Long-lasting insecticidal nets fail at household level to reduce abundance of sandfly vector Phlebotomus argentipes in treated houses in Bihar (India)

Tropical Medicine and International Health doi:10.1111/j.1365-3156.2008.02096.x volume 13 no 7 pp 953 958 july 2008 Long-lasting insecticidal nets fail at household level to reduce abundance of sandfly vector Phlebotomus argentipes in treated houses in Bihar (India) Diwakar S. Dinesh 1, Pradeep Das 1, Albert Picado 2, Clive Davies 2, Niko Speybroeck 3,4, Bart Ostyn 5, Marleen Boelaert 5 and Marc Coosemans 6,7 1 Rajendra Memorial Research Institute of Medical Sciences, Indian Council of Medical Research, Agamkuan Patna, India 2 Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, UK 3 Department of Animal Health, Institute of Tropical Medicine, Antwerp, Belgium 4 Ecole de Santé publique, Université catholique de Louvain, Bruxelles, Belgium 5 Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium 6 Department of Parasitology, Institute of Tropical Medicine, Antwerp, Belgium 7 Department of Biomedical Sciences, Faculty of Pharmaceutical, Veterinary and Biomedical Sciences, University of Antwerp, Antwerp, Belgium Summary objective To determine whether the use of long-lasting insecticidal nets (LLINS) at household level are effective in reducing the abundance of Phlebotomus argentipes, vector of anthroponotic visceral leishmaniasis in India, Nepal and Bangladesh. methods The impact of two long-lasting nets (Olyset and PermaNet) on indoor sandfly abundance was evaluated in selected houses of three endemic hamlets in Bihar (India). It was assumed that most sandflies breed inside the houses and that LLINs would progressively reduce the indoor density during the reproduction season. A campaign of indoor spraying with dichloro-diphenyl-trichloroethane (DDT) interfered with the trial but did not affect the sandfly population. results Only the density of males of P. argentipes was significantly reduced by both the LLINs but not the. conclusions These findings suggest that most female sandflies are coming from outside and that LLINs do not reduce their entry rate. keywords Phlebotomus argentipes, sandfly control, visceral leishmaniasis, long-lasting nets, indoor residual spraying, Indian subcontinent Introduction Based on positive experiences in the past, the efficacy of indoor residual spraying (IRS) with DDT to control Phlebotomus argentipes, vector of visceral leishmaniasis (VL) in India, Nepal and Bangladesh is commonly accepted, although this strategy has been poorly documented (Kaul et al. 1994). However, it requires an important logistical support that is often absent. Insecticide-treated nets (ITN) may represent a more sustainable method of reducing domestic transmission of Leishmania donovani. Insecticide-treated nets are effective in reducing malaria morbidity and mortality (Lengeler 2004) and may also provide community protection through mass impact on vector mosquito population, when used at a high coverage (Coosemans & Carnevale 1995; Curtis et al. 2003). Pyrethroids are the only insecticides currently recommended by the WHO for treatment of nets owing their strong insecticidal activity at low concentration and their safety for the users (Zaim et al. 2000). But low re-impregnation rates are a major constraint for control programmes. Effective long-lasting insecticide nets (LLINs), which do not require re-treatment, even after washing, are an interesting alternative. LLINs are now largely promoted to control malaria (WHO 2007b). Considering safety, efficacy and wash-resistance, WHOPES (WHO Pesticide Evaluation Scheme) recommends five brands of LLINs for the prevention and control of malaria: PermaNet 2.0, Olyset, Interceptor, Duranet and NetProtect (WHO 2001, 2004, 2007a). Insecticide-treated nets have been evaluated against sandflies and Leishmania sp. in several countries (reviewed by Alexander & Maroli 2003) but always at a small-scale ª 2008 Blackwell Publishing Ltd 953

level and not against L. donovani in Asia. A large distribution of ITNs in eastern Sudan reduced the incidence of VL but more evidence is required through randomized trials (Ritmeijer et al. 2007). The LLIN Olyset was effective in reducing the incidence of cutaneous leishmaniasis in Iran (Emami et al. 2005). To measure the impact of LLINs on incidence of VL, a large-scale evaluation is now going on in Nepal and in India (http://www.kalanetproject.org). However, it would also be appropriate to evaluate the protection of LLINs at the household level, as P. argentipes is classically reported to breed inside the houses and in cattle sheds (Kumar et al. 1995). The use of insecticide-treated materials may then affect the vector density inside these houses without requiring coverage. In this study, we assessed the efficacy of two LLINs (OLYSETÒ and PermaNetÒ) to reduce the sandfly density inside selected houses of three VL endemic hamlets in Bihar (India). Materials and methods Study design The study was conducted in the state of Bihar (India) where two major epidemics of Kala-azar occurred, one in 1977 and one in 1991 1992, after the Malaria Eradication Programme was supposed to have eliminated the disease in 1964. Both outbreaks were controlled by spraying campaigns with DDT and treatment of cases. Apart from some local spraying in villages with outbreaks, no systematic DDT spraying was performed after 1995 (Thakur 2007). In 2005, a Regional Kala-azar Elimination Initiative was launched with an attack phase foreseen in 2007 (WHO 2005a). In Bihar, P. argentipes has well-defined seasonal patterns with an increase of the population from March to May and crash down during the monsoon (June to September). A second but lower peak occurs around November (Dinesh et al. 2001). Three hamlets endemic for VL were selected: Gulmehiya Bagh in Patna district, and Rasoolpur and Majlishpur, both located in Vaishali district. No systematic indoor spraying had been performed during the last 10 years. In each hamlet, 16 houses were purposively selected: eight human dwellings without cattle inside the house but with cattle within the compound and eight mixed dwellings where cattle and humans are sharing the same roof. For both the categories and in each hamlet, two houses were randomly allocated to one of the four treatments: two with long-lasting insecticide-treated nets (PermaNet PT, Olyset OT) and two with untreated nets (PermaNet Control PC, local polyester untreated nets LC). OLYSETÒ is a wide mesh (4 mm 4 mm) polyethylene net, blended with permethrin 2% (size 172 129 150.5 cm). PermaNetÒ 2.0 has small meshes (25 holes cm 2 ) with polyester fibres coated with resin containing deltamethrin with a target dosage of 55 mg m 2 (size 200 128 165 cm). Two types of nets were used as controls: untreated PermaNets (PC) and untreated local made polyester nets (25 holes cm 2 ; size: 185 125 165 cm) (LC). Treatment allocation was not stratified on the numbers of sandflies collected during the baseline study, but baseline study differences were taken into account in the models during the analysis stage. In each house, three corresponding nets were installed after the first survey (S 0 ). The house holders were asked to exchange their own nets (all untreated locally purchased nets) for the ones provided by the investigators. No rotation of treatment arms among the houses was foreseen, considering that treated nets may have an impact on the vectors breeding inside the house. During the study period, the three hamlets were sprayed with DDT by the Governmental Control Programme at a dosage of 1 g active ingredient m 2, between surveys 1 and 2. These control activities were not initially planned nor foreseen within the study protocol, but for ethical reasons no interference with spraying activities was exerted by the study team. Surveys Entomological efficacy was estimated on sandfly abundance during one sandfly season (April June 2006) (Dinesh et al. 2001). Surveys were done before the intervention (week 0), 3, 6 and 9 weeks after installing the nets. For each survey, a centers for disease control (CDC) light trap (Miniature Incandescent Light Trap, Model 1012; JW Hock Company, Gainesville, FL, USA) collection was performed during one night (from 6 pm to 6 am) followed by an aspiration collection in the morning (between 6 and 8 am) by one trained person for 30 min. CDC light traps are performing well in collecting males, unfed and gravid of P. argentipes, but not fed for which the aspirator method is more effective (Dinesh et al. 2008). One CDC light trap per house was installed near the corner of a bedroom 6 inches above the ground and 1 inch away from the wall. Collected sandflies were determined under a binocular microscope. Sex and physiological status for (unfed, fed and gravid) was noted. Statistical analysis Crude geometric means of sandfly counts were calculated per survey and per treatment as log (count + 1). Incidence 954 ª 2008 Blackwell Publishing Ltd

rate ratios (IRR) were estimated to assess the effect size of the different treatments compared with the untreated control nets (LC), by fitting a random-effects negative binomial model to the data. Explanatory variables used were effect of survey, which compares the results in the (combined) surveys to the baseline survey, treatment, type of house (Mixed dwellings vs. Human dwellings), effectiveness of overnight CDC light traps and differences between hamlets. Including an interaction of the survey effect and the treatment effect in the model allows investigating whether households showed different insect counts by species, sex and physiological status according to their allocation of treatment (OT, PC and PT when compared with LC) at the baseline study; and investigating whether there is a change in counts after the baseline study through the interaction terms (effect of OT, PT and PC when compared with LC). Negative binomial models account for the overdispersion in the count data. The use of a random-effects model approach can be motivated by arguing that collections within a household share the same environment (physical location), as well as specific characteristics (Speybroeck et al. 2003). These shared factors, whose effects can change from household to household, create dependencies between responses observed for the individual collections. The xtnbreg command in Stata 9 was used to fit the random-effects negative binomial regression models. Ethical considerations The householders were informed on the purposes of the study and an informed consent was signed by each householder. Ethical clearance for this study was obtained from the Indian institutional review boards (Indian Council of Medical Research). Results A total of 1039 P. argentipes (42.3%) and 1415 Sergentomyia spp. (57.7%) were collected inside the 36 houses over four surveys of one night. The IRS with DDT between surveys 1 and 2 (i.e. during the sandfly reproduction season) did not affect the sandfly population (Figure 1). Significantly higher numbers of P. argentipes males were noted during the baseline survey in PT (IRR: 5.70; P = 0.008) and OT (IRR: 4.63; P = 0.028) allocated houses than in LC houses. Larger numbers of of Sergentomyia, mainly unfed, were observed in OT allocated houses (IRR: 1.96; P = 0.0480). Differences in insect counts across treatment arms at the base line study were controlled for when assessing the impact of the different treatments. Incidence rate ratios are given in Table 1 for surveys 1, 2 and 3 vs. baseline survey and for treatment effect (OT, PT and PC) compared with local untreated nets during surveys 1, 2 and 3. Household was included in the model as a random effect. After distribution of nets, no significant differences were observed between the two control arms (PC and LC). Compared with LC, there is significant reductions in males of P. argentipes with OT (IRR: 0.14; P = 0.0070) and PT (IRR: 0.23; P = 0.0340). These reductions are not significant for of P. argentipes. For Sergentomyia, only OT results in a significant reduction in males (IRR: 0.32; P = 0,014) and (IRR: 0.43; P = 0.0330). The latter reduction is mainly due to a reduced number of unfed Sergentomyia (IRR: 0.39; P = 0.0410). OT results in a significant reduction of the pooled P. argentipes and Sergentomyia (IRR: 0.43; P = 0.0250). The model shows that while controlling for the possible differences between surveys, treatment, type of house (mixed vs. human dwellings) and differences between hamlets, CDC light trap collections performed better than aspirator collection for all the groups except for fed of P. argentipes and Sergentomyia for which aspirator collection in the morning was more effective (results not shown), which confirms previous findings (Dinesh et al. 2008). Furthermore, no significant differences were observed between the types of houses (mixed vs. human dwellings results not shown). Discussion This study failed to demonstrate any impact at the household level of both long-lasting insecticidal nets on female densities of P. argentipes, although LLINs were slightly effective in reducing the male population. Sergentomyia densities (males and unfed ) were affected only by Olyset and not by PermaNet. Moreover, the indoor spraying with DDT, conducted independently by the Indian control programme, had no impact on the sandfly population. This contrasts with the findings of Kaul et al. (1994) who found no sandflies after IRS-DDT treatment. Quality of the IRS carried out in the study site can be questioned as the deposit of the insecticide on the wall was not uniform. DDT resistance tolerance has been reported in the region (Kishore et al. 2004) but cannot fully explain the low performance of this intervention. Both the LLINs have been shown to be highly effective against malaria vectors when tested in experimental huts (WHO 2001, 2004) or in large field trials. Different reasons may explain the low impact observed on sandfly densities in the current study. DDT resistance has been often described in the region and cross resistance with ª 2008 Blackwell Publishing Ltd 955

Olyset PermaNet PermaNet control Local control net 12 Phlebotomus argentipes 10 8 Crude geometric mean of collected sandflies per house 6 4 2 0 12 10 8 6 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 Sergentomyia spp. 4 2 0 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 Surveys Males Females Figure 1 Unadjusted geometric mean (with 95% CI) of males and of Phlebotomus argentipes and Sergentomyia spp. collected per house before (survey 0) and 3, 6 and 9 weeks (surveys 1, 2 and 3) after installing three nets per house. For each treatment arm, 12 houses were selected (4 houses per hamlet in three hamlets). Indoor spraying with DDT was performed in all houses between surveys 1 and 2. pyrethroids cannot be excluded. However, a recent study in Bihar (Dhiman et al. 2003) showed that P. argentipes was resistant to DDT but not to deltamethrin. A better explanation is provided by the design of the study assumed that most sandflies are breeding inside the houses and the presence of treated nets inside a single house would progressively reduce the sandfly population in that house during the reproduction season. As such decline was not observed, we could postulate that most of the sandflies are breeding outdoors. Available studies on breeding sites of P. argentipes are biased by prospecting only cattle sheds, human and mixed dwelling (Sivagnaname & Amalraj 1997; Feliciangeli 2004) and only one study mention the presence of immatures outside, e.g. in debris and shrubs (Rahman et al. 1986). Moreover, in our study, sandfly density was similar in human houses and houses shared with cattle. If most sandflies are coming from outside, they may be repelled by the treated nets as it is the case for anophelines. Such deterrent or repellent effect was not observed in Amazon Brazil (Courtenay et al. 2007) where a similar number of sandflies was recorded in houses with treated or untreated nets. The absence of deterrent effect in sandflies could explain the low impact of LLINs on vector abundance inside the treated houses. Nonetheless, LLINs should certainly not yet be dismissed as a potential tool in the fight of kala-azar based on the above data, as a possible mass effect, as well as their impact on morbidity, mortality and personnel protection 956 ª 2008 Blackwell Publishing Ltd

Table 1 Effect of Olyset (OT), PermaNet (PT) and PermaNet untreated controls (PC) on sandfly densities when compared with local untreated nets (LC) Phlebotomus argentipes Sergentomyia Males Unfed Fed Gravid Total Males Unfed Fed Gravid Total ALL Survey vs. baseline 3.8089 2.8310 0.5387 1.1822 1.5378 1.0845 0.8556 0.3244 1.2770 0.7486 0.8203 0.0280 0.0530 0.1800 0.7990 0.2420 0.8150 0.6670 0.0590 0.7270 0.3290 0.4670 Effect of PC 0.3019 0.5138 2.0162 0.5514 0.7459 1.2504 1.5894 6.2828 0.8863 1.8676 1.5937 vs. LC survey 0.1010 0.3450 0.3780 0.5830 0.5760 0.6650 0.3730 0.1360 0.8800 0.1730 0.2530 Effect of OT 0.1442 0.5340 1.4527 0.9110 0.7839 0.3164 0.3884 0.8278 0.6596 0.4245 0.4334 vs. LC survey 0.0070 0.3540 0.5990 0.9150 0.6116 0.0140 0.0410 0.8040 0.5910 0.0330 0.0250 Effect of PT 0.2321 0.4076 1.2101 2.2033 0.6796 0.9213 0.8541 0.8915 5.5602 0.9529 0.9201 vs. LC survey 0.0340 0.1770 0.7490 0.3510 0.3970 0.8620 0.7350 0.8760 0.1430 0.9030 0.8180 ln_r 1.4920 2.7827 1.7980 1.6716 2.1581 2.6787 1.5110 2.1653 6.7572 1.8055 1.9381 ln_s 2.4543 3.5391 1.6485 1.6992 3.0470 3.6606 2.6394 2.6211 7.5753 3.2319 3.7180 Incidence rate ratios are obtained through a random effects negative binomial regression with household as a random effect (with corresponding P-values in italics). The model also includes baseline survey in OT, PT, PC allocated houses when compared with LC ones, CDC light trap vs. aspirator collection, mixed houses vs. human houses, hamlets 2 vs. 1, and hamlets 3 vs. 1 (results not shown). The inverse of the dispersion is assumed to follow a Beta(r, s) distribution. was not measured in this study. The latter can only be assessed by human landing collections, which is ethically unacceptable in an endemic focus. ITNs were shown to affect not only the survival of sandflies but also their behaviour in the same room. Feeding inhibition, providing a protection of non-net users in a room containing ITN has been documented (Alexander et al. 1995) and was associated with reduced human landing rates and a doubling in proportion of sandflies alighting on walls in ITN rooms (Courtenay et al. 2007). Moreover, reduction in leishmania incidence attributable to ITN use is not always associated with a significant reduction of vector density (Jalouk et al. 2007). Experimental huts as used for evaluating ITNs and LLINs against mosquitoes (WHO 2005b) would be more appropriate to evaluate the induced mortality and exophily, feeding inhibition, deterrency but this requires that all sandflies are coming from outside, which we still cannot confirm. Only randomized community-based trials can then be helpful to assess the efficacy of LLINs against P. argentipes and its transmission of Leishmania donovani in the community. Acknowledgements The authors are grateful to the European Union for providing financial support vide Proposal Contract No.: 015374 (KALANET), to the Indian Council of medical research for providing facilities and getting approval from the Government of India Health and family Welfare, New Delhi and to the staff of the division of Vector Biology and Control as well as the KALANET project staff for laboratory and field assistance. The CDC light traps purchase was sponsored by Mr Guy Deckers (Konhef, Belgium). References Alexander B & Maroli M (2003) Control of phlebotomine sandflies. Medical and Veterinary Entomology 17, 1 18. Alexander B, Usma MC, Cadena H et al. (1995) Evaluation of deltamethrin-impregnated bednets and curtains against phlebotomine sandflies in Valle del Caucu, Colombia. Medical and Veterinary Entomology 9, 279 283. Coosemans M & Carnevale P (1995) Malaria vector control: a critical review on chemical methods and insecticides. Annales de la Société belge de Médecine tropicale 75, 13 31. Courtenay O, Gillingwater K, Gomes PAF, Garcez LM & Davies CR (2007) Deltamethrin-impregnated bednets reduce human landing rates of sandfly vector Lutzomyia longipalpis in Amazon households. Medical and Veterinary Entomology 21, 168 176. Curtis CF, Jana-Kara B & Maxwell CA (2003) Insecticide treated nets: impact on vector populations and relevance on initial intensity of transmission and pyrethroid resistance. Journal of Vector Borne Diseases 40, 1 8. Dhiman RC, Raghavendra K, Kumar V, Kesari S & Kishore K (2003) Susceptibility status of Phlebotomus argentipes to insecticides in districts Vaishaii and Patna (Bihar). The Journal of Communicable Diseases 35, 49 51. Dinesh DS, Ranjan A, Palit A, Kishore K & Kar SK (2001) Seasonal and nocturnal landing biting behaviour of Phlebotomus argentipes (Diptera: Psychodidae). Annals of Tropical Medicine and Parasitology 95, 197 202. ª 2008 Blackwell Publishing Ltd 957

Dinesh D, Das P, Picado A, Davies C (2008) The efficacy of indoor CDC light traps for collecting the sandfly Phlebotomus argentipes, vector of Leishmania donovani. Medical and Veterinary Entomology 22, 120 123. Emami MM, Yazdi M & Bashardoust N (2005) Impact of Olyset Long Lasting Nets to Control Transmission of Anthroponotic Cutaneous Leishmaniasis in Central Iran. 1 16. 2002. Final Technical Report Project No: SGS04-76, WHO, Eastern Mediterranean Region. Feliciangeli MD (2004) Natural breeding places of phlebotomine sandflies. Medical and Veterinary Entomology 18, 71 80. Jalouk L, Ahmed MA, Gradoni L & Maroli M (2007) Insecticidetreated bednets to prevent anthroponotic cutaneous leishmaniasis in Aleppo Governorate, Syria: results of two trials. Transactions of the Royal Society of Tropical Medicine and Hygiene 101, 360 367. Kaul SM, Sharma RS, Dey KP, Rai RN & Verghese T (1994) Impact of DDT indoor residual spraying on Phlebotomus argentipes in Kala-azar endemic village in eastern Uttar Pradesh. Bulletin of the World Health Organization 72, 79 81. Kishore K, Kumar V, Kesari S, Bhattacharya SK & Das P (2004) Susceptibility of Phlebotomus argentipes against DDT in endemic districts of North Bihar, India. The Journal of Communicable Diseases 36, 41 44. Kumar V, Kesari SK, Sinha NK et al. (1995) Field trail of an ecological approach to control Phlebotomus argentipes using mud & lime plaster. The Indian Journal of Medical Research 101, 154 156. Lengeler C (2004) Insecticide-treated bednets and curtains for preventing malaria (Cochrane Review). Cochrane Database of Systematic Reviews (2), CD000363 Review. Rahman SJ, Menon PKM, Rajagopal R & Mathur KK (1986) Behaviour of Phlebotomus argentipes in the foothills of Nilgiris (Tamil Nadu), South India. The Journal of Communicable Diseases 18, 35 44. Ritmeijer K, Davies C, van Zorge R et al. (2007) Evaluation of a mass distribution programme for fine-mesh impregnated bednets against visceral leishmaniasis in eastern Sudan. Tropical Medicine and International Health 12, 404 414. Sivagnaname N & Amalraj DD (1997) Breeding habitats of vector sandflies and their control in India. The Journal of Communicable Diseases 29, 153 159. Speybroeck N, Boelaert F, Renard D et al. (2003) Design-based analysis of surveys: a bovine herpesvirus 1 case study. Epidemiology and Infection 131, 991 1002. Thakur CP (2007) A new strategy for elimination of kala-azar from rural Bihar. The Indian Journal of Medical Research 126, 447 451. WHO (2001) Report of the Fifth WHOPES Working Group Meeting. (WHO CDS WHOPES 2001.4) WHO, Geneva, pp. 4 16. WHO (2004) Report of the Seventh WHOPES Working Group Meeting. (WHO CDS WHOPES 2004.8) WHO, Geneva, pp. 29 56. WHO (2005a) Regional Strategic Framework for Elimination of Kala-azar from South-East Asia Region (2005 2015). (SEA-VBC-85 Rev.1), WHO, New Delhi, 22 pp. WHO (2005b) Guidelines for Laboratory and Field Testing of Long-Lasting Insecticidal Mosquito Nets. (WHO CDS WHO- PES GCDPP 2005.11.) WHO, Geneva, 18pp. WHO (2007a) Report of the Tenth WHOPES Working Group Meeting. (WHO CDS WHOPES 2007.1) WHO, Geneva, pp. 53 61. WHO (2007b) Insecticide treated mosquito nets: a position statement. http://www.who.int/malaria/docs/itn/itnspospaperfinal.pdf (10pp). Zaim M, Aito A & Nakashima N (2000) Safety of pyrethroidtreated mosquito nets. Medical and Veterinary Entomology 14, 1 5. Corresponding Author Marc Coosemans, Department of Parasitology, Institute of Tropical Medicine of Antwerp, Nationalestraat 155, B-2000Antwerpen, Belgium. Tel.: 32 3 247 63 12; Fax: 32 3 247 63 59; E-mail: mcoosemans@itg.be 958 ª 2008 Blackwell Publishing Ltd