If faced with an injured or trapped bat, the majority of respondents (73%; 600/821) reported that they would not handle the bat and that they would phone an animal welfare organization, ignore/avoid contact with the bat or euthanize it at a distance (Table 4 refers to responses not respondents). However, a quarter of respondents (25%; 206/821) indicated that they would handle the bat. Of the 206 respondents who indicated that they would handle the injured or trapped bat, 17% (36/206) stated that they would use their bare hands. Of those that stated they would not use their bare hands, 58% (95/165) said they would use any glove or readily available hand covering (including blankets and towels), while only 8.5% (14/165) indicated that they would use “thick, industrial gloves”. Among those who had previously handled bats, 32% (41/127) indicated that they would again handle a trapped or injured bat compared with 24% (165/694) of those who had never previously handled a bat (χ2 = 4.13, p = 0.0421). Of those respondents who reported previously hearing a warning about bats, 24% (83/345) stated that they would handle a trapped or injured bat. Surprisingly, 22% (27/121) of respondents who said that the main communicated message relating to ABLV was “not to handle bats” stated that they would handle a trapped or injured bat if they encountered one.

At the population level, rabies is the quintessential bat EID that has been studied most intensively. Public health guidelines recommend rabies vaccination for humans in high-risk groups, vaccination of pets as well as animals on public display, isolation of domestic animals from the wildlife reservoirs of rabies, and public health education on appropriate precautions. Current guidelines recommend that pre-exposure prophylaxis be offered to those in high-risk groups including veterinarians, animal handlers, rabies researchers, and some laboratory workers. In addition, the vaccine can be offered to long-term travelers to endemic areas, especially if immediate medical attention will be unavailable (148, 152). Routine vaccination of the general population is currently not recommended, mostly due to cost.

Despite advances in determining best practices for animal vaccination, control of rabies in domestic and wild reservoirs remains challenging in resource-limited settings. Control of rabies in bats has proven challenging. Bat rabies has been reported in every state except Hawaii and 1,806 rabid bats were documented in the United States during 2009 (19). Of all animals, bats in particular pose a serious risk for rabies and should be excluded from structures to prevent contact with humans (148, 152). However, widespread reductions in bat populations to control rabies is neither feasible nor desirable. Instead, some novel methods have been explored to control infection in bat populations. Vampire bats can efficiently digest only coagulated blood and they die if the consumed blood is not coagulated. Application of anticoagulant-containing ointment on the fur of captured vampire bats (with their subsequent release) leads to consumption of the coagulant by several roost mates via mutual grooming. Similarly, anticoagulation of livestock is another useful approach to control vampire bat populations where rabies is a threat (reviewed in Kuzmin and Rupprecht (7)). As another approach, it has been suggested that oral vaccination of wildlife may limit the spread of rabies by bats (153). Finally, we know that some species of moths are able to disrupt bat echolocation using ultrasonic clicks of their own (154, 155). The use of similar, artificially produced, sounds to ward off bats from human and livestock habitats should be explored.

Respondents were asked what actions they would take if they received a minor scratch (one without blood) and a major scratch or bite (one with blood) (Table 5). For minor scratches, 14% (117/821) of respondents indicated that they would ignore the scratch, while 10% (85/821) reported they would wash the wound with water and 21% (171/821) would use an antiseptic. Overall, 38% (311/821) said they would immediately seek medical care. For major scratches or bites, 0.5% (4/821) of respondents indicated that they would ignore the scratch, 10% (80/821) reported they would wash the wound with water and 12% (95/821) indicated that they would wash the wound with antiseptic. In total, 74% (603/821) said they would immediately seek medical care. Participants who had previously not handled bats (75%; 521/694) were more likely to immediately seek medical care than those who indicated that they had previously handled bats (65%; 82/127) (p value = 0.0137). Those who reported hearing warnings about bats also commonly reported that they would immediately seek medical care (80%; 275/345).

Tree and cave roosts could expose hanging and resting bats to direct contact with a potential maintenance host. However, as a first observation, the upside-down vertical position of bat roosting does not really favour disease transmission from an alternative host. For bat species roosting in tree-holes, the situation can be different as they can share temporally or directly their nest space with other animals. Secondly, the density of bats roosting in caves prevents the presence of many other potential hosts in the cave roof (but, for example, snakes can predate on bats in caves). During their feeding behaviour, frugivorous bats could be in direct contact with other hosts attracted by the fruits. Their nocturnal habits will limit the diversity of host they can interact with. We are not aware of any extensive study on the network of potential contacts between bats and other animals during their roosting and feeding behaviour. The majority of studies investigated potential of infectious contact from bats to other organisms. Novel technologies, such as camera traps equipped with nocturnal vision, could provide opportunities for more research on this topic.

No specific medical therapy has proven beneficial once people become ill from bat EIDs (at least of viral origin). For example, although rabies is an ancient disease, effective therapeutic treatment of rabies in humans continues to be very challenging. Rapid early diagnosis in the biting animal is critical, since identification of rabies before its fulminant stage allows for effective prophylaxis. Fulminant rabies continues to carry a very poor prognosis. The first case of the successful experimental treatment of rabies in a naïve patient was a 15-year-old girl bitten by a bat in 2004 (145). However, extension of the ‘Milwaukee Protocol’ (i.e., therapeutic coma, antiviral drugs, intensive medical care) in other patients has been much less successful (see for example Rupprecht (146) and Rubin et al. (147)).

Prophylaxis, after exposure but well in advance of illness, has a much higher success rate. Appropriate post-exposure wound cleansing has been shown to reduce significantly the likelihood of RABV transmission (148). Besides washing the wound with soap and water, unvaccinated persons should receive both rabies immune globulin and four doses of cell-culture vaccine. Globally, more than 12 million persons receive post-exposure prophylaxis each year (149).

Besides rabies, novel treatment strategies are being developed for other bat EIDs. The use of RNA interference has been suggested for the treatment of henipaviruses (150). These currently untreatable infections may be ameliorated by the introduction of small interfering RNA molecules homologous to the RNA in these pathogens. While promising in theory for many agents, this line of treatment is still in its preliminary stages, and issues such as efficacy in humans, delivery, and cost have yet to be addressed.

The potential for filoviruses to be used as bioweapons has spurred research efforts for an effective vaccine that could be used in an outbreak. For example, in a mouse model of hemorrhagic EBOV infection, a vesicular stomatitis virus-based vaccine has been shown to be safe and effective in preventing clinical presentation of disease (151). Furthermore, the possibility that this vaccine may be deliverable through mucosal surfaces offers potential as a rapid vaccination agent during an outbreak.

The incidence rate of leishmaniasis among the United States Armed Forces was 7.2 cases per 100,000 person-years for the period of 2001–2016, with the majority of cases being CL (Stahlman, Williams & Taubman, 2017). The reduced incidence rate of recent years has been attributed to better equipment and an emphasis on personal protective measures (Rowland et al., 2015). Nevertheless, there is still a cause for concern when troops are newly deployed to endemic regions; supportive resources may not be fully in place, deployed personnel may have limited knowledge, and a culture of preventative measure necessity may not have yet developed (Coleman, Burkett & Putnam, 2006). All of these factors can lead to an initial high caseload (Oré et al., 2015). The first line of protection from CL is through the use of personal protection techniques. However, there are no effective chemoprophylaxis drugs and no fully developed vaccines, and thus prevention of CL can be extremely difficult (Ghorbani & Farhoudi, 2018).

The preventative techniques that do exist focus on avoiding the bite of infected sandflies. Uniforms are impregnated with a type of pyrethroid (usually permethrin), insect repellants are recommended, and personnel are given pyrethrin-treated bed nets (Orsborne et al., 2016). Theoretically, these measures should bring the incidence rate to near-zero levels, but this has not been the case so far. In 2003 alone, the incidence rate was estimated at 200 per 1,000 soldiers (Gonzalez, Solís-Soto & Radon, 2017). Efforts have also begun to be focused on reducing the population of sandflies, in order to mitigate transmission risk. Cyfluthrin, a pyrethroid insecticide, has been used to decimate sand fly populations, and chloropiricin, a wide-spectrum nematicide and insecticide, has been used to reduce the rodent (leishmania reservoir) population (Aronson et al., 1998; Crum, 2005).

Fortunately, modern medicine has afforded CL cure rates up to 91%. The most effective treatment is sodium stibogluconate, given intravenously at doses of 20 mg per kilogram of body weight, for 20 days (Mitchell, Silvitz & Black, 2007; Stahlman, Williams & Taubman, 2017), with side effects such as fatigue, arthralgia, myalgia, headaches, and chemical pancreatitis. Sodium stibogluconate is efficacious, but development of new drugs is imperative due to these side effect, the threat of drug resistance, and the high cost ($100 per 100 mL) (Aronson et al., 1998). The threat posed by recent reports of treatment failures in South and Latin America leishmaniasis cases, including the induction of transmissible skin microbiota that significantly promotes inflammation, should be a concern for all in the infectious disease community, particularly the military (Mans et al., 2016; Obonaga et al., 2014; Gimblet et al., 2017).

Of the 572 cats with recorded test results, 29 (5.1%) were positive for FeLV antigen only, 91 (15.9%) were positive for FIV antibody only and 12 (2.1%) were positive for both FIV and FeLV. The remaining 440 cats (76.9%) tested negative for both diseases. Descriptive statistics on the signalment and clinical variables stratified by test status are presented in Table 2 along with the results from the univariable analyses. Only 2/41 FeLV-positive cats (4.9%) were retested with a confirmatory FeLV PCR. One was PCR positive and the other was PCR negative. Two of the FIV-positive cats were retested with a repeat FIV antibody SNAP test, both of which remained positive.

The majority of FeLV-positive cats (95.1%; n = 39/41) had at least one compatible clinical sign (mean 3.6, median 4, range 0–7). The most common clinical signs were lethargy (65.9%; n = 27/41), anorexia (61.0%; n = 25/41) and weight loss (48.8%; n = 20/41). Only 36.6% (n = 15/41) of FeLV-positive cats had a recorded history of treated cat bites. In the multivariable model (Table 3), the presence of anaemia, pyrexia and immunosuppression were significantly associated with FeLV positivity. Sex was also found to be significant, with male cats being 4.63 times more likely to be positive than females (95% CI 1.60–17.00).

The majority of FIV-positive cats (80.6%; n = 83/103) had at least one compatible clinical sign (mean 2.2, median 2, range 0–7) and, similar to FeLV, the most common clinical signs were lethargy (31/103; n = 30.1%), anorexia (40/103; n = 38.8%) and weight loss (35.9%; n = 37/103). Only 45.6% (n = 47/103) of FIV-positive cats had a history of treated cat bites. The multivariable model (Table 4) showed that being >5 years of age was associated with increased risk of FIV test positivity. Additionally, male cats, domestic breeds and cats with the presence of immunosuppression were also at significantly increased risk. In contrast, cats that were specifically being tested for FIV prior to vaccination were significantly less likely to be seropositive than cats that were tested for other reasons (OR 0.26, 95% CI 0.08–0.67; P = 0.013).

The median survival times of FeLV-positive and FIV-positive cats were 10 days (95% CI 0–16) and 650 days (95% CI 431–993), respectively. Median survival time of negative cats could not be calculated in this study as 264/440 negative cats (60.0%) were still alive at the end of the observation period. The majority of FeLV-positive cats (n = 29/41; 70.7%) were euthanased within 14 days of diagnosis (Figure 1). In contrast, only 21/102 FIV-positive cats (20.6%) died or were euthanased within 14 days of diagnosis. Long-term survival of the remainder of the FIV-positive cats was lower than FIV-negative cats (Figure 1). The log-rank tests showed that there was a significant difference in survival times between FeLV test positive cats and cats that were test negative for both FeLV and FIV (P <0.001) and between FeLV antigen-positive cats and cats that were test positive for FIV only (P <0.001).

Considering the scenario B and C in Figure 1, that bats are not the maintenance hosts of EBOV or that they are not the only host involved in the maintenance of EBOV, helps in focusing EBOV research protocols on a reduced range of potential transmission routes and potential alternative hosts interacting with bats in their specific and limited habitats. This means that if bats are not the maintenance hosts for EBOV, then there is only a limited number of candidate species to play the role of alternative maintenance hosts. This limited number of alternative maintenance hosts is defined by the ecology of bats that imposes on those alternative maintenance hosts only a few possible EBOV transmission pathways towards bats. From the biodiversity of African forest and the full web of interactions between species, a set of secondary hypotheses indicated in Table 1 can be tested through protocols presented to further investigate the role of different maintenance host candidates for EBOV. The observation of this limited number of hosts calls for testing them, even if only to exclude them from the list of hypotheses and strengthen the main hypothesis. As warned above, the EBOV multi-host maintenance system could include a complex network of interacting bat species (Figure 1A2) and to proceed by elimination of alternative hypotheses may be a way to zoom-in on the maintenance community. The hypothesis of human playing a role in ebolavirus maintenance has not been addressed here, even if persistence of EBOV in previously infected humans has been recently proven. This scenario would be more indicating of a change in the evolutionary trajectory of the pathogen (as moving from Step 4 to 5 in Figure 1 of Wolfe et al.) than of the natural maintenance of ebolaviruses that is considered here.

In order for these protocols to be efficient and well designed, insights from behavioural ecology, plant phenology, and molecular biology (amongst other disciplines) will be necessary. Integrated approaches to health have been proposed recently and, in EBOV ecology, they should promote the integration of ecological sciences into health sciences that are usually at the forefront of epidemiological investigations. For example, a lot of sampling of potential alternative hosts has been undertaken during ebolaviruses outbreaks (e.g.,). These investigations concerned mainly the search for “what transmits ebolaviruses to people” as they were implemented during a human (or great ape) outbreak, and in the vicinity of outbreaks. This does not mean that they can automatically inform on “what maintains ebolaviruses”. When looking for the maintenance host, investigations should also target the same and other alternative hosts during inter-outbreak periods with ecologically driven hypotheses. This is what is currently done for bats following the main maintenance hypothesis (e.g.,), but not often for alternative hosts. Experimental trials should also concentrate on the environmental conditions occurring in bat-specific habitats, which can be very different from human outbreak conditions.

The transmission routes towards bats represent interhost contacts of unknown intensity and frequency, and it would be difficult to compare their relative importance. However, one can prioritize some transmission routes based on the current knowledge. The insect food-borne and vector-borne routes of transmission need, surely, to be further investigated, as they can expose bats to numerous other hosts. Previous works on insects have mainly concentrated on sampling insects in the human outbreaks’ surroundings (e.g.,). When searching for a maintenance host that can transmit EBOV to bats, protocols should concentrate on insects in interaction with known-exposed bat species. This would mean combining bat behavioural ecology and arthropod capture protocols to detect their potential carriage of EBOV, as well as protocols exploring bat feeding habits (e.g., molecular detection of prey DNA in bat’s guano). For example, insect captures should be targeted where insects can bite bats, in caves or at canopy level, and not at ground level where bats may not occur. Studying host interaction networks at fruit feeding sites is also an interesting avenue to explore direct, environmental, and fruit-borne routes of transmission. Behavioural ecology could inform and help targeting protocols. Chimpanzees and monkeys can feed at the same height as bats. Some rodent species feed on fruits, but the selection of the arboricolous species feeding at the same height as bats can reduce the list drastically. Camera trap protocols could inform host interaction networks placing bat species in symmetric or asymmetric interactions with other potential alternative hosts.

Under field reality, and especially in rainforests, this list of protocols will need a carefully designed programme to be successful, rooted in interdisciplinarity. As bats, and especially those species that have been exposed to ebolaviruses, are the entry point of most of these alternative hypotheses (i.e., alternative host need to be in contact with bats), the behavioural and community ecology of targeted bat species will need to be locally understood. Data recorders, such as vector or camera traps, will need to be deployed where bats are currently roosting or feeding. This can be a difficult task. Understanding which feeding resources attract bats at a specific season requires a good understanding of indigenous and domesticated tree phenology (e.g.,). Prior to this work, a guano-based dietary analysis of the feeding behaviour of bats could help to map locally where and when bats will be present. Then, simultaneous protocols on bats and sympatric alternative hosts can be implemented, and a biological search for antibodies or antigens can be implemented. Combining protocols to test the main and alternative hypotheses could provide cost-effective and synergetic options.

To conclude, alternative hypotheses presented here should be explored alongside efforts to confirm bat species as maintenance hosts for EBOV. The ecology of those bat species already known to be exposed should be used to design protocols in order to target relevant alternative maintenance hosts. Given the number of species already involved/exposed to EBOV, the ecology of EBOV and its maintenance system can be expected to be complex, ecosystem dependent, and dynamic, due to global changes. The Ebola maintenance system, once isolated in the forests, is now interacting with humans and their modified environments and will adapt to it. Aiming at this moving target will require out-of-the-box thinking and interdisciplinary collaboration.

Food producing animals in stock has reached a total of more than 200 million (cattle, pigs, sheep, goats, and chicken) living on farms in Europe (see text footnote 1). It has been demonstrated that farm animals are reservoirs of many zoonotic pathogens to humans (34, 43). However, annually, a large amount of drugs are being used worldwide to sufficient quantities of food to feed a rapidly growing world human population (44–47). The farm animals consume worldwide approximately eight million kilograms of antibiotics annually (70% of which is used for non-therapeutic purposes such as growth promotion; forbidden in the EU from January 2006, and disease prevention) compared with only approximately one million kilogram per year used in human medicine (7). Antibiotics are routinely fed to livestock as growth promoters to increase profits and to ward off potential bacterial infections in the stressed and crowded livestock factory environment (48–52).

Despite large differences in methodology, most results demonstrate that not so long after the introduction of an antibiotic in veterinary practice, resistance in pathogenic zoonotic bacteria and/or the fecal flora increases. In particular, the wide-spread use of antibiotics in animals has resulted in an increased emergence of bacterial resistance to antibiotics, in zoonotic organisms such as Salmonella, Campylobacter, Shigella, Yersinia, Listeria, and Enterococcus genera, as well as the E. coli species. These zoonotic bacterial organisms are propagated primarily among animals and subsequently infect people (53–56). Humans can be infected by contact with animals and their dung or droppings or consumption of infected food. Severe diarrhea may develop, sometimes with blood in the feces. At all ages, but especially in children under 5 years and adults over 65 years, very serious illnesses often occur. These complications can result in loss of life or permanent kidney damage. According to the latest epidemiological trends, Salmonellosis and Campylobacteriosis are indicated as the most frequent food-borne bacterial zoonoses in Europe. The main food sources were eggs and mixed foods (57).

Furthermore, the recent emergence of ESBL-producing and carbapenemase positive Enterobacteriaceae bacteria in animal production (58), the emergence of farm associated MRSA ST398 (the main pig associated clone) (59–61), and of plasmid-mediated quinolone resistance in animal isolates and food products (62, 63) are great threat for public health. Unfortunately, these antimicrobial resistant “superbugs” are not only confined to hospital environments where antimicrobial use was high and many pathogens were prevalent. They are already widespread in the European community and animal populations that have a great hazard on public health (64, 65).

The causative agent of bovine tuberculosis, Mycobacterium bovis (M. bovis) has been identified worldwide. Thanks to decades of disease control measures that the occurrence of the infection has been greatly reduced. But there are still hundreds of new cases of human tuberculosis reported in the USA (66). Experience in Europe and the USA, has shown that M. bovis can be controlled when it is restricted in livestock; however, eradication is almost impossible once it has spread into wildlife as follows; the European badger in the United Kingdom (67), elk in Canada (68) and white tailed deer in the USA (69).

In the last decade, Q fever caused by Coxiella burnetii was one of the most devastating farm animal originated bacterial zoonotic bacteria in Europe. The Netherlands, in particular, has experienced several outbreaks from 2007 to 2010 following identification of a Q fever outbreak on various dairy farms in 2007. Infected humans were mainly located within the intensive dairy goat farms (<5 km) (70). The infection is spread by ticks, inhalation of the organism from the placental fluids, urine, and consumption of unpasteurized milk – meat products of sheep, goats, and cattle. The clinical signs in humans might be severe flu-like syndrome that may last for months (71).

To capture temperature and R0 variability featured dengue burden risk (DALYs), we began with estimating the season-varied temperature and R0. Figure 6A shows temperature distributions for seasons summer (TS), fall (TF), and winter (TW) as well as the whole temperatures (TAll) in the period 2001–2014. Among all temperature distributions, TS has the lowest variation in temperature estimates of 29.16°C (95% CI: 28.05–30.32) followed by TW, TF, and TAll, respectively, with estimates of 20.04°C (17.28–23.26), 26.07°C (21.97–30.87), and 24.98°C (18.37–34.14). Accordingly, variations of R0 estimates follow the same trend that R0,F > R0,All > R0,W > R0,S with medians and range values of 2.28 (95% CI: 1.67–2.47), 2.26 (1.21–2.47), 2.09 (1.46–2.43), and 1.92 (1.73–2.10), respectively (Figure 6B).

Figure 6C–F displays ER profiles conceived with DALYs-based dengue health burden given temperature and R0 variability. Generally, most dengue cases occurred in the fall with a risk probability of 50% (ER=0.5) and had DALYs estimates of 323 (95% CI: 267–379) (Figure 6D and Table 4), whereas the least dengue cases were likely to occur in winter with DALYs estimates of 60 (45–75) (Figure 6E and Table 4). Taking all seasons into consideration, DALYs estimates were more likely (ER=0.8), likely (ER= 0.5), and less likely (ER =0.2) to be 127 (95% CI: 91–163), 148 (115–181), and 419 (358–479), respectively (Figure 6F and Table 4).

Transmission of infectious agents is highly dynamic in bats and is associated with significant changes in bat population structure (e.g., birth pulse in maternity colonies). Periods of high prevalence of infected bats with Hendra and Marburg viruses have been shown to coincide with the timing of infectious agent spillover to other hosts. Although several studies have focused on these aspects, a precise assessment of the diversity of transmission routes involved in disease epidemiology in bats is still lacking, especially when considering the extreme diversity of bat species and associated ecological and biological features. Such information is not only relevant from a fundamental perspective but can provide major information for the development of biosafety measures, therefore limiting emergence risk.

Although communication and education on the risk associated with bat-borne pathogens has increased over the past decade, the benefits of bats in ecosystem functions (e.g., pollination, soil fertilization, and crops pest control) tend to be disregarded. Gaining knowledge on disease epidemiology and bat ecology is critical to fully assess the challenges associated with human health and bat conservation. In this context, implementation of One Health approaches seems essential and beneficial for a sustainable development, particularly for populations living in hotspots of bat-borne disease emergence.

Visceral leishmaniasis (VL) is a severe, often lethal zoonosis caused by the intracellular protozoa L. infantum. This disease is endemic in the Mediterranean Basin, South America, and parts of Asia (1, 2). The presence of domestic dogs, considered the main domestic reservoir of the etiological agent of VL (L. infantum), in endemic areas is a known risk factor for human infection (3–5).

Parasite transmission to the vertebrate host occurs through the bite of infected sand flies during blood feeding (6). Parasites inoculation occurs in conjunction with sand fly saliva (7, 8) composed of pharmacologically potent molecules exerting anticoagulant, vasodilating, and anti-inflammatory activity, which can directly affect hemostasis and the inflammatory and immune response of the vertebrate host (9–11). A previous study by our group demonstrated that DNA plasmids encoding the L. longipalpis LJM19 salivary protein induced a strong DTH response in immunized hamsters that provided protection against lethal L. infantum infection. In addition, LJM19-immunized hamsters exhibited increased production of IFN-γ and IL-10 after exposure to uninfected sand fly bites, suggesting that this DTH response could be a marker of protection against infection by L. infantum (12).

Our collaborators identified two L. longipalpis salivary proteins(LJM17 and LJL143) that elicit DTH reactions in dogs subjected to repeated sand fly bites. These authors subsequently vaccinated dogs using these salivary proteins in a very complex immunization strategy involving three intramuscular injections of DNA plasmids codifying LJM17 or LJL143, followed by one dose by intradermal route and two subsequent doses by intramuscular route coupled to electroporation of the same antigens. They then intradermally injected the two antigens as purified recombinant salivary proteins associated with CpG, and finally boosted with recombinant Canarypoxvirus expressing LJL143 or LJM17. The intention of these investigators was to elicit a strong response against the two proteins in an attempt to determine whether the generated immune response was capable of protecting the animals from L. infantum infection. These authors observed that LJM17 immunization induced enhanced RNA synthesis of IL-12, while the LJL143 protein provoked a mixed response as evidenced by the expression of IL-12 and IL-4 at the vaccine inoculation site. Moreover, in an in vitro killing assay, dogs immunized with LJM17 or LJL143 presented a reduction in the number of infected macrophages in the presence of autologous lymphocytes (13).

Due to the promising nature of these results, we chose to adapt this immunization strategy to formulate a simpler version more compatible with future potential use in the field, consisting of three intramuscular doses. Therefore, a single dose of plasmids encoding the two L. longipalpis salivary proteins (LJM17 or LJL143) was followed by two booster doses of rCanarypoxviruses vector expressing one of the two aforementioned L. longipalpis salivary protein genes. In addition to evaluating the immune response developed by immunized dogs, we also analyzed immunological, parasitological, and clinical parameters during the follow up of the dogs corresponding to 10 months after experimental infection with L. infantum in the presence of L. longipalpis saliva.

Bacterial zoonoses have a major impact on global public health. Both emerging and re-emerging bacterial zoonoses have gained increasing national and international attention in recent years. The closer contact with companion animals and rapid socioeconomic changes in food production system has increased the number of animal-borne bacterial zoonoses.

Animal bite injuries in daily human-animal contact are not surprising, especially for the school-aged children. Most of these wounds are medicated by patients as first aid and not registered in health systems. In more developed countries, most of the victims with moderate to severe bite injuries will seek professional medical treatment. Regardless, all bites should be treated as serious, especially if the skin is broken. Prompt diagnostic and treatment can prevent wound complications. The possibility to form biofilms by previously mentioned wound microorganisms is quite high, may cause severe tissue damage and protect the bacteria from innate-immune response and antimicrobials. The most of the commercial topical agents and wound dressings are ineffective against the biofilm matrix. Surgical repair (for example, CO2 surgical laser techniques, Leon Cantas, personal research notes 2014), which is usually used to obtain a better cosmetic result might be needed to remove biofilm formed bite infections. This mechanical debridement is essential in the eradication of a wound biofilm. Antimicrobials may be more effective in the treatment of the wound after debridement in the prevention of biofilm reformation. Despite the use of currently optimal culturing methods, approximately 7% of infected wounds yield no bacterial growth. In such cases, some other fastidious pathogens, i.e., Chlamydia spp., Mycoplasma spp., and even viruses should be investigated. New advanced molecular diagnostic techniques are needed. Prevention strategies for animal bites include close supervision of child–animal interactions, stronger animal control laws, better reporting of animal bites, and public education for better ownership of pets. Regular nail trimming, routine oral examinations under annual health checks and comprehensive dental treatments of the companion animals (i.e., routine removal of the teeth tartar and plaques) by veterinarians will reduce the bacterial mass exposure to humans in case of direct contacts or animal bites.

It is important to realize that enteropathogenic zoonoses may be contracted from both clinically sick and apparently healthy companion animals. Feeding of pets with raw food diets is a potential source of Salmonella, Campylobacter, and other important bacterial zoonoses; however, some recalls of commercial pet food diets have also occurred as a result of contamination with those microorganisms. Pig ear dog treats, in particular, have been implicated as an important source of Salmonella infection for dogs, which can also serve as a source of infection to humans.

Nevertheless, it can be said that easy-to-use personal hygiene rules should be applied by companion animal owners. Thorough hand washing with soap after handling of a companion animal and before eating or drinking, avoiding mouth-to-mouth contact, avoiding aerosolization of dusty fecal matter will help to prevent transmission of the zoonotic disease to humans. The animals with diarrhea should be isolated immediately and veterinary advice should be sought. The household should be cleaned with agents and kept as clean as possible.

On the other hand, the prevalence of antimicrobial resistance in small animal pathogens is increasing globally due to overuse of broad spectrum antibiotics by veterinarians. There is an immediate need for worldwide smarter use of antimicrobials that have some positive effect on the recovery of animals from life threatening diseases. National veterinary antimicrobial treatment guidelines should be established by the local authorities according to the updated routine surveillance results.

Chronic diarrhea, dermatitis, ear and eye infections of pets caused by microbes demand longer durations of antimicrobial remedies at home. More frequent use of advanced laboratory tests, such as; feed/insect/mould allergy tests and differential diagnosis of the other relevant auto-immune disorders may help to investigate the main underlying cause of the such reactions which can be managed in various alternative treatment methods (i.e., hypoallergenic diets) rather than antibiotics solely. Herein, pet specific auto-immune vaccines against allergens and auto-Lactobacillales (Auto-Lac, Leon Cantas, personal research notes, 2011–2014) as dietary supplements can also be more frequently administered within the preventative veterinary practice measures. Owners should be encouraged to insure their family animals to afford such costly veterinary services contradictory to the cheaper and sometimes life-long medical (i.e., antibiotic) treatment demanding options. Veterinarians should also spear more time to educate the pet owners under consultations to handle infected-antimicrobial treated animals with precaution due to irreversible consequences of the antimicrobial resistance development and its spread in households. Proper hand washing and use of gloves are strictly recommended while handling antimicrobial in veterinary clinics. Veterinarians should prescribe broad spectrum and synthetic antimicrobials preferably after culturing with extreme precautions (i.e., dosage, dosing intervals and length of the treatment). Reduced antibiotic use will hinder the development of antibiotic resistance in animal microbiota which might cause zoonotic infections in humans (50, 52).

Food-borne zoonoses are an important public health concern worldwide and every year a large number of people affected by diseases due to contaminated animal originated food consumption. Food hygiene education of the consumers is an important competent of food-borne diseases prevention. However, main prevention of food-borne zoonoses must begin at the farm level with in the concept of “One Health.” Herein, control of the production stress especially in intensive livestock industry, with the development of better animal health management routines (i.e., routine vaccinations, immune stimulants: pre-, probiotic feed additives) and the increased animal welfare programs, will contribute eventually to an optimal production of animal health. Increased antimicrobial resistance among emerging and re-emerging farm-borne bacterial pathogens in crowded settings (i.e., poultry, pig farms) is a growing problem. Restrictive antimicrobial choice with better animal welfare managements are needed to control the spread of antibiotic resistance elements.

In the EU, the use of avoparcin was banned in 1997 and the use of spiramycin, tylosin, and virginiamycin for growth promotion were banned in 1998. All other growth promoters used in feeding of food producing animals were banned from January 1, 2006 after a few national bans the years ahead3. In the U.S., politicians are still discussing to introduce a similar ban (S-742, 109th U.S. Congress (Preservation of Antibiotics for Medical Treatment Act). Despite the ban on the use of all antibiotics as growth promoters in the EU and a ban on the use of quinolones as growth promoters in the poultry feed in the US medical, important antibiotics are still routinely fed to livestock prophylactically to increase profits and to ward-off potential bacterial infections in the stressed and crowded livestock and aquaculture environments in some parts of the world (50, 90, 91). Because stress lowers the immune system function in animals, antibiotics are seen as especially useful in intensive animal confinements (92). The non-therapeutic use of antibiotics involves low-level exposure in feed over long periods – an ideal way to enrich resistant bacterial population (93, 94). Moreover, antibiotic resistance has been detected in different aquatic environments (95). Fish pathogenic bacteria often produce devastating infections in fish farms where dense populations of fish are intensively reared. Bacterial infections in fish are regularly treated with antibiotics in medicated feed. So far, most of the fish pathogenic bacteria with a history in diseased fish farms have developed drug resistance (96). Modern fish farming relies increasingly on vaccination procedures and improved management to avoid infections (97). For example, the Norwegian aquaculture industry has produced over one million tons farmed fish4 by using improved vaccines, management techniques, and only 649 kg of antimicrobials in 2011 (98).

Vector-borne and zoonotic bacterial pathogens are a major source of emerging diseases, and since the time of Hippocrates, weather and climate are linked to the incidence of such infectious diseases. Complexity of epidemiology and adoptive capacity of microorganisms and the arthropods make the vector-borne disease almost impossible to eradicate. Insect repellants, routine tick checks after outdoor activity in risk regions, prompt-proper tick removal, use of long sleeves and trousers (light-colored), and routine insecticide treatment of pets are recommended as general preventative measures (99). Herein, Lyme disease, tick-borne illness, is vastly underestimated over past decades and clearly the urgent prevention is needed. Besides individual awareness of such vector-borne diseases, better national surveillance and reporting programs will contribute to improved the disease control strategies. Clinicians have an important role in the effective management of vector-borne zoonotic diseases, with enhanced differential diagnostic skills based on clinical symptoms and rapid molecular identification techniques (100–103). Most of the time, the clinicians are on the first line of detection of these epidemics due to large group of patients with novel sets of similar symptoms. Increased medical networking via online databases offer a broad overview to followers with regard to changes in temporal patterns of illness in real time, which helps faster detection of new epidemics (104).

Identification and control of emergent zoonotic bacterial diseases require a “One Health” approach, which demands combined efforts of physicians, veterinarians, epidemiologists, public health workers, and urban planners. Collaborative international routine surveillance strategies, prompt – reliable agent identification techniques, and optimization of the treatment regiments will ensure the prevention and management of such infections.

Pathogen transmission from NHP to humans can occur by air droplets, fecal-oral contamination, cutaneous contact, bite or by an arthropod vector (Figure 7). There are serious risks that humans can be bitten by monkeys when they keep them as pets, when scientists maintain monkeys for medical research, when staff members handle NHP in zoos or national parks, when travelers visit sites with high prevalence of free-roaming monkeys, and when ecotourists escape from conventional sightseeing to meet great apes in Africa. NHP bites remain poorly documented to date. Incidence and type of NHP bites will depend on geographic location, industrialized vs. developing country, ecosystems, and cultural factors (117). According to WHO, monkey bites accounts for 2 to 21% of animal bites injuries worldwide. Among 332 patients who sought medical attention for bite wounds in 1975 in USA, five (1.7%) claimed to be injured by monkeys (118). A retrospective analysis of incidents caused by NHP in the UK on primatologists indicated that 85 Cynomolgus monkey bites had occurred over a 6 year period (119) [Tribe and Norenux2c 1983]). In 1 year, 55 patients presented to St Bernard's Hospital, Gibraltar, with a primate bite (120).

Most frequently (75%), bites are located on the hands, arms and legs (121). Depending on the force of the animal bite it might cause a crush injury with a variable amount of tissue damage. The severity of the injury ranges from superficial abrasion to crush wounds and major tissue loss. The risk of infection increases with the size of tissue destruction. Bacteria isolated from humans bitten by monkeys cover a large spectrum of species including Streptococcus spp., Enterococcus spp., Staphylococcus spp., Neisseria spp., Haemophilus spp., Bacteroides spp., and Fusobacterium spp. (122). A bacteriological analysis of 17 rhesus monkeys indicated that Neisseria spp., Streptococcus spp. and Haemophilus parainfluenzae, were the species most frequently isolated from the tongue microbiota of these animals (123). The sub-gingival microbiota of macaques (M. mulatta) was found to include Haemophilus spp., Fusobacterium cleatum, Peptostreptococcus micros, Streptococcus spp., Actinobacillus actinomycetemcomitans, Wolinella spp., Campylobacter spp., Eikenella corrodens, and spirochetes (124). Regarding respiratory pathogens, Klebsiella pneumoniae was found in the nose and throat of NHP (125). Tuberculosis is rare in wild NHP, yet animals carrying Mycobacterium tuberculosis could infect back humans (Table 1). Natural infections with Mycobacterium leprae was reported in chimpanzees and sooty mangabeys (Cercocebus atys) (126). Recently, different strains of Mycobacterium leprae were isolated from NHP, including chimpanzees, sooty mangabeys and cynomolgus macaques (127). Mycobacterium orygis was found in captured rhesus monkeys (128). As far the fecal-oral transmission, Shigella flexneri and S. sonnei infections are common in NHP, as well as enteropathogenic Escherichia coli, Salmonella enteritidis, S. typhimurium, Campylobacter fetus, C. jejuni, Helicobacter pylori, and many others (129). Special attention should be drawn to endemic Treponema pallidum infection with genital stigmata in NHP from Guinea, Senegal, and Tanzania. Many NHP in Africa, including Papio papio, P. anubis, P. cynocephalus, Chlorocebus pygerythrus, and Cercopithecus mitis, were found to suffer from treponematoses (130). An isolate called Fribourg-Blanc obtained from a baboon lymph node was found to be genetically linked to Treponema pallidum pertenue, the causative agent of human yaws and, could be transmitted to humans by contact (131, 132). Vigilance is required regarding Bacillus anthracis, since anthrax killed chimpanzees and gorillas in West and Central Africa (133).

Regarding parasites, many NHP have been found infected with Trypanosoma cruzi, a pathogen isolated in 1924 from South American squirrel monkeys (Chrysothrix sciureus), that causes anorexia, weight loss, dehydration in monkeys and Chagas disease in humans. Natural T. cruzi infection has been reported for several NHP such as marmosets, spider, cebus, rhesus monkeys, and gibbons (134) and could be transmitted to humans via triatomine bugs after feeding on infected animals. Giardia lamblia, an enteric flagellate, induces diarrhea in monkeys and children (135). The parasite Entamoeba histolitica, common in OWM, has been reported in most NHP including the great apes (gibbons, orangutans, and chimpanzees) as a cause of severe enteric disease. It can infect humans as well, leading to dysentery. OWM such as mangabeys and great apes such as chimpanzees can carry either Schistosoma mansoni or S. haematobium (136). Leishmania major has been identified in wild gorillas' feces (137).

Several other parasites such as Amoeba, Toxoplasma, Babesia, Cryptosporidia, Coccidia, nematodes, and cestodes can be found in NHP possibly presenting a risk for humans (138).

The Plasmodia that infect great apes are usually of a different group than those found in OWM, and they are related to parasites inducing malaria in humans. Indeed, characterization of Laverania spp. found in various apes identified lineages in eastern chimpanzees as well as western lowland gorillas that were nearly identical to P. falciparum and P. vivax (139, 140). Among others, P. knowlesi that circulates in cynomolgus, leaf monkey and pig-tailed macaques in Southeast Asia inducing moderate symptoms in these natural hosts, can be fatal for rhesus monkeys. In contrast, P. cynomolgi, found in cynomolgus, toque monkeys, pig-tailed macaques, Formosan rock macaque and leaf monkeys, induces moderate symptoms in rhesus monkeys (141). To note, P. cynomolgi, P. siminovale and P. inui are related to P. vivax, P. ovale and P. malariae in humans, respectively. Cross infection of P. knowlesi has been documented in humans (70, 142, 143). Other plasmodia are found in great apes such as P. pitheci in orangutans in Borneo, and P. rodhaini in chimpanzees and gorillas. African apes can be considered as the source of parasites responsible for the human malaria.

Other blood sucking insects such as flies, ticks, fleas, sandflies, lice could also transmit pathogens from NHP to human; tsetse flies might transfer trypanosomiasis and lice can transfer Bertiella mucronata tapeworm (144).

The number of animal species which can serve as Leishmania reservoir hosts is ever increasing, including rodents, canines, and farm animals (Stephens et al., 2016). Leishmaniasis was not traditionally considered endemic to the United States, although recent epidemiologic findings reveal this status may be changing, secondary to globalization and autochthonous infections leading to persistent endemicity, especially along the southern border and new animal hosts (Wright et al., 2008; Barry et al., 2013; McIlwee, Weis & Hosler, 2018). Incidence and disease burden are higher in societies where people live in close proximity to host animals (De Vries, Reedijk & Schallig, 2015). Similar to malaria, it is a vector-borne disease, requiring a phlebotomine sand fly to pick up amastigotes during a bloodmeal from an infected (reservoir) host. Amastigotes undergoes development and maturity in the fly, which then inoculates infective promastigotes into a new mammalian host during the next blood meal. Disease manifests as three distinct clinical forms: cutaneous (including diffuse cutaneous form), mucocutaneous, or visceral. Cutaneous leishmaniasis (CL) has been the most diagnosed of the three among deployed United States servicemen and women (Weil, 2010; Beaumier et al., 2013), with L. major, the most prevalent (Herwaldt, 1999). Overall, 90% of global CL cases occur in Afghanistan, Brazil, Iran, Peru, Saudi Arabia, and Syria. With the United States military active involvement and troop deployment to these locations, increased cases of leishmaniasis were recorded, until drawdown when the number of deployed soldiers reduced dramatically (Shirian et al., 2013; De Vries, Reedijk & Schallig, 2015). Expectedly, this is not unique to the US military, with cases of L. major infection and multiple reports of cutaneous disease among British, Dutch, and German soldiers as well (Faulde et al., 2008; Bailey et al., 2012).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Dirofilaria immitis does not meet the criteria used to categorize other vector‐borne pathogens because transfusion of microfilaria from an infected donor cannot lead to heartworm disease in the recipient. However, filaremic blood transfused to a recipient has the potential to interfere with diagnostic testing, can be infectious to mosquito vectors, and can carry Wolbachia spp.2 In addition, a donor infected with D. immitis would not be considered a healthy donor, and collection of large amounts of blood from such a donor could be unsafe. Therefore, it is recommended that dogs and cats to be used as blood donors in heartworm endemic areas be screened for D. immitis infection and placed on heartworm prophylaxis.

Anaplasma spp.

A. phagocytophilum and A. platys are the causative agents of canine granulocytic anaplasmosis and infectious canine cyclic thrombocytopenia, respectively. Transmission of A. phagocytophilum occurs via Ixodes scapularis and Ixodes pacificus ticks in the United States. Widespread subclinical infections followed by pathogen clearance appear common in both humans and dogs. A further pathway for transmission is via infected blood, either experimentally or by blood transfusion.3 In human medicine, several reports of transfusion‐transmission of A. phagocytophilum via different blood products (non‐leukoreduced/leukoreduced RBCs, leukoreduced platelets) and also transfusion‐transmitted granulocytic anaplasmosis have been documented.4, 5 Widespread subclinical infections followed by pathogen clearance appear common in both humans and dogs, and in immunocompromised or elderly people such infections can cause severe disease. Donation screening or inactivation by pathogen reduction technologies is considered in human medicine.5 Anaplasmosis occurred in a splenectomized dog on chemotherapy after a packed RBC transfusion; both the donor and recipient tested positive by PCR (Kohn unpublished data). PCR positive dogs can be seronegative and can have clinical and hematologic variables within reference intervals.6, 7 Antibodies to Anaplasma species can be detected using IFA assays, automated fluorescence‐based systems, a point‐of‐care lateral flow ELISA assay,1 or laboratory‐based ELISA assays.2,7, 8 Serologic cross‐reactivity among Anaplasma species occurs in some assays, but not all (Table 3). The seroprevalence (IFA or ELISA) is high in endemic areas (up to 50%) and antibody titers may persist for several months or even years.6, 9

The extent to which A. phagocytophilum can persist in tissues and contribute to chronic disease manifestations in humans and dogs currently is unknown. In 1 study, treatment of experimentally infected dogs with prednisolone up to 6 months after infection was followed by development of positive PCR results for the organism, and in some dogs, thrombocytopenia and reappearance of morulae on blood smears.10 In another study, dogs infected with A. phagocytophilum by exposure to wild‐caught Ixodes scapularis ticks were PCR positive for at least 12 weeks.7 In light of the above information, the panel recommends that optimal standards are to screen donors using both serology and PCR, and dogs that test positive with 1 or both assays should be excluded. Exclusion of all seropositive dogs might limit the donor pool in endemic areas (see comments in Table 1).

Anaplasma platys is thought to be transmitted by Rhipicephalus sanguineus, and infections are common in regions endemic for this tick. Anaplasma platys can establish a chronic, persistent subclinical infection, sometimes accompanied by mild thrombocytopenia.11 Dogs infected experimentally with blood developed severe thrombocytopenia within 7 days after inoculation.12 Transfusion‐transmitted infection or disease has been reported neither in humans nor in dogs. An IFA assay for detection of serum antibodies is commercially available, but cross‐reactions occur with A. phagocytophilum.13 Species‐specific PCR testing of blood samples is the diagnostic method of choice. Assays for A. phagocytophilum antibodies may or may not detect A. platys antibodies (Table 3).7 Donor dogs that are negative for antibodies and negative using species‐specific PCR are optimal (Table 1).

Our current study found that the occurrence of FeLV was 2.6% and the occurrence of FIV was 18.5% among predominantly client-owned cats across New Zealand. These results were markedly different from the 5.5% FeLV prevalence and 10% FIV prevalence reported in the previous cross-sectional study of random blood samples submitted to a diagnostic laboratory in New Zealand27 but broadly consistent with estimates from Australia, which were obtained from demographically similar populations.13,32

There are several possible reasons for our higher observed occurrence of FIV. First, veterinarians who responded to our survey indicated that the majority of cats were only tested because there was a high index of suspicion for disease, such as the presence of compatible clinical signs, known exposure to infected animals and presentation for bite wounds. These have previously been identified as risk factors for FIV,2,8,13 and we would therefore expect a higher occurrence of disease in this population compared with cats selected at random. Even though most cats were not retested for confirmation, previous research has demonstrated a good correlation between the presence of FIV antibodies on ELISA and the presence of proviral DNA on blood PCR, which is indicative of active infection.33 Second, false-positive reactions can occur on the FIV antibody tests used in practice owing to previous FIV vaccination or the presence of maternal antibodies.34–36 Although it is unlikely that veterinarians would have tested a known vaccinated cat, many patients present with an incomplete medical history and it is therefore possible that some of the tested cats were previously vaccinated. Third, in the free-text comments of our survey, several veterinarians indicated that they had stopped testing for FIV and FeLV in-house because clinical cases were rare in their practice regions. The study sample may therefore have been biased towards practices with higher clinical disease occurrence.

The occurrence of FeLV test positivity in cats from the first-opinion veterinary practice was marginally greater than expected at 7.2% vs the 2.6% observed in the general population across New Zealand. With FeLV antigen ELISA testing, there is no known interference from previous vaccination or maternal antibodies that could lead to false-positive reactions. However, the sensitivity and specificity have been reported at 92.3% (95% CI 79.7–97.3) and 97.3% (95% CI 95.5–98.4%), respectively.23 Given the relatively low occurrence of FeLV and the fact that <5% of positive cats were retested for confirmation, we cannot rule out the possibility that some of these cats were false-positive reactors on the in-house ELISA. Several respondents to the national clinic survey also anecdotally reported isolated incidents of particularly severe clinical FeLV, which suggests there could be regional differences in FeLV occurrence and epidemiology. A future prospective study with unbiased sampling methods is needed to explore this further.

In guidelines published by the American Association of Feline Practitioners (AAFP), it was recommended that all cats be routinely screened for FeLV and FIV when they are first acquired, when they have compatible clinical signs and when they have high-risk lifestyles, which include known exposures to infected cats, evidence of bite wounds and access to the outdoors.19 Previous research has documented that the majority of client-owned cats in New Zealand are free-roaming,37 and at least 36% of cats from our first-opinion practice data had a known history of medical treatment for cat bite wounds, which would place them in the high-risk category. However, <2% of client-owned cats across New Zealand were tested for FeLV and FIV over a single calendar year. Similar low levels of compliance with testing guidelines have been reported in the USA, even when testing was offered at no cost to the client.24 In the free-text comments of the national clinic survey, one veterinarian indicated that routine testing was not performed because a positive diagnosis was unlikely to change the management or clinical outcome for the patient, particularly since many clients were reluctant to confine their cats. To our knowledge, there has been little research to date in the field of FeLV and FIV epidemiology exploring client perceptions and responses to positive diagnoses. This could have significant implications on our ability to control these viruses at the population level and therefore warrants further investigation.

Cats that tested positive for FeLV and/or FIV on the in-house ELISA frequently presented with lethargy, anorexia, weight loss and immunosuppression, which may have indicated that their retroviral diseases had progressed to an advanced clinical stage. Given that relatively few cats were retested for confirmation, we cannot rule out the possibility that these clinical signs were associated with other underlying disease processes. However, we speculate within the limitations of reviewing historical medical records that many of these cats would have been euthansed anyway, regardless of confirmation, owing to the poor clinical prognosis. This likely explains why the survival times for positive cats were significantly lower in our study compared with the survival times reported in other studies where cats were still clinically healthy at the time of diagnosis.20–22,38 Other factors that may contribute to lower survival times could include coinfection with both FIV and FeLV,39 or other opportunistic pathogens,40,41 as well as differences in the potential virulence of common circulating strains in New Zealand compared with other countries.42

The preliminary investigation into risk factors for FeLV and FIV found no evidence that previous bite wounds, known exposure to infected cats or being sexually intact increased the risk of test positivity, despite these being identified in other studies as important risk factors.9,10,12 While we cannot rule out biases due to missing information in historical medical records or biases due to data only being obtained from a single practice, it is also possible that transmission dynamics in New Zealand are different from other populations owing to the majority of cats being free-roaming. For example, FeLV is traditionally considered a pathogen that requires prolonged close friendly contact like mutual grooming or shared water for transmission given the poor survivability of the virus in the environment.43,44 Housemates that are indoor–outdoor may have fewer close contacts. Similar to other studies from the USA,8 Australia32 and Germany,2 we found that being male significantly increased the risk of FeLV and FIV test positivity. From a behavioural perspective, male cats have been shown to have more aggressive tendencies, leading to a greater risk of bite wounds.24,45 The majority of cats in our study were also already neutered, which may explain why this was not found to be a significant risk factor.

The AAFP guidelines also advocate the use of vaccines in cats with high-risk lifestyles.19 This has become more complicated in New Zealand now that the FeLV vaccine is no longer commercially available. Although vaccination rates were still low prior to March 2016 when the vaccine was removed,30 our study found that FeLV vaccines were primarily administered to high-risk cats. It will be important to monitor how the epidemiology of the disease changes now that these cats can no longer be protected. Several practices declined to participate in the national survey because they were unable to easily query their practice data to obtain summary data on the number of feline patients and results from the diagnostic testing. New initiatives are currently underway in New Zealand to improve data capture from electronic medical records that will hopefully make these type of monitoring studies easier in the future.46,47

Rhabdoviridae contain six genera, including Lyssavirus, the most important bat-associated virus. At least 14 species of the Lyssavirus genus can be detected in bats, which are considered the ancestral hosts for these viruses. Lyssaviruses can be found worldwide and can be classified using different criteria, such as genetic distance, antigenic patterns, geographical distribution and host range.10,11 The characteristic bullet-shaped virus, transmitted to humans through the bite of infected animals, causes an acute, and frequently fatal, encephalitic disease.

The first report of a transmission of a viral disease from bats to humans was in 1911 and related to the rabies virus (RABV) belonging to the Lyssavirus genus.12 Carini12 suggested a link between rabies infection and hematophagous bats, known as vampires, in Central and South America. Several years later, rabies was also detected in non-hematophagous bat species.13 Although RABV is found worldwide in several terrestrial hosts, its presence in bats is observed only in the Americas. In Europe, four different Lyssaviruses have been isolated from bats: European Bat Lyssavirus type 1 (EBLV-1) and European Bat Lyssavirus type 2 (EBLV-2), Bokeloh Bat Lyssavirus (BBLV) and West Caucasian Bat Virus (WCBV).14 Recently, a new putative Lyssavirus in bat, named Lleida Bat Lyssavirus (LLBV), was found in Spain.15 To date, no human exposure to LLBV has been reported. EBLV-1, with the sub-types EBLV-1a and EBLV-1b, is the most isolated type throughout Europe. In addition, spillover infections by EBLV-1 in other mammals were also observed.13,14 The type 2 of EBLV is believed to be less virulent than type 113 and is found less frequently being present only in few countries and human contamination has been reported only in two cases.14 Two other members of this family are found in bats but significantly less frequently than the previous ones: BBLV isolated in Germany and France13,16,17; WCBV isolated once in the Caucasus Mountains but also detected in Kenya in seropositive bats, suggesting a greater geographical distribution.14,18 Australian Bat Lyssavirus (ABLV) is the first endemic lyssavirus identified in Australia and is phylogenetically related to RABV and EBLV1.10,19 ABLV has been identified in all flying fox species on Australia’s land mass. Three fatal human infections by ABLV have been reported. Additionally, other viruses of this family detected in bats are summarized in Table 1.

Dogs can be infected with several hemoplasma species, including Mycoplasma haemocanis, “Candidatus M. haematoparvum,” and possibly also “Candidatus M. haemominutum” or a related organism.29, 30, 31 Although ticks have been implicated in transmission of M. haemocanis, the mechanism of transmission has not been proven. Kenneled dogs and research animals appear to be at higher risk for infection by M. haemocanis. Diagnosis of infection is based on PCR assay of whole blood.32 No serologic assay for hemoplasma infection is commercially available. In general, dogs are subclinically infected with these organisms, but M. haemocanis can cause anemia in splenectomized dogs, with a few case reports of infected dogs with other immunocompromising comorbidities. Only a single clinical infection with “Candidatus M. haematoparvum” has been reported in a splenectomized dog with hemic neoplasia being treated with chemotherapy, and it was unclear to what extent the hemoplasma played a role in development of anemia.33 This dog received several units of blood products, some of which tested positive using PCR for “Candidatus M. haematoparvum,” and tested negative before transfusion (Sykes et al, unpublished data). Therefore, optimally, donor dogs should be screened for all hemoplasma by PCR assay and excluded if positive (Table 1). However, until more is learned about the risk of transfusing blood testing positive for “Candidatus M. haemominutum” and “Candidatus M. haematoparvum”, testing for these pathogens could be considered optional. It also should be kept in mind that the prevalence of hemoplasma infection in the general client‐owned pet dog population in North America appears to be low (<5%).34 Whether the viability of canine hemoplasma species is lost during storage of blood products (see feline hemoplasmas) requires further study. Because antimicrobial therapy does not reliably eliminate hemoplasmas, the panel does not recommend treating potential donors with antimicrobials in an attempt to eliminate infection.

Thirty beagle dogs of both sexes, aged 2 to 3 months, were acquired in Paraná, Brazil. Throughout the study, these animals were housed at the Experimentation Kennel facility in the region of Monte Gordo, located in the municipality of Camaçari, Bahia-Brazil. All dogs received routine vaccinations (rabies, distemper, hepatitis/Adenovirus type2, leptospirosis, Parvovirus and Coronavirus) and were dewormed. Blood samples were collected for serological evaluation and no dogs showed any detectable levels of antibodies against Leishmania, L. longipalpis saliva, Ehrlichia canis, Borrelia burgdorferi, or Dirofilaria immitis. Dogs were kept under surveillance and received veterinary medical care, a balanced feed, water ad libitum, and were housed in covered kennels with stalls protected by thin netting to prevent natural exposure to vectors, with animals grouped up to five per stall, separated according to gender. These procedures were kept during for all the period of the study (immunization, infection, and following up).

Two features of human behavior throughout history are striking. The first has been the domestication of numerous animals for food production, as working animals for activities such as transport, and for hunting purposes. They are also, particularly in modern times, used as a source of companionship, and in the USA alone there are estimated to be over 77 million dogs.1 The second feature is our propensity to travel. This trait has accelerated in the modern age through technical innovations such as the motor car, air travel, and international shipping.2 Human travel through migration, leisure, or business continues to increase in parallel with a population increase and the spread of affluence. This provides opportunities for pathogens to move to new areas and cause outbreaks of disease. Commercial air travel has been particularly effective at transporting pathogens such as SARS coronavirus and influenza viruses.3 It is also contributing to the spread of disease vectors and the diseases they transmit.4 As we travel, we take animals with us either through trade in livestock or in the movement of companion animals.

In Europe, the free movement of people within the continent has been enabled by the formation of the European Union (EU) and liberalization of border controls. This was formalized by the Schengen Agreement signed by member states in 1985 and implemented in 1995. This created the Schengen Area that reduced border controls between member states and allowed free movement of people between countries. Some countries, including the UK and Ireland, have in the past, agreed opt-outs that retain limitations of movement into those countries.

When humans travel, they often take their companion animals, particularly dogs, and these in turn can relocate the pathogens and vectors they harbor. Canine parvovirus emerged in 1978 as a new disease in dogs causing hemorrhagic enteritis. Retrospective serology suggested that the disease appeared in Europe in 1976 and spread throughout the world by 1978.5 The mechanism that enabled global dissemination of the virus has been attributed to contaminated footwear. There are numerous canine-associated diseases but relatively few are significant zoonoses. However, of these, a number are fatal to humans and control is essential to protect public health. Globally, the most significant is the rabies virus. Infection leads to fatal encephalitis, for which there is no treatment.6,7 Pre-exposure vaccination protects against the disease in mammals, and due to the extended incubation period between contact with a rabid animal and development of disease, often measured in months, timely post-exposure vaccination is also effective in humans.8 The dog is the most important reservoir for the virus, and contact with dogs is responsible for virtually all human cases of the disease. Where efforts have been concentrated on controlling dog rabies, the reduction in human cases of disease has been dramatic.9 Dog rabies has been virtually eliminated from Europe, although there are examples of reintroduction10,11 and cross-border movement of rabid animals.12

Alveolar echinococcosis is caused by the tapeworm Echinococcus multilocularis. It is a fatal condition that is relatively rare in Europe although there are clear areas of endemicity, resulting in human infections in an area of Europe ranging from France in the west to Austria in the east.13 The disease manifests as a tumor-like growth of cysts containing the larval stage of the parasite in organs such as the liver. Detection of cysts often occurs many years after initial infection, and without intervention such as surgery to remove the cyst(s), the disease is fatal. Dogs act as the definitive host for the adult form of the parasite and movement of infected dogs can lead to the spread of tapeworm eggs and introduction of the disease into new areas. As the adult worm is very small (less than 5 mm) and does not cause clinical signs in the definitive host, spreading of eggs and human infection can remain undetected until clinical symptoms develop.

Leishmaniasis is caused by protozoa belonging to the genus Leishmania. Two forms are recognized, ie, cutaneous leishmaniasis causing skin lesions and the more serious visceral disease involving multiple organs. Natural transmission is through the bite of phlebotomine sandflies belonging to the genera Phlebotomus in the Old World.14 Distribution of leishmaniasis is limited by the presence of the vector. In Europe, the vector is indigenous to those countries around the Mediterranean Sea. The major mammalian reservoir is the domestic dog. Estimates of autochthonous human cases in Europe are approximately 700 a year.15 Numbers of cases in Turkey are higher, with over 3,000 annually. Non-endemic countries in Europe do encounter cases of canine leishmaniasis as a result of pet movements.16

In the absence of border restrictions, it is difficult to establish the true extent of regional movement of animals either for trade or through holiday travel. Monitoring of dogs and cats entering the UK through its pet travel scheme indicated that almost 100,000 animals entered the country annually (Table 1). A similar situation is likely to exist for most EU member states. In addition to this, there is the problem of illegal movement of animals either by organized groups for commercial purposes or inadvertent contraventions of legal requirements by holidaymakers through importation of their animals. Quantification of this is by its very nature difficult, but detection of noncompliance with legislation or deliberately smuggled animals is a regular occurrence,17 and the incidence of disease often highlights this activity.

The following sections review important zoonotic diseases of pet origin and the policies currently in place to control zoonotic diseases of companion animals and their limitations, concluding with recommendations on what more could be done.

Virus group and independent virus-specific models were used to predict probabilities for species to host flaviviruses, given data on known hosts and their respective ecological traits. Known hosts for each modeled zoonotic flavivirus, along with model-based predicted probabilities are shown in Supplementary Data file 3. The predicted probabilities for species for all the models were not highly correlated with any of the geographical traits of species indicating that model predictions were not geographically biased towards any regions of human outbreaks (Supplementary Figure 11). Out of 112 predicted host species (top 5% of species) for YFV and ZIKV (Group 1), 21 were primate species. These species on an average had 20% probability (SD = 17.0) for being a host for YFV and ZIKV. Nine of the 21 primate species predicted have not yet been detected with Group 1 viruses by confirmatory tests such as virus isolation, PCR, or validated serology (PRNT). Among the nine predicted new hosts, only the Guinea baboon (Papio papio) has had YFV previously detected by serology (Supplementary Data file 3)13. Species from Rodentia (4% mean probability, SD = 6.3, n = 97) showed significantly lower probabilities for being YFV and ZIKV hosts compared to Primates (Mann–Whitney–Wilcoxon P < 0.005, Supplementary Figure 12). The geographical ranges of predicted host species for Group 1 viruses were more widely distributed than the geographical ranges of serologically-confirmed and antigenically-confirmed positive species and showed a higher host diversity in Africa, Asia, and South America (Fig. 3). Areas in North America and Europe indicate the distributions of predicted rodent species as hosts for YFV and ZIKV, but rodent species had very low probabilities, compared to primate species in other regions. When geographical distributions of species were weighted by model-predicted YFV and ZIKV host probability values, the probability map showed hotspots with high probabilities for YFV and ZIKV sylvatic hosts in South Asia and mid-eastern Africa (South Sudan, Ethiopia, Kenya, and Somalia), and a small hotspot in the eastern coast of South America (Fig. 4).

The model predicted a total of 708 novel hosts for WNV, SLEV, and USUV, which have 254 currently recognized hosts (WNV = 194 hosts, SLEV = 72 hosts, and USUV = 65 hosts). Only species from the order Charadriiformes (12% mean probability, SD = 10.0, n = 95) showed a significantly different probability of hosting Group 2 viruses than species from the order Accipitriformes (39% mean probability, SD = 31.0, n = 40, Mann–Whitney–Wilcoxon Bonferroni-adjusted P < 0.005) and species from Strigiformes (38%, SD = 29.7, n = 19, Mann–Whitney–Wilcoxon Bonferroni-adjusted P = 0.01, Supplementary Figure 12). Orders that showed high predicted mean probabilities for being hosts for WNV, SLEV, or USUV were Cathartiformes (56%, SD = 36.4, n = 4), and Columbiformes (34%, SD = 28.7, n = 13). Of non-avian species, Equus ferus (order Perissodactyla) was in the top 5% of predicted hosts, with a probability of 8% for being a host for Group 2 viruses. None of the primates species were in the top 5% of predicted species even though Macaca sylvanus is a known host for WNV14,15 and Ateles paniscus, and Sapajus apella have been detected positive for SLEV15,16 (Supplementary Data file 3). Overall, two regions, North America and central Europe, showed high species richness of predicted hosts (including correctly identified known hosts) of Group 2 viruses compared to other regions (Figs. 3, 4). When the model for WNV, SLEV, and USUV was run by relabeling species as positive only when they were found positive by PCR or virus isolation, the model predicted higher mean probabilities for species from the orders Accipitriformes, Anseriformes, Carnivora, Charadriiformes, Galliformes, Passeriformes, and Rodentia (Mann–Whitney–Wilcoxon Bonferroni-adjusted P < 0.05). The changes in the probabilities were due to changed number of Orders included in the modeling procedure as some orders dropped out from the model as they only had species positive by PRNT (Supplementary Table 2). Predicted species in the alternate model with species confirmed only by PCR or virus isolation, revealed hotspots in the same geographical region of North America (Supplementary Figure 13).

The model for TBEV predicted 494 hosts in addition to the already recognized 75 hosts for this virus (Supplementary Data file 3). Only species from the order Passeriformes (19% mean probability, SD = 27.6, n = 156) showed a significantly higher probability of hosting TBEV than species from order Charadriiformes (4% mean probability, SD = 13.0, n = 75, Mann–Whitney–Wilcoxon Bonferroni-adjusted P = 0.04, Supplementary Figure 12). Mean probabilities for other Orders were not significantly different from each other (Mann–Whitney–Wilcoxon Bonferroni-adjusted P > 0.05, Supplementary Figure 12). The geographical distribution of these species showed high species diversity clusters across Europe and Russia (Fig. 3), with similar hotspots when geographical distributions were weighted for the predicted host probabilities (Fig. 4).

Results for the Group 3 (RBV, ENTV, and DBV) model indicated that top 5% of species with the highest predicted probabilities included 42 bat species that have not yet been detected positive for Group 3 flaviviruses. Mean predicted probabilities for hosting Group 3 viruses for families within Chiroptera were not significantly different from each other (Mann–Whitney–Wilcoxon Bonferroni-adjusted P > 0.05), but leaf-nosed bat species (Phyllostomidae) had the highest mean probability of 26% (SD = 24.2, n = 10, Supplementary Figure 12). Additionally, model results showed two geographical clusters with high predicted host diversity: one extending across central sub-Saharan Africa to the southeastern coast of Africa, and the second encompassing the Neotropic region and Central Americas (Figs. 3, 4).

The model for DENV predicted 173 host species, of which 139 are new, potentially unrecognized host species. The average probability for being a host for DENV among species in orders Didelphimorphia (mean = 18%, SD = 26.4, n = 16), Chiroptera (mean = 9%, SD = 14.2, n = 75), Rodentia (mean = 6%, SD = 11.3, n = 53), Primates (mean = 5%, SD = 2.9, n = 12), and Perissodactyla (mean = 3%, SD = 0.9, n = 13,) were not statistically different from each other (Mann–Whitney–Wilcoxon Bonferroni-adjusted P > 0.05, Supplementary Figure 12). Predicted species for DENV showed high, combined host species richness in the neotropical region stretching from the Central American tropics in the north to the northern Bolivian and Paraguayan regions in the south, covering all of Brazil (Figs. 3, 4).

The JEV model predicted 408 host species, of which 388 would be new hosts. Their distribution showed high species richness in Southeast Asia, central Europe, and Australia (Figs. 3, 4). Order Cetartiodactyla (mean = 3%, SD = 7.0, n = 45,) showed significantly higher average mean predicted probabilities than Passeriformes (mean = 2%, SD = 7.0, n = 135, Mann–Whitney–Wilcoxon Bonferroni-adjusted P = 0.048) and Primates (mean = 2%, SD = 3.0, n = 12, Mann–Whitney–Wilcoxon Bonferroni-adjusted P = 0.023). Avian orders Columbiformes (mean = 5%, SD = 13.8, n = 12), mammalian orders Perissodactyla (mean = 7%, SD = 18.4, n = 8), Chiroptera (mean = 1%, SD = 2.9, n = 185), and Eulipotyphla (mean = 1%, SD = 1.1, n = 11) showed similar mean probabilities for being hosts for JEV (Mann–Whitney–Wilcoxon Bonferroni-adjusted P > 0.05 Supplementary Figure 12).

Although bat bites may be the main transmission route coming to mind, pathogen transmission involving bat bites has been documented mostly for rabies virus (Rhabdoviridae). The common vampire bat (Desmodus rotundus) can, for instance, naturally transmit rabies to other species when biting to feed on blood, particularly to livestock and sometimes to humans. Mycoplasma has also been detected in common vampire bat blood and saliva and might be transmitted between bats, for instance, during aggressive behaviors. Obligate blood-feeding bats are, however, restricted to Central and South America and represent only a very small proportion of the bat species diversity (<0.005%; 3/1,200). Most bat species do not naturally bite humans unless intentional contacts occur (e.g., veterinarian and field biologists involved in bat capture and handling, people trying to remove bats from houses).

Contact with bat body fluids (saliva, urine, and feces) is increasingly recognized as an important mechanism of pathogen spillover to humans. Human encroachment into bat habitats as well as increasing urbanization, which facilitates bat roosting in artificial structures, are likely to increase contact with bat body fluids. For example, Nipah virus (Paramyxoviridae) human infection cases reported in Bangladesh were associated with the consumption of raw sap from date palm trees contaminated with fruit bat saliva and urine. In the case of Marburg virus (Filoviridae), experimental studies indicate that bat-to-bat transmission may occur via saliva and aerosols, suggesting that the virus may be transmitted to other hosts by a similar mechanism. This hypothesis is supported by investigations revealing that most humans infected with Marburg virus had entered bat (Rousettus aegyptiacus) caves before becoming sick and reported regular contacts with bats or their secretions.

Hunting, preparation, and consumption of bats as bushmeat have also been pointed out as a potential source of infection, especially for Ebola virus. For instance, the putative first human case of the 2007 Ebola outbreak in the Democratic Republic of Congo would have bought freshly killed bats for consumption. The fruit bat Eidolon helvum, which is the most frequently hunted and traded bat species in many African countries (e.g., more than 120,000 E. helvum are sold yearly in markets in Ghana), has been shown to be infected with Henipa-related viruses. This highlights the substantial exposure of local hunters and consumers to viruses of potential zoonotic importance.

While it is often recommended that a detailed understanding of dog ecology is needed for effective canine rabies control, the consistency of research findings generated over the past 30 years allows us to be confident in concluding that mass dog vaccination is feasible across a wide range of settings and campaigns can and should be initiated without delay. In some cases, more nuanced understanding may be required to improve coverage, but these insights can be often be gained through implementation of control measures and used to progressively improve the design and delivery of subsequent interventions. Key considerations include the nature and degree of community engagement, timing of campaigns, placement of vaccination stations and whether or not to charge owner fees [62–64]. The costs of implementing campaigns free of charge may exceed those readily available to government veterinary services, but many approaches can still be explored to improve affordability, acceptability and cost-effectiveness.

While there is widespread agreement about the central importance of mass dog vaccination in canine rabies control and elimination, the role of dog population management remains the subject of debate. There is a rich literature around fertility control for management of roaming dog and wildlife populations. However, as rabies transmission varies little with dog density, reproductive control measures carried out with the aim of reducing dog density are not likely to be effective for rabies control. In theory, reducing population turnover (e.g. through improving life expectancy and/or reducing fecundity) could help sustain population immunity between campaigns and improve cost-effectiveness. However, there is little empirical evidence that dog population management tools have been able to achieve this. Furthermore, even in populations with a high turnover, achieving a 70% coverage during annual campaigns has been sufficient to sustain population immunity above critical thresholds determined by R0. The relatively high cost of sterilization also means that strategies which combine vaccination and sterilization are less cost-effective in terms of achieving human health outcomes than strategies based on dog vaccination alone, even in populations with a large proportion of roaming dogs. Improved dog population management is undoubtedly a desirable longer-term goal for animal health and welfare and may have important secondary benefits for rabies control, for example by enhancing community or political support. However, a focus on mass dog vaccination currently remains the most pragmatic and cost-effective approach to canine rabies control and elimination.

The limited availability and quality of routine animal rabies surveillance data in LMICs has been an obstacle to the application of the analytical approaches from which we have learned so much about wildlife rabies. ‘Gold standard’ surveillance data based on laboratory-confirmed diagnosis is hampered not only by limited laboratory infrastructure but also by the practical challenges of locating, sampling and submitting specimens. However, pragmatic approaches to improving rabies surveillance have yielded rich insights. In addition to providing a foundation for burden of disease estimates, data on animal-bite injuries have been a used as a reliable indicator of canine rabies incidence, revealing new understanding of rabies metapopulation dynamics, as well as improving detection of animal rabies cases, the management of animal bites and the cost-effectiveness of PEP.

Pragmatic solutions are also being found to improve rabies diagnosis in settings with limited laboratory infrastructure, including techniques to support decentralized laboratory testing (e.g. direct rapid immunohistochemical test, dRIT) [73–76] and field diagnosis (e.g. immunochromatographic tests) [77–79]. These have great potential for empowering field staff to engage in rabies surveillance and respond more effectively to surveillance data, but standardization and quality control of field diagnostic kits still needs improvement. Given the rapid advances in metagenomic sequencing methods, future approaches may include real-time genomic surveillance. However, even simple technologies such as mobile phones can serve as leapfrogging technology that can dramatically improve the extent and resolution of rabies surveillance data.