The impact of indoor residual spraying of deltamethrin on dengue vector populations in the Peruvian Amazon

Abstract

Dengue is an important public health problem in the Amazon area of Peru, resulting in significant morbidity each year. As in other areas of the world, ultra-low volume (ULV) application of insecticides is the main strategy to reduce adult populations of the dengue vector Aedes aegypti, despite growing evidence of its limitations as a single control method. This study investigated the efficacy of deltamethrin S.C. applied through indoor residual spraying (IRS) of dwellings in reducing A. aegypti populations. The residual effect of the insecticide was tested by monthly bioassays on the three most common indoor surfaces found in the Amazon area: painted wood, unpainted wood and brick. The results showed that in an area with moderate levels of A. aegypti infestation, IRS dramatically reduced all immature indices the first week after deltamethrin IRS application and the adult index from 18.5 to 3.1, four weeks after intervention (p < 0.05). Even though housing conditions facilitated reinfestation with A. aegypti (100% of the houses have open roof eaves, 31.5% lack sewage systems, and 60.4% collected rain in open containers), indices remained low compared to baseline 16 weeks after insecticide application. Bioassays showed that deltamethrin S.C. caused mortalities >80% 8 weeks after application on all types of surfaces. The residual effect of the insecticide was greater on brick than on wooden walls (p < 0.05). Our results demonstrate that IRS can have both an immediate and sustained effect on reducing adult and immature A. aegypti populations and should be considered as an adult mosquito control strategy by dengue vector control programs.

1. Introduction

Dengue fever, also known as break-bone fever, is a vector-borne viral disease transmitted primarily by the Aedes aegypti mosquito. Dengue fever is a major public health problem in Peru. After its first report in the 1990s (Phillips et al., 1993), numerous dengue outbreaks have taken place and the number of cases of both classic dengue and its more severe manifestation, dengue haemorrhagic fever (DHF), have increased dramatically. The majority of the cases occur in poor areas of the country, particularly affecting urban and suburban communities in the Amazon region (Guzmán, 2014).

Without a vaccine, vector control remains the main strategy to prevent and control dengue (Cattand et al., 2006). Vector control interventions that target immature stages of the mosquito mainly rely on the use of chemical larvicides in water holding containers located in close proximity to human dwellings. Although the prevailing control paradigm has been focused on the immature stages, these interventions have limited impact on adult mosquitoes, since most larvae will die before reaching the adult stage (Morrison et al., 2008). Interventions targeting the adult stage of the mosquito are often only used reactively in response to reported dengue cases. Space spraying with aerosol insecticides through ultra low-volume applications (ULV) (Lofgren, 1970) is one of the main strategies used against adult A. aegypti during dengue outbreaks in Peru. Although its main objective is to reduce on-going epidemics by reducing adult mosquito densities in the short-term, its effect on the incidence and transmission of dengue is limited (Newton and Reiter, 1992; Koenraadt et al., 2007; Esu et al., 2010).

Indoor residual spraying (IRS) is one of the main vector control strategies for targeting malaria vectors, and is based on the spraying of residual formulations of insecticides on all indoor wall surfaces in human habitations (WHO, 2006b). Although IRS is not currently recommended for dengue control, there is evidence of its potential efficacy in reducing populations of A. aegypti (Suleman et al., 1996; Rozilawati et al., 2005; Lenhart et al., 2008; Esu et al., 2010) and its application as an alternative for dengue vector control has been discussed (Morrison et al., 2008; Chadee, 2013). Indeed, IRS has demonstrated efficacy in the past to control dengue vectors in several areas of the world (Giglioli, 1948; Ritchie et al., 2002; WHO, 2006c; Vazquez-Prokopec et al., 2010).

Dengue is associated with population growth and uncontrolled urbanization in tropical countries (Kyle and Harris, 2008). In recent years, the city of Iquitos in the Amazon region of Peru has grown rapidly in a chaotic way, triggered by the limited resources in surrounding rural areas (Gobierno_Regional_de_Loreto, 2009). As a result, a high percentage of the population lives in semi-urban communities that lack basic services such as adequate sanitation, drinking water and sewage disposal. As water availability is erratic, residents store water in containers that serve as breeding sites for mosquitoes, increasing the risk of acquiring dengue, malaria and other infectious diseases (Institute of Medicine, 2008). Residents of these areas build their houses with any available materials and living areas are often partially exposed to the outdoors. In addition, houses rarely have curtains in the windows (Schneider et al., 2004), which may result in higher exposure to sunlight of insecticides applied indoors. Additionally, the high temperatures inside of houses could also potentially affect insecticidal activity (Gurtler et al., 2004).

Taking these factors into account, it was important to determine if the localities studied here would be suitable for large-scale IRS interventions. To address this question, we tested the residual effect of deltamethrin S.C. on A. aegypti adults in a subset of houses that were representative of the three most common wall surfaces in the city of Iquitos. We hope these findings will contribute to the growing body of information regarding the efficacy of vector control interventions targeting adult dengue vectors.

2. Methods

2.1. Study site

The study was conducted in the suburban locality of San Juan Bautista, in the southern suburbs of the city of Iquitos in the Loreto region of Peru (Fig. 1). Iquitos is the largest city in the Peruvian Amazon and the fifth most populated city in Peru, with an estimated 430,000 inhabitants (INEI, 2012). Availability of piped water is irregular in the study area and the collection and storing of rain-water is a common practice. Poverty indices are high in Loreto, with 56% of the population living under the poverty line (INEI, 2010). Houses in the study area are built from a variety of materials, with wood and bricks (commonly un-plastered) as the most commonly used construction materials. Roofs are made of woven palms or metal sheeting, leaving open spaces around the eaves to facilitate air flow, due to the hot and humid climate. Due to its location in a tropical rainforest, Iquitos is characterized by high temperatures throughout the year, with an average daily temperature of 25.8 °C (ranging 21.9 °C–34.4 °C) and heavy rainfalls, with an annual precipitation of 3.4 m (Morrison et al., 2010).

Fig. 1.

The study site is located in the southern part of Iquitos city, in the department of Loreto, in the Amazon region of Peru. The study houses where bioassays were carried out are shaded according to their wall surface type. Untreated control houses are represented with an “x”.

2.2. Housing condition survey

Trials were conducted from March to July 2012 and follow-up occurred in 36 houses: 12 constructed with painted wood, 12 with unpainted wood and 12 with unpainted brick. In addition, three houses (one per type of material) were used as untreated controls. Control houses had to be free of previous insecticide use for at least 6 months for IRS and 1 month for ULV and were maintained unsprayed during the trial period. We used questionnaires to obtain information at baseline regarding structural and behavioural factors that could affect treatment results, such as the state of the roofs and eaves (open/closed), type of kitchen, construction materials, the use of domestic insecticides, the type of water supply and methods of waste disposal. In addition, all water-holding containers on the premises were inspected for the presence of A. aegypti larvae and pupae. An authorized board from the Ministry of Health of Peru reviewed and approved our procedures.

2.3. IRS intervention

The IRS intervention was carried out in March 2012, to coincide with the rainy season (September–May), when A. aegypti densities are especially high. Deltamethrin was applied to all interior wall surfaces and also outdoors, in the walls of the peridomestic environment. Spraying was carried out following the guidelines of the Ministry of Health in Iquitos: 100 ml deltamethrin (Delta 5% SC; Farmex Corp., Lima, Peru) diluted in 8 l of water was sprayed at 757 ml/min with a pressure of 3 bar with a Hudson X-Pert sprayer (H.D. Hudson Manufacturing, Chicago, IL). IRS was applied in all houses in the neighborhood, with the exception of the control houses. Follow-up was conducted in a subset of 36 houses representing the 3 main types of construction materials. Inhabitants were informed of the IRS objectives and procedures during initial household surveys and before application. To ensure the correct application of insecticides, we followed WHO recommendations: the equipment was verified prior to intervention and the application process was supervised at all times by trained staff. In addition, appropriate safety measures were taken according to WHO guidelines (WHO, 2002).

2.4. Bioassays

The bio efficacy of deltamethrin on the sprayed walls was tested using the insecticide-susceptible New Orleans A. aegypti strain maintained in the insectary of the Entomology Department of the Health Directorate of Iquitos (Dirección de Salud de Iquitos). Following the World Health Organization protocol (WHO, 2006a), two standard cones were firmly fixed at each of three different heights on each tested wall: 0.5, 1.0 and 1.5 m from the floor (six cones per wall), and five female mosquitoes were introduced into each cone using a manual aspirator. Mosquitoes were exposed to the treated surfaces for 1 h and then transferred to holding chambers and maintained with sucrose solution under controlled conditions (27 °C and 90% RH humidity). Knockdown was recorded at 1 h post-exposure (KD60) and mortality was assessed after 24 h. Residual efficacy of insecticides was monitored 1, 4, 8, 12 and 16 weeks after IRS application. The first three monitoring activities were conducted during March–May, which corresponded to the end of the rainy season (September–May), whereas the last two were conducted during June and July, which corresponded to the dry season (June–September). Bioassays were also carried out in the untreated control houses, per the same methodology.

2.5. Entomological surveys

Surveys were conducted in intradomestic and peridomestic environments to determine the following entomological indices: Breteau index (BI; number of A. aegypti-positive containers per 100 houses), house index (HI; percent of houses with A. aegypti-positive containers) and container index (CI; % of water holding containers containing immature A. aegypti). For these indices, surveys were carried out at baseline and 4, 8, 12 and 16 weeks post IRS. The adult index (AI) was calculated as the percentage of houses positive for A. aegypti adults. A. aegypti adults were collected from inside houses using Prokopack aspirators (Vazquez-Prokopec et al., 2009). For adult indices, surveys were carried out at baseline and 1, 4, 8, 12 and 16 weeks post intervention. A shortcoming of the study was that the number of houses included in each post-treatment survey varied: 4 weeks after treatment, n = 35 houses; 8 weeks, n = 36; 12 weeks, n = 35; and 16 weeks, n = 31. At baseline, a total of 57 houses were surveyed, possibly because this survey was carried out during the school holiday season and residents of the neighbourhood were keen to participate.

2.6. Data analysis

As the data of bioassays were not normally distributed, non-parametric Kruskal–Wallis and Mann–Whitney tests were conducted using Statgraphics Plus V. 3.0 software. Differences between groups were considered significant at P < 0.05. Entomological indices were compared with ANOVA, using SPSS (version 19).

3. Results

3.1. Housing and climate conditions

Of an estimated 85 inhabited dwellings in the study site, 54 were included in the initial survey of the area to determine housing characteristics, although not all 54 households provided data for all questions on the housing survey. All 54 houses surveyed had open eaves. Only 41.5% (22/53) of the houses had concrete floors, whereas the remaining 58.5% (31/53) had dirt floors. Although the majority of the houses (68.5%, 37/54) had a sewage system, 24.1% (13/54) used latrines. Open sewage ditches were present in 3.7% (2/54) of the houses. Although 98.1% (52/53) of the households surveyed had piped water, 60.4% (32/53) acknowledged collecting rain in water storage containers. Wood was the preferred fuel used in kitchens (42.6%; 23/54), whereas 25.9% (14/54) used gas and 29.6% (16/54) used both types of fuel.

There was not a high degree of variation in temperature during the study period, with an average mean temperature of 25.7 °C in March that slowly dropped to 25 °C in July. Relative humidity decreased slightly from an average of 84.8 in March to 82.9 in July, while precipitation dropped from approximately 278 mm in March to 80 mm in July (NCDC, 2011).

Regardless of our instructions, a few homeowners applied additional insecticides within the houses. These houses were excluded from results and analysis.

3.2. Residual bioefficacy of deltamethrin on different wall surfaces

During the first twelve weeks after treatment, the number of mosquitoes knocked down 1 h after exposure (KD60) was greater than 85% on all tested surfaces, however, brick walls and painted wood walls had the highest KD60 (>90%; Table 1). At week 16, a dramatic decrease of more than 20% in the number of mosquitoes knocked down after 60 min was observed for all treated surfaces. KD60 in the absence of deltamethrin (control group) was 0% for all tested surfaces at all time points.

Table 1.

Knockdown effect 60 min post-exposure (KD60) and mortality (after 24 h) of A. aegypti exposed to deltamethrin on different wall substrates. Number of mosquitoes exposed for each surface type/week was 360.

Weeks	Surface type
	Brick		Unpainted wood		Painted wood
	KD60 ± SD	Mortality ± SD	KD60 ± SD	Mortality ± SD	N (%)	Mortality ± SD
1	90.2% ± 17.8	96.1% ± 8.7	86.9% ± 13.9	82.8% ± 17.3	80.6% ± 28	86.1% ± 15.6
4	93.6% ± 12.5	88.6% ± 14.2	90.3% ± 12.6	83.9% ± 15.2	89.4% ± 12.9	85% ± 19.3
8	91.9% ± 13.3	81.1% ± 19.5	89.4% ± 16.8	81.9% ± 15.7	95.8% ± 11.1	82.2% ± 18.9
12	93.3% ± 12.1	76.4% ± 23.1	85% ± 14.9	63.9% ± 27.4	91.7% ± 12.9	70.3% ± 28.9
16	61.1% ± 14.6	51.4% ± 21.9	52.7% ± 16.9	55.8% ± 31.1	59.2% ± 17.9	44.2% ± 25.6

SD: Standard deviation.

Statistical analyses show that the knockdown effect was significantly greater for brick walls than unpainted wood walls (p < 0.0002), and for painted walls as compared to unpainted walls (p = 0.02) No significant difference in KD60 was detected when comparing walls made of brick and painted wood.

One week after treatment, mortality recorded 24 h after exposure was 96.1% for mosquitoes exposed to brick walls sprayed with deltamethrin. This was significantly higher than the mortality recorded for painted (86.1%) and unpainted wood (82.8%) (p < 0.05). After 8 weeks, mortality remained greater than 80%, and similar results were obtained for all three surfaces. By week 12, mortality decreased notably on painted wood (from 82.2% to 70.3%) and unpainted wood (from 82.2% to 63.9%); whereas it decreased less for brick walls, where 76.4% were still killed 24 h after exposure. At 12 weeks, significantly more mosquitoes were killed after exposure to treated brick walls than unpainted wood walls (p < 0.05). At the final bioassay, carried out 16 weeks after treatment, mortality was higher for unpainted wood (55.8%) compared to painted wood. Although mortality for unpainted wood was also higher than brick (51.4%), the difference was not statistically significant.

When it occurred, mortality in controls was negligible (3.3%). For this reason no correction factor was applied. No significant differences were detected between bioassays carried out at 0.5, 1 and 1.5 m.

3.3. Entomological surveys

Baseline survey results showed that the study site had a moderate infestation of A. aegypti (BI = 14.3, HI = 8.9, CI = 3.6, AI = 18.5). Although the adult index only dropped slightly from 18.5 at baseline to 16.3 one week after IRS, it then decreased dramatically to a low of 3.1 at 4-weeks post-IRS (p < 0.05) and remained significantly lower than baseline (p < 0.05) 12 weeks after intervention. IRS reduced all immature indices from the first survey through two months post-intervention. Results show a substantial reduction of all immature indices, which remained low 16 weeks after insecticide spraying (Fig. 2). Nevertheless, statistical analyses show no significant differences when compared with baseline. On the other hand, 8 weeks after treatment a significant increase (p < 0.05) was detected in all immature indices in control houses, when compared to treated ones (Fig. 2).

Fig. 2.

Entomological indices before and after the application of deltamethrin S.C. through indoor residual spraying in treated houses (solid lines) and control houses (dashed lines). Gray arrows indicate significant differences in treated houses from baseline (p < 0.05). Dark dashed arrows indicate significant differences (p < 0.05) between treated and control houses.

4. Discussion

When dengue outbreaks are reported in the Peruvian Amazon region, ultra-low volume (ULV) insecticide sprays are typically applied around the houses of cases as the main strategy to reduce adult populations of A. aegypti. However, considering its transient efficacy, its long term impact is limited when used as a single intervention. While there are some reports of peridomestic ULV leading to reductions in entomological indices similar to what we report here, those impacts are not sustained and indices tend to recover within 1–2 weeks (Esu et al., 2010). The failure of peridomestic ULV spraying to sustainably impact on dengue vectors in the Americas has been attributed to the resting behavior of A. aegypti (Perich et al., 2000; Chadee, 2013). As the vector rests mainly indoors, only a fraction of the insecticide sprayed outside reaches indoor resting areas. Among the alternative methodologies to reduce infestations of adult A. aegypti, IRS could be a promising tool, as it targets mosquitoes precisely during their indoor resting periods.

This is the first study that assesses the effect on A. aegypti of deltamethrin S.C. applied through IRS in the Peruvian Amazon, which is a hyperendemic dengue transmission setting. Although housing conditions in the study site could be considered challenging from a vector control perspective (100% of the houses had open roof eaves, 27.8% lacked a sewage system, and 60.4% collected rain-water in open containers), our results indicate that IRS had an immediate impact on reducing A. aegypti populations. The effect of IRS on immature indices was evident from the first week after insecticide spraying and remained low throughout the 16 weeks of follow-up (Fig. 2). Eight weeks post IRS, we found a significant increase of immature indices in control houses which was not observed in treated houses. However, we consider that the number of houses used as controls (3) was insufficient to draw further conclusions.

Due to the labor required to in collecting adult A. aegypti, adult indices are often not used to estimate A. aegypti densities in field studies, despite being a priority target for vector control (Morrison et al., 2008). Since IRS targets adult mosquitoes, we included adult mosquito collections which allowed us to calculate the adult index among our entomological indicators. In this case, AI was not reduced significantly the first week after treatment; but a more substantial reduction was observed from the 4th week until the end of the study period (p < 0.05). This could at least partially result from the fact that IRS would not necessarily have affected any existing immature stages at the time of treatment, which may have led to the emergence of adults that were still detected one week after treatment (Fig. 2). But due to the fact that immature indices had also dropped by 1-week post-treatment, other explanations such as the immigration of mosquitoes from neighboring, untreated areas could also have contributed.

The residual effect of IRS was evaluated through bioassays of treated surfaces using an insecticide-susceptible laboratory strain of A. aegypti. The surfaces that were studied represented the most common materials used in housing construction in the Peruvian Amazon: brick, unpainted wood and painted wood. The immediate effect of the IRS (measured 1 week after spraying) was significantly higher for brick walls than wood walls (p < 0.05), and its residual effect lasted significantly longer (p < 0.05) than on unpainted wood (Table 1). The knockdown effect of the IRS was consistently higher on brick walls than in unpainted wood walls (p < 0.05). This is consistent with results obtained by Etang et al. (2011) in Cameroon, which found that the residual effect of IRS using deltamethrin WG on malaria vectors lasted 26 weeks on brick walls and 15 weeks on wood walls. The presence of open roof eaves to provide ventilation in this tropical environment may negatively affect the residuality of pyrethroid insecticides used in IRS, as insecticide degradation can be triggered by climatic conditions, such as high temperatures and UV light exposure (Sibanda et al., 2011).

Considering the ever-increasing body of evidence that traditional A. aegypti control strategies targeting immature mosquito stages do not provide adequate protection to control and prevent dengue outbreaks, it is essential to look for alternative strategies that target adult mosquitoes (Morrison et al., 2008). Substantial reductions in entomological indices have been reported using insecticide treated materials (ITMs) against dengue vectors, primarily through their application as window curtains, window screens and water storage container covers, although the impact varies depending on coverage and the suitability of housing structures (Kroeger et al., 2006; Lenhart et al., 2013; Vanlerberghe et al., 2013; Manrique-Saide et al., 2015). Insecticide treated bednets have also been reported to reduce A. aegypti infestations, and the authors hypothesized that the reduction could potentially be attributed to the fully extended bednets providing large insecticide treated surface areas inside houses, akin to what occurs when IRS is applied (Lenhart et al., 2008).

There have been very few studies on the efficacy of IRS as control strategy against A. aegypti. In Australia, the targeted spraying of common indoor A. aegypti resting sites with residual insecticides has been shown to reduce dengue transmission risk when coverage is sufficiently high (Ritchie et al., 2002; Vazquez-Prokopec et al., 2010). Rozilawati et al. (2005) did not find outdoor residual spraying to be effective at reducing either A. aegypti or Aedes albopictus, despite the residual effect of the insecticide. This was likely because deltamethrin was only applied outdoors whereas in our study, IRS was applied both outdoors and indoors.

Although the shortcomings of the routine use of IRS for dengue vector control include practical difficulties such as gaining access to homes as well as time, human, and financial resource limitations (Morrison et al., 2008), its role as part of an integrated vector management (IVM) strategy should be considered.

Furthermore, the fact that A. aegypti rest for long periods inside houses (Chadee, 2013) reinforces the idea of IRS as a plausible tool for dengue vector control. As it has been established for malaria, IRS could be carried out as part of an IVM approach, (WHO, 2015). IVM would be particularly important in an area like Iquitos, which is endemic to more than one vector-borne disease.

As dengue continues to increase as a global public health problem, novel vector control strategies are increasingly needed to suppress mosquito populations and control dengue transmission. Herein, we present compelling data that suggest that in certain circumstances, IRS may be an effective and appropriate tool in reducing A. aegypti infestations at the house level.