The impact of a small-scale irrigation scheme on malaria transmission in Ziway area, Central Ethiopia
2009; Wiley; Linguagem: Inglês
10.1111/j.1365-3156.2009.02423.x
ISSN1365-3156
AutoresSolomon Kibret, Yihenew Alemu, Eline Boelee, Habte Tekie, Dawit Alemu, Beyene Petros,
Tópico(s)Insect Resistance and Genetics
ResumoObjective To assess the impact of a small-scale irrigation scheme in Ziway area, a semi-arid area in the Central Ethiopian Rift Valley, on malaria transmission. Method Parasitological, entomological and socio-economic studies were conducted in a village with and a village without irrigation. Blood smear samples were taken from individuals during the dry and wet seasons of 2005/2006. Socio-economic data were collected from household heads and key agricultural and health informants through interviews and questionnaires. Larval and adult mosquitoes were sampled during the dry and short wet seasons of 2006. Female anopheline mosquitoes were tested by enzyme-linked immunosorbent assay for blood meal sources and sporozoite infections. Results Malaria prevalence was higher in the irrigated village (19%, P < 0.05) than the non-irrigated village (16%). In the irrigated village, malaria prevalence was higher in the dry season than in the wet season while the reverse occurred in the non-irrigated village. Households with access to irrigation had larger farm land sizes and higher incomes, but also higher prevalence of malaria. Larval and adult abundance of the malaria vectors, Anopheles arabiensis and Anopheles pharoensis, was higher in the irrigated than in the non-irrigated village throughout the study period. Furthermore, the abundance of An. pharoensis was significantly higher than that of An. arabiensis during the dry irrigated period of the year. Canal leakage pools, irrigated fields and irrigation canals were the major breeding habitats of the two vector mosquitoes. Plasmodium falciparum sporozoite infection rates of 1.18% and 0.66% were determined for An. arabiensis and An. pharoensis in the irrigated village. Peak biting activities of the vectors occurred before 22:00 h, which is a source of concern that the effectiveness of ITNs may be compromised as the mosquitoes feed on blood before people go to bed. Conclusion Irrigation schemes along the Ethiopian Rift Valley may intensify malaria by increasing the level of prevalence during the dry season. To reduce the intensity of malaria transmission in the small-scale irrigation schemes currently in operation in Ethiopia, year-round source reduction by using proper irrigation water management, coupled with health education, needs to be incorporated into the existing malaria control strategies. Objectif: Evaluer l'impact d'un régime d'irrigation à petite échelle dans la région de Ziway, une zone semi-aride dans la vallée du Rift dans le centre de l'Ethiopie, sur la transmission de la malaria. Méthode: Des études parasitologiques, entomologiques et socio-économiques ont été conduites dans un village avec et un village sans irrigation. Des frottis d'échantillons de sang ont été obtenus des personnes pendant les saisons sèches et humides de 2005/2006. Les données socio-économiques ont été recueillies auprès des chefs de ménage et des informateurs clés agricoles et de la santé au moyen d'interviews et de questionnaires. Des larves et des moustiques adultes ont étééchantillonnés au cours de la saison sèche et de la courte saison humide de 2006. Les moustiques anophèles femelles ont été testés par ELISA pour les sources d'origine du sang et pour les infections à sporozoaires. Résultats: La prévalence de la malaria était plus élevée dans le village irrigué (19,2%; P < 0,05) que dans le village non-irrigué (16,0%). Dans le village irrigué, la prévalence de la malaria était plus élevée pendant la saison sèche que pendant la saison des pluies alors que l'inverse était observé dans le village non-irrigué. Les ménages ayant accès à l'irrigation avaient des terrains agricoles de plus grandes tailles et des revenus plus élevés, mais aussi une plus forte prévalence de malaria. Une abondance de larves et d'adultes des vecteurs de la malaria, An. arabiensis et An. pharoensisétait plus élevée dans le village irrigué que dans le non-irrigué tout au long de la période étudiée. En outre, l'abondance d'An. pharoensisétait significativement plus élevée que celle d'An. arabiensis durant la période sèche irriguée de l'année. Les canaux de fuites des bassins, les champs irrigués et les canaux d'irrigation étaient les principaux habitats de reproduction des deux moustiques vecteurs. Des taux d'infection à sporozoïtes de P. falciparum de 1,18% et 0,66% ont été déterminés pour An. arabiensis et An. pharoensis dans le village irrigué. Le pic des activités de piqûre des vecteurs a eu lieu avant 22h00, ce qui est une source de préoccupation que l'efficacité des moustiquaires à insecticide pourrait être compromise car les moustiques se nourrissent de sang avant l'heure de coucher des gens. Conclusion: Les systèmes d'irrigation dans la vallée du Rift éthiopien pourraient intensifier la malaria en augmentant le niveau de prévalence au cours de la saison sèche. Afin de réduire l'intensité de la transmission de la malaria dans les petits systèmes d'irrigation à petite échelle actuellement en opération en Éthiopie, une réduction des sources tout au long de l'année par l'utilisation d'une gestion appropriée des eaux d'irrigation, associée à l'éducation sanitaire, doit être intégrée dans les stratégies existantes de lutte contre la malaria. Objetivo: Evaluar el impacto sobre la transmisión de malaria de un esquema de irrigación a pequeña escala en el área de Ziway, un área semi-árida del Gran Valle del Rift de Etiopía Central. Método: Se realizaron estudios parasitológicos, entomológicos y socio-económicos en un poblado con y un poblado sin irrigación. Durante las estaciones seca y lluviosa del 2005/2006 se tomaron muestras de sangre para láminas de los individuos del estudio. Mediante entrevistas y cuestionarios a los cabeza de familia e informantes clave agricultores y sanitarios, se recolectaron los datos socio-económicos. Se muestrearon mosquitos adultos y larvas durante las estaciones seca y lluviosa del 2006. Los mosquitos anofelinos hembra fueron testados mediante un ensayo ELISA para fuentes de comida e infecciones por esporozoitos. Resultados: La prevalencia de malaria era más alta en el poblado con irrigación (19.2%, P < 0.05) que en el no irrigado (16.0%). En el poblado con irrigación la prevalencia de malaria era más alta en la estación seca que en la lluviosa, mientras que lo contrario ocurría en el poblado no irrigado. Los hogares con acceso a irrigación tenían terrenos cultivables más grandes y mayores ingresos, pero también una mayor prevalencia de malaria. La abundancia de larvas y adultos de los vectores de malaria, An. arabiensis y An. pharoensis fue mayor en el poblado irrigado que en el no irrigado durante todo el periodo de estudio. Más aún, la abundancia de An. pharoensis era significativamente mayor que la de An. arabiensis durante el periodo seco e irrigado del año. Las piscinas de drenaje de los canales, los campos irrigados y los canales de irrigación fueron los principales hábitats de reproducción de los dos mosquitos vectores. Las tasas de infección de esporozoitos de P. falciparum de 1.18 y 0.66% se determinaron para An. arabiensis y An. pharoensis en el poblado irrigado. Los picos de actividad de mordedura de los vectores ocurría antes de las 22:00 horas, siendo una fuente de preocupación ya que la efectividad de las mosquiteras puede estar comprometida puesto que los mosquitos se alimentan de sangre antes de que las personas se vayan a la cama. Conclusión: Los esquemas de irrigación a lo largo del Valle de Rift Etiope podrían intensificar la malaria al aumentar los niveles de prevalencia durante la estación seca. Con el fin de reducir la intensidad de la transmisión de malaria en los esquemas de irrigación a pequeña escala en operación en Etiopía, se requiere incorporar a las estrategias de control de la malaria ya existentes, una reducción de las fuentes durante todo el año manejando adecuadamente la irrigación apropiada, junto con una educación sanitaria. Development of irrigation schemes is widely recognized as a key for promoting economic growth, ensuring food security and alleviating poverty in most developing countries (Lipton et al. 2003). However, past experience shows that inadequate consideration of public health aspects can seriously undermine the sustainability of such schemes (Hunter 1993; McCartney et al. 2007). Key among the potential negative impacts is the link between irrigation and malaria – a disease that affects between 300 and 500 million people each year globally and claims the lives of 1.5–2.5 million people annually (Keiser et al. 2005; WHO 2008). Especially in areas where malaria is unstable, irrigation may alter the malaria transmission pattern from seasonal to annual (Marramaa et al. 2004;Sissoko et al. 2004). However, the relationship between irrigation and malaria is not straightforward (Lindsay et al. 1991; Boudin et al. 1992; Henry et al. 2003). In some cases, the most anthropophilic malaria vector Anopheles funestus is replaced by Anopheles arabiensis that has lower vectorial capacity and thrives better in irrigated fields (Ijumba & Lindsay 2001). Higher mosquito densities in irrigated areas may lead to increased inter-species competition and reduced longevity (Dolo et al. 2004). Finally, communities near irrigation schemes may benefit from the greater wealth created by the schemes and get better access to improved health care and protective (Klinkenberg et al. 2004). In Ethiopia, where three-quarters of the land are potentially malarious, introduction or expansion of irrigation schemes can increase the burden of malaria in the country. In the highlands of Tigray, northern Ethiopia, malaria incidence in young children was sevenfold higher in communities near small dams than in those farther away (Ghebreyesus et al. 1999). However, despite extensive development of irrigation schemes in semi-arid fertile areas of the country with unstable disease transmission (Ministry of Water Resources 2002), detailed understanding of the link between irrigation and malaria in such settings is lacking. The main objective of this study was to analyse the impact of small-scale irrigation on malaria in a semi-arid area with seasonal disease transmission. The study was undertaken between September 2005 and May 2006 in two rural farming villages, Abene-Girmamo and Woshgulla, located in the Ziway area in the Ethiopian Rift Valley, Central Ethiopia, 165 km south of Addis Ababa. The area receives between 700 and 800 mm of annual rainfall, with the main rains from June to September and short rains in April and May (National Meteorological Agency, unpublished data). The mean annual temperature is 20 °C. Malaria transmission in Ziway is generally unstable and seasonal, with peak transmission between September and November, immediately after the main rainy season and a secondary transmission period between April and May in the short rainy season. Plasmodium falciparum is the most prevalent malaria parasite, responsible for 60–70% of malaria cases, followed by Plasmodium vivax (Abose et al. 1998b). Anopheles arabiensis is the primary malaria vector, while Anopheles pharoensis plays a secondary role (Rishikesh 1966; Abose et al. 1998a,b; Ye-ebiyo et al. 2000). Abene-Girmamo, the 'irrigated village', is situated at an altitude of 1647 m. The village was inhabited by 934 people in 2005, mainly dependent on irrigated subsistence farming during the dry season. Most families owned livestock (mainly bovine, ovine and equine), with a human to cattle ratio of 1:0.4. Irrigating farmers often stayed outdoors late, working on their field. In the rainy season, farmers practice rain-fed agriculture. Woshgulla, the 'non-irrigated village', is an agricultural village without irrigation, situated at an altitude of 1654 m, with a population size of 741 in 2005. The village is located 8 km away from the irrigation scheme in Abene-Girmamo. The inhabitants were dependent on subsistent rain-fed agriculture during the wet seasons and on livestock (mainly bovines and some equines) rearing. The mean human to cattle ratio in the village was 1:0.6. In both study villages, livestock spent the night either indoors in the same homesteads as the owners or outdoors in open cattle enclosures. Borrow pits for making mud bricks were commonly found at the backyards of houses. Each village had a water-harvesting pool (approximately 2 by 2 m wide and 6 m deep) covered with corrugated iron. Four different types of houses were distinguished in the area based on combinations of walls made of mud bricks or of wood plastered with mud, and roofs made of corrugated iron sheets or grass tops. The water source for irrigation in Abene-Girmamo is Lake Ziway, located 5–6 km away from the scheme. Water is pumped from the lake through three long underground pipes to the unlined elevated surface canals. These main irrigation canals feed furrows that cover the agricultural fields. Onion, cabbage, tomato and maize were the main irrigated crops throughout most of the year. Previous parasitological surveys conducted in the same study area found 3.5–12.6% prevalence during the main transmission period (July–December) (Abose et al. 1998a). Combined with time series recorded malaria data (1999–2005) from the Ziway Health Center, an average sample size was determined as 500 people per village by using Win Episcope 2.0 computer software (Wageningen University, Netherlands). Accordingly, 220 households (on average 4–5 people per household) were randomly selected for the parasitological surveys: 114 (51.8%) in the irrigated village and 106 (48.2%) in the non-irrigated village. In the study villages, finger prick blood samples were taken by experienced health technicians twice from a total of 2435 individuals, during the long rainy season (September/October 2005) and the dry season (January/February 2006). The slides with thick and thin blood smears were prepared, properly labelled, air dried and examined under a 100× oil immersion objective. Ethical clearance was provided by the Federal Ministry of Health under the umbrella arrangement for Addis Ababa University, and positive individuals were treated following national guidelines (MoH 2004). For the socio-economic study, a household was recorded as positive when at least one of the family members was found positive for malaria. A socio-economic survey was performed using a structured questionnaire in the same households selected for the parasitological study, after the household heads had signed a consent form. The questions elicited information on family size, composition of households (age and sex), farm size, average annual income per household, type of agricultural system (irrigation or rainfed), livestock ownership and housing condition (based on Klinkenberg et al. 2005). Anopheline larvae were sampled from the study villages for 8 days each in February and March in the dry season and in April and May in the short rainy season of 2006. At each survey, all available potential mosquito-breeding habitats were surveyed using a quantitative method (Amerasinghe & Munasingha 1988). Because of poor construction and lack of maintenance, most of the irrigation canals were leaking, creating permanent pools at undesirable places. Sometimes, poorly laid out furrows formed waterlogged areas in the field. The surface area of each potential mosquito-breeding site was estimated in square metres (m2), and sampling was performed at a rate of 6 dips/m2 (Amerasinghe et al. 2001). Larval anopheline samples were transferred to separate vials and killed by gently heating and preserved in 70% alcohol. Adult anophelines were sampled indoors and outdoors for 10 consecutive nights each between February and March in the dry season and between April and May in the short rainy season of 2006. A total of 12 (six indoors and six outdoors) Centre for Disease Control and Prevention (CDC) light traps (Model 512; J. W. Hock Co., Atlanta, USA) was operated in each village from 1800 to 0700 h throughout each sampling night, collecting samples hourly on some days. Each indoor light trap was hung near a person sleeping under an untreated bed net (Yohannes et al. 2005). Outdoor light traps were hung on trees at close proximity to open cattle enclosures where some individuals spent the evening. Hourly mosquito collections were also conducted by light traps indoors and outdoors. Preserved anopheline larval samples were counted, and third and fourth larval instars individually mounted on microscope slides for species identification according to Verrone (1962b). All collected adult anophelines were sorted out into species (Verrone 1962a). Female anophelines were stored in a silica gel desiccator, transported to the Biomedical Science Laboratory at Addis Ababa University and kept there at room temperature (19–22 °C) for further processing. The head–thorax portion of each dried female anopheline was tested for the presence of P. falciparum and P. vivax sporozoite antigens using enzyme-linked immunosorbent assay (ELISA) (Wirtz et al. 1987). The direct ELISA procedure described by Beier et al. (1988) was used to determine the sources of blood meals (human vs. bovine) of the engorged female anophelines. Chi-square test was used to compare variations in malaria prevalence between villages and seasons. To identify household factors associated with malaria prevalence, logistic regression was applied. The abundance of larval and adult anophelines was compared between villages and seasons using non-parametric Mann–Whitney U-test. The relative abundance of Anopheles species in the irrigated village was compared using Kruskal–Wallis Test. The sporozoite infection rate of each Anopheles species was expressed as the proportion of mosquitoes containing malaria sporozoite antigen in the total samples of a species tested by ELISA. The human blood index (HBI) for each Anopheles species was calculated as the proportion of samples positive for human blood in the total blood meals of a particular species tested. All analyses were carried out using Microsoft Excel 2003 and statistical software, spss version 13 (SPSS Inc, Chicago, IL, USA). Table 1 shows the seasonal prevalence of malaria in the irrigated and non-irrigated study villages. Malaria prevalence was higher in the irrigated village (18.9%; P < 0.05) than in the non-irrigated village (16.1%) during the study period. However, there was seasonal variation in malaria prevalence between the two study villages. In the irrigated village, prevalence of P. falciparum infection was higher in the dry season (21%; P < 0.05) than in the wet season (9%) while prevalence of P. vivax infection became higher in the wet season (7.0%; P < 0.05) than the dry season (1.7%). In the non-irrigated village, prevalence of P. falciparum and P. vivax was higher in the wet season than in the dry season (Table 1). As age increased, malaria prevalence decreased in both villages. However, malaria prevalence was higher in all age groups during the dry season in the irrigated village than in the wet season and the reverse for the non-irrigated village (Table 2). Malaria prevalence was higher in households that use irrigation (mean malaria prevalence = 74.2; 95% CI = 69.9–78.5; P < 0.001) than in those using only rain fed in the irrigated (60.6%; 95% CI = 54.5–66.5) or non-irrigated (40.2; 95% CI = 36.7–43.7) villages (Figure 1). Among 114 households surveyed in the irrigated village, 97 (85.1%) were farmers and the rest (n = 17; 14.9%) were non-farmers (e.g. traders and government employees). Among 106 household surveyed in the non-irrigated village, 92 (86.8%) were farmers and the rest (n = 14; 13.2%) were government employees (5) or without a specific job (9). Among farming households in the irrigated village, 66 (68.0%) practiced rain-fed agriculture and 31 (32.0%) were both rain-fed user and irrigation user. Household malaria prevalence in 2006 by mode of agriculture in irrigated and non-irrigated villages in Ziway, Ethiopia (vertical bars show 95% confidence interval). In the non-irrigated village, bigger farm size (more than 1.5 hectares) was associated with a decrease in household malaria prevalence. In the irrigated areas, the household malaria prevalence hardly varied over three categories of farm size (Figure 2). On average, income per person per day of rain-fed producers and irrigation users was 0.15 and 0.56 US American dollar (USD), respectively. A change in house type from grass top to corrugated iron roof was associated with a decrease in malaria prevalence (Table 3). Surprisingly, an increase in the total annual income of households was positively associated with an increase in malaria infection prevalence. A history of malaria in the household malaria was also positively associated with malaria infection prevalence. Household malaria infection prevalence in different farm size categories in irrigated and non-irrigated areas in Ziway, Ethiopia. (vertical bars show 95% confidence intervals) (1 kert is 0.25 hectar). Four-times more positive Anopheles larval sites were encountered in the irrigated village (n = 51) compared to the non-irrigated village (n = 12) during the study period (Table 4). Consequently, a higher Anopheles larval density was found in the irrigated village (36.0 larvae per 100 dips; z = −3.196, P < 0.001) than in the non-irrigated village (14.9 per 100 dips) throughout the study period. In the irrigated village, Anopheles breeding occurred both in the dry and in the wet seasons while larval occurrence was limited to the wet season in the non-irrigated village (Table 4). Five species anopheline larvae were collected, among which An. arabiensis, An. pharoensis and Anopheles coustani were the major ones (Table 5). In the irrigated village, canal leakage pools and irrigated field puddles were the most important sources of An. arabiensis, whereas canal leakage pools and irrigation canals were the major larval habitats for An. pharoensis. Overall, around 80% of the total Anopheles larval occurrence in the irrigated village was from habitats associated with the irrigation scheme. In the non-irrigated village, borrow pits and rain pools were the only larval habitats (Table 5). A total of 1271 adult anophelines were collected from the two study villages during the study period, of which 94% (n = 1213) and 6% (n = 58) were from the irrigated and non-irrigated villages, respectively (Table 6). In the irrigated village, An. pharoensis was the major species in the dry season (56.9%; n = 340; χ2 = 52.294; df = 2; P < 0.001) while An. arabiensis dominated in the short rainy season (50.2%; n = 309; χ2 = 17.751, df = 2, P < 0.001). The majority (65.5%, n = 38) of anophelines in the non-irrigated village were An. arabiensis. No adult mosquitoes were caught in the non-irrigated village during the dry season. Peak indoor and outdoor activities of An. arabiensis were observed during the early period of the night, between 18:00–19:00 and 19:00–20:00 h, respectively (Figure 3). Thereafter, its activity steadily decreased both indoors and outdoors throughout the rest of the night. Peak indoor and outdoor activities of An. pharoensis occurred between 20:00–21:00 and 19:00–20:00 h, respectively, which declined thereafter, but with a remarkable second peak between 22:00–23:00 h outdoors. For An. coustani, peak indoor and outdoor activities were recorded between 18:00–19:00 h, which sharply dropped thereafter but with a remarkable secondary peak between 22:00–23:00 and 05:00–06:00 h, indoors and outdoors, respectively. Overall, about 75%, 66% and 69% of the bite by An. arabiensis, An. pharoensis and An. coustani, respectively, occurred during the early period of the night (before 22:00 h), before the villagers retire to bed (Figure 2). Hourly activity of Anopheles arabiensis (top), Anopheles pharoensis (middle) and Anopheles coustani (bottom) indoors and outdoors from light trap catches (as percentage of mosquitoes collected each hour; n = 24 light trap nights in all cases) in an irrigated area in Ziway, February to May 2006. Among 120 blood-fed An. arabiensis specimens, 70.8% (n = 85) and 14.2% (n = 17) were found positive for only human and bovine blood, respectively. Some (7.5%, n = 9) had mixed blood meals of human and cattle, and the remaining were unidentified (7.5%, n = 9). The HBI for An. arabiensis was found to be 0.78. Of 142 blood-engorged female An. pharoensis specimens, 61.3% (n = 87) and 20.4% (n = 29) had human and bovine blood meals, respectively. Some blood meals (7.7%, n = 11) were mixed, and the remaining blood meals were not identified. The HBI for An. pharoensis was 0.69. Of 16 blood-engorged An. coustani specimens, only one (6.2%; HBI 6%) was positive for human blood while the majority (75%, n = 12) gave positive results for bovine blood. In the irrigated village, among 424 female An. arabiensis and 509 An. pharoensis specimens tested for P. falciparum sporozoites, five (1.18%) and three (0.59%) were positive, respectively. Seasonally, higher P. falciparum sporozoite rate of An. arabiensis was recorded in the short rainy season (1.47%; 4/272) than in the dry season (0.66%; 1/152). The P. falciparum sporozoite rate of An. pharoensis was 0.92% (3/325) in dry season, while none (0/184) were positive in the short rainy season. None of the tested samples from the irrigated village were positive for P. vivax sporozoite, and none of the samples collected from the non-irrigated villages were positive for malaria sporozoites. Our study found higher larval and adult vector abundance and a higher prevalence of malaria in the irrigated village than the non-irrigated village particularly during the dry season – implying the continuation of malaria transmission in this season where transmission is normally unexpected. The major finding of the present study is that because of breeding of malaria vectors (An. arabiensis and An. pharoensis) in the small-scale irrigation scheme in Ziway, the period of malaria transmission in the irrigated village has extended into the dry season. Higher malaria prevalence in the dry season, as observed in households that use irrigation, could be because irrigation activities continue in the evening when the major malaria vectors' peak biting activities occur. Thus, irrigating farmers would be at higher risk of infective mosquito bites during the dry season. Lower prevalence in the wet season in the irrigated village could be explained by density-dependent mechanisms, such as competition, which might affect the proportion of sporozoite-infected mosquitoes, even resulting in reduced transmission at higher densities (Dolo et al. 2004). Among household factors evaluated in the socio-economic survey, quality of housing was an important health determinant. A decreasing trend in malaria prevalence was observed when the housing quality improved from grass top to corrugated iron roof. This finding is in agreement with previous studies elsewhere (Konradsen et al. 2003; Deressa et al. 2007). Household malaria prevalence was also positively associated with previous malaria infection history, which could be because of the clustering of mosquitoes to specific houses close to mosquito-breeding sites within a village (Ribeiro et al. 1996; Brooker et al. 2004). The entomological survey indicated the most important prolific Anopheles larval habitats to be poorly constructed irrigation canals (with stagnant water), canal leakage pools and puddles in waterlogged fields, especially during the dry season. The same breeding habitats have been productive breeding grounds for An. arabiensis elsewhere (Ijumba et al. 1990; Yohannes et al. 2005; Muturi et al. 2006). In the non-irrigated village, only larvae and no adult mosquitoes were found during the dry season, possibly because CDC traps are not sensitive enough for sampling low-density populations (Line et al. 1990). We found peak indoor and outdoor activities of An. arabiensis, An. pharoensis and An. coustani during the early period of the night (before 22:00 h), coinciding with the evening activities of the people in the study area. Similar early biting behaviour was previously reported elsewhere in Ethiopia (Abose et al. 1998b; Taye et al. 2006). The early biting activity of An. arabiensis could be a consequence of long-term application of residual insecticides, particularly DDT, as it has also been suggested for northern Ethiopia (Yohannes et al. 2005). Such early biting activities of the malaria vector populations are likely to compromise the efficacy of insecticide-treated bed nets as most bites occurred before the villagers go to sleep. This could partly explain why higher income households with larger farm size in the irrigated village did not show lower malaria prevalence because even if they would invest in protective measures such as bednets, these are of little advantage. The reported high HBI for An. arabiensis (0.78) and An. pharoensis (0.69) in the present study reaffirmed the importance of these species in malaria transmission in Ziway area. Comparable HBI has been reported in other parts of the country (Adugna & Petros 1996; Yohannes et al. 2005) suggesting their antropophagic behaviour. The P. falciparum sporozoite rate of An. arabiensis in the present study (1.18%) is comparable to that reported from southern Ethiopia (1.1%, Habtewold et al. 2001; Taye et al. 2006; Tirados et al. 2006), but lower than the sporozoite rate in the adjacent Wonji area (1.5%, Ameneshewa 1995). In our study, the P. falciparum sporozoite rate of An. pharoensis in the dry season confirms the vectorial role of this species in malaria transmission in the irrigated villages of Ziway area. Anopheles arabiensis was infected with P. falciparum sporozoites both in the dry and in the short rainy seasons, suggesting that this species plays a significant role in malaria transmission throughout the year. These findings could be the first report of confirmed P. falciparum infection in vector mosquitoes in the dry season and also for An. pharoensis that could play an important role in transmitting P. falciparum. In conclusion, although development of irrigation schemes is of paramount importance to increase crop yield and enhance food security in Ethiopia, our findings underscore the importance of irrigation schemes in semi-arid areas like Ziway in maintaining malaria transmission particularly during the dry season, when mosquito abundance is normally limited. The findings also suggest a need for health education so that irrigators would invest in personal protection measures (e.g. protective clothing and repellents in the evening), rather than to rely on bednets that are no longer very effective with mosquitoes biting earlier. Proper water management and canal maintenance for source reduction through environmental management could help to reduce mosquito-breeding sites and thus malaria transmission. This study was partly funded by the Austrian Government through a collaborative research project on 'Impact of Irrigation Development on Rural Poverty and the Environment' and partly by the Dutch Government through the 'Systemwide Initiative on Malaria and Agriculture' (activity 4644) and Addis Ababa University. We are grateful to the Ethiopia Health and Nutrition Institute for allowing us to use the Vector Biology Laboratory. We thank Mr Deribe Feyissa, Mr Eba Tucho and Mr Hailu Engida and Mr Asegid Taye for their assistance in the field and in the laboratory. We also acknowledge the kindly cooperation of the inhabitants of Abene-Girmamo and Woshgulla during the field work.
Referência(s)