Live wildlife trade in markets - a scoping review to inform risk assessment of emerging infectious diseases

Wet markets are important for food security in many regions worldwide but have come under scrutiny due to their potential role in the emergence of infectious diseases. The sale of live wildlife has been highlighted as a particular risk, and the World Health Organisation has called for the banning of live, wild-caught mammalian species in markets unless risk assessment and effective regulations are in place. Following PRISMA guidelines, we conducted a global scoping review of peer-reviewed information about the sale of live, terrestrial wildlife in markets that are likely to sell fresh food, and collated data about the characteristics of such markets, activities involving live wildlife, the species sold, their purpose, and animal, human, and environmental health risks that were identified. Of the 59 peer-reviewed records within scope, only 25% (n = 14) focussed on disease risks; the rest focused on the impact of wildlife sale on conservation. Although there were some global patterns (for example, the types of markets and purpose of sale of wildlife), there was wide diversity and huge epistemic uncertainty in all aspects associated with live, terrestrial wildlife sale in markets such that the feasibility of accurate assessment of the risk of emerging infectious disease associated with live wildlife trade in markets is limited. Given the value of both wet markets and wildlife trade and the need to support food affordability and accessibility, conservation, public health, and the social and economic aspects of livelihoods of often vulnerable people, there are major information gaps that need to be addressed to develop evidence-based policy in this environment. This review identifies these gaps and provides a foundation from which information for risk assessments can be collected.


1
Data that were extracted about markets included the type of market (for example, wildlife or 2 0 5 food), and the physical structure, frequency, and types of people and activities at the markets. Hygiene practices and temporal changes (for example, increased sale of live animal species) 2 0 7 were also recorded. Data that were extracted about animals at the markets included the species of live animals sold at the market were also extracted, as well as the total volume of all wildlife and the 2 1 2 purposes of wildlife sale. Any non-wildlife species sold at the market were recorded, as well 2 1 3 as contacts that were documented between wildlife species and any other live animal species. Data that were extracted about risks included specific pathogens and diseases, measures of 2 1 5 their frequency and how they were identified, and any other risks associated with human and 2 1 6 animal health. Due to the longer form at Level 3, four additional reviewers were included, and agreement 2 1 8 amongst all reviewers was initially assessed with three records (selected based on diversity of 2 1 9 study type). Modifications were made to the form following reviewer feedback. The form 2 2 0 was then tested by all reviewers with two more records to assess agreement. All conflicts 2 2 1 were discussed. Once consensus was achieved for all five records and further modifications 2 2 2 to the Level 3 form were made to improve clarity and increase agreement, the full Level 3 2 2 3 review was conducted. Conflicting opinions about extracted data were discussed between 2 2 4 each record's pair of reviewers and a third reviewer to determine a consensus prior to 2 2 5 synthesis of the extracted data. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;

2
Verification of the search strategy was undertaken by screening the bibliography list of three 2 2 8 records that were retained for data analysis in Level 3 (one for each decade of the search 2 2 9 window). Records which were potentially relevant to the study were included in the review 2 3 0 process using the same methods as records identified in the initial search. Topics of interest were tabulated according to region, and information about each topic was 2 3 3 then summarized and presented as a narrative. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;    (1980−2020) associated with live, terrestrial, vertebrate wildlife sold for any purpose at markets likely to sell 2 5 0 food. Of 56 records included in Level 3, all were single studies. A spreadsheet of the Level 3 2 5 4 studies and the data charted for each study is included in the Supporting Information. One record was sourced from a thesis (Edwards, 2012), and all other records were published 2 5 6 in 35 peer-reviewed journals (Supplementary Material 3: Figure S1). The most frequently 2 5 7

Results
represented journal was TRAFFIC Bulletin (20%; n = 11), followed by Biodiversity and frequently represented study region was Southeast Asia (n = 22; mode country Lao PDR, n = 2 6 0 6), followed by China (here, classified as a region due to its large geographic size and number 2 6 1 of studies; n = 17; mode province Guangdong, n = 7), Africa (n = 16; mode country Nigeria, 2 6 2 n = 4), and South America (n = 8; mode country Brazil, n = 6). Two studies were conducted 2 6 3 in more than one region: Europe and Africa (Spanish enclaves of Ceuta and Melilla, and Dianmo (1998)). publications in the early 2000s was frequently collected prior to 1997; therefore, research 2 7 0 intensity did not appear to be influenced by particular EID events ( Figure 3). The most common focus for studies was investigation of the wildlife trade in relation to 2 7 2 conservation concerns such as species depletion and loss of biodiversity (75%; n = 42). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;

5
Disease risk was the primary focus of study in only 14 studies (25%), and there was no 2 7 4 association between the focus of studies and global region (Χ 2 = 9.5, d.f. = 8, P = 0.3). wildlife sold for any purpose at markets likely to sell food. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;  The most frequent market types (according to author definitions and descriptions) were  Table 1. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 Market physical structure was described in 20 records (36%); a range of permanent and 2 9 4 temporary structures were reported, either indoor or outdoor (for example, on the sides of 2 9 5 roads or in car parks), and often in locations accessible for the arrival and distribution of 2 9 6 goods. For example, in Belém, Brazil, market trade takes advantage of riverside ports (da 2 9 7 Nobrega Alves and Pereira Filho, 2007), and in the Niger Delta, hunters deliver goods via 2 9 8 rivercraft to one of the largest bushmeat markets in the region which is also accessible via recently constructed roads (Akani et al., 2015). In central Lao PDR, authors reported that the 3 0 0 most active wildlife markets were located close to major roads which facilitated trade across  of live, terrestrial, vertebrate wildlife at markets likely to sell food globally. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint were a feature of some markets; for example, in Yunnan Province, China, wildlife trade could 3 1 0 occur at irregular times and sites, and sometimes, markets appeared to be the sites of 3 1 1 advertising and negotiation as vendors displayed their goods (including live wildlife in cages) 3 1 2 whilst transfer of goods occurred elsewhere (Haibin and Kunming, 1999).
The types of people involved in markets were vendors and consumers -also called traders and buyers -as well as hunters and trappers, middlemen (also termed intermediaries and healers and drivers. Middlemen could be involved in the chain of events before or after the 3 1 7 market between wildlife acquisition and the end-user. For example, some middlemen traded 3 1 8 between hunters and market vendors (these people could also be termed 'collectors' because  trade was also associated with families as younger generations inherited the family business 3 2 2 (for example, turtle collectors in Kerala, India; Krishnakumar et al. (2009)).

2 3
Illegal trade was reported in most records (68 %, n = 38), particularly if the focus was of illegal trade included the sale of wildlife species that were prohibited for sale by national 3 2 7 or international laws, sourced from protected areas such as National Parks (for example, in CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.13.21263377 doi: medRxiv preprint example, wildlife from Vietnam sold in Yunnan Province, China; Zhang et al. (2008)).

2
Sometimes, laws were open to misinterpretation because they varied according to animal  (2019)) or were inconsistent 3 3 5 between regional and national levels (for example, in Brazil; Fernandes-Ferreira et al. Lao (Schweikhard et al., 2019), and turtle collectors in Kerala, India who were aware that 3 3 9 trade was illegal, but were not deterred by small fines from local police and forest officials 3 4 0 (Krishnakumar et al., 2009). Government officials were reportedly sometimes corrupt and 3 4 1 therefore, not necessarily promoting regulatory activities (Haibin and Kunming, 1999).

4 2
Sometimes illegal trade was clandestine. In Lao PDR, Davenport and Heatwole (2013) 3 4 3 suspected that as live wildlife on display were sold, they were replenished from bags that wildlife and their products were concealed unless trade was likely (Williams et al., 2016).

4 6
Other than trade, activities directly associated with live wildlife included slaughter and . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.13.21263377 doi: medRxiv preprint indication of a large catchment area of wildlife as some have previously suggested). Others 3 5 7 also noted lack of refrigeration of fresh meat in studies in Asia (Pruvot et al., 2019;Shepherd 3 5 8 and Nijman, 2007). In an assessment of zoonotic disease risk in markets throughout Lao 3 5 9 PDR, (Greatorex et al., 2016) noted that hygiene practices such as hand-washing were rare, 3 6 0 sanitation was poor (butchers knives and tables were rarely cleaned; blood and entrails were 3 6 1 on the floor) and wildlife contact with fresh food was high in most markets. In a study from 3 6 2 Nigeria and Egypt in which biosecurity compliance was evaluated (Fasanmi et al., 2016), measures were poor at many of the markets. Temporal trends were mentioned in 18 records (32%), with changes in activity and species Nigeria (Akani et al., 2015) and in winter in Guangdong, China (Dong et al., 2007), and Island (Cronin et al., 2015;Fa et al., 2000), and decreased in Togo, Mozambique and Papua has increased at Mong La market in Myanmar (Shepherd and Nijman, 2007). Trade also CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint these, the most frequent orders were Testudines (37%, n = 197), followed by Passeriformes  Viverridae (large Indian, lesser Indian and masked palm civets) were reported in 4 records 3 9 1 from China (Chow et al., 2014;Dong et al., 2007;Haibin and Kunming, 1999;Yiming and 3 9 2 Dianmo, 1998). Pholidota (pangolins) were reported in 10 records from Africa (n = 5), 2007; Sodeinde and Soewu, 1999;Soewu and Ayodele, 2009;Yiming and Dianmo, 1998).

9 6
Primate species were observed in markets in Africa, China and Southeast Asia, including a 3 9 7 live chimpanzee in a bushmeat market in Liberia (Greengrass, 2016). Mustelidae species 3 9 8 were observed in markets in China (Dong et al., 2007), and squirrels (Sciuridae) were 3 9 9 observed in Southeast Asia and China (Pruvot et al., 2019;Yiming and Dianmo, 1998).  Table S2). Of these, wildlife of class 4 0 8 Aves were most frequently reported. Kong, Shenzhen and Guangzhou in southern China (Cheung and Dudgeon, 2006). Large Eid (2014)). However, nearly half of the records only reported one live wildlife species (48%, 4 1 8 n = 26; Figure S3). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  Wildlife (live or products) were mostly sold for food and medicine globally, as well as pets in 4 2 6 Southeast Asia and China, and spiritual purposes in South America and Africa ( Figure 6).

2 7
The reported volume of live wildlife and total wildlife products sold was dependent on the in some markets due to covert trading -information from records about volume of wildlife is . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. reported research (1980−2020) associated with the sale of live wildlife at markets likely to sell food globally.

5 0
Main plot: classes of wildlife that were identified by species. Inset plot: classes of wildlife that were identified 4 5 1 by common name, genus or order. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review) preprint
The copyright holder for this this version posted September 16, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 wildlife species were sampled (Dong et al., 2007). CoV was detected in several Asian leopard CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Zoonotic disease risks that were mentioned other than those associated with disease and  (Edmunds et al., 2011), and concern about zoonoses such as Ebola virus disease, rabies, and 4 9 0 tuberculosis associated with handling primates in Benin, Africa (Sogbohossou et al., 2018).

CC-BY 4.0 International license
It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Discussion
This review highlighted enormous global variability and epistemic uncertainty in all aspects regarding circumstances in which live, terrestrial wildlife are sold in markets that are likely to sell fresh food. Specific insights included the limited research focus on disease associated with live wildlife in markets, the wide variety of live wildlife species traded at markets especially in China and Southeast Asia, and the high accessibility of markets (both socially and geographically). These insights, as well as the variability and uncertainty related to the markets, the wildlife, and associated risks, greatly influence the feasibility of accurate assessment of the risk of emerging infectious disease associated with live wildlife trade in markets, and the development of effective, sustainable policy in this environment.
Most studies in this review focused on the impact of live wildlife trade in markets on conservation, particularly biodiversity loss. Whilst this can indirectly inform the scale of biodiversity loss as a broad driver of emerging infectious disease, focus on potential pathogens and disease is needed for risk analysis frameworks, which require information about specific hazards and exposure pathways. Many zoonotic and foodborne microorganisms have been identified in items other than live wildlife (including fresh meat, vegetables, and the environment) in wet markets worldwide, such as Leptospira spp., Vibrio spp., Toxoplasma spp., Campylobacter spp., and Salmonella spp. (Hamid et al., 2020;Kottawattage et al., 2017;Ngan et al., 2020;Sekoai et al., 2020), yet only three potentially zoonotic microorganisms (Spirometra species, Coronaviridae, and avian influenza virus) were reported from live wildlife in this review. There was also limited information about water, sanitation, and hygiene (WASH; https://www.who.int/health-topics/water-sanitationand-hygiene-wash, accessed 20 August 2021) conditions associated with live wildlife trade, such as butchering and slaughter of wildlife, contamination of other food products or the environment, or contact between wildlife and other species, including people. Although . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 market regulations are often aimed at reducing the risk of zoonotic and foodborne disease, and some countries -for example China, Indonesia, and Thailand -have specific WASH regulations applied to wet markets (The Law Library of Congress, 2020), there appears to be a gap in knowledge about the range of microorganisms and WASH risks associated with live wildlife. A greater understanding of current WASH conditions around activities associated with live wildlife trade (including market infrastructure such as hand washing facilities, adequate drainage, and separation of animals) is required to inform risk assessment exposure pathways. In addition, previous recommendations to increase surveillance of wildlife to identify potential pathogens and to detect and respond to emerging infectious disease (Grogan et al., 2014;Kuiken et al., 2005;Vrbova et al., 2010) appear to have been largely overlooked in the context of live wildlife in markets and should be adopted more comprehensively.
The greatest number of live wildlife species was reported from studies in markets in China and Southeast Asia. Wide species diversity might provide more opportunity for EID events due to a greater number of potential reservoir-spillover host pairs. However, following previous patterns of disease emergence, species from classes Mammalia and Aves have been considered a higher disease-risk due to their association with potentially pathogenic microorganisms (Olival et al., 2017;White and Razgour, 2020;Woolhouse et al., 2012), and a greater focus on these classes of wildlife might be warranted. Consistent with this, the World Health Organization have recommended suspension of trade of live-caught mammalian wildlife in traditional food markets (World Health Organisation, 2021). In this review, live mammalian wildlife species that could be considered high risk were identified in studies from markets in Africa, Southeast Asia, and China, and wild birds were also reported in markets from South America and the Middle East in addition to these regions. However, the likelihood of spillover will also depend on additional characteristics of the wildlifepotential spillover host interface, including its magnitude (amount of live wildlife traded and . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 contact with the potential spillover species) and factors that affect transmission, such as stressors that might influence the shedding of microorganisms (including travel, overcrowded housing, comorbidities and the physical condition of the wildlife), the type of contact occurring and the susceptibility of the potential spillover host (Plowright et al., 2017;Wikramanayake et al., 2021). The scale of interfaces could not be determined in this review, and although most studies and species were identified from China and Southeast Asia, this might not be representative of the global distribution of markets or volume of live wildlife sold. Ultimately, resources for mitigation strategies will be limited, and development of effective policy will require assessment of relative risk dependent on the diverse global contexts that were apparent in this review, such as market type (for example, markets in this review were predominantly bushmeat in Africa, wildlife in Southeast Asia, and food in China) and species sold, as well as local, market-specific factors.
The scope of policy to mitigate EID risks associated with trade of live wildlife in markets is also important to define. If the goal is to prevent dissemination of disease from markets and an epidemic -and not only to prevent spillover of microorganisms from live wildlife reservoir hosts -social and geographic pathways from markets also need to be considered in risk assessments. Short value-chains that support local producers and ensure a supply of fresh food at low cost -due to the limited requirement for transport and refrigeration -have been considered an advantage of wet markets (Goldman et al., 1999;Zhang and Pan, 2013).
However, several studies in this review described how markets were located to facilitate trade and distribution, and highlighted that wildlife trade is a high-value, global activity with long value chains and movement of wildlife across many regions. In 2017, the value of the global wildlife trade was an estimated USD23 billion (van Uhm and Wong, 2019). They also described a multitude of people associated with live wildlife trade, consistent with the typologies defined by Phelps et al. (2016). Long value-chains which involve many people . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.13.21263377 doi: medRxiv preprint might be specific features of markets selling live wildlife that could be related to the purpose and value of wildlife. Although quantitative methods can be used to characterise risks and their pathways within markets (such as species abundance and contact durations), social science and ethnographic methods, such as those described by Edwards (2012) in the context of markets in Mali, Africa, are likely to be of greater value to understand complex wildlife trading networks. Understanding the drivers and risks behind network structure is valuable to identify potential control points (Phelps et al., 2016), and will also serve to understand the socio-economic impacts of regulations to mitigate risks. Networks involving people and transport are often small-world networks and it is possible that wildlife trade networks are of this type; such networks can be robust in the face of disruption (Hu and Verma, 2011). Therefore, as well as not addressing the drivers of wildlife consumption, banning trade of live mammalian wildlife in markets might not be an effective strategy to mitigate EID risks. It has already been suggested that those whose livelihoods depend on wildlife trade are likely to find alternative trade routes that avoid regulations; trade diverted to hidden networks could inadvertently pose a greater risk to public health (Roe et al., 2020).
Overall, this review demonstrates that accurate assessment of the risk of EIDs associated with live wildlife trade in markets is currently infeasible due to massive epistemic uncertainty. The magnitude of live wildlife-potential spillover host interfaces is unknown, high-risk species and microorganism pairs are largely unknown, there is limited information about activities that could influence exposure to microorganisms, and there is limited evidence to associate EID events with wet markets. Although this review did not identify records in languages other than English, the small proportion of records which focused on disease indicates that assessing EID risks in markets has had less research focus than the conservation impacts of live wildlife trade. This does not mean that markets are not a risk for EID emergence and that a precautionary approach is unwarranted. Ultimately however, policy will need to be . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.13.21263377 doi: medRxiv preprint evidence-based and commensurate with risk so that it can be implemented and sustained. The high-profile association of early COVID-19 cases with a wet market in Wuhan and the current recommendation to ban trade of live mammalian species in markets has catalysed the need to understand these risks. This review provides a baseline of currently available, peerreviewed information about live wildlife trade in markets globally. Addressing the gaps in knowledge to accurately assess risks across the wide range of market types, activities and wildlife species purpose globally will provide the framework in which policy to promote biodiversity, protection of livelihoods and prevention of EIDs can be supported.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 16, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021