Is it safe to take melatonin every night?
For healthy adults, melatonin is safe with short-term use, but as a sleep aid for chronic insomnia it works considerably less well than standard sleep medications; pregnant women, breastfeeding women and children outside a medical indication are better off avoiding it.
Short-term melatonin use in adults is relatively well studied. The side effects reported in randomised trials are mild -- think dizziness, headache, nausea and daytime sleepiness. They are comparable to what people report with a placebo, and serious adverse effects have not been identified in the available randomised trials1.
Daily long-term use in adults shows a similar picture: side effects remain mild and are no greater than with placebo. Nevertheless, the volume of long-term data is considerably smaller than for short-term use, which warrants caution. At high doses there is insufficient evidence to guarantee safety1,2. In older adults, residual daytime sedation is a real risk, even at low doses3.
For pregnant women, breastfeeding women, children and adolescents outside specific medical indications, melatonin is explicitly discouraged or at least insufficiently studied. There are simply too few human safety studies available to properly assess risks for the mother, the child or the developing adolescent1.
A practical point of concern is that melatonin is an unregulated supplement in many countries. The actual dose in a product can differ considerably from what is stated on the label, making dosing unpredictable.
As a sleep aid for chronic insomnia, melatonin scores substantially worse than common agents such as benzodiazepines, zolpidem or orexin receptor antagonists in large network meta-analyses. It has some effect on sleep onset latency, but the clinical benefit in adults with chronic insomnia is limited4,5. An exception is prolonged-release melatonin in children with autism spectrum disorder: after 52 weeks, studies showed significant improvements in sleep duration and sleep onset latency, with fatigue (5.3%) and mood swings (3.2%) as the main side effects6.
All claims are based on one primary review (PMID 26692007) and two network meta-analyses (PMID 35843245, 36701954), supplemented by a study in older adults (PMID 27751669), an RCT in children with autism (PMID 30132686) and a review with a commercial conflict of interest (PMID 36235587). The evidence base for long-term safety is moderate to limited.