BACKGROUND: Night shift work is associated with sleep disturbances, obesity, and cardiometabolic diseases. Disruption of the circadian clock system has been suggested to be an independent cause of type 2 diabetes and cardiovascular disease in shift workers. We aimed to improve alignment of circadian timing with social and environmental factors with administration of melatonin. METHODS: In a randomized, placebo-controlled, prospective study, we analysed the effects of 2 mg of sustained-release melatonin versus placebo on glucose tolerance, insulin resistance indices, sleep quality, circadian profiles of plasma melatonin and cortisol, and diurnal blood pressure profiles in 24 rotating night shift workers during 12 weeks of treatment, followed by 12 weeks of wash-out. In a novel design, the time of melatonin administration (at night or in the morning) depended upon the shift schedule. We also compared the baseline profiles of the night shift (NS) workers with 12 healthy non-night shift (NNS)-working controls. RESULTS: We found significantly impaired indices of insulin resistance at baseline in NS versus NNS (p < 0.05), but no differences in oral glucose tolerance tests nor in the diurnal profiles of melatonin, cortisol, or blood pressure. Twelve weeks of melatonin treatment did not significantly improve insulin resistance, nor did it significantly affect diurnal blood pressure or melatonin and cortisol profiles. Melatonin administration, however, caused a significant improvement in sleep quality which was significantly impaired in NS versus NNS at baseline (p < 0.001). CONCLUSIONS: Rotating night shift work causes mild-to-moderate impairment of sleep quality and insulin resistance. Melatonin treatment at bedtime improves sleep quality, but does not significantly affect insulin resistance in rotating night shift workers after 12 weeks of administration.