Transmission of Opisthorchis viverrini, Schistosoma mekongi and soil-transmitted helminthes on the Mekong Islands, Southern Lao PDR

Background Prevalence of Opisthorchis viverrini, Schistosoma mekongi and soil-transmitted helminths (STH) remains high in Lao People’s Democratic Republic (Lao PDR), despite control efforts including mass-drug administration, education and communication campaigns. New approaches are required to advance helminth control. Methods An ecohealth study was conducted on two Mekong islands in Southern Laos. Demographic and behavioural data were collected by questionnaire. Human and animal reservoir stools were examined. Bithynia spp. and Neotricula aperta snails were examined using shedding. Fresh water fish were examined using digestion technique. Multivariate random-effects analysis was used to find risk factors associated with helminth infections. Results Human infection rates with O. viverrini, hookworm, S. mekongi, Trichuris trichiura, Ascaris lumbricoides and Taenia spp. were 60.7%, 44.1%, 22.2%, 4.1%, 0.6% and 0.1%, respectively. Heavy intensity infections were 4.2%, 3.6% and 1.8% for O. viverrini, S. mekongi and hookworm, respectively. O. viverrini and S. mekongi infection rates among dogs and cats were 25.0% and 14.7%, respectively. Of the cats tested, 53.1% were infected with O. viverrini. Prevalence of O. viverrini and S. mekongi in snails was 0.3% and 0.01%, respectively. Overall prevalence of O. viverrini infection in fresh water fish was 26.9%, with the highest infection rates occurring in Hampala dispa (87.1%), Cyclocheilichthys apogon (85.7%) and Puntius brevis (40.0%). Illiteracy and lower socioeconomic status increased the risk of O. viverrini infection, while those aged 10–16 years and possessing latrines at home were less likely to be infected. Household dogs and cats that consumed raw fish were significantly and positively associated with O. viverrini infection of the household members. For S. mekongi, children under 9 years old were exposed significantly to this infection, compared to older age groups. Conclusions There is a pressing need to design and implement an integrated helminth control intervention on the Mekong Islands in southern Lao PDR. Given the highly dynamic transmission of O. viverrini, S. mekongi, STH and extended multiparasitism, annual mass-drug administration is warranted along with environmental modifications, health education and improved access to clean water and adequate sanitation to consolidate morbidity control and move towards elimination. Trail registration number Our findings presented here are from a cross-sectional study, therefore, it has not been registered. Electronic supplementary material The online version of this article (doi:10.1186/s40249-017-0343-x) contains supplementary material, which is available to authorized users.


Multilingual abstract
Please see Additional file 1 for translations of the abstract into the five official working languages of the United Nations.

Background
Helminthiases are neglected tropical diseases (NTDs) of major public health concern in many low-and middleincome countries (LMIC) in the tropics and sub-tropics, including in Lao People's Democratic Republic (Lao PDR) [1][2][3][4]. Liver flukes (Opisthorchis viverrini), blood flukes (Schistosoma mekongi) and soil-transmitted helminths (STH) such as round worm (Ascaris lumbricoides), whipworm (Trichuris trichiura) and two-hookworm species (Ancylostoma duodenale, Necator americanus) are among the most prevalent infections in Lao PDR. O. viverrini is endemic nationwide but is most prevalent in the central and southern parts of the country. It occurs in the lowlands, along the Mekong River, where fish are abundant and local inhabitants prefer to consume traditional dishes prepared with raw fish [1,[4][5][6]. S. mekongi is only endemic in two districts of the most southern province, Champasack, bordering Cambodia [7][8][9][10]. STH are highly prevalent in the northern part of the country and in the mountainous areas along Lao-Vietnamese border [4,11].
Helminths have complex life cycles; O. viverrini, for example, involves two aquatic intermediate hosts, namely freshwater snails (of the genus Bithynia) and freshwater fish (of the Cyprinidae family). Humans and other mammals are infected by eating raw or undercooked fish [21]. The life cycle of S. mekongi involves humans and other mammals (such as dogs, pigs and possibly rats) [22,23]. The Neotricula aperta snail, which lives in the crevices of submerged rocks in the Mekong River, serves as intermediate host.
The cercariae emerge from the infected snails during the daytime and lie under the water surface [9,24]. Humans and animals are infected with this parasite via skin penetration when they come into contact with infested waters [8]. Lao PDR adheres to the preventive chemotherapy control strategy promoted by WHO [3,25,26]. Over the last decade, considerable efforts were employed to implement this strategy through deworming programmes targeting schoolchildren [27] and through mass-drug administration (MDA) alongside information, education and communication (IEC) campaigns in high risk provinces of the country [28]. Despite these efforts, the prevalence of helminth infections, including multiple infections, remains high in many places [4,18,26,[29][30][31]. Given the complexity of the transmission cycle of helminth infections and the risky behaviour of humans in endemic communities, it may be necessary to adapt the control strategy to improve the effectiveness of interventions.
Ecohealth research is an emerging field of research studying human health in close connectivity with the ecosystem [32]. It is increasingly conducted to strengthen the sustainability of infectious disease control programmes [33][34][35] and was widely introduced in Southeast Asia (SEA) by the Canadian International Development Research Centre (IDRC) in the late 2000s [36,37]. Ecohealth has been defined as follows: i) "EcoHealth involves research and practice to promote sustainability of individuals, animals and biodiversity by linking complex interaction of ecosystem, socio-cultural and economic factors" and ii) "Ecohealth is a comprehensive approach to understanding health at its human, animal and environmental interface in a socioecological systems context". Here, we employ an ecohealth approach to determine the prevalence and risk factors of O. viverrini, S. mekongi and STH infections in humans in the ecological environment of Khong district, where potential animal reservoir and intermediate hosts, like molluscs and fish, live in close connectivity.

Study area
Khong district is an island district located at the Southern border of Champasack Province, Lao PDR (Fig. 1a). It has an estimated population of 100,000 people and comprises a few dozen islands in the Mekong River (geographical coordinates: 13.57°-14.14°N latitude and 105.44°-106.08°E longitude). The district is a known endemic area for O. viverrini, S. mekongi and STH. Done Khon and Done Som are among the biggest islands and are popular tourist destinations. Done Khon has about 260 households with a total population of 1560 people, while Done Som has some 378 households with a total population of 2344 people.

Study design and population surveyed
Our cross-sectional study was carried out between October 2011 and August 2012 on Done Khon and Done Som islands. These study sites were selected based on a three-stage random sampling. First, we randomly selected two islands out of 10 known endemic islands for the targeted diseases. For each island, 323 study participants were required based on our sample calculation using a formula of simple random sampling, e.g., Z 1-α/ 2 2 × p(1-p)/d 2 with a 30% proportion and 5% precision. Based on previous experiences, about 40% of all study participants (129 persons) failed to submit complete stool samples when they were asked to submit multiple stool samples (i.e., at least two). With this in mind, at least 904 study participants from both islands were required for this study. Second, two villages were selected on each island. Finally, about 30 households in each village were randomly selected to meet the required sample size. All members of the selected households, aged 2 years and older and available on the survey day, were invited to participate in the study.
Potential animal reservoir hosts, i.e., dogs, cats, pigs and buffaloes, from selected households were also enrolled and examined for helminth infections. Due to the small numbers of these animals in the study villages (0.4 animals per household; from village record), we examined all of those present during the survey. Village health volunteers helped to identify the domestic animals and conduct follow-up examinations. We collected intermediate hosts for O. viverrini (Bithynia spp. snails and Cyprinoid fish) and for S. mekongi (Neotricula aperta snails) from selected sites in the study villages and examined them for infection (Fig. 1b).
Snails of the genus Bithynia spp. were collected with a scoop [38] from water bodies near the study villages (e.g., ponds, canals, and rice fields). From each water body, 5-10 sites with an area of 1 × 1 m were identified as collecting points. All Bithynia snails collected from each site were counted, recorded and examined separately. Cyprinoid freshwater fish were captured from the same selected water bodies as well as from the Mekong using a fishing net. Each captured fish was measured for length and weight and were examined at the field station for the presence of O. viverrini metacercariae.
N. aperta snails [39] live in the rocky area of the Mekong River. We identified 10 sites along the Mekong River, where water was frequently used by study villagers for their daily needs. Submerged stones were dredged and snails were hand-picked from them [38]. At each site, N. aperta snails were collected for 20 min by five malacologists. All collected snails were counted, placed in a plastic bag and carried to the field station for examination.

Field procedures and laboratory examinations
In each village, a house, school or temple was identified as a field study station. Two questionnaires were administered to all participating households. A household questionnaire was administered to the heads of households for collecting data on household characteristics (e.g., building type, toilette and water supply), asset ownership (e.g., farm engine, boat, car, motorbike, electricity, television, bicycle, telephone and agriculture land) and animal ownership (e.g., buffalo, cow, goat and pig). An individual questionnaire was used to interview all household members to collect demographic data (e.g., age, sex, educational attainment and professional activities and behavioural risks (e.g., food consumption habits, water contact, animal raising and personal hygiene). Parents or legal guardians answered for children under 10 years of age.
Eligible study participants were invited to submit two stool samples over consecutive days for parasitological analysis. The first stool container (pre-labelled with participant's name, unique identity number, age and date of collection) was handed to the study participants on the registration day, along with a detailed explanation of stool collection. The second empty container was handed out after study participants returned the first filled container.
Two Kato-Katz (KK) thick smears [40] were prepared from each stool sample (i.e. four smears per person) and examined under light microscopes by an experienced technician within 1 h of sample preparation. Eggs were counted and recorded for each helminth species separately.
We collected faecal samples from potential domestic reservoir animals owed by study households, namely cats, dogs, pigs and water buffaloes. To collect fresh faecal samples [41] from small animals (cats, dogs and pigs), rectal enemas were performed using Sodium Chloride (NaCl) solution and petroleum jelly lubricant. Faecal samples from water buffaloes were collected by rectal swab. All faecal samples were immediately preserved in a 10% formalin solution and transported to the National Institute of Public Health (NIOPH), Vientiane, for processing using the formalin ether concentration technique (FECT) [42].
Bithynia spp. and N. aperta snails were examined for the presence of cercariae infection using the shedding test, previously described by Sri-Aroon and colleagues [43]. In summary, the fresh water snails were put into a transparent plastic container filled with Mekong water and exposed to artificial light. After 2 h, the container was examined under a stereoscope for the presence of cercariae. The infected snails were identified, counted and recorded separately.
The species identification of captured Cyprinoid fish was performed based on guidelines available at FishBase website [44,45]. Fish digestion was performed using the pepsin enzyme digestion technique [25]. The residue was examined for the presence of O. viverrini metacercariae. The metacercariae were counted and recorded for each infected fish.

Data management and analysis
Information from questionnaires and data forms were double entered into EpiData, version 3.1 (EpiData Association; Odense, Denmark) and validated for their correctness and completeness. Statistical analyses were performed with STATA, version 13.1 (StataCorp., College Station, USA). Only study participants with at least two KK thick smear examinations and with complete questionnaires were retained in the final analysis. Participants were stratified into five age groups: (i) ≤ 9 years, (ii) 10-16 years, (iii) 17-36 years, (iv) 37-50 years, and (v) ≥ 51 years. Socioeconomic status (SES) of the household was calculated using an asset-based method. Indicator data were defined by principal component analysis (PCA). The procedure is widely used and details can be found elsewhere [5,46,47]. SES conditions in the household were categorized into one of five wealth quintiles, namely (i) most poor, (ii) very poor, (iii) poor, (iv) less poor, and (v) least poor according to their cumulative standardized asset scores. Details of this widely used approach have been presented elsewhere [5].
The intensity of helminth egg counts was expressed as eggs per gram of stool ( [25,31,48]. Prevalence of parasitic infections was determined and stratified by age, sex and study area (Done Khon versus Done Som). Chi-square test was used to examine the association among categorical variables. The geometric mean for helminth egg counts was calculated for infected individuals. Univariate random-effects logistic regression analysis was used to associate O. viverrini and S. mekongi infections (outcome) with potential risk factors (predictors). The crude odds ratio (cOR), 95% confidence interval (95% CI) and P-value were calculated. Explanatory variables with a P-value of <15% were included in the stepwise multivariate random-effects logistic regression model. Adjusted odds ratio (aOR) was calculated. Smoothed age distribution of O. viverrini, S. mekongi, hookworm and T. trichiura infections by gender was established. Statistical significance was defined as yielding a P-value smaller than 0.05.

Characteristics of the study participants
A total of 994 study participants were included in this final analysis (Fig. 2). Of these, 475 (47.8%) were from Done Khon and 519 (52.2%) from Done Som. There were slightly more female than male participants (51.8% vs 48.2%). Age ranged from 2 to 88 years (median age 29.8 years). The schooling rates did not differ between the two study islands. Subsistent rice farming and fishing were the main professional activities (60.0%). Less than half of the study participants reported having access to a latrine at home (Done Khon 49.7%, Done Som 38.9%). People living in Done Som had a lower socioeconomic status than in Done Khon (Most poor, 25.8% vs 16.4%, respectively). The sociodemographic characteristics of study participants are summarized in Table 1.

Helminth infections in humans
Helminth infections were very frequent on the two islands. O. viverrini, hookworm, S. mekongi, and T. trichiura were found in 60.7%, 44.1%, 22.2% and 4.1% of the participants, respectively. Very few participants were infected with A. lumbricoides (0.6%) and Taenia spp. (0.1%). The prevalence of O. viverrini was almost two-times higher in Done Som compared to Done Khon (77.3% vs. 42.5%, P < 0.001). S. mekongi prevalence was similar on both islands (P = 0.329). Multi-parasitism was diagnosed in 40.5% of the study participants. Details of the helminth infections are given in Table 2. Figure 3 displays the smoothed age prevalence of helminth infections by gender. O. viverrini infection appears to be acquired at a young age, with prevalence increasing gradually (Fig. 3a). Hookworm infection is acquired at a very young age. For males, the prevalence peaked among adolescents aged 10-20 years and plateaued among older age groups. For females, prevalence peaked between 10 and 20 years old and again after 50 years old (Fig. 3b). For males, two prevalence peaks were observed; the first among children under 10 years old and the second among adults between 40 and 50 years old. For females, only one peak was seen among children under 10 years old. T. trichiura prevalence was distributed similarly among males and females independent of age (Fig. 3c). S. mekongi prevalence was differently distributed among males and females (Fig. 3d).
Human helminth infection intensities are summarized in Table 3. Most helminth infections were categorized as light infections. Nevertheless, O. viverrini, S. mekongi and hookworm accounted for infections of heavy intensity in some cases (4.2%, 3.6% and 1.8%, respectively).

Discussion
The Khong District, with its dozens of islands in the Mekong, has a distinct ecological setting (Fig. 1). Human settlements line the island shores, while the rest of the island is used for agricultural activities, particularly rice farming. The Mekong River as well as the diverse water bodies on the islands represent a rich ecosystem for fish and mollusc populations. On two Mekong islands, highly endemic for multiple species of helminth infections, we studied the transmission of O. viverrini, S. mekongi and STH using an ecohealth approach [32,37] to better assess the relation of human infection status to environmentally present reservoir and intermediate hosts. Heavy infections and multiparasitism were prevalent among the human population and age-gender distributions revealed parasite-specific patterns. Examination of potential animal reservoir hosts from the study participants' households (cats, dogs, pigs and buffaloes) yielded ten different helminth species, with many of them having zoonotic capacity. Infection rates of intermediate snail hosts Bithynia sp. and N. aperta were low but reflect on-going transmission. In addition, infection rates of locally caught cyprinoid fish with O. viverrini and minute intestinal fluke (MIF) metacercariae were very high, pointing to a high risk of infection when they are consumed raw or undercooked.
In this study, we document high infection rates of O. viverrini, S. mekongi and selected species of STH, namely hookworm infections. The high infection rates are a surprise given that MDA campaigns were conducted annually between 2008 and 2013 [26], in which praziquantel (40 mg/kg BW single dose) and albendazole (400 mg single dose) were provided to the entire population (older than 4 years). In addition, biannual deworming (with mebendazole) takes place in all Lao primary schools [27]. Local health authorities confirmed that all Mekong islands were targeted, but we could not find coherent information on the number of treatment rounds conducted on our study islands. Nevertheless, our results indicate that the impact of the intervention is insufficient.
The Ministry of Health's objective is to eliminate S. mekongi as a public health problem in Lao PDR by 2016. On our study islands, S. mekongi cannot be considered eliminated given the high infection rates. Our data indicate that S. mekongi infection in dogs may fuel the transmission by constantly infecting Neotricula populations in the Mekong. Of similar importance are cats and dogs for the transmission of O. viverrini. Hence, animal reservoirs in households should also be a target of integrated parasite control on the Mekong islands, and throughout Lao PDR.
Several factors might account for the persisting high O. viverrini infection rates among humans on the Mekong islands. One such factor is the high infection prevalence among cyprinoid fish. More than 80 species of the Cyprinidae family and at least 13 species of other families can serve as a secondary intermediate host [25]. In our study, O. viverrini metacercariae were identified in 11 cyprinoid fish species, while some had particularly high O. viverrini metacercariae infection rates, e.g. in 87.1% of Hampala dispa. All the cyprinoid species in which we detected an infection are known to be good O. viverrini transmitting species [49][50][51][52]. They were identified in all water bodies examined in this study. Fish are mostly likely infected while small and living in rice fields, canals and ponds. The metacercariae remain alive as the fish grow and move into the Mekong.
Cyprinoid fish accumulate the metacercariae over a long time. Low infection rates in Bithynia snails may be sufficient for transmission [53]. We found a low infection rate of 0.3% in Bithynia sp. snails. Other studies have detected infection rates between 0.3-8.3% [54]. But infection rates may vary considerably, depending on sampling locality and season [54,55]. It is important to note that even low infection prevalence rates are sufficient for maintaining transmission.   We observed low S. mekongi infection rates in N. aperta (0.02%) compared to other reports. The presence of infected molluscs gives evidence that S. mekongi transmission is currently on-going. Therefore, abandoning control activities would inevitably lead to an increase in infection rates among humans. There are many more S. mekongi endemic Mekong islands, which might display a different N. aperta population distribution and infection pattern [9,10].
A major finding from our study is the dramatically high helminth infection rates among domestic cats, dogs, pigs and buffaloes. Ten different parasite species were detected in these animal hosts residing in the households of our study participants. By using FECT, we could distinguish O. viverrini eggs in dogs and cats from other small trematode eggs. Our results showed higher rates than Aunpromma et al. (2012) found in neighbouring Thailand, where 0.37% and 35.5% of the dogs and cats were infected, respectively [56]. The infection rate among dogs, in particular, was 20 times higher than that found in the study of Aunpromma et al. (2012). Through observation and from interviewing animal owners in both communities, it appears that most of the dogs and cats were free-roaming and usually accompanied their owners to the rice field where they caught and ate fish directly from the canals or rice fields. Moreover, raw and undercooked fish were often fed to these animals. These phenomena, in combination with the high infection rates of dogs and cats, likely maintain the transmission of O. viverrini and other fish-borne trematode infections in the communities.
Only dogs were diagnosed with S. mekongi in this study, which is consistent with other study findings [9,22]. We did not find any S. mekongi eggs in pigs or water buffaloes, though both animals were found to be infected in earlier investigations [57]. However, they are not of importance for transmission on our study islands. On other Mekong islands where these animals are more free-roaming, their infection status could be higher and, thus, their contribution to transmission of greater importance.
The results of our risk factor analysis for O. viverrini infection differed from many previous studies [5,30,49]. More than half of our risk factors dropped out after multivariate analysis, whereas the initial univariate analysis showed significant associations between infection and age group, occupation, socioeconomic status, latrine availability, history of open defecation this year, and eating raw and/or undercooked fish ( Table 6). The association between O. viverrini and socioeconomic status was not clear for our study population. The study area was geographically very small. Therefore, the variation in socioeconomic status and living conditions might not have varied enough to results in risk differentiation. Furthermore, control activities such as the annual treatments between 2008 and 2013, have had an impact on infection status, which in turn might have blurred important associations. For example, eating raw/ undercooked fish was not significantly associated with O. viverrini infection, although deeply rooted habits of eating raw or improperly cooked fish is a well-known factor in sustaining helminth infections in humans and difficult to control [30,53,58].
In our multivariable analysis, we did not find any association between S. mekongi infection and risk factors, except for age. Children under 9 years old had a higher risk of infection than older study participants. This result is likely due to MDA over the years having reduced infection rates among older villagers. Therefore, controls targeting lower age groups could further contribute to eliminating S. mekongi on the Mekong islands.
Our study suffers from some limitations. Our diagnostic procedure most likely underestimated the true infection burden. Although examining a duplicate Kato-Katz thick smear per faecal sample has a considerably higher sensitivity than a single smear, the egg detection rate remains far below that of a multiple stool sample diagnostic procedure [5,18]. Furthermore, the Kato-Katz technique cannot differentiate small trematode eggs [59]. It is therefore possible that some of the infections in humans were counted as O. viverrini infections instead of MIF.