Study site
This study was conducted in Gasabo District which occupies the northern half of Kigali City approximately 9.8Kms from Kigali city with fifteen administrative sectors. It has an area of 430.30 km2 of which a big portion (84%) is rural while the small portion (16%) represents the developed urban area (16%), with a population of 879,505 residents in 2022 with 81.2% of its population residing in urban areas (National Institute of Statistics of Rwanda et al., n.d.). The elevation is about 1456 meters above sea level. Rainfall is generally bimodal, March to May is marked with long rains, and short rains from September to November. The short dry season starts from December to February and long dry season start from June to mid-September (Muhire et al., 2015). The annual average rainfall received is 927 mm and the temperature ranges from 17°C to 28°C. Most of the population in Gasabo District are employed in Agriculture (31%), Trade (17%), Government (11%) (Gasabo District, 2013). Malaria in Gasabo District is mesoendemic and main vectors are Anopheles gambiae ss and An. Arabiensis.
Rice farming is a major agricultural activity. The first farming cycle is from January to June and the second cycle starts in July and ending in December of each year. The expected mosquito breeding sites are mainly made of the stagnant water in the rice fields, expected to be permanent for the first three months of the rice cultivation (July to September and January to March). Other potential mosquito breeding sites are mainly after rain season, include inter-crops water drains, the pits and puddles from mining activities, water dams for harvesting rainwater for irrigation, stagnant water in the peri-domestic, water in different containers in use or unused.
Study design
The study was non-randomized with control involving a total of five blocks of marshlands located into five sectors of the District of Gasabo. Four blocks of marshlands located in the sectors of Jabana, Gisozi, Gatsata, and Kinyinya, with a total area of 336Ha was the experimental arm and received Bti application (Figure 1&2). The control arm was in the sector of Nduba with 78 Ha and did not receive any Bti application. Using drones, maps of all water bodies in the experimental and control parts were generated prior to the intervention and each time before Bti application.
The experimental sites were treated with 3000 ITU/mg Bacillus thuringiensis israelensis (Bti) strain AM 65-52 every 2 weeks using the Unmanned Aerial Vehicles “drones”. Rice farmers were trained to spray Bti into the mosquito breeding sites not accessible by drones mainly the peri-domestic breeding sites and other areas identified as non-eligible for aerial spraying.
Considering the recommended dosage of Bti 3000 ITU/mg, strain AM 65-52, commercially traded as VectoBac®, Water-Dispersible Granules (VectoBac® WDG), 300 grams were diluted in 10 liters and covering one ha with aerial spraying with drone in 15 minutes, one drone was estimated to cover between 15 to 20 ha per day. For the supplemental hand application, sprayer pumps were calibrated for releasing 30 liters of 300 grams diluted Bti in one ha of water body.
Mosquito larval sampling methods
To measure the impact of larval control using Bti, baseline surveys on mosquito breeding habitats and larvae were carried out one week before the application of Bti, key entomological indicators were measured on larval densities. Larval sampling continued every two weeks for ten months, starting from two to three days post Bti spraying. The larval monitoring was performed in selected sampling plots purposively chosen using a Global Position System coordinate, sampling points marked at every 100 meters alongside the marshlands in three line transects, middle and two ridges of the marshlands (Figure 3). The overview of the numbers of sampling plots is as following (table 1):
Sampling was done using standard dippers (350ml) to make five or ten dips (depending on the type of habitats) in each water body, the presence that were marked as positive or absence marked as negative of mosquito larvae were recorded and categorized according to their development stage.
Mosquito adult sampling methods
Indoor mosquito collections were done for two successive nights using battery powered CDC miniature light traps and pyrethrum spray collection (PSC) to sample endophagic and endophilic vectors respectively. Female adult mosquitoes were collected in twenty randomly selected houses from five different sites adjacent to the marshlands, three sites (12 houses) in neighborhood of intervention area and two (Eight houses) in control area (Figure 2). Identification of mosquito species ad sprorozoite infection (methods used: morphological and PCR, and ELISA for detection of sporozoite infection)
The sampled adult mosquitoes were counted and identified using morphological features such as wing patterns, size, abdomen markings, mesopleural and thoracic hairs as described in the Gillies and Coetzee identification keys (M.T. Gillies and M. Coetzee, 2020). Each collected female anopheline mosquito sample was kept individually in a labeled micro-centrifuge tube with a lid for airtight locking with a desiccant.
Siblings of the collected female An. gambiae s.l. were characterized using PCR technique based on DNA which utilizes a mixture of 20 base oligonucleotides primers that target the species nucleotide sequences in the ribosomal DNA (rDNA) intergenic spacers (IGS) (Scott et al., 1993), after genomic DNA extraction from legs and wings using the Cetyl Trimethyl Ammonium Bromide (CTAB) based protocol (de la Cruz-Ramos et al., 2019). To determine presence of a circumsporozoite proteins (CSP), ELISA tests was done on the head and thorax using the CS microplate ELISA to detect P. falciparum, Optical Density (OD) measured using spectrophotometer (Appawu et al., 2003). Blood-meal sources of all collected blood-fed female samples captured by PSC were analyzed using a direct ELISA, using antihost Monoclonal antibodies (IgG) conjugate against human, cattle, goat, sheep, chicken proteins, OD measured using spectrophotometer (Getachew et al., 2019).
The balance checks for study
The balance check was done to assess the abundance and density of anopheles, and culicines mosquito larvae and pupae in the study sampling plots before the intervention. The initial density was also conducted on adult mosquitoes collected using CDC-LT and the Pyrethrum Spraying Collection methods.
Determination of malaria incidence
Community health workers (CHWs) patients’ registers were used for monthly data collection of malaria cases from the contingent villages to the study area. Fifteen villages in proximity to the marshlands of the study area, aggregated into twelve nearby the intervention and three in the control area respectively.
Statistical Analysis
Data were recorded in Excel and transferred into statistical software, version R 4.0.2 for statistical analysis. In preliminary, descriptive analysis was performed to generate tables and curves. The intervention and control groups were constructed to visualize differences in the responses between the two groups. We used mean, median, and standard errors (SE) to present continuous variables and frequencies for categorical variables.
We performed a balance check using the T-test to compare means of treatment and control groups for independent variables at baseline (Round 0). Further, the difference analysis was calculated using regression analysis, adjusted with time, to evaluate the effect of larviciding on breeding habitats with anopheles and culicines mosquito larvae and pupae as well as the adult mosquitodensities. The evaluation of differences on malaria incidences between intervention and control sites, a non-parametric test was used, Mann Whitney test.
Ethical consideration
The study was presented to Rwanda Biomedical Centre, Division of Research, Innovation and Data Sciences for review and clearance and received approval Ref: No 225/RBC/2020. The importation and usage of Bti was authorized by the Ministry of Health, Department of Food and Drug Authority. Before application of Bti and larval monitoring, verbal consent was obtained from local leaders, the head of the rice farmer cooperative, owners of houses used for adult collections and the entomology technicians involved in entomology monitoring.