Aedes aegypti abundance and insecticide resistance profiles in the Applying Wolbachia to Eliminate Dengue trial

Abstract

The Applying Wolbachia to Eliminate Dengue (AWED) trial was a parallel cluster randomised trial that demonstrated Wolbachia (wMel) introgression into Ae. aegypti populations reduced dengue incidence. In this predefined substudy, we compared between treatment arms, the relative abundance of Ae. aegypti and Ae. albopictus before, during and after wMel-introgression. Between March 2015 and March 2020, 60,084 BG trap collections yielded 478,254 Ae. aegypti and 17,623 Ae. albopictus. Between treatment arms there was no measurable difference in Ae. aegypti relative abundance before or after wMel-deployments, with a count ratio of 0.96 (95% CI 0.76, 1.21) and 1.00 (95% CI 0.85, 1.17) respectively. More Ae. aegypti were caught per trap per week in the wMel-intervention arm compared to the control arm during wMel deployments (count ratio 1.23 (95% CI 1.03, 1.46)). Between treatment arms there was no measurable difference in the Ae. albopictus population size before, during or after wMel-deployment (overall count ratio 1.10 (95% CI 0.89, 1.35)). We also compared insecticide resistance phenotypes of Ae. aegypti in the first and second years after wMel-deployments. Ae. aegypti field populations from wMel-treated and untreated arms were similarly resistant to malathion (0.8%), permethrin (1.25%) and cyfluthrin (0.15%) in year 1 and year 2 of the trial. In summary, we found no between-arm differences in the relative abundance of Ae. aegypti or Ae. albopictus prior to or after wMel introgression, and no between-arm difference in Ae. aegypti insecticide resistance phenotypes. These data suggest neither Aedes abundance, nor insecticide resistance, confounded the epidemiological outcomes of the AWED trial.

Fig 1. Ae. aegypti mosquitoes caught per week by study arm.

Panel A is the mean number of Ae. aegypti mosquitoes caught per BG trap per week and panel B is the ratio of female to male Ae. aegypti caught. Releases of wMel-infected Ae. aegypti mosquitoes occurred from March 2017 –December 2017 (yellow shading). The first nine weeks of releases are excluded from analyses due to unequal trapping density in the treated and untreated clusters. There was an average of 4 BG traps/km² in wMel-treated clusters and 5 BG traps/km² in untreated clusters in the pre-release period, and 16 BG traps/km² and 15 BG traps/km² during and post-release. The warm, wet season in Yogyakarta is from November to April.

Fig 2. Mean number of Ae. aegypti mosquitoes caught per trap per week by study arm (circles) adjusted for season (wet/dry).

Analysis of the total study period was additionally adjusted for release status (pre-release [March 2015 –February 2017], during release [March 2017 –December 2017], post-release [January 2018 –March 2020]). Bars are the number of trap collections in each arm. The first nine weeks of releases were excluded from analyses due to unequal trapping density in the treated and untreated clusters. There was an average of 4 BG traps/km² in wMel-treated clusters and 5 BG traps/km² in untreated clusters in the pre-release period, and 16 BG traps/km² and 15 BG traps/km² during and post-release. More Ae. aegypti were caught per trap per week in the wMel-treated arm compared to the untreated arm during wMel deployments (count ratio 1.23 (95% CI 1.03, 1.46)). When formally compared, Ae. aegypti abundance did not differ between the wMel-treated and untreated arm before (count ratio 0.96 (95% CI 0.76, 1.21), p = 0.74) or after wMel-deployments (count ratio 1.00 (95% CI 0.85, 1.17), p = 0.99). Similar results were found for the post-intervention period when wMel contamination was accounted for in a sensitivity analysis (count ratio 0.94 (95% CI 0.86, 1.03), p = 0.18).

Fig 3. Ae. albopictus mosquitoes caught per week by study arm.

Panel A is the mean number of Ae. albopictus mosquitoes caught per BG trap per week and panel B is the ratio of female to male Ae. albopictus caught. Releases of wMel-infected Ae. aegypti mosquitoes occurred from March 2017 –December 2017 (yellow shading). The first nine weeks of releases are excluded from analyses due to unequal trapping density in the treated and untreated clusters. No male and very few female Ae. albopictus mosquitoes (median (interquartile range) 0 (0–0), range 0–97) were caught between mid-2016 and mid-2017. There was an average of 4 BG traps/km² in wMel-treated clusters and 5 BG traps/km² in untreated clusters in the pre-release period, and 16 BG traps/km² and 15 BG traps/km² during and post-release.

Fig 4. Mean number of Ae. albopictus mosquitoes caught per trap per week by study arm (circles) adjusted for season (wet/dry).

Analysis of the total study period was additionally adjusted for release status (pre-release [March 2015 –February 2017], during release [March 2017 –December 2017], post-release [January 2018 –March 2020]). Bars are the number of trap catches in each arm. The first nine weeks of releases were excluded from analyses due to unequal trapping density in the treated and untreated clusters. There was an average of 4 BG traps/km² in wMel-treated clusters and 5 BG traps/km² in untreated clusters in the pre-release period, and 16 BG traps/km² and 15 BG traps/km² during and post-release. There was no significant difference in Ae. albopictus abundance before (count ratio 1.00 (95% CI 0.68, 1.49)), during (1.08 (0.78, 1.51)) or after (1.12 (0.87, 1.45)) wMel-deployments.

Fig 5. Ratio of Ae. aegypti to Ae. albopictus mosquitoes caught per week by study arm.

Weekly counts of Ae. aegypti and Ae. albopictus were aggregated across 12 wMel-treated clusters and 12 untreated clusters. Releases of wMel-infected Ae. aegypti mosquitoes occurred from March 2017 –December 2017 (yellow shading). The first nine weeks of releases are excluded from analyses due to unequal trapping density in the treated and untreated clusters. There was an average of 4 BG traps/km² in wMel-treated clusters and 5 BG traps/km² in untreated clusters in the pre-release period, and 16 BG traps/km² and 15 BG traps/km² during and post-release.

Fig 6. Percentage mortality of Ae. aegypti in the WHO bioassay for various insecticides by study arm in 2018 and 2019.

Boxes are median and interquartile range for percentage mortality in cohorts of mosquitoes derived from 12 wMel-treated clusters and 12 untreated clusters. Whiskers indicate the range, and circles indicate outliers. Insecticide type and concentrations used were in line with recommendations for Ae. aegypti mosquitoes and followed the WHO standard method [18]. The median (25^th– 75^th percentile) percentage mortality for the susceptible strain (Rockefeller) in 2019 is as follows: 100 (100–100) for cyfluthrin 0.15%; 91 (76–99) for malathion 0.8%; and 100 (97–100) for permethrin 1.25%.