Background: Malaria remains a serious threat in the Greater Mekong Sub-region (GMS), not just from the direct impact on human health, but also from the emergence and spread of resistance to artemisinin, the last remaining effective antimalarial. Malaria control in this region is therefore a high priority on a global as well as local scale. In the southern region of the Lao People's Democratic Republic (Lao PDR) as across much of the GMS malaria vectors are found biting outdoors in the early evening before people are protected by long-lasting insecticidal nets (LLlNs). Therefore there is a need for additional malaria control tools that can protect people during these evening hours. Methods: Human landing catches in a village setting in southern Lao PDR were used to evaluate the protection from evening biting given by repellent lotions containing 10-20% N,Ndiethyl-m-toluamide (DEET). A randomised controlled trial was designed to test the effect of 15% DEET against malaria. A total of 1,597 households were recruited and randomised to either the repellent or a placebo lotion. All households were also provided with LLiNs. The acceptance and compliance with the repellent lotion was assessed through exit questionnaires and focus group discussions (FGDs). A meta-analysis was then carried out to put the results from the Lao PDR in context with other repellent trials. Findings: All DEET concentrations provided at least 96% protection from evening biting over five hours. However 15% DEET was determined to be the best choice of intervention over 10% DEET after also considering the results of other similar trials. Intention to treat analysis of the randomised controlled trial found no difference between treatment arms after accounting for gender and socio-economic status (incidence rate ratio 0.96, 95% confidence interval 0.54- 1.71, p=O.886). According to protocol analyses of participants who used the lotions over 90% of the time also found no effect from repellent use after other factors had been taken into account (incidence rate ratio 1.45, 95% confidence interval 0.53-3.99, p=0.467). The most important predictor of malaria incidence was socio-economic score which indicated that lower wealth was significantly associated with an increased malaria risk. Although the repellent was well received with over 90% of participants reporting that they liked using the lotions, compliance was still low with fewer than 60% of participants using the lotions more than 90% of the time. It emerged from FGDs that the assumption that local populations were protected from night biting if they were provided with LLiNs was not always true. Adult men and children reported spending time outdoors at night hunting and fishing. The protection from malaria by repellent use in this trial was lower than in other randomised controlled trials carried out in Bolivia, Pakistan and Tanzania. The meta-analysis found that repellent use was associated with a 33% reduction in P. falciparum incidence (95% CI 0.42-1.09, p=O.l1) and a 35% reduction in P. vivax incidence (95% CI 0.18-2.34, p=0.51), however neither figure reach significance. Interpretation: Limitations of this trial include the compliance level which was lower than in other trials. In addition the variability inherent in topical repellents may make them unsuitable for use as an intervention. The outcome of this trial shows that topical insect repellent is not a suitable wide-scale intervention against malaria and does not provide significant protection over and above LLiNs in an area of outdoor biting. However, repellents do undoubtedly reduce biting and therefore their potential to be effective intervention tools remains. Future work should concentrate on forms of repellent that can be better standardised such as impregnated clothing. If successful then further research into mosquito response to repellent is recommended including, where best to apply and the potential for the development of resistance.