The aim of this study was to evaluate the efficacy and safety of ramelteon 4, 8, 16 or 32 mg and placebo in Japanese patients with chronic insomnia using a randomized, double-blind, five-period crossover design. A total of 65 Japanese patients with chronic primary insomnia received ramelteon or placebo for two nights each in sleep laboratories. Changes in sleep parameters were assessed objectively by polysomnography and subjectively by postsleep questionnaires. Safety and tolerability was evaluated by assessment of the occurrence of adverse events, next-day residual effects and laboratory and ECG investigations. Ramelteon 8 and 32 mg significantly shortened the mean latency to persistent sleep in comparison with placebo, and there was a statistically significant trend for linear dose-response for this sleep parameter. Overall changes in sleep architecture were modest (<3% changes vs placebo), with increases in stage 1 and decreases in stage 3/4. Ramelteon was well tolerated, the most common adverse effect being somnolence, which was similar to placebo at doses up to 8 mg, but increased with higher doses. Next-day residual effects occurred no more frequently with ramelteon at any dose than with placebo. When compared with sleep latency data from a similarly-designed US study, there was no evidence of any ethnic differences in the efficacy of ramelteon between Japanese and US patients. Overall, ramelteon 8 mg showed the most favorable balance between sleep-promoting effects and tolerability. The unique efficacy profile of ramelteon, promoting sleep initiation without affecting other sleep parameters, may be due to its circadian shifting effect.