacteria are non-culturable on artificial laboratory media, which could result in an underestimation of the quantity of viable bacteria present. CFUs are determined using a standard culturing protocol (i.e., growth media, time, temperature), dispersal of a fixed quantity (i.e., volume, weight, or sample collected from a fixed volume of air) of sample on the growth media, and counting the number of visible growth patterns of the same type. By diluting or disrupting the sample and then repeating the standard protocol, it is possible to release bacteria that are clumped together to determine the total number of viable agent entities in the sample. 2.1.2 Viruses Viruses are minute infectious agents that lack an independent metabolism and are able to replicate only within a living host cell. An individual viral particle (a virion) consists of nucleic acid (either DNA or RNA) and a protein capsid shell that contains and protects the nucleic acid; the shell may be multilayered. The host range of viruses is extremely broad; virions may infect bacteria (i.e., bacteriophages), plants, animals, and humans. Individual virions range from approximately 20 to 200 nanometers in size. However, without special treatment virions are usually associated with larger particles. Since viruses depend on a viable host cell for replication, naturally derived virions are often found in a mixed matrix with material in which they were grown. Viruses can be cultivated under controlled conditions in animals, eggs, and cell cultures. The material associated with virus-containing aerosol particles depends on how the virus was grown and processed before aerosolization. Therefore, aerosol particles containing viruses can be very diverse in size, chemical make-up and number of virions contained in an aerosol particle. Further complexity arises from the variation in potential host responses to viruses presented in such diverse size distributions and chemical matrices. Examples of viral pathogens and their associated diseases are: • Variola major—smallpox • SARS associated corona virus—severe acute respiratory syndrome (SARS) • Ebola virus—ebola hemorrhagic fever Aerosol particles containing viruses can be characterized using physical and chemical means as well as biological activity. Physical and chemical characterization methods are used to identify the presence of virus component materials. Viral nucleic acid is routinely analyzed by polymerase chain reaction (PCR). Capsid surface features are routinely identified by molecular recognition assays involving antibodies and other recognition moieties. These techniques do not assess whether the viruses are biologically active. (The ability of a virus to infect and replicate in a cell can be compromised by many environmental factors, including heat, sunlight, and humidity; different viruses are vulnerable to different environmental conditions.) Knowing whether the aerosol particle can establish an infection in a host cell culture or animal model is critical to determining the health hazard. Methods that are specific to a number of types of virus have been developed to characterize biological activity. These methods employ challenge of a known susceptible host cell culture, whole animal model, or other suitable host tissue, such as 

17 embryonated eggs. The challenge is followed by incubation of the host or cell culture for a period of time after which biological effects are assessed. The biological effect is often an obvious end point, such as host death or cell death that produces a plaque (an area of dead cells) on a tissue culture monolayer. The results of tissue culture assays are presented in such units as plaque-forming units (PFU) per unit volume for tissue culture assays, while such units as 50 percent lethal dose (LD50) or infective dose (ID50+) per unit volume, or sero-conversion, are used for whole animal models. These assays involve surrogate host cells and animal models that are meant to simulate or predict infectivity, morbidity, and mortality in humans. The above brief description of viruses, methods by which they can be characterized, and the challenges in predicting their impact on human health illustrates the difficulties associated with developing a unit of measure for virus-containing aerosol particles. • A very large number of virions can be contained in a single respirable aerosol particle and the minute size of the viruses make optical counting or physical dissociation of viruses impractical; • The virions may or may not be infective for a human host; direct assessment of human infectivity is currently virtually impossible to accomplish; and • While a quantitative analysis of viral constituents (e.g., nucleic acids, surface epitopes) can be readily accomplished, relating analyte mass to infectivity is extremely difficult, even for well-characterized viral strains. 2.1.3 Toxins Biotoxins, the third class of BWAs, are biologically derived substances that have a deleterious effect on an organism. Biotoxins range in size and composition from small molecular weight nonprotein toxins (e.g., aflatoxins) to large molecular weight protein toxins (e.g., botulinum). Biotoxins are naturally produced by a wide range of organisms, such as bacteria, fungi, algae, snails, and higher organisms. Advances in chemical synthesis and bioengineering have provided additional approaches to producing toxins that were originally isolated from naturally occurring organisms. The mode of production of a toxin and how it is processed affects the composition of the particle or other environment in which it is found. Unless a great effort is made to purify the toxin, it may be a minor fraction of the total mass of an aerosol particle. This presents a sensitivity challenge for direct detection of the toxin, but opens up additional opportunities for detection of toxin-related substances (e.g., castor bean DNA associated with ricin). Toxins can be quantitatively characterized in terms of chemical composition and physical structure or biological activity. A number of analytical techniques, such as mass spectrometry, provide great selectivity and sensitivity in analyzing the chemical composition of toxins; their physical structure can be recognized by antibody-based assays. These techniques determine the mass of sampled toxin, from which the mass concentration (mass/volume) can be calculated. Once again, it is their biological activity that is of principal importance. In some cases, toxins can be deemed to be present by chemical analysis even after biological activity has been lost due to conformational changes or inhibition. Toxins perturb host processes in many different ways (e.g., diphtheria toxin inhibits protein synthesis whereas botulinum toxin blocks neuronal 

18 transmissions); biological activity must be assessed by a method that mechanistically mirrors the toxicity pathway of the particular toxin in humans. Such in vivo and in vitro assays have been developed and the units of measure are often expressed in effective dose 50 percent (ED50) per unit volume. Toxin-containing aerosol particles present a relatively simpler case for characterization than that of bacteria or viruses. Nonetheless, complexity exists in host response depending deposition site in the respiratory tract, so aerosol particle size remains important. The host dose response curve and the effects of duration of exposure relative to total dose add additional uncertainty that must be addressed. At present it is not clear how the distribution of toxin molecules on aerosol particles will affect the health outcome. 2.1.4 Sampling and Analysis of Biological Warfare Agents There are two somewhat disparate elements contributing to the challenge of bioaerosol assessment, which sensor development efforts attempt to combine: microbiological measurement and aerosol measurement. Starting with microbiological measurements, three approaches have been employed historically to indicate microbial organism presence or identity: (1) culturing, (2) nucleotide sequencing or sequence recognition, and (3) antibody-antigen reactivity (immunological response). In these broad categories many specific measurement techniques have been developed. For example, enzyme-linked immunosorbent assay (ELISA) and electrochemiluminescence (ECL) are two of the many possible examples of distinctive measurement protocols that share a common detection mechanism of antibody-epitope binding. Of these three basic categories, only culturing conveys an assessment of viability, or more generally, biological activity. Results from culturing can be difficult to quantify or automate and often require lengthy incubation times. The combination of these features makes this measurement approach impractical for use in autonomous, continuous-operation detection systems. It should be noted that, as a measure of viability, the ability to culture an organism is limited by the inherent challenges of growing viruses or bacteria in artificial environments. Because of its acceptance as a legal precedent for microbial identification, culturing will likely continue to be used for confirmatory analysis beyond the immediate detection system capability. On the aerosol measurement side, three methods have been used historically to collect an aerosol sample from a volume of air: (1) filtration, (2) impaction, and (3) impingement. Filtration involves collection onto a filter substrate and is used to determine a time-averaged aerosol concentration. Filtration methods with adequate pore size will collect most particles larger than a minimum penetration size. Impaction uses the principle that an aerosol particle has significantly greater inertia than aerodynamic drag, making it unable to follow the air as it is deflected by an obstacle. This is exploited to separate the particle from the gas, typically onto a substrate. Impingement involves the transfer of a particle suspended in air into a liquid by impaction or by bubbling the air through the liquid. As with microbiological measurements, there are many specific embodiments of these three generic aerosol sample collection mechanisms and each has its unique characteristics. However, all methods require trade-offs that must be considered when developing and testing bioaerosol detectors. A simple bioaerosol detection system will consist of one of the above microbiological measurement methods combined with an appropriate aerosol collection method. This basic approach is illustrated by the Department of Homeland Security (DHS) Biowatch program, 

19 which couples aerosol filter collection with PCR, a technique that detects and amplifies specific DNA sequences. A strategy in which some aerosol property is monitored continuously is used for military applications. This monitoring does not necessarily require the collection of particles. Detection of a change in this property consistent with the presence of bioaerosols triggers the more intensive process of aerosol sample collection and microbiological measurement. One of the most successful aerosol monitoring methods to date has used optical spectroscopic properties of biological aerosols, especially ultraviolet excited fluorescence in conjunction with elastic scattering. Similar optical spectroscopic properties have also been exploited to provide a preliminary indication of the presence of bioaerosols at extended distances (i.e., provide some remote sensing or standoff detection capability). It should be noted that current capabilities are constantly changing due to fairly intensive research and development efforts to introduce and validate alternative measurement techniques. On the microbiological side, proteomics has been pursued to create an alternative to antibody- based identification such as measurement of an array of specific epitope binding units other with aptamers and antimicrobial peptides. There are also several physical-chemistry-based approaches involving mass spectrometry, ion mobility spectrometry, and higher resolution optical spectroscopies, such as infrared absorption or Raman scattering. Some of these developmental measurement approaches can be used both for rapid detection of single particles and with a collected macroscopic sample. It is possible that one of these approaches, or combination of approaches, could exhibit sufficient discrimination to be used as the main microbiological detection mechanism, displacing the microbial measurement methods listed previously. For aerosol measurements, recent developments in air-to-air particle concentrators, particle charging and manipulation, and selective particle collection may also augment current BW detection system capabilities. For the purposes of this report, it is important to remember that whichever collection method is used and whichever physical characteristics of the agent are measured, it is necessary to be able to relate the measurements made by a detector to the health hazard posed by the aerosolized agent. 2.2 PROPERTIES OF AEROSOLS 2.2.1 Transport and Dissemination The likelihood of disease in a person exposed to a BWA aerosol is dependent in large part on the properties and characteristics of the aerosol itself. Biological agents, including biotoxins, behave as particles when they are disseminated in air. Airborne particle size depends on the formulation of the agents and the method of dissemination. Large particles sediment more quickly than smaller ones, reducing the time they are likely to remain airborne.(Seinfeld and Pandis 2006) Settling velocity, while important, does not entirely determine how long particles will remain airborne, or how far they will be transported as shown in Figure 2.1. It has been suggested that based on their high settling velocities, large particles containing BWAs do not pose a threat except very near the source. If this were true, pollen-induced allergic rhinitis (e.g., hay fever) would be less common, because pollen grains range from 20 

20 µm to 100 µm in diameter, corresponding to settling velocities ranging from 1 to 30 cm/s.2 However, atmospheric turbulence and convection enable long-range transport by raising particles to altitudes well above those at which they were released. While most large particles may land close to the source, some are transported much farther than the terminal settling velocity would suggest. FIGURE 2.1 Particle sedimentation over distance. The top figure shows a cross-section through the center of the plume from a point release of 25 µm particles at a height of 100 m with a mean wind speed of 3 m/s and an eddy diffusivity of 10 m2/s. Shading denotes relative particle concentration. The gray line shows the trajectory of the particles in the absence of atmospheric turbulence. The bottom figure shows the concentrations at ground level. Turbulent dispersion causes the ground level concentration to peak well upstream of the point predicted by the kinematic model, but also allows particles to reach distances well downstream of the point estimated by fall time alone, and to be dispersed well to the side of the original release trajectory. SOURCE: Courtesy of Richard C. Flagan. Aerodynamic diameter (Dae) depends on several particle characteristics, including size, morphology, relative hygroscopicity, and density. The longer a particle remains airborne, the greater the potential for human respiration, so the aerodynamic diameter is directly related to risk of bioaerosol exposure. Diameter also greatly influences where a particle will be deposited in the 2 Birch and other tree pollens have been detected in Finland weeks before the trees there begin to flower, but after flowering has occurred farther south in Europe. Long-range transport of the pollen grains is the likely cause for this early appearance.(Hjelmroos 1991), (Matikainen and Rantio-Lehtimäki 1998) 

21 respiratory tract and thereby further links this physical characteristic with health risk. A number of aerosol collectors have been designed to collect fractions of the aerosol that crudely relate to their behavior in the respiratory tract; others (such as cascade impactors, aerodynamic particle sizers, optical particle counters, and aerosol mass spectrometers) segregate or measure particles in a number of size fractions from which more detailed assessments of respiratory tract deposition patterns can be determined. 2.2.2 Deposition in the Human Respiratory Tract Particle size is a major factor determining the probability that an inhaled particle will be deposited in a specific region of the human respiratory tract. The other two major factors affecting the deposition pattern of inhaled particles are the airway morphology and breathing physiology. The human respiratory tract can be divided into three anatomical regions as shown in Figure 2.2: (1) the extrathoracic (ET), or head airway, is the entry to the respiratory tract and the first defense against hazardous inhaled material; (2) the tracheobronchial (TB) tree, or conducting airway includes the trachea and 16 generations of branching airways. Gas exchange takes place in (3) the pulmonary region (P), which consists of alveolar ducts and alveolar sacs. Most people breathe through the nose during rest or light exercise, but switch to a combination of oral and nasal breathing during heavy exercise or work, because resistance through the oral airways is much lower than through the nasal airways. Some people are habitual oral breathers even at rest. When an airborne particle is transported near a person, it may be inhaled and enter the respiratory tract through either nasal or oral passages. The ability of the particle to enter the head airway, labeled its inhalability, is a function of its aerodynamic diameter.(Phalen et al. 1986) Based on experimental data obtained largely in aerosol wind tunnels, the inhalability (which is a fraction of airborne particles entering the human airway) is near 100 percent for particles smaller than 5 µm. The inhalability decreases as the particle size increases and stays at 50 percent for particles greater than about 50 µm. Once a particle enters the body through either the nose or mouth, it is deposited in different regions of the respiratory tract: larger particles deposit by inertial impaction or sedimentation, smaller particles deposit by diffusion. Electrostatic effects may enhance or modify deposition of charged particles of any size. Condensation of water in the humid environment of the respiratory tract may cause particles to increase in size, changing the way they move and deposit in the respiratory tract. The efficiency of particle deposition and the spatial distribution of deposition in the human respiratory tract have been measured experimentally in human volunteers using test aerosols, usually comprised of spherical particles tagged with radiolabel. Computational models have also been developed to estimate deposition in each airway region by each of the deposition mechanisms based on known properties of the aerosol particles. These models have been used to predict the aerosol deposition patterns in human lungs. Lung deposition models, such as those developed by the International Commission on Radiation Protection (ICRP 1994) and National Council on Radiation Protection and Measurements, (NCRP 1997) are based on airway anatomy and breathing patterns of standard men. These models have been used extensively to estimate radiation dose from exposure to 

22 FIGURE 2.2 Schematics of the human respiratory tract. SOURCE: Courtesy of Yung Sung Cheng. radioactive aerosols. Figure 2.3 shows an example of the deposition fractions as a function of aerodynamic diameter in men with nasal breathing for a tidal volume of 1.25 L/breath and 20 breaths per minute. Figure 2.4 shows an example of the deposition fractions as a function of aerodynamic diameter in a man with mouth breathing for a tidal volume of 1.25 L/breath and 20 breaths per minute. With mouth breathing, deposition in the oral airway region is much lower than in the nasal passages for particles of all sizes between 0.01 and 10 than with nasal breathing. As a result, there is substantial increase in deposition in the TB and P regions when mouth breathing is the primary inhalation method. Therefore, aerosol delivery of pharmaceutical agents to treat lung diseases or to target systematic effects uses the oral delivery route. In the battlefield, oral breathing is likely during light to heavy workload, increasing the risk of lung deposition for agent-containing particles of all sizes. 2.3 BIOLOGICAL EFFECTS OF INHALED PARTICLES Because particle size determines where aerosol material deposits in the respiratory tract, it is an important factor in predicting the health consequences of exposure to aerosolized bacteria, viruses, or toxins. However, there are only a handful of experimental studies in the medical literature that describe the effect of particle size on the nature of disease resulting from inhalation 

23 100 Deposition Fraction (%) 80 ET 60 P 40 TB 20 0 0.001 0.01 0.1 1 10 100 Particle Diameter (μm) FIGURE 2.3 Distribution of particle deposition for different regions of the respiratory tract system for 100 percent nasal breathing. For bacteria, the relevant region is 1 µm and larger. For virus and toxins, dimensions much less than 1 micron are relevant. For particles larger than 0.02 µm in diameter, deposition in the extrathoracic (ET) region increases as particle size increases up to 8 µm. ET deposition also increases when particle size decreases from 0.02 µm. Increased deposition of the larger particles is driven by an inertial mechanism, and deposition of the smaller particles is a result of diffusion. ET deposition decreased slightly for particles greater than 6 µm, because of decreasing inhalability of large particles. Particles greater than 6 µm in diameter deposit primarily in the nasal airways with little penetration to the lung. Particles in the size range between 0.002 and 5 µm can penetrate into lower airways and deposit in the pulmonary regions. The data were calculated using the LUDEP software (NRPB, Oxon, UK) based on the ICRP model. (ICRP 1994) NOTE: ET = Extrathoracic region, P = Pulmonary region, TB = Tracheobronchial region. SOURCE: Courtesy of Yung Sung Cheng. 

24 100 Deposition Fraction (%) 80 60 ET 40 P TB 20 0 0.001 0.01 0.1 1 10 100 Particle Diameter (μm) FIGURE 2.4 Distribution of particle deposition for different regions of the respiratory tract system for 100 percent mouth breathing. For bacteria, the relevant region is 1 µm and larger. For virus and toxins, dimensions much less than 1 micron are relevant. The data were calculated using the LUDEP software (NRPB, Oxon, UK) based on the ICRP model (1994).(ICRP 1994) NOTE: ET = Extrathoracic region, P = Pulmonary region, TB = Tracheobronchial region. SOURCE: Courtesy of Yung Sung Cheng. of a biological aerosol and most were published before 1960. In the first report to examine this question from the point of view of BWAs, laboratory animals were exposed to radiolabeled spores of a simulant for B. anthracis, of various mass median aerodynamic diameter (MMAD).(Harper and Morton 1953) That study showed that a large fraction of particles in the 1- 4 µm MMAD (in guinea pigs) or 1-6 µm (in monkeys) size range entered the lungs, while larger particles were almost entirely retained in the head and trachea. Even though deposition patterns in humans and various animal species differ because of variations in respiratory tract anatomy and breathing dynamics, experimental studies are in agreement that the infectivity of an aerosolized pathogen is greatest when it includes significant numbers of 1-5 µm particles that can reach the bronchioles and alveolar spaces.(Druett et al. 1953; Druett et al. 1956; Druett, Henderson, and Peacock 1956; Schlesinger 1985) The lower probability of disease initiation per inhaled organism for particles of greater than 5 µm results from a combination of factors. First, the upper airway presents a much smaller target for deposition than the huge total surface area of the bronchioalveolar spaces. In addition, particles that deposit above the level of the respiratory bronchioles encounter a layer of mucus that limits their contact with epithelial lining cells. Those depositing above the respiratory bronchioles and below the trachea are also subject to the sweeping action of cilia that propel them upward out of the lung for removal by coughing or swallowing. These larger particles cleared from the ET and P respiratory regions will mostly be ingested. This introduces another disease pathway that strongly overlaps alternative BW attack modes of food and water source contamination. For this reason, pathogenesis of the digestive 

25 system was not specifically explored in this report, but may be particularly relevant to toxin agents. Although this section focuses on respiratory system response as a main pathway for airborne agents, the framework recommended here is easily extended to additional routes and mechanisms of disease. Similar considerations of efficacy as a function of particle size for cutaneous or ingestive routes likely exist. Prior literature or experimental data on these routes are even less prevalent than for respiratory effects. By contrast, the greater pathogenicity of a microbe or toxin delivered in a 1-5 µm particle reflects its ability to reach the large exposed surface area of the alveoli that lacks both a mucus barrier and ciliated epithelium. Instead, such particles are cleared by local macrophages that may actually serve as sites of replication for such pathogens as Bacillus anthracis, Francisella tularensis, Mycobacterium tuberculosis, and a number of viral agents. In the case of inhalational anthrax, when the anthrax spores reach alveolar spaces, infected cells are cleared by the immune system into the mediastinal lymph notes, resulting in more severe disease than when the agent is confined to the upper respiratory tract and the agent is cleared to the cervical and peribronchial lymph nodes.(Druett et al. 1953) Even though pathogen-containing particles larger than 5 µm MMAD have a lower probability per particle of initiating infection than smaller particles, they are still capable of causing severe or fatal disease. The illness induced by large particles in laboratory animals differs in its clinical features from that produced by a small-particle aerosol. It is generally characterized by a lower mortality rate and usually, but not always, by a longer incubation period. For example, a 1-5 µm MMAD aerosol of B. anthracis caused hemorrhagic mediastinitis in rhesus macaques, similar to inhalational anthrax of humans, but delivery of the same agent in 12 µm particles resulted in massive edema of the soft tissues of the head and neck, presumably as a result of the initial spread of infection to cervical lymph nodes.(Druett et al. 1953) Similarly, while both small- and large-particle aerosols of Yersinia pestis caused fatal disease in guinea pigs, an aerosol composed of single organisms produced bronchopneumonia, leading to septicemia and death. A 12 µm MMAD aerosol actually produced septicemia more rapidly, but without causing pulmonary disease.(Druett, Henderson, and Peacock 1956) Experiments using aerosolized Brucella suis ranging in size from single organisms through 12 µm particles showed that the smallest particles were 600-fold more likely to initiate disease, presumably because they were able to replicate more readily after depositing on the vast exposed surface of the pulmonary alveolar epithelium.(Druett et al. 1956) Similarly, monkeys exposed to aerosols of F. tularensis in which mean particle sizes were 2.1 or 7.5 µm became ill and died more quickly than those exposed to the same agent in 12- or 24-µm particles.(Day and Berendt 1972) As might have been predicted, the first group developed a diffuse pneumonitis, while the second showed massive infection of the upper airway. Another study compared the effect of small- and large-particle aerosols of influenza virus in mice and found that a much lower delivered dose of the small-particle aerosol was required to initiate disease.(Hatch 1961) Particle size is also a factor that determines the biological effects of an aerosolized toxin, especially if the substance exerts its damaging effects directly on the lining of the respiratory tract. A recent study using the toxin ricin showed that a small-particle aerosol caused lethal illness in guinea pigs, while the same dose of delivered material in the form of 12-µm particles did not cause death.(Roy et al. 2003) Although ricin-containing particles presumably cause mucosal injury wherever they are deposited in the respiratory tract, deposition of large particles in the upper airway was apparently less damaging to the animals' ability to breathe than the deposition of small particles in the linings of small bronchi and alveoli, which led to diffuse airway obstruction. 

26 Once particles deposit on the surface of the respiratory tract, they do not necessarily remain intact; they may absorb water and disperse into their smaller component parts. Therefore, it follows that the threat to health posed by an aerosol of bacteria, virions, or toxins would be greater as the average number of infectious agents or the mass of toxin contained in each delivered particle increases. This logical conclusion was verified in the only experiment of its kind to be found in the medical literature; guinea pigs were exposed to aerosols of B. anthracis of different particle sizes, and the average number of spores per particle was controlled by adding dextrin to the suspension before aerosolization.(Druett et al. 1953) The investigators found that different combinations of particle size and the number of spores per particle could produce similar health effects, and concluded that “ to achieve the same mortality, the total number of spores presented to the animal, in particles of a given size, must be the same, irrespective of the number of particles carrying them.” This result suggests many different factors contribute to the nature of the disease produced by an aerosolized microbe or toxin, including the identity of the agent, the vulnerability of the host, and the size of the aerosol particles. The severity of the disease will increase with the average number or microbes or mass of toxin in each particle. However, when an aerosol contains a variety of particle sizes, those in the range of 1-5 µm will always pose the greatest threat, even if they represent only a minor fraction of the inhaled material.(Hatch 1961) As demonstrated in the above section, a limited number of reports, mostly published before 1960, have shown that particle size plays an important role in determining the probable health outcome after exposure to BWA aerosols. More research in this area using modern laboratory methods is clearly needed to better understand that role. For the detector testing community, this has implications for test design. As more information becomes available, the committee recommends that test guidelines be updated, published, and peer reviewed to reflect the improved understanding of particle size and health effects. 2.4 PARTICLE SIZE AND AEROSOL COLLECTION Aerosols are described physically in terms of the aerodynamic diameter distribution of particles. For biological aerosols, the diameter and viability may be greatly influenced by air temperature and relative humidity. The biological content per particle is also important, as both size and biological content are needed to link to health hazard and potential mitigation strategies. Monitoring particle size distribution presents technical challenges for testing of bioaerosol detectors. Most current detector collection systems capture particles over a range of diameters, for example particles smaller than 10 µm diameter (PM10), which can penetrate beyond the head and thoracic regions during mouth breathing, and particles smaller than 2.5µm diameter (PM2.5), which are respirable aerosols that can penetrate into the lower airways during nasal breathing. These characterizations do not, unfortunately, provide sufficient information to assess where in the respiratory tract inhaled particles will actually deposit. Bioaerosol samplers based on inertial deposition (e.g., slit-to-agar samplers) collect only particles larger than a critical aerodynamic diameter, while excluding smaller particles that may contain spores, virions, or toxins. In the study of hazardous components of the environmental aerosols, smaller particles have been shown to have disproportionately large health impacts. Several studies have suggested that surface area (Oberdorster 2000) or number concentration (number/cm3)(Donaldson et al. 2000) may better represent the health impacts of such particles than does the particle mass. Once a collection system has concentrated the aerosol into a single sample, it is no longer 

27 possible to reconstruct the size distribution. Hence, the potential distribution of deposited particles in the respiratory tract cannot be accurately estimated or used to inform response priorities. The collection system can be conceptualized mathematically as a transfer function that blurs or integrates over a range of particle sizes. Therefore, performance of a detection system depends on the distribution of particle sizes, and this system bias should be determined as part of test and evaluation. Figure 2.5 illustrates a model system with good performance to a uniformly distributed 1-5 µm aerosol but with unacceptable performance to a 1-10 µm aerosol. FIGURE 2.5 Illustration of a transfer function of a hypothetical detection system. After collection it is not possible to reconstruct the size distribution. In this example, particles with an aerodynamic diameter less than 1 or greater than 5 μm will not contribute to the signal. 2.5 CURRENT UNITS OF MEASURE AND HEALTH HAZARD Conceptually, it is now possible to see more clearly the limitations of the current standard unit of measure for biological aerosols. The Agent Containing Particle per Liter of Air (ACPLA) unit has become the de facto standard for measuring performance of biodefense aerosol detection systems. Independent of specific implementation, ACPLA is the number of particles that contain at least one biological agent in a liter of air. ACPLA appropriately recognizes that particles with less than one biological agent (i.e., zero) are not a biodefense concern. However, there are limitations to using this unit of measure. Consider whether ACPLA accurately estimates exposure to biological warfare agents. ACPLA is in practice inferred from measurements of CFU, PFU or micrograms of toxin. Measures of CFU or PFU provide information about the number of particles containing culturable agents, but this may not yield the total number of agent units (exposure or dose) that is important for estimating health consequences. As an example, a measure of 10 ACPLA in reference to anthrax could equivalently describe either 10 individual spores per liter or 10 particles of size 7 µm per liter, each containing over 100 spores. In terms of irreducible units (spores), the concentration can vary by orders of magnitude for the same ACPLA designation. 

28 The distribution of the biological content is also important. A thousand bacteria in a single 10 µm particle may represent a different health risk than a thousand 1 µm particles each with one bacterium. Obviously, if the bacteria are not viable or have lost the ability to be infective or pathogenic, the health hazard has changed dramatically. Current research is suggesting that nonviable content in the aerosols—including lysed cellular material—may contribute to the health risk or, to the contrary, provide benefit to the immune response in some cases. The currently used unit of measure, ACPLA, and current state-of-the-art detection capabilities, cannot take these factors into account. A further limitation of ACPLA is that it does not contain any indication of particle size and particle size distribution, characteristics of the aerosol clearly important in evaluating health hazard to humans. For aerosolized particles with biological content, the aerodynamic diameter is a critical parameter in determining the final site of deposition in the respiratory tract and must be considered. In light of these limitations, the committee developed a new, robust framework for evaluating the health hazard posed by biological aerosols. The next chapter describes the recommended framework and its mathematical and conceptual reasoning.