Insecticides

Novaluron was obtained as a commercial emulsifiable concentrate containing 100 g active ingredient (a.i.)/l. Pyriproxyfen was obtained as a commercial emulsifiable concentrate containing 103 g a.i./l. Spinosad used in ovitraps was obtained as a 2.5% a.i. granular formulation of spinosad applied to oviposition traps, whereas an extended-release tablet formulation containing 7.48% a.i. spinosad was used in water tanks. Temephos was obtained as a generic mineral granular formulation comprising 1% a.i. provided by the Ministry of Health (Secretaría de Salud), Mexico City.

Study area

Experiments were performed in an area of ~12 hectares in the town of Huixtla (population ~51 000), Chiapas, Mexico (15° 08’ N; 92° 28’ W, altitude 40 m), with an average annual temperature of 27 °C and annual precipitation of 2 500 mm that occurs mainly from May to October. The study area contained ~550 houses with ~2 500 inhabitants. Verbal consent was obtained from the heads of 75 households distributed in the area. All the activities described in this study were authorized by the Ethics Committee of the National Institute of Public Health (Instituto Nacional de Salud Pública [INSP]).

Persistence of larvicides in oviposition traps

The experiment to evaluate the persistence of larvicides ran for 13 weeks from May 27 to August 19, 2015, during the rainy season. Houses were assigned to treatments following a balanced blocks design with five treatments and 15 repetitions. A circular black plastic oviposition trap, 10 cm in diameter by 20 cm in height, containing 1 l water, was placed in each house. Each oviposition trap was assigned to one of five treatments: 1) 0.5 mg a.i./l pyriproxyfen, 2) 0.5 mg a.i./l novaluron, 3) three granules of spinosad (mean ± SE: 5.5 ± 0.1 mg); 4) 0.1 g temephos granules, 5) untreated water as control. These treatments corresponded to the concentrations recommended by the National Center for Preventative Programs and Disease Control for mosquito control in Mexico.⁷ To avoid loss of temephos granules, these were placed in a 1.5 ml perforated plastic tube that was removed prior to each sample. Weekly counts were performed to determine the number of living larvae and pupae present in each container. All insects (living and dead) were removed and discarded at each sampling, and the water that had evaporated was replaced in all treatments. Oviposition was monitored by placing a strip of filter paper in contact with the water, attached to the top of the container. Filter papers were replaced during weekly sampling and taken to the laboratory where eggs were counted. The prevalence of eggs on each filter paper that had hatched in the week since the previous sample, was recorded as proportion of egg hatch. Unhatched eggs were placed in water and larvae that emerged were reared to adulthood using a powdered laboratory rodent diet¹⁴ and identified to species. A single oviposition trap was lost during the study in the control and pyriproxyfen treatments, compared with two traps lost in the spinosad and novaluron treatments, and three traps, in the temephos treatment.

Efficacy of larvicides in domestic water tanks

This experiment was performed in 40 houses distributed in the same area of Huixtla as the previous experiment. The houses were assigned following a balanced blocks design with four treatments, each with 10 replicates (houses). The selected houses had an internal brick-built water tank with an average capacity of 1 668 liters.

The experiment began with three weeks of pretreatment sampling from August 25 until September 15, 2015. The same day (September 15), one of the following treatments was assigned to each tank, based on its estimated capacity: 1) 0.5 mg a.i./l pyriproxyfen (mean ± SE volume: 1 677±158 liters; range: 714-2 445 liters), 2) 0.5 mg a.i./l novaluron (mean: 1 762±162 liters; range: 825-2 379 liters), 3) spinosad - one tablet for every 200 liters water (mean: 1 606±175 liters; range: 590-2 420 liters), and 4) 10 g temephos granules for every 100 liters water (mean: 1 655±232 liters; range: 140-2 342 liters). To monitor the efficacy of larvicides in water tanks over time, two ovitraps with untreated water were placed in each house. Untreated water tanks were not used as a control because cases of chikungunya and dengue fever were present in the study zone.

In order to improve the persistence of liquid larvicides, pyriproxyfen and novaluron treatments were adsorbed onto 70 g of crushed pumice for a 24 h period prior to application. This resulted in fine granules, similar to the granular formulation of the temephos treatment. Each treatment (spinosad tablets, temephos granules and pyriproxyfen-treated or novaluron-treated pumice) was placed in a 6x22 cm perforated nylon bag, tied to an empty 600 mL PET soda bottle that floated on the water, with the gauze bag suspended ~15 cm below the surface.

Water tanks and untreated oviposition traps were monitored weekly to evaluate the persistence of larvicides during 12 weeks post-treatment from September 22 to December 8, 2015. Tanks were sampled using cone shaped nets, with an orifice diameter of 25 cm and a pore size of 0.7×0.17 mm. Two samples were taken from each tank; a perimeter sample was taken by dragging the net around the edge and another X-shaped sample was taken from corner-to-corner. Immature mosquitoes were counted, identified, registered according to genus and a sub-sample (~10 larvae) was placed in a plastic bag with water, transported to the laboratory and reared to adulthood on powdered rodent diet for species identification. For oviposition traps, larvae + pupae were counted, paper filters were changed and evaporated water replaced, as described in the previous section. The number of eggs on each of the paper filters and the proportion of hatched eggs were determined by direct observation in the laboratory.

Statistical analyses

Mixed effects models were fitted to each of the variables measured using R v. 3.4.0. Egg counts from oviposition traps in both experiments were analyzed by specifying a negative binomial error distribution. The prevalence of egg hatching was analyzed by specifying a binomial error distribution for treated oviposition traps (first experiment) and a normal error distribution for untreated oviposition traps (second experiment). The numbers of pupae were generally low, so that, for analysis, counts of larvae and pupae were pooled within replicates for each sample time and were analyzed by specifying a Poisson error distribution. Sample times were considered as a categorical variable.

Persistence of larvicides in oviposition traps

The mean (±SE) air temperature during this study was 32.6±0.3°C (range 26.0-41.1°C), while relative humidity averaged 57.6±1.0% (range 33-74%). Average container water temperature at the moment of sampling was 28.5±0.08°C (range 20-39°C).

A total of 52 563 eggs of Aedesspp. were registered in the oviposition traps in houses (figure 1A). Average weekly egg counts ranged from 48.9±4.5 eggs/trap in the control to 83.8±7.2 eggs/trap in novaluron-treated traps, but did not differ significantly among treatments (χ2=2.93, df=4, p=0.57); however, egg numbers/trap fluctuated significantly over time (treatment*time: χ2=63.5, df=44, p=0.029).

Overall, a proportion of 0.405 eggs (N=21 296) from oviposition traps were found to have hatched in the preceding week since the previous sample was taken (figure 1B). The mean proportion of egg hatch ranged from 0.318±0.023 in the control to 0.495±0.024 in pyriproxyfen-treated traps and differed significantly between treatments (c2=17.1, df=4, p=0.002) and between sampling dates (treatment*time: c2=1 610, df=44, p<0.001). The prevalence of egg hatching in pyriproxyfen-treated ovitraps was significantly higher than that observed in the control in eight sampling points and was higher than all other larvicide treatments in six out of 12 sampling points (figure 1B). None of the larvicides exhibited marked ovicidal effects.

In total, 3 473 Aedes spp. larvae + pupae were observed developing in experimental containers. The number of larvae + pupae differed significantly between treatments (c2=1 413, df=4, p<0.001) and between sampling points (c2=1 715, df=11, p<0.001). The average number of immature Aedes spp. in the control was 8.9±1.4 insects/container over the 12-week period (figure 1C). Among the larvicide treatments, the average numbers of larvae + pupae were highest in the pyriproxyfen treatment followed by the temephos treatment, with average densities of 6.3±0.5 and 4.2±1.0 insects/container, respectively. These larvicides provided 3-4 weeks of complete inhibition of immature development during the study. In contrast, the presence of larvae + pupae in spinosad- and novaluron-treated containers was lowest, with an average density <1.5 insects/container, although densities in these treatments increased at 9-12 weeks post-treatment. The spinosad treatment provided seven weeks of complete inhibition of Aedes spp. in oviposition traps, while the presence of immature mosquitoes was sporadic, but always low, in the novaluron treatment (figure 1C).

Laboratory rearing revealed that 95% of insects (N=354) in control oviposition traps were Ae. aegypti, compared with 94-100% (N=30-254, depending on treatment) of this species in larvicide-treated traps. All other individuals were identified as Ae. albopictus. As IGRs can affect the production of pupae, we calculated the percentage of emergence inhibition (IE) during laboratory rearing (table I).¹⁵ Small numbers of pupae first appeared in pyriproxyfen-treated traps at 4 weeks post-treatment and were first reared to adulthood at seven weeks post-treatment, resulting in an IE value of 94%. In the novaluron treatment, pupae were first observed at 10 weeks post-treatment, and none of the insects collected developed to adulthood (IE=100%). Pupae were observed in the temephos and spinosad tablet treatments at four and eight weeks, respectively, and laboratory rearing resulted in IE values of 51% in both treatments (table I).

Figure 1. Efficacy of larvicides in oviposition traps in terms of (A) mean weekly counts of eggs, (B) proportion of hatched eggs and (C) mean number of larvae + pupae in larvicide-treated traps placed in houses in Huixtla, Chiapas, Mexico, during a 12 week study (May-August, 2015). Vertical bars indicate SE; for clarity, only half the bar is shown at some points 

Efficacy of larvicides in water tanks

During the sampling period, the average air temperature was 31.5±0.1 °C (range 27-34 °C) and the mean of relative humidity was 61.2±0.8% (range 42-92%). The average tank water temperature was 26.8±0.05 °C (range 24-31 °C).

The average number of eggs in oviposition traps was significantly higher in the pre-treatment samples (range: 36.7-75.0 eggs/trap/week depending on treatment) compared with samples taken post-treatment (range: 28.9-44.3 eggs/trap/week) (z =-4.114, p< 0.001). Following treatment of water tanks, the average number of Aedes eggs in untreated oviposition traps fluctuated significantly over time (c2=8 800, df =1, p< 0.001) (figure 2A), but was significantly lower in houses that received the novaluron and spinosad treatments compared with those treated with temephos or pyriproxyfen (χ2=14.4, df=3, p=0.002). Overall, novaluron was significantly more effective than spinosad tablets in reducing Aedes spp. oviposition in traps (z=2.379, p=0.017).

The proportion of hatched eggs observed at each sample time (table II, figure 2B) varied between 0.268±0.020 in oviposition traps located in pyriproxyfen treatment houses and 0.363 ± 0.024 in traps in temephos treatment houses, and was intermediate in the spinosad and novaluron treatment houses (χ2=8.31, df=3, p=0.04).

Mean numbers of larvae + pupae in untreated oviposition traps, taken as an indicator of Aedes spp. population density during the experiment, fluctuated between 3.4 and 26.7 insects/trap during the post-treatment period (figure 2C). Larvicide treatment resulted in a significant reduction in numbers of larvae + pupae in water tanks (χ2=2 614, df=4, p<0.001). Temephos treatment provided just one week of absolute inhibition of immature mosquitoes (figure 2C), compared with seven weeks for spinosad and eight weeks for novaluron, although the numbers of larvae + pupae remained extremely low in both these treatments until the end of the study. Pyriproxyfen did not eliminate mosquito development at any time point, except at week 12. The mean number of Aedes larvae + pupae sampled in treated water tanks was highest in the pyriproxyfen treatment (3.11±0.7 larvae + pupae/tank), lowest in the novaluron and spinosad treatments (0.09 - 0.68 larvae + pupae/tank) and intermediate in the temephos treatment (1.84±0.4) (table II). Overall, the efficacy of spinosad tablets did not differ significantly from that of novaluron (z=0.788, p=0.43).

Of the insects reared from water tanks, a proportion of 0.965 (N=726) and 0.958 (N=1 350), were Ae. aegypti in pre-treatment and post-treatment samples, respectively. The remainder were Ae. albopictus and very small numbers of Culex spp. (N=15-42 insects) (table II). Pupae were first observed in the pyriproxyfen-treated water tanks at three weeks post-treatment and were first reared to adulthood at seven weeks post-treatment (table I). Inhibition of adult emergence (IE) in laboratory reared insects was 89% in this treatment. In contrast, no pupae were observed in the novaluron or spinosad tablet treatments. IE values were 71% and 37% in the spinosad tablet and temephos treatments respectively, based on laboratory rearing of small numbers of larvae (table I).

Figure 2. Efficacy of larvicides in water tanks in terms of (A) mean weekly counts of eggs and (B) proportion of hatched eggs in untreated oviposition traps placed in houses with larvicide-treated water tanks. (C) Mean number of larvae + pupae sampled in treated water tanks in houses in Huixtla, Chiapas, Mexico, during a 12 week study (August-December, 2015). Untreated oviposition traps acted as controls in (C). Vertical bars indicate SE; for clarity, only half the bar is shown at some points