In 2006, severe outbreaks of chikungunya occurred in the villages in Karnataka, South India. Chikungunya is a rare arboviral infection transmitted by Aedes mosquitoes. It shares the same vector and geographical distribution as dengue fever, and has similar symptoms with the exception of incapacitated arthralgia [1]. Over a period of three decades, chikungunya outbreaks have struck the southwestern islands of the Indian Ocean region. In India, outbreaks occurred late-2005 mainly in the southwestern states and were primarily transmitted by Aedes aegypti [2]. Approximately1.4 million suspected cases were reported in 2006 and over 0.7 million cases were from Karnataka alone [3]. Larvivorous fish in natural habitats that feed on Anopheles larvae have been successfully used in malaria control [4], but their use in confined domestic containers against Aedes larvae has been very limited [5]. Based on our previous experience using two potential poeciliid larvivorous fish guppy (Poecilia reticulata) and mosquitofish (Gambusia affinis) for malaria control in Karnataka [4, 6], we aimed to conduct two small-scale community-based feasibility studies combined with information, education and communication (IEC) to test the comparative efficacy of these fish in containing Ae. aegypti larval infestation and reducing chikungunya. Use of Poecilia fish combined with IEC offered the most effective means of controlling Aedes mosquito populations in our study.

In brief, our studies were carried out within Karnataka in two villages (Domatmari and Srinivaspura) in Tumkur District and in one village (Balmanda) in Kolar District, affected by chikungunya. A survey on knowledge, attitude and practice (KAP) on chikungunya was initially conducted in Domatmari and Srinivaspura. IEC campaigns were performed in Domatmari and Balmanda, before and after fish introductions. Aedes larval surveys were carried out in all three villages to assess the impact of the intervention measures undertaken. Poecilia was released in Domatmari and Srinivaspura, and Gambusia in Balmanda. Fish survival was also monitored during the larval surveys. Anonymised data on chikungunya cases were collected from the respective district health offices.

Before the study began, more than 3000 cases of fever were reported in villages within the Tumkur and Kolar Districts. Subsequent laboratory investigations confirmed these outbreaks were caused by chikungunya [2]. Thus, the present study was devised to identify an alternative method of vector control, as no other immediate options were available. Accordingly, we decided to use larvivorous fish as a biocontrol method against Aedes larvae as this was found very effective against malaria vectors [4, 6]. While we ensured matched control villages were included in our study as per standard experimental protocol, during the course of the trial, participants within the control villages released fish into their tanks when they discovered that the fish were responsible for controlling on-going outbreaks in the study villages. Thus, after consulting a statistician, the impact of intervention was assessed considering the data before and after fish introductions.

Self-sustained populations of Poecilia and Gambusia are the most preferred poeciliid larvivorous fish in malaria control in India [4, 6]. Besides their predatory nature, the presence of such fish may also inhibit Ae. aegypti oviposition in domestic containers [7]. Maintaining these fish in confined habitats is important in an Aedes control programme. In the one-week post-fish release period, an almost 100% fish survival rate was recorded in all villages. In the following one-month period, fish populations were better sustained in Domatmari than in Srinivaspura and Balmanda.

Poecilia is an omnivorous species that better survives in confined habitats, namely open dug wells. In the absence of larvae, this fish can survive on other food available within the ecosystem. It also grows equally well in small containers with minimum care. In the malaria control programme we observed that villagers offered cooked rice or Ragi (a type of millet locally grown) as a food supplement, which helped to sustain and propagate their fish populations. This information was further told to the villagers as part of the IEC programme. In contrast, Gambusia is a cannibalistic species that feeds on zooplankton as its preferred food source, and populations of this fish are not sustained in small water bodies for long periods of time [8]. This was possibly one of the reasons that this fish did not survive even for a month in cement tanks. Thus, Gambusia is not a preferred fish in small water habitats, and is preferable for large water bodies such as ponds and lakes [6]. While Poecilia reproduced in some of the study tanks, indicating sufficient conditions for sustained populations in these containers, this was not observed for Gambusia. We have also observed this finding in the laboratory (unpublished observation).

Many biocontrol agents have been tested against Aedes larvae. Recent experience in Vietnam has been remarkably successful, with members of the community being closely engaged in vector control efforts by cleaning public areas and using biocontrol agents in water storage tanks [9]. There is an abundance of local Mesocyclops spp. (copepods) in Vietnam that can be incorporated into specifically designed community-based control programmes aided by Micronecta water bugs and fish [10]. In a village in French Polynesia, Mesocyclops aspericornis and poeciliid fish released in ponds and tanks successfully controlled mosquito species other than Aedes, which do not breed in these habitats [11]. In a coastal village in Taiwan, integrated control of Aedes used G. affinis, P. reticulata, Tilapia mossambica and Sarotherodon niloticus in potable water containers. They were later replaced with Cyprinus carassius because of constant availability and adaptability [12]. In 1980, Chinese catfish was used to control Ae. aegypti larval breeding when a dengue outbreak occurred in fishing villages among Chinese coastal provinces [13]. In Thailand, the most effective method of Ae. aegypti control was maintaining fish in rectangular tanks and correctly covering water storing containers with lids [14]. In Southern Mexico, five indigenous fish species, namely Lepisosteus tropicus, Astyanax fasciatus, Brycon guatemalensis, Ictalurus meridionalis and P. reticulata, were significantly effective as biocontrol agents against Ae. aegypti larvae in water storage tanks [5]. More recently, the Dengue Control Programme in the northeastern Brazilian state of Ceará has used five non-native larvivorous fish species (Betta splendens, Trichogaster trichopteros, Astyanax fasciatus, P. sphenops, and P. reticulata) to combat Ae. aegypti larval infestation [15]. Another example was reported in Cambodian villages, where P. reticulata reduced dengue-carrying Ae. aegypti larval infestation by 79%, compared to control villages [16].

Water tanks and water storing practices

In each village, water is supplied via a deep bore well system operated by the local Gram Panchayats. Due to irregular water supplies, the villagers store water in indoor cement tanks for washing, bathing and also for drinking. Aedes mosquitoes mainly breed within these tanks. On an average 2.6 ± 0.6 (range 0-6) indoor cement tanks were recorded in each household. Two-thirds of villagers cleaned their tanks once in a week (as per the KAP survey), and IEC alone did not reduce the level of Aedes larval breeding. This is because of a faulty design in these tanks, in which the base of each tank is lower than the ground level making it difficult to completely empty. A few inches of water is retained at the bottom of the tanks, allowing Aedes larvae to rest at the bottom during the cleaning of mosquito-positive tanks. Moreover, each tank is attached to an oven for heating water for bathing (Figure 2). This helps maintain a favourable temperature range of 24-26°C, which is conducive for the growth and development of mosquito larvae, especially Aedes. Other breeding habitats, namely outdoor cement tanks that were placed under shades for cattle feeding, small plastic containers, and earthen pots where fish could not be introduced, also supported Aedes larval breeding. This knowledge was conveyed to villagers through IEC campaign in an effort to prevent larval breeding (Figure 3).

Currently, commercially available plastic moulded mosquito-proof water storage tanks may be used for the prevention of mosquito breeding (Figure 4). One-month observations provided by our study suggested that Poecilia introductions in water storage tanks combined with IEC is an alternative method of Aedes control. IEC improved the survival and handling of this fish at the community level. This was not observed when using Gambusia. Poecilia are available in many villages in Karnataka used in the malaria control programme, and can be easily grown in wells and small tanks. Furthermore, the local religious trust propagated Poecilia in their garden tanks and helped supply them to the local villagers. After completion of the trials, Poecilia was reintroduced after IEC in Srinivaspura and Gambusia was replaced with Poecilia in Balmanda for control of Aedes larvae.

During the control programme, a weekly application of an organophosphorus compound (Temephos; 1 ppm per litre) is recommended for the control of Aedes larvae in potable water. Being a chemical insecticide, this is not liked by many and is logistically difficult to use including its prohibitive costs. Moreover, Aedes can potentially develop resistance to this compound if used for a long period, as reported in Brazil after 30 years of its use [17]. Larvivorous fish are therefore suggested as the best option for controlling Aedes larval infestation, as they are both sustainable and cost effective. The operational cost was calculated at 0.50 (US$ 0.011, 1 US$= 47) per capita per application. Thus, monthly monitoring and application of Poecilia may be recommended after proper IEC. Consequently, the community is likely to facilitate the long-term control of mosquitoes, thereby preventing the diseases transmitted by these vectors.

Our study successfully determined the comparative efficacy of artificially maintaining populations of two non-native fish to control mosquito vectors. The use of Poecilia + IEC is further highly recommended for areas still affected by chikungunya in Karnataka [18], while Gambusia should be avoided as populations of this species are not sustained in domestic containers. Effective Aedes control should be formulated on the basis of local situations. Therefore, day-to-day water storing practices, 'vector sanitation' involving the National Rural Health Mission, the Gram Panchayat, and local community are essential for a successful and sustainable programme.

¹ taluka is a secondary revenue division of a district.