Role of Melatonin in Bone Remodeling around Titanium Dental Implants: Meta-Analysis

The theory, known as the “brain-bone axis” theory, involves the central nervous system in bone remodeling. The alteration of the nervous system could lead to abnormal bone remodeling. Melatonin produced by the pineal gland is a hormone that is characterized by its antioxidant properties. The aim of this meta-analysis was to examine the role of melatonin in the growth of new bone around titanium dental implants in vivo. A manual search of the PubMed and Web of Science databases was conducted to identify scientific studies published until November 2020. We included randomized clinical trials (RCTs) and animal studies where melatonin was used with titanium implants. Fourteen studies met the inclusion criteria. Quality was assessed using the Jadad scale and SYRCLE’s risk of bias tool. Our meta-analysis revealed that the use of melatonin during implant placement improves bone-to-implant contact percentages in animals (difference of means, random effects: 9.59 [95% CI: 5.53–13.65]), reducing crestal bone loss in humans (difference in means, random effects: −0.55 [95% CI: 1.10–0.00]). In animals, titanium implants using melatonin increase bone-to-implant contact surface 2–6 weeks after their placement and reduce crestal bone loss in humans following six months. The results of this meta-analysis should be taken with caution, due to the small samples and the large heterogeneity among studies.


Introduction
The first description of osseointegration was provided by Brånemark and colleagues [1] more than 50 years ago, and to date, this process still remains unexplored. One of the theories posed in recent years, referred to as the "brain-bone axis theory" by certain authors [2], has drawn particular interest. It suggests the involvement of the sympathetic nervous system (SNS) in bone remodeling, claiming the need for the autonomic nervous system to be undamaged in order to contribute to the maintenance of healthy bone tissue, with its alteration leading to possible anomalies in bone remodeling [3,4]. This remodeling process would be mainly controlled by neurotransmitters (noradrenaline, serotonin and dopamine), and growth hormones secreted by the pituitary gland could stimulate osteoblast and osteoclast proliferation, which plays a crucial part in the bone formationdestruction balance [5,6].
It has been proven that the group of glucose-sensing neurons in the hypothalamic arcuate nucleus makes control of the skeleton by the brain possible [7], and that long-term use of certain central nervous depressant drugs causes a reduction in bone mass that results in osteoporosis, and therefore in high rates of dental implant failure in patients under treatment with such drugs [8].
The biofunctionalization of a certain biomaterial consists of modifying the physicochemical characteristics of its surface, which improves a body's biological response when it comes into contact with it, despite the fact that there are other factors that influence

Protocol
The Population, Intervention, Comparison and Outcome framework (PICO) was used as a basis to formulate the research question, which was: "The inclusion of melatonin in dental implant surgery: does it influence osseointegration?".

Data Sources and Search Strategy
The PubMed and Web of Science electronic databases were searched for findings published in the last 15 years until November 2020. The search terms used were: "titanium dental implants AND melatonin surface"; "melatonin AND dental implants"; and "melatonin AND dental implants AND bone formation". The Boolean operator "AND" was used to combine the searches.

Inclusion and Exclusion Criteria
The inclusion criteria for the study selection were: • Randomized clinical trials and animal studies on Ti implants, with and without the incorporation of melatonin. • Randomized clinical trials and animal studies that reported bone-implant contact percentages, with and without the incorporation of melatonin.

•
Studies with a minimum of six implants/group.
The exclusion criteria for the study selection were: • In vitro studies. • Narrative and systematic reviews. • Clinical cases. • Studies that assessed the effectiveness of melatonin in bone regeneration without including dental implants, duplicates and those that failed to meet the inclusion criteria.

Data Extraction and Analysis
Those articles that failed to address the research question were removed. The correponding titles and abstracts of the eligible articles were taken, and two reviewers (NL-V and AL-V) separately drew up a selection of them. The reviewers discussed and solved the discrepancies over the choice of articles that arose. The full versions of the chosen articles were then obtained for review and inclusion.

Risk of Bias (RoB) of the Selected Articles
The methodology of the scientific evidence gathered in the selected studies was assessed using SYRCLE's risk of bias tool (an adapted version of the Cochrane RoB tool, with specific biases in animal studies) [25].

Quality of the Reports of the Selected Articles
This assessment was based on the provided ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines [26], with a total of 23 items. The reviewers, N.L.-V. and A.L.-V., allocated each item a score of 0 (not reported) or 1 (reported), including an overall inventory of all the selected studies.

Quality of the Reports of the Included Randomized Clinical Trials
The assessment was carried out using the Jadad scale [27], which reveals the methodological quality of a study based on how it describes randomization, blinding and dropouts (withdrawals). The scale goes from 0 to 5, with a score of ≤2 meaning poor quality and a score of ≥3 meaning that the report meets high quality standards.

Statistical Analysis
The meta-analysis was conducted using the RevMan 5 program (Review Manager (RevMan) [Computer program]. Version 5.4. Copenhagen, Denmark: The Nordic Cochrane Centre, The Cochrane Collaboration, London, UK, 2014). Animal studies were assessed for BIC [28][29][30] between 2 and 6 weeks after placement, and crestal bone loss or MBL of implants was assessed in RCTs 6 months after placement [31,32]. Mean difference (MD) and standard deviation (SD) were used for the assessment of continuous variables (BIC and crestal bone loss), weighting by inverse variance with a 95% confidence interval (CI). The threshold for statistical significance was established at p < 0.05. Heterogeneity was assessed by calculating I2 and Chi-square, using a random effects model in both cases. The sensitivity analysis was conducted by excluding one study at a time to check whether there were changes in the results. A funnel plot graph was used to assess publication bias.

Characteristics of the Studies
Until November 2020, a total of 135 studies were identified for subsequent assessment by the reviewers. After a first screening, 41 duplicate studies were removed. In a second screening, 26 studies that did not clearly meet the inclusion criteria, and were therefore considered inadequate, were removed (Figure 1, Flowchart). Table 1 provide the evaluation of the ARRIVE criteria in animal studies. Tables 2 and 3 provide a general description of the details corresponding to the RCTs and experimental animal studies, respectively. Table 4 provides the Jadad quality score in RCTs.  Mode Value: 16.36 ± 0.88 (Mean value and standard deviation); Each item was allocated a score of "0" (not reported) or "1" (reported). The total score of each of the included studies was also recorded.  1. Title

Risk of Bias and Quality Assessment of Animals Included Studies
The risk of bias assessment results for the animal studies are shown in Figure 2. Although allocation to blinding was mentioned in several articles, the lack of information resulted in a high and unclear risk of bias for most of the items. The ARRIVE checklist criteria for the animal studies [26] included are shown in Table 1. The mean score of the studies was 16.36 ± 0.88. All the studies provided adequate information in terms of title, abstract, introduction, ethical declaration, species, surgical procedure, assessment of results and statistical analysis. None of the studies reported items 5 (Reasons for animal models), 19 (3Rs, Replace, Reduce and Refine), 20 (Adverse events), 21 (Study limitations) or 22 (Generalization/applicability). The topical application of melatonin could be a good treatment option for dental implants in the posterior maxilla.

Methodological Quality of the Included Randomized Studies
The mean score of the studies was 3.3 ± 1.2. Two of the included studies [44,46] scored ≥3; the study by Hazzaa et al. [45] obtained the lowest score (Table 4).

Methodological Quality of the Included Randomized Studies
The mean score of the studies was 3.3 ± 1.2. Two of the included studies [44,46] scored ≥3; the study by Hazzaa et al. [45] obtained the lowest score (Table 4). Each study was assigned a score of 0-5; Mode value: 3.3 ± 1.2 (Mean value and standard deviation).
Three animal studies were excluded from the quantitative analysis: that by Palin et al. [33], as it did not measure bone-implant contact; the study by Muñoz et al. [38], which combined melatonin with growth hormone; and the study conducted by Calvo-Guirado et al. in 2010 [40], where melatonin was combined with pig bone.
Heterogeneity in the results was very high (I 2 = 96%; Chi-square = 193.87; 95% CI), which is why a random effects model was chosen, assuming that the differences among studies were due to heterogeneity, and that the effect of small studies on the result of the meta-analysis was relevant. The results of the sensitivity analysis did not suggest the exclusion of any study to be the cause for heterogeneity, the latter being always above 90%; however, it appeared to be the cause for significant changes in the direction or size of the effect. Since in this case the large effect size suggested a positive result (higher percentage of bone-implant contact), the forest plot's labels were inverted.
The study of the forest plot ( Figure 3) revealed no significant differences between the two groups (melatonin vs. placebo) in the studies of Guardia et al. [39], Calvo-Guirado et al. en 2015 [36] and Salomó-Coll et al. [34], since confidence intervals at 95% overlap and cross the line of no effect. In the remaining studies [33,35,37,38,[40][41][42][43], the experimental group (melatonin) achieves significantly better results than the placebo group. There were also no noticeable differences among studies, since the confidence intervals of the selected studies overlap, except in the case of Takechi et al. [43], where results were more favorable to melatonin than in the other studies.
Coatings 2021, 11, x FOR PEER REVIEW 9 of 14 cross the line of no effect. In the remaining studies [33,35,37,38,[40][41][42][43], the experimental group (melatonin) achieves significantly better results than the placebo group. There were also no noticeable differences among studies, since the confidence intervals of the selected studies overlap, except in the case of Takechi et al. [43], where results were more favorable to melatonin than in the other studies.
The meta-analysis revealed that treatment with melatonin is associated with greater contact between the implant's surface and the bone in the assessment carried out between 2 and 6 months after implant placement (difference in means, random effects: 9.59 [95% CI: 5.53-13.65]). As regards RCTs, three studies [44][45][46] assessed crestal bone loss 6 months after placement were selected.
Heterogeneity was high (I 2 = 95%; Chi-square = 41, 50; 95% CI), so a random effects model was chosen, assuming that the differences among studies were due to heterogeneity rather than chance. The study of the forest plot (Figure 4) revealed that the difference between the two study groups (melatonin vs. placebo) was not significant in the study by El-Gammal et al. [46], while in the two studies by Hazzaa et al. [44,45], the experimental group (melatonin) achieved better results than the placebo group. The meta-analysis revealed that treatment with melatonin is associated with greater contact between the implant's surface and the bone in the assessment carried out between 2 and 6 months after implant placement (difference in means, random effects: 9.59 [95% CI: 5.53-13.65]).
Heterogeneity was high (I 2 = 95%; Chi-square = 41, 50; 95% CI), so a random effects model was chosen, assuming that the differences among studies were due to heterogeneity rather than chance. The study of the forest plot ( Figure 4) revealed that the difference between the two study groups (melatonin vs. placebo) was not significant in the study by El-Gammal et al. [46], while in the two studies by Hazzaa et al. [44,45], the experimental group (melatonin) achieved better results than the placebo group. placement were selected.
Heterogeneity was high (I 2 = 95%; Chi-square = 41, 50; 95% CI), so a random effects model was chosen, assuming that the differences among studies were due to heterogeneity rather than chance. The study of the forest plot ( Figure 4) revealed that the difference between the two study groups (melatonin vs. placebo) was not significant in the study by El-Gammal et al. [46], while in the two studies by Hazzaa et al. [44,45], the experimental group (melatonin) achieved better results than the placebo group.
The meta-analysis also proved that, after 6 months, the implants placed in the experimental group (melatonin) presented less marginal bone loss than those placed in the control group (difference in means, random effects: −0.55 [95% CI: 1.10-0.00]). The load increase in the study by El-Gammal et al. [46] led to a widening of the confidence interval for the overall effect size. Nevertheless, because of the small size of the sample and the large heterogeneity among studies, the results of this meta-analysis should be taken with caution.  The meta-analysis also proved that, after 6 months, the implants placed in the experimental group (melatonin) presented less marginal bone loss than those placed in the control group (difference in means, random effects: −0.55 [95% CI: 1.10-0.00]). The load increase in the study by El-Gammal et al. [46] led to a widening of the confidence interval for the overall effect size. Nevertheless, because of the small size of the sample and the large heterogeneity among studies, the results of this meta-analysis should be taken with caution.

Publication Bias and Heterogeneity
The experimental studies show graphical signs of publication bias, as can be observed in the Funnel Plot ( Figure 5).

Publication Bias and Heterogeneity
The experimental studies show graphical signs of publication bias, as can be observed in the Funnel Plot ( Figure 5).

Discussion
The purpose of this study was to explore the role of MT in bone growth and remodeling around Ti dental implants, both in RCTs and in experimental animal trials.
Bone remodeling involves hormones, cytokines, growth factors and other molecules [47], with MT being one of the hormones that modulates bone formation and absorption.
Certain studies have reported that MT stimulates the osteogenic activity of bone tissue, increasing human osteoblast differentiation in vitro and inducing the formation of cortical bone in mice in [48].
The relationship between MT and bone metabolism has been demonstrated in several

Discussion
The purpose of this study was to explore the role of MT in bone growth and remodeling around Ti dental implants, both in RCTs and in experimental animal trials.
Bone remodeling involves hormones, cytokines, growth factors and other molecules [47], with MT being one of the hormones that modulates bone formation and absorption.
Certain studies have reported that MT stimulates the osteogenic activity of bone tissue, increasing human osteoblast differentiation in vitro and inducing the formation of cortical bone in mice in [48].
The relationship between MT and bone metabolism has been demonstrated in several studies [37,41,45], with evidence of its effect as a precursor of bone cells in rat bone marrow [49]. Koyama and colleagues [50] were the first to prove that the administration of MT in young rats during their growth period increased spongy bone mass and inhibited bone resorption. If such findings were to be confirmed using adult animals with no endogenous MT, they could be useful to explain the concept of osseointegration. Satomura and colleagues proved that MT accelerates osteoblastic differentiation in humans and rodents, suggesting its possible application as pharmaceutical agent to promote bone regeneration [51]. Nevertheless, in an in vitro study on the effect of melatonin on adipogenesis and osteogenesis in human mesenchymal stem cells, Zhang and colleagues [52] reported an early increase in adhesion and proliferation but found no differences in extended culture periods.
The experimental studies included in our meta-analysis [34][35][36][37]39,[41][42][43] show the beneficial effects of MT as regards bone regeneration around Ti dental implants, be it topically applied on implant beds [34,37,39,41,42], coating the implant [35,36], or injected around the implants at the time of placement [43]. However, although certain studies reported a reduction in osteoclastogenesis [53] when topically applied to the alveolar socket after extraction prior to surgery [23], Cobo-Vázquez and colleagues [54], in a pilot study using a sample of 10 patients, found no differences regarding bone density when MT was applied to post-extraction alveolar sockets of retained mandibular third molars. According to our systematic review, Tresguerres and colleagues [37] presented the most thorough histological results, reporting changes in the cortical and medullary regions, and a larger amount of trabecular tissue in contact with the implants in the group treated with MT. The remaining studies only reported histomorphometric results [34][35][36]39,40,42,43].
Current studies regard the skeleton as a true endocrine organ controlled by the hypothalamus [55]. Protein degradation mediated by the ubiquitin-proteasome pathway is essential to regulate the balance between bone formation and bone resorption through certain signal transduction pathways, which regulate the activity of mature osteoblasts and osteoclasts [56].
Apart from contributing to synchronize biological rhythms and its antioxidant and inflammatory effects, among others, MT has immunomodulatory effects and induces apoptosis [57]. However, the role of MT in the formation of new bone is not fully defined, its reduction being regarded as proportional to skeletal maturation and with contraindications concerning its function, such as the fact that certain individuals with different defects in osteoblast function are at a greater risk for osteosarcoma. Conversely, MT can improve the normal functions of osteoblasts, and would therefore play a protective role against bone cancer [58,59].
Because of the ethical implications associated with histopathological examinations, the RCTs included in our meta-analysis exclusively reported macroscopic and radiologic results; nevertheless, the reduced number of included studies prevented an adequate and conclusive meta-analysis. Moreover, the study by Hazzaa and colleagues [45] presented certain limitations in its design and in how it was conducted, such as group randomization, implant location, reason for removal, demographic characteristics of the participants and the degree to which they balanced between groups that reduce the reliability of its results [60]. Nonetheless, our meta-analysis found that MT stimulates the formation of new bone and increases bone density around Ti dental implants, although it presented serious limitations, mainly associated with the heterogeneity among the selected studies and the scarcity of RCTs. On the other hand, there were significant discrepancies among animal studies regarding measurement of the parameters of bone surface in contact with the implant, as had been observed in the radiologic measurements in the RCTs. There were also major differences regarding the amounts, preparation, forms of application, concentrations and application timing of MT. Another important limitation was that none of the included studies considered factors such as bone quality or the surgical technique used-these factors would provide biases in obtaining results.
The study by Takechi and colleagues [43], which is consistent with that of Satomura and colleagues [51], was the only one where melatonin was used systemically (intraperitoneal, 100 mg/Kg weight) until the animals were sacrificed (4 weeks after implant placement): this form of administration and dosage conflicts with those used by the other authors included in the review [33][34][35][36][37][38][39][40][41][42][43]. The systemic administration of MT requires large doses of the drug, which increases the possibility of side effects; therefore, the topical application of MT is preferred over its systemic administration [61]. This same discrepancy in form of administration and dosage could be observed in the RCTs, while Hazzaa and colleagues and El-Gammal and colleagues [44,46] used MT in the form of topical gel, in doses of 1.2 mg., and Hazzaa and colleagues [45] used it in powder form, not specifying the dosage. Hence, there is no consensus as to the best route of administration for this molecule, and the dosage required to achieve the desired effect.

Conclusions
Bearing in mind the limitations of most of the studies, all of those included in this meta-analysis reported that the topical application of MT on the ostectomy site at the time of implant placement can induce greater bone-to-implant contact, as well as greater bone mass and density around Ti dental implants, especially in the earliest stages of healing, thus favoring osseointegration. Nevertheless, to clearly confirm such affirmations, further research using broader, well-designed samples with long-term monitoring and standardized protocols for application, MT dosage and assessment of bone parameters is required; all with the purpose of ensuring predictable and reliable outcomes.

Conflicts of Interest:
The authors declare no conflict of interest.