Rozerem is a selective melatonin agonist. It acts on melatonin receptors at the suprachiasmatic nucleus. It is prescribed as a non-addictive sleep aid for people having difficulties with the onset of sleep, i.e., falling asleep in the evening.
While melatonin itself appears unlikely to be a molecule that directly induces sleep, it does have phase-resetting effects on the circadian clock. Thus, Rozerem appears ideal as an aid for extreme "owls" to help them fall asleep (if they need to wake up early in the morning, as some jobs require). By mimicking melatonin, it would phase-advance the clock by a couple of hours and make it easier to fall asleep at a more socially acceptable time.
I did not really spend much time thinking about this, but my wife just told me something interesting. Last night, she took Rozerem and fell asleep practically instantly and had a great night of sleep. But, last week, when she tried using Rozerem in the morning after coming back home from her 12-hour night shift, there was no effect.
Perhaps Rozerem, just like melatonin, is incapable of inducing instant 12-hour phase-shifts of the circadian clock in the SCN. Shifting by a couple of hours is fine, but shifting the cycle by 180 degrees is a different story altogether - it may take several days to accomplish.
Perhaps the clock in the SCN has its own Phase-Response Curve (PRC) to melatonin - presence of the molecule induces smaller or bigger (or not at all) phase-advances or phase-delays depending on the phase of the cycle in which it is applied.
So, perhaps Rozerem given in the evening hits the PRC at the phase in which a large (i.e., 2-3 hours) phase-advance is induced, thus placing the body at the right time for the onset of sleep. By the same logic, application of Rozerem in the morning may hit the PRC at the phase in which it has no effect or an effect in the wrong direction - phase-advance again, which would bring the body at the proper time for brunch! I guess I'll need to dig through the old literature on the melatonin PRCs in humans to see how it looks like and if this notion may be correct.
Perhaps if one works a night shift and repeatedly applies Rozerem every morning, there would be a gradual shift of the clock, over several days, until the correct phase is achieved at which the drug would work as advertised. Also, avoiding any conflicting cues to the clock (e.g., light, noise, etc.) would seem important to ensure that a morning dose of Rozerem has its intended effect for night-shift workers. All of this is late-night speculation, though, so do not take my word for it. It is based on a single data-point. More research is needed....
