Use of Insecticide-Treated House Screens to Reduce Infestations of Dengue Virus Vectors, Mexico

Dengue prevention efforts rely on control of virus vectors. We investigated use of insecticide-treated screens permanently affixed to windows and doors in Mexico and found that the screens significantly reduced infestations of Aedes aegypti mosquitoes in treated houses. Our findings demonstrate the value of this method for dengue virus vector control.


DISPATCHES
Dengue prevention efforts rely on control of virus vectors. We investigated use of insecticide-treated screens permanently affixed to windows and doors in Mexico and found that the screens significantly reduced infestations of Aedes aegypti mosquitoes in treated houses. Our findings demonstrate the value of this method for dengue virus vector control. V ector control is the primary method for prevention and control of the increasingly frequent dengue outbreaks that threaten more than half the global human population (1). Existing approaches target breeding sites or attack adult mosquitoes by insecticide space-spraying, but these methods, at best, offer only immediate solutions and are rarely effective or sustainable for the long term (2). Methods that target the largely endophilic adult female Aedes aegypti mosquito vectors within buildings where they rest and bloodfeed have greater potential for sustained results and acceptance at the community level. One such method, long-lasting insecticidal-net (LLIN) curtains hung at windows or doors, can greatly reduce vector populations at high coverage rates (3)(4)(5), but efforts are compromised when curtains remain open during daytime or when all house entry points cannot be protected (6,7). Fixed or permanent screens covering doors and windows could eliminate this problem. Mosquito-proofing of houses is effective in malaria control (8), and reduced risk for dengue has been associated with the use of untreated (9) and insecticide-treated (3,10) screens.
We used a cluster-randomized sampling design constructed on the basis of earlier studies (4-6,11) to select 20 clusters (10 treatment, 10 control; 100 households/cluster) from a possible 30 clusters by using digital maps (Google Earth software; Google Inc., Mountain View, CA, USA) ( Figure 1). Sample size was determined by using a 2-level hierarchical model to achieve 80% power at a 5% level of significance. Thus, for a negative binomial distribution with a dispersion coefficient of 0.02 and intracluster coefficient of 0.05, a minimum of 8.9 clusters/arm were required. Written informed consent was obtained from participating households; the WHO Ethical Review Committee (WHO reference no. 2010/82951-0, unit reference no. A90297) and Guerrero State Ministry of Health granted ethical permission for the study.
Participating households in the treatment arm were instructed on LLIS maintenance during installation (April-December 2012). Control houses received no treatment. Indoor resting adult mosquitoes were collected by using modified CDC backpack aspirators (John W. Hock Co., Gainesville, FL, USA) from all houses in a cluster on the same day during 9 am-3 pm. Indices for Ae. aegypti mosquitoes (the only Aedes species found) were calculated to quantify house infestation (percent of all houses positive) and infestation density (numbers per infested house) for all mosquitoes, all females, all blood-fed females, and males.
For presence-absence data, we performed logistic regression models with a single predictor variable identifying houses with LLIS and control houses (coded as 1 and 0, respectively) and accounting for each house membership in a given sampling cluster (cluster-robust SE calculation). Odds ratios (ORs) and 95% CIs indicating the effect of LLIS on each entomologic indicator were calculated. Overdispersed index data were compared between arms by using the Mann-Whitney U test. The effect of treatment on each metric was analyzed by negative-binomial regression using, as with the logistic models, treatment as the sole predictor variable (1 and 0 coding). Negative binomial models also accounted for membership of a house in a sampling cluster (cluster-robust SE calculation). ORs and incidence rate ratios (IRRs) were calculated with 95% CIs; significance was set at p<0.05. Analyses were performed by using Stata 12.0 (StataCorp, College Station, TX, USA).

Conclusions
In our study, the entomologic effect of LLIS was greater than that detected in a recent study of deltamethrin-treated window curtains (12), in which a 27% reduction of adult Ae. aegypti mosquitoes was only sustained for a short time after curtain installation. Other studies of insecticide-treated curtains in Latin America have reported entomologic effects by using immature stage indicators alone (4,5,7). Whether these reductions were sufficient to affect dengue transmission is unknown, and the overall effect on dengue infections remains to be evaluated.
Our results are encouraging in view of high levels of insecticide resistance in Ae. aegypti mosquitoes in Acapulco. Although resistance to α-cypermethrin has yet to be reported in Guerrero, high frequencies of mutations in the voltage-gated sodium channel gene, which is associated with pyrethroid resistance in Ae. aegypti mosquitoes, have been reported (13). If insecticide resistance began to reduce the efficacy of the method we describe, the screens could be treated with different insecticide classes.
We found the use of LLIS was a popular intervention, and perceived efficacy was reinforced by a reduction in other domestic pests ( Figure 1) (14). The likely effects on other peridomestic disease vectors could promote increased adoption of the intervention with additional cost benefits. The polyethylene netting was durable on windows; it was often damaged on the lower sections of doors (14) but readily repaired by reinforcement with metal mesh.
Dengue vector control programs using house screens are ongoing in selected cities in Mexico and Brazil. These results were obtained during an exploratory phase of that initiative. Stakeholders in other countries may also consider evaluating this novel approach for dengue vector control.