Spatial repellents and malaria transmission in an endemic area of Cambodia with high mosquito net usage

Abstract

Background: The spread of artemisinin resistant malaria from SE Asia to the rest of the world remains a threat that will only be ended by eliminating malaria from the region. Novel control approaches are required to mitigate this threat. Spatial repellents (SR) are one such approach. We therefore conducted a multiple cross-over experiment from April 2013 - April 2014, in which all houses in one of two villages in Mondolkiri Province, Cambodia were alternately supplied with an emanator of the spatial repellent metofluthrin per 30 m³ of protected area to cover all potential peridomestic areas where people might spend their time before sleeping. Emanators were replaced every month for a three-month period.

Material and methods: Mosquito densities were simultaneously monitored in each village for two weeks every month using six CDC light-traps/night run from 18.00 to 07.00 hrs inside bedrooms and malaria prevalence, seroconversion and gSG6 protein rates assessed from prevalence surveys. After emanators were installed in the first village they were installed in the second village for a further three-month period and following that were again used in the initial village for a further three months. Surveys were undertaken before the initial installation of the emanators and at each cross-over point.

Results: Anopheles dirus densities were highest in houses closest to the forest. Transmission rates were low even before the application of the emanators. Perhaps due to the low levels of malaria transmission in Mondolkiri no significant relationships were found in Plasmodium cases or seroconversion rates between villages, surveys or by intervention. Adult males, who might spend more time unprotected in the forest at night, appeared to be at greater risk of becoming infected with P. falciparum malaria as compared to women or young children.

Conclusion: At the malaria transmission levels present in Mondolkiri the metofluthrin emanators evaluated had no observable effect on malaria prevalence. This may be due to confounding by low prevalence rates.

Figure 1

Species of anophelines collected from Ou Chra in rank order (on a log scale). A – CDC light-traps inside houses, B– CDC light-traps over a pigsty, C– from an MMX trap. Note the preponderance of An. dirus (in red) inside houses compared to the other trap types.

Figure 2

Mean numbers of An. dirus collected in CDC light-traps indoors per three-month period by village when metofluthrin emanators were or were not present in the village. (The asterisk indicates that metofluthrin was present in the village).

Figure 3

Box plot describing the range of normalised optical densities of antibody responses to PfAMA-1 (green), PfMSP-119 (orange), PvAMA-1 (blue) and PvMSP-119 (red). Red boxes denote the application of the intervention.

Figure 4

Box plots describing the range of corrected optical densities (OD minus the mean of two blank wells) of antibody responses to the gSG6 salivary protein. Red boxes denote the application of the intervention.

Figure 5

Reverse cumulative distribution plots of antibody responses to gSG6 salivary protein and PvMSP-119 at baseline (green line) and post intervention (red line). The graphs are constructed by plotting on the Y axis the proportion of subjects having an OD equal to or greater the level shown at each point along the x axis.

Figure 6

Profile likelihood plot (PLP) for both Pf antigens for Ou Chra and Pu Cha. PLP show the log likelihood of a catalytic conversion model allowing for a change in transmission occurring at iterative years. The maximum log likelihood is the time point at which a change in transmission is most likely to have occurred. The plot suggests a change in transmission occurred approximately 13 years ago.

Figure 7

Anti-malaria seroconversion curve for antibodies to either PfAMA-1 or PfMSP-119 for participants in Ou Chra and Pu Cha across all surveys. A reversible catalytic model was fitted to age adjusted data using maximum likelihood to create seroconversion curves. Solid lines represent the fitted probability for being positive, dotted lines represent the 95% confidence interval for these fits, and triangles represent the observed proportion of seropositives per age decile. Seroconversion rates (SCR; λ) are presented on the graph. Change points were identified using profile likelihood plots. The plots identified a change point of 15 years for Ou Chra and 13 years for Pu Cha, models allowing for a change in SCR were significant (Likelihood ratio test: <0.05) compared to the model that did not allow a change.

Location of Ou Chra and Pu Cha villages (Mondolkiri Province, Cambodia)

Ou Chra village with georeferenced houses

Pu Cha village with georeferenced houses