Portal de Periódicos da CAPES

Laboratory and field testing of bednet traps for mosquito (Diptera: Culicidae) sampling in West Java, Indonesia

2010; Wiley; Volume: 35; Issue: 1 Linguagem: Inglês

10.1111/j.1948-7134.2010.00076.x

1948-7134

Craig A. Stoops, Yoyo R. Gionar, Saptoro Rusmiarto, Dwiko Susapto, Heri Andris, Iqbal Elyazar, Kathryn A. Barbara, Amrul Munif,

Dengue and Mosquito Control Research

Surveillance of medically important mosquitoes is critical to determine the risk of mosquito-borne disease transmission. The purpose of this research was to test self-supporting, exposure-free bednet traps to survey mosquitoes. In the laboratory we tested human-baited and unbaited CDC light trap/cot bednet (CDCBN) combinations against three types of traps: the Mbita Trap (MIBITA), a Tent Trap (TENT), and a modified Townes style Malaise trap (TSM). In the laboratory, 16 runs comparing MBITA, TSM, and TENT to the CDCBN were conducted for a total of 48 runs of the experiment using 13,600 mosquitoes. The TENT trap collected significantly more mosquitoes than the CDCBN. The CDCBN collected significantly more than the MBITA and there was no difference between the TSM and the CDCBN. Two field trials were conducted in Cibuntu, Sukabumi, West Java, Indonesia. The first test compared human-baited and unbaited CDCBN, TENT, and TSM traps during six nights over two consecutive weeks per month from January, 2007 to September, 2007 for a total of 54 trapnights. A total of 8,474 mosquitoes representing 33 species were collected using the six trapping methods. The TENT-baited trap collected significantly more mosquitoes than both the CDCBN and the TSM. The second field trial was a comparison of the baited and unbaited TENT and CDCBN traps and Human Landing Collections (HLCs). The trial was carried out from January, 2008 to May, 2008 for a total of 30 trap nights. A total of 11,923 mosquitoes were collected representing 24 species. Human Landing Collections captured significantly more mosquitoes than either the TENT or the CDCBN. The baited and unbaited TENT collected significantly more mosquitoes than the CDCBN. The TENT trap was found to be an effective, light-weight substitute for the CDC light-trap, bednet combination in the field and should be considered for use in surveys of mosquito-borne diseases such as malaria, arboviruses, and filariasis. Collecting mosquitoes during mosquito-borne disease surveillance programs is an essential task in uncovering phenology, seasonality, and population dynamics of vector species (Moorehouse and Wharton 1965, WHO 1966). Vector incrimination and planning effective anti-vector interventions are often based on the information gathered during these sampling activities (Burkett et al. 2001). The gold standard for sampling mosquitoes during disease surveillance programs is the Human Landing Collection (HLC) (Service 1993). An HLC is where a person acts as bait and collects the mosquitoes landing on an exposed leg or arm themselves, or serves only as bait and the mosquitoes are collected by another individual (Service 1993). The mosquitoes collected are then assumed to have been seeking to take a blood meal and are considered anthropophilic. If the species collected are found to be positive for the disease agent of interest, the collected species are thought to be important in pathogen transmission. This sampling method has provided the foundation for our understanding of pathogen transmission by mosquitoes for diseases such as malaria and dengue (Reid 1961). Despite the power of this sampling method, there are several disadvantages to using HLCs, including the skill necessary to collect landing mosquitoes, the need for close supervision of collectors, and the direct exposure to pathogens to which there may be no effective medicines (Trape 2001). It is also possible that HLCs also overestimate the biting rate when calculating Entomologic Infection Rates (Mathenge et al. 2002). To minimize the risk and overcome the problems associated with HLCs, an amazing number of trap types have been developed (Service 1993). Currently the most often used mechanical trap for collecting mosquitoes is the CDC miniature light trap (Davies et al. 1995). There are disadvantages to using mechanical traps. Not all vector species are attracted to these traps, mechanical traps need a power source such as dry or gel cell batteries, and diligent care must be taken for the traps to remain in a working condition. To increase the number of mosquitoes collected, the traps can be baited with a source of CO2, such as dry ice. In many regions of the world sources of CO2, such as dry ice, are not readily available. In these instances traps can be baited with animals, with the potential bias of over-emphasizing the importance of zoophilic species, or be baited with human volunteers, with a risk of exposure of the bait to the pathogen of interest. To use humans as bait for host-seeking mosquitoes but protect them from bites, nets draped over or enclosing the bait have been used in several regions of the world to collect mosquitoes (Service 1977, Govella et al. 2009, Kweka et al 2009, Odiere et al. 2007). In Southeast Asia, Gater (1935), cited in Colless (1958), was reported to have designed the Malayan human-bait trap which was used across Malaysia for over 20 years. And recently a bednet trap design, called the Mbita trap, was found to be effective for sampling malaria vectors in Africa (Mathenge et al. 2002). Service (1993) felt that as long as the bednet trap is used as a sampling device and not used as a quantitative measure of population size or mosquito density it is effective for trapping anthropophilic mosquitoes in a low-cost, low-effort manner. Other net traps such as Malaise traps and tent traps that sample the aerial populations of mosquitoes are effective at capturing mosquitoes and other medically important arthropods (Malaise 1936, Smith et al. 1965). These trap types can be baited with humans or animals to increase trap catch (Wallace et al. 1977). Because CDC light traps are not always practical, suitable alternatives need to be discovered and tested. The objective of this study was to find a self-supporting, exposure-free bednet trap that collects as many or more mosquitoes than a CDC light trap placed beside a person sleeping in an untreated bednet. The trap had to be lightweight, easy to assemble, and provide protection to the person in the bednet while trapping vectors that can be identified and tested for human pathogens. Four bednet trap designs were tested in the laboratory. The first bednet design tested was the Mbita trap (MBITA) built to specifications found in Mathenge et al. (2002). See Mathenge et al. (2002) for a figure of the Mbita trap. The second trap was a commercially available self-supporting MegaBugTM tent developed by Black Diamond (TENT) with the U.S. DoD low profile bednet inside (Figure 1A). The MegaBugTM tent is a pyramid-shaped screen-only tent (mesh size of 300 holes/cm2) held up by a single pole that is placed inside the tent that extends from the floor to the apex of the pyramid. Total dimensions of the MegaBugTM tent are 2.3 m × 2.3 m × 1.4 m. The U.S. DoD low profile bednet is 2 m long, 0.8m wide, and 0.68 m high with a mesh size of 410 holes/cm2. The net is factory treated with permethrin to a concentration of 120 mg/m2 (Frances et al. 2003). The second trap was a Townes-style malaise trap (TSM) (John W. Hock Company) with the U.S. DoD low profile self-supporting bednet placed parallel and adjacent to the malaise trap (Figure 1B). A standard CDC miniature light trap hung beside a standard U.S. military bednet cot combination (CDCBN) was used as a control (Figure 1C). The standard U.S. military bednet (2.4 m × 1.2 m × 1.5 m ≈113 holes/cm2) attached to poles 91 cm in length placed into the four corners of a folding cot. Illustrations of A) the baited tent trap (TENT) trap, B) the baited Townes style Malaise trap (TSM) and C) the baited CDC light trap hung beside a standard DoD bednet cot combination (CDCBN) tested in the laboratory and field, West Java, Indonesia, 2007 and 2008. Four screened enclosures were constructed, each with the dimensions 2.6 m × 1.25 m × 2.5 m. One trap type, baited and unbaited, was compared to a baited and unbaited CDCBN. For example, one run of the experiment was TSM-baited, TSM-unbaited, CDCBN-unbaited, CDCBN-baited. Each trap was randomly assigned to one of four enclosures and one enclosure held only one of the baited or unbaited traps. People serving as bait were rotated between the trap types to overcome potential differences in an individual's attractiveness to mosquitoes (Lindsay et al. 1993). The TENT trap was not set up to its full floor dimensions to leave space beside the tent on either side of the tent in the testing cages. After the person serving as bait inside the bednet was in place, each enclosure received 100 to 200 laboratory-reared, non-blood-fed female Anopheles (Anopheles farauti Laveran or Anopheles maculatus Theo.) mosquitoes, depending on the number of adult mosquitoes available from the NAMRU-2 mosquito colony on the date of the experiment. Each enclosure received the same number of mosquitoes. The room with the screen cages was kept in total darkness. After two hours the mosquitoes were aspirated out of the trap (recorded as collected) and out of the cages (recorded as uncollected) and placed in containers for enumeration. The number of mosquitoes collected by each trap type was compared, as was the number of individuals collected between unbaited and baited of each individual trap type. The two traps that collected the most mosquitoes over the 16 trials were considered the best performing and were brought forward for field trials. Java island is approximately 127,000, km2 with a human population of approximately 115 million (Whitten et al. 1996). The testing was done in the subdistrict of Simpenan, village of Cibuntu. The village is a collection of houses surrounded by rice fields, approximately 500 m from a beach on the Indian Ocean. Rainfall at the site ranged from 0 to 220 mm per month, with most of the rain falling between November and April. Two trap types were brought forward from the laboratory experiment for field testing: the tent trap (TENT) and the Townes-style Malaise (TSM). The CDC light trap/DoD cot bednet combination (CDCBN) was again used as a control. In a 6 × 6 Latin Square design, traps were tested once per month from January, 2007 to September, 2007 for six nights every other night over a 12-day period. Six houses were chosen, and on the first night of the experiment one trap type was randomly assigned to a house. Traps were then rotated between the six houses every other night for six nights over 12 nights with each trap being tested one night at each house. Traps were tested beginning at 18:00 and mosquitoes were collected at 06:00. The person serving as bait entered the bednet at 18:00 and remained until the mosquitoes were collected at 06:00. The people in the bednets were rotated between the bednet types to mediate potential differences in an individual's attractiveness to mosquitoes (Lindsay et al. 1993). The field experiment was run a second time with the above rotation using baited and unbaited CDCBN, baited and unbaited TENT traps, and Human Landing Collections (HLCs). Two HLCs were conducted outdoors. Each person was rotated between bednet types and HLCs to mediate potential differences in attractiveness and skill at conducting HLCs (Lindsay et al. 1993). The trial was from January, 2008 to May, 2008. Collection containers were removed from the TSM and CDCBN traps, and mosquitoes in the TENT trap were collected using a battery-powered aspirator and placed into screened collection cups. Mosquitoes in the HLCs were collected using a mouth aspirator and placed into screened collection cups. All mosquitoes in the containers were returned to the laboratory for identification using morphological characters (O'Connor and Soepanto 1989). For the second field trial, blood meal status (fed vs unfed) was recorded and specimens were dissected to determine parity (Detinova 1962). Voucher specimens are stored at the U.S. Naval Medical Research Unit 2 laboratory in Jakarta and will be made available for study in coordination with the Indonesian Ministry of Health. For the laboratory study, Fisher's exact tests were used to detect differences in trap catch among TENT, MBITA, and TSM traps and the control trap (CDCBN). In the field study, trap, day, replicate, and location effects were evaluated with Analysis of Variance for each of the six consecutive night sampling periods and nine trap evaluations. Trap sampling data were transformed to log10 (x+1) before analysis to limit heteroscadicity (Kline et al. 2006). Dunnett t-tests were used to compare TENT-baited and unbaited and TSM-baited and unbaited and HLCs to the control traps CDCBN-baited and unbaited. Back-transformed means are reported in Table 3 and Table 4. Chi-square was used to compare the number of parous, nulliparous, and blood-fed and non-blood-fed individuals collected by each trap type. Stata 9 (StataCorp LB College Station, TX) and SPSS 14 (SPSS Inc. Chicago, IL) were used for all statistical analyses. In the laboratory, 16 runs comparing MBITA, TSM, and TENT to the CDCBN were conducted for a total of 48 runs of the experiment using 13,600 mosquitoes. Table 1 provides the trap catches of each trap type compared to the CDCBN trap. The test trap that collected the most mosquitoes over the 16 trials was the TENT-baited trap (n=468) followed by the TENT-unbaited (n=263), the TSM-baited (n=170), the TSM-unbaited (n= 51), the MBITA-baited (n=35), and the MBITA-unbaited (n=1). The TENT-baited and unbaited collected a significantly greater number (P<0.01) of mosquitoes than the CDCBN-baited and CDCBN-unbaited. There was no significant difference between the baited and unbaited CDCBN and TSM traps (P=1), although the CDCBN-baited and -unbaited collected a significantly greater number of mosquitoes than the MBITA-baited and -unbaited traps (P<0.01). The first set of field tests were conducted during six nights of two consecutive weeks per month from January, 2007 to September, 2007 for a total of 54 trapnights. A total of 8,474 mosquitoes representing 33 species was collected using the six trapping methods (Table 2). Baited traps of all three types collected more mosquitoes (n=5,588) than unbaited traps (n=2,886). The TENT-baited trap collected the most (n=3,325) of all of the baited and unbaited trap combinations. The TENT-unbaited collected more (n=1,622) than the CDCBN-baited (n=1,480) and the TSM-baited (n=783). The CDCBN-unbaited collected more mosquitoes (n=840) than the TSM-baited. The trap type that collected the fewest mosquitoes was the TSM-unbaited (n=424). Daily variation in total trap catch each month did not show a significant difference (F5,318=0.25, p=0.942) for total number of mosquitoes collected during each six-day sampling period. Mosquito populations were not uniform during the nine months, and there were significant differences in the total catch between months (F8,315=18.7, p 0.05). The CDCBN-unbaited collected a significantly greater number of mosquitoes than the TSM-unbaited (4.47 vs 2.12, p=0.002). There was no significant difference among the traps for the total number of Anopheles collected. The TENT-baited collected a greater number of Culex spp. vs the CDCBN-baited (14.28 vs 6.96, p=0.007) and the CDCBN-baited collected significantly more than the TSM-baited (6.96 vs 2.45, p=0.003). The CDCBN-unbaited collected significantly more Culex spp. than the TSM-unbaited (3.49 vs 1.34, p=0.004). There were differences in the number of individual species collected by each trap (Table 3). The CDCBN-baited collected significantly more An. aconitus than both the TENT-baited (0.41 vs 0.1, p=0.001) and the TSM-baited (0.41 vs 0.03, P<0.001). The CDCBN-unbaited collected significantly more An. vagus than the TSM-unbaited (0.06 vs 0, p=0.041). The TENT-baited collected significantly more Cx. pseudovishnui (0.36 vs 0.07, p= 0.023), Cx. quinquefasciatus (0.34 vs 0.07, p <0.001), and Cx. vishnui (9.8 vs 4.49, p=0.03) than the CDCBN-baited. And the CDCBN-baited collected significantly more Cx. bitaeniorhynchus (0.27 vs 0.04, p=0.016) and Cx. vishnui (4.49 vs 1.17, p=0.002) than the TSM-baited. The TENT-unbaited collected significantly more Cx. quinquefasciatus (0.25 vs 0, p<0.001) than the CDCBN-unbaited. The CDCBN-unbaited collected significantly more Cx. gelidus (0.17 vs 0.03, p=0.012) and Cx. vishnui (2.11 vs 0.63, p=0.013) than the TSM-unbaited. The second set of field tests were conducted during six nights of two consecutive weeks per month from January, 2008 to May, 2008 for a total of 30 trapnights. A total of 11,923 mosquitoes was collected representing 24 species (Table 2). Human landing collections captured the highest number of mosquitoes (n=8,983) followed by the TENT (n=1,153) and the CDCBN (n=787). As in the laboratory and first field trial, baited traps of both trap types collected more mosquitoes (n=1,651) than unbaited traps (n=1,289). The TENT-baited trap collected the most (n=1,151) of the baited and unbaited trap combinations. The TENT-unbaited collected more (n=1,002) than the CDCBN-baited (n=500) and the CDCBN-unbaited (n=287). Culex vishuni (n=9,593) represented 88% of Culex spp. collected and 80% of all individuals collected. Most were collected by the HLCs (n=7,693) representing 85% of individuals collected using this method. The TENT trap collected more Cx. vishnui (n=1,480, 69% of all individuals collected by the TENT) than the CDCBN (n=474, 60% of all individuals collected by the CDCBN). Daily variation in total trap catch each month showed a significant difference (F4,145=50.9 p=0) for total number of mosquitoes collected during each six-day sampling period and there was a difference between the number of mosquitoes collected at each house (F 5,174= 3.16, p = 0.009). But there were no significant differences in the total catch between months (F4,145=1.19, p=0.318). Human landing collections (216.4 vs 6.96, p<0.01) and the TENT trap (18.8 vs 6.96 p<0.05) collected significantly more mosquitoes than the CDCBN (Table 4). Human landing collections collected significantly more Anopheles (p<0.01), Culex (p<0.05), and all other genera combined (p<0.05) than both the TENT and CDCBN (Table 4). Human landing collections also collected significantly more An. aconitus, Cx. bitaeniorhynchus, Cx. quinquefasciatus, Cx. sitiens, Cx. tritaeniorhynchus, Cx. vishnui, and Ae. albopictus (Table 4). The TENT collected a greater number of all Culex spp. combined (15.03 vs 4.89, p<0.01) and a greater number of Cx. vishnui (12.66 vs 3.80 p<0.01). Table 5 provides the number of blood fed, non-blood fed, parous, and nulliparous individuals collected by each method during the second field trial. Human landing collections collected a significantly greater number of blood fed and non-blood fed individuals (p 0.05). When compared individually in a separate test, the CDCBN baited collected a significantly greater number of unfed individuals than the CDCBN unbaited (p 0.05). The objective of our research was to find a self-supporting, exposure-free bednet trap that can serve as a substitute for CDC light traps. The traps were not designed to focus on sampling of a specific genus of mosquitoes but were designed for general surveillance of mosquitoes that may serve as vectors of endemic pathogens. The bednet traps tested needed to collect medically important species in the same or greater numbers than the CDC light trap bednet combination (CDCBN). The Black Diamond MegaBugTM and the DoD low-profile self-supporting bednet combination (TENT) trap met the objective. The two other bednet traps, the Mbita (MBITA) and the Townes Style Malaise (TSM) tested in the laboratory, and the TSM tested in the field did not meet the objective. Both traps collected fewer mosquitoes than the CDCBN. The MBITA trap performed poorly at collecting mosquitoes in the laboratory. This supports data reported by Mathenge et al. (2004) and Mathenge et al. (2005) from two trials in Kenya, where the MBITA trap was found to be less sensitive for Anopheles spp. and Culex spp. than CDC light traps. Laganier et al. (2003) also found the MBITA trap performed poorly when tested against HLCs in the highlands of Madgascar. Based on our data and the data reported in the literature (Laganier et al. 2003, Mathenge et al. 2005), we did not bring the MBITA trap to the field for testing. The TSM trap did better than the MBITA trap in the laboratory and collected medically important species in the field trial but was outperformed by both the TENT and the CDCBN taps. Despite the performance in our study, malaise traps should still be considered for use in surveys of medically important arthropods especially in areas where general faunal surveys are being conducted (Smith et al. 1965). Of the trap types field tested (CDCBN, TSM, and TENT), the TENT trap was the best performing for general sampling of mosquitoes. The TENT collected a greater number of individuals when compared to the CDCBN. In the second field trail as expected, the human landing collections (HLCs) collected significantly more mosquitoes than all of the trap types. Despite the overall performance of the TENT over the CDCBN, there were differences in the number of individuals collected between the TENT, CDCBN, and the HLCs for certain species. The three species collected most often in both field trials regardless of sampling method were Cx. vishnui, Cx. tritaeniorhyncus, and An. vagus. The HLCs collected a significantly greater number of Cx. vishnui and Cx. tritaeniorhyncus than both trap types, but the TENT trap collected a significantly greater number of these two species vs the CDCBN. There were no differences in the numbers collected by any of the methods in both field trials for An. vagus. The TENT trap did as well or better than the CDCBN for these three species. The HLCs collected a significantly greater number of An. aconitus, and in the first field trial, the CDCBN collected a significantly greater number of An. aconitus vs the TENT. CDC light traps have been found to be more attractive for An. aconitus in other parts of this species' range. In a comparison of four light trap types in Assam, India, Das et al. (1993) reported the CDC light trap captured a significantly greater number of An. aconitus vs other Anopheles spp. The incandescent light of the CDC light trap may provide an extra stimulus to attract this species. In our study, the TENT trap did collect An. aconitus, but at low population densities would not be as effective as HLCs and possibly CDC light traps. In the first field trial, the TENT collected a greater number of An. sundaicus (n=3) vs the CDCBN (n=1), but in the second field trial HLCs collected the only An. sundaicus (n=6). Populations of An sundaicus fluctuate greatly from season to season on Java and An. sundaicus is known to exist at low levels in the study site (Stoops et al. 2009). It is possible that the TENT and CDCBN traps are not sensitive enough to attract and capture this species unless density is high. It is important to note that Anopheles maculatus, an important malaria vector on Java, was not collected by any of the trap types or the HLCs. During an extensive larval survey of Sukabumi, no An. maculatus larvae were collected near the village where the trapping was conducted (Stoops et al. 2008). None of the sampling methods captured a greater number of parous versus nulliparous individuals so there is no bias for older or younger mosquitoes in any of the trapping methods. Even without humans acting as bait, the TENT trap may be viewed by mosquitoes as a suitable resting site following a blood meal and may act as a resting box-type trap (Service 1993) or mimic a natural resting place such as a stream bank (Chow et al. 1960). Because the number of blood fed versus non blood fed in the baited and unbaited is not significant the mosquitoes are entering the traps after taking a bloodmeal. The treated bednet inside the TENT is protecting the person acting as bait in the trap. Due to the heterogeneity of the field site in proximity to mosquito larval habitats and daytime resting locations there was a placement effect found in both field trials. The difference between houses was not unexpected and is often observed in field testing of traps (Kaufman et al. 2008). Traps, people serving as bait and HLC collectors were rotated between houses to minimize this effect (Burkett et al. 2001). For any bednet trap design, mosquito escape seems to be a major obstacle (Service 1993). With bednet traps that enclose a person sleeping in a net and a gap is left between the ground and the edge of the outer net, it seems that mosquitoes are readily able to escape (Frances et al 2003, Service 1993). The door to the TENT trap was 80% zippered closed (Figure 1). This size of entrance proved suitable for mosquito entry and to prevent escape of most individuals. Another advantage of the TENT trap is the combined weight of the TENT and DoD self-supporting bednet is only 2 kg (TENT trap weighs 1.19 kg, DoD self-supporting bednet weighs 0.9 kg). Also, no power sources (batteries) or external sources of CO2 (dry ice/compressed gas cylinders) are necessary. Mosquitoes collected in the TENT trap are alive and can be used in pathogen testing and testing for insecticide resistance and the only necessary items are an aspirator and collection cups or vials. During the trials, we timed how long it took a collector to aspirate all of the mosquitoes out of the tent and on average it took 12 min. No single sampling method or trap will work best in all situations or will be suitable for use in all studies (Service 1993). The TENT trap will not replace CDC light traps, but in areas where CDC light traps are not practical, disease transmission is high or drug resistant pathogens occur, mosquito sampling could occur while minimizing the risk to the collector (WHO 1966, Trape 2001). The person acting as bait only has to sleep inside the inner bednet, and in areas of disease transmission this person should be using a bednet with or without the addition of the outer TENT trap. If general surveillance of mosquito species in an area is the focus of a collector's efforts, the TENT trap provides a suitable alternative to CDC light traps. We thank D. Supardi, Malaria Field Coordinator for Sukabumi District, and H. B. Thahadibrata, Head of Ministry of Health for Subkabumi District, for their strong support of this project. Petty Officer Jane Nonthaveth illustrated the traps depicted in Figure 1. P.J. Obenauer, B. F. Prendergast, and A.G. Wheeler provided very helpful reviews of the manuscript. This study was supported by the U.S. Military Infectious Disease Research Program (MIDRP). The opinions or assertions expressed herein are the private views of the authors and are not to be construed as representing those of the U.S. Navy, the Department of Defense, or the Indonesian Ministry of Health.