ARTÍCULO ORIGINAL

Evaluation of community-based strategies for Aedes aegypti control inside houses

Clara B. Ocampo 1, Camila González 1, Carlos A. Morales 2, Mauricio Pérez 1, Dawn Wesson 3, Charles S. Apperson 4

1 Unidad de Entomología Médica, Centro Internacional de Entrenamiento e Investigaciones Médicas (CIDEIM), Cali, Colombia

2 Área de Salud Ambiental-Programa Dengue, Secretaría de Salud Pública Municipal, Cali, Colombia

3 Department of Tropical Medicine, Tulane University, New Orleans, Louisiana, USA

4 Medical and Veterinary Entomology, North Carolina State University, Raleigh, North Carolina, USA

Recibido: 15/04/08; aceptado:18/02/08


Introduction: Dengue viruses transmitted principally by the urban mosquito Aedes aegypti, cause one of the major public health problems confronting tropical cities. Insecticide spraying has been the mainstay of mosquito control; however, its continuous use has selected for resistance. Other important methods of control involve community participation.

Objective: This study evaluated two control methods for Ae. aegypti that can be used by the community: Lethal ovitraps (LOs) and Bacillus thuringiensis var israeliensis (Bti) briquettes.

Materials and methods: The project study was carried out in four similar neighborhoods within a representative district in the city of Cali, Colombia. Three interventions (LO, Bti, LO+Bti plus education and one control (education only) area were evaluated for efficacy in post-intervention entomological surveys. Additionally, entomological indices were also compared to results from a pre-intervention survey carried out on a sample of city blocks in the same neighborhoods. Relative vector abundance in relation to weather conditions using the same entomological sampling methods was compared.

Results: The interventions did not achieve significant differences in vector abundance among the treatments. However, the interventions achieved a significant reduction in entomological indices compared with those observed during the pre-intervention survey: House index 15.1% vs. 8.5%, mean pupae per house 1.15 vs. 0.073, and Adult index 56.3% vs. 34.8% (p<0.05).

Conclusions: The lack of significant differences among the interventions, and between treated and control blocks suggested that educational activities together with periodic visits to the houses produced similar reductions of immature and adult Aedes aegypti.

Key words: Aedes aegypti, vector control, Bacillus thuringiensis, dengue, consumer participation.


Evaluación de estrategias comunitarias para el control de Aedes aegypti en Cali, Colombia

Introducción. Los virus del dengue transmitidos principalmente por el mosquito urbano Aedes aegypti, causan uno de los mayores problemas de salud pública que confrontan las ciudades tropicales. La aplicación de insecticidas ha sido la base para el control de mosquitos; sin embargo, su continuo uso ha servido para seleccionar individuos resistentes en las poblaciones de mosquitos. Otro método importante para el control involucra la participación comunitaria.

Objetivo. Este estudio evaluó dos métodos de control para Ae. aegypti que podrían ser usados por la comunidad: las ovitrampas letales (OL) y las briquetas de Bacillus thuringiensis var israeliensis (Bti).

Materiales y métodos. El estudio se llevó a cabo en cuatro barrios similares de la Comuna 16 de Cali, Colombia. Se evaluaron tres intervenciones (OL, Bti, OL y Bti) más educación y un área control (sólo educación) para medir la eficacia de la vigilancia entomológica posterior a la intervención. Además, los índices entomológicos se compararon con los resultados de una vigilancia antes de la intervención llevada a cabo en bloques de casas seleccionadas aleatoriamente en los mismos barrios. La abundancia relativa del vector en relación con las condiciones climáticas se comparó usando los mismos métodos del muestreo entomológico.

Resultados. Las intervenciones no produjeron diferencias significativas entre los tratamientos en la abundancia del vector. Sin embargo, las intervenciones lograron una reducción significativa de los índices entomológicos comparados con los observados en la vigilancia antes de la intervención: índice de casa, de 15,1% a 8,5%; promedio de pupas por casa, de 1,15 a 0,073, e índice de adultos, de 56,3% a 34,8% (p<0,05).

Conclusiones. La ausencia de diferencias significativas entre las intervenciones y el bloque control sugiere que las actividades educacionales junto con las visitas periódicas a las casas producen reducciones similares de los estadios inmaduros y adultos de Ae. aegypti.

Palabras clave: Aedes aegypti, control vectorial, Bacillus thuringiensis, dengue, participación comunitaria.


Dengue virus transmission (serotypes 1-4) by Aedes (Stegomyia) mosquitoes is a public health problem that principally affects tropical countries. Almost half of the global population lives in high risk areas and currently more than 100 countries experience dengue fever and dengue hemorrhagic fever (1,2). Each year, an estimated 50 to 100 million new cases of dengue occur around the world (3,4); of these, 500,000 cases correspond to dengue hemorrhagic fever with a mortality of 5% (25,000 cases) (1,2,4,5). Due to the lack of a vaccine or an antiviral treatment, disease prevention is based on vector control (1).

The main vector of dengue virus in the Americas is Aedes aegypti because of its anthropophilic feeding behavior (6,7); its persistence and resting behavior inside houses (8), and, its capability to inhabit most water holding containers (6,9). Rapid, poorly planned urbanization in association with weak regulatory policies for discharge of solid waste has resulted in the accumulation of discarded containers in most developing countries. These accumulations have favored the establishment and geographic spread of this mosquito. The strategies to control dengue transmission used by the public vector control programs have not been adequate in most countries. The emergence of insecticide resistance, the difficulty of eliminating larval production through environmental sanitation, and lack of efficacy of ultra-low volume insecticide spraying to control adults are factors which have limited the effectiveness of vector control programs (10,11).

For these reasons, since 1997, in its hemispheric plan to control Aedes aegypti, the Pan American Health Organization (PAHO) recognized the need to redirect the control strategy from the exclusive use of chemical control. In particular, PAHO emphasized the need to achieve community participation based on appreciation of the problem together with social communication and education. The community-based strategy requires changes in social conduct that will result in the elimination of mosquito production sites by community action. However, change in behavior and elimination of production sites will be difficult to achieve in the short term. In addition, other vector control strategies are necessary for use by the community that do not use insecticide applications that result in selection pressure for insecticide resistance.

In Colombia, the city of Cali has hyperendemic transmission of all four serotypes of dengue virus (12,13). Here, vector control is based principally on two methods of control. First, immature stages in permanent production sites, such as catch basins (14), are controlled by application of chemical and bacterial larvicides twice a month, and (2) adults are controlled during outbreaks or high infestation levels by use of organophosphate and pyrethroid insecticides (Municipal Secretary of Health, personal communication). Community-based projects are focused mainly on multiunit residences and businesses. However, in spite of these vector control strategies, dengue virus transmission persists, and, furthermore, emergence of insecticide resistance has been detected to the organophosphate temephos (15). Additionally, enhancement of mixed-function oxidases and non-specific esterases have been described as two biochemical mechanisms involved in insecticide resistance in these mosquito populations (16,17). The occurrence of insecticide resistance requires a search for alternative strategies for improved vector control.

In the current study four different treatments for Ae. aegypti control were evaluated, with community participation and education as additional variables examined. The interventions included (a) the lethal ovitrap, that targets adult gravid females, (b) the application of Bacillus thurigiensis var israelensis (Bti) briquettes, biological larvicide that target larvae stages, (c) use of both treatments simultaneously, and (d) control (no pesticide treatment).

Materials and methods

Rationale

A main objective was to evaluate alternatives to the application of chemical larvicides to containers that were environmentally benign, low in cost and less likely to produce insecticide resistance. The lethal oviposition trap (LO) (U.S. patent Nº 5,983,557, November 11th, 1999) consists of a black plastic cup fitted with a paper strip that is impregnated with the fast-acting, pyrethroid insecticide, deltamethrin. The LO has been found effective against Ae. aegypti in the laboratory (18) and in the field (19). The LO is designed to attract and kill gravid Ae. aegypti females (18,19). Females depositing eggs contact the deltamethrin-treated paper strip and obtain a lethal dose of insecticide. Since the LO kills gravid Ae. aegypti females, the trap targets the epidemiologically relevant portion of vector populations and thereby decrease transmission of dengue virus. The LO method is environmentally benign, and appears to be a promising strategy to integrate with vector control programs based on community participation. The briquettes composed of Bti control larval stages of Ae. aegypti (20). The floating briquette formulation (Bactimos Briquettes, Summit Chemical Co., Baltimore, MD, EPA registration Nº 6218-47) releases a constant, low dose of bacterial toxins for a period of a month or more, and has been used to control Ae. aegypti in its main production sites, water storage containers in houses. Bti is a biological insecticide that is environmentally safe and the briquettes can be used by the community after minimal instruction (21).

Study design

The study was continued over a 2-year period from 2004 to 2005. The study was centered in Comuna 16 (an administrative division for public services that correspond to several neighborhoods) because numerous cases of dengue had occurred there during the previous five years, with an average of 85 cases per 100,000 inhabitants per year. From this comuna, four neighborhoods were selected: República de Israel, Mariano Ramos, Unión de Vivienda Popular and Antonio Nariño (figure 1). All of these neighborhoods had similar demographic parameters in terms of property size, population density and socio-economic strata (these were classified as level 2 of 6 social stratification levels, as defined by the Congreso de Colombia. Ley 142 de 1994 (11 de julio), artículo 102). These neighborhoods were representative of more than half of the Cali population with basic public services and houses constructed of bricks and a clay tile roofing. The regularly spaced configuration of city blocks permitted a symmetrical experimental design.

Four treatments were evaluated on four city blocks within each neighborhood: (a) LO, (b) Bti briquettes, (c) LO+Bti, and (d) control (no pesticide intervention). One block was selected in each neighborhood using random numbers generator; however, the randomly selected blocks were accepted if they had no commercial activity and were located at a minimum of 500 m apart from the previously selected block. Before activating the interventions, all families from the selected blocks were informed about dengue and the study goal (see education activities). Informed consent was obtained from an adult in the house, who agreed to participate as well as to allow biweekly visits to his/her house. In addition, in the selected blocks, all the water in the containers were discarded explaining to the owner the reason for doing it, and suggesting to keep them dry and clean or dispose them if possible.

After this process, one intervention methodology was randomly assigned to each block, including the block that did not receive any intervention (control block). The control block was expected to serve as a comparison of the effects of education, the initial container removal and the biweekly visits by the project team. The interventions were applied to every house in the selected blocks (approximately, 40 houses per block in an area of 100 m2) and in every house along the treatment block edge; the latter houses were designated as the buffer zone and consisted of approximately 40 houses (10 houses by side) (figure 2). The buffer zone was expected to provide a barrier to re-infestation of Ae. aegypti into the intervention block by eliminating females immigrating from the outside. Once the interventions were initiated, 10 houses in each experimental block were randomly sampled every 15 days to estimate Ae. aegypti densities. The interventions were continued for a period of four months (April-July, 2005); this included a high (April-May, 313 mm) and low (June-July, 116.5 mm) precipitation period. The sample size of 10 houses per intervention block was chosen to avoid visiting the same house at regular intervals. Data from each house consisted of establishing the entomological indices (house, pupal and adult) and identifying the mosquito species.

Before the intervention period, weekly entomo-logical surveillance (house, pupal and adult indices) during the previous year (May to October, 2004) was carried out in randomly selected blocks of the Comuna 16, using the same sample size (10 houses) per block. This surveillance permitted an estimation of the entomological indices for the area to see if they were of similar magnitude among the neighborhoods, as well as to characterize the types of mosquito-positive containers present. This period also included high and low precipitation periods –April-May (360 mm) and June-July (108 mm). The use of the same sample size during the pre-intervention and intervention periods permitted a comparison of the entomological indices and changes in mosquito densities due to weather changes. Before the intervention, laboratory tests were conducted to measure the lethality of the lethal ovitrap on gravid females and first instar larvae. The tests were conducted for a one month period, using local captured mosquitoes in indoor and outdoor ambient conditions.

Mosquito sampling

The entomological survey was conducted in the selected houses to estimate the abundance of Ae. aegypti larvae, pupae and adults. At each site, every water-holding container was inspected. Containers were placed in the following categories: high (rooftop) water storage tanks, low (groundlevel) water storage tanks (washbasins), tires, flower pots or plant in water, bottles, cans, small miscellaneous (<500 ml volume) and large miscellaneous (>500 ml). For each container, the water volume was measured and the presence of larvae and/or pupae was recorded. Containers with Aedes larvae and/or pupae were recorded as positive. All the pupae were collected, counted, packed in separate tubes labeled by site, and transported to the laboratory at CIDEIM for species identification after adult emergence. Larvae were left in the containers during the intervention period to avoid a reduction in the mosquito population. Sampling of adult mosquitoes inside each house was accomplished with modified electric manual aspirators (Insect Vac aspirator, Bioquip) as described by Perich et al. (19). Captured mosquitoes from each house were placed in a separate cage and transported to the laboratory for identification.

Entomological data were used to calculate the following indices: house index (percentage of houses infested with larvae and/or pupae); pupal index (mean number of pupae per house); and adult index (percentage of houses infested with adult mosquitoes). Entomological surveys were completed twice a month in identical fashion for all interventions and the control area.

In addition to the household surveys, catch basins were inspected for mosquitoes during the survey and intervention. Catch basins (stormwater collection along streets) are important production sites for mosquitoes in Cali because they hold water permanently (14). The catch basins are cement boxes of 1 m x 0.5 m x 0.8 m of size that maintains an average of 100 to 150 liters of water permanently. Two catch basins, present in the area where the houses were sampled, were evaluated for mosquito larvae. Each catch basin was opened and sampled by taking 20 dip collections with a 300 ml cup and recording the presence of larvae and pupae. All pupae collected in each catch basin were counted and identified after their emergence in the laboratory. During the study, the Secretary of Public Health of Cali separately continued their ongoing control program in the catch basins with the insect growth regular Triflumuron® applied biweekly.

Efficacy of the lethal ovitraps after environmental exposure

Before the intervention, LO efficacy was evaluated under local conditions. Traps were placed in an empty lot next to the CIDEIM laboratory to simulate the environmental conditions in which the LO would be exposed at the house sites. Ae. aegypti Rockefeller susceptible strain and a local strain collected from the study area were used to bioassay the toxicity of LOs that had been exposed to two environmental conditions: inside the house and in an open area (backyard). In the bioassays, the mortality rate of blood fed females three days after feeding and second instars larvae was evaluated after exposure to LOs that had been placed in the two environmental conditions for periods of 10, 20 and 30 days. Each assay consisted of the following tests: (a) three replicate cages with 10 bloodfed females and 10 larvae, each containing a cup with a deltamethrin (0.1 mg) strip, (b) two replicate control cages, each with a cup containing a new strip (positive controls), and (c) a cage without strip (negative control). After 24 hours, survival rates of larvae and adults in each cage were recorded. Additionally, adult mortality was recorded at 45-minute intervals for the first 5-8 hours of each assay in order to compare the killing rates of each lethal ovitrap.

Lethal ovitraps

During the intervention, 10 LOs were placed in every house of the block chosen for this treatment. Five LOs were placed inside the house and five were located in open areas (backyard and front yard). The deltamethrin strips in the LOs were changed twice a month due to loss of activity observed in laboratory susceptibility tests of adult Ae. aegypti from the study zone (see results). Hay infusion (10% dilution) was added to the LO as an attractant on the first day the traps were set (22). When LOs were serviced and ovistrips were replaced, cups were emptied and refilled with water without cleaning the cup leaving bacterial contents on the wall of the cup. At the time the ovistrips were replaced, the presence of eggs was recorded and the LOs were categorized according to the eggs number (1-10; 11-30; 31-80; >80).

Bti briquettes

In each house, a Bti briquette was placed in the main larval production sites. These were principally ground-level water storage tanks, with the water used for clearing or clothes washing. According to manufacturer’s recommendations, a quarter (¼) briquette was placed in each tank, usually of 0.1 to 0.5 m2 in surface area. A net bag was hung from the tank faucet. Its design allowed maintaining the briquette in the tank with or without water. The briquettes retained their effectiveness even after drying. The briquettes were changed once per month and the condition of the tank and briquette recorded.

Education of the study population before intervention

Two weeks before the intervention, each house in the study neighborhoods was visited by two project members who informed the residents about the dengue problem, the biology of the vector, and the Ae. aegypti control methods to be used in the study. A brochure with this information was left in each house. After receiving informed consent to proceed, the project team located mosquito-infested, water-holding containers present in the house, showed them to the residents, and explained the mosquito life cycle. The mosquitoes were removed and the water discarded as part of the clean-up campaign. Families were informed about the re-inspection schedule. The treatment blocks were selected randomly after the visits.

At the end of the 4-month intervention period, a questionnaire was given to an adult family member in the house to evaluate the usefulness of the control methods to the community and to evaluate basic knowledge about dengue virus transmission.

Statistical analysis

Data were entered in an Access database (Microsoft© Office Access, Microsoft Corp., Seattle, WA) and then exported into ArcMap (ver. 8.3, Esreri, Redlands, CA), a geographic information system. For statistical analysis, the house was the sample unit. Data were analyzed using SPSS 7.5 software for windows (SPSS Inc, Chicago 1996) and Stata 6.0 (Stata Corp., 1999). Entomological indices (House and Adult) were expressed as percentages. Comparisons were tested for significance among interventions (treatments vs. control) and between years (2004 vs. 2005) using Poisson regression. The pupal index was analyzed with a negative binomial regression to compare between the interventions and the first survey. Significance level was set at p ≤0.05 for all statistical analyses.

Results

Efficacy of the lethal ovitraps

Lethal ovitraps produced 97-100% mortality of larvae and adults of the Ae. aegypti Rockefeller strain (susceptible strain) for all exposure times (10, 20, 30 days) evaluated under the two environmental conditions (inside the house and open area: shade and sunlight) (table 1). This finding confirmed that water quality did not affect LO insecticide activity. However, when Ae. aegypti from the study site (Cali strain) were evaluated, approximately 90% mortality of adults and larvae was achieved at time 0, but after 10 days of exposure to an open environment, mortality significantly decreased (table 2). LOs exposed to both environmental conditions for 10 days prior to efficacy evaluation produced a mortality rate significantly lower for Cali strain mosquitoes (adults and larvae) compared with the Rockefeller strain (table 2). Mortality rates for the Cali strain of Ae. aegypti after 10 days of exposure were: 70.0% for adults and 60.0% for larvae inside a house; and 58.3% for adults and 20.8% for larvae in an open area exposed to sunlight. When mortality was recorded at 45-minute intervals during the assay with the LO exposed 10 days in an open area, significant declines in mortality occurred, as well as an increase in the time required to achieve mortality for the Cali strain compared to the Rockefeller strain (figure 3). Due to the low mortality rates with the Cali strain, the longer exposure times were not evaluated. Biochemical levels for non-specific esterases, mixed function oxidases and insensitive acetylcholinesterase were tested with the Ae. aegypti Cali strain, but no significant differences between Rockefeller and Cali strain were noted (data not shown). These results led to a strategy of replacing the LOs deltamethrin strips twice a month during the intervention phase in order to decrease the possibility that the traps would become mosquito production sites (see methodology).

Entomological surveillance pre-intervention

During the six-month survey period (May to October, 2004), 255 houses from Comuna 16 were sampled. The houses sampled from each district were as follows: 82 houses from “República de Israel” (32.2%), 70 from “Mariano Ramos” (27.5%), 63 from “Unión de Vivienda Popular” (24.7%) and 40 from “Antonio Nariño” (15.7%) (figure 1). The entomological indices varied with precipitation (figure 4). Entomological indices for immature stages of Ae. aegypti (house and pupal indices) varied directly with precipitation; in contrast, higher values for the adult index were found in the months with less precipitation (June, July and August) (figure 4). Entomological indices did not differ among the neighborhoods (p<0.01).

During the 2004 survey, 69 of 327 containers with water inspected were found positive for immature mosquitos in 49 houses (21.1%). The most common container with water found was low water storage tanks (47%). The percentage of positive containers for larvae and pupae in the houses were principally flower pots or plants in water (33%), low water storage tanks (21.7%), and large miscellaneous containers (20.3%) that were mainly plastic or metallic tanks for water storage. Small volume miscellaneous containers represented only 17.4% of the mosquito-positive containers and comprised of ceramic plates under the flower pots or plastic containers abandoned in house yards. A few tires (2.9%), bottles (2.9%) and tin cans (1.4%) were also found with water and were positive for Ae. aegypti larvae or pupae.

When the containers were analyzed regarding pupal production, the importance of the containers changed. Only 19 of 69 positive containers (27%) contained pupae and were mainly low water storage tanks (36.8%) and large volume miscellaneous containers (31.6%). Flower pots or plants in water comprised only 15.8% (n=3) of the mosquito-positive containers. The pupal index averaged 1.15 pupae per house or 0.2 pupae per person. In the containers, Ae. aegypti and Cx. quinquefasciatus were found in a ratio of 18.75:1. However, the adult index was 50.2% for Ae. aegypti and 41.2% for Cx. quinquefasciatus. A mean of 0.33 Ae. aegypti adults per person was calculated.

Entomological surveillance during the intervention

During the intervention, reductions in entomo-logical indices achieved in the untreated block was not significantly different (p>0.05) from those blocks in which the control measures were implemented (table 3). When the monthly entomological indices obtained from the initial survey in 2004 were compared with the control block in 2005, no significant differences were found. However, when the overall data from the intervened blocks were compared with the 2004 survey, a significant reduction in the number of pupae and adults was detected (p<0.05) (table 4). The major effect of all interventions was a reduction in production of pupae. The adult index during the intervention period did not increase later in the season during the low precipitation period, in contrast to the observations in the initial survey (2004). The type of positive containers for immature stages varied. Positive containers consisted principally of plants in water (50%), cans and small volume (<500 ml) miscellaneous containers (27.5%). Only 2 out of 103 filled low water storage tanks were positive for Ae. aegypti immature stages. As expected with this type of positive containers, pupae were not frequently found. Only 18 pupae were collected from 4 houses during the intervention study.

In the treatment blocks with LOs (LO and LO+Bti), 961 were placed in 98 houses. From 4,492 service visits of LOs, 45.2% of the strips were positive for Ae. aegypti eggs, but only 1.15% of the LOs contained larvae and 0.31% contained pupae; indicating that the immature stages were susceptible to deltamethrin during the fifteen days of the toxic strip evaluation.

The Bti briquettes were placed in 76 low water storage tanks and replaced monthly. During the study evaluation (341 visits) only one tank was positive for larvae. In 64% of the visits, the briquette was present in the tank and in 60% of the briquettes were submersed in the water.

Catch basins

Catch basins were evaluated during the pre-intervention and intervention period (table 5).. During the pre-intervention period a higher percentage of the catch basins were positive for pupae and more pupae were collected. In both periods the number of catch basins with Ae. aegypti pupae were low. Most of the positive containers with Ae. aegypti were found in June.

Knowledge survey

At the end of the study, a brief questionnaire was administered to an adult member of each household that participated in the study. A total of 179 surveys were completed. In terms of dengue knowledge, 81% of the people knew that dengue is transmitted by mosquitoes, 66.5% answered that Ae. aegypti is produced in water-holding containers inside the house, and 68% knew that not storing water in their houses was one strategy of mosquito control. With respect to knowledge of interventions, only 29.6% with LOs and 29.5% of the people associated the Bti in their houses with control of Ae. aegypti. However, 75.3% with LOs and 68.2% of the people with Bti perceived reductions in adult mosquito numbers inside their houses; furthermore, 75.3% (LO) and 71.6% (Bti) answered that they would continue using these methods for mosquito elimination.

Discussion

Four interventions in the community using lethal ovitraps, Bti briquettes, the combination of both methods and community education were evaluated. The entomological indices obtained during the intervention were not significantly different among the treatments or controls. The entomological indices in the untreated area were not significantly different from the pesticide treatments. This was possibly due to an effect of the education and cleanup campaign at the beginning of the intervention study and the continuous visits by the research team (twice a month). However, a significant reduction in pupae and adult indices was observed between the pre-intervention survey (2004) and the intervention period (2005). Unfortunately, the decrease in abundance of pupae did not eliminate the presence of adult mosquitoes in the houses to the levels sufficiently low to block transmission. This suggests that the buffer areas surrounding the intervention areas were not large enough to prevent immigration of adult mosquitoes; possibly additional production occurred in cryptic larval habitats not identified in the surveys.

In spite of the small sample size, variation in the types of positive containers, positive houses and adult densities in relation to the intervention and precipitation was observed. Although the sample size is a limitation of the study, it allowed us to identify entomological characteristics from the study sites so that we could directly compare results. During the intervention, the untreated area was affected by some of the same variables affecting mosquito production in the treated blocks (education, cleaning campaign, and periodic visits), which allowed us to decrease the bias and measure the real effect of the pesticide methods.

During the pre-intervention survey, the entomolo-gical indices varied over time and were principally affected by precipitation. The indices related to immature stages (house and pupal indices) varied in direct relation with the precipitation. In contrast, the adult index increased during the period of low rainfall. The high abundance of adults suggested that there were other production sites outside the houses during the dry period. This speculation was also supported during the intervention period when pupal production in houses was very low (mean=0.073 pupae per house) but adults persisted in the houses and a reduction of only 30% was observed. The production sites observed for Ae. aegypti outside the houses were the catch basins. They have been previously described as permanent breeding sites for Cx. quinquefasciatus and Ae. aegypti because they permanently maintain water (19,23). In this study, the catch basins evaluated were highly positive for larvae and pupae of Cx. quinquefasciatus (71.4% reintervention and 57.6%-intervention) and low percentage of them were positive for Ae. aegypti pupae (14.2% and 15.3%, respectively), principally in June. The low occurrence of Ae. aegypti in catch basins during the study, could be explained by the maintenance of the control program with the insecticide Triflumuron which is effective against Ae. aegypti but not against Cx. quinquefasciatus (24). However, the production of Ae. aegypti in catch basins observed in this study does not explain completely the increase of adult frequency in the houses since Aedes pupae was found principally in June but not during the entire dry period. Possible explanations are that the number of catch basins examined was low (2 per 10 houses), the presence of other produc-tion sites outside the houses (not found in the area covered), or mosquito adult migration from outside buffer zone to treated zone during the intervention period. This last one, suggests the need of a larger buffer zone to prevent immigra-tion and entrance of adults to the study houses. It is well known that Ae. aegypti adults mosquitoes can fly long distances (>500 m) (25-27).

During the intervention, the adult index was not significantly different between blocks with LOs and the control block, suggesting that the LO treatment did not reduce the abundance of adults in the houses in comparison with houses with no insecticide treatment but with educational intervention. The presence of adults in treatments areas could be explained as mentioned before due to the small area treated (block and buffer zone) that did not prevent the immigration of adults and/or because of low adult mortality found in the LO bioassays with the local strain of Ae. aegypti prior to the intervention study. However, even with the low LO efficacy observed on adults, we observed that the LOs were effective in killing immature stages in the field since 40.8% of the strips were positive for eggs but only 1.4% of LOs contained larvae or pupae which escaped the effect of the insecticide on the strip. This finding suggested that the LO must be exerting a controlling effect on Aedes populations, due to the low emergence of larvae observed, but effects on abundance of adults could not be detected in the small area treated. In previous studies using LOs against Ae. aegypti in Brazil and Thailand, some adult reduction was found in treated areas, indicating that LOs do not completely eliminate adults but could be combined with other strategies to achieve effective vector control (28). Our study did not show a clear reduction of adults in LO treated areas in comparison with the control area (education only); however, the low susceptibility of Ae. aegypti adults from Cali to the deltamethrin treated strips in LOs suggests that future studies with this methodology should use insecticides with higher efficacy in the laboratory and that LOs should be placed over a larger area to prevent A. aegypti immigration.

The Bti briquette was effective in eliminating Ae. aegypti larvae in the low water storage containers since we only found a container with larvae during the intervention period with this control method. Unfortunately, in a high percentage of houses (40%), the briquettes were not used routinely because when they dissolved a visible residue remained in the water, causing people to remove the briquette. This could be the reason for the presence of larvae in the positive container. Since the project team left the briquette and checked them only during the entomological survey, its continued use could not be confirmed. It is possible that in addition to the insecticide effect, its use may also induce a behavioral change in the house inhabitants which clean the water storage containers more often. Although there were no significant differences among treatments, it appears that the combined treatment (LO+Bti) achieved a substantial reduction in entomological indices, eliminating positive containers and reducing adult abundance in July (table 3). It is possible that the mosquito reduction was a consequence of the combination of treatments and behavioral changes of the house inhabitants. We speculate that residents learned to better identify the relationship between the presence of larvae and the production of adults and eliminated the potential mosquito breeding sites.

In the knowledge survey, the majority of the people (81%) knew that dengue is transmitted by the mosquito Ae. aegypti. However, only 66% of the people interviewed associated the presence of the adult mosquitoes with container habitats of larvae in their houses. In spite of all the activities in the project (education and periodic inspections), only 30% of the inhabitants were able to recognize positive benefits from the treatment methods for vector control. This low percentage perhaps because were not familiar with the methods or mortality of adult mosquitoes was not directly observed. Regardless, they did perceive that adult populations were reduced in their houses.

Results obtained in this study indicate that the area covered in the interventions is not enough to eliminate transmission of dengue viruses in Cali conditions. Additionally, to the small area covered by the treatments, the buffer zone did not work as a barrier for adult mosquitoes. At this level, education and permanent surveillance are as effective in reducing mosquito populations as the insecticidal control methods evaluated. This indicates that future studies should cover larger areas in terms to evaluate the effect of any potential control method to be use by the community.

The low number of pupae found inside the houses during the pre intervention and intervention which was lower than the transmission threshold of 0.5 to 1.5 Ae. aegypti per person suggested by Focks and coworkers (28), contrast with the high infestation of adults in houses. These results confirm the presence of breeding sites outside houses in Cali that requires careful investigation. Additionally, the variation of entomological observed during the different precipitation periods in the pre-intervention period, indicates the complexity of the dynamics of the immature stage and adult mosquito populations and the necessity to understand human behavior, mosquito population dynamics and the production potential of sites inhabited by immature Ae. aegypti inside and outside houses.

Acknowledgment

We thank the Municipal Secretary of Health for support of this project with technical staff, Rodrigo Rivas and Violeta Jaramillo, who participated in the conduct of the field work. We also thank CIDEIM personnel Paola Andrea Caicedo, Neila Julieth Mina, Luis Ernesto Ramírez, y Gilberto Londoño for their collaboration in the fieldwork. We gratefully acknowledge the assistance and advocacy of “Junta de Acción Comunal” from Comuna 16, in introducing the team and the study to the neighborhoods. LO were provided by the US Army and Bti briquettes were donated by Summit Chemical Co. This project was supported by Colciencias contract 2229-04-14346.

Conflict of interests

We did not identify any conflict of interest.

Financing

This project was supported by Colciencias contract 2229-04-14346

Correspondencia: Clara Beatriz Ocampo, Unidad de Entomología Médica, Centro Internacional de Entrenamiento e Investigaciones Médicas, CIDEIM, Carrera 125 Nº 19--225, Cali, Colombia. Teléfono: 555 2164; telefax: 555 2638 claraocampo@cideim.org.co

References

1. Spiegel J, Bennett S, Hattersley L, et al. Barriers and bridges to prevention and control of dengue: the need for a social-ecological approach. EcoHealth. 2005;2:273-90.

2. Guha-Sapir D, Schimmer B. Dengue fever: new paradigms for a changing epidemiology. Emerg Themes Epidemiol. 2005;2:1.

3. Gubler DJ. The changing epidemiology of yellow fever and dengue, 1900 to 2003: full circle? Comp Immunol Microbiol Infect Dis. 2004;27:319-30.

4. Gubler DJ. Resurgent vector-borne diseases as a global health problem. Emerg Infect Dis. 1998;4:442-50.

5. Guzman MG, Kouri G. Dengue: an update. Lancet Infect Dis. 2002;2:33-42.

6. Christopher RS. Aedes aegypti (L.), the yellow fever mosquito; its life history, bionomics and structure. London: Cambridge Univ. Press; 1960.

7. Morrison AC, Costero A, Edman JD, Clark GG, Scott TW. Increased fecundity of Aedes aegypti fed human blood before release in a mark-recapture study in Puerto Rico. J Am Mosq Control Assoc. 1999;15:98-104.

8. Perich MJ, Davila G, Turner A, Garcia A, Nelson M. Behavior of resting Aedes aegypti (Culicidae: Diptera) and its relation to ultra-low volume adulticide efficacy in Panama City, Panama. J Med Entomol. 2000;37:541-6.

9. Focks DA, Chadee DD. Pupal survey: an epidemiologically significant surveillance method for Aedes aegypti: an example using data from Trinidad. Am. J Trop Med Hyg. 1997;56:159-67.

10. Gubler DJ, Clark GG. Dengue/Dengue Hemorraghic Fever. The emergence of a global health problem. Emerg Infect Dis. 1995;1:55-7

11. Gubler D. The emergence of epidemic dengue fever and dengue hemorrhagic fever in the Americas: a case of failed public health policy. Rev Panam Salud Pública. 2005;17:221-4.

12. Ministerio de Salud de Colombia. Prevención y control del dengue. Boletín epidemiológico, semana 42. Sivigila, Ministerio de Salud, Colombia. Nov 2001; 2001.

13. Ministerio de Salud de Colombia. Comportamiento por regiones del dengue en el 2000. Boletín epidemiológico nacional, semana 2. Sivigila, Ministerio de Salud, Colombia; 2002:2.

14. Suárez-Rubio M, Suárez M. The use of copepod Mesocyclops longisetus as a biological control agent for Aedes aegypti in Cali, Colombia. Am J Trop Med Hyg. 2004;20:401-4.

15. Suarez MF, Gonzales R, Morales CA. Themephos resistance to Aedes aegypti in Cali, Colombia. Am J Trop Med Hyg. 1998;55(Suppl.):257.

16. Hernandez VC, Segura I, Wesson DM, Ocampo CB. Insecticide resistance dynamics in Aedes aegypti from Cali, Colombia. Am J Trop Med Hyg. 1999;61(Suppl.):436.

17. Ocampo CB, Wesson DM. Population dynamics of Aedes aegypti from dengue hyper-endemic urban setting in Colombia. Am J Trop Med Hyg. 2004;71:506-13.

18. Zeichner BC, Perich MJ. Laboratory testing of a lethal ovitrap for Aedes aegypti. Med Vet Entomol. 1999;13:234-8.

19. Perich MJ, Kardec A, Braga IA, Portal IF, Burge R, Zeichner BC. Field evaluation of the lethal ovitrap against dengue vectors in Brazil. Med Vet Entomol. 2003;17:205-10.

20. Sulaiman S, Jeffery J, Sohadi AR, Yunus H, Busparani V, Majid R. Evaluation of Bactimos wettable powder, granules and briquets against mosquito larvae in Malaysia. Acta Tropica. 1990;47:189 -95.

21. WHO. Environmental Health Criteria, No 217: Bacillus thuringiensis; 1999.

22. Reiter P, Amador MA, Colon N. Enhancement of the CDC ovitrap with hay infusions for daily monitoring of Aedes aegypti populations. Am J Mosq Control Assoc. 1991;7:52-5.

23. Sithiprasasna R, Mahapibul P, Noigamol C, et al. Field evaluation of a lethal ovitrap for the control of Aedes aegypti (Culicidae) in Thailand. J Med Entomol. 2003;40:455-62.

24. Giraldo-Calderon G, Pérez M, Morales CA, Ocampo CB. Evaluación del triflumurón y la mezcla de Bacillus thuringiensis más Bacillus sphaericus para el control de las formas inmaduras de Aedes aegypti y Culex quinquefasciatus en sumideros de Cali, Colombia. Biomédica 2008;28:224-33.

25. Reiter P, Amador MA, Anderson RA, Clark GG. Short report: dispersal of Aedes aegypti in an urban area after blood feeding as demonstrated by rubidium-marked eggs. Am J Trop Med Hyg. 1995;52:177-9.

26. Russell R, Webb C, Williams C, SA R. Mark-release-recapture study to measure dispersal of the mosquito Aedes aegypti in Cairns, Queensland, Australia. Med Vet Entomol. 2005;19:451-7.

27. Honorio NA, Silva WdC, Goncalves JM, Lounibos LP, Lourenco-de-Oliveira R. Dispersal of Aedes aegypti and Aedes albopictus (Diptera: Culicidae) in an urban endemic dengue area in the State of Rio de Janeiro, Brazil. Mem Inst Oswaldo Cruz 2003;98:191-8.

28. Focks DA, Brenner RJ, Hayes J, Daniels E. Transmission thresholds for dengue in terms of Aedes aegypti pupae per person with discussion of their utility in source reduction efforts. Am J Trop Med Hyg. 2000;62:11-8.