2.1. Literature Review

Additional searches were conducted by replacing the search terms in (i) with the 29 priority zoonoses identified in Vietnam [21] and viral families with high zoonotic potential (filo^∗ OR rhabdo^∗ OR corona^∗ OR flavi^∗ OR paramyxo^∗).

A first assessment of all records published between 1965 and August 2022 was conducted using their title and abstract (Figure 1). Selected publications were then screened using full-text documents. Records were eligible for data extraction if they were published in English and included data on pathogen screening of biological specimens collected from terrestrial wildlife species in Vietnam or exported from Vietnam. All terrestrial vertebrate species belonging to the animal classes Amphibia (amphibians), Reptilia (reptiles), Aves (birds) and Mammalia (mammals) and not commonly raised as pets (dogs and cats), livestock (cows, buffalos, swine, and goats) or poultry (chickens and ducks) in Vietnam were considered as wildlife species. Studies based on experimental infections in laboratory conditions or targeting aquatic species and insects were excluded (Figure 1). Studies based on pathogen detection from vectors (ticks, mites, and fleas, etc.) were included only if the animal host of the vector was unambiguously identified in the study.

2.2. Data Extraction From Eligible Records

Data were extracted from eligible publications using a standardized spreadsheet to create a pathogen inventory of terrestrial vertebrates in Vietnam (Supporting Information S1). The following data items were recorded for each publication: host genus, species and common name, type of pathogens targeted by the study, species name of the pathogen detected, prevalence, detection method, country, locality, year, and the type of human–wildlife interface at which the pathogen was detected.

After extraction from all publications, data were cleaned, and host and pathogen/parasite names were standardized using the NCBI taxonomic database. Host and pathogen taxonomy information were added to the spreadsheet. The latest recommendations from the International Committee on Taxonomy of Viruses (ICTVs) were used for virus names. All pathogens/parasites were classified as zoonotic (human infections recorded) or not based on literature data [22, 23]. Bacterial pathogens identified at the genus level only were considered zoonotic if their genus was known to include zoonotic species.

Interface names were also standardized and 13 types of human–wildlife interfaces were retained: wildlife farm, free-ranging animal in a farm, primate facility, restaurant, market, free-ranging animal in market, wildlife rescue center and sanctuary, zoo and recreational park, bat guano farm and guano collection, free-ranging animal in natural habitat (forests and fields), free-ranging animal in human settlements, confiscation from illegal trade, and international import. Definitions of all types of interfaces are available in Supporting Information Table S1.

2.3. Global Database of Host–Pathogen Associations

Data on bacterial and viral pathogens known to infect each host family were extracted from a global host–pathogen association database [23]. Pet (dogs and cats), cattle (cows, buffalos, swine, and goats), or poultry (chickens and ducks) species were removed from the dataset, and the numbers of pathogens were summarized for each host family. Zoonotic status of each pathogen was obtained from the same source [23].

2.4. Zoonotic Disease Risk Assessment

Due to the scarcity of data available for several host families in Vietnam that are commonly traded in the country (see Results), global data on host–pathogen associations from Shaw et al. [23] were used for this risk assessment. Our risk assessment is based on two types of risk estimates: a quantitative “taxa zoonotic risk” and a qualitative “trade point risk”. The “taxa zoonotic risk” was assigned to each wildlife taxa based on the number of zoonotic pathogens (“zoonotic risk”, high risk ≥ 50 known pathogens and ≥40 zoonotic pathogens; medium risk = 20–49 known pathogens and/or 15–39 zoonotic pathogens; low risk ≤ 19 known pathogens and ≤14 zoonotic pathogens) and the number of Vietnam priority zoonotic pathogens (“priority zoonotic risk”, high risk = ≥ 5 Vietnam priority zoonotic pathogens; medium risk = four–two Vietnam priority zoonotic pathogens; low risk = one or no Vietnam priority zoonotic pathogen) reported globally in each wildlife family. We selected the highest risk category between the zoonotic risk and the Vietnam priority zoonotic risk as the final “traded taxa risk” for each family.

For the “trade point risk”, we used an approach similar to the zoonotic disease risk assessment for wildlife trade chains and points of sale suggested by Wikramanayake et al. [24] for viral pathogens. This risk score was assessed qualitatively using three variables (transmission risk, spread potential, and zoonotic virus risk) based on market size, crowding of wildlife that creates stressful situations, hygiene conditions, number and turnover of people through the market, distance buyers may travel after visiting a market, and points along market trade chains that could allow pathogens to accumulate and amplify the potential for zoonoses [24]. As these trade conditions may also favor the spread of other types of zoonotic pathogens, such as bacteria, and are not restricted to viruses, we used the same risk scores as Wikramanayake et al. [24] in this study. Indeed, while viruses and bacteria differ in their biology and how they survive in the environment, they are transmitted in multiple ways and share several common transmission routes, such as direct contact, airborne transmission, fecal–oral routes, vector-borne transmission and indirect transmission via contaminated surfaces, that will be similarly impacted by the three variables described above. Fungi can also be transmitted through airborne spores or droplets, contact with contaminated surfaces, and fecal–oral routes. We also added three additional types of interfaces (bat guano farms, rescue centers and sanctuaries, and zoos and recreational parks) common in Vietnam and that were not included in the initial assessment of Wikramanayake et al. [24] (Supporting Information S2). Primate facilities were included within wildlife farms. Taxa zoonotic risk and trade point risk assessments were then combined in a risk matrix of traded taxa and trade points. We also assessed the cumulative zoonotic risk of raising taxa from different families in the same facility by combining the zoonotic risk of each taxa (very high risk = two high-risk taxa in the same facility; high risk = two medium-risk taxa OR one medium-risk taxa + one high-risk taxa OR one low-risk taxa + one high-risk taxa in the same facility; medium risk = two low-risk taxa OR one medium risk taxa + one low-risk taxa in the same facility).

3.1. Literature Review Results

The first screening of records obtained from the main and additional literature searches using title and abstract identified 112 potentially eligible records (Figure 1). Screening using full-text documents yielded 87 eligible records (Figure 1), including 86 peer-reviewed publications and one ONG report, from which data were extracted (Supporting Information S3). Data generated by the USAID PREDICT project in Vietnam and publicly available in the USAID Development Data Library [25] was also included in the inventory (Supporting Information S1).

Among all eligible records, 48 targeted viruses, 27 targeted bacteria, three targeted fungi, six targeted protozoans, and nine targeted macroparasites. Animal hosts targeted by those publications belong to the classes Amphibia (four publications), Reptilia (nine publications), Aves (nine publications), and Mammals (71 publications). Among mammals, Rodentia (34 publications) and Chiroptera (20 publications) were the most studied orders (Supporting Information Table S2).

3.2. Pathogens Associated With Terrestrial Vertebrates in Vietnam

A total of 162 pathogens/parasite species, including 22 parasites, 48 bacteria, two fungi, eight protozoans, and 82 viruses, were recorded in Vietnam in 46 families of terrestrial wildlife vertebrates belonging to four classes and 18 orders. Most of these pathogens, including a large proportion of viruses (81/82), were reported in Mammalia species (Table 1 and Supporting Information Table S3). It is important to note that families of some wildlife species commonly raised in wildlife farms in Vietnam, such as the Oriental rat snake (Colubridae), Siamese crocodile (Crocodylidae), sambar deer (Cervidae), and sika deer (Cervidae) [10] were not targeted by any of the publications included in the literature review and were therefore not represented in our pathogen inventory of terrestrial wildlife vertebrates in Vietnam (Table 1).

The data aggregated by host families show that the highest number of pathogens, both globally and in Vietnam, was observed in Muridae (rats and mice), followed by Cercopithecidae (Old World monkeys) (Table 1 and Figure 2). A large proportion (>50%) of the pathogens hosted by these two mammal families are zoonotic. However, only a small proportion of globally reported pathogens were detected in Vietnamese wildlife (Table 1 and Figure 2). Several bird families (Phasianidae, Columbidae, and Ardeidae) and mammal families (Suidae, Felidae, Sciuridae, Ursidae, Mustelidae, and Herpestidae) host many bacteria and viruses globally, but only a few of these pathogens were reported in Vietnam (Table 1 and Figure 2). Cervidae (deers) also host many important zoonotic and nonzoonotic bacteria and viruses globally, but no pathogen was reported in this host family in Vietnam (Table 1 and Figure 2).

3.3. Wildlife Hosts of Priority Bacterial and Viral Zoonotic Diseases in Vietnam

According to our literature review, 12 of the 29 Vietnam priority zoonoses identified by Pham-Thanh et al. [21], including one helminth, four viral, and seven bacterial diseases, were reported in Vietnam in 27 terrestrial wildlife host families (Figure 3). The Influenza A virus was reported in all bird families included in the inventory as well as in several mammal families, including Spalacidae (bamboo rats), Viverridae (civets), and Felidae (cats) (Figure 4). The Lyssavirus rabies was detected in several bat families as well as in Cercopithecidae monkeys. Paslahepevirus balayani (hepatitis E virus), Japanese encephalitis virus, and the helminth Trichuris trichiura were reported in Cercopithecidae monkeys only. The bacterial priority pathogens, including Leptospira spp., Clostridium sp., Escherichia coli, Salmonella spp., Rickettsia typhi, Rickettsia sp., and Orientia tsutsugamushi, were detected in Muridae (rats and mice), Sciuridae (squirrels), Soricidae (shrews), Ursidae (bears), Vespertilionidae (microbats), Struthionidae (ostriches), Gekkonidae (common geckos), and Pythonidae (pythons) (Figure 3).

According to our literature review, the remaining 17 priority zoonoses were not reported in any terrestrial wildlife species in Vietnam. However, according to global data [23], hosts of these 29 Vietnam priority zoonoses include 27 of the terrestrial vertebrate families included in our pathogen inventory (Figure 3). The highest numbers of priority bacterial and viral zoonotic pathogens are observed in Cercopithecidae (n = 10), Cervidae (n = 10), Suidae (n = 9), Ardeidae (n = 7), Columbidae (n = 7), Muridae (n = 6), Sciuridae (n = 6), Ursidae (n = 6), Felidae (n = 6), and Mustelidae (n = 6) (Figure 4).

3.4. Pathogens Identified at Interfaces Along the Wildlife Value Chain in Vietnam

The 13 interfaces at which pathogens and parasites were detected in Vietnam were classified under five main categories: (i) environment (free-ranging animals in natural habitat, in farms, in markets, and in human settlements), (ii) wildlife sources (wildlife farms, primate facilities and bat guano farms, and guano collection), (iii) wildlife distribution and sales (markets, confiscations from the illegal trade, and international import), (iv) wildlife consumption (restaurants), and (v) wildlife conservation and recreation (zoos and recreational parks and rescue centers and sanctuaries) (Supporting Information Table S4). Zoonotic pathogens were reported at all these interfaces in Vietnam, except in bat guano farms and in animals confiscated from the illegal trade, where the viruses detected in wildlife are not known to be able to infect humans (Figure 5 and Supporting Information Table S4). High proportions of zoonotic pathogens were observed in Cercopithecidae (Old World monkeys) and Pythonidae (pythons) raised in wildlife farms (Figure 5). Free-ranging animals sampled in animal farms, as well as in markets and human settlements, were also infected by numerous zoonotic bacteria. Most detections of priority zoonotic diseases in Vietnam were made in free-ranging animals (environment) as well as in wildlife farms and primate facilities (Figure 6).

Evidence of viral cross-taxa transmission at different interfaces was also reported in Vietnam (Supporting Information Table S5). Avian and bat coronaviruses were detected in three rodent families (Muridae, Spalacidae, and Hystricidae) in captivity or free-ranging in wildlife farms. Another bird coronavirus (duck-dominant coronavirus) was detected in Muridae rodents sold in markets, and a rhabdovirus was detected in both Cercopithecidae (Old World monkeys) and Hystricidae (porcupines) in wildlife farms. These examples confirm the risk of viral cross-taxa spillover in these facilities where different taxa are kept in close contact.

3.5. Zoonotic Disease Risk Assessment Along the Wildlife Supply Chain in Vietnam

High-risk host families include three bird families, Columbidae (pigeons and doves), Phasianidae (jungle fowl, peacock, pheasant, and quail), Ardeidae (wading birds), and eight mammal families, Cervidae (deers), Suidae (pigs), Felidae (cats), Ursidae (bears), Mustelidae (small carnivores), Cercopithecidae (Old World monkeys), Muridae (rats and mice), and Sciuridae (squirrels) (Supporting Information Figure S1). Medium-risk host families include two bird families, nine mammal families, and two reptile families, while all amphibian families are considered low-risk (Supporting Information Figure S1).

Facilities characterized by a very high risk of zoonotic spillover include bushmeat markets in cities and town, wildlife farms, restaurants, and rescue centers trading, housing and breeding birds belonging to Columbidae (pigeons and doves), Phasianidae (jungle fowl, peacock, pheasant, quail), Ardeidae (wading birds), and mammals belonging to Cervidae (deers), Suidae (pigs), Felidae (cats), Ursidae (bears), Mustelidae (small carnivores), Cercopithecidae (Old World monkeys), Muridae (rats and mice), and Sciuridae (squirrels) (Figure 7). Zoos, roadside stalls, research facilities, e-commerce and warehouse sales trading, and housing and breeding those animals are considered high-risk (Figure 7). Trading and breeding Scolopacidae (sandpipers), Struthionidae (ostriches), Giraffidae (giraffes), Herpestidae (mongooses), Viverridae (civets), Molossidae (free-tailed bats), Pteropodidae (fruit bats), Rhinolophidae (Horseshoe bats), Vespertilionidae (microbats), Soricidae (shrews), Hylobatidae (gibbons), Colubridae (snakes), and Elapidae (venomous snakes) in bushmeat markets in cities and town, wildlife farms, restaurants, and rescue centers is considered high risk (Figure 7).

We also assessed the cumulative zoonotic risk of raising taxa from different families in the same facility for 12 mammalian, avian, and reptilian families commonly raised in wildlife farms in Vietnam [10] (Figure 8). The zoonotic risk of breeding two wildlife species belonging to Suidae (pigs), Cervidae (deers), Ursidae (bears), Cercopithecidae (Old World monkeys), and Phasianidae (jungle fowl, peacock, pheasant, and quail), together in this type of facility is very high (Figure 8, below diagonal). The high numbers of facilities breeding together Cervidae or Ursidae + Cercopithecidae, Cervidae or Ursidae + Phasianidae, and Cercopithecidae + Phasianidae in Vietnam [10] indicate that these facilities represent a real zoonotic disease risk for human populations (Figure 8, above diagonal). The very high numbers of wildlife farms breeding Hystricidae + Viverridae, Elapidae + Colubridae, Colubridae + Viverridae, Elapidae + Viverridae, or Spalacidae + Viverridae also represent an important zoonotic disease risk (Figure 8, above diagonal).