Control of phlebotomine sandflies

Authors


Dr Michele Maroli, Department of Parasitology, Istituto Superiore di Sanità, Viale Regina Elena, 299 Rome, Italy. E-mail: maroli@iss.it

Abstract

Abstract. Phlebotomine sandflies (Diptera: Psychodidae) transmit many zoonotic diseases (arboviruses, bartonelloses and especially leishmaniases) of importance to human health in at least 80 countries. Measures used to control adult sandflies (Lutzomyia and Phlebotomus) include the use of insecticides (mostly pyrethroids) for residual spraying of dwellings and animal shelters, space-spraying, insecticide-treated nets, impregnated dog-collars and personal protection through application of repellents/insecticides to skin or fabrics. Because the breeding-sites of sandflies are generally unknown, control measures that act specifically against immatures are not feasible, although the effectiveness of a few biological and chemical agents has been demonstrated in laboratory evaluations. Reports of insecticide-resistance refer to only three sandfly species (P. papatasi, P. argentipes and S. shorttii) against DDT in one country (India), although there are reports of DDT-tolerance in several countries. Current knowledge of sandfly susceptibility to various insecticides is summarized. Constraints and advantages of different compounds, formulations and delivery methods for sandfly control under different environmental conditions are discussed.

Introduction

Phlebotomine sandflies are the vectors of leishmaniases, affecting people in more than 80 countries (Desjeux, 1996). In some countries, sandflies also carry and transmit other zoonoses such as bartonellosis (Birtles, 2001), phleboviruses (Tesh, 1988) and certain flaviviruses, orbiviruses and vesiculoviruses (Comer & Tesh, 1991; Ashford, 2001), causing health problems for humans and domestic animals. Despite their small size and delicateness, female sandflies are haematophagous pests so their control may be required even where they are not active as vectors. Sandfly breeding-sites are generally difficult to find in nature, so control measures that act specifically against immatures are not feasible, although the effectiveness of a few biological and chemical agents has been demonstrated in laboratory evaluations.

The sandfly vectors of various Leishmania spp. comprise more than 40 species of Phlebotomus in the Old World and 30 species of Lutzomyia in the Americas (WHO, 1990). The epidemiology of cutaneous leishmaniases (CL) in Europe, Asia and Africa differs fundamentally from that in the New World by occurring principally in open, arid habits, including urban areas. By contrast, the American leishmaniases are more associated with rural areas, often under very humid conditions. This has been reflected in the different approaches to control of sandfly vectors adopted by health authorities in the different hemispheres. Although sandfly control is accepted as practical in most Old World situations, in the Americas the perception of CLs as occupational hazards associated primarily with forests where vectors could not be targeted efficiently has resulted in emphasis being placed on treatment of these diseases rather than their prevention. American zoonotic visceral leishmaniasis (ZVL), by contrast, occurs in open, semi-arid areas within the wide geographical range of the sandfly Lu. longipalpis s.s. Similar environmental conditions are encountered in the Mediterranean region where Old World ZVL occurs and it is now widely accepted that the same parasite (Le. infantum) is involved, probably transported to South America in infected dogs of Spanish or Portuguese colonists (Mauricio et al., 2000).

In recent years, studies of transmission cycles in rural areas of the New World tropics have indicated the feasibility of sandfly control in intervention programmes against leishmaniasis. In addition, cases of both CL and ZVL have become increasingly prevalent in urban areas, including large Brazilian cities such as Belo Horizonte, Fortaleza and Recife. A similar trend has been recorded in Mediterranean foci during the last decade (Gradoni et al., 1993; Cascio et al., 1997). Based on mathematical models, insecticidal control of sandflies appears to represent a more effective way of reducing Le. infantum transmission than the present strategy of culling infected dogs (Dye, 1996) as well as being more acceptable to the human population. Since man is a dead-end host of most Leishmania species, treatment of existing human cases generally does not affect transmission. Interruption of the cycle by vector control may offer a cheaper, more practical solution to treatment and improved knowledge of the alternatives available could lead to preventative measures being undertaken in more leishmaniasis foci. Here we review current knowledge on sandfly control, with a view to assisting health authorities in selecting the most efficient and cost-effective option for their areas.

Historical overview

The first attempt to control sandflies using modern insecticides were carried out with DDT in the Rimac Valley, a Peruvian focus of bartonellosis, in January 1944 (Hertig & Fairchild, 1948). Residual spraying of two adobe huts with a 5% solution of this insecticide conferred protection against the bites of sandflies (mostly Lutzomyia verrucarum) for about a week. Later that year Hertig evaluated DDT against Phlebotomus papatasi in Naples, Italy, during an outbreak of sandfly fever before returning to conduct further experiments in Peru in 1945. Field trials were subsequently carried out in Palestine (Hertig & Fisher, 1945; Jacusiel, 1947), Italy and Greece (Hertig, 1949) when DDT became available to combat outbreaks of sandfly fever and leishmaniasis. These early trials allowed Hertig and Fairchild (1948) to establish four basic principles of sandfly control: (a) residual spraying of houses and animal shelters prevented blood-feeding by the insects, (b) such treatments denied them resting and breeding sites, (c) the effects of spraying were sharply localized, so that control measures need only be undertaken within a limited area of a few hundred metres, and (d) the relatively long life cycle of sandflies delayed recovery of populations within a treated area.

Ghosh (1950) carried out the first field evaluations of DDT and BHC against sandflies in India. Large amounts (approximately 14 g/m2) of 2.5% or 5% DDT sprayed onto mud walls of houses and cowsheds reduced populations of sandflies (P. argentipes, P. papatasi and P. minutus) for 7–8 months. BHC was much less effective and sandflies reappeared within 1 month. Both BHC and DDT were employed extensively in the Soviet Union for sandfly control (Perfil’ev 1966). Insecticidal control of visceral leishmaniasis (VL) in the People's Republic of China dates from 1958 and has been based on residual spraying of houses and caves containing livestock with DDT, the carbamate 3,5-MC, BHC and more recently deltamethrin, as well as aerial spraying with BHC (Guan, 1991).

In Brazil, attempts to control sandflies with chemical insecticides date from 1954, when Nery-Guimaraes & Bustamante (1954) evaluated spraying of houses with DDT in a focus of CL in Rio de Janeiro State. After 5 years of periodic spraying they noted a decrease both in sandfly population levels and in incidence of disease. DDT was subsequently used in spraying campaigns against Lu. longipalpis in the Brazilian states of Ceara and Minas Gerais until 1964 (Oliveira Filho, 1994). Insecticidal control of sandflies was resumed in 1980 in response to the increasing number of cases in several states (Monteiro et al., 1994). Organochlorides (principally DDT and BHC) were used until 1992, with pyrethroid applications dating from 1989, the compounds most used being cypermethrin at 125 mg a.i./m2 and deltamethrin at 25 mg/m2. Brazil accounts for 90% of leishmaniasis cases in the New World and there seems to be no other national intervention programme based on insecticidal spraying in any of the other countries of the Americas.

Lane (1991) pointed out that the single most important constraint to assessing the value of vector control in reducing leishmaniasis incidence was the lack of well-documented examples of intervention. He cited the supposed abatement of a CL outbreak in Sudan by vector and reservoir control as an example of this, noting that the epidemic may already have been in decline before these measures were initiated, perhaps due to decreasing numbers of susceptible individuals. Although circumstantial, the most dramatic evidence for the effectiveness of insecticidal spraying against sandflies as a leishmaniasis control measure is probably provided by the Indian National Malaria Eradication Programme of 1958–1970. No cases of Old World VL, or ‘kala-azar’, caused by Le. donovani were reported from the Indian state of Bihar during that period, presumably because populations of P. argentipes had been effectively suppressed by DDT house-spraying against the malaria vector Anopheles culicifacies. However, two cases of post kala-azar dermal leishmaniasis (PKDL) appeared within months of the programme being halted and 301 076 cases of VL were reported from 1977 to 1990, with a fatality rate of 2% (Thakur & Kumar, 1992). More recently, Pandya (1983) found that malathion spraying against malaria mosquitoes suppressed populations of P. argentipes in Gujarat, India, for 8–9 months.

Corradetti (1954) reviewed the earliest attempts to control Phlebotomus using DDT in Italy and noted that the malaria control campaign based on this insecticide reduced the transmission of Leishmania. Reduction of the number of leishmaniasis cases was also postulated as a collateral benefit of the malaria intervention programme in Iran (Seyedi Rashti & Nadim, 1975) although Nadim & Amini (1970) had earlier concluded that it did not interrupt Leishmania transmission in the city of Isfahan, based on catches of sandflies using sticky traps. Field-collected specimens of P. papatasi from Isfahan showed greater tolerance to DDT than populations from other areas routinely treated with DDT from 1950 to 1968, even 20 years after cessation of the antimalaria programme (Seyedi Rashti et al., 1992).

Reduced incidence of leishmaniasis has since also been reported as a collateral benefit of malaria control programmes in Bangladesh (Elias et al., 1989), Syria (Tayeh et al., 1997) and Peru (Davies et al., 1994), with a modest reduction also observed in response to BHC spraying in Azerbaidjan (Nadzharov, 1955). In addition, Tesh & Papaevangelou (1977) reported that antimalaria activities in Greece produced a reduction in the number of sandfly fever cases. No reduction in morbidity from ZVL was recorded as a result of antimalaria measures in Portugal or Greece, however, (Lane, 1991) and incidental suppression of leishmaniasis could only be expected in areas where the transmission cycle is largely anthroponotic (Saf'janova, 1971).

There are several other examples of sandflies being affected by control measures directed against other pest species. Wijers & Kiilu (1984) credited pesticide use on cotton by Kenyan farmers and its storage in human dwellings with the suppression of man-biting by both P. martini and the malaria vector An. gambiae. Urban populations of P. papatasi in Saudi Arabia were reduced by ground and aerial application of diazinon against synanthropic flies such as Musca (Büttiker, 1980). Most recently, it has been claimed that insecticidal fogging to control Aedes aegypti during the recent dengue epidemic was instrumental in lowering the incidence of ZVL in Belo Horizonte, Brazil (M. Michalik, personal communication). Interpretation of the results of control measures against ZVL in Brazil is complicated by the fact that a three-pronged strategy is used, involving treatment of human cases, elimination of infected dogs and insecticidal spraying of houses and animal shelters within a 200-m radius of the homes of infected individuals. Thus even when this strategy produces a reduction in the incidence of human cases, the contribution made by vector control cannot be determined (Magalhães et al., 1980). Alencar (1961), however, noted a 58.2% reduction in the number of human ZVL cases in municipalities where the walls of houses were sprayed with DDT at 1.5 g/m2 whereas the prevalence of the disease increased by 11.9% in untreated areas during the same period.

Review of available sandfly control methods

Residual spraying of houses and animal shelters

Residual treatment of the walls of human dwellings and or animal shelters depends on the availability of a suitable public health infrastructure, including adequate supplies of insecticide, spraying equipment and trained personnel. Ideally, such personnel should be trained in insecticide application, monitoring techniques and interpretation of sampling data, as well as safety techniques.

Joshi & Rai (1994) and Kaul et al. (1994c) found that two rounds of indoor residual spraying with DDT were effective for the control of P. argentipes in Uttar Pradesh, India, whereas Mukhopadhyay et al. (1996) noted that in West Bengal this species reappeared 9 months after a single application at 1 g/m2. Pandya (1983) found that malathion at 2.1 g/m2 controlled this same species for 8–9 months in cattle sheds in Gujarat, India. These shelters attracted large numbers of sandflies and refusal of farmers to allow spraying during the monsoon season, when the animals were kept indoors, was an important constraint on control measures (Gupta, 1975).

Morsy et al. (1993) evaluated the residual effect of four insecticides 75 days after these had been sprayed on cement walls in Egypt against P. papatasi. Even after 30 min exposure, the highest mortality produced was only 76.7%, for the carbamate propoxur. The values obtained for permethrin, malathion and BHC were all around 50%. Benzerroug et al. (1992) compared human infections with Le. infantum in Algeria before and after a house-spraying campaign based on DDT and found that annual incidence dropped from 426 cases per 100 000 inhabitants to only 17.9 one year later.

Kelly et al. (1997) studied the impact of the pyrethroid lambdacyhalothrin on the abundance and distribution of peridomestic Lu. longipalpis in the Brazilian Amazon and compared two treatments: ‘blanket coverage’, in which all animal pens within a village were treated and ‘focal coverage’ in which spraying or installation of insecticide-impregnated cotton sheets was restricted to a subset of shelters. They detected a 90% reduction in Lu. longipalpis abundance in treated sheds included in the focal intervention, with no discernible effect on numbers in untreated dining huts or houses. This differential impact may have been due to disruption of male pheromone production. By contrast, blanket coverage produced an increase in sandflies collected in untreated dining huts, possibly because of deflection away from aggregation sites.

Le Pont et al. (1989) sprayed the inner and outer walls of houses and animal shelters in Yungas, Bolivia, with deltamethrin at 25 mg/m2 and found that Lu. longipalpis disappeared for 9–10 months. No such effect was seen for Lu. nuneztovari anglesi, a vector of CL described by the same authors, presumably because of its more exophilic behaviour. Falcão et al. (1991) investigated the effect of intradomiciliary and peridomiciliary spraying of deltamethrin in a Brazilian focus of American cutaneous leishmaniasis, principally against the Leishmania vectors Lu. intermedia and Lu. migonei. Internal and external walls of houses were sprayed as well as outbuildings and trees to a radius of 10 m. They obtained a significant reduction between the number of sandflies inside houses of the treated area before and after spraying as well as when compared with the untreated area. No such reduction was observed in peridomiciliary collections. Davies et al. (2000) evaluated the effect of spraying interior walls and ceilings of houses in the Peruvian Andes with lambdacyhalothrin on transmission of Le. peruviana. They found that the LT95 value of this compound for the sandfly Lu. verrucarum was about 20 mg/m2, with no reduction in effectiveness observed over 6 months when a concentration of 25 mg/m2 was applied. Indoor spraying reduced the abundance of Lu. verrucarum by 78% and Lu. peruensis by 83%. More importantly, the proportion of susceptible householders acquiring leishmaniasis was significantly reduced by 54% as a result of spraying. Marcondes & Nascimento (1993) evaluated three different concentrations of an emulsifiable concentrate of deltamethrin applied to walls for the control of Lu. longipalpis in Brazil. Although the number of sandflies aspirated from the walls of treated houses was significantly lower than that from untreated dwellings, this effect disappeared within only 2 months and some insects were collected as early as 14 days after spraying. Alexander et al. (1995b) found that residual spraying of walls with deltamethrin in a Colombian village surrounded by forest had no perceptible effect on the number of sandflies entering houses, although the insecticidal activity of the treated surfaces was undiminished during the study period.

M. D. Feliciangeli et al. (unpublished data) carried out an indoor trial to reduce the population densities of Lu. ovallesi, the proven vector of CL in Miranda State, Venezuela. Houses were matched according to their structure (adobe, concrete and wood) and randomly assigned to a control group or treated group sprayed with 25 mg/m2 of lambdacyhalothrin, this dose having been selected based on susceptibility tests in the laboratory. Sandfly abundance was measured using CDC light traps from 7 to 79 days post-intervention, by which time the population in control houses declined to very low levels, coinciding with the end of the Leishmania transmission season. Catches of fed females, total females and males were significantly lower in sprayed than control houses immediately after treatment but numbers of the latter two groups recovered quickly, reaching the levels in control houses after 7 and 11 weeks, respectively. The short residual effect of the insecticide was confirmed by bioassays following WHO protocols on a laboratory colony. However, catches of fed females in sprayed houses did not recover during the 3 months of the trial. Previous studies of adult population dynamics showed Lu. ovallesi to be abundant in the study area for a short period each year, leading the authors to conclude that two indoor sprayings with lambdacyhalothrin, the first at the beginning of November and the second early in January, would reduce the population considerably and lower CL transmission.

However, the effectiveness of residual spraying may depend on the degree to which sandflies have adapted to man-made environments, as well as the total area treated. Thus sandfly/leishmaniasis control by this method will be much more effective in urban situations, where every house and animal shelter is treated than in rural areas, where relatively few, widely dispersed dwellings are sprayed and the insects that bite man and domestic animals represent a small proportion of the total vector population.

In rural areas where large areas must be covered and suitable spraying equipment is available (including aircraft modified for crop-spraying) application of insecticides as aerosols may represent a viable alternative to residual treatment of houses or animal shelters. Turner et al. (1965) evaluated the effectiveness of fogging with DDT and malathion against P. orientalis in the Sudan, as well as residual applications of these insecticides and BHC. Based on the numbers of female sandflies taken in human bait collections, the percentage reduction of vector populations was disappointingly low and short-lived for all three insecticides when sprayed on the walls of houses. As a fog, only 5% DDT gave any indication of control.

Sandfly control in silvatic environments

In communities where Leishmania transmission occurs in and around houses but the sandflies breed and rest in surrounding forest, barrier spraying may represent an alternative to intra- or peridomiciliary spraying. This involves treatment of tree trunks and vegetation within a predetermined radius of human dwellings. Problems here include the difficulties in achieving adequate coverage with insecticides, reduced persistence of the chemicals and dangers to non-target organisms. Chaniotis et al. (1982) obtained a 30% reduction in man-biting catches of sandflies in Panamian forest using bimonthly fogging or ulv applications of malathion, the latter applied to the tree trunks used as diurnal resting sites by the insects. Resting site collections of sandflies in fog and ulv-treated areas fell by 20.6 and 12.5%, respectively. Ready et al. (1985) sprayed DDT on the lower trunks of trees used as diurnal resting sites by Lu. umbratilis in Amazonian Brazil and found that the insecticidal effect persisted for up to 11 months. Floch (1957) advocated control of CL by clearing forest and applying residual insecticides (DDT, HCH and dieldrin) to sandfly breeding and resting sites around villages to a radius of 250 m. Initial trials in French Guiana showed that spraying was effective for up to 6 weeks, even in the rainy season. Esterre et al. (1986) monitored human cases, sandfly density and putative animal reservoirs of Leishmania over an 18-month period after creating a barrier zone of radius 400 m around a village in French Guiana by clearing forest and fogging with naled. Although they successfully reduced the incidence of leishmaniasis cases as well as vector and reservoir density in the village, the detrimental environmental effects and labour-intensiveness of such a measure would make it both impractical and undesirable in most situations. Alexander et al. (1995c) suggested that rather than cut down trees used as resting sites or spray them with residual insecticides, application of whitewash to the lower trunks (already widely used in the Neotropics to protect foliage against leaf-cutter ants) might make such trees unfavourable to sandflies. This might force the insects to seek other trees further away from villages, reducing the degree of man-sandfly contact and thus the risk of peridomiciliary transmission of Leishmania.

The width of barrier zones has been determined from mark–release–recapture studies of sandfly dispersal (e.g. Chaniotis et al., 1974; Alexander, 1987). However, the relatively small number of insects recaptured in such studies means that they can only provide limited information on the potential distances that the insects may travel, the maximum values determined by the logistically defined limits of the area that can be sampled for marked sandflies. A more effective method of evaluating dispersal capability, at least for communities surrounded by forest, would be to mark and release insects at various distances from a central point in the community studied, and calculate the proportion of insects released at each of these concentric perimeters in the forest that are recaptured entering houses to bite.

Perich et al. (1995; 2003) used barrier spraying of pyrethroids to reduce populations of man-biting sandflies in Guatemala and Brazil. A 100-m-wide barrier created by spraying vegetation with cyfluthrin significantly reduced the numbers of sandflies collected in CDC light traps for more than 80 days. As with house-spraying, application of this measure requires the involvement of adequately trained and equipped personnel; perhaps this can only be achieved by military personnel on a short-term rather than sustained basis.

Use of insecticide-treated mosquito nets (ITNs)

The use of ITNs is an effective, relatively cheap and sustainable method of malaria control (Curtis et al., 1990; Lengeler et al., 1996; Chavasse et al., 1999). The synthetic pyrethroids used for the treatment of the nets combine the properties of low to moderate mammalian toxicity (Wells et al., 1986; Zaim et al., 2000), low volatility and high insecticidal activity. ITNs have also been evaluated against phlebotomine sandflies and leishmaniasis in several countries, including Italy (Maroli & Lane, 1989), Burkina Faso (Majori et al., 1989), Syria (Tayeh et al., 1997; Desjeux, 2000), Sudan (Elnaiem et al., 1999a, 1999b), Kenya (Mutinga et al., 1992, 1993; Basimike & Mutinga, 1995), Colombia (Alexander et al., 1995d) and Venezuela (Feliciangeli et al., 1995; Kroeger et al., 2002). The use of ITNs may represent the most sustainable method of reducing intradomiciliary transmission of Leishmania in communities surrounded by forest, where the diurnal resting sites of vectors are unknown or inaccessible.

ITNs do not emit repellent vapour (Maroli & Majori, 1991) and act as ‘baited traps’ in which sandflies attracted by exhaled CO2 and host odour die after alighting on treated surfaces. Elnaiem et al. (1999b) found that only 30 s of exposure to bednets impregnated with 10 mg a.i./m2 of lambdacyhalothrin were sufficient to kill 100% of P. orientalis within 1 h. Maroli & Majori (1991) recorded over 90% mortality within 24 h in P. papatasi and P. perniciosus exposed in the laboratory to mesh treated with 1 g a.i./m2 permethrin, as well as reduced biting rates in both species.

It is often argued that the small size of sandflies means that the mesh used on bed nets would have to be extremely fine to confer protection against their bites, and that this would increase cost and limit their acceptability to potential users. However, Maroli & Lane (1989) found that permethrin-impregnated nets of 1 cm2 mesh placed over windows significantly reduced the numbers of P. perfiliewi entering houses in Italy. Majori et al. (1989) obtained a similar result using permethrin-impregnated curtains of slightly smaller mesh (5 mm2) in Burkina Faso. Both mesh sizes would be large enough to permit the passage of sandflies without touching the treated material and their protective effect must be due to sandflies alighting on the material or a volatile repellent or irritating effect operating at very close range. Feliciangeli et al. (1995), however, found that sandflies (principally Lu. ovallesi) were able to pass through 6 mm mesh treated with deltamethrin at 15 mg/m2 and bite volunteers. The insects could also rest on the impregnated surfaces for periods of up to two minutes with no significant mortality observed. When mesh size was reduced to 4 mm and concentration raised to 60 mg/m2 all sandflies were killed within 30 min after 10 min of exposure.

In Colombia, Alexander et al. (1995d) evaluated the use of bednets and curtains impregnated with deltamethrin against Lu. youngi. Although the mesh size used was large enough to permit the passage of sandflies, the biting rate on human volunteers under impregnated bednets was significantly lower than that under untreated bets. In addition, all sandflies exposed to treated mesh died within 24 h, so that the protective effect of the nets would be supplemented to some extent by reduction of sandfly population levels. However the latter is unlikely to be significant in areas where sandflies entering houses to bite represent a relatively small proportion of the total population. This probably explains the observations of Feliciangeli et al. (1995) who noted that ITNs impregnated with deltamethrin at 60 mg/m2 produced no significant reduction in the numbers of sandflies entering houses in Venezuela. The authors attributed the failure of the intervention in reducing indoor sandfly population to the low compliance of the control measure. Discomfort due to high nocturnal temperatures frequently prompted local residents to remove nets on windows and doors.

Tayeh et al. (1997) were also unable to detect a significant reduction in density of P. sergenti in Syrian villages whose inhabitants were provided with deltamethrin-impregnated bednets. However, there was a sharp and consistent reduction in the incidence of cutaneous leishmaniasis due to Le. tropica in these villages in the first year after introduction of the bednets. This effect disappeared in the remaining two years of the study.

An advantage of ITN use as a leishmaniasis control measure is that members of affected communities can treat the nets themselves, whether these are manufactured locally or supplied by health authorities. Although the pyrethroids used to treat bednets are relatively safe, their incorporation in such leishmaniasis control programmes should still involve a degree of supervision or training of users by local health authorities to ensure that impregnation is done regularly and the chemicals are not substituted by users for other, more dangerous pesticides.

Another problem as that much sandfly activity occurs around sunset, generally before people have retired for the night, so that exposure to bites may be reduced but not eliminated. Impregnated curtains hung across doors and windows may reduce sandfly access before the occupants retire for the night but since the insects are preadapted to resting in very confined spaces during the day even small gaps in the walls or roof of a dwelling would allow them to enter. Elnaiem et al. (1999a) carried out laboratory and field evaluations of curtains impregnated with permethrin on P. papatasi in the Sudan and found that exposure to mesh treated with 0.5 mg a.i./m2 for 3 min killed all the insects within 24 h. Both the human biting rate and the resting density of this species (although not its nocturnal activity) were also significantly reduced. Basimike & Mutinga (1995) evaluated screens made of polyester netting and impregnated with 0.5% permethrin, hung beside beds and each occupying an area equivalent to half the surface area of a one-roomed house. The screens were treated at 6-month intervals and percentage reduction of P. martini numbers inside houses increased to a maximum recorded value of 88.8% after eight treatments.

Mutinga et al. (1992, 1993) hung cotton wall cloths impregnated with 0.5 g/m2 permethrin inside houses in Kenya and found that these retained their insecticidal effect against P. martini and P. duboscqi for 6 months. The numbers of these species collected inside houses were reduced by 76 and 85%, respectively. Unlike bednets and curtains, the purpose of these cloths was solely to kill sandflies that would normally rest on the inside walls rather than to restrict access to houses or sleeping individuals.

Insecticidal treatment of sandfly resting and breeding sites

In open areas where the resting or breeding sites of sandflies are known, insecticidal treatment or destruction of these microhabitats present more focused alternatives to leishmaniasis control. Karapet'ian et al. (1983) used aerosol insecticidal smoke pots to reduce the population densities of P. papatasi in the burrows of great gerbils (Rhombomys opimus) in the Soviet Union. Spraying gerbil burrows used as resting and breeding sites of P. papatasi did not give long-term control of sandflies in Turkmenia however (Saf'janova, 1963). Destruction of these microhabitats represents one of the few examples of effective non-insecticidal control of sandflies and is reviewed elsewhere (Eliseev, 1980; Vioukov, 1987).

Robert & Perich (1995) evaluated residual spraying of termite mounds and animal burrows with the pyrethroid cyfluthrin as a control measure against P. martini in Kenya. Numbers of adult sandflies in light trap collections were reduced by up to 90% for 2 weeks following treatment and this measure suppressed sandfly populations at resting sites for up to 12 weeks. Similar attempts have also been made to control P. martini in both Kenya and Ethiopia using DDT and lambdacyhalothrin (Lane, 1991).

Chemical repellents

In areas where Leishmania transmission is extra-domiciliary and leishmaniases are an occupational hazard, use of insect repellents or protective clothing may be the only preventative measures available. The latter may be impractical in the Tropics or prohibitively expensive while repellents are also relatively costly and may be potentially harmful after prolonged use. As such they should only be considered for use by people who are risk of Leishmania infection only temporarily, such as tourists, soldiers on manoeuvres or hunters. Sabin (1951) found that dimethyl phthalate (DMP) and a pyrethrum cream prevented probing and feeding by P. papatasi in foci of sandfly fever. Schmidt & Schmidt (1969a) compared the effectiveness of nine repellents against this species and found that chloro-diethyl-benzamide gave the longest mean protection time (326 min). Diethyl-toluamide (DEET) was tested in two series of trials and found to be effective for over 4h.

Fossati & Maroli (1985) evaluated three tropical repellents against P. perniciosus using dose–response techniques (calculation of ED50 and ED90 values). Twenty-one human volunteers were exposed to 1800 laboratory reared nulliparous females of P. perniciosus. Indalone and MGK 11 were more effective than standard DEET and results indicated that P. perniciosus was considerably more sensitive to the repellents than other haematophagous arthropods.

Availability of repellents could be improved by encouraging people to manufacture the compounds themselves from locally available materials such as citronella or eucalyptus. Topical application of 2% neem oil provided 86.1% protection for 7 h against the bites of P. papatasi under laboratory conditions (Sharma & Dhiman, 1993) and was even more effective under field conditions in Rajastan, India, where it conferred 97.6% protection for 10 h (Dhiman & Sharma, 1994). A 5% solution of neem oil was also highly effective in reducing the number of sandfly bites on volunteers when impregnated onto cardboard mats placed on electrical heaters originally designed for use with synthetic pyrethroids.

The commercially available insect repellent DEPA (N,N-diethylphenyl acetamide) was compared with neem oil for protection against the bites of P. papatasi by Srinivasan & Kalyanasundaram (2001). Neem oil was significantly more effective than DEPA when applied to mice in the laboratory at concentrations of 1% and 2% but the repellencies of the two compounds were similar at 5%.

Jia et al. (1990) studied the efficacy of five repellents against P. alexandri in the laboratory and field in China. The most effective compound tested was a mosquito-repellent perfume (MRP) which at 0.25 µl/cm2 conferred protection for almost 8 h. The least effective was dibutyl phthal (DBP) which repelled sandflies for only 1 h.

Alexander et al. (1995a) in Colombia evaluated a relatively inexpensive soap formulation containing DEET and permethrin and found that it retained 67% of its activity (in terms of numbers of sandflies biting volunteers) up to 8 h after application. However the main drawback of this formulation was that repellency was lost if the soap was rinsed off the skin. Perfil’ev (1966) summarized experiences in the Soviet Union with older formulations of repellent soap and long-lasting repellent-treated netting that remained active for the whole season.

Insecticides or repellents applied to clothing rather than skin offer an alternative approach to personal protection against sandfly bites. However clothing impregnated with permethrin did not completely protect volunteers against sandfly bites in Panama (Schreck et al., 1982), probably because of the low vapour pressure of the insecticide and the fact that insects landing on the treated surface would be deflected to the exposed skin of the face and hands. Dees et al. (1987) found that P. papatasi exhibited probing behaviour during direct contact with permethrin-treated uniforms and readily attacked skin previously covered by treated fabric. Fryauff et al. (1996) investigated the effects of laundering and exposure time on the insecticidal activity to P. papatasi of military uniform fabric impregnated with 0.125 mg/cm2 permethrin. The insecticide remaining after three washes was toxic to sandflies exposed for as little as one minute, killing 15% of insects within 24 h. Nevertheless significant reductions in the knockdown effectiveness of treated fabric were associated with repeated laundering, 24 h mortality falling from 91% (0 washes) to 63% in sandflies exposed to treated surfaces for 10 min. Based on the disappointing results of these carefully controlled studies, use of impregnated clothing to protect non-military personnel from sandfly bites may be construed as impractical.

Impregnated dog-collars

Although certain wild animal species may be involved in the epidemiology of ZVL, domestic dogs appear to be the principal reservoir host of Le. infantum throughout the world. Much of the focus on control of ZVL is currently placed on these animals, particularly the search for a canine vaccine. However a more prosaic but potentially viable alternative method of protecting dogs would be to fit them with collars impregnated with insecticides such as deltamethrin.

Killick-Kendrick et al. (1997) first demonstrated that such collars [now marketed as Scalibor® Protector Bands (Intervet International, Boxmeer, the Netherlands)] exerted a potent antifeeding effect on P. perniciosus and killed up to 60% of the insects within 2 h of exposure. Lucientes (1999) confirmed their efficacy by showing that collared dogs were protected against 85–98% of sandfly bites over a 6-month period when compared to unprotected dogs. Collars also killed about 50% of sandflies exposed to protected dogs during the same period. Halbig et al. (2000) found that dogs wearing collars were bitten by about 80% fewer P. papatasi than unprotected animals. Mortality of sandflies exposed to dogs with collars for 20 h was however, low (18%) and not significantly different from that of insects on unprotected animals.

Based on the results of laboratory evaluations, it has been suggested that, at least in the Mediterranean and Middle East subregions, this measure could be expected to protect dogs from most sand fly bites and retain a protective and killing effect for a complete biting season (Killick-Kendrick et al., 1997). Given the long-term effect of collars (up to 34 weeks), it has been suggested that supplying them to the majority of dogs in a Le. infantum focus would reduce contact between vectors and reservoirs sufficiently to diminish the risk of infection for both dogs and humans.

Maroli et al. (2001) carried out a village-based intervention trial in the Campania region of Italy during two consecutive transmission periods and found that collars conferred up to 86% protection against Le. infantum infection in pet dogs. This was the first confirmation that delthamethrin-impregnated dog-collars could indeed protect dogs from ZVL. In a further field trial (Maroli et al., 2002b), 49/249 seronegative stray dogs were fitted with collars and confined in a kennel. The first serological evaluation performed 1–2 months after the end of the sandfly season showed seroconversion in 6.1% of these animals, vs. 24.7% of controls, a significant protective effect of 75.3%. A second round performed 7–8 months after sandfly exposure revealed a cumulative seroconversion rate of 18.2% in collared dogs, compared to 38.2% of controls, so that protection fell to 52.3%.

Preliminary results from Iran indicate that use of dog-collars there significantly reduced ZVL incidence in both dogs and children (Mazloumi-Gavgani et al., 2002) and a Brazilian study found that they significantly reduced the odds of dogs increasing their anti-Leishmania titre by 50% (R. Reithinger et al., unpublished data). The collars have a potent antifeeding and insecticidal effect against both Lu. longipalpis and Lu. migonei, New World vectors of ZVL and CL parasites, respectively (David et al., 2001). However, the effectiveness of this measure is compromised by non-replacement of collars that are lost or whose insecticidal activity runs out. Although one collar/animal/year may suffice to protect pet dogs in southern Europe, where Leishmania transmission is restricted to the summer months, in countries such as Brazil transmission occurs throughout the year and ZVL epidemiology involves large numbers of stray dogs, so that this measure would probably not be sustainable without effective monitoring by public health authorities.

A method that is fundamentally similar to control by impregnated dog-collars is the use of insecticidal baths or shampoos to treat potential Leishmania reservoirs. Jin et al. (1994) exposed P. chinensis females to hamsters infected with Le. donovani and subjected to a deltamethrin bath. All sandflies exposed to the treated hamsters died within 24 h whereas 69.1% of control group insects survived and developed infections. Xiong et al. (1995) then carried out field trials with this treatment on dogs in Sichuan province, China, and were able to interrupt leishmaniasis transmission significantly for 2 years, with no cases recorded whatsoever during the second year.

Reithinger et al. (2001) compared the susceptibility of Brazilian sandflies (Lu. intermedia and Lu. whitmani) to four topical insecticide treatments (two collars and two lotions) applied to dogs. Blood feeding and survival rate of both fed and unfed sandflies were significantly reduced by collars impregnated with deltamethrin as well as by both lotions (containing permethrin and fenthion) but diazinon-impregnated collars had no effect.

Other alternatives for sandfly control

Although sandfly larvae are susceptible in the laboratory to the bacterium Bacillus thuringiensis var. israelensis (Bti) (De Barjac et al., 1981), the difficulty of finding immatures under natural conditions precludes targeting breeding sites as a viable control measure. However, Yuval & Warburg (1989) suggested that microbial agents such as Bti could be used against adult insects by incorporating them in sugar baits sprayed onto substrates in open, dry habitats. Concentrations of 4.4 × 10−2, 10−3 and 10−4 mg/ml of Bti killed 100% of adult P. papatasi, P. argentipes and Lu. longipalpis in the laboratory.

Majori & Maroli (1983) studied the larvicidal effect of Bti serotype H-14 against P. perniciosus. After 6 days of exposure, they observed 100% mortality among larvae fed on the Bti-treated diet.

Robert et al. (1998) evaluated the mosquito larvicidal bacterium B. sphaericus against sandflies in the laboratory. Aqueous suspensions inhibited hatching of eggs of P. dubosqui and Sergentomyia schwetzi by 95% at low concentrations (0.5 and 0.11 mg/cm2, respectively). In a previous study (Robert et al., 1997) they determined that sugar solutions containing this bacterium, sprayed onto vegetation and taken up by sandflies, caused significant mortality of larvae at their breeding sites in animal burrows in Kenya. Adult sandfly populations were also significantly reduced by spraying the bacterial solution at burrow entrances, the effect persisting for up to 10 weeks after treatment.

Kassem et al. (2001) carried out laboratory evaluations of two avermectins on the sandflies P. papatasi and P. langeroni, by presenting the compounds in blood (ivermectin) or sugar meals (abamectin). Low concentrations of either avermectin killed sandflies of both species and sublethal doses of ivermectin in blood resulted in reduced survival and fecundity of adult females. The avermectins are environmentally safe compounds and could therefore be used as systemic insecticides, administered to animals used as blood meal sources by sandflies or to the plants from which they obtain sugars.

In Colombia, the entomopathogenic fungus Beauveria bassiana is employed to control infestations of the coffee berry borer (Hypothenema hampei) in coffee plantations where sandflies transmit Leishmania to man. Although Warburg (1991) selected for a strain of the fungus which killed sandflies (P. papaptasi and Lu. longipalpis) in the laboratory, Reithinger et al. (1997) found that live insects were not infected by commercial preparations of B. bassiana. Simultaneous application of mixtures of strains pathogenic to H. hampei and Lutzomyia spp. might, however, represent a viable alternative in coffee-growing areas where both borer infestions and leishmaniasis occur.

Following the discovery that certain plants (Capparis spinosa, Ricinus communis, Solanum luteum) used as sources of sugar by sandflies were toxic to Le. major (Schlein & Jacobson, 1994; Jacobson & Schlein, 1999), Schlein et al. (2001) found that certain exotic species were also able to kill the insects themselves. Planting these (Bougainvillea glabra, Ricinus communis, Solanum jasminoides) in barrier zones might therefore provide a low-cost, sustainable alternative to insecticide use in the control of sandflies and leishmaniasis. Luitgards-Moura et al. (2002) evaluated the insecticidal effects of two plant extracts used by Amazonian Indians to kill fish. Both were highly toxic to Lu. longipalpis, dried leaf extracts dissolved in water of Antonia ovata (Loganiaceae) and Derris amazonica (Papilionaceae) killing 80% and 100% of females, respectively. These plants could therefore represent a readily available alternative to commercial insecticides for sandfly control in the ZVL focus of Roraima, Brazil.

In recent years the discovery of pheromones produced by male sandflies has led to the suggestion that synthetic copies of these compounds could be used to attract females to insecticide-treated surfaces and potentiate conventional control measures (Lane, 1991). Although attractiveness is not dependent on the presence of a warm-blooded host (Morton & Ward, 1990) the range of the Lu. longipalpis pheromone is restricted to a few metres, considerably less than that exerted by CO2 or host odour (Alexander, 2000) and not comparable to that of semiochemicals currently used for the control of several lepidopteran pests. Sandfly pheromones might be better used as tropical treatments on dogs or livestock to disrupt mating, at least in the case of Lu. longipalpis.

Quesada & Montoya-Lerma (1994) evaluated the insect growth inhibitor chlorfluazuron against second and third-instar larvae of Lu. longipalpis and observed a number of lethal and sublethal effects. Larvae ceased to feed and underwent premature moults, cuticular rupturing or imperfect shedding of exuviae. Female adults that had ingested the insecticide in the larval diet were less likely to take a bloodmeal and the wings, abdomens and genitalia of treated males were significantly smaller than those in control groups.

In some situations, integrated vector control can be employed against the sandfly vector species, involving a combination of different methods. One leishmaniasis control programme in Uzbekistan used spraying with HCH and DDT, elimination of gerbils by poisoned baits, destruction of gerbil burrows/sandfly resting sites, personal protection methods and vaccine prophylaxis (Faysulin, 1980).

Susceptibility of sandflies to insecticides based on laboratory bioassays

Whereas the legs of mosquitoes are used merely for slow walking, those of phlebotomine sandflies are used to run or hop across the skin of the host or other surfaces. As such the degree of contact sandflies may make with treated surfaces is greater and this should increase their susceptibility to contact insecticides. Sandflies exposed to pyrethroids shed their legs before dying (Maroli et al., unpublished data), possibly because these insecticides are more active on a subset of tarsal Na+ channels or because sandflies have unusually high densities of these receptors on the tarsi (French-Constant, personal communication).

Published data on the susceptibility status of sandflies to various insecticides are summarized in Table 1. The first cases of insecticide resistance in sandflies were reported from Bihar, India, where P. papatasi was found to survive exposure to 4% and later 8% DDT (Kaul et al., 1978; Joshi et al., 1979). Dhanda et al. (1983) confirmed that this species was highly resistant to DDT in Bihar, only 13.3% mortality being recorded when insects were exposed to 8% for 24 h. Almost all insects exposed to 3.2% malathion and 4% BHC died within 1 h however. Chandra et al. (1995) found P. argentipes to be susceptible not only to malathion but also to both DDT and dieldrin. Although Mukhopadhyay et al. (1990; 1992) reported resistance of the latter species to DDT in the Samastipur district of Bihar, the continued susceptibility of P. argentipes in this state was later confirmed by laboratory and field observations made by Kumar et al. (1995) and Basak & Tandon (1995).

Table 1.  List of studies (in chronological order) on the susceptibility status of phlebotomine sandflies to insecticides (S, susceptible; T, tolerant; R, resistant).
Species testedStatusInsecticides testedCountry (State)References
  1. †Classes of insecticides tested comprise: carbamates (bendiocarb, propoxur); organochlorines (BHC, DDT, dieldrin, lindane, methoxychlor); organophosphates (chlorophos, chlorpyrifos, fenitrothion, malathion, pirimiphos-methyl); pyrethroids (cypermethrin deltamethrin, lambdacyhalothrin, permethrin, resmethrin). 1Egyptian lab strain showed slight tolerance. 2P. papatasi was R to DDT and S to dieldrin. 3P. papatasi was R to DDT and T to dieldrin; P. argentipes was S to both insecticides. 4R to both insecticides. 5R to DDT, S to malathion and BHC. 6S to DDT and permethrin, T to methoxychlor.7S to DDT and lindane, T to dieldrin and propoxur. 8In 1990 populations from Patna and Bhojpur districts were S, while that from Samastipur was T, in 1992 the T population from Samastipur became R for the first time. 9R to DDT, S to malathion and fenitrothion. 10S to DDT, lambdacyhalothrin, cypermethrin, T to malathion, fenthion, R to propoxur and delthametrin. 11P. papatasi was R to DDT, dieldrin and propoxur and S to malathion, fenitrothion and permethrin; S. punjabensis was S to all insecticides.12S to DDT, propoxur, malathion, deltamethrin, lambdacyhalothrin, permethrin; T to fenitrothion and pirimiphos-methyl. 13Both species were high S to bendiocarb, T to DDT and malathion and R to permethrin. For BHC, P. papatasi and P. argentipes were S and T, respectively. P. papatasi was T and R to deltamethrin and lambdacyhalothrin, respectively, while P. argentipes was reverse. 14Both populations tested (Sirora and West Bengal) were S to deltamethrin and R to DDT; population from West Bengal was also R to dieldrin and malathion.

P. papatasiS  DDT, chlorophosTurkmenistanZhogolev & Kachanova (1968)
P. papatasi1ST DDT, dieldrinEgypt, SudanSchmidt & Schmidt (1969b)
P. papatasiS  DDT, dieldrinEgyptHassan etal. (1970)
P. papatasi2S RDDT, dieldrinIndia (Bihar)Kaul etal. (1978)
P. argentipesS     
P. papatasi3 TRDDT, dieldrinIndia (Bihar)Joshi etal. (1979)
P. argentipesS     
P. papatasi4  RDDT, dieldrinIndiaRahman etal. (1982)
P. papatasi5S RDDT, BHC, malathionIndia (Bihar)Dhanda etal. (1983)
P. argentipesS  DDTIndiaHati (1983)
P. brevisS  DDTAzerbaijanArtemiev etal. (1984)
P. transcaucasicusS     
P. papatasiS  DDTUzbekistanDergacheva & Strelkova (1986)
P. papatasi6ST DDT, methoxychlor, permethrinIsraelPener & Wilamovsky (1987)
Lu. longipalpisS  DeltamethrinBrazilFalcão etal. (1988)
Lu. youngi7ST DDT, lindane, dieldrin, malathion, propoxurVenezuelaScorza & Márquez (1989)
P. argentipes T DDTIndia (Bihar)8Mukhopadhyay etal. (1990)
P. argentipes  RDDTIndia (Bihar)8Mukhopadhyay etal. (1992)
P. papatasi T DDTIranSeyedi Rashti etal. (1992)
P. papatasiS  DDT, BHC, malathion, propoxur, PermethrinEgyptRefaat etal. (1993)
P. papatasiS  DDT, BHC, permethrin, malathion, propoxurEgyptAboul Ela etal. (1993)
P. papatasi  RDDT, dieldrinIndia (Gujarat)Thapar etal. (1993)
Lu. longipalpisS  DDT, chlorpyriphos, malathion,
propoxur, deltamethrin
BrazilOliveira Filho & Melo (1994)
S. shorttii9S RDDT, malathion, fenitrothionIndia (Assam)Kaul etal. (1994a)
P. argentipesS  DDTIndia (West Bengal)Basak & Tandon (1995)
P. argentipesS  DDT, dieldrin, malathionIndia (West Bengal)Chandra etal. (1995)
P. argentipesS  DDTIndia (Bihar)Kumar etal. (1995)
Lu. youngi10STRDDT, fenthion, malathion, propoxur, cypermethrin, deltamethrin, lambdacyhalothrinVenezuelaScorza etal. (1995)
P. papatasi  RDDTIndia (Bihar)Das Gupta etal. (1995)
P. papatasi11S RDDT, dieldrin, malathion,India (Rajastan)Bansal & Singh (1996);
S. punjabensisS  fenitrothion, propoxur, permethrin Singh & Bansal (1996)
P. papatasi  RDDTIndia (W. Bengal)Mukhopadhyay etal. (1996)
Lu. longipalpis12ST DDT, propoxur, malathion, fenitrothion, pirimiphos-methyl, deltamethrin, lambdacyhalothrin, permethrinVenezuelaMazzarri etal. (1997)
P. papatasi13STRDDT, BHC, malathion, deltamethrin, permethrin, lambdacyhalothrin, bendiocarbIndia (Pondicherry)Amalraj etal. (1999)
P. argentipesS     
P. papatasi14  RDDT, dieldrin, malathion, deltamethrinIndia (West Bengal)Dhiman & Mittal (2000)
P. papatasiS  DDT, bendiocarb, cyfluthrin, malathion, permethrin, resmethrinEgyptTetreault etal. (2001)
P. bergerotiS     
P. langeroniS     
P. sergentiS     
P. papatasiS  DDT, lambdacyhalothrin, permethrinItalyMaroli etal. (2002a)
P. perniciosusS     

Mukhopadhyay et al. (1996) provided evidence of the susceptibility of P. argentipes to DDT in west Bengal whereas the reappearance of P. papatasi within one month of spraying suggested that this species had acquired resistance. Amalraj et al. (1999) screened field-collected adults of P. argentipes and P. papatasi in the laboratory for susceptibility to six insecticides (DDT, BHC, malathion, deltamethrin, permethrin, lambdacyhalothrin and bendiocarb). The Pondicherry strains of both species were resistant to permethrin and tolerant to DDT and malathion. Although P. papatasi was susceptible to BHC, P. argentipes showed tolerance to this insecticide by a factor of 1.6. The former species was tolerant to deltamethrin but the latter exhibited resistance. This situation was reversed with respect to lamdacyhalothrin and both species remained susceptible only to bendiocarb.

Bansal & Singh (1996) evaluated the susceptibility status of Sergentomyia punjabensis and P. papatasi to some insecticides in Rajasthan, India, and found that while the former was killed by exposure to all compounds tested, the latter was resistant to DDT, dieldrin and propoxur although still susceptible to malathion, fenitrothion and permethrin. Thus at least one population of P. papatasi is now resistant to two groups of insecticides (organochlorines and carbamates) which should shed some light on the type of mechanism involved.

Populations of P. papatasi from western Rajasthan were resistant to DDT alone and remained susceptible to dieldrin, malathion, fenitrothion and propoxur (Singh & Bansal, 1996). More recently, Dhiman & Mittal (2000) compared the susceptibilities of P. papatasi populations from Uttar Pradesh and West Bengal and found the former to be highly resistant to 4% DDT while the latter was unaffected by 0.4% dieldrin and somewhat resistant to malathion. Both remained susceptible to deltamethrin.

Tolerance of P. papatasi to methoxychlor was reported in Israel by Pener & Wilamovsky (1987) whereas the insects remained susceptible to DDT and permethrin. Although the biochemical mechanism involved in P. papatasi resistance to DDT has not been determined, El-Sayed et al. (1989) showed that insecticide-susceptible sandflies are able to metabolize this insecticide to DDE using the mixed function oxidase and glutathione-S-transferase (GST) mechanisms. The level of GST activity in P. papatasi was lower than that seen in susceptible adults of the mosquito Culex quinquefasciatus when expressed in terms of activity/mg soluble protein. By contrast, cytochrome P-450 was slightly higher in the former species, in both the reduced and oxidized states. These enzymes might confer some degrees of resistance to other insecticide classes.

Schmidt & Schmidt (1969b) found that mean LC50 values of Egyptian and Sudanese populations of P. papatasi exposed to DDT for 1 h ranged from 0.80 to 1.93%, whereas dieldrin was somewhat more toxic (0.63–1.07%). They noted that sandflies had a higher level of tolerance to the latter insecticide than mosquitoes. Aboul Ela et al. (1993) determined the baseline susceptibility levels of Egyptian P. papatasi to five insecticides, using both laboratory-reared and wild-caught insects. The lowest LC50 value was obtained for propoxur (0.00043%), followed in descending order by permethrin, BHC, malathion and finally DDT. The lowest LT50 value obtained was 1.7 min, for laboratory-reared sandflies exposed to 5.0% malathion. Laboratory-bred insects were more susceptible to each of the five compounds tested than wild-caught individuals.

Fahmy et al. (1996) evaluated the susceptibility of Egyptian field populations of P. papatasi to organochlorines (DDT and dieldrin), organophosphates (malathion), carbamates (propoxur) and synythetic pyrethroids (permethrin and deltamethrin) over a 2-year period. Apart from slight (less than three-fold) differences between the LT50 values of 1994 and 1995 for malathion and propoxur, there was no evidence of changes in the susceptibilities of these sandflies to insecticides or indications that resistance was developing. Tetreault et al. (2001) determined the baseline susceptibilities of four North African and Middle Eastern species of sandflies (P. bergeroti, P. langeroni, P. papatasi and P. sergenti) to six insecticides. The responses of all four species to the insecticides were remarkably similar. DDT was the least toxic compound, with an LD50 value of 3.29 mg/ml recorded for P. sergenti. The most potent overall was resmethrin, which gave an LD50 of 0.024 mg/ml for P. bergeroti. The ranges of LD50 values for malathion, permethrin, bendiocarb and cyfluthrin were 0.305–0.548, 0.280–0.387, 0.051–0.125 and 0.036–0.052 mg/ml, respectively.

In Italy, Maroli et al. (2002a) studied the susceptibility of newly established laboratory colonies of P. perniciosus and P. papatasi to some insecticides. A laboratory colony of P. papatasi, unexposed to insecticides for 10 years, was used as a reference strain. The LT50 values for P. perniciosus were 19.9, 3.2 and 6.9 min, and for P. papatasi 18.0, 7.4 and 11.0 min, using DDT, lambdacyhalothrin and permethrin, respectively. The results showed that two Italian populations of P. perniciosus and P. papatasi from Campania region and Rome remain susceptible to the insecticides tested. Lavagnino & Ansaldi (1991) tested the susceptibilities of wild-caught P. perniciosus and P. perfiliewi from Sicily to DDT, malathion and permethrin. Higher concentrations of DDT (4% vs. 1%) and exposure times to permethrin (30 min vs. 10 min) were required to kill the former species. Mortalities of both species were low when exposed to 0.5% malathion for up to 1 h, values of only 55% and 65% being recorded for P. perniciosus and P. perfiliewi, respectively.

The first report of insecticide resistance in the genus Sergentomyia involved S. shorttii in Assam (Kaul et al., 1994a). Although of no direct medical importance and partially exophilic, this species had been exposed to DDT spraying against malarial mosquitoes since 1958. When sandflies were exposed to 4% DDT for 1 h, mortality was only 54%, rising to 75–90% in insects exposed for 24 h. This species remained susceptible to the organophosphate insecticides malathion and fenitrothion however.

To date there have been no records of insecticide resistance in Lutzomyia although pyrethroids have been used for several years against urban populations of Lu. longipalpis in Brazil. Mazzari et al. (1997) reported increased (three-fold) tolerance of this species to pirimiphos-methyl, fenitrothion and permethrin in Venezuela. Falcão et al. (1988) found that only 10 mg a.i./m2 was enough to kill 100% of females exposed to treated filter papers in WHO bioassay kits. Oliveira Filho & Melo (1994) evaluated insecticides of several groups against Lu. longipalpis in the laboratory, using the same kits. They found that the concentration of insecticide (mg a.i./m2) required to produce 100% mortality of the insects was similar for DDT (642) and malathion (630) and considerably higher than for either propoxur (214) or deltamethrin (10.5). Nevertheless the concentrations of propoxur, DDT and malathion to which the insects were exposed were fractions of the doses employed in Brazil for the control of anopheline mosquitoes and triatomine bugs.

Scorza et al. (1995) exposed wild-caught individuals of the Venezuelan species Lu. youngi to seven different insecticides in the laboratory, all except lambdacyhalothrin at concentrations of 125 mg/m2. This species was most susceptible to DDT and cypermethrin, with LD95 values achieved for both insecticides after 50 min. Lambdacyhalothrin (12.5 mg/m2) required 60 min to kill all insects while malathion, fenthion, propoxur and deltamethrin required 80, 90, 120 and 120 min, respectively.

Discussion

Given the many different combinations of vector, parasite, reservoir, symptoms, ecological conditions, epidemiology and cultural practices that contribute to the transmission of Leishmania spp., none of the methods discussed above would be suitable for control of all sandfly populations. Table 2 lists several factors to consider in choosing the most appropriate control options. Desjeux (2001) recognized four major eco-epidemiological entities among the leishmaniases, while Heyneman (in Saf'janova, 1971) described four different types of interaction between man and Leishmania foci. In these interactions, different degrees of environmental modifications due to human activities lead to man becoming (a) an accidental host, (b) the principal or only host, (c) one of several hosts in a stable amphixenosis or (d) exposed to increased risk of transmission due to rapid multiplication of vectors and reservoir hosts. Each of these four types of interaction may occur in various ecological settings and affect human populations that exhibit a diversity of cultural and socio-economic conditions.

Table 2.  Phlebotomine sandfly control options and their respective suitabilities in different situations.
 Questions to be consideredRecommendations
1Has the vector been identified?Conclusive incrimination of the vector may not be possible. However identification of a species known to transmit elsewhere, discovery of natural infections in dissected specimens and marked anthropophilic behaviour may be taken as sufficient circumstancial evidence to identify the target organism.
2Is the transmission cycle partially or totally anthroponotic?If partially anthroponotic, treatment and/or personal protection methods will affect transmission. If wholly zoonotic, only reservoir or vector control methods will be effective.
3If zoonotic, is the reservoir species known?If so, reservoir control (e.g. destruction of rodent burrows) is an option. The nests of certain species may provide sandflies with resting and/or breeding sites that can be targeted for vector control (e.g. insecticidal treatments, including microbial insecticides). Non-nesting species such as dogs can be protected against the bites of sandflies (e.g. using impregnated dog-collars and insecticidal shampoos that prevent feeding and also kill the insects).
4Does transmission occur in or around houses, or in some extradomiciliary situation such as inside forest?Intra- and peridomiciliary transmission can be targeted by residual spraying of houses, domestic animal shelters and other sand fly resting sites with insecticides, maintaining the peridomicile clean (so as not to attract rodents) or situating animal shelters as distant as possible from dwellings. If transmission occurs inside houses impregnated bednets, curtains (ITNs) or wall cloths may be used ITNs can also be employed if it is necessary to sleep in an extra-domiciliary risk area, otherwise repellents or protective clothing should be used.
5Is transmission by sandflies seasonal or year-round?Use of some measures (e.g. impregnated dog-collars) more practical in areas where all transmission occurs within a period shorter than the duration of insecticidal activity.
6Is there an infrastructure present that would allow organized, sustainable measures to be used?In situations where sandfly borne infection is endemic/zoonotic, treatments such as residual spraying of houses and animal shelters will have to be repeated at regular intervals, requiring trained personnel and adequate equipment and materials. Even if these resources are not available, in some situations sandfly control may be devolved to population at risk (e.g. use of ITNs if transmission is intradomiciliary).
7Are human communities at risk willing to participate in the control measures proposed?Involvement of these communities may range from acquiescence (allowing houses and animal shelters to be sprayed by authorities) to active participation (regular impregnation of ITNs, movement or removal of domestic animal shelters).
8What methods are available and are there practical, legal, environmental or cultural constraints on their use?Practical constraints may include resistance to insecticide in target population; there are legal constraints on the use of DDT in most countries; there are environmental constraints on the use of pyrethroids in fishing communities in areas with a high water table; and examples of cultural constraints include the unacceptability of bednets to some communities.

Approximately 70 species of sandflies transmit more than 20 Leishmania spp. and other pathogens in ecological settings that vary from very humid tropical forest to deserts, from temperate cities situated at sea level to high mountain villages. Despite this diversity all sandfly species share a number of basic features. All are nocturnal, resting during the day in dark, humid microhabitats and able to insert themselves into confined spaces to avoid extremes of temperature or humidity. They generally bite various hosts and should be considered as opportunistic man-biters rather than anthropophilic. Their flight, ranges are limited to a few hundred metres. Because of their wide host range, small size and silent, non-hovering flight, people in Leishmania-endemic areas may be unaware of sandfly presence and its role in the epidemiology of the disease, a fact that may compromise leishmaniasis control efforts through community participation. All sandflies are terrestrial breeders and larvae are usually difficult to locate, so that in most areas control measures are effectively limited to targeting the adults.

Although effective in urban areas with high concentrations of sandflies, residual insecticide spraying programmes require suitable equipment and trained personnel for large-scale interventions. In rural areas where dwellings are more dispersed and surrounded by large, untargeted ‘reservoir’ populations of sandflies, residual spraying of houses may be both impractical for logistic reasons and ineffective. Barrier spraying of vegetation is uneconomical and generally unsustainable. Impregnated bednets may offer the best solution in rural areas where transmission is largely intradomiciliary. This measure has the advantage that it can be employed at the individual household level or outdoors, and affords collateral benefits such as privacy and control of other biting insects such as mosquitoes, fleas and bedbugs.

DDT is the cheapest insecticide available but, for reasons of environmental impact and other concerns, its use is no longer permitted in the majority of countries (WHO, 2001). It is also the only compound for which resistance has been recorded in sandflies. Fortunately phlebotomines remain susceptible to all other major insecticidal groups and there is no pressing need to develop new compounds specifically for sandfly control. To date most records of DDT-resistance refer to only three species (P. papatasi, P. argentipes and S. shorttii) in one country (India), although there are reports of sandflies with increased tolerance to this compound in other countries. Sandflies have been shown to possess detoxification mechanisms that could confer cross-resistance between DDT and other compounds. Insecticide-resistance may arise in other sandfly populations through selection resulting directly from leishmaniasis vector control measures, or inadvertently through exposure to insecticides applied for other purposes, e.g. agricultural or domestic pest control, peri-focal spraying against dengue vectors or residual house-spraying against vectors of malaria or Chagas disease.

More focalized measures may require increased community participation and education in preventative measures against leishmaniasis. Inadequate control may merely increase the mean age at which leishmaniasis is acquired, possibly increasing the severity of the disease (WHO, 1990). Improved information on sandfly biting behaviour, resting and breeding sites would make delivery of existing compounds more efficient, resulting in lowered costs of interventions, higher efficacy and fewer detrimental effects to the environment.

Acknowledgements

We thank Dr Philippe Desjeux for inviting us to write the present review and P. Desjeux for his comments on an earlier draft of the manuscript. We are also grateful to C. Khoury for her invaluable assistance in preparing the bibliography. We acknowledge partial financial support from Intervet Italia S.r.l.

Ancillary