Melatonin Use in Young Children

Key Points

Question

What are the long-term outcomes (safety and effectiveness) associated with melatonin use in young children?

Findings

This systematic review of 19 studies found that melatonin prescribing practices, extended use, and overdoses in young children (aged 0-6 years) increased in the past 2 decades. Five interventional trials found that young children with autism spectrum disorder or related conditions fell asleep faster after treatment, but none examined efficacy in children with typical development or measured outcomes beyond 2 years.

Meaning

These findings suggest that medical professionals should discuss behavioral interventions with young children to reduce medication usage and overdose.

This systematic review examines the long-term safety and effectiveness associated with melatonin use among young children.

Abstract

Importance

Melatonin is the leading cause of unsupervised medication exposure and overdose in emergency departments for young children (aged 0-6 years). Existing literature has documented the benefits of melatonin in older children (aged 7-18 years) with neurologic conditions but has yet to examine young children.

Objective

To examine the long-term outcomes (safety and effectiveness) associated with melatonin use in young children.

Evidence Review

Nine databases, including Ovid MEDLINE, Embase, and Web of Science; 2 clinical trial registries; existing subject matter systematic reviews; and forward and backward citation of included articles were reviewed from inception to February 26, 2025, to identify observational and interventional studies that investigated the safety and effectiveness of exogenous melatonin on sleep in young children. Methodological quality was assessed using the Downs and Black Checklist.

Findings

Nineteen articles representing 12 observational studies, 6 experimental studies (hereafter, trials), and 1 protocol published between 2000 and 2025 were included. The observational studies reported on 9 years (range, 3-21 years) of data from Nordic or Australian registries or poisoning data from the US and Portugal. The trials included 167 young children with neurologic conditions and lasted a mean of 12.7 weeks (range, 2 weeks to 2 years). For safety, observational studies documented rises in prescribing practices, extended use, and overdoses, especially in the past decade. For effectiveness, trials provided evidence for improved sleep onset in young children with neurologic conditions (eg, autism spectrum disorder), with few adverse events. However, data regarding long-term outcomes for other behaviors and health outcomes were absent, and efficacy data were not available for children with typical development. The methodological quality of included studies was poor for 3 studies, fair for 9, and good for 6.

Conclusions and Relevance

These findings suggest a global rise in prescriptions without efficacy data on use in children with typical development, underscoring the need to identify strategies to prevent and reduce melatonin use in young children, as well as to improve adherence by pediatricians to evidence-based practice standards.

Introduction

Melatonin is the leading substance in cases of unsupervised medication ingestion and overdose among young children (aged 0-5 years) in US emergency departments.^1,2,3 The frequency of these ingestions and overdoses increased fivefold from 2009 to 2021^1,3; between 2019 and 2020, the increase was 37.9%.¹ Most cases (85%) involved young children. While most ingestions (97%) result in minor adverse effects, some lead to more serious outcomes, including death.¹ Addressing melatonin ingestions may help achieve the US’s Healthy People 2030 goal of reducing emergency department visits for unintentional medication overdoses in young children.⁴

Melatonin can serve as chronobiotic for circadian clock realigment⁵ but has a short half-life (approximately 40 minutes). Thus, immediate-release melatonin formulations may support falling asleep faster, while prolonged release supports longer sleep duration.⁶ After behavioral treatment and a thorough sleep assessment, melatonin is recommended for children (aged 2-6 years) with autism spectrum disorder (ASD) and related neurodevelopmental conditions. Specific instructions include taking small doses of melatonin (0.5-1.0 mg/d) within 30 to 60 minutes of bedtime, with gradual dose escalation if indicated.^7,8 Children with ASD are more likely to have a deficit in melatonin production and function, which is associated with delayed sleep onset.⁹ Sleep problems affect 50% to 80% of youths with ASD compared with 12.5% of children with typical development.¹⁰

A recent consensus statement has noted that behavioral intervention and thorough assessment of sleep conditions should be the first approach for treating insomnia in children with typical development.^11,12 Behavioral intervention includes promoting sleep hygiene (eg, bedtime routines, eliminating screen time before bed) prior to any medication use and has been shown to be effective in managing sleep disorders in children with typical development.^13,14,15,16 Screen time can delay natural melatonin production and subsequent sleep onset.¹⁷ Following these steps, with continued behavioral intervention and medical supervision, the consensus has endorsed brief melatonin use (up to 2 mg) in children with typical development.^11,12 However, experts have expressed concerns about long-term exogenous melatonin affecting hormonal development in these children.^18,19 Melatonin use may begin in toddlerhood and continue for long periods (approximately 12 months), suggesting use as a hypnotic rather than as a chronobiotic.^20,21

Most melatonin usage reviews focused on older children (aged 7-16 years)^22,23 or limited their findings to 1 diagnosis with only short-term follow-up.²⁴ Thus, our knowledge is limited about melatonin usage and subsequent long-term outcomes in young children.^22,25 Investigating melatonin use and its impact on child health over time may help inform clinical practice, research efforts, and guidance for parents. Our review examined the long-term outcomes (ie, safety, behavioral, and health factors) of melatonin usage in young children.

Methods

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guideline.²⁶ The review methodology was established prior to commencement of the review and was registered with PROSPERO (CRD42025645422). As data are deidentified and publicly available, ethics review and informed consent were not required in accordance with the Common Rule. There were no deviations from the protocol that impacted the review scope or conduct.

In collaboration with research librarians (P.J. and P.W.), the search strategy was organized using controlled vocabulary, such as Medical Subject Heading terms, keywords, and Boolean logic, around the following concepts: melatonin and children (aged 0-6 years). The search was developed in 1 large bibliographic database (Ovid MEDLINE), peer reviewed using the Peer Review of Electronic Search Strategies process,²⁷ and appropriately translated to the following databases: Embase, CINAHL Complete, APA PsycInfo, Web of Science Core Collection (Science Citation Index Expanded, Social Sciences Citation Index, Arts and Humanities Citation Index, and Emerging Sources Citation Index [Clarivate]), Alt HealthWatch, Consumer Health Complete, Global Health (including Archive), SPORTDiscus with full text, Cochrane CENTRAL, ClinicalTrials.gov, and the International Clinical Trials Registry Platform (eTable 1 in Supplement 1). All databases were searched from inception to February 26, 2025. No database filters or limits were applied to the searches except in the International Clinical Trials Registry Platform on recruitment status. Supplemental search strategies included forward and backward citation searching and review of references of relevant reviews. Article results were imported and deduplicated using Covidence, a web-based data synthesis software program.

Detailed inclusion and exclusion criteria are outlined in Table 1. Studies that prospectively assessed melatonin usage in young children aged 0 to 6 years and presented in English-language articles in peer-reviewed literature or gray literature (eg, conference abstracts) were included. Before study selection, a prepiloted protocol was created, and reviewers reached at least 80% agreement. Thirteen independent reviewers performed abstract screening, full-text screening, and data extraction in duplicate (C.L.K., G.B., R.B., M.C., D.N.C., R.C.G., G.G., J.L., L.M., O.O., N.S., A.S.T., A.W.). A third reviewer resolved conflicts. All review stages were conducted using Covidence. Data were extracted in duplicate using a prepiloted form to record article information, study population, intervention conduct, outcome assessment, and methodological quality. Data were cross-checked for accuracy before synthesis. Methodological quality was assessed using the 27-item Downs and Black Checklist for randomized and nonrandomized studies.²⁸ The power question was modified to whether a power analysis was reported, such as in other reviews.^29,30 There was a possible score of 28 and 25 for randomized studies and nonrandomized studies, respectively. Studies were categorized into the following quality levels: excellent (26-28), good (20-25), fair (15-19), and poor (≤14).³¹

Table 1. Study Inclusion and Exclusion Criteria.

Criterion	Include	Exclude
Population	Human children aged 0-6 y of all abilities Articles with age-specific estimates for age <5 y or mean age <5 y All regions	Includes only humans aged >7 y Age-specific estimates not provided (eg, aged 0-18 y estimates) Nonhuman populations
Intervention or exposure (cohort studies)	Prescription of melatonin Sleep-related use of melatonin	Endogenous melatonin production or melatonin as biological response Individual estimates for melatonin not available (eg, combined with other over-the-counter medications) Melatonin use for nonsleep aid purposes Other sleep aids (eg, magnesium) Single assessment of melatonin (eg, prior to surgery)
Comparator	Any comparator included	No exclusion criteria
Outcome	Primary: safety-related measures (eg, overdose, prescribing practices, emergency department visits) Secondary: physical health (eg, body mass index), mental health (eg, anxiety, depression), sleep-related behavior (sleep duration, etc)	Surgery-related outcome
Time	Intervention studies (before and after, single arm, randomized clinical trials) Longitudinal studies (case-control, cohort) Electronic health record data (multiple years)	Cross-sectional studies (eg, single-year medical record review, survey)
Article type	Original research (protocol articles, intervention studies, cohort studies) Published and gray literature (abstracts, clinical trial registries, etc) No publication year restrictions	Systematic reviews or consensus articles Commentaries or opinions Qualitative original research Case reports Non-English language

Analysis

A numerical analysis (central tendencies) of study characteristics and safety, behavioral, and health outcomes was conducted using Microsoft Excel, version 2023 (Microsoft Corp). A narrative synthesis of outcomes of included studies was conducted for each category of safety, behavioral, and health outcomes. Given the heterogeneity in outcomes (eg, total sleep time, sleep onset) and small sample size of trials, we did not conduct a meta-analysis.

Results

In total, 13 751 abstracts were retrieved and 8058 screened, of which 363 were considered for full-text review. Nineteen articles representing 6 completed experimental studies (hereafter, trials),^{32,33,34,35,36,37,38} 1 trial protocol,³⁹ and 12 observational studies^{40,41,42,43,44,45,46,47,48,49,50,51} were included in the analysis (eFigure in Supplement 1). The primary exclusion reason (k = 123 of 344 [35.8%]) was missing age-specific data, which encompasses not presenting young child age-specific results⁵² or being conducted outside the age range.⁵³ A list of studies excluded at full-text screening is provided in eTable 2 in Supplement 1. Table 2 and Table 3 summarize the included trial and observational study characteristics, respectively, as well as quality level. eTable 3 in Supplement 1 documents their funding and conflicts of interest.

Table 2. Characteristics of Included Trials (n = 7).

Source (country)	Study design	Intervention length	Diagnoses included	Analytic sample, No.	Age of participants, mean (SD) [range], y	Melatonin dosage	Melatonin intervention information	Concurrent medications	Study qualitya
Giannotti et al,34 2006 (Italy)	Quasi-experimental, no comparator	6 mo	ASD	29	4.6 (2.2) [2.6-9.6]	3 mg	Timing: taken 30-40 min before the desired bedtime, at the same time every evening; after 5 d, the time of melatonin administration was fixed at 8:00 pm Brand: not specified in the article Type: controlled-release melatonin	Not allowed	Fair
Gringras et al,36,37 2017 (US, UK, France, the Netherlands, Finland)b	RCT, placebo comparator	26 wk (13-wk Initial dosing, 13-wk final dose)	ASD; other neurogenetic condition	95 (28 Aged 2-7 y)	8.7 (4.15) [2-17]	2-5 mg	Timing: 30-60 min before bedtime Brand: PedPRM Type: prolonged release	Not allowed	Good
Jalilolghadr et al,38 2022 (Iran)	RCT, parent education comparator	4 wk	Chronic insomnia	59	1-3	0.1-1 mg/kg	Timing: 9:00 pm administration Brand: not described Type: not described	Not allowed	Good
Jan et al, 200032 (Saudi Arabia)	Quasi-experimental, no comparator	4-12 mo (Mean, 7.5 mo)	Multiple disorders	10	5.4 (NA) [1-11]	3 mg	Timing: 1-2 h before the target bedtime each night Brand: not described Type: not described	Not allowed	Poor
Malow et al,35 2012 (US)	Quasi-experimental, no comparator	14 wk	ASD	24	5.9 (NA) [3-10]	1-9 mg (3-wk Escalation of 3 mg if no satisfactory response)	Timing: 30 min before bedtime Brand: Natrol (Natrol) Type: not described	Psychotropics excluded, allergy and constipation medications allowed if stable	Fair
Rolling et al,39 2025 (France)	RCT, placebo comparator	13 wk	PTSD	120 Planned	NA [2-17.5]	2 minitablets of 1 mg	Timing: not described Brand: PedPRM Type: prolonged-release melatonin	Stable doses of nonexcluded medications for ≥3 mo	TBDc
Ross et al,33 2002 (UK)	Cross over	2 wk (1 wk Before treatment and 1 wk of treatment)	Multiple sleep, neurologic, or developmental disorders	46 (17 Aged 0-6 y)	NA [1-13]	2.5-5 mg, Increased to a maximum dose of 7.5-10 mgd	Timing: not reported Brand: Penn Pharmaceuticals Type: not reported	Not clear	Poor

Abbreviations: ASD, autism spectrum disorder; NA, not available; PedPRM, pediatric prolonged-release melatonin; PTSD, posttraumatic stress disorder; RCT, randomized clinical trial; TBD, to be determined.

^a Assessed by the Downs and Black Checklist.^28,29,30,31 Quality rating: excellent (26-28), good (20-25), fair (15-19), and poor (≤14).

^b Maras et al³⁷ provided additional data for the study by Gringras et al.³⁶

^c Protocol, no quality rating given. Authors were not contacted in case of missing data.

^d Parents could provide another 5 mg if their child initially settled to sleep but later awoke.

Table 3. Characteristics of Included Observational Studies (n = 12).

Source (country)	Years included	Total y	Diagnoses included	Analytic sample	Age of participants, young child subgroup, y	Age of participants, full sample, y	Data source	Qualitya
Araújo et al,51 2025 (Portugal)	2018-2022	4	NA	Not clear	1-4	0-18	Portuguese Poisons Information Center	Poor
Barinas et al,48 2023 (US)	2000-2021	21	NA	513 (16 650 Estimate)	0-4	0-4	National Electronic Injury Surveillance System	Fair
Bliddal et al,40 2023 (Denmark)	2012-2019	8	NA	43 652	0-4	0-24	Danish National Prescription Registry	Fair
Freeman et al,49 2024 (US)	2019-2022	4	NA	295 (10 930 Estimate)	0-5	0-5	National Electronic Injury Surveillance System	Fair
Hartz et al,44 2015 (Norway)	2004-2012	9	NA	Incident users, 2221; recurrent users, 589	4-8	4-17	Norwegian Prescription Database	Fair
Kildegaard et al,41 2025 (Denmark)	2005-2023	19	NA	Aged 0-2 y (range, 113 878-130 134); aged 2-5 y (range, 237 037-265 289)	0-2, 2-5	0-18	Danish National Prescription Registry	Fair
Kimland et al,46 2021 (Sweden)	2006-2017	12	NA	4 033 085	0-4	0-17	Swedish Prescribed Drug Register and the National Patient Register	Fair
Klau et al,50 2022 (Australia)	2011-2018	8	NA	537 371	0-4	0-18	MedicineInsight	Good
Nielsen et al,42 2017 (Denmark)	2006-2012	7	NA	29 851	0-5	0-17	Register of Medicinal Product Statistics	Fair
Rasmussen et al,43 2018 (Denmark)	2010-2015	6	ASD	14 210	3-5	3-18	Danish National Patient Registry, Danish Psychiatric Central Register, Danish National Prescription Registry, and Danish Civil Registration System	Good
Stangeland et al,45 2023 (Norway)	2004-2019	6	NA	Range, 1 727 471-1 938 350 per y	0-4	0-29	Norwegian Prescription Database	Good
Tedroff et al,47 2022 (Sweden)	2016-2019	4	NA	2 267 970	0-5	0-25	Stockholm Regional Health Care Data Warehouse	Good

Abbreviations: ASD, autism spectrum disorder; NA, not available.

^a Assessed by the Downs and Black Checklist.^28,29,30,31 Quality rating: excellent (26-28), good (20-25), fair (15-19), and poor (≤14).

The 6 concluded trials included a total of 245 participants, with trial results for 167 young children (Table 2).^{32,33,34,35,36,38} Trials included children with neurodevelopmental disorders, namely ASD.^34,35,36 No trials were conducted in children with typical development. Trials lasted a mean of 12.7 weeks (range, 2 weeks to 2 years), with 2 studies reporting 1-year³⁶ and 2-year³⁴ follow-up data. Study designs were primarily randomized clinical trials^36,38,39 or quasi-experimental studies,^32,34,35 with 1 crossover design.³³ Comparison groups included placebo^36,39 or parent education.³⁸ Two studies provided a nonmelatonin intervention in the form of parent education on sleep hygiene and promoting child sleep.^34,38 Melatonin formulations were reported in 3 trials and were prolonged release,^36,39 with 1 trial using controlled or immediate release.³⁴ Studies began with a mean dose of 2 to 3 mg (range, 1-5 mg), with some escalating to 9 to 10 mg.^33,35 Except for 1 trial,³⁸ all formulations were ingested 30 to 60 minutes before bedtime, and 1 study eventually transitioned participants to a fixed time of 8:00 pm.³⁴ Most trials did not allow concurrent medications^32,34,36,38; some allowed medications for allergies and constipation³⁵ or stable doses of nonexcluded medications.³⁹

Twelve observational studies used registry data, representing 8 unique registry sources from Europe (Denmark,^40,41,42,43 Norway,^44,45 and Sweden^46,47) and Australia⁵⁰ and poisoning data from the US,^48,49 and Portugal⁵¹ (Table 3). Studies captured a mean of 9 years (range, 3-21 years) of data and spanned the past 2 decades from 2000 to 2023. Young children were primarily evaluated as an age 0- to 4-year subgroup of a pediatric population, which totaled more than 100 000 individuals in most studies. Multiple studies did not provide the exact number of children in the 0- to 4-year age groups.^{40,42,43,44,45,46,47,50} Only 1 observational study was specific to children with ASD,⁴³ and all other studies included the general population.

Long-Term Safety Outcomes of Young Children Using Melatonin

Prescribing Practices

Eight observational studies reported outcomes associated with prescribing practices, including incidence, prevalence, and defined daily dose (Table 4).^{40,41,42,44,45,46,47,50} The incidence and prevalence of melatonin prescribing increased over time in 5 studies,^{41,42,46,47,50} with some reporting a 500% increase.^46,50 The defined daily dose of 2 mg per individual doubled in Sweden⁴⁷ but remained constant in Norway.⁴⁵

Table 4. Summary of Safety Outcomes Associated With Melatonin Treatment.

Source		Country	Years included or length of trial	Prescribing practices	Extended use, %			Adverse events	Notes
					1 y	2 y	3 y
Observational studies (n = 12)
Nielsen et al,42 2017	Denmark		2006-2012	Increase	NA	NA	NA	NA	Prescribing practices: top 3 most-used drugs off label. Melatonin became the predominant off-label prescription; in 2006, it was not in the top 3. In 2012, melatonin was number 1 for girls aged 0-5 y (n = 480 [98.9%]) and boys aged 0-5 y (n = 631 [97.8%]).
Rasmussen et al,43 2018	Denmark		2010-2015	NA	55.5	46.40	NA	NA	Extended use: among 210 individuals with ASD initiating treatment, 71.9% filled a second prescription for 180 d, as did 61.9% at 180 d, 55.5% at day 365, and 46.4% at day 730.
Bliddal et al,40 2023	Denmark		2012-2019	No change	2-mo: 85	40.0	NA	NA	Prescribing practices: incidence (approximately 0.5 per 1000 person-y) and prevalence of melatonin use (approximately 1.0 per 1000 person-y) remained stable across time. Extended use: based on 2019 data, approximately 85% of children aged 0-4 y were still taking melatonin after 2 mo, and approximately 40% of children were still using melatonin after 2 y.
Kildegaard et al,41 2025	Denmark		2005-2023	Increase	NA	NA	NA	NA	Prescribing practices: since its introduction in 2007, melatonin prevalence gradually rose in children age 0-2 y and 2-5 y up until 2023. Not clear whether statistically significant.
Hartz et al,44 2015	Norway		2004-2012	No changea	75	50b	46.5b	NA	Prescribing practices: amount (milligrams) and number of prescriptions dispensed were stable across time, with no differences in age groups. No exact data for age group provided. Extended use: based on 2009 data, approximately 75% of users were still refilling prescriptions at 365 d, approximately 50% between 366 and 700 d (approximately 2 y), and approximately 46.5% between 731 and 1095 d (3 y).
Stangeland et al,45 2023	Norway		2004-2019	No change	NA	NA	NA	NA	Prescribing practices: melatonin use began at 300 DDDs per individual in 2004, with fluctuations across time, and ending with approximately 250 DDDs per individual in 2019.
Kimland et al,46 2021	Sweden		2006-2017	Increase (×5)c	25b	10b	9b	NA	Prescribing practices: incidence rose from approximately 0.2 per 1000 in 2006 to 1.0 per 1000 (girls) and 1.7 per 1000 (boys) in 2017. Prevalence rose between 2006-2017. Extended use: approximately 25% of new users were still using after 1 y, as were approximately 15% for 2 y. Estimate of use from 3 y was approximately 9% to negligible after 8 y (approximately 2%).
Tedroff et al,47 2022	Sweden		2016-2019	Increase (×3)c	NA	NA	NA	NA	Prescribing practices: number of prescriptions per individual, absolute increase of 300%; median presented, 2 in 2016, 5 in 2017, 5 in 2018, and 6 in 2019. Number of DDDs (2 mg) per individual, absolute increase of 200%; median presented, 50 in 2016, 50 in 2017, 100 in 2018, and 100 in 2019.
Klau et al,50 2022	Australia		2011-2018	Increase (×5)c	NA	NA	NA	NA	Prescribing practices: absolute increase of 534.3% (233.0%-1108.1%) across time and annual change average (26.7% [20.2%-33.6%]) from 0.3 per 1000 persons in 2011 to 2.0 per 1000 persons in 2018 in adjusted models.
Barinas et al,48 2023	US		2000-2021	NA	NA	NA	NA	Increase (×10)c	Adverse events: estimated melatonin ingestions rose steadily across time from 0 in 2000-2001; 96 (0.6%) in 2002-2006, 746 (4.5%) in 2007-2011, 5307 (31.9%) in 2012-2016, and 10 501 (63.1%) in 2017-2021.
Freeman et al,49 2024	US		2019-2022	NA	NA	NA	NA	Increase (×2)c	Adverse events: emergency department visits due to unsupervised ingestion doubled across time from 42 (estimate, 2032 [18.6%; 95% CI, 10.7%-26.4%]) in 2019, 74 (estimate, 3294 [30.1%; 95% CI, 19.8%-40.5%]) in 2020, 78 (estimate, 2135 [19.5%; 95% CI, 11.7%-27.3%]) in 2021, and 101 (estimate, 3469 [31.7%; 95% CI, 18.3%-45.2%]) in 2022.
Araújo et al,51 2025	Portugal		2018-2022	NA	NA	NA	NA	Increase (×10)c	Adverse events: melatonin-related reported poisonings rose significantly across time. This trend began with 1 in 2018, 6 in 2019, 61 in 2020, 72 in 2021, and 60 in 2022.
Experimental studies (n = 6)
Ross et al,33 2002	UK		2 wk	NA	NA	NA	NA	None	NA
Jalilolghadr et al,38 2022	Iran		4 wk	NA	NA	NA	NA	Mild	Adverse events: 1 child experienced headaches.
Malow et al,35 2012	US		14 wk	NA	NA	NA	NA	Mild	Adverse events: 1 child experienced loose stools.
Giannotti et al,34 2006	Italy		6 mo	NA	NA	NA	NA	None	NA
Jan et al,32 2000	Saudi Arabia		4 12 mo (Mean, 7.5 mo)	NA	NA	NA	NA	None	NA
Gringras et al,36 2017	US, UK, France, the Netherlands, and Finland		1 y	NA	NA	NA	NA	Mild	Adverse events: treatment-emergent and other adverse events were reported in 76.9%-85.0% of participants per group, but did not provide young child–specific data for such events.

Abbreviations: ASD, autism spectrum disorder; DDD, defined daily dose.

^a No age-specific data presented, only described in text.

^b Approximate values.

^c The parenthetical indicates a multiplier of the outcome; for example, ×5 indicates a quintupling of the outcome during the period.

Detailed information on prescribers and diagnoses of young children was available in 2 studies. Using Danish data, the primary prescriber for new treatment was a hospital pediatrician (approximately 65%) for new prescriptions and (approximately 50%) for continued treatment.⁴⁰ General practitioners, private practice specialists, and psychiatrists also were prescribers (approximately 5%-15% each) for new and continued treatment. In that same dataset, 41.9% of young children in 2012 and 36.9% in 2017 who were prescribed melatonin had a psychiatric disorder or accompanying psychotropic drug.⁴⁰ Similarly, in Norway, most melatonin users had mental or behavioral disorders (boys, 86.0%; girls, 82.7%) or hyperkinetic disorders (boys, 69.4%; girls, 57.7%) during 2008 to 2012.⁴⁴

Extended Use

In 4 observational studies, approximately 40% to 50%, but as low as 10%, of young children were still refilling prescriptions for melatonin 2 and 3 years after initial prescription.^40,43,44,46 In Denmark-based studies, young children with ASD were reported to have slightly higher rates of refills (46.4%)⁴³ compared with the entire population (40.0%).⁴⁰

Adverse Events

Three observational studies reported doubling or decoupling of melatonin-related poisonings or emergency department visits due to unsupervised ingestions, especially between 2018 and 2022.^48,49,51 In most cases (89.9%-90.2%), no additional medications were implicated.^48,49 For trials, 3 that included young children with neurologic conditions reported no adverse events,^32,33,34 and 2 observed only mild adverse events in a few children, including headaches³⁸ and loose stools.³⁵ In another trial, treatment-emergent and other adverse events occurred in the majority of patients (placebo group, 76.9%; melatonin group, 85.0%), but the study did not report these events based on age group.³⁶

Long-Term Behavioral Outcomes

A summary of 6 trials documenting behavioral sleep outcomes at 8 time points is presented in Table 5.^{32,33,34,35,36,38} For sleep onset latency or time to fall asleep parameters, 3 trials in 52 children with neurodevelopmental disorders found significant improvements after 14 weeks,³⁵ approximately 7.5 months,³⁶ and 39 weeks³² of treatment. For total sleep time, 4 trials in 91 children with neurologic conditions demonstrated varying effects of melatonin. Total sleep time increased by approximately 1 hour in 2 trials lasting approximately 7.5 months³⁶ and 39 weeks.³² One trial found an increase at 6 months in 29 children (before: mean [SD], 410 [90] minutes per night; after: mean [SD], 570 [30] minutes per night),³⁴ but this was reduced to a mean (SD) 460 (100) minutes per night after a 1-month discontinuation of melatonin. Another trial reported a nonsignificant increase at 14 weeks.³⁵

Table 5. Summary of Experimental Study Evidence of Melatonin Treatment on Behavioral Sleep Outcomes in Young Children.

Study	Time point	Analyic sample	Sleep measure	Sleep outcome				Notes
				Sleep onset latency	Total sleep time	WASO	Other sleep metric
Ross et al,33 2002	2 wk	17	Parent report	NA	NA	NA	POS	Other sleep metric: 14 of 17 participants aged 0-6 y experienced benefit (not defined).
Jalilolghadr et al,38 2022	4 wk	59	Sleep diary		POS	POS		Total sleep time: average nighttime sleep was significantly lower in control (mean [SD], 10.76 [1.79]) compared with treatment (mean [SD], 10.25 [0.67]) when comparing 4-wk measures. WASO: times of waking up during the night were significantly higher in the control group (mean [SD], 1.96 [1.7]) compared with treatment group (mean [SD], 0.90 [1.09]) when comparing 4-wk measures.
Malow et al,35 2012	14 wk	24	Actigraphy	POS	NS	NS	NS or POS	Sleep onset latency: significant reduction from acclimation (42.9 min) to end of study (22.5 min; difference, 20.4 min). Total sleep time: nonsignificant increase from acclimation (450.1 min/night) to end-of-study (457.3 min/night; difference, 7.3 min/night). WASO: no significant change (acclimation, 64.9 min; end of study, 68.6 min). Other sleep metric: no change in sleep efficiency (acclimation, 75.0%; end of study, 79.5%). Other sleep metric: improvements in CSHQ sleep duration, anxiety, and sleep total scores, but not other scores.
Giannotti et al,34 2006	6 mo	29	Sleep diary	NA	POS	POS	NA	Total sleep time: significant increase at 6 mo (before: mean [SD] 410 [90] min/night; after: mean [SD], 570 [30] min/night), but this amount was attenuated to a mean (SD) of 460 (100) min/night after 1 mo. WASO: reduced from a mean (SD) of 70 (54) min/night to 10 (15) min/night at 6 mo and rebounded to 40 (45) min/night 1 mo after discontinuing melatonin.
Jan et al,32 2000	4-12 mo (Mean, 7.5 mo)	10	Sleep diary	POS	POS	POS	NA	Sleep onset latency: delayed sleep onset reduced from 6.8 nights/wk to 0.4 nights/wk after treatment. Total sleep time: average sleep duration increased from 6.8 h/night to 7.9 h/night. WASO: reduction of night awakenings from 3 per night to 0.5 per night.
Gringras et al,36 2017	1 y	28	Sleep diary	POS	POS	NA	NA	Sleep onset: sleep onset latency significantly decreased by a mean (SD) of 48.66 (8.87) min in children aged 2-7 y. Total sleep time: significantly increased by a mean (SD) of 52.36 (26.52) min in children aged 2-7 y.
Giannotti et al,34 2006	1 y	16	Questionnaire	NA	NA	NA	POS	Other sleep metric: significant reduction in total CSHQ score (baseline, 63.0; 1 y, 44.0).
Giannotti et al,34 2006	2 y	16	Questionnaire	NA	NA	NA	POS	Other sleep metric: significant reduction in total CSHQ score (baseline, 63.0; 2-y, 44.7).

Abbreviations: CSHQ, Children’s Sleep Habits Questionnaire; NA, not available; NS, nonsignificant effect; POS, positive and significant response; WASO, wake after sleep onset.

For night awakening measures, 3 trials in 63 children with neurologic conditions found mixed evidence. One trial of 29 children demonstrated a reduction in waking after sleep onset from a mean (SD) of 70 (54) minutes per night to 10 (15) minutes per night at 6 months and rebounded to 40 (45) minutes per night 1 month after discontinuing melatonin.³⁴ The other 2 trials reported both positive³² and nonsignificant results.³⁵

Other sleep parameters assessed were sleep patterns via the Children’s Sleep Habit Questionnaire, sleep efficiency, and sleep benefit. One study of the 16 participants who continued melatonin after the end of the trial rather than discontinuing treatment found significant reductions in their total Children’s Sleep Habit Questionnaire scores (discontinuation [6 months, end of trial], 63.0; 1-year [using melatonin], 44.0; 2 years [using melatonin], 44.7); these results are from a subsample who returned to baseline sleep levels after discontinuing melatonin.³⁴ Ross et al³³ demonstrated that 14 of 17 participants in this age range (0-6 years) experienced a benefit (not defined) from melatonin after 2 weeks. One study reported no change in sleep efficiency after 14 weeks.³⁵ One 14-week trial reported improvements in child behavior, child repetitive behavior, and parenting stress, with approximately 50% of subscales of the respective questionnaires showing significantly higher scores (ie, improved) after the study.³⁵

Long-Term Health Outcomes

Two trials provided qualitative reports of long-term growth and development.^36,38 One, a 4-week study that assessed precocious puberty after treatment, found no association with development.³⁸ The other, a 1-year study, reported small increases in body mass index and body mass index z scores, with no differences between treatment groups (ie, melatonin vs control) but did not provide young child–specific results.³⁶

Methodological Quality

Methodological quality of completed studies are shown in Table 2, Table 3, and eTable 4 in Supplement 1. Two trials each had quality ratings of good,^36,38 fair,^34,35 and poor.^32,33 Ratings were limited due to low external validity, internal validity, confounding, and power calculations. Methodological quality of observational studies was good in 4 studies^43,45,47,50 and fair in 7 studies,^{40,41,42,44,46,48,49} and 1 study was rated as poor quality.⁵¹ Observational studies reported high external validity and general reporting but scored lower on internal validity and power calculations.

Discussion

This systematic review found that melatonin use in young children has increased steadily over the past decade. Observational studies documented rises in prescribing, extended use, and overdoses. Trials presented evidence for improved sleep onset with few adverse events. Data regarding long-term outcomes for other behaviors and health outcomes were generally absent. Among good- and fair-quality studies, melatonin was found to improve sleep onset in children with neurologic conditions but not in children with typical development. Our review contributes to the evidence that a clear public health problem may exist with regard to the increase of unsafe melatonin practices in young children and continued use beyond clinical recommendations.

Long-term safety outcomes revealed 2 imperfect systems for melatonin use. In countries where melatonin supplements are classified as dietary supplements (eg, US),⁵⁴ increases in overdoses occurred most notably between 2018 and 2022. Three main culprits were speculated in these studies. First, the rise may be attributable to behavioral treatment at home during the COVID-19 pandemic⁵¹ potentially due to high screen time and disrupted sleep.⁵⁵ Another consideration is the formulation, including child-friendly gummy formulations that comprised 47.3% of ingestions in 1 US-based study.⁴⁹ Further US evidence noted that 90.2% of pediatric melatonin ingestions were unintentional, suggesting that young children may be consuming parents’ melatonin¹; adult melatonin use has also increased in the past decade.⁵⁶ Finally, these overdoses may be attributed to commercial brands greatly exceeding labeled amounts,⁵⁷ which may have devastating effects for younger children.

In countries where melatonin supplementation is mainly prescription-only and with limited over-the-counter sales (eg, the Nordic countries),⁵⁴ medications were generally prescribed to the target population, but continued use (3 years) exceeded maximum trial durations (2 years). Prescribing practices do not necessarily reflect actual consumption, as some young children may consume less than the prescribed amount.⁵⁸ However, these large increases still indicate a widespread use of melatonin. This increased incidence and use may be associated with a lack of meeting practice standards. A retrospective audit of melatonin prescribing practices in the UK found that among 4151 children, 32% did not receive a nonpharmacologic intervention prior to melatonin initiation, and only 36% of practitioners met practice standards of quantifying melatonin effectiveness in the first 3 months of treatment.⁵⁹ Other clinical medical record reviews reported even lower adherence (9%) to prescribing a nonpharmacologic intervention prior to melatonin initiation in children.⁵⁹ When aligned with recommendations, trial studies revealed that melatonin was generally safe in specific populations for up to 1 to 2 years. These results align with other systematic reviews of older children, which indicated that minor adverse events occur.²² These registry studies did not report adverse events, so it is unclear whether the trial results translated to actual clinical practice.

All trials reported improvement in sleep onset latency and parent-reported questionnaire metrics, with neutral evidence for total sleep time and night awakenings. These results align with other reviews and trials of melatonin in children with neurologic conditions.^24,60 Melatonin use has clinically meaningful impacts, reducing sleep onset by at least 30 minutes, but no clinically or statistically significant increases in total sleep time have been observed.⁵² The nonsignificant sleep results may be attributed to rigorous assessment of sleep (actigraphy) compared with parent-reported sleep diaries. One longer-term study (approximately 1 year in duration) attempted actigraphy, but it was not tolerated well by the sample population.³⁶ Other longer-term results are nuanced, as they only included a clinical population of children who returned to baseline sleep levels after discontinuing melatonin.³⁴ Other behaviors, such as physical activity and diet, were not reported, but preliminary evidence showed their associations with improvement in child behavioral problems.³⁵

These findings support clinical practice of recommending melatonin for young children with ASD after behavioral intervention and medical supervision, but there was no evidence to support this practice in children with typical development. The immediate next steps are threefold. First, we must improve pediatrician and parent support of behavioral sleep practices, such as reducing nighttime screen time.¹⁶ Parent education is critically needed as parents may interpret over-the-counter availability as an indication of safety⁶¹ and use melatonin in substitution of behavioral sleep practices (eg, complete or gradual extinction practices, structured bedtime routines, reduction of screen time before bed).¹⁶ Second, pediatricians should assess child and parent supplement usage, as convenience samples of US parents have noted that not all parents may discuss their melatonin usage with their health care professionals⁶¹ but still use melatonin for extended periods.²¹ Third, even in children with ASD, long-term medical follow-up is required given the sustained use beyond 3 years. Long-term solutions are twofold. First, based on overdose increases in nonregulated countries, allocating melatonin to a prescription medication may improve use estimates, medical supervision, and accurate formulation. Second, development of discontinuation resources to safely reduce melatonin use, while promoting healthy sleep practices, is needed to meet recommendations for children with typical development.⁶²

Strengths and Limitations

The strengths of this review include a thorough assessment of various outcomes in a targeted population and timely exposure, leveraging rich pharmacologic databases⁶³ and experimental evidence, as well as an exhaustive search strategy and rigorous systematic review practices. Limitations of the current review center on generalizability and outcomes obtained. Observational study results may not generalize across borders due to varying country-level systems, and trial results may not translate to children with typical development. Trial lengths (2 weeks to 2 years) generally align with other long-term outcome reviews,²² but limited health outcome assessment (eg, growth) across extended time frames hindered our ability to make recommendations. Some included studies noted that they were unable to assess young children specifically due to small sample sizes in the group.^45,46 Furthermore, many studies lacked power calculations and accounting for confounding, which may influence results. These results encompassed exogenous melatonin and may not translate to clinical use in inpatient settings.⁶⁴

Conclusions

In this systematic review of melatonin use in young children, observational studies found a global increase in prescriptions, extended use, and adverse events. In trials, melatonin use was generally effective for sleep outcomes for specific populations in the short term; however, longer-term data in children with typical development were lacking. These findings may help identify strategies to prevent and reduce melatonin use in young children, as well as to improve adherence to practice standards.

Supplement 1.

eTable 1. Search Strategy

eFigure. PRISMA Flow Diagram

eTable 2. List of Articles Excluded at Full-Text Stage With Exclusion Reasons, Including Those Unable to Be Retrieved (n = 344)

eTable 3. Funding and Conflicts of Interest of Included Studies

eTable 4. Quality of Included Articles (n = 19)

Supplement 2.

Data Sharing Statement