Comparison of field-based xenodiagnosis and direct membrane feeding assays for evaluating host infectiousness to malaria vector Anopheles gambiae.

Several techniques are currently being used to study host infectiousness to mosquitoes, including the experimental possibility of laboratory reared mosquitoes acquiring infections through membrane feeders or directly on host skin. Here, the relative performance of the laboratory-based membrane feeding method (DMFA) and the field-based xenodiagnosis (XD) of malaria infectious hosts using wild Anopheles mosquitoes were compared. A cross-sectional survey involving a sample of 70 children (aged 3-12 years) living in a malaria endemic area in Western Burkina Faso, was carried out to measure their infectiousness to Anopheles mosquitoes using two approaches. The first approach used the xenodiagnostic procedure in which children were exposed to mosquito bites overnight, being sleeping individually in different sentinel huts from 6 pm to 6 am (4 nights per child). Anopheles sp that had acquired blood-meal on each child were subsequently collected early in the morning, and examined for Plasmodium falciparum oocyst infection on day 7 post-feeding. In the second approach, the infectiousness of the same children was estimated by whole-blood membrane feeding procedure using F0 An. gambiae s.l. that emerged from field-collected larvae cohorts. In the DMFA, 41.4% of the children successfully infected at least one mosquito with the mean oocyst prevalence of only 4.6±1.1% in the 2171 mosquitoes that were examined (mean oocyst intensity: 2.0±(std error of mean) 0.3 oocysts per infected midgut). Comparatively 78.6% of children yielded oocysts infection in mosquitoes during the XD approach (Chi square=20.11, df=1; p<0.001), with a mean rate of 19.6±2.0 in the 3752 wild caught mosquitoes (mean intensity: 3.93±0.2 oocysts per infected mosquito). The DMFA failed to reveal a portion (n=26) of infectious individuals that were sharply evidenced by the XD, particularly at low gametocyte densities or at levels that could not be detected by the classical microscopic examination of blood smears. As opposed to the resource consuming DMFA, which is often mined by technical constraints, using the XD method could be an advantage in experimental investigations of host infectiousness in areas where anopheline species cannot be conveniently reared for the experimental studies. Ethical aspects of this approach, mainly related to exposure of the human subjects to potentially infectious mosquito bites are discussed.