Melatonin for Anxiety
Melatonin has modest, moderate-certainty evidence for easing short-term anxiety right before surgery or procedures, but there is essentially no good evidence it treats generalized or chronic anxiety, and it is not approved for any anxiety use. Because these two situations point in opposite directions, the overall grade is weak and disputed.
Why this grade7-layer evidence engine
The grade is held at Weak/Disputed because the evidence splits sharply by setting. For pre-procedure anxiety, a 2020 Cochrane systematic review (PMID 33319916, n=2319) found melatonin probably lowers preoperative anxiety versus placebo (about 12 points on a 100 mm scale) and works about as well as benzodiazepines like midazolam. This is only moderate certainty, however, since most included trials had unclear or high risk of bias, so it cannot rise to a top grade.
Even within the surgical setting the newer trials disagree. A 2024 sublingual-melatonin trial before cesarean section (PMID 38841008, n=80) showed a clear benefit, but the 2024 MAGIC trial in anxious children (PMID 37953202, n=110) found melatonin clearly inferior to midazolam and was stopped early for futility, and a 2022 trial before hernia repair (PMID 36106858, n=33) found no benefit at all.
Outside surgery the case largely collapses. A 2022 review (PMID 36555831) notes that data for generalized or chronic anxiety remain limited, and regulators offer no support: the FDA finds no basis that melatonin is even safe-by-default (GRAS), EFSA's approved claims cover only sleep onset, and NHS/WHO do not back it for anxiety. Notably, Mayo, Cleveland Clinic, and Harvard list anxiety as a possible side effect of melatonin, not a treatment, which is why the verdict stops short of an endorsement for everyday anxiety.
Scoring transparency
All scores computed by a 7-layer evidence engine — fully auditable▸View the full decision path (audit trail)
- compute_raw_score — 加權公式: L2×0.30 + L3×0.25 + L5×0.25 + L11×0.10 + L1×0.10 = 0.616
- tier_from_score — 依分數區間映射至 tier letter
- apply_hec_rules — 無高階證據可裁決
- tier_strict_requirement_check — | B→C 因 scope.conflation_risk=true 且 L11 獨評較低 (B7-2 tier cap)
- detect_disputes — 偵測到 1 個 hard + 0 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status