Chinese Medical Journal 2007;120(3):237-242
Impact of overgrazing on the transmission of Echinococcus multilocularis in Tibetan pastoral communities of Sichuan Province, China

Correspondence to:WANG Qian,Sichuan Centre for Diseases Control and Prevention, Chengdu 610041, China (Tel: 86-28-85589512. Fax:.
echinococcosis, hepatic; Echinococcus multilocularis; overgrazing; Tibetan pastoral communities

Background  Overgrazing was assumed to increase the population density of small mammals that are the intermediate hosts of Echinococcus multilocularis, the pathogen of alveolar echinococcosis in the Qinghai Tibet Plateau. This research tested the hypothesis that overgrazing might promote Echinococcus multilocularis transmission through increasing populations of small mammal, intermediate hosts in Tibetan pastoral communities.
Methods  Grazing practices, small mammal indices and dog Echinococcus multilocularis infection data were collected to analyze the relation between overgrazing and Echinococcus multilocularis transmission using nonparametric tests and multiple stepwise logistic regression.
Results  In the investigated area, raising livestock was a key industry. The communal pastures existed and the available forage was deficient for grazing. Open (common) pastures were overgrazed and had higher burrow density of small mammals compared with neighboring fenced (private) pastures; this high overgrazing pressure on the open pastures measured by neighboring fenced area led to higher burrow density of small mammals in open pastures. The median burrow density of small mammals in open pastures was independently associated with nearby canine Echinococcus multilocularis infection (P=0.003, OR=1.048).  
Conclusion  Overgrazing may promote the transmission of Echinococcus multilocularis through increasing the population density of small mammals.

Human alveolar echinococcosis (AE) is an infection caused by Echinococcus multilocularis (E. multilocularis), a highly pathogenic zoonotic cestode with a life cycle involving foxes as definitive hosts (hosts of the adult stage) and small mammals as intermediate hosts (hosts of the larval stage, or metacestode). Human AE, albeit restricted to localized endemic areas, is the public health concern in central Europe.1 It is a major public health problem in western China2,3 where dogs act as the definitive hosts.4,5 Mass ultrasound surveys in Tibetan communities of Sichuan Province in 1997-1998 documented a high prevalence of AE in humans, averaging 1.9% (77/3998) in the population, and 2.3% (43/1858) in the communities raising livestock, 97.0% of whom were Tibetans.4

Two researches in this area showed that pasture type might affect E. multilocularis transmission6 and the extent of fenced pastures was positively related to human AE prevalence.7 Both assumed that grazing practice might be the underline reason increasing the density of small mammal intermediate hosts of parasite in open pastures, consequently higher level of infection in community dogs thus higher contamination pressure.

However, the relationship between grazing and small mammal populations is complex and remains poorly understood. Livestock grazing was reported to affect small mammal populations negatively in both richness and abundance in sagebrush rangeland of Idaho, USA8 and in South Africa.9 In contrast, populations of small mammals were shown to increase on grazed sites in the Thar Desert, India.10 In the Great Salt Desert (USA),

heavy grazing increased the population densities of some species, while the densities of the others decreased.11,12 On the Qinghai Tibet Plateau, some have suggested that when combined grazing of yak, sheep and horses lowered the height of vegetation, plateau pika (Ochotona curzoniae, O. curzoniae) might be found at greater densities than on the natural meadows.13,14 It was assumed that overgrazing led to outbreaks of small mammals including plateau pika, mole rats (Myospalax baileyi) and voles (Microtus fuscus, Microtus irene).15,16 However, it has also been argued that the lowered grass could be the result of Ochotona curzoniae invasion.17,18 In this and other areas, the relationship between overgrazing and population density of small mammals is thus not clearly understood.

Since the 1980s, partial fencing of pastures around Tibetan pastoral winter settlements has become common as a result of changes in land property regulations.19 Our preliminary field observation revealed that fenced pastures were privately owned and primarily used to provide forage for weak, young or pregnant livestock during the winter and early spring when forage is extremely limited. This may lead to significantly less livestock grazing pressure on the privately owned, fenced areas than in community owned open pastures. In addition, partial fencing reduces the area of community owned open pastures; this may increase grazing pressure on the open pastures. Thus, it might be hypothesized that proportionately larger fenced (private) areas create higher grazing pressure on the open (communal) pastures. The existence of both fenced (private) and open (communal) pastures in an otherwise homogeneous environment provided us with a unique opportunity to investigate the effect of overgrazing on the abundance of small mammals and its potential consequence to contamination pressure by infectious agents that depend on small mammals as reservoir host. This research was designed to study the livestock management of open and fenced pastures, the availability of forage, to compare the burrow density of small mammals in open and fenced pastures, and to find their relationship with E. multilocularis in winter settlements.


Site description
The study was carried out in three townships of Shiqu County (Ganzi Tibetan Autonomous Prefecture in northwest Sichuan Province, China) in spring and autumn 2003. The selection of the investigation sites was based on the documented high prevalence of AE in humans by mass screening and the observation of a variety of fencing practices in the area.7

Shiqu County is located on the eastern Qinghai Tibet Plateau at 97˚20’00’’-99˚15’28’’E and 32˚19’28’’-34˚20’4’’N, has a population of 62 000 (98% ethnically Tibetan) and shares a border with Qinghai Province in the east, north and west and with the Tibet Autonomous Region in the south. It covers an area of 25 100 km2, at a mean elevation of 4200 m, 19 000 km2 of which can be used as grazing pasture. The weather affected by monsoon climate, with a mean yearly rainfall of 596 mm, is characterized by wide temperature differences (average temperature -1.6 C˚). Winter is longer than summer and frost conditions persist throughout the year.19 In 1996, the livestock population was about 630 000 and consisted of yaks, sheep, goats, horses and a small number of pigs.

Data collection
The study was approved by the Ethical Committee of Sichuan Institute of Parasitic Diseases, Sichuan Province, as well as those of all collaborating investigators. The investigation was performed in 30 settlements of 3 Tibetan pastoral townships (18 villages). If a settlement was located in a narrow valley (less than 1 km width), the landscape was described as “valley”; in the entrance to a valley, it was described as “valley entrance”; at the foot of a hill and facing a flat land on the other side, it was described as “piedmont”; in a wide valley (more than 1 km width of level land), it was described as “flat land”.

A questionnaire was administered to the herdsmen after obtaining individual informed consent. In small settlements (less than 10 households), all households with herders at home were investigated. In larger settlements, a random selection of 10 households in settlements with less than 20 households, or half the households in settlements with more than 20 households was investigated. The questionnaire covered demographic information, number of livestock, livestock management practices and herdsmen’s perception of the discrepancy between needed forage and the productivity of the winter pasture.

For each settlement, the area of fenced pastures was measured using a GPS (GPS 12, Garmin International Inc., USA). Small mammal populations were monitored using index methods. These methods are based on the detection of surface indicators of small mammals, i.e. holes and faeces,20-23 from which population densities are estimated.24,25 Sampling was performed by 2 investigators walking at approximately 1 km per 1.5 hours along a 1 km transect drawn across each settlement. Along this transect, 50 counts of small mammal burrows were performed; each count covered an area 20 metres long and 10 metres wide (about 200 m2). For valley, valley entrance and piedmont settlements, transects began at a point 20 metres to the right of the first house in the direction of entrance into the settlement. If it was impossible to walk in this direction (for example, facing a very steep hill), the starting point changed to the left of the first house. For settlements on flatland, transects went north beginning from a point 20 metres away from the northernmost house of the settlement.

In settlements where transects were done in 2 townships, all dogs were sampled through arecoline purgation, according to the recommendations of OIE/WHO,26 and/or collecting ground faeces when accessible and available.5 Purgation and faecal samples were taken to the Sichuan Institute of Parasitic Diseases (SIPD) in Chengdu, where helminths were removed, counted and placed in 10% methanal saline or 85% ethanol. Species identification of worms was done at SIPD and copro-PCR testing at the School of Environmental and Life Sciences, University of Salford (UK), using species-specific primers for E. multilocularis DNA amplification based on the method of Dinkel et al27 as modified by van der Giessen et al.28

Statistical analysis
Data related to livestock management, perception of the gap between the forage needs of livestock and the production capacity of the pastures, the density of small mammal burrows, the area of fenced pastures and prevalence of canine infection in settlements were used in the analysis. Landscape was taken into consideration for the analysis because the productivity could vary among pastures with different landscapes.

The burrow density of small mammals outside the fenced areas was compared with that within the fenced areas using nonparametric tests considering landscape factor. Spearman correlation tests were used to quantify the statistical relationship between the density of small mammals outside fenced areas and the surface of fenced areas in settlements where the fenced areas were all measured, also controlling the landscape factor. Multiple stepwise logistic regression was used to relate median density of small mammal burrows to canine infection in the settlements. The dependent variable was E. multilocularis negative (0) or not (1). Independent variables included dog age and sex, ground collected faecal samples, purged faecal samples as well as median areal density of small mammal burrows. The criterion for statistical significance was 0.05 and all these analyses were done by SPSS10.0.


The Qiwu, Yiniu and Xiazha townships (populations 2238, 2515, and 2471, respectively) covered 1046 km2, 955 km2, and 834 km2 respectively. Herders from 128 households were interviewed and a total of 30 km of transect were walked over 30 settlements. All areas of fenced pastures were measured in 22 settlements and purged faeces or/and ground faeces were collected from 252 dogs in 15 settlements in Yiniu and Xiazha townships.

Livestock management and pasture ownership
In these Tibetan pastoralist communities, almost all the population was engaged in raising livestock. The distributions of numbers of person, yak, sheep or goat and horse per household were not Gaussian (P<0.05). The medians were 4.0, 23.5, 0, and 1.0 respectively. Among the 128 households, 92 (71.9%) owned fenced pastures, 2 rented fenced pastures, 94 grazed their livestock inside the fenced area in wintertime. Usually, by May, the herders had moved their livestock to summer pastures. Most herdsmen (87.2%, 82/94) only allowed young, old, sick and pregnant livestock to graze in the fenced pastures, while the rest permitted all their livestock to graze inside the fenced pastures. Most of the investigated herders (69.5% 89/128) reported that available forage could not satisfy the needs of their livestock in winter pasture; 27.3% (35/128) thought that the available forage was extremely deficient during winter months while 3.9% (5/128) had no idea on this issue.

Characteristics of the settlements
The median number of households in a settlement was 8. The smallest settlement had 3 households and the largest had 25. The average surface of fenced pastures per settlement, calculated from data of 22 settlements where the surface of all fenced pastures could be measured, was 49 900 m2. The counting of small mammal burrows based on 1500 observations was done along 30-km transect. Some fenced pastures (13%, 24 of 186 measured fenced pastures) were also shared pastures. However, only 5 among these 24 fenced pastures were owned collectively by 4 to 7 households. One transect inside the fenced area was done in a shared fenced pasture that was owned by 4 households and 2 of the 3 owners surveyed indicated that all their livestock were permitted to graze inside this fenced area from February/March to May. This fenced area was actually used communally, and consequently overgrazed. Therefore, the transect data in this settlement was not included in the following analyses.

Distribution of small mammal burrows
The distribution for small mammal burrows was highly skewed. Kolmogorov-Smirnov test indicated that in fenced pastures (P<0.001) and outside fenced pastures (P<0.001), the data did not fit a Gaussian distribution even after Box-Cox transformation.

Landscape type influenced the abundance of small mammal burrows (P<0.001). Post-hoc Tukey multiple comparison test on ranks confirmed that the densities of small mammal burrows were different among different landscape types (P<0.05, except for the comparison between flatland and piedmont, P>0.05). The density of small mammal burrows in open (common overgrazed) pasture was greater than that in fenced (private nonovergrazed) areas (P<0.001). The densities of small mammal burrows in open pastures were significantly higher than these in fenced pastures in 3 of all 4 landscapes (Table 1).

view in a new window

Table 1. Densities of small mammal in open and fenced pastures stratified by landscapes

The Spearman correlations found that the bigger fenced areas led to higher burrow density of small mammals in the open pasture in all landscape types, namely valley (rs=0.382, P<0.001), flatland (rs=0.312, P<0.001), piedmont (rs=0.471, P<0.001) and valley entrance (rs=0.296, P=0.001, Table 2). The relationship between the area of fenced pastures and the density of small mammal burrows inside the fenced pastures in the 4 landscapes, was not significant except for flatland (valley (rs=-0.079, P=0.322), flatland (rs=-0.458, P=0.024), piedmont (statistics not applicable due to 3 observations only) and valley entrance (rs=0.081, P=0.736)).

view in a new window

Table 2. Relationship between the areas of fenced pastures and the densities of small mammal burrows in open pastures

Relationship between median densities of small mammal burrows in open pastures and E. multilocularis infections in dogs
Of 252 dogs sampled for faeces 159 (63.1%) were purged in 15 settlements. Based on copro-PCR testing, the dog E. multilocularis infection rate was calculated to be 16.7% (42/252) for the 252 dogs (183 males/252; mean age 4.41 years); the infection rate for purged samples was 18.2% (29/159); and the figure for ground collected faecal samples was 14.0% (13/93). Multiple forward conditional stepwise logistic regressions showed that the median density of small mammal burrows in the open pastures was significantly positively related to dog infection (P=0.003, OR=1.048). Dog age, sex and the nature of the sample, i.e. ground sample vs. purged sample, were not significant.


Overgrazing, increased densities of small mammal populations in open pastures and their association with canine infection were the missing link that was needed to confirm our hypothesis of a relationship between overgrazing and the transmission of E. mulitilocularis, human AE suggested in previous studies.7

Overgrazing on small mammal population dynamics has not been adequately documented. Grazing and overgrazing have different consequences and allow us to understand better the transmission of small mammal borne, diseases to humans. Overgrazing is a problem that may be caused by human behaviour and the arrangement of property rights of pastures. From an ecological point of view, overgrazing is a process caused by herbivores leading to a continuous decrease of pasture productivity over time. According to Hardin,29 overgrazing is the result of overstocking of communal pastures and occurs under two conditions: 1) raising livestock is a key industry in the community where pastures are a scarce resource; 2) the existence of communal pastures. The Tibetan pastoralist communities investigated had both features. Considering that current practice of most herders permitted only weak livestock to graze in the fenced pasture in winter and early spring, one can assume that the fenced pastures were generally well protected and not overgrazed. In contrast, open pastures experienced unlimited grazing by all livestock so that vegetation was often severely degraded. Hence, the comparison between open winter pastures and fenced winter pastures was actually a comparison between overgrazed and nonovergrazed pastures.

On the eastern Tibetan plateau, the plateau pika, O. curzoniae was the most commonly observed small mammal and this observation was also supported by other surveys.15,19,30 These small lagomorphs live in large colonies and their burrows are easily seen as they occur mostly in open pastures with low vegetation cover and are noticeably associated with degraded pastures, often leading to bare ground. It has been estimated that one hectare of Tibetan alpine pasture corresponded to 2600 pika burrows, equivalent to 273 pikas.31 In this area, however, there were several potential intermediate hosts for E. multilocularis, including O. curzoniae, O. cansus, Cricetulus kamensis, Microtus leucurus, Microtus irene and Microtus oeconomus/limnophilus.30 Biases could have been introduced when counting the burrows of small mammals inside fenced pasture because of the high grass resulting from the lower grazing pressure. In our survey, fenced pastures were located on the slopes, bogs, flatlands, and riverbanks and the species of small mammals within the fenced pastures included voles of Microtus spp and pika. In fenced grassland with high grass, the most common small mammal, however, was Microtus oeconomus/limnophilus whose burrows are typically small and round (about 3 cm diameter).30 The grass around the burrows of Microtus spp was obviously lower and conspicuous runways were also evident. The mole rat, Myospalax baileyi, was also reported to be a prominent pasture resident on the Tibetan plateau.15,16 However, we found only a few burrows of this rodent which builds large molehills (diameter: about 25 cm) surrounded by short vegetation. Among the 6 species identified in this area, several species were trapped within similar habitats (O. curzoniae, O. cansus and Cricetulus kamensis in overgrazed pastures),30 making the use of small mammal surface indices particularly important. The burrow index should be taken as a crude estimate of the relative population densities of the small mammal communities (likely including the genera Ochotona, Microtus and Cricetulus, depending on the habitats).

The current study revealed that a greater proportion of fencing was associated with a higher density of small mammal, burrows and the higher median density of small mammals was linked to a significantly higher prevalence of E. multilocularis copro-PCR positive dogs in the Tibetan communities investigated. The current study supports the hypothesis that partial fencing leads to increased populations of susceptible small mammal species in open pastures, especially the pika O. curzoniae and Mofuscus, which is a recognised intermediate host of E. multilocularis on the eastern Tibetan Plateau.3,32 A previous study indicated that E. multilocularis prevalence in O. curzoniae was 5.6% (13/233) in the same study area.32 The population density of small mammals seems to be one of the key factors for the transmission of E. multilocularis in other studies because high densities were found quantitatively associated with infection of red foxes and human AE in France and in Gansu Province (China).33-36 Our data support these observations in a Tibetan pastoral area, where dogs are probably an important definitive host of E. multilocularis.4,5 Thus, we consider overgrazing on the eastern Tibetan plateau a risk for increased exposure to human AE infection amongst pastoral groups.


1. Kern P, Bardonnet K, Renner E, Auer H, Pawlowski Z, Ammann RW, et al. European echinococcosis registry: human alveolar echinococcosis, Europe, 1982-2000. Emerg Infect Dis 2003; 9: 343-349.

2. Ito A, Urbani C, Qiu JM, Vuitton DA, Qiu D, Heath DD, et al. Control of echinococcosis and cysticercosis: a public health challenge to international cooperation in PR China. Acta Tropica 2003; 86: 3-17.

3. Vuitton DA, Zhou H, Bresson-Hadni S, Wang Q, Piarroux M, Raoul F, et al. Epidemiology of alveolar echinococcosis with particular reference to China and Europe. Parasitology 2003; 127: 87-107.

4. Wang Q, Qiu JM, Schantz P, He JG, Ito A, Liu FJ. Risk factors for development of human hydatidosis among people whose family raising livestock in western Sichuan Province, China. Chin J Parasite Dis Parasitol 2001; 19: 289-293.

5. Budke CM, Campos-Ponce M, Wang Q, Torgerson PR. A canine purgation study and risk factor analysis for echinococcosis in a high endemic region of the Tibetan plateau. Vet Parasitol 2007; 127: 43-49.

6. Wang Q, Vuitton DA, Xiao YF, Budke CM, Campos-Ponce M. Pasture types and Echinococcus multilocularis, Tibetan Communities. Emerg Infect Dis 2006; 12: 1008-1009.

7. Wang Q, Vuitton DA, Qiu JM, Giraudoux P, Xiao YF, Schantz PM, et al. Fenced pasture: a possible risk factor for human alveolar echinococcosis in Tibetan pastoralist communities of Sichuan, China. Acta Tropica 2004; 90: 285-293.

8. Reynolds TD, Trost CH. The response of native vertebrate populations to crested wheatgrass planting and grazing by sheep. J Range Manage 1980; 33: 122-125.

9. Eccar JA, Walther RB, Milton SJ. How livestock grazing affects vegetation structures and small mammal distribution in the semi-arid Karoo. J Arid Environments 2000; 46: 103-106.

10. Wada N, Narita K, Kumar S, Furukawa A. Impact of overgrazing on seed predation by rodents in the Thar Desert, northwestern India. Ecological Res 1995; 10: 217-221.

11. Jones AL, Longland WS. Effects of cattle grazing on salt desert rodent communities. Am Midl Nat 1999; 141: 1-11.

12. Price MV, Brown JH. Patterns of morphology and resource use in North American desert rodent communities. Great Basin Nat Mem 1983; 7: 117-134.

13. Shi Y. On the influence of rangeland vegetation on the density of plateau pika (Ochotona curzoniae). Acta Theriologica Sin (Chin) 1983; 3: 181-187.

14. Zhang Y, Fan N, Wang Q, Jing Z. The changing ecological process of rodent communities during rodent pest managements in alpine meadow. Acta Theriologica Sin (Chin) 1998; 18: 137-143.

15. Li FM. Out of control of the rodents in Shiqu county and handling strategies. Sichuan Grassland (Chin) 1995; 3: 27-30.

16. Hou XM. The current situation of rodents and its control methods in the resource area of Qingnan pasture. Sichuan Grassland (Chin) 2001; 1: 28-31.

17. Smith AT, Foggin JM. The plateau pika (Ochotona curzoniae) as is keystone species for biodiversity on the Tibetan plateau. Anim Conserv 1999; 2: 235-240.

18. Cincotta R, Zhang Y, Zhou X. Transhumant alpine pastoralism in northeastern Qinghai Province: an evaluation of livestock population response during China’s agrarian reform. Nomad People 1992; 30: 3-25.

19. Editorial commission of Shiqu County record. Shiqu county record. Chengdu: The People’s Publication House of Sichuan Province; 2000: 55-137.

20. Fichet-Calvet E, Jomaa I, Giraudoux P, Ashford RW. Estimation of sand rat abundance by using surface indices. Acta Theriologica 1999; 44: 353-352.

21. Giraudoux P, Pradier B, Delattre P, Deblay S, Salvi D, Defaut R, et al. Estimation of water vole abundance, by using surface indices. Acta theriologica 1995; 40: 77-96.

22. Hansson L. Field signs as indicators of vole abundance. J Appl Ecol 1979; 16: 339-347.

23. Quere JP, Raoul F, Giraudoux P, Delattre P. An index method applicable at landscape scale to estimate relative population densities of the common vole (Microtus arvalis). J Ecol -Terrain Life 2000; 55: 25-32.

24. Giraudoux P, Delattre P, Habert M, Quere JP, Deblay S, Defaut R, et al. Population dynamics of fossorial water vole (Arvicola terrestris scherman): a land usage and landscape perspective. Agric Ecosyst Environ 1997; 66: 47-60.

25. Raoul F, Defaut R, Michelat D, Montadert M, Pepin D, Quere JP, et al. Landscape effects on the populations dynamics of small mammal communities: a preliminary analysis of prey-resource variations. J Ecol - Terrain Life 2001; 56: 339-352.

26. A public health problem of global concern. In: Eckert J, Gemmell MA, Meslin FX, Pawlowski ZS, eds. WHO/OIE manual on echinococcosis in humans and animals. Paris: OIE; 2001: 1-265.

27. Dinkel A, von Nickisch-Rosenegk M, Bilger B, Merli M, Lucius R, Romig T. Detection of Echinococcus multilocularis in the definitive host: coprodiagnosis by PCR as an alternative to necropsy. J Clin Microbiol 1998; 36:1871-1876.

28. van der Giessen JW, Rombout YB, Franchimont JH, Limper LP, Homan WL. Detection of Echinococcus multilocularis in foxes in the Netherlands. Vet Parasitol 1999; 82: 49-57.

29. Hardin G. The tragedy of the commons. Science 1968; 162: 1243-1248.

30. Raoul F, Quere JP, Rieffel D, Bernard N, Takahashi K. Distribution of small mammals along a grazing gradient on the Tibetan plateaus of Western Sichuan, China. Mammalia. In press 2007.

31. Xu HX. Geography in Tibet Autonomous Region. Lhasa: Tibetan Publishing House; 1986: 166-167.

32. He JG, Qiu JM, Liu FJ, Chen XW, Liu DL, et al. Epidemiology survey on hydatiodosis in Tibetan region of western Sichuan II. Infection situation among domestic and wild animals. Chin J Zoonoses (Chin) 2000; 16: 62-65.

33. Viel JF, Giraudoux P, Abrial V, Bresson-Hadni S. Water vole (Arvicola terrestris scherman) density as risk factor for human alveolar echinococcosis. Am J Trop Med Hyg 1999; 61: 559-565.

34. Craig PS, Giraudoux P, Shi D, Bartholomot B, Barnish G, Delattre P, et al. An epidemiological and ecological study of human alveolar echinococcosis transmission in south Gansu, China. Acta Trop 2000; 77: 167-177.

35. Giraudoux P, Delattre P, Takahashi K, Raoul F, Quere KP, Craig P, et al. Transmission ecology of Echinococcus multilocularis in wildlife: what can be learned from comparative studies and multiscale approaches? In: Craig P and Pawlowski Z, eds. Cestode zoonoses: echinococcosis and cysticercosis. Amsterdam: IOS Press; 2002: 251-266.

36. Giraudoux P, Craig PS, Delattre P, Bartholomot B, Bao G, Barnish G, et al. Interactions between landscape changes and host communities can regulate Echinococcus multilocularis transmission. Parasitology 2003; 127: 121-131.