Effectiveness of winter larviciding as a malaria vector control intervention in selected rural areas of Botswana and Zimbabwe

Abstract

Introduction Successful control of malaria vectors requires the control of the larval and the adult stages. There is currently enough evidence on effectiveness of adult control methods through indoor residual spraying and insecticide treated nets, and these remain the main vector control methods in both Botswana and Zimbabwe. However, the growing resistance to insecticides used for indoor residual spraying and for treating long lasting insecticide treated nets is threatening the successes towards malaria elimination in the two countries. There is therefore need for implementation of other complimentary interventions such as larviciding. Products for implementing larviciding are available but the implementation is affected by insufficient evidence on the effectiveness of this strategy, particularly in rural areas. Both Botswana and Zimbabwe implement larviciding to some extent but the national malaria control programs of the two countries have not quantified the contribution of the intervention towards the overall malaria response. This study was conducted to assess the effectiveness of larviciding in selected rural areas of Botswana and Zimbabwe, particularly on larval density and adult mosquito density. Materials and Methods An experimental study was conducted in Molalatau and Mathathane villages of Botswana and Ward 33 (also known as Birchnough Bridge) of Zimbabwe. In Botswana, Mathathane was used as an intervention village while Molalatau was a control village. In Birchnough Bridge of Zimbabwe, the northern side of the ward (also known as Pfupi village) was used as an intervention while the southern part (also known as Tamanikwa village) was used as the control. The two villages in Zimbabwe were separated by the irrigation fields which acted as a buffer. Implementation of the intervention and data collection started at the end of May in Zimbabwe and in July in Botswana. Within both the intervention and control areas, all larval habitats were identified and mapped using portable hand-held geographic positioning system devices. Habitats in the intervention areas were treated using the commercial larvicide VectoBacÒ (Bacillus thuringiensis var. israelensis [Bti]) through community volunteers identified with the help of the local national malaria control programs. Every fortnight, larval counts were made from selected breeding habitats in both the intervention and control villages/wards to determine the effectiveness of the larvicide on the larvae. Within the same interval, adult mosquito sampling was conducted using exit window traps to capture mosquitoes which would enter houses for a blood meal and then want to rest outdoors. Pyrethrum spray catches were used to capture indoor resting mosquitoes. All sampled larvae and captured adult mosquitoes were identified to genera. Additionally, interviews were conducted with members of the community to understand their perceptions on effectiveness and acceptability of larviciding. Random-effects Poisson regression was used to compare intervention with control areas with respect to larval and adult mosquito counts. This was done using Stata Release 13, (StataCorp, College Station, TX: StataCorp LP). The incidence rate ratio (IRR) for treatment was of primary importance. Thematic analysis was used to analyse qualitative data from the interviews. Findings There was a significant overall effect of 92% and 65% on mosquito larvae in study sites of Botswana and Zimbabwe respectively following the application of larvicide (p<0.001). The effect on the early and late larval stages was 77% (P<0.001) and 91% (P<0.001) respectively for both countries combined. The average marginal effect of larviciding on the mosquito larvae taking interaction with time (period) into account, was -1.94 (95% CI: -2.42 to -1.46) with incidence rate ratio of 0.14, thus, an 86% larval effect attributable to the intervention for both countries combined. The estimated effect on the adult mosquitoes was 70% (IRR=0.303, p<0.001) in the intervention areas relative to the control for both countries combined, and was 77% (IRR=0.233, p<0.001) and 63% (IRR=0.369, p<0.001) for Botswana and Zimbabwe respectively. The volunteers who conducted larviciding also demonstrated to have competency in implementing larviciding. Additionally, larviciding was found to be an acceptable intervention in both countries, and factors influencing acceptability included; its importance as a supplementary method to IRS and LLINs; the desire to be protected from mosquito bites; known effectiveness; as well as willingness to support it. Identified strengths included: ability to kill mosquitoes early; safety; and ease of implementation. Conclusion This study demonstrated that larviciding using Bti is an effective vector control intervention in semi-arid rural areas of Botswana and Zimbabwe and the two countries should consider scale-up of the intervention, only in areas that have few, fixed and findable mosquito breeding habitats. Additionally, as part of the broader roll-out, local communities should be mobilised and utilised to ensure sustainability. Implementation should also be accompanied by detailed operational manuals.