Wuchereria bancrofti transmission pattern in southern Mali prior to and following the institution of mass drug administration

Background The Global Programme to Eliminate Lymphatic Filariasis (GPELF) was launched in 2000 with the goal of stopping transmission of lymphatic filariasis (LF) through yearly mass drug administration (MDA). Although preliminary surveys of the human population in Mali suggested that Wuchereria bancrofti infection was highly endemic in the Sikasso district, baseline entomological data were required to confirm high levels of transmission prior to the selection of villages in this region for a study of the impact of MDA on transmission of LF by anopheline vectors. Methods W. bancrofti transmission was assessed in 2001 (pre-MDA) and 2002 (post-MDA) in the Central District of Sikasso in southern Mali by dissection of Anopheles mosquitoes caught using the human landing catch (HLC) method. The relative frequencies and molecular forms of An. gambiae complex were determined. Results The majority (86%) of the anopheline vectors captured were identified as An. gambiae complex, and these accounted for >90% of the entomological inoculation rate (EIR) during both years of the study. There was a dramatic decrease in the number of An. gambiae complex mosquitoes captured and in the An. gambiae complex infectivity rates following MDA, accounting for the observed decrease in EIR in 2002 (from 12.55 to 3.79 infective bites per person during the transmission season). An. funestus complex mosquitoes were responsible for a low level of transmission, which was similar during both years of the study (1.2 infective bites per person during the transmission season in 2001 and 1.03 in 2002). Conclusions Based on the entomological data from this study, the district of Sikasso was confirmed as an area of high W. bancrofti transmission. This led to the selection of this area for a multi-national study on the effects of MDA on LF transmission by anopheline vectors. Comparison of vector transmission parameters prior to and immediately following the first round of MDA demonstrated a significant decrease in overall transmission. Importantly, the dramatic variability in EIR over the transmission season suggests that the efficacy of MDA can be maximized by delivering drug at the beginning of the rainy season (just prior to the peak of transmission).

Results: The majority (86%) of the anopheline vectors captured were identified as An. gambiae complex, and these accounted for >90% of the entomological inoculation rate (EIR) during both years of the study. There was a dramatic decrease in the number of An. gambiae complex mosquitoes captured and in the An. gambiae complex infectivity rates following MDA, accounting for the observed decrease in EIR in 2002 (from 12.55 to 3.79 infective bites per person during the transmission season). An. funestus complex mosquitoes were responsible for a low level of transmission, which was similar during both years of the study (1. Conclusions: Based on the entomological data from this study, the district of Sikasso was confirmed as an area of high W. bancrofti transmission. This led to the selection of this area for a multi-national study on the effects of MDA on LF transmission by anopheline vectors. Comparison of vector transmission parameters prior to and immediately following the first round of MDA demonstrated a significant decrease in overall transmission. Importantly, the dramatic variability in EIR over the transmission season suggests that the efficacy of MDA can be maximized by delivering drug at the beginning of the rainy season (just prior to the peak of transmission).

Background
Lymphatic filariasis (LF) is a chronic debilitating infection caused by the mosquito-borne filarial nematodes, Wuchereria bancrofti (W. bancrofti), Brugia malayi and Brugia timori. Worldwide, more than one billion people are at risk of infection, among which 120 million are already infected, the majority of whom are in India, with an estimated 49.2% of the infection burden, followed by sub-Saharan Africa with 34.1% [1]. W. bancrofti is responsible for approximately 90% of LF cases worldwide and all of the cases in sub-Saharan Africa, where the most common vectors are An. gambiae and An. funestus complexes [2]. Forty-three million people are officially recognized as being disabled due to LF and millions more suffer from social and psychological problems [3]. Nevertheless, the socio-economic burden of LF is underestimated in many endemic areas that are among the poorest of the world [4].
In Mali, although the public health importance of LF was noted as early as the 1970′s [2], the prevalence and distribution were not studied again until 2002, when the National Lymphatic Filariasis Elimination Program (NLFEP) provided the first countrywide LF map. Based on antigen testing using ICT cards, all 8 administrative regions of Mali were shown to be endemic for LF with an overall prevalence of 7.07%, ranging from 1% in the north to 18.6% in the south. In anticipation of the launch of NLFEP activities in Mali, a pilot study of the impact of mass drug administration (MDA) with albendazole and ivermectin on W. bancrofti transmission was initiated in collaboration with the World Health Organization (WHO) in Ghana, Mali, and Nigeria. The present study was designed to provide baseline data on vector transmission in this highly endemic region of Mali and to assess the effects of the first round of MDA.

Study site identification and characterization
The study was initiated in the Sikasso district in southern Mali prior to the introduction of MDA for the elimination of LF. This area was historically known to be endemic for W. bancrofti, and mapping surveys performed by the National Program for the Elimination of Lymphatic Filariasis confirmed a high prevalence of circulating filarial antigen (CFA) positivity (as assessed by ICT card testing of 50-100 individuals/village) in the village of Dozanso and a neighboring village in 2001 (unpublished data). Additional ICT card surveys were subsequently carried out in the larger villages surrounding Dozanso in 2001 and led to the selection of 6 high prevalence villages (Dozanso, Gondaga Missasso, N'Torla, Niantansso and Zanadougou) for baseline entomological studies.
The study villages are comparable in terms of sociocultural indicators, health care seeking behavior and disease perception. The distance between the villages and the community health care center of Kolokoba ranges from 6 to 15 km (mean 9.5 km), occupied by cotton fields, backwaters, and trees typical of the dense Sudan Savannah vegetation. Rainfall in this region ranges between 1200 and 1500 mm per year, with a rainy season that extends from July to December. Due to the high levels of transmission documented during the first year of the study, yearly MDA was instituted in the 6 study villages in June 2002, one month prior to the second entomological survey.

Study population
A complete census, including the name, age, sex and profession of each inhabitant, was performed in all 6 villages. All dwellings were recorded and assigned an identification number. A global positioning system device (GPS) was used to produce basic maps of the locations of the 6 villages within the Central District of Sikasso. The total population of the 6 study villages was 3,681 in 2001, consisting primarily of farmers, whose main occupations are agriculture (cotton, maize, millet and peanut) and domestic animal breeding. The ICT card surveys carried out in 2001 revealed CFA prevalences varying from 81.8% in Niantanso (165/202) to 24.6% in Zanadougou (50/202) ( Table 1). The prevalence of microfilaremia was assessed in 2002 (prior to the initiation of MDA) by examination of 3 slides of 20 μl of night blood/subject, and ranged from 40% in Dozanso (48/120) to 13.8% in N'Torla (27/196).
A collective village-wide oral consent was obtained from the villages' elders, and all study participants signed individual written informed consent. The study protocol and consent forms were approved by both the Institutional Review Board (IRB) of the World Health Organization/Tropical Diseases Research (WHO/TDR) and the ethics committee of the Faculty of Medicine, University of Bamako, Mali.

Study design
This was a longitudinal study during which monthly entomological surveys were performed in 6 study villages from July to December in 2001 (prior to initation of MDA with albendazole and ivermectin) and in 2002 (one month after the first MDA).

Laboratory analysis
Entomological surveys were performed 12 days per month (2 days/village/month) by the same team. Mosquitoes were collected by two trained field personnel in each room of four different houses in each village using the HLC method. One collection team worked from 6:00 pm to midnight and the second from midnight to 6:00 am. Mosquitoes were captured using a Colluzi and Petrarca type mouth aspirator connected to a paper cup as the storage container. A supervisor retrieved the containers at two-hour intervals. The captured mosquitoes were kept overnight at ambient temperature in a paper cup under a damp cloth and dissected the following morning.
Mosquitoes were sorted morphologically for species identification (An. gambiae and An. funestus complexes). Some An. gambiae complex specimens were processed by polymerase chain reaction (PCR) method to distinguish between the 2 members of the complex (An. arabiensis and An. gambiae ss). The An. gambiae ss were further processed by PCR to identify the molecular forms, M and S, as described by Favia et al. [5,6]. The head, thorax and abdomen were dissected separately for each mosquito and recovered parasite larvae were categorized into L1, L2 or L3 stages.
Entomologic parameters assessed included infection rate, infectivity rate, human biting rate (HBR) and entomological inoculation rate (EIR) and were calculated as previously described [5,7]:   (Table 4). This high frequency of the S molecular form was observed in all of the study villages (data not shown).

Vector infection rates and transmission pattern
Both An. gambiae and An. funestus complexes were found to be harboring infective larvae during the two years of study (2001 and 2002). In July and August, 100% of the infective W. bancrofti larvae were recovered from An. gambiae complex (Figure 1). An. funestus complex harbored infective larvae with increasing rates from September to November. No infective mosquito was recovered in December 2001 (  (Table 3). In 2001, An. funestus complex was found to be carrying W. bancrofti larvae from September to November with monthly infection rates ranging from 2.7% to 6.5% while infected An. gambiae complex was recovered each month from July to December with rates ranging from 2.8% to 6.8% (Table 2). In 2002, both An. funestus and An. gambiae complexes were found infected from July to December with monthly rates ranging respectively from 4.3% to 14.3% and 1.1% to 16.7% (Table 3) In 2001, the overall monthly EIR was more than 10 fold higher for An. gambiae complex than An. funestus complex. The An. gambiae complex was responsible for 0 to 35.53 infective bites per human per month. The An.   (Table 3).

Discussion
The baseline entomological data collected in this longitudinal study confirmed measurable transmission of W. bancrofti in the 6 study villages in Sikasso prior to the initiation of MDA. As had been reported previously in Mali, An. gambiae and An. funestus complexes were the predominant vectors [2,8]. In keeping with the high prevalence of human infection in these villages, the recorded vector densities were higher, especially for An. gambiae complex, than those reported in Banambani (Sudan savannah area) in Mali, where W. bancrofti is endemic but the prevalence of infection is lower [8]. Other Anopheles species (An. pharaoensis, An. nili, An. rufipes) were collected but not systematically processed during this study because of their very low relative frequencies, precluding an epidemiologically significant role in the transmission of LF, and the fact that they have not been recognized as vectors of W. bancrofti in Mali and other neighboring West African countries [2,9]. Among the An. gambiae complex members examined by PCR for specific species identification, the S form of Anopheles gambiae s.s. was predominant. A predominance of the S form of An. gambiae s.s among vectors of LF has also been observed in Ghana [10]. Although the PCR identification of An. gambiae complex species and An. gambiae s.s. molecular forms was not performed on all the collected mosquitoes for logistical reasons, at least 76% of the mosquitoes collected each month were dissected to ensure that the samples tested were temporally and geographically representative. Mosquitoes were sent for PCR analyses without identity numbers that could link them to the dissection results precluding the determination of infection rates for the different molecular forms.
Overall, the highest monthly vector relative frequencies for An. gambiae complex were found in July and August (at least 99% of the vectors collected in the month), while those of An. funestus complex were observed in November and December (at least 33% of the vectors collected in the month). Similar variations in the relative frequencies of the two vectors were reported in Banambani [8] and are related to differences in environmental conditions during the transmission season and the breeding preferences of each species (An. funestus complex prefers shadowed and vegetated breeding sites while An. gambiae complex prefers sunny breeding sites with limited vegetation) [2,8]. The frequencies of An. gambiae s.s. and An. arabiensis, two members of the An. gambiae complex, also showed differing patterns during the transmission season.
Due to the low infection and infectivity rates, processing of pools of Anopheles vectors for W. bancrofti infection is the recommended strategy for following vector transmission rates during MDA [11,12]. A recently developed L3 specific RT-PCR allows infective pools to be distinguished from infected pools and provides a more accurate determination of the transmission potential for W. bancrofti [9,13]. In the present study, the infection and infectivity profiles of the two morphologically distinct Anopheles species complexes (An. gambiae s.l. and An. funestus) were quite different, suggesting that the two species complexes should be processed for PCR in separate pools if detailed information regarding their relative contributions to monthly transmission is desired. Nonetheless, in the setting of post MDA assessment, where human-vector contact is the main factor of interest, An. gambiae and An. funestus complexes can be processed in the same pool [11,13].
In 2002 (post MDA), the number of mosquitoes caught was approximately half that in 2001 (before MDA). This effect was most dramatic for An. gambiae complex where the number captured decreased by almost 50%. Potential reasons for this decrease in mosquito numbers include changes in climate, increased awareness of the study area population with respect to the role of mosquitoes in disease transmission (resulting in less breeding sites and increased use of insecticide treated nets), and the effect of ivermectin on mosquito survivorship. Examination of the rainfall, temperature and humidity records for the region did not show any major differences between 2001 and 2002, suggesting that climate did not play a major role in the decreased number of mosquitoes. Although decreases in mosquito numbers following the initiation of MDA [14] and an effect of ivermectin on mosquito survivorship [15,16] have both been described, these factors were not directly addressed in the present study.
Whereas the decreased vector numbers in 2002 (post-MDA) clearly contributed to the overall decrease in EIR observed for the An. gambiae complex, infectivity also declined significantly in 2002, suggesting that multiple factors may have played a role in the observed decrease in transmission including the decrease of the mf prevalence and loads consecutive to the MDA [14]. The fact that a similar decrease in EIR was not seen for An. funestus complex may have been due to the low overall numbers of An. funestus complex mosquitoes captured, although a higher degree of facilitation by An. funestus complex as compared to An. gambiae complex cannot be excluded. Unfortunately, the study was not designed to address this issue, and published data comparing facilitation between the two species are limited [17,18].
Despite the fact that the overall mosquito infection rates were relatively stable during the six months of collection in each of the two transmission seasons, the EIR for W. bancrofti varied considerably over the course of the seasons as a result of the large differences in vector densities and HBR [11]. This has important implications for the timing of MDA for LF in this region, since drug administration conducted at the beginning of the rainy season would be predicted to be most effective in decreasing transmission due to maximal reduction in mf prevalence and loads at the precise time that vector density and biting rates are beginning to rise.
Ethical approval for this study was obtained from WHO and University of Bamako. At the time that the study was performed, Human Landing Catch was considered an ethically acceptable method of mosquito collection. The collectors in this study were adult village residents normally exposed to mosquito bites. The collectors were not given antimalarial prophylaxis, but were provided access to a health practitioner (nurse) during the study in the event of malaria infection as recommended for adult subjects living in malaria endemic area. Since the goal of HLC is to collect the mosquito before it bites, the risk of infective bite is actually quite low. Although HLC is still used in some settings, research is actively ongoing in our center and others to find a comparable method that does not involve human bait [19][20][21].

Conclusions
In conclusion, the entomological data from the present study confirmed the district of Sikasso as an area of high W. bancrofti transmission. This led to the selection of this area as the site of a multi-national study on the effects of MDA on LF transmission by anopheline vectors and as the first region in Mali for implementation of MDA with ivermectin and albendazole to eliminate transmission of LF. Comparison of the vector transmission parameters prior to and immediately following the first round of MDA demonstrated a significant decrease in overall transmission after institution of MDA. Importantly, the dramatic variability in EIR over the transmission season suggests that the efficacy of MDA can be maximized by delivering drug at the beginning of the rainy season (just prior to the peak of transmission).