Background
While great strides are being made towards reducing the worldwide malaria burden, malaria still caused an estimated 584,000 deaths in 2013, mostly among African children [
1]. In Kenya, there are an estimated 6.7 million new clinical cases and 4000 deaths each year, and those living in western Kenya have an especially high risk of malaria [
2]. Malaria prevalence is highest in Kenya in the lake endemic zone (38 %), is caused primarily by
Plasmodium falciparum [
3], and remains the most common cause of child morbidity in western Kenya [
4]. In this region,
P. falciparum is vectored by
Anopheles gambiae sensu stricto
(s.s.),
An. arabiensis, and
Anopheles funestus s.s. [
5]. Given the behavioural diversity among these three vector species, a strategic investment must be made towards understanding their ecologies in order to design an effective, integrated management approach [
6].
Anopheles gambiae
s.s. and
An. funestus
s.s. are both highly anthropophilic, but
An. arabiensis feeds readily on non-human vertebrates, particularly cattle [
7‐
9]. Anthropophily in
An. arabiensis also varies significantly, ranging from a high preference for human blood in West Africa to almost exclusive zoophily in Madagascar [
7,
10‐
12]. In western Kenya, over half of the blood meals identified from
An. arabiensis came from cattle, but a small proportion of
An. gambiae
s.s. also fed on cattle [
13]. In areas where
An. arabiensis is more anthropophagic, blood feeding still occurs predominately outdoors [
14]. These behaviour traits make
An. arabiensis less likely to encounter control strategies, which target endophagic and endophilic mosquitoes.
Zhou et al. [
15] documented a resurgence of malaria parasite prevalence and malaria vectors in western Kenya despite increased usage of ITNs, which could be attributed to insecticide resistance and poor ITN coverage or usage. However, over the last 10 years,
An. gambiae s.s. and
An. arabiensis have also undergone changes in their relative abundance, likely influenced by the implementation of IRS and ITNs [
13,
16]. While these strategies have led to the reduction of
An. gambiae s.s. and
An. funestus s.s., an unintentional consequence has been a proportionate increase in
An. arabiensis [
13]. Therefore, novel control strategies are needed for use in integrated malaria management programmes that target outdoor-feeding vectors not effectively controlled by ITNs and IRS.
One such approach is the use of “endectocides”, or treatment of a vertebrate host with a systemic insecticide that haematophagous arthropod vectors would become exposed to upon blood feeding. This host-targeted insecticide strategy for vector control has already been demonstrated effective in reducing sand fly vectors of visceral and cutaneous leishmaniasis [
17‐
20], and flea vectors of plague [
21,
22]. Targeting cattle, a frequent blood host of
An. arabiensis [
7‐
9,
12,
13], with a systemic insecticide may be an efficient approach to control this vector species. Foy et al. [
23] discussed the application and potential impact of ivermectin and other endectocides on malaria control. Community-directed ivermectin treatment of humans is already main strategy for control of onchocerciasis [
24], and has been successfully used in humans for malaria control as well [
23,
25]. Many studies have demonstrated the lethal effect of ivermectin on mosquitoes after imbibing ivermectin-treated blood [
26‐
30]. Eprinomectin is commercially used for control of endoparasites of livestock [
31] and was demonstrated to be as effective as ivermectin at killing blood-feeding
An. gambiae s.s. in the laboratory [
30]. However, further investigation is needed to determine whether efficacy against mosquitoes is maintained in an in vivo system, and ascertain the duration of effectiveness. Fipronil is a broad spectrum insecticide which blocks the GABA-gated ion channels in the central nervous system [
32]. Fipronil has been used to control ectoparasites on domestic animals [
33], and as a pour-on or dip for cattle to control ticks [
34,
35]. Mosquitoes are highly susceptible to fipronil during all life stages and by different routes of exposure [
36‐
40]. However, field tests of fipronil as a systemic insecticide for mosquito control are currently lacking.
The long-term goal of the research is to create a product that can be utilized in an integrated malaria management programme, particularly to augment current control methodologies aimed at endophilic vectors by targeting more exophilic vectors with broader host utilization, such as An. arabiensis. To that end, this study examined the efficacy of ivermectin, eprinomectin and fipronil on the survivorship of adult An. arabiensis. The specific aim of this study was to determine the percent mortality of adult female An. arabiensis fed on cattle treated with different doses of ivermectin, eprinomectin, and fipronil, and determine the duration of this lethal effect post-treatment.
Methods
Study area
The study site was located 10 km west of Kisumu in the village of Kisian, Kenya (latitude −0.073220° and longitude 34.662974°).
Cattle breed selection and cattle maintenance
All animal activities were reviewed and approved by the Institutional Animal Care and Use committees at Genesis Laboratories, Inc. and the Kenya Medical Research Institute (KEMRI). Lactating Zebu cattle (Bos indicus) were leased or purchased from markets or from private individuals. Cattle were transported to the study cattle shed located on the grounds of US Centers for Disease Control and Prevention and the Kenya Medical Research Institute (KEMRI), Kisian, Kenya. Transportation permits were provided by the department of veterinary services nearest to each purchase location. Test subjects were housed in individual stalls (1.5 × 3 m) within a covered cattle shed and were allowed periodic grazing in an outdoor pen during the 12-days acclimation period.
Upon arrival to the test facility each cow received an ear tag with a unique identification number and was inspected for general health. All test subjects were provided with clean tap water ad libitum and clean feed consisting of 8 kg of chopped Napier grass (Pennisetum purpureum) and 1.3 kg of dairy meal per day as directed by project veterinarians.
Cattle (test subjects) were maintained in a semi-controlled environment with adequate ventilation and natural light. Each test subject’s general health, and the daily temperature and relative humidity of the animal facility were documented by staff during the acclimation period and the test.
Treatment randomization
A blocked randomization scheme by body weight was used to eliminate possible bias. Randomization was carried out using a random number generator service [
41]. Each of the test subjects was assigned to either a control or treatment group. For each experiment, treatment groups which received doses of insecticide (test substance) consisted of three lactating Zebu cattle each, and the control group was allocated two lactating Zebu cattle. Precautions were taken to avoid animals contacting or grooming each other. The animals were housed individually in separate pens with a minimum distance to avoid contact between animals within and between treatment groups. Control animals were separated from the treatment animals.
Administration of the test substance
Four experiments were conducted in order to evaluate multiple doses each of ivermectin, fipronil, and eprinomectin (Table
1). Experiment 1, conducted from 20 Dec 2012–22 Jan 2013, consisted or dosing cattle orally with eprinomectin at doses of 0.2 or 0.5 mg/kg, or topically with 0.5 mg/kg eprinomectin (Eprinex
®). In experiment 2, conducted between 23 Jan 2013–26 Feb 2013, cattle were dosed orally with 0.1 or 0.2 mg/kg ivermectin, or topically with 0.75 mg/kg eprinomectin. In experiment 3, conducted from 25 Apr 2013–6 Jun 2013, cattle were dosed orally with either 1.0 or 1.5 mg/kg fipronil, or topically with 1.5 mg/kg eprinomectin. Experiment 4, conducted from 8 Aug 2013–22 Sept 2013, cattle received oral doses of fipronil at either 0.25 or 0.5 mg/kg. For each of these experiments, cows were randomized into 3 cows per treatment group and 2 cows per control for a total of 11 cows per experiment. Test substance quantity was calculated using weights recorded no more than 3 days prior to dosing. Topical and oral application methods of administering eprinomectin were chosen to assess efficacy and explore differences between application routes on mosquito survivorship.
Table 1
Listing of active ingredients, concentrations, route of administrations and total number of engorged mosquitoes used in the survival analysis per experiment
1 | T0 | Control | n/a | n/a | 379 |
T1 | Eprinomectin | 0.2 mg/kg | Oral | 465 |
T2 | Eprinomectin | 0.5 mg/kg | Oral | 396 |
T3 | Eprinomectin | 0.5 mg/kg | Topical | 408 |
2 | T0 | Control | n/a | n/a | 511 |
T1 | Ivermectin | 0.1 mg/kg | Oral | 475 |
T2 | Ivermectin | 0.2 mg/kg | Oral | 416 |
T3 | Eprinomectin | 0.75 mg/kg | Topical | 522 |
3 | T0 | Control | n/a | n/a | 522 |
T1 | Eprinomectin | 1.5 mg/kg | Topical | 537 |
T2 | Fipronil | 1.0 mg/kg | Oral | 599 |
T3 | Fipronil | 1.5 mg/kg | Oral | 575 |
4 | T0 | Control | n/a | n/a | 460 |
T1 | Fipronil | 0.5 mg/kg | Oral | 429 |
T2 | Fipronil | 0.25 mg/kg | Oral | 407 |
Experiment 1 Cattle in treatment group one (T1) received an eprinomectin dose of 0.2 mg/kg orally, subjects in T2 received 0.5 mg/kg orally and subjects in T3 received 0.5 mg/kg topically. Because eprinomectin is not commercially available in oral formulations, crystalline eprinomectin was weighed in the laboratory and placed in a capsule for oral application. For T3, eprinomectin was applied topically using liquid Eprinex© (Merial Ltd., New Zealand) which was applied according to the manufacturer’s application directions. The manufacturer recommended application for Eprinex© pour-on commercial product is 1 ml/10 kg which would achieve a dosage of 0.5 mg eprinomectin/kg body weight.
Experiment 2 Treatment group T1 received a 0.1 mg/kg ivermectin orally, subjects in T2 received 0.2 mg/kg ivermectin orally and subjects in T3 received 0.75 mg/kg eprinomectin topically. Ivermectin was administered orally using boluses; ivermectin tablets were weighed in the laboratory and placed in a capsule for oral application. Eprinomectin was applied topically using liquid Eprinex© applied according to the manufacturer’s application directions, but with a higher dose.
Experiment 3 Treatment group T1 received a 1.5 mg/kg eprinomectin topically, subjects in T2 received 1.0 mg/kg fipronil orally and subjects in T3 received 1.5 mg/kg fipronil orally. Fipronil was administered orally using capsules. Technical grade fipronil was weighed in the laboratory and placed in a capsule for oral application. Eprinomectin was applied topically using liquid Eprinex©, applied as described above, but with a higher dose.
Experiment 4 Treatment group T1 received a fipronil dose of 0.5 mg/kg orally while subjects in T2 received 0.25 mg/kg orally. Fipronil was weighed in a laboratory and placed in a capsule for oral application. For each experiment, the control group (T0) was left untreated.
Clinical observations of test subjects were recorded daily by project staff during acclimation and experimentation phases of the study. In addition, a veterinarian conducted weekly health checks to more thoroughly examine test subject health. During application and experimentation periods, feed was weighed daily to assess the effects of test substances on the animals’ appetite. When spillage occurred, feed was returned to the appropriate container and weighed to the nearest 0.5 gram. Cattle weights were recorded on the final day of acclimation and weekly throughout the course of the study. Differences in appetite and body mass were compared by evaluating test subject weight means and standard deviations before and after treatment.
Mosquito bioassays
All An. arabiensis used in this study were reared at the KEMRI/CDC, Kisian station, Kenya. Efficacy of each treatment was assessed by comparing survivorship of fully blood fed An. arabiensis at 1, 3, 5, 7, 14 and 21 days post treatment in experiments 1 and 2. While in experiment 3 mosquitoes were exposed at days 1, 7, 10, 14 and 21, in experiment 4 we exposed mosquitoes in days 1, 3, 5, 7, 14 and 21.
Prior to bioassays approximately 600 An. arabiensis adults were separated into an experimental cage and starved for 12 h. The day of application, 11–12 plastic capsules were filled with approximately 50 3–4 day-old female mosquitoes. Containers were modified round paper cartons that were 9.5 cm deep and 8.5 cm in diameter, covered with nylon netting material on one end to facilitate blood feeding. Containers with mosquitoes were transported in a cooler to and from the cattle shed.
The day before application all cows had a circular patch approximately 6 inches in diameter shaved on the ventral portion of the abdomen to expose skin and facilitate feeding. One container with An. arabiensis was applied to the shaved location of each test subject and secured by wrapping an ace bandage around the torso. One test subject in the control group received one capsules to ensure that the number of cartons applied to each group was equal. Containers were attached to test subjects for 30 min, and then carefully removed, and blood-fed females were counted. Unfed females were removed from the study.
Data were only analyzed for fully-engorged female mosquitoes. Blood fed females were placed into cages, provided with a 10 % sugar source ad libitum. For each group of mosquitoes in experiment 1 and 2, mortality was monitored at 3, 6 and 24 h post feeding and then daily for approximately 12 days thereafter. In experiment 3 and 4 we followed the same scheme but mortality was recorded daily for 9 days after the first 24 h.
Statistical analysis
The statistical analysis of the survival data obtained from the control and treatment groups was conducted using the “survival” package [
42] for the software R [
43]. The package implements the Kaplan–Meier estimator, which is used to calculate the survival function of a random variable in time. A survival curve is the plot of the survival function representing the survivorship of the target population. The statistical difference between the control and the treatment was assessed using the Mantel–Haenszel test as implemented in the survival package. Values smaller than 0.05 represent a significant difference between the control and treatment group. The resulting survival functions were used to estimate the median survival time and 95 % confidence intervals for the estimate and the size of the effect of the active ingredient (Table
2).
Table 2
Median survival time and 95 % confidence interval per experiment
Exp. 1 control | 216 [168, 264] | 144 [96, 192] | 168 [144, 192] | 96 [48, 216] | n/a | 156 [96, 192] | 120 [96, 168] |
0.2 mg/Kg OE | 192 [144, 264] | 72 [72, 96] | 120 [96, 168] | 192 [168, 192] | n/a | 24 [24, 96] | 96 [96, 120] |
0.5 mg/Kg OE | 24 [24, 24] | 48 [48, 72] | 72 [72, 96] | 168 [120, 216] | n/a | 24 [24, 72] | 96 [96, 96] |
0.5 mg/Kg TE | 48 [48, 72] | 72 [48, 72] | 48 [48, 72] | 72 [72, 96] | n/a | 48 [48, 72] | 96 [96, 120] |
Exp. 2 control | 144 [120, 168] | 72 [48, 120] | 216 [168, 288] | 240 [216, 264] | n/a | 192 [168, 192] | 120 [120, 120] |
0.1 mg/Kg OI | 96 [72, 96] | 48 [48, 72] | 132 [48, 264] | 144 [96, 240] | n/a | 192 [168, 216] | 120 [120, 144] |
0.2 mg/Kg OI | 72 [72, 72] | 60 [48, 72] | 72 [72, 120] | 144 [96, 168] | n/a | 168 [144, 216] | 120 [120, 120] |
0.75 mg/Kg TE | 48 [48, 48] | 48 [24, 48] | 48 [48, 48] | 48 [48, 72] | n/a | 168 [144, 192] | 120 [120, 120] |
Exp.3 control | n/a1 [−∞, ∞] | n/a | n/a | 216 [216, 240] | 168 [144, 168] | 192 [192, 240] | n/a1 [240, ∞] |
1.5 mg/Kg TE | 24 [24, 24] | n/a | n/a | 48 [24, 48] | 72 [48, 96] | 192 [192, 240] | n/a1 [216, ∞] |
1.0 mg/Kg OF | 24 [24, 24] | n/a | n/a | 48 [48, 48] | 72 [72, 96] | 132 [72, 192] | 144 [72, 240] |
1.5 mg/Kg OF | 24 [24, 24] | n/a | n/a | 24 [24, 48] | 48 [24, 48] | 84 [48, 144] | 72 [48, 168] |
Exp. 4 control | 144 [120, 144] | 168 [168, 216] | n/a | 192 [168, 240] | n/a | n/a1 [216, ∞] | 240 [216, ∞] |
0.25 mg/Kg OF | 48 [48, 48] | 120 [96, 144] | n/a | 48 [48, 48] | n/a | 204 [144, ∞] | 216 [192, ∞] |
0.5 mg/Kg OF | 48 [48, 72] | 120 [120, 144] | n/a | 144 [144, 168] | n/a | 216 [144, ∞] | n/a1 [216, ∞] |
To compare the effect of time on the effectiveness of the test substance we did a post hoc analysis for the same concentration and delivery method (a single row on a table). For this, the significant level was adjusted using a Bonferroni correction (α/n, where α is the significance level set at 0.05 and n is the number of comparisons).
Conclusions
Ivermectin, eprinomectin, and fipronil each show promising potential as endectocides administered to cattle for lowering the survival rate of An. arabiensis mosquitoes, and hence reducing malaria transmission rates. Mosquito mortality was significantly higher than control mortality as long as 21 days post-treatment after mosquitoes fed on cattle dosed orally with 0.2 or 0.5 mg/kg eprinomectin, topically with eprinomectin at 0.5 mg/kg, or orally with either 1.0 or 1.5 mg/kg fipronil. Other components of vectorial capacity were not evaluated, and would be valuable to incorporate into future studies. Endectocidal treatments in cattle are a promising new strategy for control of residual, outdoor malaria transmission driven by vectors that feed on cattle, and could effectively augment current interventions which target more endophilic vector species.
Authors’ contributions
The project was conceived and designed by RP. Principal Investigators were RP (Genesis) and NB (KEMRI). Project consultants were JG, and EW. The field work was conducted by RR, RG, DP, JK, DB, MNB. Chemical analyses were performed by LP. The data were analyzed and figures generated by SL. The manuscript was written by RCK with contributions by SL. All authors read and approved the final manuscript.