The literature on melatonin includes reports of adverse effects reported with human melatonin use at 10 mg/day or less to include central nervous system effects (e.g., somnolence, headaches, increased frequency of seizures, nightmares), cardiovascular effects (e.g., hypotension or hypertension), gastrointestinal effects (e.g., diarrhea, abdominal pain), and dermatological effects. In addition, melatonin use at higher doses (240–1,000 mg/ day) in a small number of subjects was associated with hormonal changes that were inconsistent among the different reports. This summary explores the quality and other factors that may have contributed to serious adverse events.

The available data on melatonin safety in humans are based mostly on reports of studies with small numbers of participants that were not designed to evaluate the safety of melatonin. This monograph is based on 48 studies and reports of melatonin use in humans that included over 1,000 subjects (in the melatonin arms). The range of melatonin doses used in these studies is wide, 0.1 to 1,000 mg. These studies vary from one-time ingestion of melatonin to 6 months of daily ingestion. Many studies omit statements about adverse effects or state that no adverse effects were observed without describing the safety parameters monitored. Moreover, there is insufficient information on interactions of melatonin with drugs or other dietary supplements. Most available studies were conducted with adults, and little information is available for infants and young children regarding adverse effects, specifically concerning possible melatonin-induced alterations of pubertal development. Likewise, there is no information on safety of melatonin use by pregnant or lactating women.

The LD₅₀ of melatonin in animal models (1–3 g/kg body weight for oral doses in rats and mice) far exceeded the typical doses used as a dietary supplement in humans (0.5–10 mg/d). At a dose of 20 mg/L in drinking water, melatonin was associated with an increased rate of spontaneous tumors in one strain (CBA) of female mice. However, more recent studies by the same group showed the same amount of melatonin administered to another strain (SHR) of female mice had no effect on tumor rate. Thus the data on the effect of melatonin on tumor incidence in mice is inconclusive and this area of investigation should continue to be monitored. (This monograph focus is on understanding and interpreting these data, as limited resources were instead focused on human data.) In addition, it has been well established that melatonin has significant effects on the reproductive axis in animals (Reiter, 1991; Rivest et al., 1986). These effects might be undesirable if they occured in humans.