Twelve studies (n=225) were included in the review: 7 double-blind placebo-controlled trials and 5 uncontrolled case series.
The quality of the included studies ranged from 0.21 to 0.85 on Cho and Bero's scale, and from 0.21 to 0.74 on the sleep-specific scale.
DSPS (3 placebo-controlled trials, 5 case series).
One controlled trial showed a large effect size for the effect of melatonin on sleep onset (effect size +2.61). Two other controlled trials found results which favoured melatonin, with effect sizes of +1.77 (self-report) and +1.21 (actigraphy), respectively; however, there were methodological concerns with these trials. All the case series reported that melatonin was an effective treatment, often without clear definitions of efficacy or statistical support.
PI (4 placebo-controlled trials).
One trial reported a statistically significant (p<0.05) increase in daytime alertness and total sleep time in the melatonin group but not the placebo group; none of the other trials reported any statistically significant effects of melatonin treatment.
Few adverse events were reported; where found, these were headache, daytime tiredness and an unusual taste in the mouth.
Poor reporting of sleep parameters in the included studies meant that the planned meta-analysis was not conducted.