• View in gallery

    Location of the 20 villages studied. The villages belong to Gaojian, Chuanxing, Daxing, and Hainan townships surrounding Qionghai Lake near the city of Xichang in southwestern Sichuan Province in China. A variety of topographic conditions are represented, ranging from flatlands along the lake to steep and highly terraced landscapes at higher elevations.

  • View in gallery

    A, Age distribution of the study participants (line). Bars show the age- and sex-specific prevalence of infection with Schistosoma japonicum. Overall prevalence of infection was 29%. B, Bars show the age- and sex-specific intensity of infection. Overall intensity of infection was 25.7 eggs per gram of stool.

  • View in gallery

    A, Seasonal water contact index based on duration and frequency of different activities weighted by the fraction of body surface area in contact with water for farmers. Water contact was similar for farmers in 18 villages (error bars show the 18-village mean ± 1 SD), but different from villages 8 and 14 (shown individually). B, Average exposure for students. Water contact was roughly the same for students in 20 villages (error bars show the village mean ± 1 SD).

  • View in gallery

    Map of the irrigation system and results from the snail survey and mouse bioassay for village 10 (Chuangxing Xinlong 7). Note the import of cercariae in the fields south of the river, but isolated fields north of the river. ave. = average.

  • View in gallery

    Relationship between crop types and intensity of infection with Schistosoma japonicum. Villages are ordered left to right in increasing intensity of infection and identified by village number and eggs per gram (E.P.G.) of stool. A high EPG is generally related to larger fractions of land devoted to tobacco and vegetable farming and a correspondingly small percentage of rice farming.

  • View in gallery

    Mean village worm burden in bioassay mice versus cercarial risk predicted by the product of mean village snail density and estimated mean density of parasite eggs in the ditch environment.

  • 1

    Davis GM, Wilke T, Zhang Y, Xu XJ, Qiu CP, Spolsky C, Qiu DC, Li YS, Xia MY, Feng Z, 1999. Snail-Schistosoma, Paragonimus interactions in China: population ecology, genetic diversity, coevolution and emerging diseases. Malacologia 41 :355–377.

    • Search Google Scholar
    • Export Citation
  • 2

    Chilton NB, Bao-Zhen Q, Bogh HO, Nansen P, 1999. An electrophoretic comparison of Schistosoma japonicum (Trematoda) from different provinces in the People’s Republic of China suggests the existence of cryptic species. Parasitology 119 :375–383.

    • Search Google Scholar
    • Export Citation
  • 3

    Seto E, Xu B, Liang S, Gong P, Wu WP, Davis G, Qiu DC, Gu XG, Spear R, 2002. The use of remote sensing for predictive modeling of schistosomiasis in China. Photogrammetric Engin Remote Sensing 68 :167–174.

    • Search Google Scholar
    • Export Citation
  • 4

    Jiang Z, Zheng QS, Wang XF, Guan LZ, Hua HZ, 1997. Analysis of social factors and human behavior attributed to family distribution of schistosomiasis japonica cases. Southeast Asian J Trop Med Public Health 28 :285–290.

    • Search Google Scholar
    • Export Citation
  • 5

    Jiang Z, Zheng QS, Wang XF, Hua ZH, 1997. Influence of livestock husbandry on schistosomiasis transmission in mountainous regions of Yunnan Province. Southeast Asian J Trop Med Public Health 28 :291–295.

    • Search Google Scholar
    • Export Citation
  • 6

    The Office of Endemic Disease Control, Ministry of Health, 2000. Handbook of Schistosomiasis Control. Shanghai: Shanghai Science & Technology Press.

  • 7

    Seto E, Liang S, Qiu D, Gu X, Spear RC, 2001. A protocol for geographically randomized snail surveys in schistosomiasis fieldwork using the global positioning system. Am J Trop Med Hyg 64 :98–99.

    • Search Google Scholar
    • Export Citation
  • 8

    WHO, 1985. The Control of Schistosomiasis. Report of a WHO Expert Committee. Geneva: World Health Organization, 113.

  • 9

    Zhong B, Xiong ME, Zhao LG, Lai YH, Lai J, Chen L, Yin HZ, Sha KY, Zhang Y, Gu XG, 2001. Investigation of water contamination situation in Qionghai Lake area, Xichang, Sichuan. J Practical Parasitic Dis 9 :97–100.

    • Search Google Scholar
    • Export Citation
  • 10

    Hubbard A, Liang S, Maszle D, Qiu DC, Gu XG, Spear RC, 2002. Estimating the distribution of worm burden and egg excretion of Schistosoma japonicum by risk group in Sichuan Province, China. Parasitology 125 :221–231.

    • Search Google Scholar
    • Export Citation
  • 11

    Li YS, Sleigh AC, Ross AGP, Li Y, Williams GM, Forsyth SJ, Tanner M, McManus DP, 1999. A 2-year prospective study in China provides epidemiological evidence for resistance in humans to re-infection with Schistosoma japonicum. Ann Trop Med Parasitol 93 :629–642.

    • Search Google Scholar
    • Export Citation
  • 12

    Ross AG, Yuesheng L, Sleigh AC, Williams GM, Hartel GF, Forsyth SJ, Yi L, McManus DP, 1998. Measuring exposure to S. japonicum in China. I. Activity diaries to assess water contact and comparison to other measures. Acta Trop 71 :213–228.

    • Search Google Scholar
    • Export Citation
  • 13

    Lengeler C, Utzinger J, Tanner M, 2002. Screening for schistosomiasis with questionnaires. Trends Parasitol 18 :375–377.

  • 14

    Lengeler C, Utzinger J, Tanner M, 2002. Questionnaires for rapid screening of schistosomiasis in sub- Saharan Africa. Bull World Health Organ 80 :235–242.

    • Search Google Scholar
    • Export Citation
  • 15

    Lund CC, Browder NC, 1944. The estimate of areas of burns. Surg Gynecol Obstet 79 :352–361.

  • 16

    Mosteller RD, 1987. Simplified calculation of body-surface area (letter). N Engl J Med 317 :1098.

  • 17

    Liang S, Zhao WX, Gu XG, Xu FS, Li YX, Li XJ, Liu HC, Li YN, 1997. A report on natural water infection of schistosoma cercariae in mountainous and hilly regions of Sichuan. Ministry of Health, ed. Schistosomiasis Research and Control in China. Nanjing: Nanjing University Press, 260–262.

  • 18

    Scott JT, Diakhate M, Vereecken K, Fall A, Diop M, Ly A, De Clercq D, de Vlas SJ, Berkvens D, Kestens L, Gryseels B, 2003. Human water contacts patterns in Schistosoma mansoni epidemic foci in northern Senegal change according to age, sex and place of residence, but are not related to intensity of infection. Trop Med Int Health 8 :100–108.

    • Search Google Scholar
    • Export Citation
  • 19

    Liang S, Maszle D, Spear RC, 2002. A quantitative framework for a multi-group model of Schistosomiasis japonicum transmission dynamics and control in Sichuan, China. Acta Trop 82 :263–277.

    • Search Google Scholar
    • Export Citation

 

 

 

 

FACTORS INFLUENCING THE TRANSMISSION OF SCHISTOSOMA JAPONICUM IN THE MOUNTAINS OF SICHUAN PROVINCE OF CHINA

View More View Less
  • 1 Center for Occupational and Environmental Health, School of Public Health, University of California, Berkeley, California: Sichuan Institute of Parasitic Diseases, Chengdu, People’s Republic of China; Department of Microbiology and Tropical Medicine, The George Washington Medical Center, Washington, District of Columbia

Twenty villages in the Anning River Valley of southwestern Sichuan China were surveyed for Schistosoma japonicum infections in humans and domestic animals. Also surveyed were human water contact patterns, snail populations, cercarial risk in irrigation systems, and agricultural land use. Few animals were infected, while village prevalence of infection in humans ranged from 3% to 68% and average village eggs per gram of stool ranged from 0 to 110. Except for occupation and education, individual characteristics were not strong determinants of infection intensity within a village. Differences in human infection intensity between these villages are strongly associated with crop type, with low-intensity villages principally growing rice, in contrast to villages devoting more land to vegetables and tobacco. Cercarial risk in village irrigation systems is associated with snail density and human infection intensity through the use of manure-based fertilizer. Some of the agricultural and environmental factors associated with infection risk can be quantified using remote sensing technology.

INTRODUCTION

Here we report the results of an extensive cross-sectional survey of schistosomiasis transmission conducted in 20 villages located in Xichang County in the southern portion of the Anning River Valley in China surrounding Qionghai Lake. The objective of these studies was to identify environmental and agricultural, as well as individual and group factors, responsible for differences in transmission intensity between villages. This study is part of a research program aimed at determining how remote sensing technology can be used to quantify environmental and agricultural factors that significantly modify the potential for disease transmission. This research is motivated by concerns about the long-term effects of global warming on disease transmission and, more immediately, the scale and impact of the ecologic changes arising from completion of the Three Gorges Dam.

The villages lie within a 12 ×12-km area surrounding Qionghai Lake and are populated by approximately 3,900 individuals in four townships: Gaojian, Chuanxing, Daxing, and Hainan (Figure 1). The ethnic composition of the population is almost entirely Han with only 2% being of the Yi ethnic minority. These villages were selected to typify the environment of the 196 villages in these townships. Schistosomiasis control programs in these villages are conducted by the Xichang City Anti-Schistosomiasis Station. Thus, diagnostic and treatment protocols are uniformly administered and none of these villages had recently been the subject of special intervention efforts.

The climate is subtropical with an annual average temperature of 17°C and annual rainfall of approximately 1,000 mm, more than 90% of which falls between the beginning of June and the end of October. The landscape is largely irrigated agricultural land, with two growing seasons. The living and working style of people in a village are usually very similar, and the fields that they farm are usually adjacent to their housing areas. However, crops and farming practices do differ between villages and townships, which are largely related to differences in natural terrain, soil conditions, and economic factors. In the lowland villages near the lake (elevation = 1,520 meters), the predominant crops are rice and wheat, whereas in villages in the terraced foothills (elevation = 1,810 meters), tobacco, vegetables, and corn are more common.

In general, the agriculture typical of the river valley plains does not rely heavily on animal husbandry. Thus, the animal populations are relatively small in comparison with the high mountain valleys in the region. Animal data collected in the fall of 2000 suggest that animals play a small role in disease transmission, with an average of only six buffalo and four horses per village and infection prevalences < 5% (6 of 124) and < 3% (2 of 78), respectively. Other animals, including dogs, goats, and rats, have been surveyed periodically by county and provincial authorities, but have never been found to be important in sustaining disease transmission in this region. In all of these villages, an important factor in sustaining the disease cycle is that fertilization practices make extensive use of human and animal manure, which is moved from residential pit latrines to field storage pits without treatment and with minimal holding times.

With respect to the degree to which these villages are typical of other endemic areas in China, we note that China has long recognized five general ecosystems in which different environmental factors impact the transmission of Schistosoma japonicum. These are 1) flood plains, 2) plain regions with waterway networks, 3) swamp, marshland, and lake regions, 4) extensive canals, and 5) hilly and mountainous regions. The villages studied here belong to the hilly and mountainous class. However, within this class there are differences in the subspecies of snail involved and in various other environmental and host-related factors that Davis and others propose to be the basis of five ecogenetic modes of transmission.1 Fundamental to this alternative system is the recognition of differences in the population genetics of the snails that co-evolved with the parasite. The subspecies of the Oncomelania snail that populate the environment studied here is Oncomelania hupensis robertsoni, which evolved only at high elevations in the Yunnan and Sichuan provinces above the Three Gorges of the Yangtze River. This subspecies is highly divergent genetically from O. h. hupensis, which is found throughout eastern and southeastern China below the Three Gorges;2 however, both subspecies are found in hilly and mountainous areas.

Following either the traditional Chinese ecologically based classification or the alternative proposed by Davis and others, these 20 villages are clearly within a single class. However, all 20 villages lie within a sub-area of high priority sites for disease surveillance based on their potential to support snail populations.3 Despite this, these sites have historically exhibited a wide range of disease prevalence. They represent a range of topographic conditions from flatlands along the lake to steep and highly terraced landscapes at higher elevations. Thus, they provide an opportunity to identify environmental as well as individual and group factors responsible for differences in transmission intensity between villages. In addition, we believe this to be one of the few reports in the English language describing the environmental and agricultural factors associated with endemic schistosomiasis in the hilly and mountainous regions of southwestern China.4,5

METHODS

The survey and surveillance methods used in this investigation followed standard methods and protocols recommended by the Ministry of Health.6 These include methods for the assessment of the prevalence and intensity of human infection, the determination of the cercarial hazard of surface waters, and the assessment of snail populations, although the latter was supplemented by ditch maps generated using a Global Positioning System (GPS).7 In addition, we used water-contact questionnaires, developed by the Sichuan Institute of Parasitic Disease, and records held at the township level of acreage devoted to the various crops grown in the region. Data on agricultural land use by season was obtained from the county Land Use Information Office for calendar year 2000.

Human prevalence and intensity of infection (November).

Because of the endemic nature of schistosomiasis in these villages, the villagers are familiar with infection surveys and procedures. The availability and timing of testing is announced by village leaders and participation is encouraged although not required. Verbal consent is requested and participation is typically approximately 90% among those in the targeted age groups after follow-up to inform those who may have been out of the village at the time of the announcement.

The infection survey targeted all residents between the ages of 4 and 60 but was open to villagers of any age. At the time of diagnostic testing, an interview was administered. Among the information updated during the interview with each resident were 1) individual characteristics (e.g., sex, height, weight, natural village, education, occupation), 2) general medical history, 3) schistosomiasis infection history (e.g., number of times treated, when last treated) and, 4) among a 25% stratified random sample, location-specific water contact behaviors as discussed later in this report. Stool sampling and analysis were performed at the end of the infection season following Ministry of Health protocols.6 The first step was to carry out a miricidial hatch test that involved the collection of at least 30 grams of stool from each person on three separate days. In addition, subjects were asked to submit an entire day’s stool for quantitative egg count using the Kato-Katz thick smear procedure as recommended by the World Health Organization Steering Committee for Schistosomiasis.8 In this study, three 41.5-mg slides were examined from previously homogenized stool samples, which results in a total sample weight of 124.5 mg.

Water contact survey (November).

At the time the collection of diagnostic samples, a 25% random sample of the study population, stratified by village and occupation, was selected to participate in an interview aimed at assessing the nature, frequency, and location of their water contact by month. The questionnaire was administered orally by one member of the study team. The questionnaire is structured around the following seasonal agricultural activities: washing clothes or vegetables, washing agricultural tools, washing hands and feet, playing or swimming, ditch operation or maintenance, plowing, rice cutting, and fishing. Because respondents were asked to recall the frequency, duration, and location of these activities over a period of up to seven months in the past, responses are susceptible to recall errors.

Snail populations (June).

A ditch map of the total snail habitat was created for each village by walking the ditch network with a GPS unit. These data were then downloaded and processed with geographic information system software to generate maps that are then used to index the sampling locations.7 Subsequently, snail sampling was carried out at fixed 10-meter intervals along the ditch system. At each location, the traditional kuang sampling frame (0.11 m2) was used to collect all snails within the frame.6 The collected snails are temporarily stored in a paper envelope, coded by location, and returned to the laboratory for determination of their infection status. This is done by crushing individual snails between glass microscope slides and inspecting them using a dissecting microscope at 15–20× magnification.

Cercarial bioassay (July).

The cercarial bioassay uses a cage containing five mice suspended above an irrigation canal so that the feet and tails of the mice are in the water.6,9 This method is used in preference to filter-based methods because of water turbidity and the sticky nature of S. japonicum cercaria, which make them difficult to recover from container surfaces or tubing of conventional sampling apparatus. Exposure of the mice was for one hour in the morning and a second hour in the afternoon on five consecutive days for a total of 10 hours. Site selection criteria are determined principally by the location of infected snails as determined by the snail surveys or by frequent water contact locations, for example, near residential areas or schools. Cages are also located in ditches at village boundaries to assess import or export of cercaria. In this survey 6–16 sites per village were surveyed during late July. The timing of the exposure is conditioned on water temperature, with colder periods being avoided because of low cercarial infectivity as well as survival of the animals. The animals are held for 35 days following exposure to allow for parasite maturation. They are then dissected and the worms are counted.

RESULTS

The descriptive data on the village population, amount of land farmed, mean household size of study participants, and the primary occupation reported by the villager at the time of sample collection are shown in Table 1. The actual household size is larger because of the addition of children less than fours years old and adults greater than 60 years old, except for the relatively small number of the latter who were not targeted, but chose to participate in the survey, and are therefore included. The category “Other” in Table 1 includes domestic workers, administrators, teachers, and people working in non-agricultural jobs outside the village. Clearly, these variables suggest a relatively homogeneous population except for rather marked differences in the “other” occupation category. These differences may be due to subtle distinctions between primary and secondary occupations as reported by the villagers. The numbers in parentheses in the Farmer column of Table 1 are the percentage of villagers whose primary or secondary occupation is farmer. Using this classification, we observed that the differences between villages are decreased. The notable remaining differences are in the Gaojian villages (13, 14, and 15) which are close to Xichang City where non-agricultural employment is available.

The age distribution of participants is indicated by the line superimposed in Figure 2A. A random effects model was used to determine the ratio of the variance between villages to the total variance in age and education attributable to differences both within and between villages. These fractions were 0.00086 and 0.025 for age and education, respectively, indicating little difference in the distribution of these variables between the 20 villages.

The infection data by village and occupation with prevalence defined as a positive result in the Kato-Katz procedure or the hatch test, and by intensity as measured by eggs per gram (epg) of stool by the Kato-Katz procedure are summarized in Table 2. A detailed analysis of these data, focusing on the estimation of the parameters of the distribution of epg, as well as the performance of the diagnostic tests, has been presented elsewhere.10 There we reported that virtually all samples that were hatch test positive were Kato-Katz positive and vice versa, a somewhat surprising result, but one that has also been observed in routine surveillance work in recent years in the highly endemic areas of Sichuan. These earlier analyses also show that within risk groups defined by village and occupation, the epg counts are highly aggregated among individuals, but that estimates of the degree of that aggregation are confounded by the estimate of the day-to-day variability in egg excretion for an individual at constant worm burden. For present purposes, however, the results presented in Table 2 show that the original objective of selecting villages with a wide range of infection intensities was accomplished. The data generally show the highest intensity of infection among farmers followed by students and others. Figure 2B shows these data by age and sex. In contrast to the patterns seen in populations in endemic areas of China on the lower Yangtze River, the prevalence of infection by age and sex is more uniform and the intensity of infection only slightly more variable when averaged over all 20 villages.11

Water contact data obtained retrospectively by questionnaire are considered by some to be of dubious reliability, and by others to be the preferred method.12–14 While we share concerns in terms of detailed quantitative results, we feel more confident with the qualitative patterns of the data. To summarize the data over tasks, we used an index of water contact based on duration and frequency of tasks weighted by the fraction of the body surface estimated to be wetted.15,16 These weights were 0.05 for clothes washing, ditch cleaning, rice cutting, and plowing; 0.03 for washing agricultural tools; 0.12 for washing of hands and feet; 0.20 for playing/swimming; and 0.32 for fishing. All of these activities were in surface waters potentially harboring cercaria as opposed to fish ponds, for example, which have seldom been shown to do so in this setting.9,17 The results by village for farmers and students are shown in Figure 3. The two villages that did not conform to the pattern of the others were villages 8 and 14: Chuanxing Minhe 1 and Gaojian Zhongsuo 8. The former was heavily influenced by playing/swimming reports from the farmers, which was not replicated in other villages, and the latter by washing of hands and feet, and fishing. While we are suspicious of the remarkable uniformity of the contact intensity over time among farmers from the other villages, it does seem plausible that the season-long water contact of farmers is much more uniform than the pattern of students, which shows a distinct summer peak that is consistent with what can be observed in these villages.

A summary of the snail surveys is shown in Table 3. The number of linear meters of snail habitat is the length of the surveyed irrigation ditches in each village. The area of habitat can be approximated by multiplying the ditch length by 0.25 to 0.50 meters, these being rough estimates of the average distance above the waterline occupied by snails. For each village, a negative binomial distribution was fit to the snail data resulting in the tabled estimates of the mean density per m2 and the aggregation constant k, which is inversely proportional to the aggregation of snails. Given the large number of snails collected in these surveys, averaging 880 per village, we regard the estimate of mean density and the fraction infected to be reliable although since they are the result of a cross-sectional survey, they provide no information regarding the change in density over the season. There is evidence of an overdispersed distribution of snails in all 20 villages, such that a relatively large number of snails are found in a small number of kuang. As expected, the spatial distribution of these high-density kuang are not randomly distributed along the ditches, but clumped. This is most likely due to underlying microecologies in which more snails are found due to their active movement towards conditions they prefer or due to the passive flushing of snails downstream into pockets of high density due to rainfall and field-draining events. The nonuniform distribution of snails can be seen in Figure 4, which shows the GPS ditch maps of village 10, Chuanxing Xinlong 7, with the snail survey and the mouse bioassay results overlaid. The bioassay data will be discussed later in this report.

Despite the small numbers of infected snails observed in these surveys, an attempt was made to determine if infected snails were clustered to a greater or lesser extent than would be expected from the clustering of uninfected snails. That is, is there evidence that the probability of infection of an individual snail differs by location in the village? The answer to this question has implications for focal mollusciciding since, in one case the target is uninfected snail clusters and, in the other, infected snail clusters (if the latter were stable over time and, given the low rate of snail infection, could be identified in any practical way). The analysis used negative binomial regression, which results in a likelihood ratio test that a negative binomial model better fits the data than a Poisson model, the latter suggesting a uniform infection probability over the village. Not surprisingly, the general result was that the Poisson model cannot be rejected at the 5% level for villages with small numbers of infected snails (villages 1, 5, 8, and 12), but that the negative binomial is a better fit for villages with relatively large numbers of infected snails (villages 4 and 10). Village 9 just lacked sufficient power to choose the negative binomial model over the Poisson model (P = 0.08). These results suggest that the probability of snail infection is not uniform across the village environment. However, the exceptions are villages 2 and 3, both with a relatively high number of infected snails (22 and 19, respectively), where the Poisson model is preferred, suggesting a uniform distribution of snail infection. Since this survey was conducted in May, it reflects snail exposure over periods of several months plus infected snails that have over-wintered. Thus, we speculate that the picture in the fall might show a more uniform snail infection probability across the village. Nevertheless, the non-uniform distribution of snail infection in some villages may reflect an uneven distribution of miracidia in the environment that persists over the infection season.

The cercarial bioassay results are shown in Table 4. As noted earlier in this report, the placement of the mice on the ditch system was dictated by various considerations ranging from popular water contact sites, as identified by the local health personnel, to concerns about import of cercaria from neighboring villages. Village 3, Daxing Shian 5, and village 5, Chuanxing Jiaojia 4, both show only modest importing of cercariae from the outside and are, therefore, candidates for internal control strategies. The village depicted in Figure 4, Xinlong 7, presents a complicated picture because the village lands are on both sides of the river. All influent cercaria affect the smaller piece of land south of the river. The village itself and the main part of its lands are isolated from upstream sources.

The fraction of village land used in the spring of 2000 for rice and for tobacco plus vegetables versus village mean intensity of human infection is shown in Figure 5. Clearly, there is a tendency for villages whose spring crop is principally rice to have low average intensity of infection and conversely for those that grow tobacco and vegetables. While we have made no detailed investigation of village and/or chemical fertilizer use for a crop, a preliminary survey suggests that tobacco and vegetables are preferentially fertilized with manure whereas rice is preferentially fertilized with commercial chemical fertilizer. If confirmed, this would provide an explanation for the pattern shown in Figure 5. That is, it suggests that miracidial input resulting from manure use within the village will be lower in villages whose spring crop is predominately rice as opposed to those with tobacco and vegetables.

Because human parasite burden can accumulate over periods of years, we explored the relationship of egg excretion in humans with various individual predictor variables that might be expected to be relatively stable over time. These data were from the 25% random sample of villagers, comprising a total of 1,037, for whom monthly water contact data was available, as well as age, sex, occupation, and education. This analysis used negative binomial regression, initially with one variable at a time. Only education and occupation were significant predictors of egg excretion. These two variables were then combined into a single model and both remained independent predictors. Those with the highest level of education, through high school, showed the lowest mean infection intensity whereas farmers showed the highest levels among occupational groups. Presumably, both education and occupation are proxies for the nature and intensity of water contact in this setting although education may also correspond to increased awareness and access to treatment. However, it is not clear why the water contact variables, based on the questionnaires, were not predictive of infection intensity. However, a similar result has recently been reported by Scott and others, working in northern Senegal, who investigated the issue in detail.18

DISCUSSION

While occupation and education were significant individual predictors of infection intensity, these variables were very similarly distributed across villages. Thus, they do not account for the village differences in mean infection intensity. In exploring the relationships between other elements of these data at the village level, we regressed the mean square root worm burden in mice on the product of mean village epg and village population, assuming this product to be an index of the rate of egg/miracidium input to the snail habitat within the village. In so doing, we chose to regard epg as a determinant of cercarial risk, rather than vice versa, because epg reflects cumulative infection intensity in untreated human populations over a time frame of years versus months for infected snails.19 The resulting slope coefficient differed from zero at a high level of significance (P < 0.0005). A similar regression analysis was carried out, again with village mean square root of worms per mouse as the outcome, but with the percentage of land devoted to rice as the predictor variable as suggested by Figure 5. Again the regression was highly significant (P < 0.0005) with the worm levels in mice decreasing with increasing percentages of land used for the culture of rice.

Mean village snail density was not a significant independent predictor of the bioassay results despite the fact that there is a roughly 30-fold difference in the snail density among these villages, as can be seen from Table 3. However, all villages with a prevalence of more than 30% have snail densities of greater than 9.7/m2, which suggests a threshold effect of snail density.

For these and other village level analyses, the sample size is limited to 20. This restricts the statistical power available to explore multivariate relationships. Thus, we postulate a multivariable structure incorporating the foregoing predictor variables measured at the village level as suggested by dynamic models of schistosomiasis transmission.19 Specifically, we assume that the production of snail infections, and subsequently mean cercarial density in the village irrigation system, is proportional to the density of snails times the density of eggs along the ditches. As before, cercarial density is measured by the mouse bioassay. These assumptions yield cercarial risk = K (snail density)[(epg)(population)(1 − fraction rice)/habitat], a relationship that incorporates the several individual effects discussed earlier, but without the square root transformation on worms per mouse. As discussed earlier, the (1 - fraction rice) term is postulated to represent the fraction of the total egg-laden manure that is distributed as fertilizer. Thus, the term in brackets is the mean density of eggs if all found their way into the ditch system. A linear regression results in an R2 of 0.95, as shown in Figure 6, with an intercept not significantly different from zero. Whereas the clustering of these data points along the x-axis certainly inflates the R2 value, some further non-statistical support can be offered to the foregoing model of factors explaining cercarial risk. In particular, the predictor variables relate to snail infections attributable to processes internal to the village. They do not predict the effects of either miracidial or cercarial input carried into the village from upstream. Thus, it is notable that the two villages at the upper right of the plot, above the regression line, are the only two with high levels of cercaria in influent water at the upstream boundary of the village (Table 4). We were able to discern no evidence of imported miracidia in these analyses.

The foregoing formulation of cercarial risk can be divided into two parts, one related to the potential for disease transmission and the other to the presence of the parasite in the village population. Our short-term objective is to develop remote sensing methods to allow the estimation of village transmission potential. However, the degree to which the potential for disease transmission in a village is realized, or an established level controlled, is highly dependent on the degree to which the village snail population and/or its human and animal inhabitants are connected to external sources of the parasite. Connectivity depends on both hydrologic factors and the ease of movement of people and material in the rural environment. These factors may also be amenable to remote estimation, which is of importance because connectivity is a crucial determinant of the focus and scale of effective control strategies.

Finally, while there is little in our findings particularly surprising from broad epidemiologic or parasitologic perspectives, we believe that the apparent dominance of agricultural and environmental factors in determining human infection risk in these villages carries important implications for disease surveillance and control strategies. Specifically, this environment appears to be particularly suitable for the use of existing composting or waste treatment methods for the minimization of parasite input resulting from the use of manure-based fertilizer. Furthermore, snail control through environmental modifications and strategies involving crop selection may also be more promising approaches to disease control in this setting than in other endemic areas of China.

Table 1

Village population, amount of land farmed, mean household size of study participants, and distribution of primary occupations

VillageTotal populationFarmland (hectares)Mean participants/householdFarmer (%)*Student (%)Other (%)
* Numbers in parentheses are the percentage of villagers whose primary or secondary occupation is farming.
124017.33.3669.8 (70)29.80.4
218815.83.6461.2 (62)30.38.5
312510.52.9875.2 (77)21.63.2
421911.83.4249.8 (67)26.523.8
522313.83.3143.6 (68)27.129.3
618010.53.4653.9 (62)26.719.4
722713.33.3468.3 (69)25.16.6
823617.13.5560.1 (65)25.614.3
919413.53.4847.2 (69)23.129.7
101577.13.8170.1 (70)27.42.55
1121310.73.4960.6 (69)26.313.2
1228211.33.8167.7 (68)27.35.0
1316111.13.3528.0 (60)32.339.8
1419710.83.5250.3 (55)33.516.3
151656.43.1557.0 (56)23.020.0
1618716.92.9772.2 (74)22.55.4
171668.43.3966.3 (73)22.910.8
1827218.03.2462.1 (67)27.99.9
1915412.02.9664.3 (64)25.310.4
201147.33.1734.2 (69)27.937.8
Table 2

Prevalence and intensity of infection (eggs per gram [epg] of stool) by village*

epg by primary occupation
Village% Eligible respondingPrevalence (%)OtherFarmerStudentVillage epg
* Positive infection status was determined by either a positive Kato-Katz result or positive hatch test result. The intensity of infection is shown for each primary occupational group and the entire village.
192.5443.311.45.69.7
297.37316.3130.769.9103.7
391.26811.7105.155.490.8
495.96550.972.427.459.9
587.0170.04.92.84.1
692.8129.03.23.23.9
788.13418.828.020.425.7
889.04011.021.08.316.9
989.26212.674.1139.183.8
1090.565104.3127.059.5110.4
1179.81713.85.14.15.3
1280.9131.83.67.94.2
1387.640.00.80.60.7
1488.850.00.70.80.7
1586.7130.71.60.51.1
1679.7130.03.10.32.4
1787.4100.81.62.01.6
1887.591.43.90.02.9
1977.330.00.20.00.1
2093.060.00.31.60.7
Total2912.029.819.125.7
Table 3

Snail survey results for each village*

VillageElevation change (Δm)Habitat (m)Mean no. of snails/m2kNo. of infected snailsInfection rate with 95% confidence limits
* Small values of k indicate an overdispersed snail distribution in all 20 villages. Confidence intervals for infection rates were determined from a binomial distribution. m = meters.
169.8910,35011.60.2950.005 (0.001–0.009)
2118.437,60237.90.53220.006 (0.004–0.010)
353.003,78425.00.49190.016 (0.010–0.025)
455.028,89818.30.36360.019 (0.014–0.027)
534.504,99110.50.0950.010 (0.003–0.020)
613.334,0755.10.1500 (0–0.016)
734.5110,6919.70.1300 (0–0.004)
855.357,78024.70.3530.002 (0–0.004)
963.367,67813.50.35110.010 (0.005–0.018)
1094.893,87820.90.34380.023 (0.017–0.032)
1159.073,4178.30.1500 (0–0.012)
1217.176,3214.40.0720.007 (0.001–0.024)
1310.444,69910.60.0900 (0–0.007)
1424.684,9535.80.0600 (0–0.013)
1540.204,2781.60.0300 (0–0.054)
1621.247,06713.50.1300 (0–0.004)
1729.495,3938.60.0500 (0–0.008)
1854.457,54013.40.4000 (0–0.003)
1920.292,7647.80.3700 (0–0.016)
2057.302,59320.50.5000 (0–0.006)
Table 4

Mouse bioassay results for each village*

VillageNo. of mice assayedMortality %Mean worm burden in miceNo. of upstream boundary miceMean worm burden in boundary mice
* Boundary mice refers to cages placed in ditches influent to the village to assess incoming cercarial concentrations.
15923.452.4
276535.41415.2
3621221.452.0
46577.892.3
535132.7150.1
627100.0120.0
72930.1100.0
853121.6102.5
952106.5141.4
1059231.71526.9
1128201.4100.8
1240110.0100.0
135500.0200.0
1433270.150.0
153360.350.0
1628302.9110.0
1729280.040.0
186500.1150.2
193000.00
2027100.050.0
Figure 1.
Figure 1.

Location of the 20 villages studied. The villages belong to Gaojian, Chuanxing, Daxing, and Hainan townships surrounding Qionghai Lake near the city of Xichang in southwestern Sichuan Province in China. A variety of topographic conditions are represented, ranging from flatlands along the lake to steep and highly terraced landscapes at higher elevations.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 1; 10.4269/ajtmh.2004.70.48

Figure 2.
Figure 2.

A, Age distribution of the study participants (line). Bars show the age- and sex-specific prevalence of infection with Schistosoma japonicum. Overall prevalence of infection was 29%. B, Bars show the age- and sex-specific intensity of infection. Overall intensity of infection was 25.7 eggs per gram of stool.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 1; 10.4269/ajtmh.2004.70.48

Figure 3.
Figure 3.

A, Seasonal water contact index based on duration and frequency of different activities weighted by the fraction of body surface area in contact with water for farmers. Water contact was similar for farmers in 18 villages (error bars show the 18-village mean ± 1 SD), but different from villages 8 and 14 (shown individually). B, Average exposure for students. Water contact was roughly the same for students in 20 villages (error bars show the village mean ± 1 SD).

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 1; 10.4269/ajtmh.2004.70.48

Figure 4.
Figure 4.

Map of the irrigation system and results from the snail survey and mouse bioassay for village 10 (Chuangxing Xinlong 7). Note the import of cercariae in the fields south of the river, but isolated fields north of the river. ave. = average.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 1; 10.4269/ajtmh.2004.70.48

Figure 5.
Figure 5.

Relationship between crop types and intensity of infection with Schistosoma japonicum. Villages are ordered left to right in increasing intensity of infection and identified by village number and eggs per gram (E.P.G.) of stool. A high EPG is generally related to larger fractions of land devoted to tobacco and vegetable farming and a correspondingly small percentage of rice farming.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 1; 10.4269/ajtmh.2004.70.48

Figure 6.
Figure 6.

Mean village worm burden in bioassay mice versus cercarial risk predicted by the product of mean village snail density and estimated mean density of parasite eggs in the ditch environment.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 70, 1; 10.4269/ajtmh.2004.70.48

Authors’ addresses: Robert C. Spear, Edmund Seto, Song Liang, Merrill Birkner, and Alan Hubbard, Center for Occupational and Environmental Health, School of Public Health, University of California, 140 Earl Warren Hall, #7360, Berkeley, CA 94720-7360, Telephone: 510-642-0761, Fax: 510-642-5815, E-mail: spear@uclink4.berkeley.edu. Dongchuan Qiu, Changhong Yang, Bo Zhong, Fashen Xu, and Xueguang Gu, Sichuan Institute of Parasitic Diseases, 10 University Road, Chengdu, Sichuan 610041, People’s Republic of China. George M. Davis, Department of Microbiology and Tropical Medicine, The George Washington Medical Center, 731 Ross Hall, 2300 Eye Street NW, Washington DC, 20037.

Acknowledgments: We are indebted to Sha Kaiyou (Director) and colleagues at the Xichang County Anti-Schistosomiasis Station, and to Kang Junxin (Director) of the Endemic Diseases Office, Sichuan Centers for Disease Control (Chengdu, People’s Republic of China) for their sustained support and collaboration.

Financial support: This work was supported in part by the National Institute for Allergy and Infectious Disease (grant RO1 AI-43961).

REFERENCES

  • 1

    Davis GM, Wilke T, Zhang Y, Xu XJ, Qiu CP, Spolsky C, Qiu DC, Li YS, Xia MY, Feng Z, 1999. Snail-Schistosoma, Paragonimus interactions in China: population ecology, genetic diversity, coevolution and emerging diseases. Malacologia 41 :355–377.

    • Search Google Scholar
    • Export Citation
  • 2

    Chilton NB, Bao-Zhen Q, Bogh HO, Nansen P, 1999. An electrophoretic comparison of Schistosoma japonicum (Trematoda) from different provinces in the People’s Republic of China suggests the existence of cryptic species. Parasitology 119 :375–383.

    • Search Google Scholar
    • Export Citation
  • 3

    Seto E, Xu B, Liang S, Gong P, Wu WP, Davis G, Qiu DC, Gu XG, Spear R, 2002. The use of remote sensing for predictive modeling of schistosomiasis in China. Photogrammetric Engin Remote Sensing 68 :167–174.

    • Search Google Scholar
    • Export Citation
  • 4

    Jiang Z, Zheng QS, Wang XF, Guan LZ, Hua HZ, 1997. Analysis of social factors and human behavior attributed to family distribution of schistosomiasis japonica cases. Southeast Asian J Trop Med Public Health 28 :285–290.

    • Search Google Scholar
    • Export Citation
  • 5

    Jiang Z, Zheng QS, Wang XF, Hua ZH, 1997. Influence of livestock husbandry on schistosomiasis transmission in mountainous regions of Yunnan Province. Southeast Asian J Trop Med Public Health 28 :291–295.

    • Search Google Scholar
    • Export Citation
  • 6

    The Office of Endemic Disease Control, Ministry of Health, 2000. Handbook of Schistosomiasis Control. Shanghai: Shanghai Science & Technology Press.

  • 7

    Seto E, Liang S, Qiu D, Gu X, Spear RC, 2001. A protocol for geographically randomized snail surveys in schistosomiasis fieldwork using the global positioning system. Am J Trop Med Hyg 64 :98–99.

    • Search Google Scholar
    • Export Citation
  • 8

    WHO, 1985. The Control of Schistosomiasis. Report of a WHO Expert Committee. Geneva: World Health Organization, 113.

  • 9

    Zhong B, Xiong ME, Zhao LG, Lai YH, Lai J, Chen L, Yin HZ, Sha KY, Zhang Y, Gu XG, 2001. Investigation of water contamination situation in Qionghai Lake area, Xichang, Sichuan. J Practical Parasitic Dis 9 :97–100.

    • Search Google Scholar
    • Export Citation
  • 10

    Hubbard A, Liang S, Maszle D, Qiu DC, Gu XG, Spear RC, 2002. Estimating the distribution of worm burden and egg excretion of Schistosoma japonicum by risk group in Sichuan Province, China. Parasitology 125 :221–231.

    • Search Google Scholar
    • Export Citation
  • 11

    Li YS, Sleigh AC, Ross AGP, Li Y, Williams GM, Forsyth SJ, Tanner M, McManus DP, 1999. A 2-year prospective study in China provides epidemiological evidence for resistance in humans to re-infection with Schistosoma japonicum. Ann Trop Med Parasitol 93 :629–642.

    • Search Google Scholar
    • Export Citation
  • 12

    Ross AG, Yuesheng L, Sleigh AC, Williams GM, Hartel GF, Forsyth SJ, Yi L, McManus DP, 1998. Measuring exposure to S. japonicum in China. I. Activity diaries to assess water contact and comparison to other measures. Acta Trop 71 :213–228.

    • Search Google Scholar
    • Export Citation
  • 13

    Lengeler C, Utzinger J, Tanner M, 2002. Screening for schistosomiasis with questionnaires. Trends Parasitol 18 :375–377.

  • 14

    Lengeler C, Utzinger J, Tanner M, 2002. Questionnaires for rapid screening of schistosomiasis in sub- Saharan Africa. Bull World Health Organ 80 :235–242.

    • Search Google Scholar
    • Export Citation
  • 15

    Lund CC, Browder NC, 1944. The estimate of areas of burns. Surg Gynecol Obstet 79 :352–361.

  • 16

    Mosteller RD, 1987. Simplified calculation of body-surface area (letter). N Engl J Med 317 :1098.

  • 17

    Liang S, Zhao WX, Gu XG, Xu FS, Li YX, Li XJ, Liu HC, Li YN, 1997. A report on natural water infection of schistosoma cercariae in mountainous and hilly regions of Sichuan. Ministry of Health, ed. Schistosomiasis Research and Control in China. Nanjing: Nanjing University Press, 260–262.

  • 18

    Scott JT, Diakhate M, Vereecken K, Fall A, Diop M, Ly A, De Clercq D, de Vlas SJ, Berkvens D, Kestens L, Gryseels B, 2003. Human water contacts patterns in Schistosoma mansoni epidemic foci in northern Senegal change according to age, sex and place of residence, but are not related to intensity of infection. Trop Med Int Health 8 :100–108.

    • Search Google Scholar
    • Export Citation
  • 19

    Liang S, Maszle D, Spear RC, 2002. A quantitative framework for a multi-group model of Schistosomiasis japonicum transmission dynamics and control in Sichuan, China. Acta Trop 82 :263–277.

    • Search Google Scholar
    • Export Citation

Author Notes

Reprint requests: Robert C. Spear, Center for Occupational and Environmental Health, School of Public Health, University of California, 140 Earl Warren Hall, #7360, Berkeley, CA 94720-7360.
Save