Background Hypoxia may contribute to the development and progression of diabetic macular oedema (DMO) and diabetic retinopathy. The photoreceptors are the most metabolically active cells in the retina and the rod photoreceptors consume the maximal amount of oxygen for dark adaptation. Therefore, a lightmask emitting 500- to 505-nm light and worn at night during sleep may prevent rod-related dark adaptation and, thereby, reduce the retinal oxygen consumption and improve or prevent progression of DMO and diabetic retinopathy as evidenced in small short-term studies. Objectives The clinical effectiveness and safety of using a lightmask to prevent dark adaptation as a treatment and preventative option was evaluated for DMO over 24 months. Design A Phase III, multicentre, prospective, single-masked, randomised controlled clinical trial that evaluated the clinical effectiveness and safety of a lightmask that prevents dark adaptation to treat and prevent non-central DMO versus standard care (i.e. non-lightmask arm) at 24 months. A subset of participants also enrolled for a mechanistic substudy that evaluated the role of hypoxia in DMO. Setting Fifteen NHS clinical sites in the UK. Participants Adults with non-centre-involving DMO. Intervention The participants were randomly assigned (1 : 1) to being offered a lightmask (Noctura 400 Sleep Masks, Polyphotonix Medical Ltd, Durham, UK) to wear at night during sleep to prevent dark adaptation or to the non-lightmask arm (standard care) for 24 months. The participants were evaluated every 4 months. Main outcome measure The primary outcome was defined as a change in retinal thickness at the zone of maximum thickness measured by spectral domain optical coherence tomography at 24 months in the study eye, analysed using a linear mixed-effects model that estimated adjusted treatment effects at both 12 and 24 months. Results A total of 308 participants were recruited between April 2014 and May 2015. A total of 155 participants in the lightmask arm and 153 in the control arm contributed to the intention-to-treat strategy. The lightmask arm did not show any difference in effect compared with the non-lightmask arm [adjusted mean difference between arms –0.65 µm, 95% confidence interval –6.90 to 5.59 µm; p = 0.84). The compliance of wearing the lightmask was suboptimal, as recorded electronically, from the returned lightmasks. There were no lightmask-related severe adverse events. The mechanistic study aimed at studying the effect of improving hypoxia by inhaling 100% oxygen or by offering lightmasks to wear during sleep at night over 12 months did not demonstrate any clinically beneficial effect on DMO. Conclusion The lightmask, as offered in this trial to prevent dark adaptation, did not show any treatment or preventative effect in participants with non-central oedema or diabetic retinopathy at 24 months. No effect was seen as early as 4 months when compliance was higher. Future work Future research is needed to evaluate the role of rod-induced hypoxia in the pathogenesis of DMO and diabetic retinopathy. Trial registration Current Controlled Trials ISRCTN85596558. Funding This project was funded by the Efficacy and Mechanism Evaluation programme, a Medical Research Council and National Institute for Health Research partnership. The report will be published in full in Efficacy and Mechanism Evaluation; Vol. 6, No. 2. See the NIHR Journals Library website for further project information. The lightmasks were purchased at a discounted rate from Polyphotonix Medical Ltd.