A literature search for this study was conducted by the first author (MIH) and assisted by the second author (DFFA) using five electronic databases, specifically PubMed, Web of Science, Scopus, ProQuest, and Taylor and Francis, to identify relevant research articles. Literature searches were conducted by entering several keywords such as “Melatonin Treatment,” “REM Sleep Behavior Disorder,” and “Randomized Controlled Trial” or “RCT” and combined using Boolean operators including “OR” or “AND” to ensure comprehensive scope. In searching for studies, the main author did not set any publication date limitation in order to identify all relevant randomized controlled trial (RCT) studies holistically. The results of the literature search from each electronic database are shown in Supplementary Table 1 (in the online-only Data Supplement).

The literature search process was conducted by compiling study results from various databases and managing them using Rayyan. ai software (Rayyan). Before entering the studies obtained into Rayyan.ai software, the first author (MIH) would count the number of studies obtained from each database. The studies obtained were then entered into the Rayyan.ai software, and several studies identified as duplicates were removed. Initial selection was based on the title and abstract. Studies that passed this stage were then assessed for overall eligibility based on predetermined criteria. The reasons for excluding studies at each stage are explained transparently in the PRISMA diagram. The entire process of searching and selecting studies will be performed independently by three authors (MIH, DFFA, and ZAP), and any differences of opinion that arise will be resolved through deliberation to achieve a more objective result. Details of the selection process can be seen in Fig. 1.

This systematic review and meta-analysis applied the PICO (Population, Intervention, Comparison, Outcome) framework in developing eligibility criteria. The studies to be included in this research must meet the inclusion criteria, including 1) the population consists of individuals diagnosed with RBD based on polysomnography results, 2) the use of melatonin as treatment, 3) studies with a RCT design, and 4) studies available in English. Studies will be excluded if only the abstract is available.

Each of the authors will extract data from the included studies. The first author (MIH) will begin entering the data into a spreadsheet and ensure that the collected data is accurate. The data to be collected for this systematic review and meta-analysis includes the name of the first author and year of publication, location of the study, study design, population characteristics, sample size in the experimental and control groups, age of the population, therapeutic dose used, frequency of therapy, type of control, assessment period in the population, and mean± standard deviation (SD) values of the study outcome. The results of data extraction from each study were presented quantitatively and qualitatively in tabular format. The collected studies were assessed for quality using the Cochrane risk-of-bias for version 2 randomized trials (RoB 2) by authors MIH and DFFA. The involvement of author 3 was used to resolve differences between the assessment results of MIH and DFFA so that a more objective decision could be reached. The results of the study quality assessment will be visualized using Risk of Bias VISualization (robvis).

The quantitative analysis in this meta-analysis used R. Studio 4.5.0 software (R Foundation for Statistical Computing). The data entered into R Studio included the mean, SD, and sample size from selected studies agreed upon by authors. These datasets were then processed to generate several outputs, including a heterogeneity test and an analysis of the intervention effect represented as the mean difference (MD). The mean values extracted and the associated SD from each included study have been systematically compiled into an extraction table. The final values for mean change and SD, presented in Supplementary Tables 2 and 3 (in the online-only Data Supplement), were directly taken from the original studies.

A heterogeneity test was conducted to assess the level of variation between studies included in the meta-analysis, measured by using Higgins’ I² statistic. This statistic represents the percentage of variation between studies, interpreted as low heterogeneity when I² < 50%, and I² ≥ 50% indicating high heterogeneity.

Based on the results of the heterogeneity test, the appropriate analytical model is determined using either the fixed effect model or the random effect model. The fixed effect model is applied when heterogeneity is low (I² <50%), based on the assumption that all studies share the same true effect size and that variation in results is only due to sampling error. Conversely, when heterogeneity is high (I² ≥50%), the random effect model is used to account for significant variation between studies, which may arise from differences in population characteristics, research methodology, or variation in the interventions applied. All meta-analysis results are comprehensively visualized through forest plots to facilitate clinical interpretation. Publication bias was evaluated quantitatively using Egger’s test to provide a comprehensive insight into the potential bias in the synthesis of this research data.

The search results from each database provided a total of 1,499 study articles. A total of 192 studies identified as duplicates were then removed, leaving 1,307 studies to be selected based on their titles and abstracts. Subsequently, 1,294 studies were excluded through analysis because their titles and populations did not meet the eligibility criteria. A total of 13 remaining studies were reviewed in their entirety, leaving six studies that met the eligibility criteria. The entire results of the study search process can be seen in Fig. 1. The results of the study quality assessment by Byun et al. [10] in 2023 showed a “some concern” bias, and the results of the remaining five studies by Hadi et al. [13] in 2022, Gilat et al. [14] in 2020, Jun et al. [15] in 2019, Kunz and Mahlberg [11] in 2010, and Kunz et al. [12] in 2004 showed a low risk of bias. The quality assessment conducted on the six included studies can be viewed visually in the traffic light plot (Fig. 2) and bias summary plot (Fig. 3).

All data processing results were presented in an extraction table (Table 1). Based on six articles that fulfilled the inclusion criteria, there were 169 patients diagnosed with RBD from South Korea, Germany, Iran, and Australia. The average age of patients involved in this study was more than 60 years. Pathophysiologically, there is a relationship between people suffering from RBD and Parkinsonism, elderly people who have suffered from RBD for more than 10 years can lead to neurodegenerative diseases such as Parkinson’s [4,16].

Based on the six articles included, studies by Byun et al. [10] in 2023, Gilat et al. [14] in 2020, and Jun et al. [15] in 2019 stated that the exogenous melatonin brand used in their research was Circadin^® 2 mg tablets. Meanwhile, studies by Kunz and Mahlberg [11] in 2010 and Kunz et al. [12] in 2004 mention that the melatonin used as an intervention in their research was a product manufactured by Helsinn Advanced Synthesis SA, Biasca, Switzerland, which was then analyzed for effectiveness by the Department of Pharmacology, Freie University, Berlin, Germany. However, there is one study that does not mention the type or brand of melatonin used in its research intervention [13]. Out of six research studies, five studies mentioned that the intervention time was conducted for 4 weeks, followed by analysis at baseline and post-intervention. The study by Gilat et al. [14] in 2020 conducted research for 8 weeks.

The results of the meta-analysis for sleep questionnaire scores are presented in a forest plot (Fig. 4). For Pittsburgh Sleep Quality Index (PSQI) scores, the forest plot falls within the no effect range, indicating that there is no significant effect of melatonin (MD=0.41, 95% CI [−0.71; 1.53], p=0.4707). Additionally, the forest plot results for Epworth Sleepiness Scale (ESS) questionnaire scores also indicate no significant effect of melatonin (MD= 0.36, 95% CI [−1.30; 2.02], p=0.6716). Moreover, the meta-analysis results for Rapid Eye Movement Sleep Behavior Disorder Screening Questionnaire (RBDSQ) questionnaire scores also show no significant effect of melatonin in the statistics (MD= −0.66, 95% CI [−3.05; 1.74], p=0.5913).

Thus, the results of the meta-analysis presented in the forest plot (Fig. 4) show that melatonin administration does not have a statistically significant effect (pooled MD=0.18, 95% CI [−0.65; 1.02], p=0.6681). This is evidenced by the pooled MD whose confidence interval range crosses the no effect line (zero value). Additionally, Egger’s test performed to identify publication bias in questionnaire score analysis also resulted in insignificant findings (p=0.5089), indicating that there is no potential publication bias in the results of this study.

The results of the meta-analysis of polysomnography parameters are presented in a forest plot (Fig. 5). The results of the intervention studied showed significant changes in several aspects of sleep, including a decrease in wakefulness after sleep onset (WASO) duration (MD=−11.17; 95% CI [−13.35; −8.99]; p< 0.0001), a decrease in SOL (MD=−7.05; 95% CI [−11.40; −2.69]; p=0.0015), and an increase in time spent in REM sleep (MD= 3.93; 95% CI [2.88; 4.99]; p<0.0001). However, the intervention did not show a statistically significant effect on total sleep time (TST) (MD=24.79; 95% CI [−22.43; 72.01]; p=0.3036), sleep efficiency (SE) (MD=4.26; 95% CI [−0.78; 9.29]; p=0.0974), and REM sleep latency (MD=−7.06; 95% CI [−17.80; 3.67]; p=0.1973).

Thus, the overall meta-analysis results presented in the forest plot (Fig. 5) show that melatonin administration has a statistically significant effect. This effect is seen in two different types of outcomes, firstly (Fig. 5A) in parameters that get better when the PSG score decreases (pooled MD=−10.24; 95% CI [−12.16; −8.32]; p<0.0001), and secondly (Fig. 5B) in parameters that get better when the value increases (pooled MD=5.72; 95% CI [2.81; 8.64]; p=0.0001). The significance of these results was characterized by confidence intervals that did not cross the neutral line (zero value) and highly significant p-values. Furthermore, Egger’s test conducted to measure publication bias quantitatively in both analyses showed non-significant results (p=0.3516 and p=0.5089, respectively), thus estimating that there was no potential publication bias affecting the results of this study.

Based on the interpretation of the forest plot analyzing the side effects of melatonin therapy (Fig. 6), it could be concluded that there was no statistically significant difference in the incidence of side effects between the melatonin intervention group and the control group. This was shown by the overall effect value which was not significant (z=−1.24; p=0.2166) with an overall odds ratio (OR) estimate of 0.54 (95% CI [0.20; 1.44]). The confidence interval that includes the value 1 (null effect) indicates that melatonin did not significantly increase or decrease the overall risk of side effects. Further analysis based on subgroup side effects, namely sleeping disorders and other neurological disorders, also showed consistent results. However, there were no significant effects observed in both subgroup side effects (p=0.7748 and p=0.1788, respectively). Furthermore, there was no heterogeneity between studies (I²=0% for all analyses) and differences between subgroups were also insignificant (p=0.5101), reinforcing the conclusion that these findings are consistent and not influenced by variations between studies or types of side effects. Thus, it could be concluded that melatonin had a safety profile that was not significantly different from placebo or control in the studies analyzed.

Based on epidemiological analysis, approximately 21% of patients with a history of head trauma and posttraumatic stress disorder are at risk factors underlying the etiology of RBD [17–19]. Head trauma can cause damage to brain tissue, particularly in areas that regulate muscle tone during REM sleep, such as the locus subcoeruleus, thereby disrupting the normal mechanism of muscle atonia and causing uncontrolled motor behavior [20]. Head trauma can cause damage to brain tissue, particularly in areas that regulate muscle tone during REM sleep, such as the locus subcoeruleus, thereby disrupting the normal mechanism of muscle atonia and causing uncontrolled motor behavior [4,16]. Based on studies, the prevalence of RBD in the general population is relatively low, at around 0.5%, but increases significantly with the aging process, reaching 5%–13% in the over-60 age group [7,21]. Pathophysiologically, the ageing process is associated with nerve degeneration and the accumulation of α-synucleinopathies in the locus coeruleus, thereby disrupting its function [22]. Disruption of the physiological function of the subcoeruleus nucleus causes degeneration of glutamatergic signals from the subcoeruleus locus to the medullary reticular formation and spinal ventral horn interneurons, resulting in reduced GABAergic inhibition of spinal motor neurons, thus causing characteristic motor behavior, such as complex movements and simple twitching or contractions during REM sleep in the muscles of affected individuals [23]. Thus, both head injuries and the aging process can cause damage to the area of the brain that regulates REM sleep, thereby triggering RBD through a mechanism of progressive nerve damage.

Physiologically, when the environment begins to get dark, the suprachiasmatic nucleus (SCN) receives light signals from the environment to synchronize its circadian rhythm [24]. These signals initiate the formation of messages related to dark conditions, which are then sent to the pineal gland to produce melatonin [25]. This melatonin hormone then binds to its receptors, which are spread throughout various parts of the body, sending a signal of sleepiness to all organs and preparing the body for sleep [26]. Melatonin binds to its receptors, such as melatonin type 1 receptor (MT₁), melatonin type 2 receptor (MT₂), melatonin type 3 receptor (MT₃), and Retinoic Acid-Related Orphan Receptors (ROR) or Retinoid Z Receptor (RZR), which are located in the anterior pituitary gland pars tuberalis, SCN, hypothalamus, as well as the cortex, thalamus, substantia nigra, nucleus accumbens, amygdala, hippocampus, cerebellum, cornea, and retina to initiate a series of physiological reactions [27].

In addition, MT2 receptors are also located in the periaqueductal grey (PAG), a brainstem structure that forms the basis of RBD pathophysiology [28]. The binding of melatonin to MT2 receptors in the PAG causes a series of chemical reactions that prevent the formation of muscle twitch control to the motor cortex and pyramidal cells, thus preventing muscle twitching in the patient’s body during sleep, which would otherwise interfere with the sleep process [20].

Based on several studies, melatonin is able to penetrate cell membranes and enter the cytoplasm to interact directly with specific receptors within it, the MT₃ receptors, also known as quinone reductase 2 (QR2) enzyme [29]. As an enzyme involved in detoxification and possessing antioxidant properties, the activity of the QR2 enzyme is highly dependent on the availability of its co-substrate, melatonin [30]. Naturally, melatonin levels in the body are not constant over time, fluctuating according to the light-dark cycle, increasing with age, and depending on physical condition or disease factors [31]. Consequently, changes in melatonin production, whether physiological or pathological, may affect the activity level of the QR2 enzyme [32].

Within the cell nucleus, melatonin is also able to form bonds with nuclear receptors from the ROR or RZR groups [33]. The interaction between melatonin and these receptors triggers a series of important physiological reactions, particularly the activation of ROR, which plays a role in regulating gene transcription and increasing the expression of its target genes [34]. In addition, these receptors are also found in the pineal gland and various major tissues that essentially function to modulate the internal biological clock (circadian rhythm) by regulating the expression of circadian-related genes [35]. Thus, the interaction of melatonin with nuclear receptors not only regulates gene transcription in the body, but also plays an important role in maintaining the body’s circadian rhythm balance.

Clinically, clonazepam and melatonin are compounds that are often used to treat sleep disorders, but the AASM has designated clonazepam as the first-line therapy for treating sleep disorders [8]. Pharmacologically, clonazepam, a drug in the benzodiazepine group, is effective in treating sleep disorders but has more severe side effects [8,36]. Research from Australia in 2018 explains that individuals who use benzodiazepine drugs are at greater risk of experiencing symptoms of dementia, such as difficulty thinking and memory problems [37].

Based on a study Byun et al. [10] in 2023, clonazepam therapy was shown to reduce muscle tone in RBD patients, particularly in the chin muscles (tonic muscle tone: p=0.039; “any” muscle tone: p=0.015) and the combination of the tibialis anterior and chin muscles (p=0.031). This reduction in muscle tone can decrease the intensity of movements during sleep, thereby improving patients’ motor behavior control. Polysomnography results also showed objective improvement after 4 weeks of clonazepam use. However, subjective results from sleep questionnaires differed from polysomnography results. ESS and Insomnia Severity Index (ISI) scores decreased significantly in the melatonin group (improving sleep quality), while in the clonazepam group, ESS scores increased (p=0.002), indicating daytime sleepiness. Additionally, clonazepam was reported to cause worse side effects compared to melatonin. Thus, the results of the analysis of both therapies proved to improve the sleep quality of RBD patients, but melatonin showed better safety and tolerance based on patients’ subjective reports.

In a study conducted by Hadi et al. [13] in 2022, there was a significant correlation between melatonin administration in RBD patients and the severity of PD, as measured using the Hoehn and Yahr score (p<0.001). Further evaluation using the PSQI and RBDSQ sleep questionnaires also showed that melatonin had a significant effect. In the melatonin group, the PSQI score increased by 5 points from the baseline value, while the control group only increased by 3 points. Not only that, after 4 weeks of intervention, the RBDSQ score in the melatonin group also decreased significantly compared to the control group (p<0.001). However, this study did not report significant changes in the ESS score.

Based on a randomized, double-blind study by Gilat et al. [14] in 2020 involving 30 patients, the administration of 4 mg of melatonin was well tolerated by RBD patients with PD. However, the results of the study showed that melatonin was not effective in improving the symptoms of RBD itself. However, objective analysis through polysomnography discovered a significant reduction in the time required to fall asleep (SOL, p=0.002). Subjectively, the group receiving melatonin also reported feeling more refreshed and less fatigued based on the 36-item Sort-Form (SF) Survey questionnaire.

According to a 2019 study from South Korea, statistically insignificant results were recorded after a 4-week intervention involving the administration of 6 mg of melatonin to nine patients with RBD [15]. Although there was a decrease in scores on several questionnaires, such as the REM Sleep Behavior Disorder Screening Questionnaire-Korean Version (RBDQ-KR), PSQI, and ESS, these scores were not strong enough to meet the threshold for statistical significance. Even the group receiving melatonin showed a decrease in Clinical Global Impression-Improvement scores, but these results did not show a statistically significant difference compared to the placebo group [15].

There is also a study from Germany in 2010 which revealed that melatonin therapy can improve sleep quality in RBD patients [11]. This was proven by polysomnography results from 8 patients who experienced a significant increase in the time needed to fall asleep (SOL, p=0.05) and SE (p=0.043). Additionally, about half of the patients (four individuals) even reported that their RBD symptoms disappeared completely after undergoing therapy. Not only did they experience clinical improvement, but some patients also felt subjective benefits such as feeling more refreshed and energetic after waking up [11].

In a previous study, Kunz et al. [12] in 2004 involved 14 patients and found that melatonin administration had a significant effect on increasing the percentage of REM sleep, improving REM sleep continuity, and reducing body temperature during sleep. The findings from this crossover study design also indicate that exogenous melatonin administered at the appropriate time can normalize circadian variations in human physiology [12]. This normalization of circadian rhythms has a strong positive effect on overall health, particularly in the elderly and shift workers [12].

Based on a systematic review and meta-analysis of six studies, melatonin has been shown to be effective in improving sleep quality in RBD patients, although it does not cure RBD directly. The final results of this analysis show a statistically significant improvement in polysomnography parameters. This indicates that objective research using polysomnography reinforces the evidence that melatonin can improve sleep quality in RBD patients.

Pharmacologically, melatonin is known as a therapy for treating sleep disorders and even relatively safe with minor side effects [11,12,14,38,39]. However, a recent study from Byun et al. [10] in 2023 found one case in which melatonin administration actually worsened insomnia, which was thought to be due to the patient’s chronotype mismatch and the use of a low-dose melatonin formulation (2 mg). Although these results appear inconsistent with evidence supporting the efficacy of melatonin, it is important to remember that sleep quality assessments are highly influenced by individual subjective perceptions.

Pharmacologically, the optimal dose of melatonin is not yet known, but some studies indicate that melatonin doses of 1–6 mg are generally effective and cause only minor side effects such as headaches, fatigue, or daytime drowsiness [15,39–41]. These consistency findings were supported by a study by Hadi et al. [13] in 2022, which reported no serious side effects with a 3-mg dose of melatonin, and a study by Gilat et al. [14] in 2020, which showed only mild effects such as morning drowsiness with a 4 mg dose. Additionally, other studies, including Kunz et al. [12] in 2004 and Kunz and Mahlberg [11] in 2010, also found no significant side effects at a dose of 3 mg, although the reports on improvements in patient stamina and comfort were subjective. The use of high doses of melatonin (more than 10 mg), although relatively safe in the short term, does not show significant additional clinical benefits compared to low doses [41]. In fact, high doses or overdose can potentially cause systemic effects such as confusion, hypothermia, hypotension, and tachycardia, which require close monitoring and dose adjustment, especially in susceptible populations [42].

Based on a systematic review and meta-analysis of six RCT studies, this study focuses specifically on the effectiveness of melatonin for sleep disorders in RBD, which has not been the main focus of previous meta-analyses. Although the RCT design provides a low level of bias with a total of 169 participants, this relatively small number of subjects resulted in high heterogeneity in the analysis. Therefore, generalization of the findings requires further experimental studies with larger populations.

In conclusion, based on this systematic review and meta-analysis, melatonin therapy has a significant impact on improving sleep quality in patients with better safety. Therefore, it can be concluded that the use of melatonin can be a promising alternative treatment for RBD. However, further studies that are more homogeneous and comprehensive in conducting study analyses are needed.