Forced exposure to the ITN resulted in negligible mortality of highly resistant mosquitoes within 24h regardless of exposure period (note this contrasts to 100 % mortality of the susceptible Kisumu strain in pilot studies referred to in the methods) (Fig. 1). Furthermore, 84.7 (95 % Confidence interval (CI): 78.1 to 89.9) % of mosquitoes that were provided access to a blood meal were able to feed, irrespective of the duration of the exposure (χ2 = 1.32, df = 1, p = 0.25). Blood feeding increased mosquito longevity by approximately 4 days, with mean (± SE) survival time (post exposure at 4 days old) of blood-fed mosquitoes of 14.7 ± 0.59 days post exposure, compared with 10.5 ± 1.04 days post exposure for unfed mosquitoes having access to the arm and 10.7 ± 0.39 days post exposure for unfed mosquitoes having no access to the arm (χ2 = 29.80, df = 1, p < 0.001) (Fig. 1). When exposed for 1 to 3 min, fed mosquitoes lived an average of 2.8 more days post exposure than unfed mosquitoes [unfed: 11.1 ± 0.47 days; fed: 13.9 ± 0.59 days] and when exposed for 5 min, fed mosquitoes lived another 6.5 more days post exposure [unfed: 9.8 ± 0.58 days; fed: 16.3 ± 1.36 days] (χ2 = 14.64, df = 1, p < 0.001). While the interaction between exposure time and blood-feeding was significant on longevity, there was no significant effect of exposure duration itself (χ2 = 0.56, df = 1, p = 0.45). Subgroup analysis showed no influence of exposure duration on mean longevity for unfed mosquitoes (χ2 = 3.20, df = 1, p = 0.07). However, longer exposure to insecticide led to a longer life post-exposure for blood fed mosquitoes [13.9 ± 0.59 days for 1- and 3-min exposure and 16.3 ± 1.36 days for 5-min exposure] (χ2 = 5.78, df = 1, p = 0.02).
Two additional experimental blocks, in which the mosquitoes were all only exposed for 5 min corroborated these results. Combining all experimental blocks for the 5 min exposure showed a blood meal to extend the lifespan of mosquitoes by around 7 days post exposure [9.0 ± 0.73 days for females with no access to blood source; 9.7 ± 0.43 days for those with access to the blood source but unfed and 16.6 ± 0.87 days those with access to the blood source and fed] (χ2 = 51.03, df = 1, p < 0.001).
b. Variable exposure via individual feeding choice
The presence of insecticide reduced average contact time with the netting (F = 21.97, df = 2, p < 0.001; Fig. 2.a). Mosquitoes exposed to the UTN had an average contact time of 167.7 ± 13.06 sec, while those exposed to the ITN or ITN + PBO had average contact times of 121.4 ± 9.5 and 59.1 ± 9.25 sec, respectively (Tukey pairwise comparisons: pITN-UTN = 0.005, pITN+PBO-UTN and pITN+PBO-ITN < 0.001).
The feeding duration showed a similar pattern between net types (F = 45.30, df = 2, p < 0.001; Fig. 2.b) with average times of 219.3 ± 8.82 sec, 135.4 ± 9.80 sec and 105.0 ± 15.85 sec for UTN, ITN, ITN + PBO, respectively (Tukey pairwise comparisons: pITN-UTN and pITN+PBO-UTN < 0.001, pITN+PBO-ITN = 0.48). Blood feeding duration was longer for longer net contact time (F = 72.21, df = 1, p < 0.001), independently of the net treatment (F = 2.36, df = 2, p = 0.10).
The presence of insecticide reduced the percentage of mosquitoes that fed successfully. With the UTN, 60.5 (CI: 49.3 to 70.8) % of mosquitoes took a blood meal, while blood feeding rates were only 38.9 (CI: 29.1 to 49.5) % and 9.2 (CI: 3.46 to 19.0) % with the ITN and ITN + PBO treatments, respectively (χ2 = 45.70, df = 2, p < 0.001; Fig. 2.c). Blood fed females spent 3.6 more time in contact with nets than unfed ones [unfed = 59.8 ± 6.22 sec; fed = 218.7 ± 7.46 sec] (χ2 = 110.22, df = 1, p < 0.001). Contact times of the subset of mosquitoes that successfully took a blood meal showed a similar pattern between net types again with average contact times of 253.4 ± 7.49 sec, 185.4 ± 11.73 sec and 122.3 ± 13.68 sec for UTN, ITN, ITN + PBO, respectively (Tukey pairwise comparisons: pITN-UTN and pITN+PBO-UTN < 0.001, and pITN+PBO-ITN = 0.29). However, unfed mosquitoes exposed to an ITN spent more time in contact with the net compared to those exposed to an UTN with average time of 36.5 ± 10.82 sec for UTN vs. 80.6 ± 10.66 sec for ITN and 52.7 ± 9.73 sec for ITN + PBO (F = 11.97, df = 2, p = 0.002; Tukey pairwise comparisons: pITN-UTN = 0.04 and pITN+PBO-UTN = 0.88, and pITN+PBO-ITN = 0.24). In one experimental block, the feeding success was slightly lower compared to the other blocks (F = 5.63, df = 1, p = 0.02). Thus, while there is no difference in the time spent on the net for blood fed mosquitoes, there is some variability between two experimental blocks for the contact time of unfed mosquitoes (F = 7.92, df = 1, p = 0.005).
As observed in the first experiment, there was negligible mortality within 24h in the UTN and ITN treatments. The ITN + PBO treatment, however, caused substantial 24h mortality of 86.1 (CI: 75.3 to 93.5) %. Beyond the instantaneous effects, insecticide exposure led to a reduction in long term survival (χ2 = 146.87, df = 2, p < 0.001) (Fig. 3). Mosquitoes exposed to an UTN had an average survival time (post exposure at 4–5 days old) of 16.7 ± 0.74 days, those exposed to the ITN 11.5 ± 0.62 days, and those exposed to the ITN + PBO just 2.3 ± 0.44 days. Blood feeding increased overall longevity (χ2 = 24.53, df = 1, p < 0.001), by approximately 6 days for the UTN and 4 days for the ITN. Nonetheless, blood fed females died more quickly after an exposure to insecticide than those exposed to a UTN [14.0 ± 1.10 days for the ITN vs. 19.0 ± 1.00 days for the UTN]. For the ITN + PBO, blood fed mosquitoes had marginally shorter lifespan than non-blood feds (average survival time post exposure of 1.0 ± 0 days and 2.5 ± 0.49 days, respectively (χ2 = 10.21, df = 2, p = 0.006). The time that blood-fed mosquitoes spent on the net and feeding duration did not influence longevity. However, fed mosquitoes exposed to an UTN had a longer life when they spend more time blood feeding which is not the case for fed mosquitoes exposed to insecticide (χ2 = 6.95, df = 2, p = 0.03).
One additional experimental block was added to an analysis comparing the effects of the ITN against the UTN only (providing 4 blocks in total for this comparison (SI, Table A)). The presence of insecticide reduced the mean time spent on the net [119.6 ± 7.73 sec for the ITN vs. 173.3 ± 10.11 sec for the UTN] (F = 18,02, df = 1, p < 0.001). The ITN led to a significant reduction in blood feeding (χ2 = 17.20, df = 1, p < 0.001), with 38.2 (CI: 30.6 to 46.3) % blood fed mosquitoes in the ITN treatment compared with 61.9 (CI: 53.6 to 69.8) % in the UTN. Mosquitoes that spent a longer period on the net were proportionally more successful in taking a blood meal (χ2 = 206.56, df = 1, p < 0.001). When mosquitoes spent less than 1 min on the net, the feeding rate did not differ between ITN and UTN. However, once contact time exceeded 1 min, blood feeding increased with contact time for the UTN but did not for the ITN (χ2 = 11.31, df = 1, p < 0.001). There was an overall effect of the insecticide exposure on longevity (χ2 = 29.34, df = 1, p < 0.001). The mean survival time of unfed mosquitoes was 14.0 ± 0.83 days post exposure for the UTN and 10.0 ± 0.50 days post exposure for the ITN, while for fed mosquitoes it was 18.2 ± 0.78 days post exposure and 15.0 ± 0.88 days post exposure, respectively (χ2 = 42.09, df = 1, p < 0.001).