In this study we investigated the survival of wild-caught mosquitoes resting in homes in the early morning. These endophilic, likely endophagic, vectors were exposed to differing levels of both ITN use and prevalence of malaria infected hosts [10]. To study the impact of these exposures on wild vector populations, we reared them in cages for seven days and recorded daily mortality. We found that most female anopheles survived to the seventh day. Survival differed across species and was inversely correlated with high ITN exposure in the household, but not oocyst development.
We identified a diverse vector population dominated by two primary vectors, An. gambiae s.s. and An. funestus. Major vector species had similar survival rates, while minor species exhibited poorer survival following collection. Oocyst infection rates were high for both major vectors, and was nearly 8% in those surviving to day 7. This is similar to what was found in western Kenya closer to Lake Victoria and in our previous work in the study area [11]. If we assume an infection prevalence of 30% in the human hosts [Sumner], uniform biting rates, and that 80% of anopheles have fed in the last 24 hours [10] then we estimate that 1 in 3 bites on an infected host must be infectious to achieve an 8% oocyst rate.
Regarding the relationship between survivorship and P. falciparum oocyst development in malaria vector mosquitoes, we found that infections acquired on or near the day of capture had no effect on survival up to seven days. Very few studies have assessed how infected or uninfected mosquitoes differ in survivorship, especially for wild caught mosquitoes. Most studies have been conducted on mosquitoes reared and infected under laboratory conditions. Significant reduction in survival following infection is limited to studies with combinations of vectors and plasmodium species that are not known to occur naturally [12]. The study by Chege and Beier [13] examined the effect of malaria parasites on the longevity of wild-caught, naturally infected anopheles. The results from their study are consistent with our findings that infection status does not affect survivorship. However, in their study they observed higher survival rates and lower infection rates among An. funestus compared to An. gambaie s.l..
Leveraging detailed ITN use data recorded at every mosquito collection time point, we were able to correlate vector survival with ITN coverage at the household level. From our study we found that mosquitoes collected from households that had more than 90% of its members sleeping under a net the night before mosquito collection had a significantly lower survival rate. However, this effect was not observed at lower ITN coverage. In an experimental hut system in Burkina Faso, there was no difference in long-term survival of wild-entry anopheles when they entered a hut with an untreated net or a pyrethroid-treated net [14]. In experimental studies, it has been shown that insecticide-susceptible anopheles can feed across an ITN when the bloodmeal is touching the net. [15] Although feeding time and bloodmeals are smaller, only 15% of fed mosquitoes die in the 24 hours post-feeding indicating that the bloodmeal taken across the net may be protective against mortality. Given that 80% of the anopheles collected in this cohort show evidence of recent feeding [10] feeding success could partially explain the low mortality after foraging in a household with high ITN coverage, although it is unlikely that the majority of available hosts are touching their ITN while sleeping. The observation that feeding reduces mortality from insecticides has been reported elsewhere [16, 17] but whether this is cause (the bloodmeal reduces susceptibility to the chemical) or effect (insects are more successful at feeding if they are less sensitive to the chemical) is difficult to untangle in a natural system such as ours. It is also possible that insects which died quickly after exposure to ITNs were not aspirated from walls and the collected mosquitoes represent a surviving subset of all foraging mosquitoes. Nonetheless, our findings are consistent with epidemiological studies which show that ITNs reduce incidence of malaria in areas with populations of resistant anopheles, but protection is incomplete and transmission persists. [18]
The impact of ITN coverage on survival differed by species. While An. gambiae s.s. survival was sensitive to ITN coverage, An. funestus survival did not decline with increasing ITN coverage. This could be due to differences in biting behaviors, such as late evening or early morning biting by An. funestus, that reduce exposure to ITNs during feeding or it could be due to higher levels of pyretheroid resistance in An. funestus compared to An. gambiae s.s. which has been reported elsewhere [19]. Overall, the 48-hour mortality rate after foraging and resting in a household with 80 or 90% ITN coverage was less than 25% and differences in survival of An. gambiae s.s. exposed to high or low ITN coverage were not noticeable until after day 3. It is unclear whether these small differences in day 7 survival as a function of ITN coverage would reduce overall transmission [3].
Limitations to our study include unknown effects of collection technique on mosquito fitness and survival as well as unknown age and life history of the wild caught mosquitoes. We observed that mosquito vectors had slightly higher mortality in the first few days compared to days 3–7, which might be due to damage during mechanical aspiration of mosquitoes or poor adaptation to cages. Although we reduced the aspiration force of the prokopacks using a bespoke attachment which improved survival, they may still suffer damage during collection. In addition, since it is impossible to determine the ages of the harvested mosquitoes and it is possible that older mosquitoes were more vulnerable to the trauma of aspiration, early mortality may be biased towards older insects. This is supported by the observation that infection rates were higher in mosquitoes that died on the same day they were collected; these mosquitoes must have already survived a minimum of 5–7 days to have been infected on the day of collection. However, since our study was conducted over 24 months, we expect that we collected a representative age distribution of vectors across the study which increases the generalizability of our findings.
In conclusion, we report high infection rates among wild-caught vectors. Early stages of parasite infection up to the development of oocysts does not appear to influence survival for the two major vectors identified in this study. In contrast to An. funestus, An. gambiae s.s. exhibits increased mortality when collected from households with higher ITN coverage, although the mortality rate is still lower than would be expected. Rearing wild-caught mosquitoes gives unique insights into exposures correlated with vector survival and development of infectivity.