MMP Inhibitors and Dentin Bonding: Systematic Review and Meta-Analysis

Objectives Resin-dentin bond strength decreases over time. This reduction is related to the loss of hybrid layer integrity. Collagenolytic enzymes, especially matrix metalloproteinases (MMPs), are responsible for the degradation of the collagen matrix of the hybrid layer. Various MMP inhibitors with the ability to prevent enzymatic degradation have been identified. This study aimed to systematically review the literature for studies which evaluated the effect of MMP inhibitors on the immediate and aged dentin bond strengths. Study SelectionScreening and analysis were carried out by two reviewers. Two databases were searched, and from a total of 740 articles, 43 were accepted for full review. 21 articles with 0.2%–2% chlorhexidine (CHX) treatments were included for meta-analysis. A risk of bias assessment was performed on all studies chosen for meta-analysis. A variety of MMP inhibitors have been studied, CHX being the most widely used. Conclusions A clear trend for a lower loss of dentin bond strength was observed with different MMP inhibitors. In meta-analysis, no significant difference was seen between the CHX and control in the immediate bond strengths. Bond strengths in the CHX group were significantly higher than the control group after aging (P < 0.001). The percentage of fractures occurring at the adhesive interface increased after aging. Five out of 21 studies included in the meta-analysis had high and the rest medium risk of bias. More long-term studies with lower risks of bias should be carried out to increase the reliability of results. Clinical RelevanceThe use of MMP inhibition with chlorhexidine can be recommended to increase the longevity of resin-dentin bond strength.


Introduction
Studies have shown that the bond between the adhesive systems and dentin weakens over time.
is decrease in bond strength is related to the degradation of the hybrid layer [1,2]. e hybrid layer is the area of adhesion formed by the dentin collagen matrix and resin adhesive. After exposure to acid (etch-and-rinse adhesives) or acidic monomers (self-etch adhesives), the demineralized dentin collagen matrix is infiltrated with the applied adhesive resin [3,4]. e collagen matrix is vulnerable to enzymatic degradation by the endogenous collagenolytic enzymes, matrix metalloproteinases (MMPs) and cysteine cathepsins, and these enzymes play an important role in bond destruction [2]. MMP inhibitors can prevent the function of these enzymes, with chlorhexidine (CHX), galardin, and benzalkonium chloride being the most widely studied [5,6]. More recently, collagen cross-linker agents have also been shown to inhibit protease enzymes [7].
Systematic review and meta-analysis by Montagner et al. [8] indicated that CHX is effective in maintaining long-term bond strength.
is study aimed to provide an updated systematic review and meta-analysis of the literature involving studies which evaluated the effect of MMP inhibitors on the immediate and aged resin-dentin bond strength. e hypothesis was that MMP inhibitors have a significant effect on the longevity of bond strength, which becomes evident after 6 months.
bonding, no control group, review articles, or other interests, e.g., root canal sealers and root canal posts. If any one of these factors was present, the article was rejected. e article also had to be written in English.
e screening was performed in three stages ( Figure 1). During the first stage, only the title and the abstract of the article were used to determine whether any of the rejection criteria was present. e initial screening resulted in 126 articles from PubMed and 22 articles from Scopus. During the second stage, the complete text of all remaining articles was read and interpreted by the individual screeners (a total of 148). e same rejection criteria were used, and after careful selection, 59 articles from PubMed and seven articles from Scopus were accepted (n � 66). A final screening was performed with both reviewers present. Five of the seven articles found in Scopus were the same as on PubMed. Furthermore, 18 studies were rejected due to the rejection criteria and lack of available data, leaving 43 accepted articles. In addition, studies done on carious teeth were excluded.
Due to the wide heterogeneity between the studies using other MMP inhibitors than CHX, only data involving 0.2-2% CHX would be used for the meta-analysis. e 21 chosen articles were allocated into six different comparison groups depending on whether the samples had been aged for 6, 12, or 24 months and whether an etch-and-rinse or selfetching system had been used. No articles were found for the 12-month storage time and self-etch, so this group was excluded. e bond strength, sample size (N), and standard deviation data for the respective storage times and controls were retrieved from the articles. e N for each group represented the total number of teeth used in each comparison group. Pooled effect estimates were attained by comparing the means of each bond strength value, expressed as the raw mean difference among the groups. Statistical heterogeneity of the treatment effect was assessed via the Cochran Q test, with P < 0.05 considered significant, and the inconsistency I 2 test, in which values > 50% were considered to indicate high heterogeneity. Meta-analysis on the chosen comparison groups was carried out using the MedCalc (version 19.2.1: MedCalc Software Ltd., Ostend, Belgium).

Assessment of Risk of Bias.
e risk of bias evaluation, adapted from a previous study [8], evaluated the following parameters for the study's quality assessment: randomization, use of intact teeth, use of materials according to the instructions, adhesive procedures performed by the same operator, description of sample size calculation, and blinding of testing. e articles reporting 5 to 6 items were classified as low risk of bias, 3 or 4 as medium risk, and only 1 or 2 as high risk.

Results
From the initial 740 articles, 43 articles with altogether 240 groups were subjected to a comprehensive examination (Table 1). Altogether, 21 different enzyme inhibitors were tested. e most commonly used MMP inhibitor was CHX (32 studies), followed by BAC (seven studies). Artificial saliva and distilled water were the most commonly used modes of storage for aging. Percentage decreases in bond strength were calculated, and a general trend in the decrease of bond strength with time could be observed. A clear trend for the lower decrease in bond strength with MMP inhibitors in 35 out of 43 studies was also observed, with 13 different enzyme inhibitors showing significantly (at least 50%) lower percentage loss of bond strength compared to the respective control group. e adhesive-mixed fracture percentages for the control and CHX groups ranged from 0 to 100 at baseline, from 41 to 100 at 6 months, from 50 to 100 for the control, and from 55 to 100 for CHX at 12 months, and from 77 to 100 for the control and from 75 to 100 for CHX in the final aging group of 24 months.

Meta-Analysis.
A total of 21 articles were subjected to meta-analysis. e first analysis involved CHX vs. control at baseline (Figure 2), including 37 data sets from 21 articles. No significant difference in bond strength was present between the groups (P � 0.308). e heterogeneity between the studies was low (Cochran's Q P > 0.05, I 2 21.7%).
irty data sets from 16 articles were available for the comparison between CHX vs. control after 6 months aging (Figure 3(a)). Bond strengths in the CHX group were significantly higher than in the control group after aging (P < 0.001). e heterogeneity between the studies was high (Cochran's Q P < 0.05, I 2 82.1%).
To compare CHX vs. control after 12 months aging, 17 data sets from 10 articles were available (Figure 3(b)). Bond strengths in the CHX group were significantly higher than in the control group (P < 0.001). e heterogeneity between the studies was high (Cochran's Q P < 0.05, I 2 75.0%). e last analysis involved CHX vs. control after 24 months aging, including six data sets from three articles (Figure 3(c)). Bond strengths in the CHX group were significantly higher than in the control group (P < 0.001). e heterogeneity between the studies was extremely low (Cochran's Q P � 0.817, I 2 0%).

Risk of Bias.
Of the 21 articles selected for the metaanalysis, five were classified as having a high risk and 16 as a medium risk of bias. None of the articles had a low risk of bias ( Table 2).

Discussion
e general trend of all 43 articles with 21 different collagenolytic enzyme inhibitor protocols demonstrated markedly lower loss of bond strength with enzyme inhibition. Studies involving other MMP inhibitors were excluded from the meta-analysis due to a wide range of heterogeneity in the inhibitors used and a small number of studies for each inhibitor except for CHX. All three comparisons between the CHX-treated and controls after aging for at least six months demonstrated significantly higher bond strength with CHX. us, the hypothesis was accepted. e results of the meta-analysis showed that the use of CHX has no significant effect on immediate resin-dentin bond strength. e finding is in line with a previous study [8]. Two studies [14,23] have shown a significant decrease in the immediate bond strength of the CHX group compared to the control. After reviewing the articles, no clear explanation for this difference could be identified, although several differences in the application of CHX were noted. Giacomini and co-authors speculated that the use of acidic CHX after acid etching may have resulted in increased collagen exposure, possibly reducing the immediate bond strength [14].
After aging for 6, 12, and 24 months, the meta-analysis demonstrated significantly better bond strength with CHX compared to the control groups. Despite the heterogeneity of the studies, 6-and 12-month analyses indicate the advantage of using CHX to preserve the bond strength. In addition, with the studies evaluating the bond strength after 24 months of aging, the homogeneity of the data was striking (I 2 0%). Indeed, longer aging seems to increase the difference between the bond strengths of CHX and control groups. e immediate fracture percentage at the adhesive interface was practically the same for MMP inhibitor and control groups, although some isolated differences can be identified. It can generally be observed that as the follow-up time increases, the percentage of fractures occurring at the adhesive interface seems to increase, regardless of enzyme inhibition. is may be due to slow degradation of hybrid layer collagen despite the enzyme inhibition, the hydrolytic degradation of the resin component, or-most likely-to International Journal of Dentistry  International Journal of Dentistry 5  International Journal of Dentistry 7     Figure 3: Forest plots of studies after 6, 12, and 24 months of aging. (a) Chlorhexidine (CHX) vs. control at six months. e N for both groups was 339 samples. e total random effect SMD was 0.907 (CI 95% 0.517; 1.297). e difference was statistically significant (t � 4.568, P < 0.001). e I 2 was 82.11%. (b) CHX vs. control after 12 month aging. e N for both groups was 173 samples. e total random effect SMD was 0.821 (CI 95% 0.367; 1.275). e difference was statistically significant (t � 3.557, P < 0.001). e I 2 was 74.97%. (c) CHX vs. control after 24-month aging. e N for both groups was 40 samples. e total random effect SMD was 2.168 (CI 95% 1.627; 2.708). e difference was statistically significant (t � 7.982, P < 0.001). e I 2 was 0.00%.  [2,6]. is supports the idea that bond strength decreases over time and that the adhesive interface plays a significant role in the mode of fracture. A thorough risk of bias assessment was also carried out to identify the main factors which could affect the creditability of the findings. Five articles were classified to have a high risk of bias, and none with low risk. e results are in line with the respective previous study [8]. None of the studies mentioned sample size calculations, and all but one failed to mention the blinding of the operator performing the bond strength testing. e results may reflect the standard level of reporting of bond strength studies, but at least the blinding of the person performing the bond strength testing should be done and also reported.

Conclusions
is systematic review and meta-analysis demonstrated that studies strongly indicate the benefits of collagen-degrading enzyme inhibition on the preservation of dentin bond strength. Since CHX does not have any adverse effects on the immediate bond strength, the clinical use of CHX can be recommended to increase the longevity of resin-dentin bonds.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.