The role of light in Chagas disease infection risk in Colombia

Abstract

Background: Chagas disease is the most important vector-borne disease in Latin America and Rhodnius prolixus is the main vector in Colombia. Control strategies in this region have shown poor outcomes due to the insect's ability to disperse between the sylvatic and the domestic habitat. Because insect migration to houses is responsible to sustain contact rates between vectors and humans, understanding the risk factors that promote migration could be important in designing control strategies. In this respect, it has been reported that adult triatomines have the ability to move over long ranges at night attracted by artificial light. Thus, light bulbs could be playing a critical role in house invasion. The main objective of this study is to understand the role of artificial light, or simply light, in house infestation by R. prolixus.

Methods: To investigate the role of light, we combined fieldwork in the village of Chavinave, Casanare, Colombia and a mathematical model of Rhodnius prolixus dynamics. The model allowed us to simulate insect mobility and distribution in the village based on field results. We created 11 scenarios representing different amounts of light in the village (from 0 to 100%, with increments of 10%) with 100 simulations each for a time of 1000 days (2.7 years) and compare the results between the scenarios.

Results: None of the Gomez-Nuñez traps were positive at any stage of the study, suggesting that insects do not colonize houses. The model predicts that with current village connections the proportion of houses that have visiting insects should be around 98%. Additionally we showed that an increase in light allows for insect spreading and migration to previously un-infested areas.

Conclusions: Increments in light could increase the contact rates between vectors and humans; a two-fold increase in human cases for a 30% increase in the use and visibility of light on this particular village was estimated with the model.

Fig. 1

Study site. Chavinave is located in the municipality of Mani, Casanare (Colombia) and it has an area of 32 Ha. The village is adjacent to the Cusiana River (172 m.a.s.l.) and the landscape is characterized by Savanna and Gallery Forest. Houses and palms are shown at their exact location and are represented by the house and circle icons respectively. We developed a mathematical model that assumed houses and palms to be patches that contain meta-populations. The model consisted of an age structured population dynamics coupled with insect migration dependent on light distribution. The size of the patch is proportional to visiting index at every patch. The network shown in the figure summarizes average model output after 100 simulations run until steady state. Initial conditions assumed 20 insects per palm: 5 eggs, 5 nymphs and 10 adults and 0 insects per house. The simulation time was set to 1000 days (2.7 years)

Fig. 2

Epidemiological Indices I. Model outputs included: the average number of houses visited by insects divided by the total number of houses (proportion of infested houses - PIH) and the average number of insects per house per day or visiting index (VI). The mean and standard deviations at steady state for PIH and VI were calculated for every simulation while varying the proportion of patches that can actually see a house with light bulb at night (light on the x axis). Note that PIH exhibits a saturation effect. On the contrary, VI has a quasi-linear relation. In addition, for every light scenario we recorded network metrics: average path length (green), proportion of patches in the biggest cluster (blue) and number of clusters (red). We observed in the field that Chavinave has 274 connections out of 423 (64 %) possible. Thus, with the model we predict that Chavinave has a PIH of 0.98 and a VI of approximately 2 insects

Fig. 3

Epidemiological indexes II. Model output also included the average numbers of insects per day moving from palms to houses (FPH) and between houses (FHH) (y axis). Average fluxes and standard deviations were computed at steady state for every set of simulation for a particular value of light (x axis). Fluxes from houses to palms (FHP) were not observed in the simulations. Network metrics, as described in Fig. 2, are shown with solid colored lines with their corresponding standard deviations. For Chavinave the model predicts an FPH of 3.5 insects per day and an FHH of about 6.2 insects per day

Fig. 4

Predicted number of cases as function of light. The incidence, or number of new cases per person per year, was calculated based on the obtained VI, PIH, the proportion of infected insects [41], insect biting rate [51], feeding rate on humans [52] and probability of transmission per contact with an infected Triatomine [53]. For Chavinave, we estimate an incidence of 6.3 cases per 1000 people per year (see text for calculation). Using the current approximate population for Chavinave (122 people), the model predicts 1 new case every 18 months