Is the Mbita trap a reliable tool for evaluating the density of anopheline vectors in the highlands of Madagascar?

Background One method of collecting mosquitoes is to use human beings as bait. This is called human landing collection and is a reference method for evaluating mosquito density per person. The Mbita trap, described by Mathenge et al in the literature, consists of an entry-no return device whereby humans are used as bait but cannot be bitten. We compared the Mbita trap and human landing collection in field conditions to estimate mosquito density and malaria transmission. Methods Our study was carried out in the highlands of Madagascar in three traditional villages, for 28 nights distributed over six months, with a final comparison between 448 men-nights for human landing and 84 men-nights for Mbita trap, resulting in 6,881 and 85 collected mosquitoes, respectively. Results The number of mosquitoes collected was 15.4 per human-night and 1.0 per trap-night, i.e. an efficiency of 0.066 for Mbita trap vs. human landing. The number of anophelines was 10.30 per human-night and 0.55 per trap-night, i.e. an efficiency of 0.053. This efficiency was 0.10 for indoor Anopheles funestus, 0.24 for outdoor An. funestus, and 0.03 for Anopheles arabiensis. Large and unexplained variations in efficiency were observed between villages and months. Conclusion In the highlands of Madagascar with its unique, highly zoophilic malaria vectors, Mbita trap collection was poor and unreliable compared to human landing collections, which remains the reference method for evaluating mosquito density and malaria transmission. This conclusion, however, should not be extrapolated directly to other areas such as tropical Africa, where malaria vectors are consistently endophilic.


Background
Malaria transmission is ordinarily calculated as the product of the density of anopheline vectors per human and the infectivity of this anthropophilic fraction of mosquitoes. Up to now all-night stationary direct bait collections (also called human landing collections) with a human acting both as bait and collector has been the reference method for evaluating mosquito density per human. Most of the time, mosquitoes that land on human skin are collected before they have bitten but, clearly, this method exposes men to mosquito bites. Therefore, alternative methods are needed and there have been many attempts to develop new strategies and traps, with varying degrees of success [1].
Mathenge and collaborators [2] have published a complete description of a new trap design. The 'Mbita trap': -is baited by one human protected from mosquito bites; -allows the human to sleep ad libitum; -consists of a modified conical bednet made of white cotton cloth (not netting) that concentrates in its upper part the heat and various odours produced by the human bait; the apex is made of netting and forms a funnel with a small round hole (5 cm in diameter) at its base that permits the entrance of mosquitoes but impedes their escape; a netting panel is fixed halfway up the net to separate the upper mosquito chamber from the lower human chamber; -is inexpensive to produce, does not require any maintenance, and is simple to use.
Mathenge et al [2] provide evidence of its efficacy in trapping laboratory-reared Anopheles gambiae released in a screen-walled greenhouse in the Mbita Point ICIPE field station, near Victoria Lake, Kenya. When compared sideby-side with similar samples of mosquitoes, the Mbita trap caught 43.2 ± 10% of the number caught by human landing collections. Clearly, if such success were verified in the field with wild mosquitoes, this trap would become an attractive alternative for mosquito surveillance.
The aim of this study was to evaluate the success of Mbita traps in sampling mosquitoes in the field conditions of Malagasy highlands with special references to two indicators, the anopheline vector species and the anopheline density per human. In other words, we made a comparison of methods between the Mbita trap and human landing collection.
These three villages follow a general line NE-SW and are separated by 14 km for ANH-SOA and 17 km for SOA-AMG.
The area has one rainy season from November to April. Mean annual rainfall is 1,600 mm. The mean temperature between December and May is 23,9°C ranging from 21,9°C to 24,9°C. Rice fields generate a number of breeding sites for various mosquitoes including anophelines. Rice production is the main activity of villagers. In this region, most people (>99%) do not use bednets and zebus are kept within the village at night.
In the twentieth century, the central highlands of Madagascar have experienced large malaria outbreaks. A national programme for preventing malaria epidemics, with CAID ("Campagne d'Aspersion Intra-Domiciliaire" of insecticide) performing DDT spraying of house walls at 2 g/m 2 . The houses in the study area are normally covered by this treatment, but the last insecticide treatment was carried out pre-1998 in AMG, in 2000 in SOA and 2001 in ANH, i.e. >60, 36 and 24 months respectively before the beginning of this study.
The study protocol was approved by the Ministry of Health of Madagascar.

Human landing collections
Adult male volunteers were placed in a room ordinarily used as a bedroom or out-of-doors in places protected from the rain. According to WHO recommendations [3], mosquitoes were collected with glass tubes closed by cotton plugs as they landed on the exposed lower legs of adult humans (Fig. 3). Malaria prophylaxis was offered. In each village, collections were performed monthly for two consecutive nights from 18.00 to 06.00. Each night, four houses were used and, for each house, two men were sited indoors and two outdoors, working in six-hours shifts. The total number of men per night was 32 divided in two teams of 16.

Mbita trap collections
Mbita traps were provided by Dr. Mathenge at cost price (10 US dollars each). They were used as described [2] and baited with a man resting in bed and in the trap for 12 hours from 18.00 to 06.00 (Fig. 4). In each village, in parallel to human landing collections, three traps were used per night, with one outdoors, and two indoors in separate bedrooms without people other than this under the trap in order to avoid local competition between the trap and other more accessible people for mosquitoes. Bedrooms chosen for Mbita trap collections were used one single night each month (i.e. 4 different bedrooms per village and per month). At 06.00, when the human bait left the trap, an experienced technician collected the mosquitoes with an aspirator.

Mosquito procedures and data analysis
Mosquito species were assessed using morphological characteristics. For An. gambiae s.l., a sample of 50 females per village and per month was tested by PCR [4] (this sample was obtained by human landing, pyrethrum spray, and artificial pit shelter collections). As only Anophelesarabiensis was observed in a sample of over 1,100, hereafter, any An. gambiae s.l. were assigned to An. arabiensis. Ovaries of anopheline vectors were examined for parity using the Detinova technique [5]. The origin of the blood meal of anophelines found fed in traps was assessed by ELISA [6]).
The number of mosquitoes caught by each method was recorded. By definition, one human-night referred to the unit of human landing collections i.e. the activity of mosquitoes on one human during the whole night. A trapnight referred similarly to the activity of one trap during the whole night. The efficiency of the Mbita trap (∆) is defined as the number of mosquitoes collected per trapnight divided by the number of mosquitoes collected per human-night in similar conditions of location (indoors and/or outdoors) and time (nights of observation). A positive correlation was also searched for An. funestus samples between Mbita trap and human landing collections using the Pearson's coefficient correlation.

Results
The whole data set consists of 6,899 mosquitoes for human landing and 85 for Mbita trap collections. Mosquitoes landing on humans belonged to 26 mosquito species (10 Anophelinae and 16 Culicinae) and those collected with Mbita traps to eight species (three Anophelinae and five Culicinae) (see Additional File 1 for the complete data used to performed this analysis).
Mosquito species with less than five specimens in human landing collections were excluded from the analysis (i.e. a total of 18 mosquitoes with 2 Anophelinae and 16 Culicinae, all human landing, that represented 0,26% of the whole data set) and results presented hereafter concern 6,881 and 85 mosquitoes, respectively, belonging to 17 species (Table 1) to perform this analysis are in Tabl. 2). ∆ was 0.036 in AMG, 0.963 in SOA and 1.212 in ANH (χ 2 = 165.7, df = 2, p < 10 -4 ) with values that varied inversely with the density of this species in human landing collections. ∆ was also highly variable between months and ranged from 0 to 6.8 (maximum in February, outdoors, SOA) without clear tendencies that would provide clues to explain this variation.
Beside this analysis of efficiency, a correlation was searched for An. funestus samples between Mbita trap and human landing collections. No statistically-significant positive correlation either for the indoor or the outdoor samples was evidenced (indoor, Pearson's coefficient correlation r = -0.21, n = 14, p = 0.47; outdoor, r = 0.20, n = 14, p = 0,50). Another similar analysis using log-transformed values (+1) did not modify the conclusions.
The two An. gambiae s.l. collected in Mbita traps were from indoor trap at AMG on January. Both were nulliparous and unfed and were identified as An. arabiensis. Among the 43 An. funestus collected in the Mbita trap, one was collected fully fed in an indoor trap and had taken its blood meal from zebu. Eighteen An. funestus were examined for ovaries, 12 were parous and six were nulliparous, Indoor landing collection of mosquitoes Figure 3 Indoor landing collection of mosquitoes. The man who acts as baits actively collects mosquitoes that land on his legs.
i.e. with an excess of nullipars relative to those sampled by human landing collections (85% of parous among 1,512 mosquitoes, collected either indoors or outdoors without difference in the parity rate) (p = 0.04 by exact Fisher's test).

Discussion
The efficiency of the Mbita trap compared to human landing collections is very poor for all species of mosquitoes (with the possible exception of An. funestus which will be discussed below). This low efficiency observed in the highlands of Madagascar with wild mosquitoes is in complete contradiction with previous published results [2] obtained in semi-field conditions using laboratory reared An. gambiae and in field conditions of rural Kenya [7]. We hypothesise the main reason for these discrepancies resides in the well known zoophilic/exophilic trophic preferences of Malagasy mosquitoes [8]. During the study, the antropophilic rate for An. arabiensis was 0.00 for those collected indoors by pyrethrum spray collections (only 12 fed females tested) and 0.016 for those outdoors resting in pit shelters (318 tested, unpublished data). This zoophilic/exophagic behaviour may be antagonist to the entry in the trap that is thought as a positive response to convective heat currents and various odours produced by the human bait in the trap.
The efficiency of the Mbita trap seamed less poor for An. funestus. In some cases, a trap efficiency was observed which was higher than that of human landing collections. Figure 4 View of an outdoor Mbita trap. The man who acts as bait is sleeping and protected from mosquito bites. By design, mosquitoes enter in the trap via a funnel-shaped entry-no-return port at the bottom of the trap.