Anti-Inflammatory Effects of Melatonin: A Systematic Review and Meta-Analysis of Clinical Trials

Abstract

Chronic inflammation contributes to multiple diseases including cardiovascular diseases, autoimmune disorders, metabolic disorders, and psychiatric conditions. Melatonin, a hormone responsible for circadian rhythm, plays a complex role within the immune system, including having an anti-inflammatory effect. While there are numerous animal studies demonstrating this effect, few human clinical trials have been conducted. This systematic review of clinical trials examined whether exogenous melatonin reduces levels of inflammatory markers in humans. We searched PubMed, Embase, Cochrane Library, Scopus, and PsycINFO, and the references of the identified articles for randomized and non-randomized placebo-controlled trials. Data were extracted from the articles and meta-analyses were conducted using a random effects model to calculate standardized mean differences (SMDs, i.e., Cohen’s d). From an initial search result of 4548 references, 31 studies met the inclusion criteria and were included involving 1517 participants. Melatonin had significant anti-inflammatory effects on interleukin (IL)-1 (SMD −1.64; 95% confidence interval [CI] −2.86, −0.43; p=0.008), IL-6 (−3.84; −5.23, −2.46; p<0.001), IL-8 (−21.06; −27.27, −14.85; p<0.001), and tumor necrosis factor (TNF) (−1.54; −2.49, −0. 58; p=0.002), but not on C-reactive protein (CRP) (−0.18; −0.91, 0.55; p=0.62). Trimming outlier studies with large effect sizes eliminated publication bias, and summary effect sizes were significant for IL-1 (SMD −1.11; 95% CI −1.90, −0.32; p=0.006), IL-6 (−1.91; −2.98, −0.83; p=0.001), and IL-8 (−13.46; −18.88, −8.04; p<0.001), but not for TNF (−0.45; −1.13, 0.23; p=0.19). Exogenous melatonin reduced levels of inflammatory markers and may be useful for prevention and adjuvant treatment of inflammatory disorders. Melatonin is safe with few side effects, which makes it an excellent agent for prevention of inflammatory disorders. Because chronic inflammation increases with aging and inflammation plays a role in the etiology of numerous diseases that affect older populations, melatonin has the potential to be widely used particularly in older adults.

1. Introduction

Chronic inflammation plays a key role in the pathophysiology of numerous common diseases including cardiovascular diseases, autoimmune disorders, metabolic disorders, and psychiatric conditions (Furman et al., 2019). Currently, there are treatment modalities that directly target inflammation, such as aspirin, non-steroidal anti-inflammatory drugs (NSAIDs), corticosteroids, and biologics (e.g., infliximab and etanercept). However, chronic use of these medications can have serious adverse effects such as gastric ulcer formation and gastrointestinal bleeding with aspirin and NSAIDs, and risk of infection with corticosteroids and biologics (Benjamin and Lappin, 2019; Yasir et al., 2020). Although problematic for anyone, these adverse effects can be especially detrimental for patients with medical comorbidities. Given this limitation, these medications are not routinely used as long-term prophylactic anti-inflammatory agents in clinical practice, perhaps with the exception of aspirin used for its anti-platelet aggregation properties and corticosteroids in the setting of organ transplant (Yasir et al., 2020).

Due to the potential risks of prescription medications, various over-the-counter supplements have been studied as alternative options for the prevention and treatment of inflammatory conditions (Tabrizi et al., 2019; Xin et al., 2012). In this review, one of these supplements will be highlighted. Melatonin (N-aceiyl-5-methoxytrypamine) is a neurohormone produced by the pineal gland as a derívate of the essential amino acid tryptophan (Escribano et al., 2014). Melatonin is produced in plants, unicellular organisms, algae, bacteria, and invertebrates. The suprachiasmatic nucleus, responsible for regulating circadian rhythm, controls the production of melatonin so that it is optimized during the night (Favero et al., 2014). Melatonin has become commonly known as a supplemental sleep aid.

However, the role of melatonin is multifaceted. Melatonin has been shown to have anti-inflammatory effect through multiple mechanisms. It reduces macromolecular damage in all organs by scavenging free radicals (Escribano et al., 2014). Melatonin is also reported to activate antioxidant defenses such as superoxide dismutase (SOD), catalase (CAT), glutathione peroxidase, glutathione reductase, and gluose-6-phosphate dehydrogenase (Radogna et al., 2010). On a genetic level, melatonin prevents the translocation of nuclear factor-kappa B (NF-κB) to the nucleus to bind onto DNA, which prevents the upregulation of the transcription and translation of inflammatory cytokines, including interleukin (IL)-1, IL-6, and tumor necrosis factor (TNF) (Escribano et al., 2014). Melatonin is also capable of downregulating the expression of 5-lipoxygenase, a key player in inflammation activation (Radogna et al., 2010). It also inhibits the production of adhesion molecules that are necessary for leukocyte adhesion to endothelial cells, thereby decreasing leukocyte migration and edema, which is a key component of the innate inflammatory response (Escribano et al., 2014).

Despite these multiple anti-inflammatory mechanisms of melatonin, its role is more complex than simply one as an anti-inflammatory mediator. Prior studies have shown that melatonin’s role differs depending on the stage of inflammation. It appears to have a pro-inflammatory role in early inflammation and an anti-inflammatory role in late inflammation (Radogna et al., 2010). In the early phase of inflammation, which is necessary for healing to occur in response to an acute insult, melatonin activates proinflammatory mediators including phospholipase A2 and arachidonate 5-lipoxygenase. This activation is a transient phenomenon and extinguishes within 2–3 hours. In the late inflammation phase, which refers to chronic inflammation responsible for many disease processes, melatonin exerts anti-inflammatory effects by downregulating the aforementioned inflammatory mediators and pro-inflammatory cytokines and also by reducing oxidative stress (Radogna et al., 2010). Furthermore, whereas several animal studies have demonstrated the anti-inflammatory effect of melatonin (Garcia et al., 2015; Li et al., 2019; Tao et al., 2018), it is not yet known whether melatonin has a consistent anti-inflammatory effect in humans. Thus, we conducted a systematic review and a meta-analysis of human clinical trials of exogenous melatonin testing its anti-inflammatory effects. We hypothesized that exogenous melatonin would have anti-inflammatory effects in humans.

2. Material and Methods

2.1. Procedures

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and included the following steps.

2.1.1. Search Strategy

A comprehensive search was conducted by two reviewers with the assistance of a health sciences librarian. We searched the following five electronic databases: PubMed, Cochrane Library, Scopus, PsycINFO, and Embase. The search was conducted to include all studies published up to and including July 2015. The reference lists of relevant reviews were also used to identify potential studies. Repeat searches of PubMed were conducted in November 2015, October 2018, and June 2019 for additional studies. The search terms included “melatonin,” “inflammation” and its derivatives, and various inflammatory markers including cytokines and acute phase proteins. Appendix 1 illustrates the search terms.

2.1.2. Study Selection

One reviewer retrieved and reviewed the full text of all articles that were considered potentially relevant based upon title and abstract, and another reviewer verified the selection. To be included in this review, articles were required to meet all of the following criteria:

1. Type of studies: randomized or non-randomized controlled trials with a placebo arm;

2. Participants: humans;

3. Intervention: exogenous melatonin;

4. Outcome: inflammatory markers including cytokines and acute phase proteins;

5. Languages: any.

Articles were excluded for any of the following reasons:

1. did not measure baseline levels of inflammatory markers;

2. did not measure unstimulated levels of inflammatory markers in body fluids;

3. evaluated a non-human sample.

The studies were categorized as either “treatment” or “prevention.” “Treatment” studies were conducted on patients with an inflammatory condition, such as rheumatoid arthritis or ulcerative colitis, whose pre- and post-melatonin inflammation levels were measured. Studies were deemed “prevention” studies if an inflammatory state was inflicted on a patient, such as endotoxin administration, exercise, or surgery. One reviewer assessed if studies should be deemed “treatment” or “prevention,” and another reviewer then verified the categorization.

2.1.3. Quality Assessment

Quality was assessed using the Cochrane Collaboration’s tool for assessing risk of bias in randomized trials as described by Higgins et al. (Higgins et al., 2019). One reviewer assessed study quality, and the other reviewer then verified it.

2.1.4. Data Extraction

The following data were extracted:

1. Author and year of publication;

2. Study design including whether the study was deemed “treatment” or “prevention”;

3. Intervention details: dosage, route, and duration of melatonin administration;

4. Duration of follow-up: in cross-over trials, the actual period of each evaluation was computed rather than the whole duration of the study;

5. Total sample size;

6. Sample characteristics: male to female ratio of the sample, whether the subjects were children or adults, mean age of the sample, and whether the subjects were healthy or not;

7. Baseline levels of inflammatory markers—i.e., pre-melatonin administration—and their standard deviations;

8. Follow-up levels of inflammatory markers—i.e., post-melatonin administration—and their standard deviations.

Most studies provided one follow-up value for inflammatory marker level. However, for the studies that had multiple time points, the follow-up time point at which the inflammatory marker level was highest in the melatonin group was utilized for the most conservative approach. Data were extracted directly from the numerical values presented in the studies. When numerical values were not presented, the authors of the studies were contacted. When there was no response from the authors, data were extrapolated from the graphs presented on the manuscripts.

2.2. Statistical Analysis

Standardized mean difference (SMD)—which is equivalent to Cohen’s d—was calculated for individual inflammatory markers using the following formula:

$$SMD=\frac{\left(M_{postmelatonin}-M_{premelatonin}\right)-\left(M_{postplacebo}-M_{preplacebo}\right)}{pooledstandarddeviation}$$

where M represents the mean inflammatory marker level detected. SMD was calculated for inflammatory markers that were measured in at least 2 studies. By convention, effect sizes of 0.2, 0.4, and 0.8 are considered small, medium, and large, respectively (Cohen, 1988). To obtain pooled effect estimates, meta-analyses were conducted using a random effects model due to the study result heterogeneity. Publication bias was tested using Egger’s test (Egger et al., 1997). Meta-analyses were then repeated after removing the studies that were contributing to the publication bias. Additionally, meta-regression was performed to examine the potential contribution of the following study-level variables to the heterogeneity of meta-analyses: mean age of study participants, study design (treatment vs. prevention, i.e., melatonin administration prior to an inflammatory challenge), participant health status (healthy vs. non-healthy), administration route of melatonin (intravenous, oral, or other), melatonin dose (regardless of administration route), oral dose of melatonin, intravenous dose of melatonin, treatment duration, and measurement of sleep as an outcome (yes vs. no). All analyses were performed using STATA 16.1 (StataCorp, College Station, TX).

3. Results

3.1. Search Results and Study Characteristics

The first search conducted in July 2015 yielded 3746 references, out of which 16 studies met the inclusion criteria and were included in the meta-analysis. The most recent search, which covered the period of July 2015 to June 2019, yielded 802 references, out of which 15 studies met the inclusion criteria and were included in the meta-analysis. Thus, in total, 31 clinical trials were included, involving 1517 participants. One manuscript reported a randomized controlled trial with 3 arms—N=45 with 15 subjects in each of placebo, low-dose melatonin, and high-dose melatonin arms (Dwaich et al., 2016). The details of this process of search and selection are shown in Figure 1. As the results were reported separately for each melatonin dose, we considered this study as 2 separate trials with N=22 in the low-dose trial and N=23 in the high-dose trial, taking into account the proper weight of each trial in meta-analyses. The characteristics of the included trials are shown in Table 1.

Figure 1.

Flow diagram of the systematic review

Table 1.

Description of the individual studies included in the meta-analysis

Author & year	N	Population	Melatonin dose	Route	Studytype	Meanage	Treatmentduration	Follow-upduration	Inflammatorymarkers	Measurement of sleep
(Alamili et al., 2014a)	24	Healthy volunteers	100 mg single dose (daytime)	IV	Prevention	23	8 hours	3 hours	IL-6, TNF, IL-1	No
(Alamili et al., 2014b)	24	Healthy volunteers	100 mg single dose (nighttime)	IV	Prevention	23	8 hours	3 hours	IL-6, TNF, IL-1	No
(Bazyar et al., 2019)	44	Patients with Type II diabetes and chronic periodontitis	6 mg/night for 8 weeks	oral	Treatment	52.6	8 weeks	8 weeks	IL-6, TNF, CRP	No
(Cavalcante et al., 2012)	36	COPD patients	3 mg/night for 3 months	oral	Treatment	66.58	3 months	3 months	IL-8	No
(Celinski et al., 2014)	46	Patients with non-alcoholic fatty liver disease	10 mg/day for 14 months	oral	Treatment	32.745	14 months	14 months	IL-6, TNF, IL-1	No
(Chojnacki et al., 2011)	60	Ulcerative colitis patients	5 mg/day for 12 months	oral	Treatment	34.75	12 months	12 months	CRP	No
(Cichoz-Lach et al., 2010)	30	Patients with nonalcoholic steatohepatitis	10 mg/day for 1 month	oral	Treatment	40	1 month	1 month	IL-6, TNF, IL-1	No
(Cobo-Vazquez et al., 2014)	8	Patients who just experienced third molar extraction	3 mg single dose	local	Prevention	22.6	1 minute	1 hour	IL-6	No
(Dwaich et al., 2016)	22	Elective CABG patients	10 mg/day for 5 days	oral	Prevention	52.4	5 days prior to provocation	1 day (after provocation)	IL-1	No
(Dwaich et al., 2016)	23	Elective CABG patients	20 mg/day for 5 days	oral	Prevention	53.2	5 days prior to provocation	1 day (after provocation)	IL-1	No
(El-Gendy et al., 2018)	40	Neonates with sepsis	20 mg single dose	oral	Treatment	0	1 minute	3 days	CRP	No
(El-Sharkawy et al., 2019)	74	Patients with generalized chronic periodontitis and primary insomnia	10 mg/night for 2 months	oral	Treatment	46.1	2 months	6 months	TNF	Yes
(Forrest et al., 2007)	75	Rheumatoid arthritis patients	10 mg/day for 6 months	oral	Treatment	62.52	6 months	6 months	IL-6, TNF, IL-1, CRP	No
(Ghaderi et al., 2019)	54	Patients on methadone maintenance treatment	10 mg/night for 12 weeks	oral	Treatment	42.6	3 months	3 months	CRP	Yes
(Gitto et al., 2004a)	20	Neonates with surgical malformations	219.9 mg single dose	IV	Treatment	0	3 days	7 days	IL-6, TNF, IL-8	No
(Gitto et al., 2004b)	74	Newborns with respiratory distress syndrome	77.22 mg single dose	IV	Treatment	0	3 days	7 days	IL-6, TNF, IL-8	No
(Gitto et al., 2005)	110	Mechanically ventilated infants of 32 weeks or less gestation	120.9 mg single dose	IV	Treatment	0	3 days	7 days	IL-6, TNF, IL-8	No
(Gitto et al., 2012)	60	Mechanically ventilated infants of 32 weeks or less gestation	10.93 mg single dose	IV	Prevention	0	1 minute	1 day	IL-6, IL-8, IL-12	No
(Hernandez-Velazquez et al., 2016)	30	Patients with suspected choledocholithiasis undergoing ERCP	50 mg total dose before, during, and after procedure	sublingual	Prevention	35.8	1.5 days	1 day	IL-6, TNF	No
(Javanmard et al., 2016)	39	3-vessel coronary artery disease patients	10 mg/night for 1 month	oral	Treatment	59.3	1 month	1 month	CRP	No
(Kücükakin et al., 2010a)	41	Patients undergoing elective laparoscopic cholecystectomy	10 mg single dose	IV	Prevention	48.1	30 minutes	1 day	CRP	No
(Kücükakin et al., 2010b)	50	Patients undergoing elective open abdominal aortic aneurysm repair or aortobifemoral bypass surgery	50 mg single IV dose and 10 mg/night for 3 nights	oral and IV	Prevention	65.52	3 days	3 days	CRP	No
(Mesri Alamdari et al., 2015)	44	Obese healthy women volunteers	6 mg/day	oral	Treatment	34.36	40 days	40 days	IL-6, TNF, CRP	No
(Montero et al., 2017)	60	Patients with diabetes and periodontal disease	1% melatonin cream/night	topical	Treatment	44.28	20 days	20 days	IL-6, IL-1	No
(Ochoa et al., 2011)	20	Healthy, regularly exercising volunteers	15 mg total dose over 3 days	oral	Prevention	40.5	3 days	3 days	IL-6, TNF	No
(Pakravan et al., 2017)	97	Patients with non-alcoholic fatty liver disease	unknown	oral	Treatment	41.56	42 days	42 days	CRP	No
(Raygan et al., 2019)	60	Patients with Type II diabetes mellitus and 2-and-3 vessel coronary heart disease	10 mg/night for 3 months	oral	Treatment	66.5	3 months	3 months	CRP	No
(Rondanelli et al., 2018)	86	Patients over the age of 65 with sarcopenia	1 mg/night for 4 weeks	oral	Treatment	81.75	4 weeks	4 weeks	CRP	No
(Sanchez-Lopez et al., 2018)	36	Relapsing-remitting multiple sclerosis patients	25 mg/day for 6 months	oral	Treatment	39	6 months	6 months	IL-6, TNF, IL-1	No
(Shafiei et al., 2018)	60	Patients undergoing CABG	20 mg total dose over 1 day	oral	Prevention	61.8	1 day	1 day	TNF	No
(Taghavi Ardakani et al., 2018)	70	Children with atopic dermatitis	6 mg/night for 6 weeks	oral	Treatment	8.65	6 weeks	6 weeks	CRP	Yes

N = number of participants; IV = intravenous; IL = interleukin; TNF = tumor necrosis factor; CRP = C-reactive protein

Most of the included studies took venous blood samples and measured plasma levels of inflammatory markers utilizing enzyme-linked immunosorbent assay (ELISA) techniques. One study measured exhaled breath condensate levels of IL-8 using enzyme immune assay (Cavalcante et al., 2012); one study measured salivary levels of TNF using ELISA (El-Sharkawy et al., 2019); and one study obtained blood samples from the blood clot found in the oral socket after dental extraction (Cobo-Vazquez et al., 2014).

3.2. Main Analyses

As shown in Figure 2, meta-analyses using a random effects model showed a robust summary effect of melatonin on the reduction of IL-1 (SMD −1.64; 95% confidence interval [CI] −2.86, −0.43; p=0.008), IL-6 (−3.84; −5.23, −2.46; p<0.001), IL-8 (−21.06; −27.27, −14.85; p<0.001), and TNF (−1.54; −2.49, −0.58; p=0.002). However, the summary effect of melatonin on CRP was not significant (−0.18; −0.91, 0.55; p=0.62).

Figure 2.

Meta-analyses of melatonin effect on inflammatory markers including all individual studies

3.3. Publication Bias & Sensitivity Analyses

Egger’s tests indicated a significant publication bias for the meta-analyses of IL-1 (bias −6.30, p=0.004), IL-6 (−8.58, p=0.004), IL-8 (−10.01, p=0.05), and TNF (−9.02, p=0.001), but not for that of CRP (3.62, p=0.37). After excluding outliers, publication bias was not significant for any of IL-1 (bias −5.54, p=0.19), IL-6 (−5.35, p=0.15), IL-8 (−9.59, p=0.19), and TNF (−6.51, p=0.11). Specifically, two outliers were excluded for IL-1 (Dwaich et al., 2016), two for IL-6 (Gitto et al., 2012; Gitto et al., 2004b), one for IL-8 (Gitto et al., 2012), and three for TNF (Gitto et al., 2004b; Gitto et al., 2004a; Sanchez-Lopez et al., 2018).

As shown in Figure 3, after excluding the identical outliers as in the above Egger’s tests, sensitivity analyses using a random effects model revealed that the effect of melatonin was still significant for IL-1 (SMD −1.11; 95% CI −1.90, −0.32; p=0.006), IL-6 (−1.91; −2.98, −0.83; p=0.001), and IL-8 (−13.46; −18.88, −8.04; p<0.001) but not for TNF (−0.45; −1.13, 0.23; p=0.19).

Figure 3.

Meta-analyses of melatonin effect on inflammatory markers after eliminating publication bias by excluding outlier studies

3.4. Meta-regression

As shown in Table 2, mean age of study participants significantly contributed to the heterogeneity of meta-analyses for IL-6 and TNF (respectively, regression coefficient 0.54, p=0.008, and 0.38, p=0.043), indicating that younger age was associated with a stronger anti-inflammatory effect of melatonin. In the meta-analysis of IL-8, compared to a single prevention study, in which melatonin was administered prior to an inflammatory challenge, treatment studies were associated with a stronger anti-inflammatory effect of melatonin (−77.57, p=0.04). Intravenous route of melatonin administration predicted a stronger anti-inflammatory effect of melatonin on IL-6 (−9.07, p=0.047). None of the other study-level variables significantly contributed to the heterogeneity of meta-analyses: healthy vs. non-healthy participants, melatonin dose (regardless of administration route), oral dose of melatonin, intravenous dose of melatonin, treatment duration, and measurement of sleep as an outcome. It should be noted that meta-regression is usually based on a small number of observations (e.g., no larger than 31 in the currently reported meta-analyses) and does not take into account any covariates. Thus, these results should be interpreted with caution.

Table 2.

Meta-regression results for each inflammatory marker

Study-level variable	IL-1		IL-6		IL-8		TNF		CRP
	Reg. Coef.	P	Reg. Coef.	P	Reg. Coef.	P	Reg. Coef.	P	Reg. Coef.	P
Mean age	−0.90	0.20	0.54	<0.01	0.57	0.45	0.38	<0.05	0.07	0.37
Study type (treatment or prevention)	−26.83	0.13	−1.63	0.86	−77.57	<0.05	7.51	0.37	7.43	0.09
Participants (non-healthy or healthy)	17.44	0.44	10.76	0.34	All non-healthy	NA	5.49	0.58	All non-healthy	NA
Administration route (oral, other, or IV)	10.88	0.33	−9.07	<0.05	−18.84	0.45	−6.17	0.14	1.32	0.64
Melatonin dose (mg)	0.22	0.43	−0.03	0.75	0.16	0.54	−0.07	0.30	0.03	0.96
Oral dose (mg)	−0.73	0.77	−0.03	0.59	Insufficient observations	NA	0.00	0.93	0.05	0.93
IV dose (mg)	Insufficient observations	NA	0.23	0.18	0.35	0.20	0.09	0.75	Insufficient observations	NA
Treatment duration (days)	0.06	0.42	0.03	0.47	0.45	0.42	0.02	0.60	−0.007	0.69
Measurement of sleep as an outcome (No or Yes)	Insufficient observations	NA	Insufficient observations	NA	Insufficient observations	NA	3.64	0.82	−1.21	0.80

Reg. Coef. = regression coefficient; IV = intravenous; IL = interleukin; TNF = tumor necrosis factor; CRP = C-reactive protein; NA = not applicable

The association between younger age and a stronger anti-inflammatory effect of melatonin was consistent and more robust than the other study-level variable findings. This meta-regression finding could be due to the individual studies conducted by Gitto and colleagues measuring IL-6, IL-8, and TNF in neonates and infants (Gitto et al., 2012; Gitto et al., 2004b; Gitto et al., 2005; Gitto et al., 2004a). After excluding these individual studies, sensitivity analyses using a random effects model revealed that the effect of melatonin was smaller but still significant for IL-6 and TNF (respectively, SMD −0.72, p=0.041, and −0.78, p=0.035). However, as 4 out of the 5 studies included in the meta-analysis of IL-8 were by Gitto and colleagues, no sensitivity analysis could be conducted for IL-8.

3.5. Quality Assessment

The majority of studies had low risk of selection bias, performance bias, detection bias, attrition bias, and reporting bias as well as low risk of all other biases (Table 3). Three studies had unclear risk of selection bias due to insufficient information about the randomization and sequence generation (Cichoz-Lach et al., 2010; Gitto et al., 2004a; Ochoa et al., 2011). Seven studies had unclear risk of selection bias based on the domain of allocation concealment (Celinski et al., 2014; Cichoz-Lach et al., 2010; El-Gendy et al., 2018; Gitto et al., 2012; Gitto et al., 2004b; Gitto et al., 2005; Gitto et al., 2004a). These seven studies did not provide sufficient information about the allocation concealment process. Six studies had unclear risk of performance bias as they had insufficient information about the blinding process (Celinski et al., 2014; Cichoz-Lach et al., 2010; Gitto et al., 2012; Gitto et al., 2004b; Gitto et al., 2005; Gitto et al., 2004a). One study was non-randomized and non-blinded and thus had a high risk of selection bias and performance bias (El-Gendy et al., 2018). After excluding El-Gendy et al.’s study, which only measured CRP, sensitivity analysis using a random effects model revealed that the effect of melatonin remained non-significant (SMD −0.11, p=0.79).

Table 3.

Methodological quality of included studies (risk of bias)

	Random sequence generation	Allocation concealment (selection bias)	Blinding of participants and personnel (performance bias)	Blinding of outcome assessment (detection bias)	Incomplete outcome data addressed (attrition bias)	Selective reporting (reporting bias)	Other sources of bias (e.g. bias of study design, trial stopped early)
Alamili 2014a	+	+	+	+	+	+	+
Alamili 2014b	+	+	+	+	+	+	+
Bazyar 2018	+	+	+	+	+	+	+
Cavalcante 2012	+	+	+	+	+	+	+
Celinski 2014	+	?	?	+	+	+	+
Chojnacki 2011	+	+	+	+	+	+	+
Cichoz-Lach 2010	?	?	?	+	+	+	+
Cobo-Vasquez 2014	+	+	+	+	+	+	+
Dwaich10 2016	+	+	+	+	+	+	+
Dwaich20 2016	+	+	+	+	+	+	+
El-Gendy 2018	−	?	−	+	+	+	+
El-Sharkawy 2019	+	+	+	+	+	+	+
Forrest 2007	+	+	+	+	+	+	+
Ghaderi 2018	+	+	+	+	+	+	+
Gitto 2004a	?	?	?	+	+	+	+
Gitto 2004b	+	?	?	+	+	+	+
Gitto 2005	+	?	?	+	+	+	+
Gitto 2012	+	?	?	+	+	+	+
Hernandez-Valezquez 2016	+	+	+	+	+	+	+
Javanmard 2016	+	+	+	+	+	+	+
Kucukakin 2010a	+	+	+	+	+	+	+
Kucukakin 2010b	+	+	+	+	+	+	+
Mesri Alamdari 2014	+	+	+	+	+	+	+
Montero 2017	+	+	+	+	+	+	+
Ochoa 2011	?	+	+	+	+	+	+
Pakravan 2017	+	+	+	+	+	+	+
Raygan 2017	+	+	+	+	+	+	+
Rodanelli 2019	+	+	+	+	+	+	+
Sanchez-Lopez 2018	+	+	+	+	+	+	+
Shafiei 2018	+	+	+	+	+	+	+
Taghavi Ardakani 2018	+	+	+	+	+	+	+

4. Discussion

This systematic review and meta-analysis of 31 clinical trials involving 1517 participants in diverse populations of different ages and health conditions demonstrated significant anti-inflammatory effects of melatonin.

Melatonin had a large anti-inflammatory effect on IL-1, IL-6, and IL-8, which remained large and statistically significant even after trimming outlier studies with large effect sizes to eliminate publication bias. Usually, SMD of 0.8 is considered a large effect size (Faraone, 2008; Zlowodzki et al., 2007), and in this meta-analysis, all the SMDs for IL-1, IL-6, and IL-8 were above 1 before and after excluding outlier studies. The SMD for TNF was also above 1 but became non-significant after excluding outlier studies with large effect sizes. Although a direct comparison with other putatively anti-inflammatory agents is difficult, based on existing meta-analyses, melatonin generally had an effect that is comparable to that reported for statins (SMDs ranging from 1.3 to 4.3 for various inflammatory markers) and a larger effect than that reported for curcumin (SMDs ranging from non-significant to 2.1 for various inflammatory markers), fish oil (SMDs ranging from non-significant to 1.2 for various inflammatory markers), probiotics (SMDs ranging from non-significant to 0.5 for various inflammatory markers), and vitamin D (SMDs ranging from non-significant to 0.5 for various inflammatory markers) (Li et al., 2018; Lv et al., 2015; Milajerdi et al., 2020; Mousa et al., 2018; Tabrizi et al., 2019; Xin et al., 2012).

A previous meta-analysis on the anti-inflammatory effects of melatonin only included 6 clinical trials, with findings limited to a total of 316 patients who had metabolic syndrome or related disorders, and found significant effects on IL-6 and CRP but not on TNF (Akbari et al., 2018). Comparatively, the current meta-analysis included 31 clinical trials and involved 1517 participants of different age groups from neonates to older adults and of different health status including healthy volunteers and patients with acute and chronic diseases. The current meta-analysis also tested more inflammatory markers including IL-1, IL-6, IL-8, TNF, and CRP. Furthermore, it included not only usual therapeutic trials but also experimental trials in which melatonin was administered to healthy volunteers or patients prior to an inflammatory challenge such as endotoxin and surgical procedures to prevent inflammation.

This meta-analysis has strengths such as the large number of the included individual studies, the comprehensive list of inflammatory markers, and the diversity of study populations and study designs (i.e., treatment vs. prevention trial), all of which increase the external validity of the findings. However, there are some limitations as well. First, there was significant publication bias. To address this limitation of the literature, we conducted sensitivity analyses after excluding outlier studies with large effect sizes and demonstrated no significant publication bias once outlier effects were removed. These sensitivity analyses yielded more conservative but still robust and statistically significant results. Second, perhaps due to the diversity of the study populations and study designs, this meta-analysis had high heterogeneity. Meta-regression revealed that route of melatonin administration, study design (treatment vs. prevention), and especially mean age contributed to the heterogeneity of some results. Third, the size of the included individual studies was generally small with only one study involving more than 100 participants. Fourth, the quality of the included individual studies varied, and some studies with large effect sizes did not provide sufficient information for a complete quality assessment. Fourth, while most studies used venous blood specimens, a few of the studies used exhaled breath or salivary samples.

Exogenous melatonin reduced levels of inflammatory markers and may be useful for prevention and adjuvant treatment of inflammatory disorders. Melatonin is safe with few side effects, which makes it an excellent agent for prevention of inflammatory disorders. Notably, the current meta-analysis included 11 prevention trials of melatonin, although their duration was rather short (no longer than 3 days). Because chronic inflammation increases with aging and inflammation plays a role in the etiology of numerous diseases that affect older populations, melatonin has the potential to be widely used particularly in older adults. Furthermore, because older adults tend to be on polypharmacy due to medical comorbidities, melatonin may be especially useful given its low potential for drug-drug interactions. Additionally, there is meta-analytic evidence on the small but significant effect of melatonin in improving primary sleep disorders (Auld et al., 2017; Brzezinski et al., 2005; Buscemi et al., 2005; Ferracioli-Oda et al., 2013). Also, sleep disturbance itself increases systemic inflammation (Irwin et al., 2016). Thus, individuals with chronic inflammatory disorders and comorbid insomnia and older adults with insomnia may particularly benefit from melatonin. Larger randomized controlled trials of longer duration for the both treatment and prevention purposes are certainly warranted.

Highlights.

• This meta-analysis demonstrates anti-inflammatory effects of melatonin.

• Exogenous melatonin reduces levels of IL-1, IL-6, IL-8, and TNF.

• Melatonin, with few side effects, may be useful for inflammatory disease prevention.

Appendix 1. Search Term Strategy

1. Databases to use: PubMed, PsychInfo, Cochrane Library, Scopus, Google Scholar, EMBASE

2. Searches

   1. (“inflammation” [MeSH]) OR “inflammation” OR “inflammatory” OR “proinflammatory” OR (“C-reactive protein” [MeSH]) OR “C-reactive protein” OR “C-RP” OR “CRP” OR (“Interferons” [MeSH]) OR “Interferons” OR “Interferon” OR (“Interleukin-6” [MeSH]) OR “Interleukin-6” OR “IL-6” OR “IL6” OR “Interleukin 6” OR (“Tumor Necrosis Factor-alpha” [MeSH]) OR “Tumor Necrosis Factor-alpha” OR “Tumor necrosis factor alpha” OR “TNF-alpha” OR “Tumor Necrosis Factor-α” OR “TNF-α” OR “Tumor necrosis factor” OR “TNF” OR “TNF alpha” OR “Tumor Necrosis Factor α” OR “TNF α” OR (“Interleukin-8” [MeSH]) OR “Interleukin-8” OR “Interleukin 8” OR “IL-8” OR “IL8” OR (“Interleukin-10” [MeSH]) OR “Interleukin-10” OR “Interleukin 10” OR “IL10” OR “IL-10” OR (“Interleukin-1” [MeSH]) OR “Interleukin-1” OR “Interleukin 1” OR “IL-1” OR “IL1” OR (“NF-kappa B” [MeSH]) OR “NF-kappa B” OR “NF-κ B” OR (“Endotoxins” [MeSH]) OR “Endotoxins” OR “Endotoxin” OR (“Lipopolysaccharides” [MeSH]) OR “Lipopolysaccharides” OR “Lipopolysaccharide” OR “LPS” OR “NF kappa B” OR “NF κ B”
   2. “(Melatonin [MeSH])” OR “Melatonin” OR “N-acetyl-5-methoxy tryptamine”
   3. (a.) AND (b.)

Key word search into PsychInfo and Scopus:

((inflammation OR inflammatory OR proinflammatory OR C-reactive protein OR C-RP OR CRP OR Interferons OR interferon OR Interleukin-6 OR IL-6 OR IL6 OR Interleukin 6 OR Tumor Necrosis Factor-alpha OR Tumor necrosis factor alpha OR TNF-alpha OR Tumor Necrosis Factor-α OR TNF-α OR Tumor necrosis factor OR TNF OR TNF alpha OR Tumor Necrosis Factor α OR TNF α OR Interluekin-8 OR Interleukin 8 OR IL-8 OR IL8 OR Interleukin-10 OR Interleukin 10 OR IL10 OR IL-10 OR Interleukin-1 OR Interleukin 1 OR IL-1 OR IL1 OR NF-kappa B OR NF-κ B OR NF kappa B OR NF κ B OR Endotoxins OR Endotoxin OR Lipopolysaccharides OR Lipopolysaccharide) AND (melatonin OR N-acetyl-5-methoxy tryptamine))