A summary of published research, not medical advice. See the Disclaimer.
This page sets oral melatonin beside other supplements, over-the-counter antihistamines and prescription drugs that have been studied for sleep.
How we grade evidence · Last reviewed: November 2026
What you need to know
Direct comparisons with melatonin are few, and we found none against valerian, chamomile, magnesium, ashwagandha, L-theanine, diphenhydramine or doxylamine. In a network analysis of adults with insomnia disorder, benzodiazepines, eszopiclone, zolpidem and zopiclone beat melatonin on self-rated sleep quality in short-term use, with weak evidence, and those same four caused more side effects than placebo. Melatonin shortened the time to fall asleep a little compared with placebo (moderate evidence), by about 7 minutes in pooled trials, and made little or no difference to chronic insomnia overall (weak evidence). Comparisons between the rows for each sleep aid against placebo are indirect, because the trials differ in who was studied, doses and outcomes. In 16 healthy older adults, prolonged-release melatonin did not impair psychomotor or driving tests, zolpidem did, and the two together were worse than zolpidem alone. Melatonin trials used 0.15 to 12 mg a day, most for 4 weeks or less, so long-term comparisons are missing.
These rows come from trials or network analyses that set melatonin, or a drug class that acts on melatonin receptors, beside another sleep aid. A network analysis combines direct and indirect evidence, and the abstracts do not say how much of each sits behind a result. Our PubMed searches found no head-to-head trial of melatonin against valerian, chamomile, magnesium, ashwagandha, L-theanine, diphenhydramine or doxylamine.
| Comparison | Verdict | What was measured | Studied in |
|---|---|---|---|
| Benzodiazepines, eszopiclone, zolpidem or zopiclone give better sleep quality than melatonin | May help (weak evidence) Network analysis of 154 trials, certainty rated moderate to very low |
Self-reported | Adults with insomnia disorder |
| Orexin-blocking drugs beat melatonin receptor drugs | May help (weak evidence) Network analysis of a drug class, not melatonin itself |
Measured | Adults with primary insomnia |
| Clonazepam beats prolonged-release melatonin in REM sleep behaviour disorder | May help (weak evidence) One open-label trial of 34 people, trade-offs on both sides |
Measured | People with isolated REM sleep behaviour disorder |
| Temazepam against prolonged-release melatonin in advanced cancer | Unknown (not tested) 21 people, both beat placebo, no test of one against the other in the abstract |
Self-reported | People with advanced cancer and insomnia |
| Single doses of zopiclone, zaleplon, temazepam and melatonin | Unknown (very weak evidence) One crossover study of 23 adults, rankings changed between time points |
Measured | Adults aged 21 to 53 |
Common hypnotics against melatonin. May help (weak evidence): a network analysis of 154 double-blind trials (44,089 people) found benzodiazepines, eszopiclone, zolpidem and zopiclone more effective than melatonin, ramelteon and zaleplon for self-rated sleep quality in short-term use (standardised mean difference 0.27 to 0.71) [1]. The authors rated certainty moderate to very low, which caps the grade. This verdict is about those drugs, and the same analysis found them worse than placebo for side effects. The abstract gives no long-term melatonin result.
Orexin blockers against melatonin receptor drugs. May help (weak evidence): a network analysis of 69 studies (17,319 patients) found orexin receptor antagonists better than melatonin receptor agonists for time to fall asleep, time awake after sleep onset and sleep efficiency [2]. The abstract reports the class, not melatonin alone.
REM sleep behaviour disorder. May help (weak evidence): in an open-label trial of 34 people, 4 weeks of clonazepam reduced muscle activity during REM sleep on visual scoring, and prolonged-release melatonin did not. Daytime sleepiness and insomnia symptoms fell with melatonin but not clonazepam, and depressive symptoms rose with clonazepam [3].
Advanced cancer. Unknown (not tested): in 21 people with advanced cancer, temazepam and prolonged-release melatonin each lowered insomnia severity scores compared with placebo by day 8, by 9.1 and 9.6 points [4]. Melatonin had eight participants, temazepam nine and placebo four, and the abstract reports no test between the two active groups.
Single doses. Unknown (very weak evidence): in a crossover study of 23 adults, the order of sleep-inducing effect was zopiclone, zaleplon, melatonin, temazepam before psychomotor testing, and zopiclone, melatonin, zaleplon, temazepam after it [5]. The study was designed around military operations, and the abstract does not say participants had a sleep disorder.
This is an indirect comparison across separate trials. Populations, doses, durations and outcomes differ, so these rows cannot be ranked against each other and do not show that one option is better than another.
| Option and outcome | Verdict | What was measured | Studied in |
|---|---|---|---|
| Melatonin, time to fall asleep | Probably helps a little (moderate evidence) Pooled analysis of 19 trials in 1,683 people, about 7 minutes faster |
Measured | Adults and children with primary sleep disorders |
| Melatonin, chronic insomnia overall | May make little or no difference (weak evidence) Network analysis found no overall material benefit |
Self-reported | Adults with insomnia disorder |
| Melatonin, sleep in Alzheimer's disease | May make little or no difference (weak evidence) Cochrane review, certainty rated low |
Measured | Adults with Alzheimer's disease and sleep disturbances |
| Ramelteon, time to fall asleep | Probably helps a little (moderate evidence) Pooled trials found 9 to 14 minutes, overall benefit not shown in a network analysis |
Measured | Adults with insomnia disorder |
| Benzodiazepines and Z-drugs, sleep | Probably helps (moderate evidence) Network analyses of 153 and 154 trials, reviews rated time to sleep as moderate certainty |
Measured | Adults with insomnia |
| Orexin-blocking drugs, sleep | Helps (strong evidence) About 32 minutes more measured sleep, rated high certainty, reviews differ on tolerability |
Measured | Adults with insomnia |
| Orexin-blocking drugs, sleep in Alzheimer's disease | Probably helps (moderate evidence) Cochrane review, one trial of 274 people, 28 minutes more sleep |
Measured | Adults with mild to moderate Alzheimer's disease |
| Low-dose doxepin or trazodone, sleep quality | May help a little (weak evidence) Cochrane review, small and mostly short trials |
Self-reported | Adults with insomnia |
| Diphenhydramine or doxylamine, sleep | May help (weak evidence) Small diphenhydramine trials with few significant results, doxylamine in one network analysis |
Self-reported | Adults with mild insomnia or insomnia disorder, older adults |
| Valerian, sleep | Unclear: trials disagree (weak evidence) Reviews reach opposite conclusions, products varied |
Self-reported | Adults with sleep problems or insomnia |
| Chamomile, sleep quality | May help (weak evidence) Pooled trials varied widely, blinding rarely checked |
Self-reported | Adults, healthy or with illness |
| Magnesium, sleep | Unknown (very weak evidence) Few small trials in older adults, review flagged low quality |
Measured | Older adults with insomnia |
| Ashwagandha, sleep quality | May help a little (weak evidence) Review that pooled 5 trials with 400 people, small overall improvement |
Self-reported | Adults with and without insomnia |
| L-theanine, sleep quality | May help a little (weak evidence) Pooled self-rated results, few studies of L-theanine alone |
Self-reported | People of varied ages and health |
Melatonin. Probably helps a little (moderate evidence) for time to fall asleep: 19 trials with 1,683 participants found about 7 minutes faster sleep onset and about 8 minutes more sleep than placebo [6]. In people with insomnia the reduction was 7.2 minutes [7]. May make little or no difference (weak evidence) for chronic insomnia overall: a network analysis concluded that melatonin showed no overall material benefit [1]. The Melatonin article covers delayed sleep phase disorder and jet lag, which have no comparison here.
Alzheimer's disease. One Cochrane review covered both drug types, in separate trials. May make little or no difference (weak evidence) for melatonin: doses up to 10 mg over 8 to 10 weeks changed total night sleep by 10.68 minutes, 95% CI -16.22 to 37.59 (2 studies, n=184), low certainty [8]. Probably helps (moderate evidence) for orexin blockers: 4 weeks added 28.2 minutes of night sleep (1 study, n=274), moderate certainty [8]. The review found no trials of benzodiazepines or Z-drugs in dementia [8].
Ramelteon. Probably helps a little (moderate evidence): at 4 weeks, pooled trials found time to fall asleep shorter by 8.74 minutes self-rated and 14 minutes measured [9]. Across melatonin receptor drugs as a class, the reductions were 7.73 and 7.04 minutes, at high and moderate certainty [10]. A network analysis still found no overall material benefit for ramelteon [1], which holds the grade at moderate.
Benzodiazepines and Z-drugs. Probably helps (moderate evidence): a network analysis of 153 trials found non-benzodiazepine drugs shortened time to fall asleep by about 12 minutes measured and 10 minutes self-rated, at moderate certainty, and cut time awake after sleep onset by about 14 to 17 minutes [10]. A second analysis found benzodiazepines, eszopiclone, zolpidem and zopiclone better than placebo in short-term use [1].
Orexin-blocking drugs. Helps (strong evidence): total sleep rose by 31.81 minutes measured and 21.62 minutes self-rated, at high certainty [10]. One review rated their tolerability as superior [2], and another listed daridorexant and suvorexant among drugs with poor tolerability or missing long-term data [1]. The grade is for sleep, not tolerability.
Low-dose doxepin and trazodone. May help a little (weak evidence): a Cochrane review of 23 trials (2,806 people) found there may be a small improvement in sleep quality with short-term use, and judged safety and tolerability uncertain because adverse events were poorly reported [11].
Diphenhydramine and doxylamine. May help (weak evidence): in 184 adults with mild insomnia, diphenhydramine raised sleep efficiency compared with placebo over 14 days, and sleep studies showed no group differences [12]. In older adults, it improved only the number of awakenings, 1.7 against 2.0 [13]. Doxylamine beat placebo in one network analysis [1], and its abstract gives no separate size for it.
Valerian. Unclear: trials disagree (weak evidence): one meta-analysis of 16 studies (1,093 patients) reported a relative risk of improved sleep of 1.8 in the 6 studies with that outcome, with evidence of publication bias [14], and a 2020 review of 60 studies concluded valerian could be effective [15]. A review of 29 controlled trials found most showed no difference from placebo [16], and a 2024 umbrella review found no evidence of efficacy for insomnia, though it found subjective improvement in sleep quality [17]. An analysis of herbal medicines in insomnia found no significant difference from placebo on any of 13 efficacy measures [18].
Chamomile. May help (weak evidence): ten studies (772 participants) were reviewed, and pooled Pittsburgh Sleep Quality Index scores improved by 1.88 points, with high variation (I² = 88.4%) [19]. Sleep efficiency and daytime functioning did not improve, and only one study checked blinding [19]. An earlier review found one trial in insomnia, which showed no change in insomnia severity [20], and an analysis of herbal medicines in insomnia, chamomile included, found no significant difference from placebo [18].
Magnesium. Unknown (very weak evidence): a meta-analysis of 3 trials (151 older adults) found sleep onset about 17 minutes shorter than placebo, and the evidence was rated low to very low quality [21]. Magnesium forms compared covers single-form trials.
Ashwagandha. May help a little (weak evidence): five trials with 400 participants showed a small overall improvement, larger in people with insomnia, at 600 mg/day or more and at 8 weeks or more [22].
L-theanine. May help a little (weak evidence): a meta-analysis of 18 articles found small improvements in self-rated time to fall asleep and overall sleep quality (standardised mean differences 0.15 and 0.43), and the authors noted a lack of studies on L-theanine alone [23].
Guidelines. These are expert panels, not trials. A 2017 US panel suggested clinicians not use melatonin, valerian, diphenhydramine or tryptophan for chronic insomnia in adults, and made weak suggestions in favour of drugs including zolpidem, eszopiclone, temazepam, ramelteon, doxepin and suvorexant [24]. A 2017 European guideline advised against antihistamines, melatonin and plant-based products for insomnia, on low- to very-low-quality evidence [25].
Apart from the zolpidem and melatonin rows, which come from the same trial, this is also an indirect comparison across separate studies. The abstracts we read did not compare melatonin's side effects with other sleep aids.
| Side effect | Verdict | How often vs placebo | Source | Studied in |
|---|---|---|---|---|
| Melatonin: daytime sleepiness | May happen (weak evidence) Reported in 1.66% of trial participants, no placebo comparison given |
1.66% of trial participants; placebo rate not reported | [26] | People with sleep disorders in 37 trials |
| Benzodiazepines, eszopiclone, zolpidem, zopiclone: any side effect | Probably happens (moderate evidence) Network analysis, certainty rated high to very low |
Odds ratio 1.27 to 2.78 against placebo, doxepin, seltorexant and zaleplon | [1] | Adults with insomnia disorder |
| Zolpidem: slower psychomotor and driving performance within hours | May happen (weak evidence) One crossover trial of 16 healthy older adults |
Impaired at 1 and 4 hours; no rates reported | [27] | Healthy adults aged 55 and over |
| Prolonged-release melatonin: slower psychomotor and driving performance | May not increase it (weak evidence) One crossover trial of 16 healthy older adults |
No impairment on any task vs placebo | [27] | Healthy adults aged 55 and over |
| Orexin-blocking drugs and ramelteon: sleepiness | May happen (weak evidence) Network analysis, no certainty rating in the abstract |
Relative risk 6.57 lemborexant, 3.32 suvorexant, 2.19 ramelteon | [28] | Adults with insomnia |
| Diphenhydramine: serious events or next-day effects | May not increase it (weak evidence) Two small trials, no serious events or significant residual effects |
No serious adverse events in 184 adults; adverse event counts similar in older adults | [12][13] | Adults with mild insomnia, older adults |
| Valerian: any side effect | Unclear: trials disagree (weak evidence) Reviews differ, one counted more events than placebo |
Not reported | [14][16][18] | Adults with sleep problems or insomnia |
| Ashwagandha: liver injury with jaundice | Unknown (very weak evidence) Reports of single patients only, no controlled comparison, no rate available |
Case reports only | [29][30] | Adults in case series |
In the zolpidem trial, zolpidem impaired psychomotor and driving performance at 1 and 4 hours after dosing. Next-morning performance was unaffected, except for memory recall [27]. The melatonin review covered doses of 0.15 to 12 mg/day, with very few adverse events judged serious [26]. The side-effect network analysis lists no melatonin result in its abstract [28].
This lists what studies and labels report. A pharmacist or prescriber can say whether it applies to a specific person.
| Combined with | What happened | Verdict | Source | Studied in |
|---|---|---|---|---|
| Prolonged-release melatonin with zolpidem | Impairment from zolpidem on psychomotor, driving and early memory tests was worse with melatonin added; no pharmacokinetic interaction found | May happen (weak evidence) One crossover trial of 16 healthy older adults |
[27] | 16 healthy older adults |
The Melatonin article lists label interactions with other drugs.
This lists what studies and labels report. A pharmacist or prescriber can say whether it applies to a specific person.
These are the amounts used in the studies. They are not recommendations.
Melatonin trials used 0.15 to 12 mg/day, monitored for up to 29 weeks, most for 4 weeks or less [26]. Dementia trials used doses up to 10 mg [8]. Prolonged-release melatonin was given at 2 mg in the zolpidem and clonazepam trials [27][3], and time-released melatonin at 6 mg as a single dose [5].
Magnesium trials in older adults gave less than 1 g, up to three times a day [21]. Ashwagandha results were larger at 600 mg/day or more [22]. Valerian doses, preparations and lengths of use varied considerably [14]. Doses of prescription drugs are not listed here.