INTRODUCTION

Due to its organic–inorganic nature, dentin still remains a challenging substrate when a stable and long-lasting bond is desired. Self-etch systems bond composite resin materials to dental substrates; however, some drawbacks still need to overcome,^1–4 so, to improve their effectiveness, different functional monomers have been included in adhesive system formulae, and their effectiveness has been studied over the past decades.^5,6

In a recent systematic review (SR) and meta-analysis (MA),⁷ 2-hydroxyethyl methacrylate (HEMA)-free adhesive was compared to HEMA-containing systems, and similar clinical behaviour was found. HEMA is a monomer present in a great number of systems and plays an important role, due to its hydrophilicity, wettability, and miscibility.^8,9 However, some other monomers that have different functions are also found in adhesive systems, especially in self-etch systems.

Functional acidic monomers, for instance, present in self-etch mode, are able to chemically interact with hydroxyapatite (HAp) and are composed by specific carboxylic, phosphonic, or phosphate groups, such as: Phenyl-P, 10-methacryloyloxydecyl dihydrogen phosphate (10-MDP) and 4-methacryloyloxyethyl trimellitic acid (4-MET).¹⁰

Depending upon the acid dissociation constants (pKa values), etching aggressiveness in self-etch adhesive systems can be classified into “strong” (pH<1), “intermediately strong” (pH≈1.5), “mild” (pH≈2), and “ultramild” (pH≥2.5).¹¹ The more aggressive the system, the deeper demineralization of the tooth substrate occurs, resembling that of phosphoric acid-etching treatment.¹²

Mild self-etch adhesives bond to tooth tissue through a twofold micromechanical and chemical bonding mechanism. They only partially demineralize dentin, with some HAp remaining around the collagen within a submicron, hybrid layer.^13,14 Thus, functional acidic monomers, such as 10-MDP, that are able to react with calcium, would certainly bond electrostatically to HAp, forming calcium salt around the partially demineralized collagen fibers. This possibility has been proved by X-ray photoelectron spectroscopy (XPS).⁵

After patent expiration of the 10-MDP molecule from Kuraray, this monomer quickly became present in the compositions of several brands of so-called multimode or universal adhesives. Several aspects of the 10-MDP molecule made it attractive: firstly, it readily adhered to HAp, and this bond appeared to be very stable, as confirmed by the low dissolution rate of its calcium salt in water.¹⁵ Secondly, 10-MDP bonding versatility has allowed its use beyond dental substrates, to several others, such as zirconia, lithium-disilicate, and metals, which has been very interesting from a clinical viewpoint.^16,17 Therefore, 10-MDP was introduced very quickly into multimode adhesive system compositions (the so-called universal adhesives) globally.

It is important to mention that 10-MDP is not the only constituent in adhesive system compositions, but is present amongst a blend of monomers, solvents, initiators and, nowadays, nanoparticles, which combine to complete the adhesive process.¹⁰ However, due to the great interest in the addition of 10-MDP to adhesive systems, and the good results presented in laboratory studies,^18–23 the aim of this paper was to present a review of the clinical performance of noncarious cervical lesion (NCCL) resin restorations made with adhesives that either did, or did not, contain the 10-MDP monomer, through SR and MA.

METHODS

Information Sources, Eligibility criteria and Search Strategy

The Medical Subject Heading (MeSH) terms, synonyms, and free terms (keywords) in the search strategy were defined using the PICOS²⁵ guidelines:

1. Population (P): Adult patients with NCCLs

2. Intervention (I): Composite resin restorations placed in NCCLs in a self-etch mode, with 10-MDP-containing adhesive systems

3. Comparison (C): Composite resin restorations placed in NCCLs in a self-etch mode without 10-MDP in the adhesive system formulae

4. Outcome (O): Clinical performance of restorations (note that no outcome was used in the search strategy to maximize its results)

5. Study design (S): Randomised controlled clinical trials (RCTs) and controlled clinical trials (CCTs).

In this sense, only RCTs comparing the clinical effectiveness of self-etch adhesive systems with and without the functional monomer 10-MDP for NCCL direct resin composite restorations in the permanent dentition of adult patients (male and female) of any age group were included. Editorial letters, historical reviews, pilot studies, in vitro studies, and cohort, observational, and descriptive studies, such as case reports and case series, were excluded.

To identify articles for inclusion in this review, an electronic search of the literature was conducted up to 21 March 2019, plus an alert for recently published articles was requested from the following databases: MEDLINE via PubMed, Web of Science, Scopus, Cochrane Library, OpenGrey, Clinical Trials, Current Controlled Trials, and EU Clinical Controlled Trials. No restrictions were applied on either publication date or language, except that the study must have included a follow-up of at least 1 year.

The search strategies were defined appropriately for each database (Table 1), and searches were performed by two independent reviewers (RPO and JCPB) to identify eligible studies. Full-text versions of all articles that appeared to meet the inclusion criteria were obtained for subsequent evaluation and data extraction. Next, a hand search was performed in the reference lists of included articles to identify any additional relevant studies that had not been found during database searches.

Table 1: Search Strategies for Databases

View Fullscreen

Summary Measures and Results Synthesis

The extracted data were analysed using RevMan software (Review Manager v. 5.3, The Cochrane Collaboration, Copenhagen, Denmark). Five meta-analyses were performed according to the main parameters analysed: retention (RE), marginal adaptation (MA), marginal discoloration (MD), caries (CA), and postoperative sensitivity (POS). Each parameter and the overall effect (clinical performance) were analysed.

The outcomes were divided into acceptable or unacceptable (Table 2), according to the classification criteria used by each study. The prevalence of unacceptable (failures/events), and the total number of restorations for each group, were used to calculate the risk difference with a 95% confidence interval (CI). Random effects models were employed, and heterogeneity was tested using the I² index.

Table 2: Dichotomy of Results According to the Studies Evaluation Criteria

View Fullscreen

If some of the information needed for the MA was missing from any of the selected studies, the authors were contacted to see if the missing data could be provided. Where contact was necessary, up to five attempts at contact were made; and, if after these contact attempts no response was received from the authors, or the authors did not provide the data, the study was not included in the MA.

RESULTS

Study Selection

The selection methodology has been summarised in Figure 1. A total of 7927 citations was obtained from the database searches. Following the exclusion of duplicates, 4208 articles were identified, and then, after title screening, 215 studies remained, with this number then further reduced to 79, through careful review of the abstracts. The full texts of the remaining 79 articles were examined for eligibility, leading to an additional 68 exclusions, with the remaining 11^26–36 admitted into our SR and MA.

View Figure

• Download as Powerpoint
• Download Figure

Characteristics of the Included Articles

The characteristics of the 11 included studies, which varied in length from 1 to 4 years, have been listed in Table 3. The numbers of patients involved in these randomized clinical trials varied from 21 to 124, and their ages spanned 20–84 years. During the restoration procedure, there was generally just one operator involved, although the number of operators did range up to two. The incremental technique was the most commonly used filling method, and the most common isolation technique involved use of cotton rolls and saliva aspirators. In two studies,^32,35 exposed dentin was superficially prepared by bur roughening, and the enamel was bevelled in one study,³¹ while both the procedures (roughening and bevelling) were performed in four studies.^29,30,33,36 No preparation procedures were performed in another four studies.^26–28,34

Table 3: Summary of the Studies Selected for this Systematic Review

View Fullscreen

Table 3: Summary of the Studies Selected for this Systematic Review

View Fullscreen

In total, 11 different adhesives were tested in the included studies. The most tested adhesive without 10-MDP was GC bond, and the most tested system with 10-MDP was Clearfil SE Bond. With regard to the adhesive approach, all studies used one- or two-step, self-etch techniques. Among systems without 10-MDP, six used the one-step, self-etch approach and two used the two-step, self-etch method. Among those with 10-MDP, two used one-step, self-etch and three used the two-step, self-etch approach.

For restoration evaluation, the majority (seven) of the studies used the modified United States Public Health Service (USPHS) criteria,^{26,27,29,31,32,35,36} one used the Federation Dentaire Internationale (FDI) criteria,²⁸ two used the Modified Ryge criteria,^33,34 and one used the Vanherle and others criteria³⁰ (Table 2).

Assessment of Bias Risk

An assessment of bias risk was performed for the included articles, and the results have been presented in Figure 2. E-mails were sent to all authors, requesting further information, and nine responded.^{26,28,30–36} Among the 11 selected studies therefore, two full-text articles^27,29 were classified as having an “unclear risk of bias”, in the key domains of the Cochrane bias tool, and nine studies—those listed above whose authors responded to our request for additional information—were considered as having a low risk of bias.

View Figure

• Download as Powerpoint
• Download Figure

Meta-analysis

The meta-analysis (MA) was conducted with the data available in the “low and unclear risk of bias” studies included in this SR. Only two studies did not provide sufficient data for POS—Kubo and others²⁹ and Ruschel and others.³² The group data that had the highest prevalence of failures (events) was used for studies that included more than one group of 10-MDP-containing Zhou and others³⁶ or 10-MDP-free adhesive systems Van Dijken and others.³⁵

For this MA, the overall heterogeneity was not important (I²=0%). For each parameter, the heterogeneity ranged from 0–53% (53% for RE, 47% for MA, 12% for MD, 0% for CA, and 0% for POS), or from “absent” to “substantial.”³⁸ The 10-MDP-free adhesive group showed 35 failures out of 546 restorations evaluated for the RE parameter, 39 failures out of 511 restorations that were evaluated for MA, 13 failures from 511 restorations that were evaluated for MD, 1 failure out of 511 restorations that were evaluated for CA, and 6 failures from 418 restorations evaluated for POS.

The 10-MDP-containing adhesive group showed 24 failures from a total of 573 restorations evaluated for the RE parameter, 33 failures out of 549 restorations evaluated for MA, 19 failures from 549 restorations evaluated as MD, 1 failure from 549 restorations evaluated as CA, and 5 failures from 461 restorations evaluated for POS. The overall risk difference was −0.00 [−0.01, 0.00] (p=0.77), and was −0.01 [−0.05, 0.02] (p=0.46) for RE, −0.01 [−0.03, 0.02] (p=0.46) for MA, 0.00 [−0.01, 0.01] (p=0.90) for MD, 0.00 [−0.01, 0.01] (p=0.92) for CA, and 0.00 [−0.01, 0.01] (p=0.94) for POS (Figure 3).

View Figure

• Download as Powerpoint
• Download Figure

This meant that 10-MDP-free and 10-MDP-containing adhesive systems showed statistically similar clinical performance (overall effect) for all isolated parameters, with moderate certainty of evidence for retention and high certainty of evidence for other parameters (MA, MD, CA, POS, and for pooled results). Table 4 lists the detailed GRADE classifications for each parameter.

Table 4: Certainty of Evidence

View Fullscreen

DISCUSSION

Modern adhesive systems have exhibited good short-term bond ability; however, interface stability is still challenging where long-term behaviour is concerned.³⁷ Adequate restorative material stability, in terms of its adhesion to the dental substrate, is only possible when high-quality and long-lasting micro-mechanical and chemical interactions take place between the adhesive system and dental substrate^38,39 in a hybrid layer, as described by Nakabayashi.^40,41 The evidence for this adhesive interaction was based on deep impregnation of dental substrates, especially dentin, by the adhesive system, after strong acid conditioning.^42,43

Lately, some studies have been published that identified, inter alia, problems related to clinical performance. These issues were found as part of middle- and long-terms follow-ups, and were mostly related to inability on the part of the adhesive systems to totally fill the spaces provided by the acid etching. This was seen to have left collagen deprived of mineral protection, leaving it degradation prone, mainly by some enzymes (matrix metalloproteinases) activated during the acid etching.^31,44–47

Self-etching systems were developed not only to simplify the entire adhesive process but also to allow complete monomer infiltration, allowing tissue conditioning and resin infiltration to take place at the same time.^46,48 In these systems, acidic monomers were added to play this role, and several modifications have been made over time to self-etching adhesive systems. Changes in the number of steps (one or two) and aggressiveness (strong, middle, mild, and ultra-mild) were the principle modifications,^42–44 and, as for etch-and-rinse adhesives, strong self-etch adhesives completely remove hydroxyapatite (HAp) from dentin, resulting in relatively deep dentin hybridization that is several micrometers thick. However, part of the HAp remains in the hybrid layer, where it may negatively influence bond strength. On the other hand, mild self-etch adhesives, due to their less aggressiveness, form only submicron-thick hybrid layers, and the HAp is not attacked as observed for the strong self-etch adhesives, which allows better mechanical and chemical interactions between functional monomers and HAp. ^5,49,50

Conceptually, mild self-etch adhesives bond to tooth tissue through a twofold (micro-mechanical and chemical bonding) mechanism. Partial dentin conditioning occurs, leaving some HAp around collagen fibers within a submicron hybrid layer.^13,42 In the 80’s, Kuraray developed and patented 10-MDP, a mild functional monomer, with a hydrophobic methacrylate group at one end, capable of reacting with methacrylate-based restorative materials, and a hydrophilic polar phosphate group at the other end, which was able to react with metals, zirconium, and dental mineralized tissues.^20,21 This monomer also exhibited the interesting property of low hydrophilicity, due to its long molecular chain.^5,51

Based on the adhesion–decalcification (AD) concept, functional monomers like 10-MDP bond electrostatically to HAp, forming an insoluble MDP-calcium salt CaMHP2,^10,52,53 thereby contributing to the long-term stability of the bond.^50,54 Primary chemical bonding to HAp was first demonstrated using XPS,⁵ and later confirmed using X-ray diffraction (XRD) and nuclear magnetic resonance spectroscopy (NMR).^44,45,55 The XRD examination also showed that 10-MDP bonding to calcium formed a Ca–HAp compound, characterized by a nanolayered structure.^22,42,55 This salt protects the interface against hydrolysis, due to its hydrolytic stability.⁴⁸

The immediate and widespread interest by several manufacturers—including the makers of Clearfil SE Bond, Clearfil S3 Bond, and Clearfil Protect—in adding 10-MDP to adhesive systems was probably due to the good results observed when Kuraray systems were tested.^20,21,46 Some laboratory studies demonstrated good performance when bond strength and sealing ability were important.^{20–23,39,46} In one of them the adhesive systems were evaluated using mechanical and physical parameters, and the results showed that only systems containing 10-MDP^43,45 were able to protect the bond interface against nanoleakage, and that these systems presented a higher elastic modulus after 1 year.^10,39,45

Some clinical studies showed favourable short- and mid-term performance when using systems with formulae that included 10-MDP. It is of interest to mention here that, among the 11 RCTs included in this SR, 10 used Kuraray’s system (containing 10-MDP in their formulae), compared to other self-etching systems without this functional monomer. Only one study, published in 2018,³² used different systems, applying Scotchbond Universal (3M) and Prime & Bond Elect (Dentsply Sirona). The RCTs selected normally used retention, marginal degradation and staining, postoperative sensitivity and secondary caries as clinical criteria.

As the MA only accepted paired comparisons, two systems from each RCT were selected—one containing 10-MDP and the other not. In this study, only the adhesion of self-etching systems to dentin was considered. Dentin is a vital substrate, is very dynamic, and has a high water content,⁵⁶ and it has also been noted that some pathological changes may occur over time, which make the adhesion process very challenging.^1,57 In this particular SR, neither clinical studies that involved adhesion to enamel nor the total etch technique were included, due to the number of variables that should be considered and discussed, such as self-etch or selective approaches.

Retention is normally used as the primary outcome, and for definition purposes in this study, partial or total loss of restorations was taken as implied by their replacement.⁵⁸ The results for this clinical criterion in seven studies^{27,28,30,31,33,34,36} did not present significant differences between the two compared groups. On the other hand, statistical differences for retention criteria were reported in two studies.^26,35 When these results were subjected to the MA, however, they did not show any differences when adhesives containing 10-MDP or not were compared. Better clinical performance (retention) had been expected, when 10-MDP was present in the adhesive formulation, once this functional monomer reacted with HAp-forming Ca-MDP salt, which has a hydrolysis protective effect.^54,57–59

In the present SR, all of the 10-MDP systems included were in commercially available products with the HEMA monomer in their composition; perhaps this association could explain the MA results in which no significant difference in retention among the adhesive systems were detected. Despite the important role, in a recent SR and MA, the adhesive containing HEMA was compared to HEMA-free adhesives, and a similar clinical behaviour was found.⁷ Over time, the high hydrophilicity of this monomer promotes an increase in water acceptance that results in a hydrolytic degradation of the adhesive interface.^60,61

The presence of HEMA may also interfere with the MDP-Ca nanolayer formation, so thus the retention rate for the 10-MDP-containing systems were not significantly higher, as had been expected.^54,57–59 In a study by Tian and others,⁶² several commercial adhesive systems showed little nanolayer formation behaviour at the resin–dentin interface, while more uniform CA-MDP nanolayering was detected for the one system (G-Premio Bond/GC) that did not use HEMA in its formulations. The demineralization rate of HAp by 10-MDP is reduced by HEMA presence, and 10-MDP remains adsorbed onto the HAp surface. Using XRD and NMR to better understand the nanolayering structure in the presence of functional monomer HEMA, Yoshida and others⁵⁴ confirmed the chemical interaction between Ca dissociated from HAp and MDP to form the MDP-Ca salt-layered structure. However, when HEMA was added, MDP-Ca salt formation was remarkably decreased and part of the crystalline phase compromised. As XRD analysis can only detect the presence crystalline structures, NMR was used to investigate the chemical interaction of MDP with HAp in HEMA-containing formulations and to unravel the mechanism of how HEMA inhibits nanolayering formations. The NMR analysis showed that when HEMA concentration was increased, the interaction of MDP and HAp decreased. Assessing HEMA-free MDP formulations, shoulder peaks that indicate more consolidated crystalline structures were observed. On the other hand, when HEMA was present in the formulation, weak shoulder peaks were detected that may explain the reduction in the nanolayering formation, probably due to the affinity of HEMA molecules to the HAp, which may compete with the MDP-Ca salt formation. Therefore, one can speculate that this phenomenon could interfere in the retention rates and also in the bond durability.⁵⁴

10-MDP benefits could be demonstrated in laboratory studies, in which the adhesives were synthetized with a high degree of purity and were applied using controlled concentrations. Some studies^21,42,63 have evaluated different 10-MDP concentrations (1%, 3%, 5%, and 15%) with high purity, in terms of Ca-MDP nanolayer formation, and the 15% 10-MDP formulated adhesive exhibited more intense nanolayer deposition. Unfortunately, neither the degree of purity nor the concentration are clearly presented for commercial adhesive systems. On the other hand, Carrilho and others⁵⁹ in an SR of 73 laboratory studies stated that commercial adhesives containing 10-MDP had a proven capacity to interact to HAp; however, some clinical steps relating to application of this bonding system must be conducted with due care and attention, because of their influence on the quality of the resulting bond interface. These steps include the selective enamel etching and scrubbing technique used to apply the adhesive systems to dental substrates. The authors also emphasized the importance of allowing sufficient time for the solution to infiltrate, to hybridize, and to form the MDP-Ca salt, protecting collagen fibrils and improving adhesive stability. Perhaps these techniques have not been followed in some clinical studies as precisely as they would have been in laboratory studies, which could also have influenced the results.

Follow-up time is another point to be considered, when reviewing the equity of the various retention rate results described in the RCTs. The longest clinical tracking time was 4 years (Table 3), which may not have been sufficient to detect significant differences.

Secondary outcomes considered in this SR included marginal staining and degradation, and here again, the MA did not detect significant differences when the presence of 10-MDP was considered. Formation of CA-MDP nano layering is more closely related to retention aspects, and so plausible explanations of differences may only show over time, which in some cases may not have been sufficient to detect differences, or to differentiate between variations in the mechanical and physical properties of the restorative materials. Clearly, longer evaluation periods will need to be considered for future studies.

Secondary caries and postoperative sensitivity were also considered as secondary outcomes, and, again, no significant differences were detected in the MA for these effects. The ease with which the restored areas can be kept free of biofilm, and the favourable C-Factors of the NCCLs, may also explain why the different restorative systems showed no significant variations with respect to these characteristics.⁷

Several other aspects may affect the clinical behaviour of dental restorations over time, including the patient, the type of tooth and cavity, the operator, and the restorative system used.^28,56 When the integrity of the adhesive interface is a concern, the type of tooth, cavity, and restorative system may have more impact.^64,65 In this SR, only clinical trials that involved NCCL restorations were included, although for this type of cavity, variation in size and in the amount of exposed dentin would probably not lead to different results. However, we could not affirm the same in cases with different C-factors, or types of occlusion, or with the morphological and/or pathological dentin changes normally present in class II cavities, for instance. The NCCLs that were the subjects of these controlled clinical studies reported little variation in these aspects.^7,31

Noting that two studies considered for inclusion were regarded as unclear in some review domains, the strength of the data in those studies that were included in this SR was an important aspect that needed to be addressed; and the GRADE approach indicated that the evidence was of high quality for all evaluation criteria, except for retention, which was classified as moderate. This confirmation suggested that it would be unlikely that future studies would arrive at conclusions different from those of the present SR.

CONCLUSION

According to results from this SR and MA, strengthened by application of the GRADE approach to the evidence of selected articles, and considering the follow-up time presented in selected RCTs, the authors were able to conclude that the presence of 10-MDP monomer in self-etch adhesive systems did not influence the clinical performance of NCCL restorations.