Background
Chronic periodontitis (CP) is a multifactorial inflammatory disease caused by pathogenic microorganisms and disordered host immune inflammation that leads to bone resorption, bony defects, and ultimately tooth loss [
1]. Nonsurgical periodontal treatment reduces pocket depth (PD) and increases clinical attachment level (CAL) to some extent [
2,
3], but fails to fill the bony defect [
4]. Thus, various adjuvant therapies have been applied in nonsurgical treatment to reduce tissue destruction and to enhance periodontal reparative processes including statins.
Statins are inhibitors of 3-hydroxy-3-glutaryl-coenzyme A reductase and are primarily used to prevent hyperlipidemia and coronary artery disease [
5,
6]. However, with in-depth study of statins, additional benefits have been found in the treatment of periodontal diseases. This phenomenon may be due to the unique properties of statins that limit the pathogenesis of periodontitis, such as anti-inflammatory [
7,
8], anti-microbial [
9], bone formation promoting, bone loss inhibiting [
10,
11] and antioxidant properties [
12]. Different statins exhibit different such properties, which could lead to different treatment outcomes. For example, rosuvastatin (RSV) is thought to possess stronger anti-inflammatory potential than atorvastatin (ATV) [
13], while ATV is stronger than simvastatin (SMV) in terms of anti-inflammatory and antioxidant potential [
14,
15]. SMV is considered to be the optimal statin for controlling periodontal pathogens, such as
Porphyromonas gingivalis (
Pg) and
Aggregatibacter actinomycetemcomitans (
Aa) [
16]. However, clinical trials investigating the effects of different statins on adjuvant treatment of CP are limited. To our knowledge, there are six meta-analyses comparing statins adjunctive to scaling and root planing (SRP) with SRP alone, however, they fail to measure the relative effects of various statins on CP without other systemic diseases [
4,
17‐
21]. Therefore, a network meta-analysis which compares and ranks different statins should be beneficial to clinical practice.
This network meta-analysis aimed to study whether local statins applied adjunctively to nonsurgical periodontal treatment contribute to better clinical and histological periodontal outcomes based on randomized controlled clinical trials (RCTs) when compared to periodontal treatment alone in patients with CP. This study further ranked statins based on their adjunct efficacy with SRP.
Methods
Protocol registration
This meta-analysis was prospectively registered at the National Institute for Health Research PROSPERO, International Prospective Register of Systematic Reviews (
http://www.crd.york.ac.uk/PROSPERO, registration no.: CRD42018100753).
Inclusion criteria
Only RCTs followed up for at least 6 months were included in this network meta-analysis. PICO criteria was defined as [
22]:
-
(P) Participants: Patients with chronic periodontitis without periodontal therapy and use of antibiotics in the past 6 months.
-
(I) Interventions: The following locally delivered statins employed adjunctively to periodontal treatment were considered: SRP + ATV, SRP + SMV, and SRP + RSV
-
(C) Comparison: SRP alone
-
(O) Outcome measures: primary outcome: changes in IBD; secondary outcomes: changes in PD and CAL
Exclusion criteria
Studies that had any of the following characteristics were excluded: (a) split-mouth RCT design; (b) inclusion patients with statin allergy; (c) application of systemic statin therapy; (d) inclusion of immunocompromised individuals; (e) inclusion of former smokers; (f) systemic diseases except for type 2 diabetes.
Search methods for study identification
To identify RCTs for this network meta-analysis, we searched the Pubmed, Embase, Cochrane Library, and Web of Science databases for relevant publications published up to June 2018. The following MeSH terms/free terms and their combinations searched are described in Additional file
1. The resulting reference lists of relevant articles and relevant systematic reviews [
4,
17‐
20] was manually screened to find other potentially eligible studies.
Two researchers (R.Y.Cao & Q.L.Li) independently screened the databases search for relevant titles and abstracts. Then data was extracted and recorded relevant information from eligible studies with pre-designed data-extraction forms using the following criteria: surname of the first author, publication year, country, characteristics of participants (age, gender, smoking status, systemic diseases), sample size, type of interventions, number of application sites/patients, application mode/site, application period, periodontal probe, outcome (IBD, PD, CAL, baseline and mean change in parameters from baseline to follow-up visits). Disagreements on study inclusion or data extraction were resolved through discussion among the researchers. When necessary, a third investigator (M.F.Yao) helped to reach a consensus with all reviewers.
Risk of bias assessment
The risk of bias of the included studies was performed independently by two researchers (R.Y.Cao & Q.L.Li) using the Cochrane Collaboration tool [
23]. Any disagreements were solved by the third investigator (M.F.Yao).
Statistical analysis
The treatment outcomes were measured as the absolute difference (AD) in IBD, PD, and CAL in at least 6 months after periodontal treatment. When standard deviations (SD) for the outcomes parameters were not available, they were calculated by assuming the correlation coefficient to be 0.5 as previously described [
24]. Based on patient characteristics, the studies were divided into three subgroups (systemic healthy, T2DM, and smokers). Network meta-analysis was only applied to the systemic healthy subgroup as there were two studies in other subgroups. The same follow-up duration was used in this meta-analysis in the subgroups.
First, we developed a random-effects pairwise meta-analysis in Stata 14.2 (Stata Corporation, College Station, TX, USA). Weighted mean difference (WMD) and 95% confidence intervals (CIs) were used to compare continuous variables. Second, Bayesian network meta-analyses were performed by using a random-effect model to pool the effect sizes of both direct and indirect comparisons. Non-informative uniform and normal prior distributions were used throughout the network meta-analysis. Markov chain Monte Carlo methods with four chains of 300,000 iterations after a burn-in phase of 100,000 iterations was performed to achieve credible mean difference (MD) and 95% credible intervals (CrIs). We used CrIs beyond the null value to assess significance and ranked different treatments.
Inconsistency was assessed by comparing direct evidence with indirect evidence from the entire network at each node (node-splitting analysis) with
p < 0.05 [
25]. Moreover, we examined the pooled effects from traditional pairwise meta-analysis and network meta-analysis to further verify the consistency of the network. The goodness of fit of the model was tested by calculating the posterior mean residual deviance (Dbar). When the Dbar was similar to the number of data points in the study, the model was considered to fit the data well [
26,
27]. Heterogeneity was assessed with I
2 calculation. Sensitivity analysis was performed to verify the robustness of our analyses by excluding studies with a high risk of bias then the effect was recalculated. R 3.2.2 (R Foundation for Statistical Computing, Vienna, Austria) GeMTC 0.8 package was used to analyse all data.
Discussion
Statins, possess anti-inflammatory, anti-microbial, osteo-simulative, and antioxidant properties which may partly account for their beneficial effects in treating CP. Statins were shown to suppress inflammatory factors associated with periodontitis such as IL-6, TNF-α [
43], IL-1β [
44], as well as periodontal pathogens
Pg and
Aa [
45]. Statins could also inhibit the secretion of matrix metalloproteinases (MMPs) [
46], which are involved in the destruction of periodontal tissue. Moreover, statins increase bone regeneration by inducing the expression of BMP-2, VEGF, and OPG [
47,
48]. Thus, it is unsurprising that local use of statins provides additional benefits for periodontal parameters of CP with or without systemic disease.
Traditional meta-analyses fail to measure the relative effect as they only synthesize studies with the same pair of comparators; network meta-analyses have been proposed to overcome this drawback. In our study, we performed a Bayesian network meta-analysis to compare the relative effect of different statins and found their efficacy to be similar, consistent with a study by Muniz et al. [
20] who used meta-regression. Contrastingly, a study by Bertl et al. [
4] found that RSV was more efficacious than SMV for all parameters tested and ATV in all parameters except for residual IBD. However, Bertl et al. [
4] included patients with different characteristics and different periodontal therapy which may partially account for this inconsistency with our results. More direct evidence is needed in further test and compare the efficacy of different statins.
In addition, another advantage of network meta-analysis is that Bayesian chain assists in ranking the treatment efficacy by measuring the corresponding probability [
49], so that it could provide more evidence to guide clinicians. Though we found no difference between diverse statins, ranking can pave the way for understanding the differences in opinions on the use of either statin in periodontal disease. Our results indicate that SMV is ranked the best in PD reduction and CAL gain. SMV is considered to be the best statin against periodontal pathogens such as
Pg and
Aa. Moreover, SMV was observed to decrease the expression of MMP-1, MMP-3, MMP-8, MMP-9 and MMP-13 [
50‐
52]. RSV may be the best optimal performer in terms of IBD fill. Additionally, RSV has a greater anti-inflammatory action due to more effective suppression of C-reactive protein levels. Moreover, RSV is more effective in reducing low-density lipoprotein cholesterol which had benefits in induced periodontitis in hypertensive rats via inflammatory gene profile modulation [
53].
We also assessed the efficacy of adjunctive statins in CP with T2DM comorbidity or smoking history as these are both risk factors for CP. High levels of blood-glucose increase advanced glycation end-products (AGE) and receptor of AGE (RAGE) leading to an exaggerated inflammatory response and periodontal tissue destruction by oxidative mechanisms [
54,
55]. Smoking can similarly up-regulate the expression of RAGE [
56,
57]. Statins possess strong antioxidant properties which may improve treatment outcomes for CP patients with T2DM or those who smoke. Existing RCTs indicate that locally applied ATV or SMV adjunctive to SRP was more effective than SRP alone in CP patients with T2DM or in smokers [
32,
33,
35,
38]. The results of our traditional meta-analyses also support these findings and is consistent with another meta-analysis conducted by Ambrósio et al. [
19]. However, the sample size in these trials was too small to draw a strong conclusion and more high-quality RCTs are needed to further to validate our results.
We observed a high degree of heterogeneity in CP patients without other systemic diseases. This may be attributable to variables such as different gel doses of statins used for treatment (0.1 ml or 10 ul) in the included trials. In addition, the measurement of IBD from the conventional radiographs was not calibrated which may have caused geometric errors in assessing IBD fill.
Limitations
This network meta-analysis has several limitations that should be noted. Firstly, the length of follow-up of the included trials were relatively short. Secondly, the sample sizes (28–34) for each group were relatively small. Finally, the heterogeneity was high despite decreasing the discrepancy among the characteristics of patients. Multi-centered RCTs with larger sample size and with an extended follow-up duration up to 12 or 24 months are needed to confirm the beneficial effects of statins in combination with nonsurgical periodontal treatment for CP.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.