The impact of livestock on the abundance, resting behaviour and sporozoite rate of malaria vectors in southern Tanzania

Background: Increases in the coverage of long-lasting insecticidal nets (LLINs) have significantly reduced the abundance of Anopheles gambiae sensu stricto in several African settings, leaving its more zoophagic sibling species Anopheles arabiensis as the primary vector. This study investigated the impact of livestock ownership at the household level on the ecology and malaria infection rate of vectors in an area of Tanzania where An. arabiensis accounts for most malaria transmission. Methods: Mosquito vectors were collected resting inside houses, animal sheds and in outdoor resting boxes at households with and without livestock over three years in ten villages of the Kilombero Valley, Tanzania. Additionally, the abundance and sporozoite rate of vectors attempting to bite indoors at these households was assessed as an index of malaria exposure. Results: The mean abundance of An. gambiae s.l biting indoors was similar at houses with and without livestock. In all years but one, the relative proportion of An. arabiensis within the An. gambiae s.l species complex was higher at households with livestock. Livestock presence had a significant impact on malaria vector feeding and resting behaviour. Anopheles arabiensis were generally found resting in cattle sheds where livestock were present, and inside houses when absent. Correspondingly, the human blood index of An. arabiensis and An. funestus s.l was significant reduced at households with livestock, whereas that of An. gambiae s.s. was unaffected. Whilst there was some evidence that sporozoite rates within the indoor-biting An. gambiae s.l population was significantly reduced at households with livestock, the significance of this effect varied depending on how background spatial variation was accounted for. Conclusions: These results confirm that the presence of cattle at the household level can significantly alter the local species composition, feeding and resting behaviour of malaria vectors. However, the net impact of this livestock-associated variation in mosquito ecology on malaria exposure risk was unclear. Further investigation is required to distinguish whether the apparently lower sporozoite rates observed in An. gambiae s.l at households with livestock is really a direct effect of cattle presence, or an indirect consequence of reduced risk within areas where livestock keepers choose to live.