Study sites and chlorfenapyr dosages tested
Experiments were conducted using five concentrations of chlorfenapyr at 12.5, 25, 50, 100 and 200µg AI/bottle. Tests were conducted on wild uncharacterized An. gambiae s.l. in 10 countries in sub-Saharan Africa, out of which eight countries conducted additional tests with colonized pyrethroid susceptible strains. The countries included, The Democratic Republic of Congo (4 sites), Ethiopia (1 site), Ghana (3 sites), Kenya (2 sites), Madagascar (10 sites), Mali (11 sites), Nigeria (3 sites), Senegal (3 sites), Uganda (2 sites) and Zimbabwe (1 site). Locations are shown in Figure 1. In parallel, susceptibility tests were conducted using a provisional discriminating concentration of 100µg AI/bottle in 16 countries. This concentration was chosen based on preliminary bottle bioassay testing by CDC which established a provisional discriminating concentration of 100µg AI/bottle (Dr WG Brogdon, 2017, personal communication). All bioassays were conducted between 2017 and 2020.
Preparation of solutions
Treatment of 250ml Wheaton® bottles was conducted locally in the country of testing using technical grade chlorfenapyr dissolved in acetone. A vial containing 5g of technical grade (99.9% pure) chlorfenapyr was supplied by BASF (Ludwigshafen, Germany) to each country team and a stock solution was prepared at 1mg/ml by weighing 100mg and dissolving with 100ml acetone. The stock solution was prepared in an amber glass bottle (or clear glass bottle covered with aluminium foil) to avoid exposure to UV light and sealed with a tightly fitting lid to prevent evaporation before being stored at 4°C in a refrigerator for a maximum of three months. A test solution of 200µg/ml was prepared by performing a five times dilution by mixing 10ml of the stock solution with 40ml of acetone. Diluents were serially prepared with 2-fold dilutions of 100µg/ml, 50µg/ml, 25µg/ml and 12.5µg/ml.
Bioassay procedures
Each 250ml glass bottle and its plastic cap were coated with 1ml of insecticide solution by rolling and inverting the bottles according to CDC procedures [20]. In parallel, a negative control bottle was coated with 1ml of acetone. All bottles were dried overnight in the dark and bioassays were conducted within 24h of treating bottles. In general, a total of 80-100 female mosquitoes, aged two to five days old, were exposed for 60 minutes in four replicates of 20-25 mosquitoes, with an additional single replicate of 25 mosquitoes used for the negative control (bottle treated with 1ml acetone). A total of 13,639 wild collected An. gambiae s.l. (56 vector populations per dose) in ten countries were tested using five concentrations of chlorfenapyr. While a total of 4,494 pyrethroid susceptible insectary mosquitoes from eight colonized strains were tested. A total of 23,422 wild collected pyrethroid resistant An. gambiae s.l. (259 vector populations) were tested at the discriminating concentration of 100µg AI/bottle in 16 countries. After exposure, mosquitoes were transferred to clean paper cups and provided with 10% sugar solution. Mortality was recorded at the end of the 60 minutes exposure and at 24, 48 and 72 hours after exposure. Tests were conducted during the day time with effort made to keep testing and holding conditions within WHO guidelines of 27°C±2°C and relative humidity of 75%±10% [19]. Temperature and humidity were monitored and recorded, however, in several cases could not be accurately controlled, as tests with wild collected mosquitoes were generally conducted in improvised field insectaries which did not have robust temperature and humidity controls.
Mosquito species tested
Insectary-reared pyrethroid susceptible colonies of An. gambiaesensu stricto (s.s.) Kisumu strain were used for testing in six countries (Ghana, Kenya, Madagascar, Nigeria, Uganda, Zambia) while Anopheles coluzzii Ngousso strain was used in Mali and Anopheles arabiensis Adama strain in Ethiopia. Larval collections of wild An. gambiae s.l. were made in areas where pyrethroid resistance had previously been detected from temporary sunlit pools between 2017 and 2020 (timing varied by country) using larval dippers. Larvae were subsequently transported to a field insectary where they were reared in water collected from the field and fed with Tetramin® fish food. Emerging adult mosquitoes were provided with cotton wool pads dipped in 10% sugar solution until they were used in insecticide susceptibility tests. Wild Anopheles were identified morphologically as An. gambiae s.l. in all 16 countries using the key of Gillies and Coetzee (1987) [22]. Molecular analysis to determine species of these test mosquitoes was not conducted. However, An. gambiae s.l. collected from the same locations for other purposes were identified to species by PCR using the protocols of either Scott (1993), Santolamazza (2008), or Wilkins (2006) to determine members of the An. gambiae species complex [23-26].
Data analysis
Insecticide susceptibility results were presented as unadjusted percentage mortality at the end of 60 minutes and subsequently 24, 48 and 72 hours after bioassay exposure. If negative control mortality was greater than 20%, the data was discarded, and tests were repeated. Box plots are used to present mortality data showing the median and interquartile range, with whiskers representing one and a half times the interquartile range and small circles outside the whiskers considered outliers. PoloPlus (LeOra Software, Parma MO, USA) was used to conduct probit analysis on the logarithmic scale to calculate the concentration of chlorfenapyr needed to kill a defined proportion of mosquitoes, known as lethal concentration (LC). Mortality data (72 hours after exposure) was included for each concentration used in the analysis (12.5, 25, 50, 100, 200µg AI/bottle) to determine the LC50, LC95 and LC99 (concentration needed to achieve 50%, 95%, and 99% mortality) for wild An. gambiae s.l. The LC95 value was then multiplied by three to give a discriminating concentration (LC95×3=DC) as described by Lees et al, 2019 [27]. The WHO approach of multiplying the LC99 by two was also used to determine a discriminating concentration [28]. Probit analysis was not conducted with data generated using insectary strains as there was not a sufficient spread of data to fit the probit curve.