Background

Visceral leishmaniasis (VL), also known as kala-azar in the Indian subcontinent (ISC), is the most life-threatening form of leishmaniasis, with an estimated 50,000 to 90,000 new cases occurring worldwide every year [1]. Although India, Nepal and Bangladesh used to carry two-thirds of the global VL burden in 2004, this fraction was reduced to only 18% in 2020 [2] after a regional elimination initiative was launched in 2005. This initiative targeted the elimination of the disease from the ISC by 2015 [3], a deadline later extended to 2026 [4]. The elimination target was to reduce the annual incidence below one case per 10,000 population at the implementation unit level (district level in Nepal and sub-district level in India and Bangladesh) in all endemic areas, a target which is assumed to no longer form a public health concern [5]. VL was envisaged for elimination from the ISC because of some favorable factors like humans being the only known host or reservoir, Phlebotomus argentipes sand fly as the only vector transmitting the parasite Leishmania donovani, availability of chemical-based vector control tools, rapid disease diagnostic tools and complete treatment of the disease. Accordingly, the key strategies adopted in the region and in Nepal for disease elimination are early diagnosis and complete treatment, integrated vector management with main focus on indoor residual spraying (IRS) and effective disease and vector surveillance [6].

By 2013, the VL elimination target was achieved in all 12 officially endemic districts (districts where a full transmission cycle and therefore local transmission has been formally established, warranting the presence of the national VL elimination program) mostly situated in the central and eastern plain lowland areas (also called as “terai”) of Nepal. Nonetheless, 10 years after this first achievement, Nepal has still not been able to declare the elimination of VL as a public health problem. Several districts have breached the elimination target again since 2013, with two districts above the threshold in 2022. Also, the national VL elimination program added six districts in 2016 and five more in 2021 to the endemic district’s list, after the local transmission was epidemiologically and entomologically verified [2, 6, 7]. These newly endemic districts include hilly and mountainous areas from the eastern, central and western parts of the country. At present, 42 districts are listed as endemic, 30 districts as endemic doubtful (reports of at least one locally acquired VL case in the last 10 years and full cycle of transmission is yet to be verified) and only five districts as non-endemic (no locally acquired VL cases in the last 10 years) out of 77 total districts in Nepal (unpublished data).

The continued presence of sporadic VL cases even in the districts below the elimination threshold in the historical endemic areas sustains the risk of new outbreaks, local resurgence and transmission to neighboring unaffected areas. In this context, the elimination strategies need critical evaluation, realignment and reconsideration of relevance in the current phase of sustaining the elimination target and preventing re-emergence of disease transmission [5, 8]. One of the key strategies, i.e. vector control with synthetic pyrethroid-based IRS in human and animal shelters to reduce L. donovani transmission [9, 10], is vulnerable to poor performance because of logistic constraints, irregular operational practices and limited updated knowledge on vector bionomics [11,12,13]. Vector abundance and seasonality in an area are often related to the extent of vector-human contact and Leishmania transmission [14], and up-to-date information facilitates determining the appropriate spray timing for effective IRS implementation. Furthermore, even in the absence of reported VL cases, transmission can still be ongoing, as the vast majority of infections remain asymptomatic in humans, and clinical cases can be missed or can migrate elsewhere before diagnosis. Surveillance of the prevalence of Leishmania infection in vector sand flies can help to understand to what extent the parasite is still circulating in a certain area and provide crucial information for the prediction of the risk and expansion of leishmaniasis [15,16,17,18]. The role of a sand fly species as a vector is invariably associated with the intensity of host-vector contact for blood feeding; therefore, it is undoubtedly important to identify and quantify the source of blood meals in wild-caught vector sand flies to determine the risk of Leishmania transmission in the human population [19, 20].

Several studies investigated sand fly seasonality, prevalence of Leishmania in sand flies and host preferences in the early years of the elimination initiative, mostly to assess the Leishmania transmission dynamics and to check the efficacy of various vector control tools in Nepal [13, 17, 21,22,23]. As considerable changes in the epidemiological situation have been experienced in Nepal since then, updated information on these entomological indicators is sought to guide the VL elimination efforts for sustainable elimination of the disease in the country.

In this study, we aimed to assess the status and risk of L. donovani transmission in longstanding endemic districts by collecting data on seasonality, natural L. donovani infection rate and host preference of vector sand flies from two epidemiologically contrasting areas; villages with VL cases being reported in recent years and neighboring villages within the same endemic districts but without any such cases reported during the last 10 years in the eastern part of Nepal.

Methods

Study area

Three VL endemic districts in eastern Nepal, viz. Morang, Sunsari and Saptari (Fig. 1), were selected for field work. We selected these districts based on (i) the cumulative VL case load over the years 2014–2017 (Additional file 1: Table S1) and (ii) their proximity to the BP Koirala Institute of Health Sciences (BPKIHS), Dharan, Nepal. In each district, one village reporting VL cases in three consecutive years prior to the inception of the study was assigned as the ‘VL village’, and a second village without any reported VL cases in the last 10 years before the start of the study was selected as ‘non-VL village’. Corresponding VL and non-VL villages in each district were carefully selected within 5–10 km distance of each other to minimize the differences in spatial, ecological and housing characteristics. These VL and non-VL villages were Majhare (26.406992 N, 87.337322 E) and Pokhariya (26.401985 N, 87.363547 E) in Morang district, Amahibelha (26.466103 N, 87.188618 E) and Chimadi (26.492791 N, 87.186457 E) in Sunsari district and Belhichapena (26.524450 N, 86.677925 E) and Gamhariya Parwaha (26.541511 N, 86.663169 E) in Saptari district respectively (Fig. 1).

Fig. 1
figure 1

Map of Nepal showing visceral leishmaniasis endemicity (situation as of the year 2016–2017) (inset) and the locations of the study districts and villages. VL villages, villages reporting VL cases in three consecutive years prior to the inception of the study. Non-VL villages, villages without any reported VL cases in the last 10 years before the start of the study. The map was produced with QGIS (version 3.10.2) with open access shapefile. (https://opendatanepal.com/dataset/new-political-and-administrative-boundaries-shapefile-of-nepal)

Full size image

Sand fly collection and morphological identification

In each of the selected villages, 10 households including human dwellings and mixed dwellings (cattle sheds sharing space with humans) were purposively selected for longitudinal monitoring of the sand fly density for 12 consecutive months, from July 2017 until June 2018. Centre for Disease Control and Prevention miniature light traps (Model 2836 BQ; BioQuip Products, Rancho Dominguez, CA 90220, USA) were used as a standard method of sand fly capturing. One light trap (LT) was placed in the inner lower corner of the main room or cattle shed of each of the selected households and kept functioning for 12 h starting at 18 h to 6 h. In the consecutive morning, resting sand flies were collected using mouth aspirators spending 15 min time in the same household where LT was placed. Both LT and aspirator collections were made for three consecutive nights and mornings in all 10 houses of a village every month during the study period.

We transferred the LT and aspirator collections to the BPKIHS entomology laboratory on a daily basis for the morphological identification of the sand flies using a binocular microscope and regional keys [24, 25]. All sand flies were preserved in 80% ethanol and refrigerated at 4 °C. Male P. argentipes and other species of sand flies were stored at the village level per month. Female P. argentipes were segregated and stored at the physiological stages (unfed, fed and gravid) at the household level by collection method (light trap versus aspirator) and place of collection (human dwelling and mixed dwellings) level every month.

Molecular analyses of the P. argentipes sand flies

DNA extraction

Genomic DNA was extracted from pools of gravid P. argentipes sand flies for Leishmania detection. Pooling of the gravid P. argentipes was done by collection method (LTs and aspirator), household, village and the survey month. Each pool was comprised of a minimum one to a maximum 20 sand flies. In case of more than 20 gravid sand flies collected at an event, the number was divided into two or more pools each containing not more than 20 specimens. Besides, DNA was extracted from individual specimens of blood-fed vector sand flies for blood meal source identification. For reference purpose, DNA extractions were also obtained from blood samples (approximately 5 ml each) of vertebrate hosts like buffalo, cow, goat, pig, rabbit, chicken, duck and pigeon.

The sand flies were crushed using a sterile micro-pestle and homogenized before following the DNA extraction procedure using DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. DNA elution of 50 µl was collected. The quantity and purity of the DNA extract were checked with a NanoDrop2000 spectrophotometer (Thermo Scientific, DE). DNA extracts were preserved at − 20 °C until being subjected to specific polymerase chain reactions.

Detection of Leishmania infection in vector sand flies

Assessment of Leishmania infection was done on pools of all gravid P. argentipes collected throughout the year. A TaqMan probe-based real-time PCR assay targeting the 115 bp length of the small-subunit ribosomal RNA gene of Leishmania (SSU-rRNA) was conducted on DNA extracts from the pools of gravid P. argentipes. A Leish 18S Set (LE18S) (Integrated DNA Technologies, Belgium) was used in the assay with already described and tested primers: M18S-F-L (5′-CGT AGT TGA ACT GTG GGC TGT GC-3′) and M-18S-R-L (5′-ACT CCC GTG TTT CTT GTT TCT TTG AA-3′) [17, 26], along with a double quenched L-probe (FAM-CTGGTCGTCCCGTCCATGTCGGATT-BHQ1 ZEN). The PCR reaction was carried out in a final volume of 25 µl containing 0.4 µM M18S-F-L, 0.4 µM M-18S-R-L and 0.1 µM L-probe, 1X GoTaq Probe qPCR Master Mix (Promega, USA), 0.1 mg/ml bovine serum albumin (Promega, USA) and 2.5 µl sample DNA. The PCR amplification was conducted in the Rotor-GeneQ real-time PCR System (Qiagen, Hilden, Germany). The thermal profile optimized for the reaction was the activation of the Taq polymerase present in Go Taq Probe Master Mix at 94 °C for 15 min and then 45 cycles of denaturation at 94 °C for 30 s, annealing at 60 °C for 30 s and extension at 72 °C for 45 s followed by the final extension at 72 °C for 5 min. In each PCR run, one positive control consisting of 1 pg of L. donovani DNA (BPK282/0) and one negative control containing PCR grade water were included for quality control of the test. Additional qPCR assay targeting the conserved and highly expressed spliced-leader [27] mini-exon sequence [28] was performed to verify the outcomes of the SSU-rRNA qPCR assay.

After the detection of the Leishmania-positive pools, Leishmania species identification was performed by amplification and sequencing of part of the internal transcribed spacer 1 (ITS-1) region [29, 30]. SSU-rRNA and SL-RNA positive samples were subjected to an ITS-1 nested PCR. A 25 µl PCR reaction mixture contained 1× GoTaq® G2 Master Mix (Promega, USA), 5 µl sample and 0.5 µM forward (5ʹ-CTG GAT CAT TTT CCG ATG-3ʹ) and reverse (5ʹ-TGA TAC CAC TTA TCG CAC TT-3ʹ) primer. Samples were tested on a Bio-Rad T100 thermal cycler using a 5 min activation step at 95 °C, 35 amplification cycles under optimized conditions (30 s at 95 °C, 30 s at 53 °C and 30 s at 72 °C) and a 10 min final elongation at 72 °C. PCR products were checked by 1% agarose gel electrophoresis and purified using a QIAquick PCR Purification Kit (Qiagen, Hilden, Germany) after which they were subjected to a nested PCR reaction using the same conditions as described above. PCR samples were again checked by 1% agarose gel electrophoresis and purified with a QIAquick PCR Purification Kit after which they were sent for Sanger sequencing on an Applied Biosystems 3730XL DNA Analyzer at the Neuromics Support Facility of the University of Antwerp. Leishmania species identification was analyzed using SnapGene and NCBI BLAST.

Detection of blood meal sources

A targeted ~ 350 bp segment of vertebrate cytochrome b (Cytb) gene was amplified using the universal primers Cytb1-F (L14841: 5′-CCA TCC AAC ATC TCA GCA TGA TGA AA-3′) and Cytb2-R (H15149: 5′-GCC CCT CAG AAT GAT ATT TGT CCT CA-3′) [31, 32] for host identification. Each PCR was performed on a 25 μl reaction mixture containing 1× GoTaq Green Master Mix with 2 mM MgCl2 (Promega, USA), 0.4 μM each of two primers (Biolegio, The Netherlands) and 5 μl DNA template. Each set of PCR reactions was accompanied by a negative control having PCR grade water and a positive control having DNA extract from human blood. The PCR reaction was performed on an Eppendorf Mastercycler pro PCR System (Hamburg, Germany) setting the thermal profile as pre-activation step at 95 °C for 2 min, followed by 40 cycles of denaturation at 92 °C for 30 s, annealing at 55 °C for 45 s and extension at 72 °C for 1 min and a final extension at 72 °C for 10 min. The PCR products, positive and negative controls were loaded on 2% agarose gel (Eurogentec, Belgium), stained with ethidium bromide (Promega, USA) and examined under a UV trans-illuminator/gel documentation system (Syngene, Cambridge, UK). PCR products (amplicons) were outsourced for Sanger sequencing at the Macrogen Company, South Korea. All generated sequences were checked and edited in BioEdit Sequence Alignment Editor 7.0.5.3 [33] and queried against NCBI BLAST.

Molecular validation of P. argentipes identification

A subsample of female P. argentipes subjected to Leishmania infection assessment (n = 120) and blood meal analysis (n = 31) was verified by DNA barcoding method based on the amplification of targeted 650 bp fragment length of mitochondrial cytochrome c oxidase subunit I (COI) gene as described by Folmer et al., Hebert et al. and Kumar et al. [34,35,36].

Data management and statistical analysis

Entomological data were recorded in a pre-tested paper-based data collection tool and entered in a database developed in Epi Info version 3.5.1 [37]. All statistical analyses and graphical plotting were performed with R software version 4.1.0 [38].

Sand fly abundance and seasonality analysis

Sand fly abundance and seasonality analysis were based on the LT collections only, considering them as the standard method of vector collection. Sand flies captured by mouth aspirators were excluded from this analysis to avoid any possible collection bias. Mean number of P. argentipes sand flies captured per trap-night per house is represented in a line graph by month, type of village and district. We fitted a generalized linear model (GLM) with a negative binomial distribution to assess the sand fly abundance in function of the type of village, district and month. We used a Bayesian approach to better account for the uncertainty of the estimates using the R package “rstanarm” [39, 40]. This approach produces a 95% credible interval, which means that given the observations, there is a 95% chance that the value lies between the limits of the credible interval [41]. Results of the analysis are represented as an incidence rate ratio (IRR) and credible interval (CI) at 95%.

Estimation of the prevalence of Leishmania infection rate

For the assessment of Leishmania infection, all gravid P. argentipes sand flies collected in both LTs and aspirators were used. The prevalence of Leishmania infection in pools of unequal size was estimated by a Bayesian Markov Chain Monte Carlo (MCMC) model as described by Speybroeck et al. [42, 43]. In this method, prior knowledge of the diagnostic test (in this case, the SSU-rRNA PCR method for detection of Leishmania in pools of vector sand flies) characteristics (sensitivity and specificity) was incorporated with the test results. We assumed an imperfect sensitivity, ranging uniformly between 60 and 95%, and a perfect specificity for the PCR test [26]. The result obtained was a posterior probability distribution of the prevalence; the mean and credible intervals at 95% were then calculated from it. Prevalence was estimated using the “truePrevPools” function in the R package “prevalence” [44], and credible intervals (CI) were calculated in the R package “bayestestR” [45, 46].

Blood meal analysis

A subsample of 50 blood-fed P. argentipes collected in both light traps and aspirators from mixed dwellings of each village during the initial months of the surveys was used for blood meal analysis. Host preference was calculated as the percentage of P. argentipes sand flies fed on a specific host relative to the total number of blood-fed P. argentipes tested per village category (i.e. VL versus non-VL villages).

Results

Sand fly abundance and seasonality

A total of 39,692 Phlebotomine sand flies were captured by both CDC light traps and manual aspirators from six villages of three districts during the 12 months of the survey. The vector species P. argentipes (n = 31,382; 79.1%) was dominant compared to other sand fly species Phlebotomus papatasi (n = 1632; 4.1%) and Sergentomyia spp. (n = 6678; 16.8%).

Considering the vector sand flies collected in light traps only (n = 20,101), males (n = 11,725; 58.3%) outnumbered females (n = 8376; 41.7%). About double the number of P. argentipes sand flies were captured per trap-night per house in non-VL villages (n = 12.66; IRR = 1.90; CI at 95% = 1.66–2.17) compared to VL villages (n = 6.13). Distinct variation in vector density was observed at the district level with higher per trap-night per house collections in Saptari (n = 15.47; IRR = 1.05; CI at 95% = 0.89–1.23) and lower in Sunsari (n = 3.35; IRR = 0.36; CI at 95% = 0.31–0.41) compared to collections from Morang (n = 9.35). At district level, vector density per trap-night per house was lower in the non-VL village (n = 5.55; IRR = 0.53, CI at 95% = 0.44–0.63) than in the VL village (n = 13.12) in Morang, higher in the non-VL village (n = 3.79; IRR = 1.53, CI at 95% = 1.28–1.85) than in the VL village (n = 2.92) in Sunsari and almost 12 times higher in the non-VL village (n = 28.69; IRR = 12.7, CI at 95% = 10.31–15.76) than in the VL village (n = 2.41) in Saptari (Table 1).

Table 1 Association of the Phlebotomus argentipes sand fly density in VL and non-VL villages in aggregate (combining all three districts), at the individual district level and in monthly collections and their 95% credible intervals
Full size table

Month-wise abundance and seasonality of vector sand flies varied at different degrees in all three districts (Table 1, Fig. 2). In Morang, peaks in sand fly density were seen in July and September 2017 in the VL village and in July 2017 and June 2018 in the non-VL village (Table 1, Fig. 2). In Sunsari, peaks in sand fly density were observed during November 2017 and June 2018 in the VL village, while density peaks were not very clear in the non-VL village. In Saptari, many P. argentipes were captured in the non-VL village compared to all other villages year round, and the highest density peaks were observed in the months of July and November in 2017 and June in 2018. In the VL village of the same district, density peaks were seen in the months of March and May 2018. Finally, in the analysis based on the cumulative collections from all the villages, peaks in P. argentipes density were marked in the months of June and July and from September to November while the density dropped to almost zero in the months of January and February (Fig. 2, Additional file 2: Table S2).

Fig. 2
figure 2

Monthly distribution and seasonality of Phlebotomus argentipes in light trap collections in Morang, Sunsari and Saptari districts

Full size image

Prevalence of Leishmania infection in vector sand flies

Leishmania infection was assessed in 796 pools of 3735 gravid P. argentipes sand flies in total. Leishmania infection was detected in four (1.37%) pools out of 293 from VL villages and one (0.2%) pool out of 503 from non-VL villages. The estimated prevalence of Leishmania infection in vector sand flies was 2.2% (CI at 95% = 1.1%–3.7%) in VL villages and 0.6% (CI at 95% = 0.2%–1.3%) in non-VL villages (Table 2). These Leishmania-infected vector sand flies were collected in the months of May, September and November. Three of five SSU-rRNA-positive pools were confirmed using the SL-RNA PCR. The Leishmania species was identified and confirmed as L. donovani in two pools from VL villages and one pool from a non-VL village. ITS-1 nested PCR was unsuccessful in two positive pools from a VL village (Table 3).

Table 2 Overview of the number of gravid Phlebotomus argentipes tested, Leishmania-positive pools and estimated infection rate with 95% credible intervals in VL and non-VL villages
Full size table
Table 3 Leishmania species identification based on ITS-1 sequencing in Leishmania-specific SSU-rRNA-positive samples
Full size table

Blood meal sources and host preference patterns

A sample of 301 of 3443 fed P. argentipes collected from mixed dwellings in all six villages was assessed for identification of the source of blood meal, and vertebrate host species were successfully identified in 295. We submitted selected sequences of the cytb gene of the host species to GenBank (accession nos. OQ535519–OQ535565). The most common sources of the blood meal were humans (63.4%; 187/295), followed by domestic buffaloes (18.9%; 56/295), cows (13.9%; 41/295), goats (3.1%; 9/295) and pigs (0.7%; 2/295). In both categories of villages, P. argentipes fed primarily on humans (52.7%; 78/148 in VL vs. 74.2%; 109/147 in non-VL villages). Buffaloes (31.8%; 47/148) in VL villages and cows (16.3%; 24/147) in non-VL villages were the second preferred hosts over goats (2.7%; 4/148 in VL vs. 3.4%, 5/147 in non-VL villages). A blood meal from pigs was present in only two fed P. argentipes from VL villages (Fig. 3).

Fig. 3
figure 3

Blood meal sources of P. argentipes sand flies captured from mixed dwellings in VL and non-VL villages

Full size image

Molecular validation of P. argentipes species

The samples of morphologically identified P. argentipes sand flies subjected to the assessment of L. donovani infection and blood meal analysis were confirmed as the same species by DNA barcoding method.

Discussion

Phlebotomus argentipes sand fly density varied among districts and village types. In one of the districts, P. argentipes density was higher in the VL village than in the non-VL village, the other districts had higher collections in non-VL villages. The P. argentipes density fluctuated with the season with an increase in June and July and again in September to November. Interestingly, L. donovani-infected vector sand flies were present in both VL and non-VL villages. Furthermore, P. argentipes showed a high affinity for human blood in mixed dwellings in both types of villages.

Information on seasonality of vector sand flies is crucial to identify the peak season of human-vector contact and thus the period of maximum risk of L. donovani transmission. Based on the previous studies and information on seasonality [13, 47, 48], the national VL elimination program in Nepal has been following the schedule of two rounds of IRS using synthetic pyrethroids (alphacypermethrin, deltamethrin or lambda-cyhalothrin) in the reported VL areas or villages, the first round in the months of May–June and second in September–October [6]. The observed seasonality of vector sand fly density in this study was similar to other studies conducted in Nepal [13, 47] and bordering areas of India [49,50,51] and aligns with the existing practice of IRS. Nevertheless, the seasonality of vector abundance is conducive for Leishmania transmission throughout the year except in the months of January and February and raises concerns about the suitability of two rounds of IRS with pyrethroids, the insecticide group that is reported effective to reduce the vector sand fly density up to 4 weeks only in Nepal and Bangladesh [12, 27, 52, 53]. These findings argue for the re-evaluation of vector control strategy in terms of the number of IRS rounds to be considered to cover the long transmission season or to switch or adapt to alternative insecticides that have proven longer residual impact on the sand fly density.

The Leishmania infection in vector sand flies as another key indicator of ongoing transmission was also assessed in this study. Some of the previous studies conducted in Nepal and neighboring parts of India during the earlier years of VL elimination showed a range of 0.5% to 17.4% prevalence of Leishmania infection in vector sand flies [15,16,17, 54,55,56]. In these investigations, prevalence of infection, however, was supposedly influenced by the type of PCR test used, the statistical method to estimate infection rate and the physiological stages of the female P. argentipes (unfed, fed and gravid). Thus, the selection of only gravid P. argentipes for the Leishmania infection assessment in this study was to ensure the sand flies had taken a blood meal at least once and that the meal had already been digested. This way, we assumed the sand flies positive for Leishmania to be a proxy for infective sand flies and thus able to transmit the parasite during their next blood meal. Two PCR techniques were used for quality assurance and molecular identification confirmed the L. donovani species in the infected vector sand flies. All four L. donovani-infected pools from VL villages originated from either the house of an ex-VL patient or directly adjoining houses of the same. Interestingly, we detected an L. donovani-positive pool from a non-VL village as well, supporting the findings of the existence of asymptomatic infection in the human population in the areas with zero reported VL cases for more than a decade prior to the study [57]. The presence of infected vector sand flies even in the so-called non-VL village indicates the circulating Leishmania parasite in vector and human populations and could possibly cause resurgence of cases in these areas if left unnoticed or neglected. Thus, it is anticipated the program needs to remain vigilant for the existence of transmission even at times when VL case reporting from the longstanding endemic districts is scanty or null.

We found that vector sand flies collected from mixed dwellings had a high affinity towards humans. In mixed dwellings, sand flies were assumed to have equal access to the various types of hosts and minimal influence of biotopes on the feeding behavior. The high human blood index reported in previous studies from Nepal [21, 22] well supports the findings of our study. Despite P. argentipes being marked as an opportunistic feeder, many of the recent studies reported a high human blood index in collections from various biotopes including mixed dwellings [19, 50]. From our observations, we do not have evidence of a substantial change in feeding behavior (i.e. anthropophilic to zoophilic) of P. argentipes even after years of IRS in endemic districts; hence, the species still possesses a high potential for parasite transmission in the domestic environment.

The main strength of this study is that we assessed the risk of ongoing L. donovani transmission in VL endemic districts during the elimination era based on multiple entomological indicators. These indicators were well documented from two contrasting VL epidemiological settings, i.e. VL and VL non-villages within endemic districts, which, to our best knowledge, is the first study of this kind in Nepal. This study is also the first in 15 years to update our knowledge on seasonality, prevalence of Leishmania infection and host preference of the vector sand flies in endemic districts of Nepal. However, this entomological evidence would have been complemented by epidemiological and serological data for a complete situation analysis of L. donovani transmission in these VL endemic areas where the caseload has been reduced below the elimination threshold in recent years.

Conclusions

Our findings highlight the risk of ongoing L. donovani transmission in villages with and without reported VL cases within VL endemic districts included in the disease elimination program and that P. argentipes is still an efficient vector. The favorable entomological indicators for Leishmania transmission in non-VL villages emphasize the remaining threats to the VL elimination program and show that it needs to remain vigilant for the re-appearance of transmission even years after the last VL case has been reported from the endemic districts. In addition, we recommend that the program should strengthen the entomological and epidemiological Leishmania surveillance in Nepal to better adapt the disease elimination strategies. Also, we recommend conducting similar studies in other VL endemic districts representing various geo-ecological zones and administrative regions to monitor Leishmania transmission and track the progress towards elimination.