Residual efficacy of Actellic® 300CS (pirimiphos-methyl), 2015-17
Overall results shortly after spraying showed that the quality of spraying in 2015, 2016 and 2017 was satisfactory and mortality rates were consistent across all wall surface types sprayed by different spray operators and teams. All tests conducted <2 weeks after spray application resulted in mortality of 100%, with the exception of a lowest mortality recorded at 90.8% from one house in 2016. Monthly cone bioassay indicated a mean residual duration of 7 months post-spraying (mortality >80% WHO defined mortality threshold), with a decrease in mortality to approximately 50-70% recorded 9 months post-spraying (Fig. 2). Trends were similar for all wall substrates.
Indoor density of Anopheles gambiae sensu lato (s.l.) and Anopheles funestus s.l. by CDC light trap (2016-17)
Fig.3 presents the mean nightly indoor catch of An. gambiae s.l. and An. funestus s.l. from indoor CDC light trap collections conducted monthly for 2 years from January 2016 to December 2017. Anopheles gambiae s.l. was the predominant vector species in all sites throughout the sampling period over the 2-year period 2016-2017. Anopheles funestus s.l. indoor densities were very low in most sites, with relatively high indoor densities only recorded in Chato (June-October) and Butiama (May-June). Density peaks were generally observed following periods of significant rainfall occurring between October and April (Fig. 3).
Following IRS in February/March, indoor densities were generally low in sprayed sites, at <3 An. gambiaes.l. per trap/night between March and December (1-10 months after spraying). In the sprayed sites of Ngara (Kagera Region), Geita (Geita Region) and Kwimba (Mwanza Region), densities were particularly low year-round, never exceeding 1 per trap/night. However, in Missenyi (Kagera Region) indoor An. gambiaes.l. densities were particularly high between April and August at 4-8 per trap/night, despite IRS in February. While in Butiama a smaller indoor peak of An. funestuss.l. was reached in June at 3 per trap/night, 3 months after IRS.
In many sprayed sentinel sites, including Chato (Kagera Region), Sengerema (Mwanza Region), Musoma Rural and Butiama (Mara Region) relatively high An. gambiae s.l. indoor densities were recorded between January and February. This is 10-12 months after the previous IRS cycle, by which point residual efficacy had waned.
Biting rate for An. gambiaes.l. using CDC light trap fitted with bottle rotator (CBR)
In 2017, CBR traps were set from March to December. Data were combined for 4 sprayed sites (Sengerema, Musoma Rural, Chato, Bukoba Rural), and 4 unsprayed sites (Busega, Bukombe, Tarime, Biharamulo) to compare the mean biting rate indoors and outdoors. The total catch size per site using CBR (indoors and outdoors) over 200 trap nights per site indoors and outdoors (10 trap nights per month both indoors and outdoors for 10 months) was 4,616 An. gambiaes.l. from sprayed sites and 5,260 from unsprayed sites. The total An. gambiaes.l. collected per sprayed site was 333 from Sengerema, 290 from Musoma Rural, 3,809 from Chato, and 184 from Bukoba Rural; while for unsprayed sites the total was 83 from Tarime, 2,795 from Bukombe, 2,303 from Biharamulo, and 79 from Busega.
In sprayed sites, the An. gambiae s.l. biting rate was higher outdoors than indoors at all times of night. In unsprayed sites there was more outdoor biting, but only late at night between 22.00 and 03.00. However, it should be noted that the majority of An. gambiae s.l. in all sites were collected later in the evening when the majority of people are likely to be indoors and protected by LLINs. Nevertheless, a greater degree of outdoor biting risk was observed early in the evening in sprayed sites compared to unsprayed sites (Fig. 4).
Indoor resting densities of An. gambiaes.l. in 2017 using Prokopack aspirators
The mean number of An. gambiae s.l. collected by Prokopack aspirator resting indoors was greater in the 4 unsprayed sites of Biharamulo, Bukombe, Busega, and Tarime than in the 4 sprayed sites of Bukoba Rural, Chato, Sengerema, and Musoma Rural. In general, the highest peak in resting density was observed between May and August after the long rain season, with Chato and Busega also having a smaller peak in March (Fig. 6). There was no An. gambiaes.l. collected throughout the 2017 collection period in the sprayed site of Musoma Rural (Fig. 5).
A total of 8,957 female Anopheles collected from 2016 to 2017 were analysed by PCR for species identification. The samples consisted of 5,306 (59.2%) from sprayed and 3,651 (40.8%) from unsprayed sites, with 4,389 (49%), 2,866 (32%), and 1,702 (19%) collected from CDC light trap, CBR and Prokopack aspirator, respectively. Results confirmed vector populations in sprayed districts to be predominantly Anopheles arabiensis (71%, n=3,768) with minor proportions of Anopheles parensis (11,1%, n=589), An. funestus s.s. (11%, n=585, An. gambiae s.s. (6.8%, n=361), and Anopheles rivulorum (0.1%, n=3). There was a significantly greater mean number per year of An. funestuss.s. in unsprayed sites than sprayed sites (214.5 vs 58.5, p==0.024). The predominant vector species in unsprayed districts was An. arabiensis (66.9%, n=2,441), however there was a higher proportion of An. funestus s.s. (23.5%, n=858) and fewer An. parensis (2.3%, n=85), and similar proportion of An. gambiae s.s. (7.3%, n=267) as in sprayed sites.
Between 2016 and 2017 the overall P. falciparum sporozoite rate across all sites (sprayed and unsprayed) for all Anopheles (An.funestus, An. arabiensis, An. gambiae, and An. parensis) combined was estimated as 1.72% (286/16,670). The overall sporozoite infection rate was higher in unsprayed sites, estimated as 2.02% (115/5,686) than in sprayed sites at 1.56% (171/10,984) (Kruskal-Wallis test, H (3)=6.584, p=0.086). Mean sporozoite rates were generally less than 2% for all sprayed sites (from 2016 to 2017), with the highest rates scored at 4.5% (Ngara, 2017) and 3.9% (Biharamulo, 2016) in areas where An. funestus and An. gambiae were relatively common. See Additional file 1; Table S1 for 2016 and 2017 sporozoite rates presented by site.
Results from 2017 were disaggregated by species (from PCR results) and spray status (Table 4). This could not be done with data from 2016. Results by species showed that An. funestuss.s. had the highest sporozoite rate estimated at 4.07% (30/738) across unsprayed and sprayed sites combined. The mean An. funestuss.s. sporozoite rate estimated as 4.3% (27/630) in unsprayed sites and 2.8% (3/108) in sprayed sites, although the difference was not statistically significant (p=0.48) (Table 4), possibly due to the small sample size in sprayed sites. The predominant species, An. arabiensis exhibited a relatively lower overall sporozoite rate in 2017 estimated as 1.34% (45/3,366), with a higher sporozoite rate in unsprayed sites (2.0%; with 95% CI: 1.4-2.9) compared to sprayed sites (0.8% with 95% CI: 0.5-1.3) (p=0.003). Although not commonly considered as an important malaria vector, An. parensis had an overall sporozoite rate of 1.1% (5/435).
Blood meal analysis
A total of 194 blood-fed An. arabiensis (identified by PCR) that were collected from January to September 2017 by indoor resting collections were tested for vertebrate host blood source (human, bovine, goat, dog) with 109 from sprayed sites (Sengerema, Kwimba, Bukoba rural, Missenyi) and 85 from unsprayed sites (Bukombe and Busega). Overall, the proportion of An. arabiensis that fed on humans (including mixed blood meals on both human and animal) was 59.3% (115/194), with cattle blood being the most common non-human source. The overall human blood index was 0.62 in sprayed sites and 0.55 in unsprayed sites. An estimated 32.5% (63/194) of An. arabiensis fed on both human and animals, demonstrating opportunistic feeding behaviour, while only 26.8% fed only on humans (Additional file 1, Table S2).