Melatonin as a Chronobiotic Intervention for Advancing Sleep Timing and Enhancing Sleep Quality in Delayed Sleep-Wake Phase Disorder: A Systematic Review and Meta-Analysis on Randomized Controlled Trials Studies
Article information
Abstract
Background and Objective
Delayed Sleep-Wake Phase Disorder (DSWPD) is a sleep disorder that causes difficulty falling asleep, leading to functional impairment in daily activities. This review study analyzes the effectiveness of exogenous melatonin through the optimal timing of melatonin administration and standard dosage on improving sleep quality and shifting the sleep chronobiology of DSWPD patients.
Methods
Based on the guidelines from Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020, literature review searches were conducted using databases such as PubMed, Scopus, Web of Science, and Proquest. Methodological quality assessment was performed using Cochrane Risk of Bias 2.0 and the overall analysis was conducted using R 5.4.2 to process the data.
Results
Overall analysis of actigraphy shows that exogenous melatonin improves sleep quality, such as wakefulness after sleep onset and sleep onset latency (SOL) (z=-4.94, p<0.0001), total sleep time (TST) and time in bed (TIB) (z=3.72, p=0.0002), and shifts sleep timing (z=-8.80, p<0.0001) such as dim light melatonin onset (DLMO), sleep onset time, sleep offset time, and wake-up time. Subgroup analysis indicated that high-dose exogenous melatonin (≥3 mg) reduced SOL parameters (z=-4.11, p<0.0001), while low-dose exogenous melatonin (<3 mg) significantly increased TST parameters (z=6.52, p<0.0001). Meta-regression on melatonin administration 3–5 hours before DLMO was found to be effective (SOL parameter result β=74.6, p=0.005, and TST parameter result β=-49.60, p=0.042).
Conclusions
As a chronobiotic substance, exogenous melatonin can improve sleep quality and shift the sleep time of DSWPD patients by considering the dosage and timing of administration.
INTRODUCTION
Nowadays, children and young adults often stay up late to continue their activities at night [1,2]. This persistent habit can trigger a delay in their “habitual sleep-wake” cycle, thereby increasing the risk of developing delayed sleep disorder, now known as Delayed Sleep-Wake Phase Disorder (DSWPD) [3]. These sleep disorders cause serious health problems such as frequent fatigue, loss of concentration, decreased physical performance, and some people may even develop depression [2,4]. The delay in the “habitual sleep-wake” cycle causes DSWPD sufferers to have difficulty falling asleep and may wake up at unwanted times. Some DSWPD patients often fall asleep at 2 am and wake up after 10 am with various sleep disturbance complaints [5,6]. These characteristics align with the International Classification of Sleep Disorders, 3rd edition in 2014, approved by the American Academy of Sleep Medicine (AASM), which defines DSWPD as a circadian rhythm disorder characterized by a delay in the usual “sleep-wake” cycle of 2 hours or more [7]. Based on clinical examination results, individuals with DSWPD tend to have a delayed endogenous melatonin release onset, resulting in an unstable dim light melatonin onset (DLMO). The DLMO itself is a biomarker indicating the body’s release of endogenous melatonin [8]. The instability of the DLMO causes irregular sleep-wake patterns, leading to daytime fatigue and sleepiness in affected individuals [2].
DSWPD sleep disorders have become a serious health issue among children and young adults in this era of rapid technological development. Based on the latest global survey in 2023, approximately 82% (391 respondents) of the 479 total respondents experienced DSWPD sleep disorders, clinically complaining of difficulty sleeping and frequent late waking [9]. The younger generation in Norway shows a fairly high incidence of DSWPD, at around 3.3% (337 adolescents) of a survey of 10,220 adolescents aged 16 to 18 years [10]. Even according to an epidemiological study conducted by Sivertsen et al. [11], there are variations in the incidence of DSWPD based on age. The study involved 50,540 students aged 18 to 35 years. For students aged 18–20, the incidence of DSWPD was 3.6%, and for those aged 29–35, it was 2% of the total number of students. In addition, a report from another European country, Cyprus, stated in a national epidemiological survey that there are two sleep disorders commonly experienced by young adults, namely DSWPD and shift work sleep disorder, with incidence rates of approximately 5.1% and 6.7%, respectively, out of a total of 250 individuals who met the criteria [12]. In Japan, the prevalence of DSWPD is quite high, at around 4.3% of the total 7,810 participants aged 15 to 30 years old [13]. In fact, the latest review in 2023 shows an increase in the number of DSWPD cases for the adolescent category from 3.3% to 63.3% after the COVID-19 infection era, accompanied by several other health disorders such as heart disease, diabetes mellitus, and declining academic performance [5].
Based on the issue of circadian rhythm disorders spreading throughout the world, in 2007, the AASM developed treatment guidelines establishing light therapy as the primary treatment for improving sleep quality in patients with circadian rhythm disorders. However, this treatment has been deemed ineffective [2,14,15]. In 2015, the AASM revised the guidelines to recommend the combined use of melatonin and light therapy to treat delayed sleep disorders in children and adults, and to not recommend melatonin and sleeping pills for elderly people with dementia [16]. In addition, several studies have examined the effectiveness of melatonin in DSWPD with varying results. A previous study conducted by Nagtegaal et al. [17] encouraged subsequent researchers to conduct studies on the effectiveness of exogenous melatonin in patients with DSWPD. This study mentioned that the burden of life for patients with DSWPD would be reduced after the use of exogenous melatonin. Even long before the study conducted by Nagtegaal et al. [17], the effectiveness of exogenous melatonin on DSWPD patients had been studied by Dahlitz et al. [18] in 1991 on 8 DSWPD patients, concluding that melatonin affects the wake-sleep cycle. The same study was conducted by Sletten et al. [19] in 2018, administering 0.5 mg of fast-release (FR) exogenous melatonin 1 hour before bedtime, which had a positive effect on improving sleep quality.
Based on all studies that have been conducted, the appropriate time for clinical administration of melatonin to provide a significant effect on the sleep quality of patients with DSWPD has not yet been determined [2]. Even the 2015 AASM guidelines do not recommend the appropriate dose of exogenous melatonin administration for patients with DSPWD clinically [16]. Previous meta-analysis studies have also revealed the need for further research to evaluate the effectiveness of melatonin, including determining the optimal dosage and timing for administering therapy to patients with delayed sleep disorder [20]. Thus, practical and clinical limitations have created a scientific space to research and examine the effectiveness of exogenous melatonin through the optimal timing of melatonin administration and standard doses on improving sleep quality and sleep time shifts in DSWPD patients with disrupted sleep patterns. Therefore, this systematic review and meta-analysis study was designed to examine and analyze melatonin in DSWPD patients in improving the sleep quality of DSWPD patients. This study will be conducted by systematically reviewing several previous studies, followed by statistical analysis to provide more clinically valid and accurate results.
METHODS
This study was prepared in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [21]. The research protocol has been registered with The International Prospective Register for Systematic Reviews and has been assigned registration number CRD420251126443. The results of this research protocol registration can be accessed via https://www.crd.york.ac.uk/PROSPERO/view/CRD420251126443.
Eligibility Criteria
The main author designed the Patient or Population, Intervention, Control, and Outcome (PICO) approach to identify eligible studies in the form of sleep quality analysis in DSWPD patients with melatonin intervention, making the study search and selection processes more accurate and effective [22,23]. Patients (P) involved in this study were clinically identified DSWPD sufferers. The main intervention (I) in this study was exogenous melatonin, and the control (C) used was a placebo, other compounds, or even compounds similar to melatonin. For the Outcome (O), the authors will include studies that evaluate the sleep quality of DSWPD patients after exogenous melatonin intervention. In addition, the main authors will exclude studies in the form of reviews, editorials, case reports, observational studies, research protocols, articles not available in English, and studies on experimental animals.
Search Strategy
The search for study articles was conducted by several authors (MIH and ZAL) in several databases from PubMed, Scopus, Web of Science, and Proquest. To obtain good randomized controlled trials (RCT) studies and evaluate the sleep quality of DSWPD patients, the main author did not impose any restrictions on the year of publication, given the limited availability of RCT studies. The study search strategy began by entering “search terms” such as “Melatonin,” “Delayed Sleep-Wake Phase Disorder,” and “Randomized” into several databases. To standardize medical terms and make the study search more accurate and consistent, Medical Subject Headings (MeSH) operators and the “tiab” operator were used so that the study search could cover various relevant term variations. To refine the study articles’ search and make it more sensitive and specific, the main author used the Boolean operator “OR” for terms with the same meaning and “AND” to combine search terms with different meanings. In addition, to anticipate less sensitive and specific search terms that could result in some study articles not being indexed in the search strategy, a manual search was conducted using Google to find study articles relevant to the eligibility criteria.
Study Selection
The entire study selection process involved three authors (MIH, ZAL, CNIA) to identify the appropriate studies. The study selection process began by entering all articles obtained from the search strategy into Rayyan software and counting the number of articles that had been input. Duplicates were then filtered out to remove articles with similarities. Next, studies were selected based on titles and abstracts that met the eligibility criteria. After that, the articles that will be read in full text will be left and screened based on eligibility criteria to identify data suitability. If there are differences of opinion among the authors during the study selection process, the three authors will invite the fourth author (AA) to exchange opinions and make decisions so that there are no subjective assessments.
Data Extraction and Quality Assessment
The data extraction process began with MIH entering the necessary data from the included articles into the extraction table. The data extraction fields for this study include the name of the lead author, year of publication, study design, country, characteristics of the study population (sample size of the experimental and control groups and age), intervention dose, duration of the study, time of intervention, type of compound in the control group, mean DLMO time, and mean±standard deviation (SD) data for sleep quality components and sleep time or chronobiotic time. The mean and SD data obtained from each included study article will be presented in a table to facilitate data processing in the synthesis of results. The results of reporting side effects of melatonin and the overall results of data extraction will be presented in qualitative and quantitative extraction tables.
In addition, data from several sleep parameters such as sleep efficiency (SE), total sleep time (TST), and total time in bed (TIB) have a systematic relationship as stated by Reed and Sacco in 2016 [24]. Mathematically, SE can be calculated as SE (%)=(TST/TIB)×100. Based on this formula, if only two of the three components are provided in the inclusion study, the unknown component can be calculated using this systematic relationship. This approach was used to find TIB parameter data, which is often unavailable in some studies, in order to strengthen quantitative analysis and reduce potential data exclusion in this meta-analysis. However, this approach may cause potential estimation bias and propagation errors in measurement method variations and differences in the definitions of TST and SE in each study. MIH, assisted by FF and WRI, has been careful in involving the parameters to be calculated and interpreted, so that estimation bias and consistency with the operational definitions of these sleep parameters can be accommodated properly.
The research methodology was assessed by MIH and ZAL for the included articles using the Cochrane guidelines in the form of Cochrane risk-of-bias for version 2 randomized trials to evaluate bias in the randomization process, bias in intervention deviation, bias due to data loss, bias in measuring study results, and assessing bias in selecting the results to be reported in the article [25]. The results of this assessment will classify each study into the categories “low,” “high,” and “some concern.” MIH and ZAL conducted independent methodological assessments, and if there were any differences in their decisions, they would make a decision by exchanging opinions and inviting several authors (CNIA and AA) to obtain a more objective assessment of the study bias. The results of the methodological quality assessment will be visualized using Risk of Bias Visualization and presented in a traffic plot and weighted bar plot to clarify the assessment results [26].
Data Synthesis and Statistical Analysis
All quantitative analyses in this study will be processed using the meta package feature of R 5.4.2 software (R Foundation for Statistical Computing) by entering the mean, SD, and sample size data from each study by the main author (MIH) [27]. All data will be combined to obtain a valid effect size using a fixed effect model analysis for heterogeneity less than 50% and a random effect model for heterogeneity greater than 50% [28]. The main methodology used to calculate the effect size between the experimental and control groups is mean difference (MD) and is expressed in a 95% confidence interval (CI) range. In the forest plot, the effect size calculation for the reported results of exogenous melatonin side effects will use the odds ratio (OR) methodology. Statistically, the p-value analysis result is considered significant if it is less than 5% and the critical z-value is greater than +1.96 or less than -1.96 [29].
Heterogeneity analysis will be expressed in I-Higgins (I2) to evaluate the distribution, and the results of the heterogeneity test will be classified as negligible (<25%), low (25%–49%), moderate (50%–74%), or high (≥75%) [30]. Assessment of bias in research results will be evaluated quantitatively for each component using the Egger test to describe the potential bias of statistically synthesized data [31].
In this study, the main author (MIH) used meta-regression to determine several factors, such as age, dosage, treatment duration, melatonin formulation, and time of administration, on the effectiveness of exogenous melatonin in improving sleep quality and synchronizing chronobiotic time in patients with DSWPD. Separate analyses will be used to estimate these variations in detail and to refine the conclusions [32]. The analysis model will be used on sleep quality components that experienced a decrease in score and an increase in score in the actigraphy tool.
RESULT
Search Result and Quality Assessment Result
This review identified 1,197 articles through the PubMed, Scopus, Web of Science, and Proquest databases, and the remaining 20 articles were identified through Google, which had been published until December 2025. The results of the search strategy are presented in tables that can be accessed in Supplementary Tables 1 and 2 (in the online-only Data Supplement). All articles obtained were deduplicated, leaving 775 articles to be selected based on their titles and abstracts. A total of 43 study articles were reviewed in their entirety by reading the full text. The final results of the study selection are presented in the PRISMA diagram (Fig. 1), with 18 study articles (Dahlitz et al. [18] in 1991, Eckerberg et al. [33] in 2012, Kayumov et al. [34] in 2001, Magee et al. [35] in 2020, Mundey et al. [36] in 2005, Nagtegaal et al. [37] in 1998, Rahman et al. [38] in 2010, Saxvig et al. [39] in 2014, Sletten et al. [19] in 2018, Smits et al. [40] in 2001, Smits et al. [41] in 2003, Swanson et al. [42] in 2024, van Andel et al. [43] in 2022, Van der Heijden et al. [44] in 2007, van Maanen et al. [45] in 2017, van Wieringen et al. [46] in 2001, Wasdell et al. [47] in 2008, Wilhelmsen-Langeland A et al. [48] in 2013) meeting the eligibility criteria, consisting of 12 study articles [18,19,33-38,42,43,47,48] from the search strategy and 6 study articles [39-41,44-46] from grey literature.
Preferred reporting items for systematic reviews and meta-analyses flowchart.
The results of the methodological quality assessment for 18 RCT studies using the Cochrane risk-of-bias for version 2 randomized trials were visualized in traffic light plots and weighted bar plots (Figs. 2 and 3). RCT studies are often referred to as the best research design, but every study has limitations that cause bias in study quality [49]. In assessing the methodological quality of this study, 17 articles were categorized as “low” and 1 study was categorized as “some concern,” namely the study by Nagtegaal et al. [37] in 1998. RCT studies with an open-label design often cause methodological bias because the patients and caregivers involved in the study know the type of compound used and will report more subjective results.
Traffic light plot in risk of bias assessment.
Weighted bar plot in risk of bias assessment.
General Characteristics of Included Studies
The general characteristics of the included studies are presented in Table 1. Based on the study selection results, 18 RCT studies were found, the majority of them (11 studies) originating from Europe [18,33,37,39-41,43-46,48], followed by 5 studies from America [34,36,38,42,47], and 2 studies from Australia [19,35]. The results of data extraction are presented in Table 2. A total of 18 articles found 784 patients with clinically diagnosed delayed sleep. The average age of patients involved in this study varied greatly, ranging from children to adults. Exogenous melatonin in this study was taken orally in capsule form, standardized by each researcher with varying doses ranging from 0.3 mg to 5 mg. Interestingly, the 18 studies showed that FR melatonin was more commonly used for patients with delayed sleep than prolonged-release (PR) melatonin. In addition, the duration of treatment in this study was mostly carried out within a period of 1 month, but there were several studies that carried out treatment for less than 1 month. There were even 4 study articles that carried out treatment for more than 1 month, with the longest duration carried out by Wasdell et al. [47] in 2008, which was for 12 weeks. In this study, some researchers instructed patients with DSWPD to take exogenous melatonin in the evening, several hours before each patient’s DLMO.
General characteristics of included studies
Data extraction of baseline study characteristics
Analysis Result by Actigraphy
In this study, almost all studies used actigraphy to evaluate sleep quality and sleep duration in patients with DSWPD. Raw data for statistical processing with actigraphy are presented in Supplementary Tables 3 and 4 (in the online-only Data Supplement), and the results of statistical analysis of sleep quality and sleep duration are presented in forest plots (Figs. 4 and 5, respectively).
Forest plot analysis of actigraphy in sleep quality. A: Actigraphy with descending outcomes. B: Actigraphy with ascending outcomes. C: Actigraphy with sleep efficiency outcome. WASO, wakefulness after sleep onset; SOL, sleep onset latency; SD, standard deviation; MD, mean difference; CI, confidence interval; TST, total sleep time; TIB, time in bed.
Forest plot analysis of actigraphy in sleep timing. DLMO, dim light melatonin onset; SOT, sleep onset time; SOffT, sleep offset time; WUT, wake-up time; SD, standard deviation; MD, mean difference; CI, confidence interval.
In addition, in the actigraphy scores for sleep quality, where the parameters improve as the scores decrease (Fig. 4A), wakefulness after sleep onset (WASO) and sleep onset latency (SOL), a statistically significant effect of melatonin was found (pooled MD=-13.52, 95% CI [-18.88; -8.16], z=-4.94, p<0.0001) with a high-risk heterogeneity test result (I2=84.6%, p<0.0001) and the Egger test showing no statistical bias in the findings (p=0.9948). In addition, subgroup analysis was also performed in these parameters. For the WASO component, a reduction in wake time from actual sleep was found, as evidenced by a CI that did not cross the zero line (MD=-6.30, 95% CI [-10.41; -2.20], z=-3.01, p=0.0026). While the SOL component, which was part of the subgroup analysis of this parameter, showed a significant effect of exogenous melatonin (MD=-17.33, 95% CI [-24.06; -10.61], z=-5.05, p<0.0001) in reducing the time to begin sleeping in bed.
In the actigraphy results for sleep quality, where the parameters improve as the score increases (Fig. 4B), TST and TIB, the CI range did not cross the null effect line (zero value) and it was confirmed by a significant pooled MD (pooled MD=8.71, 95% CI [4.12; 13.31], z=3.72, p=0.0002) with a high level of heterogeneity (I2=74%, p<0.0001). Even the Egger test showed no significant bias (p=0.8959), indicating no bias in the findings. Subgroup analysis was also performed on these parameters for the TST and TIB components to estimate the effect of exogenous melatonin, and the results were statistically significant (MD=9.52, 95% CI [2.73; 16.31], z=2.75, p=0.0060; MD=6.95, 95% CI [1.30; 12.60], z=2.41, p=0.0160, respectively).
As an essential measure of sleep quality, the SE component was analyzed using single-group statistics to provide an explanation of proper sleep consolidation. The results of the analysis for the SE component (Fig. 4C) showed that exogenous melatonin had an effect on preventing sleep fragmentation and was statistically significant (MD=4.06, 95% CI [0.16; 7.97], z=2.04, p=0.0414) with a high heterogeneity test (I2=94%, p<0.0001). In addition, the Egger test was used to confirm the bias effect of these findings and obtained insignificant results (p=0.2427).
As well as evaluating sleep quality, actigraphy results were also used to analyze changes and shifts in sleep time in patients with DSWPD. The results of the meta-analysis evaluating sleep time are presented in a forest plot (Fig. 5). Based on the overall analysis results, the use of melatonin has a statistically significant effect on changes in the sleep time of DSWPD patients (pooled MD=-26.02, 95% CI [-31.81; -20.23], z=-8.80, p<0.0001) with a heterogeneity level of I2=86.5% (p<0.0001) and the Egger test results showed no statistical bias in the findings (p=0.5379). Furthermore, based on the results of all subgroup analyses for sleep time parameters, exogenous melatonin had a clinically significant effect with a CI that did not cross the no-effect line on sleep components such as DLMO, which showed an earlier shift in melatonin release (MD=-26.57, 95% CI [-39.99; -13.15], z=-3.88, p=0.0001), and sleep onset time (SOT), which showed a change in SOT to 25 minutes earlier (MD=-25.19, 95% CI [-33.58; -16.80], z=-5.88, p<0.0001), sleep offset time (SOffT) showing a shift in wake-up time (WUT) to earlier (MD=-12.36, 95% CI [-19.74; -4.98], z=-3.28, p=0.0010), and WUT, which indicates a shift in the time of waking up to start activities earlier (MD=-31.80, 95% CI [-44.28; -19.32], z=-4.99, p<0.0001).
Analysis Result with Sleep Questionnaire Score and Polysomnography
All raw data to be processed for meta-analysis on sleep questionnaire scores and polysomnography (PSG) are presented in Supplementary Table 5 (in the online-only Data Supplement). The results of the sleep quality analysis on the sleep questionnaire, including the Pittsburgh Sleep Quality Index (PSQI), Epworth Sleepiness Scale (ESS), and Insomnia Severity Index (ISI), are presented in a forest plot (Fig. 6). The overall questionnaire score analysis showed a significant reduction in sleep disorder symptoms with exogenous melatonin administration, with a pooled MD of 91.97 and a CI that did not cross the zero line (pooled MD=-1.97, 95% CI [-2.55; -1.40], z=-6.70, p<0.0001) and heterogeneity results classified as low (I2=39.5%, p=0.1046) and the Egger test showed insignificant results (p=0.3642), indicating no bias in the findings. Subgroup analysis of the PSQI questionnaire (MD=-2.15, 95% CI [-2.95; -1.36], z=-5.30, p<0.0001) and ISI (MD=-3.37, 95% CI [-5.30; -2.23], z=-4.80, p<0.0001) questionnaires showed a clinically significant effect of melatonin. However, based on the ESS questionnaire subgroup analysis, a decrease in score of -0.93 was found, and this finding was not statistically significant (MD=-0.93, 95% CI [-1.93; 0.06], z=-1.83, p=0.0667).
Forest plot analysis of sleep questionnaire. PSQI, Pittsburgh Sleep Quality Index; ESS, Epworth Sleepiness Scale; ISI, Insomnia Severity Index; SD, standard deviation; MD, mean difference; CI, confidence interval.
The use of PSG in this study was limited to the rapid eye movement (REM) sleep latency parameter. Only three studies in total evaluated this parameter, including Dahlitz et al. [18] in 1991, Kayumov et al. [34] in 2001, and Rahman et al. [38] in 2010. Based on the study conducted by Dahlitz et al. [18] in 1991, there was an 11-minute reduction in REM sleep latency in the DSWPD group receiving melatonin. The results of the meta-analysis (Fig. 7) on this parameter indicate a significant effect of melatonin with a decrease in REM sleep latency of 3.81 minutes, but these findings are not statistically significant (MD=-3.81, 95% CI [-8.19; 0.57], z=-1.70, p=0.0885), and interestingly, the heterogeneity test was negligible (I2=0%, p=0.6084).
Forest plot analysis of polysomnography in rapid eye movement sleep latency. SD, standard deviation; MD, mean difference; CI, confidence interval.
Subgroup and Meta-Regression Analyses
Quantitative synthesis results in subgroup analysis of 9 studies using high doses of exogenous melatonin (≥3 mg) showed a significant decrease in SOL (Fig. 8A) (MD=-18.97, 95% CI [-28.02; -9.92], z=-4.11, p<0.0001) compared to low-dose melatonin (<3 mg).
Forest plot of exogenous melatonin at doses ≥3 mg versus <3 mg. A: Subgroup analysis of exogenous melatonin at doses ≥3 mg versus <3 mg on sleep onset latency scores. B: Subgroup analysis of exogenous melatonin at doses ≥3 mg versus <3 mg on total sleep time scores. MD, mean difference; CI, confidence interval.
In this study, meta-regression analysis (Table 3) showed that melatonin effect (β=-105.0, p<0.001), melatonin dose (β=47.1, p=0.018), treatment duration (β=78.9, p<0.001), and time of administration (β=74.6, p=0.005) are the main indicators influencing the improvement in SOL sleep quality in DSWPD patients. Pharmacologically, exogenous melatonin causes DSWPD patients to fall asleep 105 minutes earlier than placebo recipients. Additionally, the use of high-dose exogenous melatonin (≥3 mg) showed a greater reduction in SOL than the use of low-dose exogenous melatonin (<3 mg), while a treatment duration of more than 4 months and melatonin administration at night, 3–5 hours before DLMO, were considered more effective and beneficial in making DSWPD patients feel sleepy sooner so they could fall asleep earlier.
Meta-regression for subgroup analysis of exogenous melatonin with SOL and TST parameters as dependent variables
Meanwhile, the use of low-dose melatonin (<3 mg) showed a significant increase in TST (Fig. 8B) (MD=51.90, 95% CI [36.31; 67.49], z=6.52, p<0.0001) compared to when high-dose melatonin (≥3 mg) was administered.
Meta-regression in Table 3 shows that the improvement in sleep quality in DSWPD patients was strongly influenced by patient age (β=39.30, p=0.020), dose (β=49.90, p=0.009), and time of melatonin administration (β=-49.60, p=0.042). In pediatric patients, exogenous melatonin use was found to prolong sleep time more effectively, although an increase in TST was also observed in adult patients, although not statistically significant. Interestingly, high-dose melatonin (<3 mg) was more effective in increasing TST in DSWPD patients than low-dose melatonin (≥3 mg). Additionally, administering melatonin 3–5 hours before DLMO at night can prolong TST in DSWPD patients.
Analysis of Melatonin’s Side Effects
A meta-analysis of melatonin side effects was conducted on 9 studies categorized into neurological, gastrointestinal, and systemic effects for better subgroup analysis (Fig. 9). The overall quantitative analysis results are shown by an effect size OR of 0.25 (95% CI [0.16; 0.38], z=4.15, p<0.0001) with low heterogeneity (I2=37%, p=0.0585). These results indicate that recipients of exogenous melatonin have a 75% lower chance of experiencing side effects.
Forest plot of melatonin exogenous side effect. OR, odds ratio; CI, confidence interval.
Subgroup analysis for neurological and systemic effects showed significant reporting of side effects (OR=0.30, 95% CI [0.17; 0.53], z=3.66, p=0.0003 and OR=0.30, 95% CI [0.12; 0.74], z=2.37, p=0.0177, respectively). This indicates that melatonin recipients had a 70% lower chance of experiencing side effects such as headaches, drowsiness, insomnia, and itchy skin.
Meanwhile, for gastrointestinal side effects, the effect was not significant (OR=0.12, 95% CI [0.05; 0.31], z=0.38, p=0.7050). Although the chance of experiencing nausea and vomiting was 88% lower in melatonin recipients, this was not statistically significant.
DISCUSSION
Based on what we have found in this study in the context of demographics, we found various DSWPD patients from several regions. Based on the results of a global epidemiological survey, Europe is a region with a fairly high incidence of sleep disorders, with around 30% of young people in Europe experiencing sleep disorders and around 3.3% of patients experiencing DSWPD. Around 5.8%–34.8% of young people over the age of 18 experience insomnia, and around 22.7% experience a fairly progressive obstructive sleep apnoea [50,51]. Approximately 11 of our 18 studies originated from Europe, indicating that our findings align with previous studies, which reported and researched numerous cases of sleep disorders in Europe. Given that Europe is home to many developed countries, this indicates that socioeconomic factors can be a trigger for prolonged stress, causing people to experience health problems that affect their sleep function [52,53]. In addition, good public health awareness and access to health services will enable individuals with DSWPD to be easily diagnosed [53]. However, cases of DSWPD are not only found in Europe; several regions around the world have also reported similar findings [2,9,13,35,36].
In this study, approximately 12 of the 18 included studies examined DSWPD in children and the remaining 6 studies examined it in adults. In children, DSWPD is often caused by exposure to blue light from playing with gadgets at night, which can delay the release of melatonin, creating a disruption in the biological signals that initiate sleep [54]. A systematic review and meta-analysis conducted by Andersen et al. [55] in 2025 stated that puberty can affect circadian rhythm shifts by delaying melatonin release by 1–2 hours, making it difficult for people with DSWPD to fall asleep and even causing them to wake up in the morning. In adults who have had difficulty initiating sleep since childhood, there are indications that they have a variation in the Period Circadian Regulator 3 (PER3) gene that can physiologically delay melatonin release [35,56].
Sleep Quality and Sleep Chronobiology in DSWPD Patients
One of our main findings in this study was clinically and objectively proven improvement in sleep quality in DSWPD patients after exogenous melatonin administration. Several previous studies have demonstrated that exogenous melatonin improves sleep quality in patients with sleep disorders. An earlier study by van Geijlswijk et al. [20] in 2010 had previously demonstrated the administration of exogenous melatonin in DSWPD patients, with results showing a decrease in SOL of 16.04 minutes in pediatric patients and 30.28 minutes in adult patients, as well as an increase in TST in DSWPD patients of 28.39 minutes in children and 0.77 minutes in adults. Even the results of a meta-analysis by Hanafi et al. [57] in 2025 stated that exogenous melatonin was quite effective in reducing WASO (p<0.0001) in patients with sleep disorders. A previous meta-analysis study by Brzezinski et al. [58] in 2005 demonstrated an increase in SE of approximately 2.2% in patients with sleep disorders. As a chronobiotic compound capable of shifting the circadian rhythm, melatonin can shift the DLMO time in adult DSWPD patients to 1.69 hours earlier in adults and 1.13 hours earlier in pediatric patients, a shift that is more advanced than a few days earlier [20].
In addition, several studies have reported a reduction in sleep disturbance symptoms such as fatigue, difficulty falling asleep, and frequent late waking. A study by Wilhelmsen-Langeland et al. [48] in 2013 demonstrated a reduction in sleepiness during daily activities, as indicated by a decrease in ESS scores. Studies by Magee et al. [35], Saxvig et al. [39], Sletten et al. [19], and Swanson et al. [42] mention an improvement in the sleep quality of DSWPD patients based on a decrease in PSQI scores after receiving exogenous melatonin. Based on all these findings and analysis results, exogenous melatonin is one of the compounds that helps improve slow-wave sleep, which helps the body undergo the rest process, thereby creating comfortable sleeping conditions and preventing early awakening [57,59,60]. The improved sleep quality in DSWPD patients who took exogenous melatonin reflects a reduction in sleep fragmentation, resulting in good and efficient sleep consolidation [15,61,62]. Even insomnia symptoms such as feeling restless and difficulty falling asleep in DSWPD patients began to decrease after consuming exogenous melatonin for a treatment duration of 1 month [19,35].
Thus, exogenous melatonin has efficacy, therapeutic potential, and beneficial chronobiotic effects in initiating sleep and regulating circadian rhythms, thereby improving the sleep quality of patients with DSWPD. These findings can be used as initial clinical considerations and as a basis for more standardized research.
Clinical Treatment Recommendations for DSWPD Patients
Analysis of sleep quality and sleep duration in DSWPD patients undergoing melatonin treatment is greatly influenced by several factors that can affect the success and effectiveness of exogenous melatonin compounds. Various studies have examined key indicators such as melatonin dosage and timing of administration in relation to the success of melatonin treatment in DSWPD patients [63].
Pharmacologically, the use of melatonin at certain doses can have varying effects on users. According to the Food and Drug Administration (FDA), melatonin is not classified as a drug, so dosage requirements have not been officially determined [64]. To date, many studies have conducted melatonin interventions on patients with sleep disorders to improve sleep quality using various doses [65]. As far as we know, the lowest dose of melatonin commonly used by clinicians to treat patients with sleep disorders is 0.1 mg [66].
Clinically, several RCT studies have confirmed that the use of low-dose (0.3–1 mg) and high-dose (≥3 mg) melatonin can improve sleep quality in patients with DSWPD [33,35,36,43]. The use of low-dose exogenous melatonin (1 mg) is more recommended in increasing TST for pediatric patients with chronic sleep-onset insomnia, delayed melatonin secretion under dim light conditions, and delayed sleep phase syndrome, due to the slow metabolic process in children [67]. On the other hand, high-dose melatonin (≥3 mg) leads to a reduction in SOT in patients [59]. Pharmacologically, the use of melatonin at a dose of 3 mg to 5 mg creates a supraphysiological plasma concentration 10 times higher than endogenous melatonin [68]. A strong surge in melatonin concentration within 30–60 minutes causes the threshold for hypnotic effects to be exceeded, so that drowsiness is felt more quickly [69]. All of these clinical findings are highly relevant to the results of the meta-analysis and meta-regression in our study. However, these clinical findings exhibit moderate to high heterogeneity between studies and varying population characteristics, meaning that clinical conclusions must be interpreted with caution and careful monitoring.
Biologically, the human body has a phase response curve (PRC) to melatonin, a pattern of circadian phase shift in response to the time of melatonin administration [70]. Exogenous melatonin administration that begins when it starts to get dark until close to DLMO is called phase advance [71]. To regulate the circadian rhythm in DSWPD patients, exogenous melatonin is often administered in phase advance [33,37,43-45]. Exogenous melatonin administration 30–60 minutes before bedtime may not effectively shift the circadian rhythm in DSWPD patients but may only provide a sedative effect that reduces anxiety levels and calms the mind, resulting in a slow onset of hypnotic effects [72]. Meanwhile, exogenous melatonin administered during the day will indeed provide a sedative effect, but this effect will quickly fade during the day because melatonin is rapidly metabolized by the body when sunlight stimulation is still present [73]. As the first study to analyze the relationship between the timing of exogenous melatonin administration and the PRC to melatonin, Mundey et al. [36] in 2005 noted a significant systematic correlation (r2=0.94, p<0.0001) between phase advance at DLMO time and exogenous melatonin administration time. The study also noted that the use of 0.3 mg of exogenous melatonin administered 1.5 to 6.5 hours before DLMO time resulted in a favorable chronobiological shift in sleep for patients with DSWPD. These findings were also consistent with the results of meta-regression in our study. The moderate-to-high level of heterogeneity requires clinicians to exercise caution when using melatonin in clinical practice.
Thus, the dose and timing of melatonin administration can have a significant effect on improving the sleep quality of patients with DSWPD. However, all of these findings have a high degree of study variability and heterogeneity, so the use of melatonin dosage and administration 3–5 hours before DLMO can be considered a very promising approach, given that several studies have proven the potential of exogenous melatonin compounds in improving the sleep quality of DSWPD patients.
Adverse Effect of Exogenous Melatonin
After reviewing the effectiveness of melatonin, we also analyzed the safety of melatonin use. In the meta-analysis in this study, we found that the overall chance of adverse side effects from melatonin use was small, at 75%. In Table 4, we summarize studies that reported side effects commonly reported by DSWPD patients after taking exogenous melatonin. Approximately 54% of the total studies in this research reported neurological disorders such as headache, dizziness, sleepiness, nightmares, agitation, epileptic seizures, decreased mood, and hyperactivity after using exogenous melatonin. Approximately 14% of exogenous melatonin users reported digestive disorders such as diarrhea, nausea, and gastrointestinal illness, while 32% reported harmless systemic symptoms. In 2016, the French National Agency for the Safety of Medicines and Health Products examined various side effects commonly caused by exogenous melatonin, with approximately 43% mentioning neurological disorders such as headaches, followed by 19% reporting systemic symptoms such as rashes and maculopapular rashes, and another 19% reporting digestive disorders such as constipation, nausea, and acute pancreatitis [74]. Interestingly, an RCT study by Jun et al. [75] in 2019 demonstrated that the use of 2 mg PR melatonin did not cause any side effects in the study subjects. In fact, the study subjects reported an improvement in sleep disorder symptoms, unlike the placebo group, who reported persistent sleep disorder symptoms. Furthermore, a study by Kunz et al. [76] in 2004 also noted no side effects after using FR melatonin 3 mg, and instead reported an improvement in sleep disorder symptoms.
Adverse effect of exogenous melatonin
Now, the FDA’s decision to classify melatonin as a non-drug means that exogenous melatonin is widely available on the open market in supplement form so that its clinical benefits can be enjoyed to treat sleep disorders [64]. Interestingly, these supplements do not contain exogenous melatonin alone; often, a single supplement capsule contains vitamins and several micronutrients that enhance the work of exogenous melatonin [77,78]. These vitamins and micronutrients enhance the efficacy of melatonin and reduce certain side effects that often disrupt patients’ activities, such as headaches [79].
Strengths and Limitation of the Research
In writing this review, we are well aware that it has several limitations. In writing this paper, we are well aware that this review has several limitations. The results of this study have a moderate to high level of heterogeneity, so that clinical findings on the benefits of exogenous melatonin may have different effects in some studies. Therefore, all clinical findings, such as chronobiotic effects and the benefits of melatonin, melatonin dosage, and timing of melatonin administration in patients with DSWPD, cannot yet be used as definitive clinical references. We have attempted to use subgroup analysis and random effects model methods and 95% prediction interval calculations to accommodate variations in characteristics between studies. Thus, all clinical findings obtained in this study can be used as initial clinical considerations and a more potential approach to improve the sleep quality of DSWPD patients.
In addition, several studies used varying standard criteria in establishing a clinical diagnosis of DSWPD, resulting in inaccurate reporting. Technically, from the 18 included studies, 784 DSWPD patients were identified, resulting in a high degree of heterogeneity in each statistical analysis, requiring caution in interpreting the results for clinical implementation. Furthermore, several studies used open-label methodologies, which could influence outcome measurements and cause methodological bias. In addition, among all the included studies covering several continents, we did not find any RCTs originating from Africa or Asia. This could be due to the small number of DSWPD patients in these regions or even the absence of RCTs conducted there.
Based on several previous meta-analyses, our study is a meta-analysis that analyses sleep quality and sleep duration along with the selection of standard melatonin dosage and timing for DSWPD patients in several RCT studies classified as primary research with the most robust evidence. During the preparation of this review, we have aligned the writing of the systematic review and meta-analysis with the PRISMA checklist.
In conclusion, based on all the evidence from qualitative and quantitative analyzes in this study, we found that the timing of melatonin administration and the dose of exogenous melatonin greatly determine the success of melatonin treatment in improving sleep quality and shifting the sleep time of DSWPD patients. However, clinical recommendations regarding the dose and timing of melatonin administration still require more comparative clinical analysis.
Supplementary Materials
The online-only Data Supplement is available with this article at https://doi.org/10.17241/smr.2026.03426.
Supplementary Table 1.
Search strategy using PubMed database
Supplementary Table 2.
Search strategy using Scopus, Web of Science, and Proquest database
Supplementary Table 3.
Changes in the mean and standard deviation of sleep quality on actigraphy
Supplementary Table 4.
Changes in the mean and standard deviation of sleep timing on actigraphy
Supplementary Table 5.
Changes in the mean and standard deviation of sleep quality on sleep questionnaire score and polysomnography
Notes
Availability of Data and Material
All data generated or analyzed during the study are available in the published article.
Author Contributions
Conceptualization: Moh Ikhsan Hanafi. Data curation: Moh Ikhsan Hanafi. Formal analysis: Moh Ikhsan Hanafi. Investigation: Moh Ikhsan Hanafi, Zalfa Anindya Laksana. Methodology: Moh Ikhsan Hanafi. Project administration: Moh Ikhsan Hanafi. Resources: Moh Ikhsan Hanafi, Zalfa Anindya Laksana, Chandra Nur Iman Ardiyan. Supervision: Wardah Rahmatul Islamiyah, Fidiana Fidiana. Validation: Moh Ikhsan Hanafi, Wardah Rahmatul Islamiyah. Visualization: Moh Ikhsan Hanafi. Writing—original draft: Moh Ikhsan Hanafi. Writing—review & editing: all authors.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Funding Statement
None
Acknowledgements
None