Goldberg‐Stern 2012.

Methods	Randomized, double‐blind, cross‐over, placebo‐controlled trial
Participants	Inclusion criteria: aged ≥ 4 years; failure of ≥ 3 different AEDs to control seizures; therapeutic stability for 2 months prior to the study; (iv) ≥ 4 seizures in the 3 weeks prior to the initiation of the study Exclusion criteria: children with vasculitis, cardiovascular anomalies, or blindness N: 12 M: 5; F: 7 Age: 9‐32 years (range) All participants had epilepsy and received AEDs Seizure type: symptomatic generalized (4); symptomatic partial (1); symptomatic partial with secondary generalization (1); cryptogenic partial with secondary generalization (3); idiopathic partial (1) Duration of the trial (including follow‐up): 7 weeks
Interventions	Melatonin (10 mg daily at bedtime) compared with placebo (see text for more details)
Outcomes	Seizure freedom/seizure frequency: mean number of diurnal seizures was 7.75 during placebo treatment and 4.6 during melatonin treatment. Three participants showed a decrease of ≥ 50% in diurnal seizures during melatonin treatment compared to placebo Adverse events: none reported Quality of life: not systematically evaluated
Notes	5/10 participants analyzed were aged < 18 years. All adults had childhood‐onset epilepsy
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Authors did not specify how randomization was performed. Randomization was made by a pharmacist, not connected to the study
Allocation concealment (selection bias)	Low risk	Placebo and melatonin tablets identical in size and appearance
Blinding (performance bias and detection bias) All outcomes	Low risk	Placebo and melatonin tablets were identical in size and appearance
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	2 participants were excluded from the analysis because of lack of compliance in completing the diaries to assess seizure occurrence, and behaviour and sleep features. No further information on these 2 participants were reported
Selective reporting (reporting bias)	Low risk	All data regarding outcomes of interest for this systematic review were reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Placebo and melatonin tablets were identical in size and appearance
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Placebo and melatonin tablets were identical in size and appearance