Outdoor malaria transmission in forested villages of Cambodia

Abstract

Background: Despite progress in malaria control, malaria remains an important public health concern in Cambodia, mostly linked to forested areas. Large-scale vector control interventions in Cambodia are based on the free distribution of long-lasting insecticidal nets (LLINs), targeting indoor- and late-biting malaria vectors only. The present study evaluated the vector density, early biting activity and malaria transmission of outdoor-biting malaria vectors in two forested regions in Cambodia.

Methods: In 2005 two entomological surveys were conducted in 12 villages and their related forest plots in the east and west of Cambodia. Mosquitoes were collected outdoors by human landing collections and subjected to enzyme-linked immunosorbent assay (ELISA) to detect Plasmodium sporozoites after morphological identification. Blood samples were collected in the same villages for serological analyses. Collected data were analysed by the classification and regression tree (CART) method and linear regression analysis.

Results: A total of 11,826 anophelines were recorded landing in 787 man-night collections. The majority (82.9%) were the known primary and secondary vectors. Most of the variability in vector densities and early biting rates was explained by geographical factors, mainly at village level. Vector densities were similar between forest and village sites. Based on ELISA results, 29% out of 17 Plasmodium-positive bites occurred before sleeping time, and 65% in the forest plots. The entomological inoculation rates of survey 1 were important predictors of the respective seroconversion rates in survey 2, whereas the mosquito densities were not.

Discussion: In Cambodia, outdoor malaria transmission in villages and forest plots is important. In this context, deforestation might result in lower densities of the primary vectors, but also in higher densities of secondary vectors invading deforested areas. Moreover, higher accessibility of the forest could result in a higher man-vector contact. Therefore, additional vector control measures should be developed to target outdoor- and early-biting vectors.

Figure 1

Overview of Cambodia with the four districts indicated. Pailin and Pursat are located in the west, O’chum and Borkeo are located in the north-east. (Google Maps).

Figure 2

Meteorological data in Ratanakkiri province (A), Pursat province (B) and Pailin province (C) in 2005. This includes the monthly rainfall in mm (left axis) and the minimum and maximum temperature in degrees Celsius (right axis). The timing of survey 1 (S1) and survey 2 (S2) are indicated by arrows.

Figure 3

Average man biting rate of the most abundant anopheline species in the different districts. Surveys (S1 = survey 1, S2 = survey 2) and collection sites (Fo = forest, Vi = village).

Figure 4

Early biting rate for the different vector species complexes per collection site. (F = forest, V = village), districts (B = Borkeo, O = O’Chum, P = Pailin, V = Pursat), and region (E = east, W = west).

Figure 5

Regression tree representing the important determinants for Anopheles dirus s.l. early biting proportion (EBP), expressed as percentage of vectors biting before 22.00. The selected splitter variables (village, survey, district, site) are shown in the nodes.

Figure 6

Cumulative number of infected Anopheles mosquitoes (n = 17) during the night outdoors (all data pooled). The approximate time for bed net use is between 22.00 and 05.00 (marked by arrows).