Effects of permethrin-impregnated bednets on Anopheles gambiae s.l. in the Gambia.

A National Impregnated Bednet Programme (NIBP) was implemented in The Gambia covering all the villages with a primary health centre (PHC). The evaluation of the programme included epidemiology, entomology and social aspects. This thesis is part o f the entomological evaluation. Populations of Anopheles gambiae s.l. were found to be as long lived, abundant and infective in villages with permethrin-impregnated bednets as in villages with untreated bednets, confirming the lack of a "mass-killing effect", and the absence of a repellent effect of treated bednets at village level. Movements of mosquitoes between PHC villages was assessed and found that overall, 17% (C.I. 11.02%-26.71%) were immigrant mosquitoes from neighbouring villages. This amount of movements between villages may explain the difference between The Gambia and other African countries in which a clear "mass-killing effect" has been seen after the introduction of impregnated bednets. In the absence of a "mass-killing effect" the question arises of how protection against malaria in children is therefore achieved. Studies on the biting behaviour of the vector population were made to see if mosquitoes were diverted to bite outdoors, shifted the biting time, were prevented or delayed to feed or were diverted to bite other hosts rather than children. No evidence was found for a difference between treated and untreated villages in biting location ( Indoors:Outdoors), mean biting time or human blood index of indoor resting mosquitoes. An indication was observed that the gonotrophic cycle length of An. gambiae s.l. was 2 days in untreated villages. No evidence of a change in the gonotrophic cycle length was found in the presence of treated bednets, although the number of mosquitoes were low to be conclusive. The density of indoor resting mosquitoes was significantly lower in the presence of treated bednets than in the presence of untreated ones, difference probably due to the excito-repellent effect of the insecticide. A study in which mosquito bloodmeals from children could be differentiated from bloodmeals from other hosts (adults and animals) was undertaken. A significant reduction in the proportion of bloodmeals that were apparently taken from children was found in the presence of treated bednets. However, it is not clear whether these results are reliable. Assuming that they are, this could explain the way in which children are protected against malaria by the use of treated bednets. A comparative study was carried out in two villages to see if entomological factors were part of the reasons that caused the NIBP not to be as effective in Zone 5 as in the other Zones. Some entomological differences were found between Zones 5 and 3 and those included: a higher proportion of An. arabiensis in Zone 5 (40%) than in Zone 3 (20%), a significantly higher frequency of the inversions j and d in the chromosome 2R in An. gambiae s.s.. an exophilic tendency of fed females in Zone 5 and lower densities but higher sporozoite rates, resulting in a higher EIR in Zone 5 than in Zone 3. Other parameters evaluated showed no significant difference including biting cycle, Indoor:Outdoor ratios, persistence of the insecticide in the bednets and no evidence was found that the treatment had a differential effect regarding the type of rooms. Levels of bednet usage were probably more significant than these entomological parameters as a cause of the observed differences between zones of the epidemiological impact of treated bednets. The relative sampling efficiency of human landing collections were compared with light traps, exit traps and pyrethrum spray catches. In general there were few significant correlations between methods, due probably to night to night and house to house vernation and a limited sampling in both duration and geographical extension. DNA probes were compared with cytogenetics for the determination of the species of the An. gambiae complex. For the identification of An. gambiae s.s. and An. arabiensis DNA probes were shown to be as reliable as cytogenetics. However, a limitation was the variability between batches of probes. Also, the cross-reaction between the probes pAngss and pAnM14 made difficult the differentiation of An. gambiae s.s. and An. melas.