Species identification and vector incrimination
Phlebotomus orientalis is a member of the Larroussius (Nitzulesco, 1931) subgenus of Phlebotomus sand flies (previously known as the major group) that contains most principal vectors of Leishmania donovani complex in the old world (Abonnenc 1972). Interestingly, P. orientalis is the only Larroussius sand fly found in the kala-azar endemic foci in the lowlands of East Africa. All other East African Larroussius species are found in the highlands of Ethiopia and Kenya.
Parrot (1936) made the first description of Phlebotmous orientalis, as a variety of Phlebotomus langeroni (P. langeroni orientalis). Ten years later, Parrot and Clastrier (1946) raised P. l. orientalis to specific rank.
Kirk and Lewis (1955) provided the first evidence that P. orientalis is the main L. donovani vector in Sudan. Initially these authors had some doubts as to whether P. orientalis existed in sufficient numbers to maintain transmission of L. donovani. However, they later related the scarcity of this species to a deficiency in the collection methods used, which relied mainly on oiled-paper traps (Kirk and Lewis 1940, 1947). When human-baited and animal-baited traps were used instead of oiled-paper traps, the catches contained higher proportions of P. orientalis and other Phlebotomus sand flies (Lewis and Kirk 1954).
Further evidence supporting the role of P. orientalis as the vector of L. donovani in Sudan came from experiments that compared the susceptibility of P. orientalis and other abundant man-biting sand flies, namely Phlebotomus papatasi (Scopoli) and Sergentomyia clydei Sinton, to L. donovani infection. When Kirk and Lewis (1955) fed wild-caught flies of the three species on subjects with post kala-azar dermal leishmaniasis (PKDL) syndrome, only P. orientalis was shown to be capable of harboring mature infections of L. donovani. Therefore, the authors concluded that in addition to being of the “major” (now Larroussius) group, P. orientalis was the only species that fulfilled the criterion of anterior-station development of L. donovani, and hence appeared to be the probable vector of kala-azar in the Blue Nile and eastern Sudan (Kirk and Lewis 1955). The authors suggested that P. martini and P. lesley Kirk & Lewis may be the vectors of L. donovani in other areas of Sudan, where P. orientalis is not found, e.g., the Kapoeta focus of VL in the extreme south–east, near the Kenyan borders.
Vector incrimination of P. orientalis was supported by the findings of the NAMRU-3 (U.S. Navy Medical Research Unit Number 3) group in the Paloich area of the Upper Nile Province, southern Sudan (Hoogstraal and Dietlein 1963, 1964, Quate 1964, Hoogstraal and Heyneman 1969). In these studies, P. orientalis was the only human-biting sand fly found naturally harboring promastigotes of L. donovani, with an infection rate ranging between 1.9% and 5% and averaging 2.5% in 4,553 flies dissected in 1962–64. The NAMRU-3 group also conducted comparative experimental infections of L. donovani in P. orientali and P. papatasi, the only other abundant man-biting sand fly in the area. The results of these experiments confirmed the findings of Kirk and Lewis (1955) that P. orientalis is the only man-biting sand fly in the area that can harbor mature infections of L. donovani (Heyneman 1963, McConnel 1964).
More recently, P. orientalis was incriminated as the vector of L. donovani in a new focus of visceral leishmaniasis in the Western Upper Nile region of southern Sudan (Schorscher and Goris 1992) and in woodlands and villages of eastern Sudan (Elnaiem et al. 1998a, Elnaiem and Osman 1998, Hassan et al. 2004). Again, a high infection rate (9.6%) with L. donovani, was detected in P. orientalis females captured in Western Upper Nile Province (Schorcher and Goris 1991). Similarly, in a thicket of Acacia seyal in Dinder National Park, the infection rates of L. donovani in P. orientalis ranged between 6.9% in March (middle of the dry season) to 3.6% in June (end of the dry season and beginning of the rainy season) (Elnaiem et al. 1998a).
Earlier reports on sand fly vectors of L. donovani in Kenya indicate that P. orientalis is not found in numbers sufficient to be involved in the transmission of L. donovani anywhere in the country (Johnson et al. 1993, Perkins et al. 1988, Minter et al. 1962). However, recently Ngumbi et al. (2010) suggested that P. orientalis is the only probable vector of VL in villages situated close to Acacia seyal – Balanites aegyptiaca woodland in Isiolo District, where an outbreak of the disease had occurred. The authors found a relatively high abundance of P. orientalis in the area, with a complete absence of P. martini and other human-biting sand flies. However, no attempts were made to detect natural infections of L. donovani in sand flies captured during the study.
Vector habitat and risk-mapping
Phlebtomus orientalis has been reported to exist in wide areas in most countries of East Africa (Figure 1). However, within these areas it is confined to specific habitats, with high abundance of Acacia seyal and B. aegyptiaca trees that are growing on black cotton soil (Kirk and Lewis 1955, Quate 1964, Dietlein 1964, Hoogstraal and Heyneman 1969, Zeese and Franke 1987, Ashford and Thomson 1991, Schorscher and Goris 1992, Elnaiem et al. 1997, 1998, 1999a, Thomson et al. 1999) (Figure 2). As documented in the Upper Nile province, P. orietnalis was scarce or totally absent from villages except those situated at the edge of the forest (Quate 1964, Hoogstraal and Hyeneman 1969). From various reports, it appears that the association of P. orientalis with Acacia-seyal Balanites aeyptiaca is so strong that the vector abundance is inversely proportional to distances from the Acacia seyal- B. aegyptiaca woodland. Quate (1964) and Hoogstraal and Heyneman (1969) showed that in villages situated at a distance of 1.3 km or more from Acacia-Balanites forests, either no human-biting sand flies could be found, or P. papatasi was the predominant species. In villages surrounded by grassland but situated at a distance of 0.8 km from the Acacia-Balanites forest, P. papatasi was also the predominant man biter, while only a small number of P. orientalis appeared during the peak months in the dry season. Elnaiem et al. (1997) demonstrated that within the Dinder National Park woodland, eastern Sudan, the collection of the vector dropped from 67.0 sand flies per light trap night in a thicket of A. seyal, to 3.8 sand flies at 100–200 m away from the trees.
Figure 2. Photographs of typical habitats of P. orientalis in eastern Sudan: A. Dense thicket of Acacia seyal in Dinder National Park, with high abundance of P. orientalis; B. Kala-azar endemic village in eastern Sudan with high density of Balanites aegyptiaca trees and moderate abundance of P. orientalis.
Download figure to PowerPoint
Although P. orientalis was encountered mainly in Acacia seyal–Balanites aegyptiaca woodland area, it is not clear whether this association is due to specific adaptation of the vector to these trees or due to soil or other microclimatic factors in places where these trees are abundant. To address this question and determine the specificity of the association of P. orientalis with Acacia- seyal and B. aegyptiaca woodland, Elnaiem et al. (1999a) conducted two sand fly surveys: one in a 1-km transect in a typical riverine-village-sorghum field habitat of eastern Sudan, and another one in four vegetation types dominated by either A. seyal trees, B. aegyptica trees, Combretum kordofanum trees or riverine bushes of Hypaena thaibaica and Zyzphus spinachristae. The proximity of the vegetation types and the fact that the soil at all of these study sites was black cotton soil reduced the impact of confounding factors, other than vegetation parameters. The results of the surveys in the village habitat showed extreme scarcity of P. orientalis with no significant difference in its abundance, which was extremely low in all study zones. In contrast, comparisons of the different vegetation types in Dinder National Park, showed that abundance of P. orientalis was significantly higher in the thicket of A. seyal than in thickets dominated by B. aeyptiaca, Comretum kordofanum or revierine vegetation. The authors suggested that the high abundance of P. orientalis in the A. seyal vegetation was probably due to many factors acting together or independently to provide a favorable habitat for the vector. These factors included: 1) tree density, with its possible effects on temperature and humidity of microhabitats; 2) associations with mound-building termites or preferred animals hosts; and 3) sources of sugar meals. As a follow up study, Hamilton and Elnaiem (2000) compared profiles of sugars found in wild-caught P. orientalis with sugars found in extracts of leaves, fruits and exudates of plants in the study area. Although no conclusive results were obtained in this study, the authors provided evidence that P. orientalis may obtain sugar meal from fruits of B. aegyptica or some aphid and coccid secretions that were not included in the analysis. The latter hypothesis was based on the fact that sugars found in the samples of P. orientalis included melizitose, which is, known to be a characteristic sugar in aphid and coccid honeydew (Moore et al. 1987).
Although the association of P. orientalis and visceral leishmaniasis with specific habitats has been well-established from a number of studies, difficulties continue to exist in predicting places where the vector is found and where the disease is contracted. Considering the large size of kala-azar endemic countries of East Africa and logistic problems in this region, it would be extremely difficult to establish this knowledge by sampling sand flies from each location. To develop an alternative map based on environmental characteristic of P. orientalis habitat, Elnaiem et al (1998a) and Thomson et al. (1999) conducted a ground-based and a satellite-based survey in 44 woodland sites in the savannah belt of Sudan, extending from the Ethiopian border in the east to areas in Darfur in the west. Based on CDC Light trap collections, data on presence and absence of the vector were correlated with different environmental variables and then used to develop a GIS Model to map the areas where P. orientalis thrives. Based on this analysis, Elnaiem et al. (1998b) reported that the presence of P. orientalis in the Savannah zone of central Sudan is positively correlated with the presence of chromic vertisol soils (black-cotton soil), rainfall, minimum and maximum NDVI (Normalized Vegetation Difference Index), annual mean maximum temperature, minimum daily temperature and the presence of A. seyal and B. aegyptiaca trees. Using a logistic regression multivariate model, Thomson et al. (1999) showed that the annual mean maximum daily temperature and soil type were found to be the best predictors of the presence and absence of P. orientalis and hence they were used to produce a probability risk-map for the vector in Sudan. The authors pointed out that although the analysis had a number of limitations, the model was still able to accurately predict sites known for presence of P. orientalis both inside and outside the study area. In a later study, Gebre-Michael et al. (2004) used environmental variables and historical sand fly data on presence and absence of P. orientalis from 111 sites in East African regions other than Sudan, mainly Ethiopia, Kenya and Somalia, to produce a GIS risk map for this vector in the whole region. In contrast to the findings of Elnaiem et al. (1998b) and Thomson et al. (1999), the analysis by these authors showed that soil type is not a predictor for the distribution of P. orientalis in East Africa. Instead, they found wet season NDVI of 0.05–0.28 and LST of 23–36 as major predictors. Although these variables correctly predicted the positive sites of P. orientalis, they also included vast areas in the desert of northern Sudan and along the Red Sea Coast where P. orientalis is unlikely to be found. It is difficult at this stage to point out the reasons for the discrepancy in the findings of the two studies, which warrant further investigations. As suggested by Gebre-Michael et al. (2004), there is an urgent need to re-visit the models of distribution of P. orientalis by conducting current surveys that take into account the actual density of the vector. Furthermore, it would also be important to verify the nature of soils found at positive and negative sites and provide accurate definition of black cotton soil. Although reliance on the FAO digital soil maps is valuable when plotting the risk-maps, caution must be given when using them to define soil at a specific sand fly collection site.
Indoor vs outdoor vector density and the risk of infection in the villages
Several reports stressed that P. orientalis is rarely found resting or nocturnally active or biting humans inside houses (Quate 1964, Hoogstraal and Heyneman 1969, Elnaiem et al. 1997). However, recently Lambert et al. (2002) reported that 75% of P. orientalis captured in Berber El Fugarra village, near the Atbara River in eastern Sudan, were found indoors. This report contrasts with the findings of Elnaiem et al. (1997) on sand flies in Umsalala Village, located 100 km South West of Berber El fuggara, where the vector was rarely encountered inside huts. These differences may be due to variations in construction materials used in building houses, or other microclimatic conditions resulting from the relatively high abundance of shade trees in Berber El Fugarra. In all cases, the findings of Lambert et al. (2002) indicate that some populations of the vector are more adapted to domestic habitats. Due to their implications in the epidemiology and control of visceral leishmaniasis, it is important that future studies should examine these variations in the behavior of the vector.
Seasonality of P. orientalis was investigated in relatively few studies (Quate 1964, Schorcher and Goris 1991, Elnaiem et al. 1997). Quate (1964) classified the sand flies of the Paloich area of the Upper Nile region in southern Sudan into two groups: “seasonal species”, which appeared only in the dry season, and non-seasonal species”, that were present throughout the whole year. Phlebotomus orientalis was found to be a seasonal species appearing in small numbers at the beginning of the dry season (January-February) and reaching its peak abundance during the second half of the dry season (April) (Quate 1964, Hoogstraal and Dietlein 1964). In eastern Sudan, Elnaiem et al. (1997) investigated the seasonal distribution of P. orientalis for a whole year in the small village of Umsalala, and seven months in the dry and early rainy season in the Dinder National Park (Elnaiem et al. 1997). In the village habitat, the density of the vector was found in small numbers from February to June and disappeared completely in the period August-November. In the Dinder National Park, P. orientalis showed a rising abundance through the early dry season, a marked reduction in the middle of the dry season, and then a sudden flare up of the population at the beginning of the rainy season. Simple correlations with climatic data showed that the fluctuation of the vector population during the dry season is governed by changes in the ambient temperature and relative humidity.
Nocturnal activity and biting rates
Several authors studied the nocturnal activity and presented contrasting data on the biting rates of P. orientalis at different periods of the night. Quate (1964) reported that at wind velocities not higher than 2 m/second, in the Upper Nile region, P. orientalis biting activity took place between 18:30 and 20:30. Sometimes, the biting persisted to a later hour, and occasionally a good collection could be obtained up to 21:30 or 22:00 (Quate 1964). Renewed biting activities were recorded during the cool early mornings at 05:45 and continued up to 07:00 (one h after sunrise). It was also found that the biting occurred in waves spaced at intervals varying from four to 30 min, a rhythm that could not be correlated with temperature, wind velocity, or direction (Quate 1964).
Ashford (1974) showed that in Arabya, Ethiopia, P. orientalis human-biting activity reached a peak shortly after dark, although some flies continued to arrive to bite until the end of the four-h observation period. The author showed that the biting activity was considerably reduced when the temperature fell below 16° C, which was four h after sunset. During warmer nights, the catches were reduced when wind activity was increased.
More recently, Schorscher (personal observation) reported that in Western Upper Nile areas, P. orientalis was found biting throughout the night. Similarly, we observed that in Dinder National Park in eastern Sudan, the hourly light trap and human-landing collections of P. orientalis continued until late in the night (Elnaiem and Hassan unpublished data).
Dispersal, breeding, and resting sites
Dispersal and flight range of P. orientalis were studied on only one occasion. Using mark/release/recapture experiments, Quate (1964) showed that although the majority of flies remained within 300 m, some individuals could cover up to 700 m from a point of release.
Despite extensive searches, little is known about the breeding and the resting sites of P. orientalis in Sudan. The research team of NAMRU-3 exerted heroic efforts to find immature stages of P. orientalis and other sand flies. However, examination of large amounts of black cotton soils yielded no results. Although P. orientalis was found resting in tree cavities and in a porcupine burrow in Upper Nile and Ethiopia (Quate 1964, Ashford 1974), no consistent findings have been reported for specific outdoor resting sites of P. orientalis. In one occasion in Dinder National Park in eastern Sudan, P. orientalis was collected exiting from two abandoned termite mounds of Macrotermes herus (Elnaiem et al. 1997). However, the distribution of these mounds is restricted to a small range of the distribution of P. orientalis and no further data were obtained to support these findings. Similarly, although soil cracks have long been suspected as resting/breeding sites for P. orientalis, collections from these places did not provide consistent results. Studies carried out in southern and eastern Sudan ruled out human dwellings as resting sites of the species (Hoogstraal and Heyneman 1969, Elnaiem et al. 1997).
Sources of blood-meal and associations with potential reservoir hosts of Leishmania donovoni
Several studies indicated that, in many foci in East Africa, VL is contracted through intrusion in a zoonotic cycle; but no conclusions were drawn on reservoir hosts (Kirk 1939, Hoogstraal and Heyneman 1969, Elnaiem et al. 2001). As a prelude to discovering the reservoir host of L. donovani, studies were conducted on blood-meal sources and natural hosts of P. orientalis. Quate (1964) conducted a series of experiments in which various small vertebrate hosts were individually confined with wild caught sand flies introduced in small wire-mesh cylindrical cages. Higher feeding rates of P. orientalis females occurred on Rattus rattus (11 out of 18) and Arvicanthis niloticus lectuousus (14 of 57 or 26%) than on gecko, skink, and lizard (average of 0–3%) . Measured by the percentage of engorged females, none of the animals was found particularly attractive to P. orientalis. Recently Hassan et al. (2009) conducted host attractiveness studies on natural populations of P. orientalis in eastern Sudan. The vector showed clear preference to the domestic dog, followed by the Egyptian mongoose and then the genet (Genetta genetta). More recently, Gebre-Michael et al. (2010) reported that 91.6% of blood-fed P. orientalis captured in outdoor and peri-domiciliary sites of Humera-Metema area obtained their blood meals from cattle.
The transmission cycle of kala-azar and possible places and times of infection
Hoogstraal and Heyneman (1969) presented a general model for possible transmission cycles of L. donovani in the Paloich-Malakal area of Upper Nile region of southern Sudan. In this model, the principal zoonotic cycle occurs in the Acacia-Balanites forests. Humans contract the infection when intruding on this environment. Accordingly, the population of Dinka people contracts the infection when travelling with their cattle through the forest to and from the flood plains (Toich). This hypothesis was supported by observations of Schorscher (unpublished data) who reported that people of the Western Upper Nile region also contract the infection when travelling through the forests at night or in villages or temporary cattle camps situated in the forest.
In eastern Sudan conclusive evidence has been provided for the existence of a zoonotic cycle of L. donovani in the woodlands of Dinder National Park (Elnaiem et al. 1998a), with a possibility of the Egyptian mongoose (Hyrpestes ichneumon) acting as a sylvatic reservoir host (Elnaiem et al. 2001). Transmission has also been demonstrated to take place in the villages (Elnaiem and Osman 1998, Hassan et al. 2004). However, it is not clear whether the infection cycle in this habitat also involves zoonosis or is strictly anthroponotic. In light of the high incidence of PKDL cases and the previous experimental infections of P. orientalis on VL patients (McConnel 1964), it may be assumed that transmission from human to human is common.
As in other sand fly vectors of L. donovani, feasible and sustainable methods for control of P. orietnalis are not yet available. The sylvatic behavior of the adults completely eliminates options of vector control through spraying of residual insecticides, and the lack of knowledge on resting and breeding sites precludes any plan to attack the adults or immature stages at outdoor sites. With these considerations, the few attempts to design control measures against P. orietnalis and L. donovani transmission have focused on insecticide fogging of thickets of A. seyal and the use of insecticide-impregnated bednets and insect repellents.
The only report on insecticide fogging of P. orientalis in a thicket of A. seyal, indicated a transient reduction in the vector density, with rapid re-invasion from neighboring sites (Turner et al. 1965). The impact of pyrethroid-impregnated bed-nets in protecting against the bites of P. orientalis was also tested once (Elnaiem et al. 1999b). In this trial, the authors showed that no P. orientalis landed on volunteers using wide-mesh mosquito bednets impregnated with 10 mg a.i/m2, as compared to an average of 6.9 and 32.0 landings on control volunteers using unimpregnated bednet or no bednets, respectively. Socio-behavioral observations on the bed time of people living in the study area indicated that the use of impregnated bednets against P. orientalis would give more potential protection for children than for adults. Therefore, it was concluded that bednets would provide an effective means for control of VL, especially considering the fact that the disease occurs mainly in this age group.
Evaluation of repellents against P. orientalis was conducted in two trials. In an unpublished report for Medicins San Frontier – Holland (MSF-Holland), Schorcher et al. (1994) presented data that the use of oil extracts of neem (Azadirachta indica L.) seed as a repellent provided three h full protection against the bites of P. orientalis. More recently, Kebede et al. (2009) showed that in laboratory and field experiments, concentrations of 2% and 5% of neem (Azadirachta indica L.) and Chinaberry oils (Milia azadarach L.) oils in coconut oils provided strong repellent effects against P. orientalis.