Canine distemper in Nepal's Annapurna Conservation Area – Implications of dog husbandry and human behaviour for wildlife disease

Dogs are often commensal with human settlements. In areas where settlements are adjacent to wildlife habitat, the management of dogs can affect risk of spillover of disease to wildlife. We assess dog husbandry practices, and measure the prevalence of Canine Distemper Virus (CDV) in dogs, in 10 villages in Nepal’s Annapurna Conservation Area (ACA), an important region for Himalayan wildlife. A high proportion (58%) of owned dogs were allowed by their owners to roam freely, and many village dogs originated from urban areas outside the region. CDV antibodies, indicating past exposure, were detected in 70% of dogs, and 13% were positive for P-gene, suggesting current circulation of CDV. This is the first detection of canine distemper virus in a National Park in Nepal Himalaya. Dogs were generally in good condition, and none exhibited clinical signs of CDV infection, which suggests that infections were asymptomatic. CDV exposure varied with village location and age of dogs, but this variation was minor, consistent with high rates of movement of dogs across the region maintaining high seroprevalence. Residents reported the occurrence of several species of wild carnivores in or close to villages. These results suggest a high potential for transmission of CDV from village dogs to wild carnivores in ACA. We suggest that control of dog immigration, along with vaccination and neutering of dogs could mitigate the risk of CDV spillover into wild carnivore populations.

6 121 122 We approached every dog owning household (DOHH) across the 10 villages. Permission to 123 participate in the study was sought from the dog's owner. DOHH were identified through 124 information supplied by the mayor of each village. Only dogs with owners who agreed to 125 respond to the questionnaire were included in the serological analyses. One dog owner in 126 Manang refused to respond to the questionnaire and was thus excluded from the study. Two 127 dogs from separate owners, also in Manang, were excluded from the study as both dogs were 128 old and did not respond well to attempts to collect blood. To avoid further distress to the dog 129 owners and to ensure welfare of the dogs, attempts at blood collection were halted after a few 130 tries. The owners of both dogs responded to the questionnaire but these responses were 131 excluded from analyses due of an absence of serological data. A total of 76 households were 132 interviewed, representing 12.8% of the total number of households in the study area and 144 Participants were asked to report the births and mortality of dogs in the household in the past 7 146 behaviour; sterilisation status; vaccination history; age; sex; and source/origin . As a measure   147 of care afforded to dogs, owners were asked if their dog had visited a veterinarian or been   148 dewormed within the past 12 months.   149 150 Interviewees asked to report livestock that had been killed by wildlife in the past 12 months, 151 and to identify, by pointing to photographs (S1 File) supplied with the questionnaire, native   (Table 3). This is consistent with findings from a study that investigated the diet and prey selection of snow 362 leopards in the same area (48). One dog owner had lost 35 goats to attacks from snow leopards, foxes, jackals and wolves. Another indicated that he had lost over 100 goats to 364 various predators. One rancher reported giving up ranching goats as he was unable to cope with the financial loss to predators. Respondents explained that the type of predator was 366 determined by assessing bite marks and scratch marks on prey. They described that snow leopards would bite the necks of prey while canids, like jackal and wolves, would attack the Table 3. Variety and number of livestock that were reported killed by wild predators 372 and domestic dogs within a 12-month period. Numbers reflect reported livestock deaths associated with the corresponding predator. One rancher reported losing a total of 35 goats to 374 a four different predators, but was unable to separate the number of livestock deaths associated with each predator. His report has been excluded from the table below. Also 376 excluded from the table is the report by one rancher who lost more than 100 goats. The "Unknown" category describes reported livestock deaths associated with a predator attack of 378 indiscernible identity.  The finding that most dogs were free-roaming was expected and was common to studies 396 conducted in similarly rural sites in developing countries (49-52). These free-roaming dogs were also observed moving in packs both within and between villages, and owners reported 398 that it was customary for their dogs to accompany them on journeys either on foot or by vehicle to adjacent villages for work or social visits. Owners reported sourcing their dogs 400 from locations as far away as Kathmandu (Fig 3). Manang village had the largest proportion of dogs that were sourced from outside of the study area. A majority of the dogs in Pisang 402 were sourced from within the village, with some dogs from Kathmandu. Despite the relatively small population of dogs in Bhraka, the village was a source of dogs in Pisang and 404 Ngarwal. Similarly, Ngarwal was a source of dogs for villagers in Manang, Bhraka, and villages external to the study area. This human-assisted movement of dogs between urban and 406 rural locations, and the local movement of dogs due to the unrestrained culture of dog keeping, results in a high degree of mixing. The high movement of dogs suggests that the 408 dogs in the study area are effectively a single population. This would further explain why household-and village-level factors had no effect on abundance: local availability of food 410 resources would not limit dogs able to roam freely to access resources from adjacent villages.

Predator
That over a quarter of dogs in the study population originated from locations outside the 412 study area (i.e. Chame, Tanchok, Timang, Sapche, Nar, Ghorka and Kathmandu) is important information for design of management interventions to limit the size of this dog population or 414 manage health of dogs and disease risks to wildlife.
416 The age and sex of dogs was strongly skewed toward adults and males, respectively. The male-biased sex proportion was higher than the mean (62.6 ± 11.5) described by a meta-developing countries (6), and was consistent with other studies of rural dog populations 420 (42,(53)(54)(55)(56). The adult-skewed age ratio was also consistent with findings from the metaanalysis by Gompper (6).

422
The high proportion of dogs in ideal body condition recorded in this study was contrary to 424 expectations for dogs in a developing country (57). Sterilisation rates were also unusually high for a rural area and was well above the mean (11.6%) and median (6.2%) rates described 426 by Gompper (6) for dogs generally, in which all rates above 10% were from studies in 452 Residents reported observing direct interactions between dogs and wild carnivores, suggesting potential for pathogen transmission between dogs and susceptible wild carnivores. Given the 454 high prevalence of CDV in dogs, it is conceivable that spillover might already have occurred.
In the Annapurna Conservation Area, where wildlife conservation is an objective, the 456 management of domestic dogs should be considered when developing and implementing wildlife conservation plans. There is also the risk of hybridisation between dogs and wild 458 canids (66,67). To address this, measures to promote vaccination and limit number of dogs per household and their roaming behaviour could be implemented. It would also be valuable for 460 further research to investigate CDV exposure in sympatric wild carnivores.

Conclusion
There is high CDV prevalence among dogs in Nepal's Annapurna Conservation Area, and 464 spillover of disease from dogs to native wild carnivores is a potential threat to susceptible species. An outbreak among wild carnivores in the area, which includes endangered species, objectives. A mass vaccination of domestic dogs could avoid the impact of an outbreak to the 468 local economy, safeguard the wellbeing of dog owners, and buttress conservation objectives.