Vitamin D and Periodontitis: A Systematic Review and Meta-Analysis

To explore the vitamin D levels of periodontitis patients in comparison with periodontally healthy ones, and to assess the influence of vitamin D supplementation as an adjunctive during nonsurgical periodontal treatment (NSPT). Five databases (Pubmed, Embase, Scholar, Web of Sciences, and Cochrane Library) were searched until May 2020. Mean difference (MD) meta-analysis with corresponding 95% confidence interval (95% CI) and sensitivity tests via meta-regression were used. We followed Strength of Recommendation Taxonomy (SORT) to appraise the strength and quality of the evidence. Sixteen articles were included, fourteen case-control and two intervention studies, all reporting 25-hydroxyvitamin D (25(OH)D) levels. Compared with the healthy controls, the circulating 25(OH)D levels were significantly lower in chronic periodontitis patients (pooled MD = −6.80, 95% CI: −10.59 to −3.02). Subgroup analysis revealed differences among 25(OH)D measurements, with liquid chromatography-mass spectrometry being the most homogeneous method (pooled MD = −2.05, 95% CI: −3.40 to −0.71). Salivary levels of 25(OH)D showed no differences between groups. Due to the low number of studies, conclusions on aggressive periodontitis and in the effect of vitamin D supplementation after NSPT were not possible to ascribe. Compared with healthy controls, 25(OH)D serum levels are significantly lower in chronic periodontitis patients, with an overall SORT A recommendation. Future studies are needed to clarify the effect of vitamin D supplementation and the biological mechanisms linking vitamin D to the periodontium.


Introduction
Vitamin D is a fat-soluble hormone primarily obtained from exposure to sunlight, and additionally from the diet and nutritional supplements [1][2][3]. Vitamin D is a universal term employed to describe the compound that exhibits the biological activity of cholecalciferol in animals (vitamin D3). This vitamin is a key factor to the calcium-phosphate homeostasis regulation and mineral bone metabolism [4,5]. In this sense, vitamin D increases the intestinal absorption of calcium and decreases the secretion of parathyroid hormone, which consequently decreases systemic bone resorption [6,7]. In addition, vitamin D stimulates osteoblastic bone production and alkaline phosphatase activity, optimizes bone remodeling and covers bone mass by increasing bone matrix proteins [2,3,8,9].  Inclusion criteria were determined as follows: • Design: Intervention trials (randomized clinical trials (RCTs) and non-randomized studies of interventions (NRSI)) and observational studies (case-control, cohort studies); • Studies in humans reporting 25(OH)D levels in patients with and without periodontitis; The search was conducted without any restrictions regarding year of publication or language.

Search Strategy
We searched Pubmed, MEDLINE (Medical Literature Analysis and Retrieval System Online), CENTRAL (The Cochrane Central Register of Controlled Trials), EMBASE (The Excerpta Medica Database), Web of Science from the earliest data available until May 2020. We merged keywords and the Medical Subject Headings (MeSHs) regarding the periodontal disease (periodontitis OR gingivitis OR periodontal health OR periodontal diseases [MeSH]) and Vitamin D (vitamin D OR (vitamin D [MeSH]) OR 25-hydroxy-vitamin OR calcitriol OR vitamin D supplementation OR Vitamin D deficiency OR vitamin D receptor) in accordance with the thesaurus of each database. Grey literature was searched through the OpenGrey portal [48]. Additional appropriate literature was included after a manual search of the reference lists of the final included articles. Periodontology-and nutrition-specific journals were hand-searched to identify additional articles.

Study Selection and Data Extraction
Two researchers (S.L. and V.M.) independently selected the relevant articles through titles and abstracts and excluded unrelated studies. A third author (J.B.) checked the eligible studies and any disagreement was resolved through discussion. If there were multiple publications for the same study, data from the largest sample were used.
Two researchers (S.L. and V.M.) independently extracted the relevant data from the studies. Any disagreement was resolved through discussion with a third researcher (J.B). A predefined table was used to extract necessary data from each eligible study, including the first author's name, publication year, the country where the study was conducted, exclusion criteria, number of participants, gender, mean age, percentage of smokers, periodontal case definition, sample type (saliva or serum), measurement of 25(OH)D levels and laboratory analysis. Clinical periodontal measures included PPD, CAL and BoP. All data were independently extracted by two reviewers with a consensus in all aspects. The authors were contacted for additional data clarification when necessary.

Risk of Bias (RoB) in Individual Studies
Methodological quality assessment was independently performed by two calibrated authors (V.M. and J.B.) using the Cochrane risk-of-bias tool 2 (RoB2) for RCTs [46] or ROBINS-I tool for NRSI [49]. For case-control and cohort studies, we used the Newcastle-Ottawa Scale (NOS). Regarding this last tool, we scored across three categories: studies with 7-9 stars were deemed of low RoB, studies with 5-6 stars of moderate RoB, whilst studies with less than 5 stars were deemed of high RoB. Any doubt was resolved by discussion with a third author.

Statistical Analysis
Statistical analysis was performed using R version 4.0.0 (R Studio Team 2018). For continuous data, mean values and standard deviations (SD) were used and analyzed with mean differences (MD) and correspondent estimates by 95% confidence interval (95% CI). The unit of measurement used in the MD meta-analysis was ng/mL. In the case of median and interquartile range report, we converted to mean and SD following Hozo et al. [50]. DerSimonian-Laird random-effects meta-analysis [51] and forest plots were performed using the 'meta' package [52]. Statistical heterogeneity was inspected through the I 2 index and Cochrane's Q statistic (p < 0.1). The overall homogeneity was calculated through the χ2 test [53]. All tests were two-tailed, with alpha set at 0.05. Further, the weight percentage given to each study in each analysis was provided in the forest plots. Meta-regression was performed towards the influence of smoking in serum 25(OH)D levels. Publication bias was planned if at least 10 Nutrients 2020, 12, 2177 4 of 18 or more studies were included [53]. In the case of lacking data amenable to meta-analysis, we followed the Synthesis Without Meta-analysis (SWiM) guidelines to synthesize quantitative data [54].

Strength of Recommendations
We employed the Strength of Recommendation Taxonomy (SORT) to appraise the strength and quality of the evidence [55]. The outcomes of the present systematic review, clinical recommendations, and future necessary research were discussed.

Study Selection
Overall, the search yielded a total of 1531 studies ( Figure 1). After duplicate removal, 1460 were screened for titles and abstracts, and 125 articles fulfilled the inclusion criteria (1335 were excluded). These 125 articles were subjected to full paper review eligibility and 109 were excluded as they did not approach the research questions (Supplementary Material Table S2). Then, 16 articles were included for the qualitative analysis, of which two concerned Vitamin D supplementation. A total of 13 studies were included in the quantitative synthesis regarding 25(OH)D levels in patients with and without periodontitis ( Figure 1). we followed the Synthesis Without Meta-analysis (SWiM) guidelines to synthesize quantitative data [54].

Strength of Recommendations
We employed the Strength of Recommendation Taxonomy (SORT) to appraise the strength and quality of the evidence [55]. The outcomes of the present systematic review, clinical recommendations, and future necessary research were discussed.

Study Selection
Overall, the search yielded a total of 1531 studies ( Figure 1). After duplicate removal, 1460 were screened for titles and abstracts, and 125 articles fulfilled the inclusion criteria (1335 were excluded). These 125 articles were subjected to full paper review eligibility and 109 were excluded as they did not approach the research questions (Supplementary Material Table S2). Then, 16 articles were included for the qualitative analysis, of which two concerned Vitamin D supplementation. A total of 13 studies were included in the quantitative synthesis regarding 25(OH)D levels in patients with and without periodontitis ( Figure 1).
In Vitamin D supplementation on NSPT, two studies were included [63,64]. Overall, 557 patients suffering from chronic periodontitis were submitted to NSPT and 276 participants were medicated with Vitamin D3 supplements. Gao et al. [63] searched the effect of two different concentrations of Vitamin D3 supplements in a controlled design (Table 2).

Risk of Bias within Studies
Thirteen articles presented with low RoB (four with 9/9, four with 8/9 and five with 7/9 scores) ( Figure 2, Supplementary Material Table S3). Only one article presented moderate risk of bias, with an overall score of 6/9 [32]. The main reason for bias arose from the representativeness of the cases. Overall, articles adopted an adequate periodontitis case definition (100%, n = 14) and definition of control. A considerable number of studies failed to include representative samples (57.1%, n = 8) and the selection of controls (42.9%, n = 6). In the ascertainment of exposure, usability of the same method of ascertainment for cases and controls, and non-response rate, all articles presented low RoB (100.0%, n = 14). Two intervention trials presented low risk of bias, one RCT (Supplementary Material  Table S4) and one NRSI (Supplementary Material Table S5).

Vitamin D Levels and Periodontitis
Serum Levels  Table S6).
Due to the existence of only two articles [28,56] regarding the comparison of serum levels of 25(OH)D of patients diagnosed with aggressive periodontitis compared to healthy periodontal patients, the meta-analysis was not deemed possible (Table 1). While one study [28] reported lower levels of 25(OH)D serum levels, the other, based on salivary measurements, showed opposite results [56].

Salivary Levels
Regarding salivary 25(OH)D levels, our analysis did not report differences comparing chronic periodontitis to healthy periodontal subjects (MD of 0.45, 95% CI: −1.05; 1.96) (Figure 4).   Overall, articles adopted an adequate periodontitis case definition (100%, n = 14) and definition of control. A considerable number of studies failed to include representative samples (57.1%, n = 8) and the selection of controls (42.9%, n = 6). In the ascertainment of exposure, usability of the same method of ascertainment for cases and controls, and non-response rate, all articles presented low RoB (100.0%, n = 14). Two intervention trials presented low risk of bias, one RCT (Supplementary Material Table  S4) and one NRSI (Supplementary Material Table S5).  Table S3).

Vitamin D Levels and Periodontitis
Serum Levels  Table S6).
Due to the existence of only two articles [28,56] regarding the comparison of serum levels of 25(OH)D of patients diagnosed with aggressive periodontitis compared to healthy periodontal patients, the meta-analysis was not deemed possible (Table 1). While one study [28] reported lower levels of 25(OH)D serum levels, the other, based on salivary measurements, showed opposite results [56].  (Table S3).

Vitamin D Supplementation as an Adjunctive in NSPT
Due to the lack of data amenable to perform a meta-analysis about vitamin D supplementation on NSPT, we therefore decided to synthesize evidence without analysis. Two studies fulfilled the

Vitamin D Supplementation as an Adjunctive in NSPT
Due to the lack of data amenable to perform a meta-analysis about vitamin D supplementation on NSPT, we therefore decided to synthesize evidence without analysis. Two studies fulfilled the inclusion criteria, one RCT (Gao et al., 2020) and one NRSI (Perayil et al., 2015), both of low RoB (Supplementary Material Tables S4 and S5).
Gao et al. [63] applied vitamin D supplementation on 360 patients with moderate or severe periodontitis following NSPT. Patients were randomly assigned to 2000 international units (IU)/d vitamin D3, 1000 IU/d vitamin D3, or placebo. The effect of vitamin D supplementation tended to be modest with limited periodontal clinical relevance and long-term efficacy towards PPD and CAL (Gao et al., 2020).
Furthermore, Perayil et al. [64] investigated, in 82 moderate chronic periodontitis patients, if vitamin D supplementation plus calcium (Shelcal-D 500 mg calcium + 250 IU vitamin D once daily) compared to placebo following NSPT. Despite PPD and CAL having no differences between groups, the authors reported significantly better results for the vitamin D group in relation to gingival inflammation and bone density (measured using panoramic x-ray).

Additional Analyses
We confirmed that no publication bias was assessed in meta-analysis regarding the serum 25(OH)D levels (p = 0.1174) ( Figure 5).
Nutrients 2020, 12, x FOR PEER REVIEW 9 of 18 modest with limited periodontal clinical relevance and long-term efficacy towards PPD and CAL (Gao et al., 2020). Furthermore, Perayil et al. [64] investigated, in 82 moderate chronic periodontitis patients, if vitamin D supplementation plus calcium (Shelcal-D 500 mg calcium + 250 IU vitamin D once daily) compared to placebo following NSPT. Despite PPD and CAL having no differences between groups, the authors reported significantly better results for the vitamin D group in relation to gingival inflammation and bone density (measured using panoramic x-ray).

Additional Analyses
We confirmed that no publication bias was assessed in meta-analysis regarding the serum 25(OH)D levels (p = 0.1174) ( Figure 5).

Reporting on Strength of Recommendation
Using the SORT recommendation, we concluded that chronic periodontitis is strongly associated with lower serum levels of 25(OH)D (SORT A) [55].

Summary of Main Findings and Quality of the Evidence
This systematic review supported an association between serum vitamin D levels (measured in ng/mL of 25(OH)D) and chronic periodontitis, with an overall SORT A recommendation. Within the lack of the available studies, salivary levels of 25(OH)D did not present an association with chronic periodontitis. In addition, there was a scarcity of studies regarding the association of aggressive periodontitis and 25(OH)D levels, and the influence of vitamin D supplementation precluded any definitive conclusion. Analyzing the impact of smoking in serum 25(OH)D level changes in chronic periodontitis, meta-regression analysis revealed that smoking had no meaningful impact.
Overall, the results of this systematic review support a link between 25(OH)D serum levels and chronic periodontitis. That is, patients diagnosed with chronic periodontitis presented lower serum levels of 25(OH)D than periodontal healthy patients. These results are clinically relevant considering the crucial role of vitamin D in bone maintenance and in the immune system [2,3,8,65,66]. In addition, several studies have unveiled the potential harmful impact of vitamin D deficiency on the periodontium, especially after periodontal surgery where this baseline decrease might result in

Reporting on Strength of Recommendation
Using the SORT recommendation, we concluded that chronic periodontitis is strongly associated with lower serum levels of 25(OH)D (SORT A) [55].

Summary of Main Findings and Quality of the Evidence
This systematic review supported an association between serum vitamin D levels (measured in ng/mL of 25(OH)D) and chronic periodontitis, with an overall SORT A recommendation. Within the lack of the available studies, salivary levels of 25(OH)D did not present an association with chronic periodontitis. In addition, there was a scarcity of studies regarding the association of aggressive periodontitis and 25(OH)D levels, and the influence of vitamin D supplementation precluded any definitive conclusion. Analyzing the impact of smoking in serum 25(OH)D level changes in chronic periodontitis, meta-regression analysis revealed that smoking had no meaningful impact.
Overall, the results of this systematic review support a link between 25(OH)D serum levels and chronic periodontitis. That is, patients diagnosed with chronic periodontitis presented lower serum levels of 25(OH)D than periodontal healthy patients. These results are clinically relevant considering the crucial role of vitamin D in bone maintenance and in the immune system [2,3,8,65,66]. In addition, several studies have unveiled the potential harmful impact of vitamin D deficiency on the periodontium, especially after periodontal surgery where this baseline decrease might result in undesirable outcomes [67,68].
A possible mechanism for this association whereby vitamin D reduces the risk of periodontitis is through the induction of cathelicidin [69][70][71][72][73]. The vitamin D pathway has been shown to exist in human gingival fibroblasts and periodontal ligaments cells, playing an important role in immune defense in periodontal soft tissues via the activation of the human cationic antimicrobial protein cathelicidin [69][70][71][72]. Recently, serum 25(OH)D deficiency was associated with decreased hBD−2 and cathelicidin levels in periodontal tissues in gingivitis and chronic periodontitis [73].
Several limitations, however, should be reported. Firstly, the level of heterogeneity observed was high and can limit the validity and robustness of these quantitative analyses. The lack of consistency in the periodontitis case definition precluded more robust analyses of the extent and severity of periodontitis with 25(OH)D serum levels. Thus, future studies should accommodate the up-to-date consensus [74], because of its upgraded characteristics [75,76], as well providing more in-depth data on the extent and severity of periodontitis and its relevant periodontal clinical measures (such as PPD, CAL and BoP). Secondly, the main analysis in this systematic review was derived from observational studies, that only inform the association between periodontitis and 25(OH)D serum changes. Therefore, studies with longer follow-ups are mandatory to clarify this matter. Furthermore, the included studies showed multiple quantification method of 25(OH)D levels, and this may have contributed to the heterogeneity. In the future, studies should harmonize the measurement method of 25(OH)D levels. On the other hand, there are strengths in this evidence-based study. This review was designed a priori and followed a strict protocol, updated international reporting guidelines, and had an extensive literature search.
Considering the existing evidence, this is the first review to render a magnitude effect on the association between serum 25(OH)D levels and chronic periodontitis. Overall, four systematic reviews have analyzed such association [42][43][44]77], but without success in pursuing meta-analysis. Further, these reviews have reported contradictory conclusions, wherein Van der Putten et al. (2009) found no evidence of an association of vitamin D with periodontal disease in non-institutionalized elderly people, Pinto et al. [44] and Perić et al. [43] found insufficient data to provide a conclusion, while Varela-López et al. [77] reported a potential association.
With regard to periodontal treatment, our narrative synthesis provides a small view on the potential characteristics of vitamin D supplementation on NSPT. One the one hand, the shortage of literature was already highlighted, which precludes any conclusion on the effect of serum vitamin D levels on periodontal treatment [43,78]. On the other hand, a previous review found that baseline vitamin D deficiency at the time of the periodontal treatment, especially in surgical procedures, negatively affected treatment outcomes [45]. However, more randomized clinical trials are warranted to provide a robust conclusion.

Clinical and Research Implications
Public awareness of vitamin D levels is high due to the worrisome prevalence of vitamin D deficiency worldwide [3,10,11]. Nevertheless, it is important to highlight that to confirm the decreased total levels of 25(OH)D in a more complete way, studies may combine total 25(OH)D with parathyroid hormone (PTH), calcium and phosphate levels [79]. Thereafter, the results of this systematic review are clinically relevant because they link these low vitamin D levels to chronic periodontitis, an inflammatory condition that figures as one of the most prevalent disease in the world [18,80]. In other words, periodontitis patients are very likely to have lower serum levels of 25(OH)D, though the clinical consequences are still unclear in their entirety, and for this reason larger and well-designed RCTs are warranted.
Our results can also serve as clinical and research guidelines for expected differences in vitamin D levels in periodontitis patients, as well as clarifying the importance of knowing the method to measure vitamin D, as this may impact vitamin D differences. At this stage, the LC-MS measurement method showed moderate heterogeneity (I2 = 41%), regardless of the multiple periodontal diagnostic criteria observed in this subgroup; therefore, we might expect more homogeneous results when studies employ the same case definition [28,30,32]. In view of these results, LC-MS might be seen as the most consistent method to measure 25(OH)D serum levels, both in clinical practice as well as in research studies, and is in agreement with a previous reliability study on vitamin D quantification [106].
While 25(OH)D serum levels of periodontitis patients were decreased compared to healthy controls, the 25(OH)D levels in whole saliva had no significant differences, and the results of the included studies are quite heterogeneous in saliva. A reasonable explanation for such differences can be the expression of vitamin-D binding protein (DBP) locally versus systemically. In this sense, a previous study on healthy periodontal patients showed that the levels of DBP in gingival crevicular fluid were higher than serum levels [107], and therefore periodontal tissue might be another source of DBP [108]. Furthermore, the levels of DBP were found to be lower in periodontitis cases, presumably due to the lack of effective production or an increase in local consumption [107]. The results on salivary levels show some heterogeneity among the studies, in particular [33,58], as well a low number of participants, which limits the validity of this finding. A possible explanation for the discrepancy between Miricescu et al., 2014 [58] and Costantini et al., 2020 [33] is the difference between mean age, gender distribution and periodontal diagnosis between these two studies.
Notwithstanding these issues, more studies are warranted to explore the role of DBP expression on the salivary and serum levels of 25(OH)D on periodontal patients.
Mindful of the inflammation surrounding teeth, peri-implantitis is a pathological condition characterized by the progressive loss of supporting peri-implant bone [109], and strong evidence has suggested that PD is a risk factor for implant loss [110]. A single study found that 25(OH)D levels are significantly decreased in peri-implantitis patients comparing with peri-implant healthy patients and, consequently, might be important indicators of peri-implant diseases [111]. Nevertheless, further studies are needed to confirm if such an association with vitamin D follows the same fashion as PD.
Importantly, future studies on the impact of vitamin D supplementation should bear in mind the baseline status of the patients. Currently, vitamin D supplementation studies support the arbitrary application of vitamin D supplements as an adjunct in NSPT, though studies have seemed to disregard the baseline vitamin D levels, and this might lead to inevitable bias of analysis. In this sense, future studies should define a priori which patients have vitamin D deficiency (<20 ng/mL) or insufficiency (<30 ng/mL) to clarify whether the restoration of DV levels will result in superior periodontal clinical results. Moreover, studies must consider the initial periodontal status and the interplay of 25(OH)D with key periodontal clinical measures. Therefore, intervention studies using vitamin D supplementation should define with a clear threshold alike patients according to the baseline 25(OH)D levels and the periodontal status.

Conclusions
Periodontitis is associated with lower 25(OH)D serum levels. The effect of vitamin D supplementation as an adjunct of nonsurgical periodontal treatment remains unclear due to the shortage of available studies. Future studies are needed regarding the effect of vitamin D supplementation and the biological mechanisms linking vitamin D to the periodontium.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/12/8/2177/s1, Table S1. PRISMA 2009 Checklist; Table S2. List of potentially relevant studies not included in the systematic review, along with the reasons for exclusion; Table S3. Newcastle-Ottawa Scale; Table S4. RoB2 Tool; Table S5. ROBINS-I Tool; Table S6. Summary of estimates of meta-regression to assess the influence of smoking on 25(OH)D serum level.