Risk of Importing Zoonotic Diseases through Wildlife Trade, United States

The United States is the world's largest wildlife importer, and imported wild animals represent a potential source of zoonotic pathogens. Using data on mammals imported during 2000-2005, we assessed their potential to host 27 selected risk zoonoses and created a risk assessment that could inform policy making for wildlife importation and zoonotic disease surveillance. A total of 246,772 mammals in 190 genera (68 families) were imported. The most widespread agents of risk zoonoses were rabies virus (in 78 genera of mammals), Bacillus anthracis (57), Mycobacterium tuberculosis complex (48), Echinococcus spp. (41), and Leptospira spp. (35). Genera capable of harboring the greatest number of risk zoonoses were Canis and Felis (14 each), Rattus (13), Equus (11), and Macaca and Lepus (10 each). These findings demonstrate the myriad opportunities for zoonotic pathogens to be imported and suggest that, to ensure public safety, immediate proactive changes are needed at multiple levels.

M ost emerging infectious diseases are caused by zoonotic pathogens (1,2). The number and proportion of these diseases that originate in wild animals in particular has increased substantially in the past few decades, even after accounting for increased reports of new emerging infectious diseases (1). This trend and recent pandemics of wildlife-origin infectious diseases (e.g., HIV, severe acute respiratory syndrome) suggest that targeted surveillance efforts should focus on activities that bring humans and wildlife in close contact (1,3).
The United States is among the world's largest importers of live wild animals (4) and imported >1 billion indi-vidual animals during 2000-2004 (5). Little disease surveillance is conducted for imported animals; quarantine is required for only wild birds, primates, and some ungulates arriving in the United States, and mandatory testing exists for only a few diseases (psittacosis, foot and mouth disease, Newcastle disease, avian infl uenza). Other animals are typically only screened for physical signs of disease, and pathogen testing is delegated to either the US Department of Agriculture (for livestock) or the importer (6). The process of preimport housing and importation often involves keeping animals at high density and in unnatural groupings of species, providing opportunities for crossspecies transmission and amplifi cation of known and unknown pathogens. Thus, imported wildlife remain a major public health threat, as exemplifi ed by the importation of Ebola virus in primates from the Philippines (7), monkeypox from imported African rodents (8), and possibly HIV from chimpanzees in central Africa (9). Wildlife importation also poses a great threat to domestic wildlife and the US agriculture industry (5).
To analyze the volume and diversity of live mammals that have been imported into the United States in recent years, we used data from the US Fish and Wildlife Service Law Enforcement Management Information System. We focused on mammals because of the frequency and severity of previously reported mammal-borne zoonoses and because of the frequent close association between humans and many mammalian species (e.g., as pets). We then assessed the zoonotic diseases that imported mammals are known to host. Our results may be used to inform policy decisions about wildlife importation and zoonotic disease surveillance and may alert clinicians to the broad range of possible zoonoses that may be encountered in patients who have been exposed to imported animals. the United States, or it must have the potential for new epidemiology with regard to transmission (e.g., Yersinia pestis is presently found in wild rodents in the western United States, but it is not expected to be found in animals sold as pets); and 5) if the pathogen uses an intermediate vector, competent vectors must exist in the United States. The resulting list comprised 30 risk zoonoses (20 viral diseases, 9 bacterial diseases, and 1 helminthic disease); no fungal, protozoal, or prion diseases were on the list, and thus they were not analyzed.
Determination of the host range of the risk zoonoses was accomplished through systematic genus-driven and pathogen-driven searches of PubMed databases (www. pubmed.gov), the Google search engine (www.google. com), and references within published works. Confi rmed presence was defi ned as either isolation of the pathogen from an animal or serologic evidence of past infection. For all animals identifi ed in the literature as carrying a risk zoonosis, genus and family were recorded. The host ranges of all of the risk zoonoses were then cross-referenced against the imported genera to generate tables showing diseases found in each imported genus (affected genera). If the disease was found in a different genus within the same family, this was also noted (potentially affected genera). The justifi cation for this expanded risk assessment is the host nonspecifi city of many infectious diseases; lack of evidence for the presence of a given disease in a given host should not be construed as evidence against its presence.

Results
During 2000-2005, a total of 4,067 shipment fractions of mammals were imported (a shipment fraction is the sum of all animals of a single species in a given shipment; a single shipment may contain several shipment fractions), totaling 246,772 individual mammals and representing 190 genera and 68 families. The average number of animals per shipment fraction was 61 (range 1-8,000). The most common declared purpose for importation was commercial use (not classifi ed according to pet trade, food, traditional medicine, etc.), accounting for 66% (163, (305). These values must be interpreted cautiously, however, because many animals are imported and then reexported; thus, their true origin may become obscured. For example, a "wild-caught" chinchilla with a "country of origin" of Czech Republic must have originated elsewhere because chinchillas are native to Chile. A comparison between the natural geographic range of all wild-caught animals and their stated countries of origin showed that >25% of the pairings were impossible (i.e., the animals could not have come from their stated country of origin). This limitation is inherent in the way US Fish and Wildlife Service Law Enforcement Management Information System data are collected, and we were unable to correct these data.
The source of the animals was largely uninterpretable because 49% of all individuals were declared as being sourced from "animal derivatives and parts," despite the fact that we had selected only live animals for our analysis, and despite the fact that "animal derivatives and parts" is not one of the permitted responses to this question. Another 29% were declared as "captive-bred" and 15% as "wild-caught." For the fi nal list of risk zoonoses, 3 of the original 30 agents (Hendra virus, Menangle virus, and Rickettsia prowazekii) were removed because few, if any, genera were found to harbor these infections; the fi nal tables therefore include 27 diseases (Tables 1-3). The risk zoonoses capable of infecting the greatest number of genera were: rabies viruses, in 78 genera; Bacillus anthracis, the causative agent of anthrax, in 57 genera; Mycobacterium tuberculosis complex, in 48 genera; Echinococcus spp., the agents of hydatid cyst disease, in 41 genera; Leptospira spp., in 35 genera; Brucella spp., the agents of undulant fever, in 32 genera; Francisella tularensis, the agent of tularemia, in 31 genera; Crimean-Congo hemorrhagic fever virus, in 27 genera; Y. pestis, the agent of plague, in 24 genera; and Coxiella burnetii, the agent of Q fever, in 20 genera (Table  2; online Technical Appendix, available from www.cdc. gov/EID/content/15/11/1721-Techapp.pdf).
If each genus within affected families is counted as potentially capable of harboring a risk zoonosis (according to the principle that many diseases are not entirely host specifi c), the number of genera potentially capable of harboring rabies viruses rises to 155 (82% of all imported taxa); potential carriers of Leptospira spp. increase to 131; M. tuberculosis complex to 124; F. tularensis to 115; B. anthracis to 113; C. burnetii to 108; and Y. pestis to 101.

Discussion
Our data demonstrate that myriad opportunities exist for key zoonotic pathogens to be imported into the United States or, if already present, to be introduced in a new context (e.g., in an animal sold as a pet). Imported animals of a large number of taxa were found to be capable of carrying risk zoonoses; these diseases include such serious public health threats as rabies, the fi lovirus hemorrhagic fevers, tuberculosis, and highly pathogenic avian infl uenza.
This study likely underestimates the broad nature of risk associated with the importation of wild animals. We examined only families in the class Mammalia that have been shown to harbor risk zoonoses; however, many pathogens routinely cross boundaries at least as high as the class level (e.g., human psittacosis from birds), if not higher. Furthermore, we included only live animals in this analysis; recent outbreaks associated with animal products (e.g., cutaneous anthrax from an imported goat hide used for making drums) attest to the risks associated even with dead animals (13). Finally, the study can neither estimate the risk for unknown pathogens, which may be imported but not yet identifi ed, nor assess the volume and zoonotic risk created by illegal wildlife trade. Animals may be smuggled specifi cally because they have been banned from trade as a result of perceived or recognized health threats. Some animals on our list of risk zoonoses have already been banned from importation (e.g., masked palm civets, birds from countries affected by highly pathogenic avian infl uenza [H5N1]) (14). However, pathogens have been identifi ed in illegally imported wildlife; e.g., a pair of crested hawkeagles (Spizaetus nipalensis) smuggled from Thailand and recently confi scated in Belgium were infected with highly pathogenic avian infl uenza (H5N1) (15).
We did not quantitatively assess the risk for transmission of each pathogen at each importation event. Rather, we attempted to demonstrate the breadth of risk associated with importations of wild animals in general. Quantitative prevalence of the various pathogens in each wildlife host is highly variable, and determining it is beyond the scope of our analysis. Some genera represent the primary reservoirs of certain pathogens (e.g., Peromyscus for certain hantaviruses), whereas proof of the permissiveness of other genera to certain pathogens is limited to isolated case reports (e.g., Ebola Zaire virus in the duiker Cephalophus). Perhaps the greatest unknown associated with quantifying risks for each of the zoonoses is a pathogen's infectivity in various hosts. Some pathogens may increase to a high enough load in their hosts to be infectious; others may cause nothing more than a measurable serologic response in what is otherwise a dead-end host (though explicitly known dead-end hosts have been excluded from these analyses).
Our analysis highlights several ways that the US Fish and Wildlife Service could improve data collection. To enhance public health offi cials' ability to trace back the sources of imported zoonotic diseases, record keeping of the point of origin of shipments could be expanded to include not just their most recent and previous point of origin (as is currently done with the "Country of Origin" and "Country of Importation/Exportation/Re-importation") but  also their actual origin. Accurate recording of the source of the animals (e.g., wild-caught, captive-bred) is also needed. Our results showed that half of all individuals had a declared source that was not one of the allowed choices (e.g., wild-caught, captive-bred). The source of an animal affects not only the likely level of risk (i.e., one would expect captive-bred individuals to carry fewer zoonotic diseases than wild-caught individuals) but also mitigation strategies when zoonotic diseases are identifi ed (e.g., euthanizing a colony vs. improving quarantine after capture).
The potential for importation of zoonoses that would pose a major public health threat suggests that increased surveillance should be applied to imported wildlife in the United States. One opportunity to reduce this threat is restriction of importation of key high-risk species, as was done when the Centers for Disease Control and Prevention used emergency powers to restrict importation of Gambian pouched rats during the monkeypox outbreak (14). Given the great diversity of animals identifi ed by our analysis as potentially hazardous, broad importation bans would likely be necessary if the goal were to substantially decrease the overall risk. Political or social support may be limited for such broad bans, both in the United States (as one of the world's largest purchasers of wildlife) and abroad (where wildlife trade can have profound economic benefi ts).
Furthermore, illegalizing trade may only increase underground (illicit) trade, thereby eliminating the possibility of screening shipments for potential hazards. A more effective and acceptable strategy would be enhancing surveillance for the specifi c pathogens noted for the key risk genera (those harboring the greatest number of risk zoonoses, i.e., Canis, Felis, Rattus, Equus, Macaca, Lepus, Ovis, and Vulpes). Notably, the numbers of shipments of mammals is low relative to other wildlife groups (e.g., fi sh and reptiles). Lawmakers' interests in protecting our borders from external bioterrorism threats intersect with the need to protect ourselves from zoonotic diseases; many Category A bioterrorism threats (e.g., anthrax, plague, tularemia, and the viral hemorrhagic fevers) (10) are represented in the risk zoonoses outlined above. Finally, to facilitate the standardization of surveillance and detection of infection events, the Council of State and Territorial Epidemiologists should include all of the risk zoonoses among their states' notifi able diseases (most of which are already included).
Perhaps one of the simplest practical interventions for minimizing zoonotic disease risk is reduction of opportunities for transmission from wildlife to humans. Although a large proportion of imported animals are destined for biomedical research (in which potential occupational risks are largely understood and quarantine procedures likely mitigate risk), a greater proportion (even among the high-risk genera) are destined for commercial use and therefore could expose a wider group of persons to zoonotic diseases. Education of professionals likely to come in close contact with imported animals (e.g., veterinarians, importers, pet store employees), as well as the general public, should emphasize the risks for contracting zoonotic diseases from wildlife and pets (16) and the need for proper hygiene, safety procedures, and personal protective equipment (17).
The recommendations above mirror others that exist in policy documents by the Defenders of Wildlife (5)

Felis
Cat Felidae X X X X X X X X o X X X X X X 14 15