BACKGROUND: Although effective topical repellents provide personal protection against malaria, whether mass use of topical repellents in addition to long-lasting insecticidal nets can contribute to a further decline of malaria is not known, particularly in areas where outdoor transmission occurs. We aimed to assess the epidemiological efficacy of a highly effective topical repellent in addition to long-lasting insecticidal nets in reducing malaria prevalence in this setting. METHODS: A cluster randomised controlled trial was done in the 117 most endemic villages in Ratanakiri province, Cambodia, to assess the efficacy of topical repellents in addition to long-lasting insecticidal nets in controlling malaria in a low-endemic setting. We did a pre-trial assessment of village accessibility and excluded four villages because of their inaccessibility during the rainy season. Another 25 villages were grouped because of their proximity to each other, resulting in 98 study clusters (comprising either a single village or multiple neighbouring villages). Clusters were randomly assigned (1:1) to either a control (long-lasting insecticidal nets) or intervention (long-lasting insecticidal nets plus topical repellent) study group after a restricted randomisation. All clusters received one long-lasting insecticidal net per individual, whereas those in the intervention group also received safe and effective topical repellents (picaridin KBR3023, SC Johnson, Racine, WI, USA), along with instruction and promotion of its daily use. Cross-sectional surveys of 65 randomly selected individuals per cluster were done at the beginning and end of the malaria transmission season in 2012 and 2013. The primary outcome was Plasmodium species-specific prevalence in participants obtained by real-time PCR, assessed in the intention-to-treat population. Complete safety analysis data will be published seperately; any ad-hoc adverse events are reported here. This trial is registered with ClinicalTrials.gov, number NCT01663831. FINDINGS: Of the 98 clusters that villages were split into, 49 were assigned to the control group and 49 were assigned to the intervention group. Despite having a successful distribution system, the daily use of repellents was suboptimum. No post-intervention differences in PCR plasmodium prevalence were observed between study groups in 2012 (4·91% in the control group vs 4·86% in the intervention group; adjusted odds ratio [aOR] 1·01 [95% CI 0·60-1·70]; p=0·975) or in 2013 (2·96% in the control group vs 3·85% in the intervention group; aOR 1·31 [0·81-2·11]; p=0·266). Similar results were obtained according to Plasmodium species (1·33% of participants in the intervention group vs 1·10% in the intervention group were infected with Plasmodium falciparum; aOR 0·83 [0·44-1·56]; p=0·561; and 1·85% in the control group vs 2·67% in the intervention group were infected with Plasmodium vivax; aOR 1·51 [0·88-2·57]; p=0·133). 41 adverse event notifications from nine villages were received, of which 33 were classified as adverse reactions (11 of these 33 were cases of repellent abuse through oral ingestion, either accidental or not). All participants with adverse reactions fully recovered and 17 were advised to permanently stop using the repellent. INTERPRETATION: Mass distribution of highly effective topical repellents in resource-sufficient conditions did not contribute to a further decline in malaria endemicity in a pre-elimination setting in the Greater Mekong subregion. Daily compliance and appropriate use of the repellents remains the main obstacle.