Field Evaluation Of A Lethal Ovitrap For The Control Of Aedes Aegypti

VECTOR CONTROL, PEST MANAGEMENT, RESISTANCE, REPELLENTS

Field Evaluation of a Lethal Ovitrap for the Control of Aedes aegypti (Diptera: Culicidae) in Thailand RATANA SITHIPRASASNA, PRADITH MAHAPIBUL, CHUMNONG NOIGAMOL, MICHAEL J. PERICH,1 BRIAN C. ZEICHNER,2 BOB BURGE,3 SARAH L. W. NORRIS,4 JAMES W. JONES,5 SONYA S. SCHLEICH,3 AND RUSSELL E. COLEMAN Department of Entomology, U.S. Army Medical Component, Armed Forces Research Institute of Medical Sciences, Bangkok, Thailand

J. Med. Entomol. 40(4): 455Ð462 (2003)

ABSTRACT In 1999 and 2000 we evaluated a lethal ovitrap (LO) for the control of Aedes aegypti (L.) in three villages in Ratchaburi Province, Thailand. Two blocks of 50 houses (a minimum of 250 m apart) served as treatment and control sites in each village, with each house in the treatment area receiving 10 LOs. Thirty houses in the center of each treatment and control block were selected as sampling sites, with larval and adult mosquito sampling initiated when LOs were placed. Sampling was conducted weekly in 10 of the 30 houses at each site, with each block of 10 houses sampled every third week. Sampling continued for 30 wk. EfÞcacy of the LO was evaluated by determining number of containers with larvae and/or pupae per house and number of adult mosquitoes collected inside each house. In 1999, the LO had a negligible impact on all measures of Ae. aegypti abundance that were assessed; however, fungal contamination of insecticide-impregnated strips may have been responsible for the low efÞcacy. In 2000, signiÞcant suppression was achieved based on changes in multiple entomologic criteria (containers with larvae, containers with pupae, and number of adult Ae. aegypti); however, control was not absolute and neither immature nor adult Ae. aegypti were ever eliminated completely. We conclude that the LO can reduce adult Ae. aegypti populations in Thailand; however, efÞcacy of the LO is lower than desired due primarily to the high number of alternative oviposition sites. LO efÞcacy may be improved when used as part of an integrated control program that places emphasis on reduction of adjacent larval habitats. Further studies are required to assess this issue. KEY WORDS Aedes mosquitoes, surveillance, control, dengue, lethal ovitrap

DENGUE AND DENGUE HEMORRHAGIC fever (DHF) occur throughout the tropical and semitropical regions of the world (Gubler and Kuno 1997). Aedes aegypti (L.), the primary vector of dengue viruses, is an urban mosquito that relies on artiÞcial containers such as ßower-pots, small cisterns, discarded tires and cans as breeding sites (Christopher 1960, Kittayapong and Strickman 1993). Aedes aegypti primarily feeds on humans and frequently rests in secluded locations inside homes where conventional insecticide spraying is minimally effective (Perich et al. 1990, 2000, Tonn et al. 1969). Sustained control of Ae. aegypti requires a combination of source reduction and insecticide treatment (Gratz 1991, 1993, PAHO 1994). However, not The views of the authors do not purport to represent the position of the Department of the Army or the Department of Defense. 1 Louisiana State University, Baton Rouge, LA. 2 US Army Center for Health Promotion and Preventive Medicine, Aberdeen Proving Ground, MD. 3 Walter Reed Army Institute of Research, Silver Spring, MD. 4 U.S. Army Medical Research Institute of Infectious Diseases, Fort Detrick, MD. 5 E-mail: [email protected].

all breeding sites can be totally eliminated or made mosquito-proof, and it is difÞcult to involve all members in the community in a sustained clean-up campaign. In addition, neither adulticides nor larvicides are completely effective against Ae. aegypti. The development of novel, effective methods for the control of dengue vectors are therefore urgently needed, with particular emphasis on methods that are environmentally benign, cost-effective, and suitable for integration into community-based control programs (Chunsuttiwat and Wasakarawa 1994, Service 1992, Swaddiwudhipong et al. 1992). Fay and Perry (1965) and Fay and Eliason (1966) Þrst reported the development of an ovitrap that could be used as a tool to conduct Ae. aegypti surveillance. The ovitrap has subsequently been used throughout the world (Chadee et al. 1995, Kaul and Geevarghese 1979, Mogi et al. 1988) and is particularly effective in areas with low vector populations (Service 1993). Ovitraps have also proven useful for surveillance of other container-inhabiting mosquitoes, such as Ae. albopictus (Skuse) (Bellini et al. 1996, Mogi et al. 1988).

456

JOURNAL OF MEDICAL ENTOMOLOGY

The ovitrap was Þrst used to control Ae. aegypti in 1969 in Singapore (Chan 1973). Chan et al. (1977) subsequently modiÞed the ovitrap by incorporating an autocidal screenÑthis modiÞed ovitrap attracted more Ae. aegypti than did other domestic container habitats in Þeld tests. Studies by Lok et al. (1977) and Cheng et al. (1982) provided further evidence that ovitraps could be used for the control of dengue vectors. Zeichner and Perich (1999) modiÞed the ovitrap by incorporating an insecticide-impregnated oviposition strip into the design. This lethal ovitrap (LO) was evaluated against a laboratory strain of Ae. aegypti (Zeichner and Perich 1999) and against Þeld populations of Ae. aegypti in Brazil (Perich et al. 2003). The study in Brazil demonstrated that the LO could reduce both sub-adult and adult populations of Ae. aegypti. The goal of the current study was to evaluate the efÞcacy of the LO in controlling immature and adult Ae. aegypti populations in Thailand. Materials and Methods Study Site. Field tests were conducted in three villages (designated Vil-1, Vil-2, and Vil-3) in Chom Bung District, Ratchaburi Province, Thailand. Tests were conducted from 7 April to 29 October in 1999 and from 24 May to 15 December in 2000. Study Design. A split-plot design was used, with village as the block factor, area (treated versus control groups) as the whole-plot factor, and sampling schedule of houses as the split-plot factor. Within each of the three villages, two groups of 50 houses each were selected as study sites. The two groups of houses were separated by a minimum of 250 m. One group of houses was randomly assigned as the treated area with the other group serving as an untreated control. All 50 houses in the treated group received 10 LOs each, with Þve LOs placed inside the house and Þve outside. Ovitraps were placed in locations where they were unlikely to be disturbed (i.e., behind furniture, under tables, or next to walls). Thirty houses in the center of each group of 50 houses were selected as sampling sites. The 20 houses in each group that were not sampled created a buffer around the sampling siteÑthis buffer was intended to limit movement of Ae. aegypti into the treated block of houses from surrounding areas. Sampling was conducted weekly in 10 of the 30 houses (one-third of the houses in each group were sampled each week, with all 30 houses sampled once during a 3-wk period). Ovitrap Design. The LO (U.S. patent number 5,983,557, 11 November 1999) used was the same black polyethylene cup as that described by Zeichner and Perich (1999) and Perich et al. (2003). A red-velour paper strip (11 cm ⫻ 2.5 cm) was treated with deltamethrin (1.0 mg active ingredient per strip) and attached to the cup using a paper clip (Zeichner and Perich 1999). This strip served as the oviposition substrate. The 10 LOs in each treated house were Þlled within 2 cm of the top with a 10% hay infusion-water solution as described by Reiter et al. (1991). Use of the hay infusion-water solution was intended to enhance

Vol. 40, no. 4

the attractiveness of the LOs. Ovitraps were checked weekly, and 10% hay infusion-water added as needed and deltamethrin-impregnated strips replaced if found missing. Treated strips were replaced with freshly treated strips monthly for the Þrst 8 wk of the study in 1999; however, strips left in the ovitraps for 1-mo developed a fungal growth that physically prevented female mosquitoes from directly contacting the strips. To prevent this problem, treated strips were replaced bi-weekly after the Þrst 8 wk of the study. Surveillance Before Placement of Lethal Ovitraps. In 1999, a survey of larval containers was conducted from 25Ð28 January, 1Ð4 February, and 8Ð11 February in Vil-1, Vil-2, and Vil-3, respectively. In 2000, preliminary surveys were conducted on 27Ð31 March, 5Ð13 April, and 13Ð27 April. The types of containers, total number of containers, number of containers with water, and number of containers with larvae or pupae were determined in 30 houses in each control and treated block in each of the three villages. The houses sampled were the same houses used in the actual evaluation of the LO. Postplacement Surveillance. We conducted weekly assessment of larval and adult mosquito populations in the control and treated blocks of houses in each of the three villages to evaluate the efÞcacy of the LO. Each week, we determined (1) the total number of containers (excluding LOs) in and around each house, (2) the number of these containers with water, (3) the total number of these containers with immature (larvae and/or pupae) mosquitoes, and (4) the number of adult mosquitoes in each house. Adult mosquitoes were collected for 10 min within each house using hand-held aspirators. Mosquitoes were transported to the Armed Forces Institute of Medical Sciences laboratory in Bangkok, sorted by species and sex, and enumerated. Statistics. For the preplacement surveillance, means of entomologic variables (number of water containers, number of water containers with water, and number of water containers with immature mosquitoes) were compared for the treatment group (treated versus control) across villages by analysis of variance (ANOVA). The Tukey multiple comparison procedure (PROC ANOVA; SAS Institute 2001) was used to separate signiÞcantly different means. For the postplacement surveillance, the following procedures were used. During the course of the study, each house was sampled 10 times over the 30-wk period. For each house by year, weekly counts for each variable of interest were summed over the study period. Means and standard deviations for each treatment group, village and year were then calculated based on these 30-wk cumulative totals. All analyses were performed on these 30-wk totals for each house instead of the weekly totals for each house. Frequency distributions for numbers of containers per house, containers with larvae or pupae, and adult mosquito counts were positively skewed. LOG10 transformations were applied. After transformation, all variables except counts for adult Ae. aegypti met assumptions of normality and homogeneity of variance. A mixed model ANOVA for split plot experiments (Littell et al. 1996) was used to evaluate the main and

July 2003

SITHIPRASASNA ET AL.: FIELD EVALUATION OF A LETHAL OVITRAP

457

Table 1. Preliminary survey of water containers in treated and control blocks of three villages in Ratchaburi Province, Thailand. A total of 30 houses each were checked for total water containers, containers with water, and containers with larvae in each control and treated block in each village Block

Village 1 Jan 1999

Village 2 Mar 2000

Feb 1999

Village 3 Apr 2000

Feb 1999

Apr 2000

Control Treated

24.0 (1.61) 24.2 (2.29)

Mean number of water containers per house (⫾SEM) 29.9 (2.48) 23.2 (2.26) 18.6 (1.84) 22.9 (1.94) 26.7 (1.94) 23.0 (2.30) 20.8 (2.14) 26.3 (2.73)

Control Treated

16.4 (1.44) 16.0 (1.50)

Mean number of water containers per house with water (⫾SEM) 19.9 (1.91) 13.7 (1.30) 10.6 (0.99) 17.0 (1.50) 18.2 (1.76) 13.0 (1.36) 10.5 (1.04) 13.7 (1.39)

20.0 (1.40) 16.9 (1.68)

Control Treated

3.6 (0.54) 3.5 (0.41)

Mean number of water containers per house with larvae (⫾SEM) 0.5 (0.21) 3.7 (0.73) 0.4 (0.18) 4.2 (0.70) 3.0 (0.66) 0.9 (0.23) 3.0 (0.52) 2.1 (0.36)a

1.0 (0.26) 1.1 (0.42)

24.7 (1.79) 24.2 (2.93)

a Values in the treated block are signiÞcantly different (ANOVA, with TukeyÕs mean separation procedure, P ⬍ 0.05) than corresponding values in the control block.

interaction effects of treatment variables in the study. Treatment group (control versus test) and weekly sampling sequence of houses were treated as Þxed effects, while village was treated as a random effect. Interactions between treatment group and sampling sequence were also included in the model. Data for each year were analyzed separately. In examining the datasets for each village there were obvious differences in the size of mosquito populations between treatment groups across villages. Therefore, subsequent analyses were conducted to evaluate effects of LOs on entomologic measures for the treatment group and sampling sequence within each village separately. For variables that failed to normalize after transformation (i.e., adult Ae. Aegypti counts), nonparametric Wilcoxon rank-sum tests (PROC NPAR1WAY; SAS Institute 2001) were used to compare treatment groups overall each year as well as treatment groups within each village each year. Unless stated otherwise, all comparisons were made at the P ⬍ 0.05 level of signiÞcance. Results Surveillance Before Placement of Lethal Ovitraps. Pretreatment surveillance focused on assessing the type and number of water containers found in each village and whether the various containers contained immature mosquitoes. A variety of water containers were identiÞed during the preliminary survey, including water jars, ant traps, ßower pots, wash basins, cisterns, tires, coconut husks, empty glass bottles, and tin cans. Twenty-four percent (1,045/4,353) and 76% (3,308/4,353) of the containers were found outdoors and indoors, respectively. Almost 45% (1,946/4,353) of all water containers were earthen water jars of various sizes. Only 18% (185/1,045) of the water containers found indoors were water jars, whereas 53% (1,761/3,308) of the outdoor water containers were water jars. Data from the preliminary survey is presented in Table 1. Except for a higher mean number of water containers per house with larvae in the treated block of Village-1 (Vil-1) in 2000, there were no signiÞcant differences between control and treated houses in any

of the three villages for any of the criteria assessed in either 1999 or 2000. Postplacement Surveillance, 1999. Figure 1 presents the overall effects of the lethal ovitrap on immature mosquito populations, while the effects of the lethal ovitrap on adult mosquito populations is presented in Fig. 2. The data in Figs. 1 and 2 represent mean values recorded over each 3-wk period. In general, fewer water containers with larvae were found in treated blocks than in control blocks; however, absolute (100%) control was never achieved. The LO had less of an effect on adult mosquitoes, with a clear, sustained reduction in Ae. aegypti populations only observed in Vil-1 in 1999. In 1999, there was no signiÞcant difference in the total number of water containers or the number of containers with water found in the treated blocks compared with the control blocks (Table 2). Although the mean number of containers with larvae and mean number of containers with pupae overall were both lower in the treatment blocks than the control blocks, these differences were not signiÞcant. Within villages, only in Vil-1 was a signiÞcantly lower (F11.56; df ⫽ 1,54; P ⫽ 0.0013) number of containers with larvae found (Tables 2 and 3) in the treated block (19.6 ⫾ SE ⫽ 14.9) compared with the control block (30.6 ⫾ 16.7) (Table 2). In Vil-1 a significant (P ⫽ 0.0006) difference in the number of containers with pupae (Tables 2 and 3) was observed also, with the treatment block having a lower mean number of containers (6.9 ⫾ 7.6) compared with the control block (13.1 ⫾ 9.1). Overall, the mean number of adult Ae. aegypti mosquitoes in 1999 was signiÞcantly lower (Z ⫽ 3.09, P ⫽ 0.0020) in the treatment blocks (21.1 ⫾ 29.0) than in the control blocks (39.2 ⫾ 42.4) (Table 2). Although the majority of mosquitoes were male, similar effects were observed with both male and female mosquitoes. Within villages, there was a signiÞcant difference between treatment groups in Vil-1 (Z ⫽ 4.6l8, P ⬍ 0.0001), with the treated block having a lower mean number of mosquitoes (26.4 ⫾ 18.8) than the control block (62.2 ⫾ 29.7). However, differences in the number of adult Ae. aegypti found in the treated blocks of

458

JOURNAL OF MEDICAL ENTOMOLOGY

Vol. 40, no. 4

Fig. 1. Effect of the lethal ovitrap on larval container-inhabiting mosquito populations in three villages in Chom Bung District, Ratchaburi Province, Thailand.

Vil-2 or Vil-3 compared with the control blocks were not signiÞcant (Table 2). In 1999, across all villages there were no signiÞcant differences in the overall number of non-Aedes mosquitoes found in the treated blocks compared with the control blocks (Tables 2 and 3). Within villages, there was a signiÞcantly higher (F ⫽ 4.58; df ⫽ 1,54; P ⫽ 0.0368) mean total number of non-Aedes mosquitoes collected per house in the treated block (65.4 ⫾ 57.9) than in the control block (39.4 ⫾ 29.3) of houses for Vil-1. Postplacement Surveillance, 2000. There were no signiÞcant differences in 2000 between treated and control blocks in the overall numbers of water containers or containers with water (Table 4). SigniÞcant differences were found between treatment groups for the number of containers with larvae (F ⫽ 63.23; df ⫽ 1,174; P ⬍ 0.0001), as well as in number of containers with pupae (F ⫽ 22.26; df ⫽ 1,4; P ⫽ 0.0092) (Tables 4 and 5). SigniÞcant differences for these variables were also seen within each village (Tables 4 and 5). Overall and in each village, the mean number of containers with larvae and the mean number of containers with pupae were lower in the treated blocks compared with the control blocks.

The mean number of total adult (male and female) Ae. aegypti mosquitoes was signiÞcantly lower (Z ⫽ 4.55, P ⬍ 0.0001) in the treatment group (12.4 ⫾ 18.5) than in the control group (23.4 ⫾ 24.1) (Table 4). Within each village, treated groups of houses had signiÞcantly lower mean numbers of adult Ae. aegypti mosquitoes than control groups. In Vil-1, the treated group had a signiÞcantly lower (Z ⫽ 2.02, P ⫽ 0.0435) mean number of mosquitoes (19.3 ⫾ 23. 9) than the control group (25. 7 ⫾ 18.9). In Vil-2, the mean number of mosquitoes was also signiÞcantly lower (Z ⫽ 3.59, P ⫽ 0.0003) in the treatment group (4.0 ⫾ 4.5) than the control group (13.8 ⫾ 12.9), as was the case in Vil-3 (Z ⫽ 2.64, P ⫽ 0.0082) (13.9 ⫾ 18.3 in the treated group compared with 30.7 ⫾ 33.2 in the control group). The LO had less of an effect on adult female Ae. aegypti than on all adults or on male adults (Table 4). Although there were fewer adult females observed in the treated blocks in each of the three villages, these differences were only signiÞcant in Vil-2. No signiÞcant difference was found between treatment groups in the overall number of other (nonAedes) mosquitoes (Table 4). However, within each village there were signiÞcant differences between

July 2003

SITHIPRASASNA ET AL.: FIELD EVALUATION OF A LETHAL OVITRAP

459

Fig. 2. Effect of the lethal ovitrap on the number of adult Ae. aegypti collected in houses in three villages in Chom Bung District, Ratchaburi Province, Thailand.

treated and control blocks of houses. In Vil-1, the mean number of other species of mosquitoes was signiÞcantly lower (F ⫽ 68.87; df ⫽ 1,54; P ⬍ 0.0001) in the treated group (62.07 ⫾ 26.68) than in the control group (188.0 ⫾ 90.3) (Table 4 and 5). In Vil-3, the Table 2.

Effect of the lethal ovitrap on entomologic indices of mosquito abundance in 1999

Variable Containers with water Containers with larvae Containers with pupae Total adult mosquitoes Non-Aedes mosquitoes Total adult Ae. aegypti Adult female Ae. aegypti Adult male Ae. aegypti

a

mean number of other non-Aedes mosquitoes was also lower (F ⫽ 37.90; df ⫽ 1,54; P ⬍ 0.0001) in the treated group (96.1 ⫾ 50.0) than in the control group (189.9 ⫾ 85.1); however, in Vil-2 the mean number of other mosquitoes was higher (F ⫽ 144.79, df ⫽ 1,54; P ⬍

Group Control Treated Control Treated Control Treated Control Treated Control Treated Control Treated Control Treated Control Treated

Mean values (⫾SD) per house recorded overall and in each village Overall

Village 1

Village 2

Village 3

162.80 (74.60) 155.46 (84.96) 23.62 (16.37) 17.13 (12.73) 9.31 (8.29) 6.56 (7.06) 131.31 (80.42) 137.26 (100.29) 92.10 (83.60) 116.12 (100.78) 39.16 (42.42) 21.11 (29.00)a 7.72 (9.10) 5.17 (9.85)a 31.43 (35.20) 15.94 (20.61)a

175.27 (72.31) 185.47 (94.23) 30.63 (16.67) 19.63 (14.91)a 13.07 (9.10) 6.93 (7.59)a 101.57 (37.14) 91.90 (63.20) 39.37 (29.28) 65.43 (57.89)a 62.20 (29.67) 26.43 (18.75)a 9.90 (6.79) 6.27 (7.07)a 52.30 (26.08) 20.17 (15.04)a

138.17 (73.15) 128.20 (70.95) 17.60 (14.48) 12.77 (11.03) 5.60 (5.22) 5.03 (5.72) 161.57 (84.62) 197.80 (115.54) 142.67 (92.44) 188.70 (122.08) 18.90 (30.40) 9.07 (12.45) 4.27 (7.42) 1.90 (3.01) 14.63 (24.17) 7.17 (10.08)

174.97 (74.55) 152.70 (80.81) 22.63 (15.67) 19.00 (11.11) 9.27 (8.48) 7.70 (7.68) 130.80 (97.18) 122.07 (85.83) 94.27 (80.73) 94.23 (65.65) 36.37 (52.31) 27.83 (43.00) 9.00 (11.59) 7.33 (14.89) 27.37 (42.18) 20.50 (29.25)

Treated values are signiÞcantly different (P ⬍ 0.05) from control values.

460

JOURNAL OF MEDICAL ENTOMOLOGY

Table 3. 1999

Vol. 40, no. 4

Mixed model ANOVA evaluation of the effect of the lethal ovitrap on three entomologic indices of mosquito abundance in

Overall

Source

df

F

Village 1 df

P

F

Village 2 P

df

F

Village 3 P

df

F

P

0.1943 0.0133 0.9622

1,54 2,54 2,54

0.61 0.62 4.96

0.4373 0.5417 0.0105

Treatment House sequence Treatment*House sequence

1,2 2,170 2,170

5.79 1.89 1.73

0.1378 0.1546 0.1808

1,54 2,54 2,54

Containers with larvae 11.56 0.0013 1,54 1.73 0.06 0.9406 2,54 4.65 2.48 0.0932 2,54 0.04

Treatment House sequence Treatment*House sequence

1,2 1,170 2,170

2.39 3.51 1.09

0.2622 0.0321 0.3385

1,54 2,54 2,54

Containers with pupae 13.45 0.0006 1,54 0.33 0.25 0.7783 2,54 8.33 2.00 0.1459 2,54 0.17

0.5695 0.0007 0.8403

1,54 2,54 2,54

0.34 0.14 4.50

0.5642 0.8739 0.0156

Treatment House sequence Treatment*House sequence

1,2 2,170 2,170

4.53 15.85 0.38

0.1670 0.0001 0.6832

1,54 2,54 2,54

Adult (non-Aedes) mosquitoes 4.58 0.0368 1,54 3.88 0.44 0.6438 2,54 80.44 0.02 0.9848 2,54 3.30

0.0539 0.0001 0.0445

1,54 2,54 2,54

0.10 1.75 0.00

0.7532 0.1827 0.9982

0.0001) in the treated group (608.4 ⫾ 344.0) than in the control group (113.1 ⫾ 68.8). Discussion Results from our study in three villages in western Thailand suggest that the LO with an oviposition strip treated with 1.0 mg of deltamethrin reduced natural populations of adult Ae. aegypti. Perich et al. (2003) reported that use of the LO in Brazil had a signiÞcant impact on Ae. aegypti populations as measured by positive container index, mean number of pupae/ house, and adult mosquito aspiration collections. Perich et al. (2003) suggested that the LO could aid in the control of Ae. aegypti when integrated into community participation programs that target removal of containers that serve as breeding sites for this mosquito. This conclusion was supported by the fact that a longer treatment period was required to impact on vector populations (positive container index and mean number of pupae/house) in the village with the greatest number of containers. These “competing” containers presumably reduced the effectiveness of Table 4.

Effect of the lethal ovitrap on measures of mosquito abundance in 2000

Variable Containers with water Containers with larvae Containers with pupae Total adult mosquitoes Non-Aedes mosquitoes Total adult Ae. aegypti Adult female Ae. aegypti Adult male Ae. aegypti

a

the lethal ovitrap by diverting gravid females from visiting LOs. Perich et al. (2003) suggested the LO was not designed to be a stand-alone devise for the control of Ae. aegypti, but rather would be most effective when incorporated into an integrated control program. Data from our pretreatment surveillance demonstrated that there were no signiÞcant differences (except for a signiÞcantly higher number of water containers with larvae in the treated block of Vil-1 than in the control block) between control and treated blocks of houses before placement of the LO. Although we did not assess adult mosquito populations during the pretreatment surveillance, these data suggest that the control and treated blocks of the study sites were comparable. In addition, at no point during either year (1999 or 2000) of the study were there signiÞcant differences in the number of water containers in the treated or control blocks of any of the villages. During 1999, the LO had a minimal impact on containers with water, containers with larvae, containers with pupae, total adult mosquito populations, and adult populations of non-Aedes mosquitoes, with signiÞcant differences noted only in Vil-1 and not in Vil-2

Group Control Treated Control Treated Control Treated Control Treated Control Treated Control Treated Control Treated Control Treated

SigniÞcantly different (P ⱕ 0.05).

Mean values (⫾SD) per house recorded overall and in each village Overall

Village 1

Village 2

Village 3

168.14 (73.49) 168.24 (79.18) 16.76 (7.58) 8.47 (5.47)a 5.38 (3.88) 2.37 (2.26)a 187.31 (93.05) 266.37 (306.66) 163.64 (88.60) 255.54 (320.57) 23.38 (24.09) 12.39 (18.48)a 6.62 (7.77) 4.37 (6.88)a 16.76 (17.80) 8.02 (12.41)a

186.00 (74.49) 197.70 (92.05) 17.73 (7.45) 9.00 (5.56)a 5.97 (3.34) 3.03 (2.57)a 213.83 (81.93) 82.07 (25.42)a 187.97 (90.27) 62.07 (26.68)a 25.67 (18.86) 19.27 (23.89)a 6.83 (6.58) 6.33 (8.36) 18.83 (15.10) 12.93 (16.54)a

126.57 (65.86) 139.50 (61.28) 14.77 (6.32) 8.70 (5.58)a 4.43 (2.79) 2.37 (2.14)a 127.10 (70.49) 606.73 (323.99)a 113.10 (68.78) 608.43 (344.01)a 13.80 (12.89) 4.00 (4.49)a 4.57 (5.19) 1.57 (2.05)a 9.23 (9.15) 2.43 (2.93)a

191.87 (63.14) 167.53 (72.52) 17.77 (8.64) 7.70 (5.36)a 5.73 (5.08) 1.70 (1.88)a 221.00 (96.12) 110.30 (53.04)a 189.87 (85.13) 96.13 (49.96)a 30.67 (33.23) 13.90 (18.26)a 8.47 (10.35) 5.20 (7.63) 22.20 (23.82) 8.70 (11.54)a

July 2003 Table 5.

SITHIPRASASNA ET AL.: FIELD EVALUATION OF A LETHAL OVITRAP

461

Mixed model ANOVA evaluation of the effect of the lethal ovitrap on three key indices of mosquito abundance in 2000

Variable and Effect

Overall F

df

Village 1 P

df

F

Village 2 P

F

df

Village 3 P

F

df

P

Treatment House sequence Treatment*House sequence

1,174 2,174 2,174

63.23 ⬍0.0001 0.76 0.4708 0.33 0.7217

1,54 2,54 2,54

Containers with larvae 22.71 ⬍0.0001 1,54 13.16 1.58 0.2150 2,54 0.23 0.24 0.7888 2,54 0.07

Treatment House sequence Treatment*House sequence

1,4 2,170 2,170

22.26 1.37 0.85

0.0092 0.2564 0.4282

1,54 2,54 2,54

14.58 0.71 0.30

Containers with pupae 0.0003 1,54 11.52 0.4976 2,54 0.43 0.7387 2,54 0.76

0.0013 0.6549 0.4740

1,54 2,54 2,54

27.48 2.98 0.15

0.0001 0.0594 0.8638

Treatment House sequence Treatment*House sequence

1,4 2,170 2,170

0.01 1.59 0.17

0.94467 0.2068 0.8480

1,54 2,54 2,54

68.77 3.34 0.26

Non-Aedes mosquitoes ⬍0.0001 1,54 144.79 0.0430 2,54 4.43 0.7751 2,54 0.25

⬍0.0001 0.0165 0.7761

1,54 2,54 2,54

37.90 6.88 1.17

⬍0.0001 0.0022 0.3191

or Vil-3. There was a signiÞcant difference in the overall number of adult Ae. aegypti collected in treated versus control blocks. Although fewer adults were collected in the treated blocks of all three villages, this difference was only signiÞcant in Vil-1. During 2000, there were signiÞcant reductions in containers with larvae, containers with pupae and in total populations of all adult Ae. aegypti. However, when sex of the adult Ae. aegypti was evaluated, the LO had less of an effect on female mosquitoes than on male mosquitoes, with signiÞcant reductions in female populations only noted in Vil-2 and not in Vil-1 or Vil-3. In contrast to the general reduction in adult Ae. aegypti populations, the LO appeared to have no overall impact on the number of total adult mosquitoes collected or in the number of non-Aedes mosquitoes. Although we would not expect that the LO would reduce populations of non-Aedes mosquitoes, the fact that these populations were not affected whereas Ae. aegypti populations were reduced further supports the argument that the LO was speciÞcally targeting Ae. aegypti. Although there were signiÞcant differences in numbers of non-Aedes mosquitoes in each village, in some villages there was an increase in these mosquito populations while in others there was a decrease. Although we are not certain why the performance of the LO was signiÞcantly lower in 1999 than in 2000, we noted fungal growth on the insecticide impregnated strips during the Þrst 8 wk of the study in 1999. To eliminate this fungal growth, we replaced the insecticide-impregnated strips bi-weekly rather than monthly, beginning 8 wk after the study commenced in 1999. This fungal growth may have affected the efÞcacy of the LO in 1999. We did not experience this problem in 2000. Although we found that the LO had a signiÞcant effect on mosquito populations, particularly in 2000, this effect was not as long-lasting or as marked as would be observed when applying either larvicides or adulticides (Gratz 1991, 1993). Perich et al. (2003) suggested that the LO would be most useful when incorporated into an integrated control program. Our data supports this suggestion, as control achieved in this study was lower than desired.

0.0060 0.7970 0.9299

1,54 2,54 2,54

28.50 1.38 1.60

⬍0.0001 0.2602 0.2103

Attraction of gravid Ae. aegypti away from LO to adjacent containers is not the only possible explanation for the lower than desired control achieved in this study. Although our study design included a buffer zone of houses with LO around the central surveillance area, it is possible that adult Ae. aegypti may have immigrated from beyond the buffer zone into the treated areas to replace females killed by the LO. Edman et al. (1998) reported that availability of oviposition sites was inversely correlated with the potential for female Ae. aegypti to disperse, suggesting that gravid females in the treated areas in this study may have been less likely to disperse than females in areas without LO or with fewer water containers. Conversely, although the LO may have reduced populations of gravid females in the treated areas, this reduction may have led to lower rates of oviposition in containers and corresponding decreases in density dependent larval mortality. A decrease in larval mortality may have resulted in an increased production of adults to or near preplacement levels obfuscating effects of the LO on both immature and adult Ae. aegypti populations (Hawley 1985, Service 1985). The number of water containers per house in this study was substantially higher than reported by Perich et al. (2003) in Brazil and may have diluted the effect of the LO. Although it remains to be determined how signiÞcant a role the LO could play in an integrated control program, the results of our study provide evidence that the LO can reduce Ae. aegypti populations. Active community participation in a clean-up program has been regarded as an effective method of reducing populations of Ae. aegypti in Thailand (Chunsuttiwat and Wasakarawa 1994, Eamchan et al. 1989, Swaddiwudhipong et al. 1992). Cheng et al. (1982) reported that the use of a modiÞed ovitrap resulted in a decrease in the Breteau index of 36% in the control area in Houston, TX compared with a nearly 500% increase in the nonovitrap area; however, use of the ovitrap was accompanied by removal of all other breeding habitats. Incorporation of the LO into a clean-up program would presumably result in more rapid and a greater level of control than clean-up alone, particularly if adjacent water containers were eliminated; however, further studies will be required to conÞrm this.

462

JOURNAL OF MEDICAL ENTOMOLOGY Acknowledgments

Funding for this project was provided by the Military Infectious Diseases Research Program of the U.S. Army Medical Research and Materiel Command, Fort Detrick, MD.

References Cited Bellini, R., M. Carrieri, G. Burgio, and M. Bacchi. 1996. EfÞcacy of different ovitraps and binomial sampling in Aedes albopictus surveillance activity. J. Am. Mosq. Control Assoc. 12: 632Ð636. Chadee, D. D., P. S. Corbet, and H. Talbot. 1995. Proportions of eggs laid by Aedes aegypti on different substrates within an ovitrap in Trinidad, West Indies. Med. Vet. Entomol. 9: 66Ð70. Chan, K. L. 1973. The eradication of Aedes aegypti at the Singapore Paya Lebar International Airport: vector control in southeast Asia. Ministry of Health, Singapore. Chan, K. L., N. S. Kiat, and T. K. Koth. 1977. An autocidal ovitrap for the control and possible eradication of Aedes aegypti. Southeast Asian J. Trop. Med. Publ. Health 8: 56Ð61. Cheng, M. L., B. C. Ho, R. E. Bartnett, and N. Goodwin. 1982. Role of a modiÞed ovitrap in the control of Aedes aegypti in Houston, Texas, USA. Bull. W.H.O. 60: 291Ð296. Christopher, R. S. 1960. Aedes aegypti (L.), The yellow fever mosquito: its life history, bionomics and structure, 739 pp. Cambridge University Press, London. Chunsuttiwat, S., and S. Wasakarawa. 1994. Dengue vector control in Thailand: development towards environmental protection. Gaoxiong Yi Xue Ke Xue Za Zhi. 10: 122Ð123. Eamchan, P., A. Nisalak, H. M. Foy, and O. A. Chareonsook. 1989. Epidemiology and control of dengue virus infections in Thai villages in 1987. Am. J. Trop. Med. Hyg. 41: 95Ð101. Edman, J. D., T. W. Scott, A. Costero, A. C. Morrison, L. C. Harrington, and G. G. Clark. 1998. Aedes aegypti (Diptera: Culicidae) movement inßuenced by availability of oviposition sites. J. Med. Entomol. 35: 578 Ð583. Fay, R. W., and A. S. Perry. 1965. Laboratory studies of ovipositional preferences of Aedes aegypti. Mosq. News 24: 276 Ð281. Fay, R. W., and D. A. Eliason. 1966. A preferred oviposition site as a surveillance method for Aedes aegypti. Mosq. News 26: 531Ð535. Gratz, N. G. 1991. Emergency control of Aedes aegypti as a disease vector in urban areas. J. Am. Mosq. Control Assoc. 7: 353Ð365. Gratz, N. G. 1993. Lessons of Aedes aegypti control in Thailand. Med. Vet. Entomol. 7: 1Ð10. Gubler, D., and G. Kuno. 1997. Dengue and dengue hemorrhagic fever. CAB International, Fort Collins, CO. Hawley, W. A. 1985. The effect of larval density on adult longevity of a mosquito, Aedes sierrensis: epidemiological consequences. J. Anim. Ecol. 54: 955Ð964. Kaul, H. N., and G. Geevarghese. 1979. A comparative study of ovitrap and single larva survey methods for the surveillance of Aedes aegypti in Pune city. Indian J. Med. Res. 69: 71Ð74. Kittayapong, P., and D. Strickman. 1993. Distribution of container-inhabiting Aedes larvae (Diptera: Culicidae) at a dengue focus in Thailand. J. Med. Entomol. 30: 601Ð 606. Lane, R. P., and R. W. Crooskey. 1993. Medical insects and arachnids, 723 pp. Chapman & Hall, London, England. Littell, R. C., G. A. Milliken, W. W. Stroup, and R. D. Wolfinger. 1996. SAS system for mixed models, pp. 42Ð 68. Cary, NC.

Vol. 40, no. 4

Lok, C. K., N. S. Kiat, and T. K. Koh. 1977. An autocidal ovitrap for the control and possible eradication of Aedes aegypti. Southeast Asian J. Trop. Med. Public Health 8: 56 Ð 62. Mogi, M., C. Khamboonruang, W. Choochote, and P. Suwanpanit. 1988. Ovitrap surveys of dengue vector mosquitoes in Chiang Mai, northern Thailand: seasonal shifts in relative abundance of Aedes albopictus and Ae. aegypti. Med. Vet. Entomol. 2: 319 Ð324. [PAHO] Pan American Health Organization. 1994. Dengue and dengue hemoragic fever in the Americas: guidelines for prevention and control. Pan American Health Organization Publication, Washington, DC. Pant, C. P., S. Jatanasen, and M. Yasuno. 1973. Prevalence of Aedes aegypti and Aedes albopictus and observations on the ecology of dengue haemorrhagic fever in several areas of Thailand. Southeast Asian J. Trop. Med. Public Health 4: 113Ð121. Perich, M. J., G. Davila, A. Turner, A. Garcia, and M. Nelson. 2000. Behavior of resting Aedes aegypti (Culicidae:Diptera) and its relation to ultra-low volume adulticide efÞcacy in Panama City, Panama. J. Med. Entomol. 37: 541Ð546. Perich, M. J., M. A. Tidwell, D. C. Williams, M. R. Sardelis, C. J. Pena, D. Mandeville, and L. R. Boobar. 1990. Comparison of ground and aerial ultra-low volume applications of malathion against Aedes aegypti in Santo Domingo, Dominican Republic. J. Am. Mosq. Control Assoc. 6: 1Ð 6. Perich, M. J., A. Kardec, I. A. Braga, I. F. Portal, R. Burge, B. C. Zeichner, W. A. Brogdon, and R. A. Wirtrz. 2003. Field evaluation of a lethal ovitrap against dengue vectors in Brazil. Med. Vet. Entomol. (in press). Reiter, P., M. A. Amador, and N. Colon. 1991. Enhancement of the CDC ovitrap with hay infusions for daily monitoring of Aedes aegypti populations. J. Am. Mosq. Control Assoc. 7: 52Ð55. SAS Institute. 2001. SAS/STAT usersÕ guide. SAS Institute, Cary, NC. Service, M. W. 1985. Population dynamics and mortalities of mosquito preadults, pp. 185Ð201. In: Lounibos, L. P., J. R. Rey, and J. H. Frank [eds.]). Ecology of mosquitoes: proceedings of a workshop. Florida Medical Entomology Laboratory, Vero Beach, FL. Service, M. W. 1992. Importance of ecology in Aedes aegypti control. Southeast Asian J. Trop. Med. Public Health 23: 681Ð 690. Service, M. W. 1993. Mosquito Ecology Field Sampling Methods, 2nd ed, 988 pp. Elsevier Applied Science, London, England. Swaddiwudhipong, W., P. Lerdlukanavonge, P. Khumklam, S. Koonchote, P. Nguntra, and C. Chaovakiratipong. 1992. A survey of knowledge, attitude and practice of the prevention of dengue hemorrhagic fever in an urban community of Thailand. Southeast Asian J. Trop. Med. Public Health 23: 207Ð211. Tonn, R. J., P. M. Sheppard, W. W. Macdonald, and Y. H. Bang. 1969. Replicate surveys of larval habitats of Aedes aegypti in relation to dengue haemorrhagic fever in Bangkok, Thailand. Bull. World Health Organ. 40: 819Ð829. Zeichner, B. C., and M. J. Perich. 1999. Laboratory testing of a lethal ovitrap for Aedes aegypti. Med. Vet. Entomol. 13: 234 Ð238. Received for publication 11 March 2002; accepted 27 January 2003.