The pastoral area of the eastern Tibetan plateau, as represented by Shiqu County, has been recognized as one of the most serious echinococcosis endemic regions in the world [8, 19]. Thus, Shiqu has been listed by the Chinese government as a pilot area for the prevention and control of echinococcosis in China [20]. The increasing prevention awareness of echinococcosis in communities has resulted in a better implementation of dog population management, regular deworming and changes to better dog feeding. As an important part of the pilot project, the dog management work has significantly reduced the prevalence of canine echinococcosis since its implementation in 2015 (Table 5).
Effectiveness of the Dog Management Regulations
According to the NDRC in 2016, the population size of stray dogs must be controlled and decreased and the canine echinococcosis prevalence in dog populations should be less than 5% in endemic areas by the end of 2020 [7]. The dog management regulations were strictly implemented in Rizha Village right after the beginning of the Shiqu countywide implementation of the echinococcosis prevention and control program in November 2015 [14]. The questionnaire and copro-PCR results showed very high Echinococcus prevalence in owned dogs in Rizha in the summer of 2015, which was significantly decreased by 2017, two years after the implementation of the dog management regulations (Table 4). Unlike Rizha, a trial of the dog management regulations has been carried out in Eduoma Village since 2014, earlier than the countywide implementation, so the low Echinococcus prevalence in Eduoma detected in 2016 and 2017 (Table 4) was not unexpected. The overall Echinococcus prevalence in owned dogs in Eduoma was less than 5% in 2017 (Table 4), which met the standard defined by NDRC (2016) [7] and an obvious decreasing trend in Echinococcus prevalence in the dog population was confirmed by our surveillance data.
The importance of dog population control has been studied in detail by many reports [19, 21]. Although the regulations of the dog management program have been implemented in Shiqu County for many years [22, 23], it needs time to cover all the remote areas of the county. For example, numbers of owned dogs per household of the two visited villages were less than two on average and below the number permitted by the dog management regulations and were not different between villages and between sampling years. However, stray dog populations were different (Table 2). In Eduoma Village, the dog population has been strictly controlled since 2014, where unleashed dogs were rare during the entire sampling period of this research (Table 2), but in the more remote Rizha Village, unleashed dogs were still popular in 2015, where the majority of the unleashed dog feces and unleashed dogs in this study were recorded (Table 2). Numbers of unleashed dog feces were significantly decreased and unleashed dogs were not observed in 2017 (Table 2). Although unleashed dog feces might also have come from free-moving owned dogs, judging by the high rate of awareness of echinococcosis control and prevention in local people (Table 3) and the synchronization between the reduction numbers of unleashed dogs and their feces collected, a better control of the unowned stray dog population size in local areas of the Shiqu County can be confirmed.
In general, the research data showed that the dog management measures did significantly reduce Echinococcus species prevalence in local populations of dogs in Shiqu County and are effective in echinococcosis control and prevention on the pastoral area of the Tibetan Plateau.
Factors influencing Echinococcus Prevalence in Owned Dog Populations
All of the three Echinococcus species discovered in China were detected infecting dogs in our study (Table 4). Echinococcus granulosus and E. multilocularis are the two Echinococcus species confirmed as infecting humans. The prevalence of E. multilocularis was significantly higher than E. granulosus in dogs in our study (Table 4) and this was similar to conclusions by Budke et al. [11] and Moss et al. [24]. The prevalence of E. shiquicus could be as high as E. multilocularis in dogs (Table 4), which further supported the possibility of dogs as a viable definitive host species of E. shiquicus, as suggested by Boufana et al. [13]. Although no transmission to humans has been reported, E. shiquicus shared sylvatic transmission cycles with E. multilocularis between canid and small mammal species [13, 25, 26]. Echinococcus multilocularis was the main Echinococcus species and the only species continuously detected in dogs, especially in owned dogs in the two visited villages in all sampling years (Table 4). Compared with preventing owned dogs from becoming infected with E. granulosus by ingesting viscera of large livestock, it would seem more difficult to stop the trophic connection between owned dogs and small mammals to prevent infection with E. multilocularis. Factors influencing the Echinococcus prevalence in owned dogs are valuable knowledge for dog management to benefit the control and prevention of echinococcosis in humans in the pastoral areas of the Tibetan plateau.
Dog Roaming Behavior Restriction in Villages
Free roaming has been judged as a significant risk for owned dogs becoming infected with E. multilocularis and E. shiquicus, but not E. granulosus in Shiqu County because of their high chances of contact with and preying on wild intermediate host animals [11, 24]. Free dogs can have active spatial activities around villages, as shown by Vaniscotte et al. [12] who reported that a released dog could move 1500 m away from the village with an average activity area of 77 ± 59.4 ha. Such an active spatial behavior pattern enables a free roaming dog to visit areas where wild small mammal intermediate host species are distributed. The average worm lifespan of E. granulosus and E. multilocularis can be ten and five months respectively [1], so theoretically preventing contact with intermediate hosts, deworming an infected dog and restraining roaming behavior are judged as effective methods to control Echinococcus prevalence in owned dogs [7]. Because almost all households from the two villages claimed to leash their dogs 24 hours per day according to questionnaire results (Table 3), there were not enough negative samples to result in the dog roaming behavior being judged as an insignificant variable by the logistic regression analysis. The questionnaire results also suggested that, as an important and basic part of the public echinococcosis prevention education, leashing dogs has been generally accepted and followed by local people.
However, the fact that infection of E. multilocularis was continually detected in owned dogs (Table 4) implied that owned dogs still have chances to come into contact with wild small mammal host species. Small mammal bones were detected in feces of a few owned dogs and a minority of households released dogs at night (Table 3). Even if local people follow the dog roaming behavior control regulation strictly, leashed dogs may also be able to prey on small mammals. At least six widespread small mammal species, mainly voles and pikas were identified as intermediate hosts of E. multilocularis [19] and the prevalence in voles was significantly higher than that in pikas [22, 30, 31]. Mu [29] confirmed that the population density of small mammals especially vole species can be high as in the field less than 500 m away from Rizha Village. In fact, evidence of small mammal presence could be as near as 32 m away from households in villages of Shiqu County [11]. Therefore infected small mammals may have the opportunity to access the villages, which may provide leashed dogs opportunities to prey on them. Moreover, when herding on the summer pasture, owned dogs are usually unleashed all the time and these dogs could be infected by preying on small mammals before they come back to the village. Therefore, although restraining dogs is considered a fundamental measure to decrease Echinococcus prevalence in owned dogs, proactive measures such as regular dosing with praziquantel are still needed.
Regular Dog Deworming
Regular supervised dog praziquantel dosing has been considered to be the pivotal measure for echinococcosis control and prevention in the pastoral areas of the Tibetan plateau [6], starting in northwest Sichuan Province in 2006 [32]. The logistic regression model revealed that monthly dosing was significantly more powerful than irregular or no dosing to decrease Echinococcus prevalence in owned dogs (Table 5). Protoscoleces of E. granulosus s.s. and E. multilocularis usually need four to six weeks to develop to adult tapeworms after infection [1], so monthly dog deworming has been adopted as the most important control measure.
However, application of this regulation in remote settled and semi-nomadic communities is challenging. In the more remote Rizha village, none of the owned dogs surveyed were monthly dosed in 2015 and one household stated that they had not received any praziquantel for more than half a year in 2017 (Table 3). Because of the obvious difficulty in seasonal traffic restrictions, communication and the highly mobile semi-nomadic living styles of local people despite permanent settlements being provided, administration of monthly dosing in remote communities is still difficult to enforce in all families. Although monthly dosing regulation can be effectively supervised in settlements as demonstrated by our data in 2016 and 2017 (Table 3), the effectiveness of monthly dog dosing cannot be enforced in summer pasture areas where nomadic families scatter on the vast grassland freely and supervised dosing is unrealistic. This implies the extreme importance of long term supervised monthly dosing of owned dogs especially when all the semi-nomadic families gather in settlements in villages from late September to late May the next year.
Long term supervised dosing programs can be extremely costly and resource demanding. The several successful application of regular dog praziquantel dosing were usually about E. granulosus control in more developed agricultural areas [33, 34, 35], where E. granulosus is mainly transmitted between large herbivorous livestock and dogs, so coupled with livestock slaughter and viscera management, long term dog dosing can result in a significant impact to the transmission cycle [36]. With regards to the more detrimental E. multilocularis in the eastern Tibetan plateau, more complex wildlife transmitting cycles are present involving wild canids such as the Tibetan fox (Vulpe ferrilata) and the red fox (V. vulpes) [16] and small mammals, such as pika and small rodent species [26]. The large populations, wide distribution of small mammals [26, 28, 37], and the potential contact with dogs even tied up 24 hours per day, suggest that the possibility of E. multilocularis spreading from the wildlife reservoir to the human environment always exists (Table 4). Once a regular dosing program stops, Echinococcus prevalence in dogs can return to pretreatment levels in less than ten months [24], so in order to keep up the long term regular dog dosing program to cover the vast western pastoral areas of China going, He et al. [30] suggested decreasing the dosing frequency from once per month to once per every two or three months as recommended by WHO [33] to balance the expenditure and deworming effect. Their recommendation refers to dog re-infection studies of E. granulosus and not enough empirical data from E. multilocularis re-infection in dogs is available yet. Our study suggested stopping dosing for several months can significantly decrease the power of the praziquantel dosing. Because of its shorter prepatent infection period and more complex transmission cycles compared with E. granulosus, whether the optimization of the supervised regular dosing regulations for E. multilocularis with the consideration of both medical effects and economic and managing feasibility is still open to reasoning.
A single dose of praziquantel is recommended to be 5 mg/kg for animals [1]. At present, the one-dog-one-pill (0.1 g praziquantel) dosage per month neglects individual weight difference among dogs. Local people frequently expressed their worry about the negative side effect of the medicine to dogs during surveys in villages. More effective and safe medicines like slow release praziquantel [39] may be a better choice for future large scale implementation.
Dog Keeping Traditions
Dogs being male and older were two significant dog keeping factors associated with higher Echinococcus spp. infection [11]. The significant effect of being male dogs was detected by the logistic regression model in this study (Table 5). Traditionally, male dogs (figure 2) are preferred by nomadic people for better property and livestock protection and easier dog population management. Since the development of new settlements and the cooperative pasturing in nomadic local communities, people no longer need so many dogs as before. Controlling breeding activities by only keeping male dogs is usually one of the most feasible methods for remote and developing areas [21], so the proportion of male dogs is significantly higher than female dogs in local communities of Shiqu County [24, 40] ( Fig. 2; Table 3). Compared with female dogs, male dogs are more likely to maintain territories and hunt, increasing the chances of infection. However, male territorial aggression and hunting are only important when dogs are unleashed. Therefore, if dogs were tied up well as reported by the most visited households, the fact that all canine echinococcosis infections were detected in male dogs in this study should be the result of male dogs being in the majority (Table 5) but not a significant infection risk. As to the age bias, the infection burden of E. granulosus could be significantly higher in dogs over five years old, but not significant for E. multilocularis infection [11, 19]. In our study, the fact that most of sampled dogs were less than five years old (Fig. 2) and E. multilocularis was the main Echinococcus species but not E. granulosus (Table 5) might explain the insignificant effect of the dog age.
Not feeding dogs with livestock viscera was frequently recommended as one of the most effective methods to control domestic dogs infecting E. granulosus (Schantz et al., 2003; Wang et al., 2014; Yuan et al., 2017). Since we did not detect E. granulosus infection from viscera fed household fecal samples, the importance of dog feeding habits cannot be evaluated directly by the logistic regression model analysis. Nevertheless, the importance of not feeding dogs with livestock viscera is still significant in this study. In fact, most visited households reported being aware of and did not feed dogs with viscera (Table 3), and the prevalence of E. granuluosus decreased dramatically and was not detected in both villages in the last sampling year in this study (Table 4). All of these results suggested that the regulation to stop feeding dogs with livestock viscera has been well proceeded in local Tibetan communities and received expected effect.