Environ. Sci. Technol. 2008, 42, 3145–3154
Exposure Assessment of Carcass Disposal Options in the Event of a Notifiable Exotic Animal Disease: Application to Avian Influenza Virus SIMON J. T. POLLARD,† G O R D O N A . W . H I C K M A N , * ,‡ PHIL IRVING,§ RUPERT L. HOUGH,† DANIEL M. GAUNTLETT,‡ SIMON F. HOWSON,‡ ALWYN HART,| PAUL GAYFORD,⊥ AND NICK GENT# Cranfield University, Centre for Resource Management & Efficiency, School of Applied Sciences, Cranfield, Bedfordshire, U.K., MK43 0AL, Animal Health, Department for Environment, Food & Rural Affairs, Contingency Planning Division, Contingency Plans & Disposals Branch, 17 Smith Square, London, U.K., SW1P 3JR, Environment Agency, King’s Meadow House, King’s Meadow Road, Reading, Berkshire, U.K., RG1 8DQ, Environment Agency, Olton Court, 10 Warwick Road, Solihull, U.K., B92 7HX, Surveillance, Zoonoses & Emerging Issues Division, Department for Environment, Food & Rural Affairs, Nobel House, 17 Smith Square, London, U.K., SW1P 3JR, and Emergency Response Division, Health Protection Agency, Porton Down, Salisbury, Wiltshire, U.K., SP4 0JG
Received November 21, 2007. Revised manuscript received January 17, 2008. Accepted January 28, 2008.
We present a generalized exposure assessment of 28 disposal options for poultry carcasses in the event of a highly pathogenic avian influenza (HPAI) outbreak. The analysis supports a hereto unverified disposal hierarchy for animal carcasses, placing waste processing (e.g., incineration and rendering) above controlled disposal (e.g., landfill), above uncontrolled disposal (e.g., burial on-farm). We illustrate that early stages of the disposal chain (on-farm) pose greater opportunities for exposure to hazardous agents than later stages, where agents are generally contained, wastes are treated, and residues are managed by regulated processes. In selecting carcass disposal options, practitioners are advised to consider the full range of hazards rather than focusing solely on the HPAI agent, and to give preference to technologies that (i) offer high destruction efficiencies for target pathogens; (ii) do not give rise to significant releases of other pathogenic organisms; and (iii) do not release unacceptable concentrations of toxic chemicals. The approach offers an exposure assessment perspective for carcass disposal, thus providing a risk-informed basis for contingency planning and operational intervention.
* Corresponding author phone: +44 (0)20 7238 5111; fax: +44 (0)20 7238 1250; e-mail:
[email protected]. † Cranfield University. ‡ Animal Health, Department for Environment, Food & Rural Affairs. § Environment Agency King’s Meadow House. | Environment Agency, Olton Court. ⊥ Surveillance, Zoonoses & Emerging Issues Division, Department for Environment, Food & Rural Affairs. # Health Protection Agency. 10.1021/es702918d CCC: $40.75
Published on Web 03/28/2008
2008 American Chemical Society
The authors recognize that relevant legislation, public perception, available capacity, and cost also need to be considered when selecting disposal options in the event of HPAI.
Introduction Policy Context. Western governments are now managing outbreaks of the highly pathogenic, avian influenza (HPAI) A virus and its transmission to commercial poultry flocks (1–3). Of greatest concern is the HPAI H5N1 strain, since it causes serious disease in birds and can also infect humans who have been in close contact with infected birds. Contingency planning is essential and some governments are in a heightened state of alert and preparedness (4, 5). Policies are expected to be proportionate, risk-informed, evidencebased, and justified. Previous outbreaks of infectious animal disease have required the disposal of significant numbers of infected or culled stock (6–11) and, while the likelihood of humans acquiring HPAI in the United Kingdom (UK) is considered to be very small at present, carcass disposal cannot be performed without risks to people, animal health, and the environment (11, 12). Decisions on carcass disposal should help secure contingency plan objectives (biosecurity) and address issues of cost and regulatory compliance. Governments should not be concerned solely with either the cheapest or the safest disposal option at any cost; rather with developing a hierarchy of responses, topped by the optimal response. The development of a disposal hierarchy can be supported by (i) a general analysis of the exposures posed by a range of possible carcass disposal options, so that decisions on their suitability are risk-informed (this study); (ii) site-specific risk assessments, so that local environmental settings and exposures during planned or actual disposals are accounted for; and (iii) live, expert advice from health departments, health agencies, veterinary officials, environment agencies, emergency planners, researchers, and local operational partners, so that risk management is sensitive to conditions “on the ground”. Critically, the range of hazards in these situations extends well beyond individual notifiable pathogens, to include all hazards associated with a carcass, its byproduct, and the associated infrastructure of disposal, including detergents, disinfectants, veterinary medicines, other animal-borne pathogens, and amenity hazards such as odor and noise. The intensification of agriculture since the 1960s has demanded new strategies for contingency planning. Great Britain (GB) has a high human population density and carcass disposal is often in close proximity to communities and in the public view. We now have individual poultry units housing >1 million chickens on a single site. When culls are required, operational staff must kill animals quickly and humanely and have them disposed of safely and responsibly, without undue delay. The practical difficulties of maintaining biosecurity on the ground require that operational procedures are risk-informed. Calls for the adoption of historically acceptable practices, such as “on-farm” burial, become challenged by the sheer mass of culled stock and modern environmental legislation, which demands that greater consideration is given to the long-term environmental impacts of disposal. International Context. Governments, companies, and individuals managing mass animal culls must make credible, evidence-based decisions in the face of intense political, media, and public scrutiny. The scale of carcass disposals in large outbreaks may represent a “shock load” to the waste management infrastructure of countries managing these events. Since 1997, in Asia alone, over 200 million chickens VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3145
and ducks have been killed by HPAI H5N1, or culled in efforts to contain the disease (13). In 2004, the Canadian Ministry of Agriculture, Food and Fisheries incinerated and composted ca. 17 million culled birds following an outbreak of the low pathogenic H7N3 AI virus in British Columbia (14). Similar disposals, adopting a range of waste technologies have been required (15, 16) in the United States (1983–1984; H5N2 strain; ca. 17 million birds culled), in Hong Kong (1997; H5N1; ca. 1.5 million chickens), in Italy (1999–2001; H7N1; ca. 13 million birds), in Virginia (2002; low pathogenic H7N2; 5 million birds), in The Netherlands, Belgium, and Germany (2003; highly pathogenic H7N7; ca. 28 million poultry rendered and landfilled), in Hungary (2006; highly pathogenic H5N1; ca. 13 000 geese buried) and most recently in the UK (2006; low pathogenic H7N3; ca. 49 000 poultry; incinerated and rendered; February 2007; highly pathogenic H5N1, ca. 159 000 turkeys rendered; May 2007; low pathogenic H7N2; ca. 60 backyard poultry incinerated (17)). The selection of carcass disposal options in these outbreaks, other than the recent 2007 H5N1 UK outbreak, appears to have been based on the scale of the requirement, the available waste technology capacity, and the logistics of local disposal, rather than on any examination of potential exposures to public and animal health and to the environment. Many responses in South East Asia have adopted uncontrolled disposal without apparent reference to the risks across the full process chain and, to our knowledge, no exposure assessment for carcass disposal in the event of an AI outbreak, addressing all the potential hazards, has been published. Carcass disposal raises a suite of logistical, legal, economic, and sustainability issues that are reviewed elsewhere. Ellis (18) provides a valuable critique of carcass disposal, particularly following natural disasters in the United States, in which a range of waste technologies have been applied. Morgan (19) reviews the U.S. Department of Agriculture’s (USDA’s) emergency planning system for AI. The National Agricultural Biosecurity Center’s (Kansas State University) Carcass Disposal Working Group has published an authoritative review (20) of carcass disposal options and cross-cutting policy issues for those managing disposals. A high priority for contingency plans is to identify protective measures so that tested disposal strategies can be applied when outbreaks occur. Examples of the practical considerations required to maintain biosecurity include (i) securing derogations from existing legislation to allow the rapid development of storage, treatment, and disposal facilities, and (ii) the rapid development of operational procedures. Study Rationale. The UK has experienced two overlapping and significant epidemics of animal disease in the last 12 years: bovine spongiform encalopathy (BSE; (21)) and foot and mouth disease (FMD; (22)). Both had substantive, longlasting consequences for individuals, farming communities, animal health, the environment, and other stakeholders and receptors. In February 2007, the UK successfully managed an outbreak of HPAI (H5N1) in a flock of commercial turkeys, 159 000 of which were rendered more than 300 km from the site of infection in Eastern England. In preparation for threats of this kind, the Department for Environment, Food and Rural Affairs’ (Defra) and its’ executive agency, Animal Health, formerly the State Veterinary Service (SVS), prepared a generic contingency plan for exotic animal disease outbreaks covering FMD, AI, Newcastle disease, and classical swine fever (4). The version of the plan laid before Parliament in December 2006 (23, 24) proposes a disposal hierarchy (Figure 1) for animal carcasses informed by a qualitative analysis of risks to public health from FMD disposals undertaken in 2001 (25). There has since been a desire within Government to expose the hierarchy to greater scrutiny and test its robustness for application beyond FMD disposals. It may be tempting to suggest such an analysis is not required. However, given 3146
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008
FIGURE 1. Disposal heirarchy presented in the exotic animal disease generic contingency plan, presented in decreasing order of preference (4). the range of carcass disposal hazards, the differences in environmental fate posed by the sources of these hazards, and the substantive volume of stock culled during these outbreaks, GB has considered it prudent to review its disposal strategy by reference to the potential exposures that could occur. The objective of this research was to contribute to this test by undertaking an exposure assessment of disposal options for AI-infected bird carcasses. We have drawn on the following: Defra’s guidelines on environmental risk assessment and management (26–28); a review of the prior work and critical analysis of good practice (29); prior work contained within Defra’s generic contingency plan (2005; (4)); development of a methodology for reviewing disposal options in the event of an outbreak of a notifiable exotic animal disease that informed this study (2006); the methodology within the Department of Health’s rapid qualitative assessment of possible risks to public health from current foot and mouth disposal options (2001; (25)); a series of meetings and workshops with 14 policy, academic and regulatory professionals, including a facilitated 2-day exposure analysis workshop (2006); internal consultation of study group members (2006); an independent peer review (November 2006); and consultations on the approach with a range of stakeholders (January 2007). We were concerned with a generic assessment of exposures posed by the disposal of animal carcasses prior to their mitigation and, at the analysis stage, did not wish to be constrained by considerations of legislative compliance, costs, or public perception, though these important factors are discussed later. One reason for this is the ability of governments, should they desire, to request derogations from the current statute in emergencies; for example, on waste incineration and burial.
Methodology Carcass Disposal in Great Britain. Carcass disposal is a multistage process (4, 11, 12, 20) initiated by the collection of fallen stock or the culling of live stock (e.g., birds), with subsequent disinfection of carcasses, their transport to localized collection points, and the potential off-site transport to waste processing (e.g., incineration, rendering) facilities. UK poultry farms have ca. 90 000 birds on average, though larger farms have up to 1 million birds and there are >258 million poultry in GB held at 23 762 premises (as of January 17, 2007; 30, 31). Preventive measures for AI are prescribed by European statute and the domestic regulations that transpose European Directives (32). In the event of a need for off-site disposal, carcasses are loaded into specialist leakproof vehicles whose movements are licensed by Animal Health (4). Vehicles are leak-tested and travel to a disposal facility by a predetermined route. Each load is covered and escorted to ensure the vehicle does not develop a leak. Transport of infected carcasses is also covered by the International Carriage of Dangerous Goods by Road (ADR) regulations. At the disposal facility, carcasses are disposed
of under Animal Health supervision. Following the initial removal of the carcasses from the site of discovery, all litter, equipment, and areas occupied by the affected poultry are sprayed down with an approved disinfectant and left to dry for 24 h, this being referred to as preliminary “cleansing and disinfection” (C&D). The owner/keeper of the poultry is then responsible for undertaking a thorough cleansing and disinfection of the premises (final C&D; under Animal Health supervision) and for the treatment and controlled disposal of waste litter. This involves removing the litter and other organic material and a full “wash down”, followed by degreasing and application of an approved disinfectant at a recommended dilution. Restocking can commence a minimum of 21 days after final C&D is complete. The characteristics of the early stages of this process are common to most disposal options. Differences arise later, mainly in the extent to which waste processing alters the containment of hazardous agents, their relative destruction, and the generation of post-treatment waste residues (e.g., ash, or meat and bone meal) requiring onward disposal. The carcass disposal “process chain” has five key stages: (1) disinfection and collection; (2) transport; (3) preprocessing; (4) processing; and (5) dealing with residues. A consideration of disposal options must recognize two critical factors: (i) the range and distribution of opportunities for exposure along the length of the process chain; and (ii) the available reductions in exposure, and thus risk, offered by individual technologies during waste processing. Both are important because the stages that pose the greatest opportunities for exposure may be those that precede waste processing. These exposures can be reduced, but not eliminated, by good operating practices and the proper use of protective measures (wearing personal protective equipment), so that exposures to front-line workers and the environment are minimized. These issues must be considered and risk management measures must be put in place as routine. Exposure Assessment Overview. A methodology was developed and applied during an expert workshop facilitated by Cranfield University in March 2006. The intent was to focus on AI, as one example of an exotic animal disease, and test the methodology by undertaking an exposure assessment of disposal options for poultry carcasses. The research was desk-based and informed by a panel of fourteen crossGovernment scientists, researchers, and policy specialists working within the Government guidelines on environmental risk management (26–28). Workshop attendees undertook six structured tasks: (i) hazard identification, screening, and prioritization; (ii) representation of the process chain for a series of disposal options and consideration of the efficacy of barriers for risk reduction along each process chain; (iii) review of a master list of exposure pathways for hazards from carcasses (25); (iv) development of conceptual exposure models for a series of process chains and a semiquantitative exposure assessment for each option; (v) development of a draft disposal hierarchy informed by (i)-(iv); and (vi) a preliminary screening of operating constraints acting on the disposal options. (i) Hazard Identification, Screening, and Prioritization. The approach first identified a set of hazardous agents using expert veterinary, environmental, and public health expertise. We applied a successive series of filters (Figure 2) to a pooled set of hazards to identify the most significant hazardous agents. The pooled set included the following biological and chemical agents and amenity issues associated with an avian influenza outbreak, a key aspect of the analysis being this review of all potential hazards associated with a HPAI outbreak and not just the HPAI agent: (i) avian influenza (H5N1), avian influenza (other strains), Campylobacter spp, Salmonella spp, Yersinia (Yersinia enterocolitica, Yersinia fredriksenii, Yersinia pseudotuberculosis, and Yersinia kris-
FIGURE 2. Approach to hazard screening. The application of three progressive filters to identify significant hazardous agents of concern (adapted from ref (13)). tensenii), Q fever and bioaerosols (including fungal spores); (ii) ammonia, ammonium, disinfectants, veterinary medicines (pesticides, antibacterials, coccidiostats, barbiturates), methane, organic carbon, hydrogen sulfide, carbon monoxide, benzene, toluene, ethylbenzenes, and xylenes, wood resins, dioxins/furans and dioxin-like polychlorinated biphenyls (PCBs), polynuclear aromatic hydrocarbons (PAHs), oxides of nitrogen and sulfur (NOx, SOx) airborne particles (PM10), heavy metals, hydrogen chloride, metal salts; and (iii) odor, smoke, and noise, as amenity impacts. Application of the filters (Figure 2) produced a subset of hazardous agents (Table 1) posing potentially significant human health, animal health, or environmental impacts; capable of evading destruction; and capable of presenting at sufficient doses to be of potential concern. We assumed those agents passing filter 3 could pose significant harm were exposure to occur within a typical (though generalized) setting. (ii) Representation of the Process Chain. Having identified the key hazards, the likelihood of exposure to these agents for each process chain was evaluated, qualitatively in the first instance (Figure 3). The process chain for a series of disposal options was described by reference to five main stages, further broken down into twelve component stages (Figure 3): disinfection and collection: containment in situ, destruction at source, gather carcasses to local store, contained storage of gathered carcasses, containment during local processing, and destruction during local processing; transport: biosecure transport to facility; preprocessing: containment prior to processing; processing: containment during off-site processing and destruction during processing; dealing with residues: residuals containment and residuals destruction. The length of the process chain is option-specific (Figure 3). Options may have very short process chains that terminate early (e.g., do nothing; leave carcass in situ), or longer chains VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3147
TABLE 1. Hazard Screening and Prioritization: Hazardous Agents Passing Filter 3 hazardous agent
qualifying comment on amount and/or timing of exposure to agent
avian influenza (H5N1)
for animal health (low dose exposure will cause disease in birds); a negligible public health risk
avian influenza (other strains)
for animal health (low dose exposure will cause disease in birds); a negligible public health risk for H5 and H7 strains
Campylobacter spp
prevalent pathogen; several potential pathways; low dose exposure will cause disease in animals and humans
Salmonella spp
present in UK pigs, cattle, and sheep; prevalence falling in recent years but still seen as an important pathogen (control of salmonella strains in poultry is very good)
veterinary medicines (in bird): pesticides, antibacterials, coccidiostats, barbiturates
if poultry are killed before they would normally have been, there is a possibility that veterinary products may be present in the body at elevated levels (within the normal withdrawal period); if killing is by lethal injection there may be barbiturates present, potential impact on the environment
ammoniacal nitrogen/ ammonia gas
present naturally and in high-strength leachate from decaying carcasses; significant potential impact on water bodies, especially groundwaters
disinfectants
used routinely on poultry units for biosecurity, but in the event of an outbreak of avian influenza usage will increase; carcasses will be sprayed with disinfectant prior to disposal; most disinfectants contain listed priority substances and should not be discharged to controlled waters
detergents
from cleansing and disinfection, potential impact on the aquatic environment through percolation and runoff
benzene, toluene, ethylbenzenes, xylenes
combustion byproducts, potential impact on air quality
wood resins and wood preservation products
poor combustion could release hazards, impact on air quality and ash quality
particulates
combustion products, nuisance and public health impact
odor
amenity and public health impact on local communities
smoke
amenity and public health impact on local communities
noise
amenity and public health impact on local communities
where carcasses are transported some distance to regional facilities, treated, and the residues then disposed of at some distance again from the treatment facility (e.g., incineration). The extent to which each stage in the process chain offers opportunities for the release of a hazardous agent, and thus exposure, was assessed qualitatively, by the expert panel considering the likely efficacy of existing barriers to exposure at each stage along the process chain. In general (Figure 3; red notation), early stages of a process chain offer more opportunities for the release of hazardous agents, thus more potential for exposure. Subsequent waste processing, where this is a component, offers differing degrees of effectiveness depending on whether destruction of the hazardous agent can be assured or not (amber, blue, green notation, Figure 3), and the degree of engineering process control. Exposures in the latter stages of the process chain are governed by the relative containment of process residues. (iii) Semiquantitative Exposure Assessment. A semiquantitative exposure assessment was completed, first by reference to a master list of potential exposure pathways for hazards from carcasses developed during the 2001 FMD outbreak (25). The master list was developed through an understanding of the conceptual exposure models (Figure 4, for incineration) for different parts of the process chain. Having verified the 3148
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008
suitability of this list to the case of AI, a series of exposure schematics was developed in Excel spreadsheet layers for a number of the process chains (not presented for brevity). This included exposure pathways associated with the transport of pathogens and chemicals through the environment and exposure routes such as the ingestion of water, inhalation of volatiles, direct contact with animals and people, human and animal consumption of fish or shellfish (raised in contaminated water), crop irrigation and consumption, the ingestion of irrigation water and the consumption of animals (Figure 4). Exposure pathways that are more available to hazardous agents are more likely to result in an exposure. Such exposures are of interest as they could be associated with significant levels of risk as only those hazards with the potential to cause serious health effects have been considered (see Figure 2 and Table 1). Potential pathways were ranked for human, animal, and environmental receptors using a linear scale by the expert panel, as being (nominal ranks in brackets): not available (0); of negligible availability (1); of low availability (2); of medium availability (3); or of high availability (4). Ranking was achieved by considering the environmental and microbiological characteristics of the key hazards, and applying expert knowledge of the principal environmental
FIGURE 3. Qualitative assessment of process chains incorporating a range of waste processing technologies and categorization of carcass disposal options (WID ) Waste Incineration Directive approved; PPC ) Pollution Prevention and Control Directive approved). pathways for these hazards. The consequences of impacts at the receptor were not considered in detail and were not weighted. For each disposal option, the exposure assessment represents an evaluation of the relative likelihood of exposure to the key hazards identified for humans, animals, and the environment across a multistage process chain. A number of metrics were developed for comparing exposure opportunities for carcass disposal options (Table 2): (i) the total number of available pathways, representing the number of opportunities for exposure to occur; (ii) the pathway score for human, animal, and environmental receptors, and then for all pathways; a surrogate for the overall likelihood of exposure, accounting for the relative likelihood of all
pathways; (iii) the total number (n) of medium to high availability pathways, representing the number of opportunities for significant exposures to occur; and (iv) the percentage (%) of medium to high availability pathways across the full process chain, a surrogate for the likelihood of significant exposure. At this level of analysis, it was not possible to infer a dose, nor identify the consequences of exposure, because individual receptors were not considered, as they would be in a sitespecific analysis. Neither does this analysis take account of the additional measures that might be taken in the event of an outbreak to mitigate the risks of exposure. What it does provide is insight into the opportunities that exist for releases VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3149
FIGURE 4. Generalized conceptual exposure model for incineration, used as a precursor for Excel spreadsheet layers. of hazardous agents, and the distribution of potential exposures along a process chain. This is critical because risk management, in the form of sound contingency measures, needs to be targeted at critical control points; that is, at those points where the risk of exposure is greatest and where the opportunities for intervention are clear. (iv) Development of a Risk-Informed Disposal Hierarchy. The results of the qualitative analysis (ii) and semiquantitative exposure assessment (iii) were critically reviewed and synthesized and used to derive a draft disposal hierarchy.
Results and Discussion Application of the three filters (Figure 2) produced a subset of hazardous agents posing potentially significant human health effects, animal health effects, or environmental impacts, capable of evading destruction and presenting at sufficient doses to be of potential concern. Expert comments (Table 1) support the inclusion of these agents and offer qualifying statements as appropriate. The qualitative analysis is an assessment, across the complete process chain, of the relative effectiveness of each disposal option to contain hazards (Figure 3). Process chains ending with stages that are effective (or mostly effective) at reducing exposures are preferred because these indicate minimal residual exposure. As one descends Figure 3, however, there are fewer options terminating with effective barriers for reducing exposure, hence a residual risk remains. All options will, if selected and used in practice, be subject to regulatory controls to ensure that risks are mitigated. Risk management measures adopted by operators need to focus on stages in the process chain where exposures could otherwise take place. We do, however, recognize the practicality that during the adoption of contingency measures, regulatory and operator controls may not be as completely effective as under normal conditions and so some aspects of the process chains (e.g., containment prior to destruction) account for this. The central insight is the incremental loss 3150
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008
of process control on exposure, as one descends the list of options in Figure 3. An example individual conceptual model is presented in Figure 4, from which Excel representations were constructed and pathways were ranked. Rankings were aggregated for each Excel layer and presented using the chosen metrics (Table 2). Disposal options (Table 2) with (i) a large proportion of highly available pathways (“leaky” process chains, irrespective of length); and/or (ii) large numbers of pathways (typically, longer process chains) are more likely to result in significant exposures. These options pose a greater potential risk to public health, animal health, and the environment but, while this is the case, it is important to note that regulatory frameworks are in place to mitigate these risks. The metrics (Table 2) allow a broad categorization of these options, which appear to group into four categories. We propose a distinction where the proportion of medium to high availability exposure pathways exceeds 1/6 of the total number of pathways across all receptors. For this screening analysis, undesirable options were judged to be those where the proportion of medium to high availability exposure pathways exceeded 1/3 of the total. The categorization is broadly consistent with the desirability of, first, waste process options, then controlled disposal, and finally uncontrolled disposal. The semiquantitative exposure assessment allows consideration of the distribution of potential exposures along the process chains of the options considered. There is a tradeoff between process chain length and overall process control, which usually occurs off-site. A selective analysis of this for five options is presented in Tables 3 and 4, which illustrate that greater opportunities for exposure to hazardous agents occur in the early stages of the process chain, reinforcing the need for protective procedures throughout, but particularly at the outset of the management of carcasses. This disaggregation of exposure supports the qualitative analysis in Figure 2.
TABLE 2. Semiquantitative Exposure Assessment Metricsa
option and process chain This is a preferred option
composting (in vessel) alkaline hydrolysis gasification biogas atmospheric rendering pressure rendering incineration (generic) co-combustion (cement) mobile incinerator composting (windrow) air curtain incineration landfill
This is a secondary option
mass burial
This is a possible option mass pyre but not recommended on-farm pyre
anaerobic digestion (pits) on-farm burial do nothing
This is a generally undesirable option
medium and medium and high availability high availability pathways pathways environacross all across all human animal ment total receptors (n) receptors (%)
metric no. of pathways pathway scoresb no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores
142 237 149 244 149 245 157 252 157 256 157 256 149 246 122 210 127 198
135 220 142 235 142 235 149 242 149 242 149 242 142 235 115 202 120 186
17 38 19 44 19 44 21 50 21 52 21 52 19 45 17 44 12 28
294 495 310 523 310 524 327 544 327 550 327 550 310 526 254 456 259 412
33
11
44
14
44
14
45
14
45
14
45
14
44
14
44
17
44
17
no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores
163 284 163 323 109 169 109 190
156 267 154 307 103 171 103 192
17 40 18 51 12 28 12 30
336 591 335 681 224 368 224 412
62
18
62
19
46
21
46
21
no. of pathways pathway scores no. of pathways pathway scores
164 341 164 358
154 324 154 332
18 57 18 58
336 722 336 748
90
27
96
29
no. of pathways pathway scores no. of pathways pathway scores no. of pathways pathway scores
101 214 101 214 36 90
95 203 95 203 34 92
10 30 10 30 4 14
206 447 206 447 74 196
73
35
73
35
41
55
a The data presented herein should be interpreted to apply to the relative risks associated with each of the disposal options. They do not indicate that significant risks will actually be associated with these options should they be used in practice. Appropriate regulatory frameworks provide the means for assessing and managing site-specific risks. b Pathway scores were generated by summing the probability scores (0-4) assigned to each complete pathway for each disposal option.
TABLE 3. Distribution of Potential Exposure Pathways by Stage in the Process Chain process stage
1
2
3
4
5
description
disinfection/collection
transport
preprocessing
processing
residuals
disposal option incineration rendering landfill on-farm pyres do nothing a
no. of pathways
no. of pathways
no. of pathways
no. of pathways
no. of pathways
total
significanta
total
significant
total
significant
total
significant
total
significant
74 74 74 74 74
38 38 38 38 41
74 74 74 74 0
2 2 2 7 0
89 89 0 75 0
4 4 0 33 0
17 32 76 56 0
3 3 6 13 0
56 58 0 57 0
0 2 0 8 0
Pathways assessed to have a medium or high likelihood of being complete.
Care is required in interpreting these data and only generalized trends can be used to infer conclusions. Rankings used to assess the relative availability of exposure pathways were made using a linear scale (0, 1, 2, 3, 4) and elicited in the expert workshop. This said, the metrics for the semi-
quantitative exposure assessment (Tables 2-4) appear to provide a sound basis for assessing process chains in support of decisions on disposal such as those promoted within the 2006 international Terrestrial Animal Health Code (33), for example. With respect to identifying key risk drivers, this VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3151
TABLE 4. Risk-Based Carcass Disposal Hierarchy in the Event of an Avian Influenza H5N1 Outbreaka disposal optiona
statement of preference
alkaline hydrolysis; atmospheric rendering; biogas; composting (in vessel); controlled incineration; co-combustion in cement kilns; gasification; mobile incineration; pressure rendering
these are preferred options
air curtain incinerator; composting (windrow); mass burial; permitted landfill
these are secondary options
mass pyre; on-farm pyre
these are possible options, but not recommended
anaerobic digestion (pits), on farm burial, do nothing
these are generally undesirable options
a Options listed within each of the four categories are not presented in any order of preference.
analysis reveals little difference between the risks arising from individual waste technologies when exposures from their full process chains are considered. For differences that do exist, waste management options that offer engineered process control, containment, and high levels of destruction are preferred over open, on-site treatment because they destroy hazardous agents (e.g., avian influenza viruses and other pathogens are sensitive to heat), contain hazards within engineered systems, and minimize residual risk. The distribution of opportunities for exposure by process stage is more important (Table 3), as is the proportion of medium and high availability pathways across the process chain. Even where process chains are short, the observation that these options harbor a high proportion of pathways readily available to hazardous agents (Table 2; on-farm burial; do-nothing) means they pose significant opportunities for exposure. The greatest opportunities (Table 2) are where there is poor or no process control (Table 3). Where this is the case, exposures are not being actively managed. Therefore, the longer carcasses remain in this process stage unmanaged, the greater the likelihood for exposure and thus the greater the potential risk. Our analysis supports this implicit, but hereto unsubstantiated, conclusion providing important credence to the necessity and priority for early operational controls during outbreaks. The analysis is now informing (i) the annual review of the department’s contingency plans; (ii) operational protocols on the disposal of carcasses and their associated waste streams during outbreaks (34); (iii) stakeholder perceptions of the basis for the GB disposal hierarchy; and (iv) future policy options. It is valuable to compare this analysis with previous studies and risk assessments. Government bodies in the UK have commissioned generalized evaluations of risk during both the BSE and FMD crises. Quantitative risk assessments (35) have made use of generalized event trees with attending conditional probabilities to estimate the environmental fate of BSE prions in the environment; specifically with the aim of informing the targeted environmental regulation of industrial premises processing meat. These analyses have been powerful in providing detailed sensitivity analysis, but hugely resource intensive in terms of time and cost for a single hazardous agent to achieve these ends. Arguably, quantitative, probabilistic event-tree analysis for multiple 3152
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008
public health and environmental agents is not supportable for effective decision-making in a rapidly developing policy climate. During the 2001 FMD crisis, a qualitative risk analysis of hazards to public health and the environment was undertaken to support an independent evaluation of disposal options. Here (25), qualitative reasoning using the familiar “source-pathway-receptor” approach to risk analysis (36), supported by scientific judgments on hazard potency, pathway availability, and receptor sensitivity, allowed the scrutiny of the existing disposal hierarchy. The analysis presented in this paper seeks to strike a middle ground between the two historic approaches, offering a supportable level of analysis suited to a generalized risk assessment of disposal options, a hazard screen for multiple agents, and a process-chain approach that presents a more complete conceptual model. Other researchers have also successfully tracked the distribution of risk along the length of process chains (as presented in Table 3), most notably along processing lines within the food manufacturing industry, so as to inform hazard analysis critical control point (HACCP) protocols for microbiological risk management (37). The pragmatism advanced here is not without uncertainty however, and key issues for further consideration remain the appropriateness of the linear ranking of pathway availability, the generalized representations of each process chain, and the formalization of elicited judgments of barrier effectiveness. Effective mechanisms for addressing these issues without resorting to complex, lengthy evaluations are being sought. This analysis has not been constrained by the range of important considerations that biosecurity experts must also consider in devising contingency measures. We sought deliberately to free the exposure analysis of these factors in line with Government guidance (26) prior to their broader consideration. They do, however, merit discussion. Issues of practicality, economics, social implications, public acceptability, industrial capacity, and the legality of options all impact the workability of a contingency plan. Certain waste processing options are constrained because of legality or capacity, for example. The composting of carcasses killed for disease control purposes, for example, is currently not allowed under EU legislation. Similarly, irrespective of its potential utility, GB has very little biogas or alkaline hydrolysis capacity at present. Practitioners will also need to consider the distance of a disposal facility from a particular outbreak, and it is possible that although there may be “spare” capacity nationally, that the risks of transporting carcasses for several hundred kilometers may not be justified. The cost of certain technologies may be considered prohibitively expensive or disproportionate to the risk. Such issues will need to be considered by those responsible for carcass disposal policy when formulating their plans. Animal Health have subsequently taken the results of this exposure analysis and applied these additional constraints to the list of 28 disposal options. This additional analysis has, as indicated above, ruled out all forms of composting (on the basis of legality) as well as mobile incineration, biogas, alkaline hydrolysis, gasification, and pyrolysis (on the basis of capacity) and also ruled out mass burning and “do nothing” on the basis of risk and public perception. These results are not presented here since they will only be relevant to the particular circumstances in GB at the time the analysis was undertaken. To our knowledge, this is the first exposure assessment of this kind. This generic analysis provides credence to the existing GB disposal hierarchy (Figure 1) and the intuitive, though not previously justified, rationale placing waste processing (e.g., incineration and rendering) above controlled disposal (e.g., landfill), above uncontrolled disposal (e.g., burial on-farm or pyres). Beyond this are issues of practicality,
economics, social implications, public acceptability, industrial capacity, and legality of these options. We conclude the following: (1) The disposal hierarchy presented in the Defra contingency plan is supported by the hierarchy generated in this assessment. The hierarchy places waste processing (e.g., incineration and rendering) above controlled disposal (e.g., landfill) which is above uncontrolled disposal (e.g., burial on-farm or pyres). (2) A number of technologies not mentioned in the Defra contingency plan are potentially suitable on risk grounds for use in the event of an outbreak of avian influenza. Some of these (e.g., alkaline hydrolysis and gasification) are preferred over options in the existing hierarchy (even though they are currently ruled out due to lack of available capacity). (3) Carcass disposal comprises a series of stages (a process chain), with each stage offering opportunities for exposure to hazards. Typically, the process chains for disposal options offer hundreds of individual exposure pathways for hazards, although the majority are not likely to be of concern. (4) The range of hazards of concern extends well beyond the avian influenza H5N1 agent itself, because of the other pathogens that may be present, the chemicals used to disinfect carcasses, the veterinary products within birds, and the byproducts of carcass decomposition and waste processing. HPAI is not the only hazard of concern, neither is it the most significant hazard of concern in these disposals. Other pathogens, chemical agents, and carcass byproduct may dominate exposures. (5) This analysis indicates that the early stages of the process chain, on-farm, pose more opportunities for exposure to hazards than the later stages where hazards are generally contained, wastes are being treated, and residues are managed within regulated processes (Tables 2 and 3). This is a key consideration and evidence for this assertion is provided by the semiquantitative analysis (Tables 2-4). The potential risks of onward infection through fomite spread (by front-line workers, on the feet of site visitors, vehicles, etc.) must be considered and risk management measures must be put in place as routine. (6) For the process chains associated with disposal options, risks are reduced where hazards are contained rather than left in an open environment; technologies ensure the destruction of hazards rather than partial treatment; technologies have engineered process control rather than systems open to the environment; residues are minimized; and process chain are short rather then extended (in time or distance). (7) The approach represents an important advance in perspective for carcass disposal, away from one historically concerned with which waste technologies should be adopted for disposal (process mindset), toward one driven by where exposures are likely to dominate (risk-management mindset), thus providing a risk-informed basis for contingency planning and operational intervention.
Acknowledgments We thank Animal Health and Defra for funding this work and the workshop participants from Animal Health, Defra, the Health Protection Agency, the Environment Agency, the Scottish Environment Protection Agency, the Welsh Assembly, the Scottish Executive, and the Department of Health. Jens Evans, Nina Sweet, and Mark Okuniewski (Environment Agency) provided valuable critique, and Professor Roy Harrison (University of Birmingham) undertook an independent peer review prior to submission. Rupert Hough was funded on a BBSRC postdoctoral research fellowship (BB/ B512432/1) held at Cranfield University, UK.
Literature Cited (1) Rappole, J. H.; Hublek, Z. Birds and influenza H5N1 virus movement to and within North America. Emerging Infect. Dis. 2006, 12, 1486–1492. (2) Lahariya, C.; Sharma, A. K.; Pradhan, S. K. Avian flu and possible human pandemic. Indian Pediatr. 2006, 43, 317–325. (3) Woods, C. R.; Abramson, J. S. The next influenza pandemic: will we be ready to care for our children? J. Pediatr. 2005, 147–155. (4) Department for Environment, Food and Rural Affairs. Exotic Animal Disease Generic Contingency Plan Covering Foot and Mouth Disease, Avian Influenza, Newcastle Disease & Classical Swine Fever; Department for Environment, Food and Rural Affairs: London, UK, 2006. (5) United States Department of Homeland Security. Pandemic Influenza Preparedness, Response and Recovery Guide for Critical Infrastructure and Key Resources; U.S. Department of Homeland Security: Washington DC, 2006, pp. 79. (6) McQuiston, J. H.; Garber, L. P.; Porter-Spalding, B. A.; Hahn, J. W.; Pierson, F. W.; Wainwright, S. H.; Senne, D. A; Brignole, T. J.; Akey, B. L.; Holt, T. J. Evaluation of risk factors for the spread of low pathogenicity H7N2 avian influenza virus among commercial poultry farms. J. Am. Vet. Med. Assoc. 2005, 226, 767–772. (7) Scudamore, J. M.; Trevelyan, G. M.; Tas, M. V.; Varley, E. M.; Hickman, G. A. W. Carcass disposal: lessons from Great Britain following the foot and mouth disease outbreaks of 2001. Rev. Sci. Tech. Off. Int. Epiz. 2002, 775–787. (8) De Klerk, P. F. Carcass disposal: lessons from the Netherlands after the foot and mouth disease outbreak of 2001. Rev. Sci. Tech. Oie. 2002, 21, 789–796. (9) Dragon, D. C.; Elkin, B. T. An overview of early anthrax outbreaks in northern Canada. Arctic 2001, 54, 32–40. (10) Fichtner, G. In Proc. 88th Annual Meeting of the United States Animal Health Association; U.S. Animal Health Association: Richmond, VA, 1984, pp 415–429. (11) Hickman, G.; Hughes, N. Carcase disposal: a major problem of the 2001 FMD outbreak. State Vet. J. 2002, 12, 27–32. (12) Annells, D. H.; Williams, G. H.; Coulton, R.; McKelvey, P.; Tas, M. FMD burial site leachate disposal - a technical and commercial appraisal. In Proc. Aqua Enviro. Management of Wastewater Conference, 15–17 April 2002, Edinburgh, Scotland, 2002; Part 2, pp 245–258. (13) World Health Organization. Avian influenza. http://www. who.int/csr/disease/avian_influenza/en/ (accessed 17th January, 2007). (14) Province of British Columbia. Backgrounder. Frequently asked questions - avian influenza disposal; Ministry of Agriculture, Food and Fisheries, 2004, pp. 5. (15) Centres for Disease Control and Prevention. Questions and answers about avian influenza (bird flu) and avian influenza A (H5N1) virus; http://www.cdc.gov/flu/avian/gen-info/qa.htm (accessed 26th March 2007). (16) Organisation Modiale de la Santé Animale (World Organisation for Animal Health). Update on avian influenza in animals (Type H5); http://www.oie.int/downld/AVIAN%20INFLUENZA/A_ AI-Asia.htm (accessed 26th March, 2007). (17) Department for Environment, Food and Rural Affairs. Avian influenza (bird flu) latest situation: H5N1 Avian Influenza in poultry,UpperHolton,Suffolk;http://www.defra.gov.uk/animalh/ diseases/notifiable/disease/ai/latest-situation/index.htm (accessed 26th March, 2007). (18) Ellis, D.B. Carcass disposal issues in recent disasters, accepted methods and suggested plan to mitigate future events. MBA thesis, Texas State University, 2001, pp. 128. (19) Morgan, A. Avian influenza: an agricultural perspective. J. Infect. Dis. 2006, 194, S139–146. (20) National Agricultural Biosecurity Center. Carcass disposal: a comprehensive review, Report prepared for the United States Department of Agriculture Animal and Plant Health Inspection Service, National Agricultural Biosecurity Center; Kansas State University, Manhattan, KS, 2004. (21) The BSE Inquiry. The inquiry into BSE and variant CJD in the United Kingdom, Lessons learnt; Volume 1, Chapter 14, BSE Inquiry, 2000, pp 249–274. (22) Environment Agency. Environmental impact of foot and mouth disease: an interim assessment; Environment Agency: Bristol, 2001, pp. 44. (23) Department for Environment, Food and Rural Affairs (Defra). Defra’s overview of emergency preparedness for exotic animal diseases (supplementing the Defra framework response plan for exotic animal diseases); Presented to Parliament pursuant to VOL. 42, NO. 9, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3153
(24)
(25)
(26)
(27)
(28)
(29) (30)
3154
s14A of the Animal Health Act, 2002, v1.0; Defra: London, December 2006, 55 pp;available at http://www.defra.gov.uk/ animalh/diseases/control/contingency/exotic.htm. Department for Environment, Food and Rural Affairs (Defra). Defra’s framework response plan for exotic animal diseases (for foot and mouth disease, avian influenza, Newcastle disease, classical swine fever, African swine fever and swine vesicular disease); Presented to Parliament pursuant to s14A of the Animal Health Act, 2002, v1.0; Defra: London, December 2006, 128 pp;available at http://www.defra.gov.uk/animalh/diseases/ control/contingency/exotic.htm. Department of Health. A Rapid Qualitative Assessment of Possible Risks to Public Health from Current Foot & Mouth Disposal Options. Main Report June 2001; Department of Health: London, 2001. Department of the Environment, Transport and the Regions (DETR), Environment Agency and Institute for Environment and Health. Guidelines for environmental risk assessment and management; The Stationary Office: London, 2000, pp. 88. Department for Environment, Food and Rural Affairs (Defra). Risk management strategy; Department for Environment, Food and Rural Affairs: London, 2002, pp. 29. Pollard, S. J.; Yearsley, R.; Reynard, N.; Meadowcroft, I. C.; DuarteDavidson, R.; Duerden, S. L. Current directions in the practice of environmental risk assessment in the United Kingdom. Environ. Sci. Technol. 2002, 36, 530–538. Environment Agency of England and Wales. Avian influenza reference guide v1.0; Environment Agency: Bristol, UK, 42. pp. Department for Environment, Food and Rural Affairs (Defra). Animal health and welfare: Veterinary surveillance: Great Britain poultry register; 2007; available at http://defraweb/
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 9, 2008
(31)
(32) (33) (34)
(35)
(36) (37)
animalh/diseases/vetsurveillance/poultry/index.htm (accessed 17th January 2007). Department for Environment, Food and Rural Affairs (Defra). Density of poultry and premises registered in the GB poultry register; Surveillance, Zoonoses & Emerging Issues Division, Defra: London, 11 September 2006, pp. 9. Statutory Instrument No. 2701. The Avian Influenza (Preventative Measures) (England) Regulations; 2006;ISBN 0 11 075187 6; pp. 16. World Organization for Animal Health (OIE). Terrestrial Animal Health Code (2006); World organization for animal health: Paris, 2006, pp. 9. Department for Environment, Food and Rural Affairs (Defra). Protocol for the use of ABPR approved, PPC permitted rendering plants for the disposal of animal carcasses and associated wastes during an exostic animal disease outbreak, version 1.4; Surveillance, Zoonoses & Emerging Issues Division, Defra: London, 2007, pp. 41. Environment Agency. Overview of risks from BSE via environmental pathways; Report C7243 prepared by DNV Technica for the Environment Agency; Environment Agency: Bristol, 1997; with Appendices I-VI. Pollard, S. J. T.; Smith, R.; Longhurst, P. J.; Eduljee, G. H.; Hall, D. Recent developments in the application of risk analysis to waste technologies. Environ. Int. 2006, 32, 1010–1020. Arvanitoyannis, I. S.; Varzakas, T. H. Application of failure mode and effect analysis (FMEA), cause and effect analysis and Pareto diagram in conjunction with HACCP to a potato chips manufacturing plant. Int. J. Food Sci. Technol. 2007, 42, 1424–1442.
ES702918D
