INTRODUCTION

Noncarious cervical lesions (NCCLs) are commonly observed in clinical dental practice and are characterized by the cervical loss of hard dental tissue, with no occurrence of dental caries or trauma. They may be superficial or deep and can present as extensive defects with various shapes.¹ The etiology of NCCLs has been described as multifactorial, resulting from the combination of different processes, including stress (abfraction), friction (abrasion), and biocorrosion (chemical, biochemical, and electrochemical degradation).^2-4 Initially, NCCLs develop in enamel with slow progression into dentin, leading to gradual dentin sclerosis.^1,5 Sclerosis occurs as a response to low intensity and chronic stimuli from physiologic aging, which is a characteristic consistent with the more frequent occurrence of NCCLs in an elderly population.^6-8

NCCLs are often restored to treat dental hypersensitivity, to improve esthetics, and to prevent further loss of dental tissues.^9,10 However, NCCL restorations have a high index of failure, resulting from loss of retention, secondary caries, marginal discoloration, and marginal adaptation, compromising the longevity of the restorative treatment.^5,11 Problems with NCCL restorations include nonretentive characteristics and the location of lesion margins in dentin or cementum, as well as the presence of dentin sclerosis, which compromises the adhesive process.^10,12-14

Approaches to improving the clinical performance of NCCL restorations, focusing mainly on restorative materials and dentin substrate treatment, have been described.^1,14-21 Although these findings are promising, cavity geometry (shape and size) has been reported to play an important role in the longevity of NCCL restorations.^5,22-25 Also, dentin sclerosis,²⁵ tooth location,^26,27 and the presence of wear facets²⁶ have been reported to be relevant to the retention of NCCL restorations. Others^28,29 have reported that these factors do not influence retention. Therefore, because of the inconsistent results available, a systematic review was conducted with the focused question: Do tooth- and cavity-related aspects of NCCLs affect the retention of resin composite restorations in adults?

METHODS

Information Sources and Search Strategy

To define the search strategy, a preliminary search for studies was conducted using some specific keywords on the characteristics of NCCLs (shape, depth). However, the search did not retrieve relevant results. Therefore, we decided to use terms related to lesions in general and restoration with resin composite. Information on aspects of interest was only determined during the full-text reading of the studies.

The search strategy was conducted using multiple combinations of MeSH terms and free keywords (Table 1). An electronic search was performed in MEDLINE via PubMed, citation databases (Scopus and Web of Science), Latin American and Caribbean Health Sciences Literature database (LILACS), Brazilian Library in Dentistry (BBO), Cochrane Library, and ongoing trial databases, including ClinicalTrials.gov (www.clinicaltrials.gov) and ReBEC (The Brazilian Clinical Trials Registry; www.rebec.gov.br). The non–peer-reviewed literature was searched using the database System for Information on Grey Literature in Europe (SIGLE). Additionally, the reference lists of the included studies were checked to identify possible relevant studies. An expert librarian (DM) supervised the search strategy. No restrictions were placed on the publication date or language. The search strategy was appropriately modified for each database to identify eligible studies.

Table 1 Electronic Database and Search Strategy (5 December 2016, Updated in July 2018)

View Fullscreen

A “Search Alert” with the search strategy in the PubMed, Scopus, Web of Science databases, and Cochrane Library was created, and the search was updated weekly until July 2018.

Study Selection and Data Collection Process

The studies were selected by title and abstract according to the described search strategy (Table 1). Articles that appeared in more than one database were considered only once. Full-text articles were also obtained when the title and abstract contained insufficient information for a clear decision. The search in the Cochrane Library retrieved some abstracts from conferences. The authors were contacted for further information about the publication of the entire study. Subsequently, two reviewers classified those that met the inclusion criteria. Each eligible article received a study identification number, which combined the first author name and year of publication.

Relevant information about the study design, participants, number of treated teeth, follow-up time, dropouts, and factors reported were extracted independently using customized extraction forms by two authors (Table 2); in cases of disagreement, a decision was reached by consensus.

Table 2 Data Extraction From Selected Studies

View Fullscreen

Table 2 Data Extraction From Selected Studies (ext.)

View Fullscreen

When multiple reports of the same study (eg, reports with different follow-up times) were found, the study with the highest long-term follow-up time was included for data extraction. When more than one resin composite or adhesive was included in the study, their values were combined to make a single entry. Three attempts to contact the authors by e-mail were made when data not described in the articles were necessary.

Regarding the retention rate, we collected data from the studies and grouped them according to the factors reported: arch distribution, tooth location, wear facets, dentin sclerosis, shape, size, depth, occluso-gingival distance, and margin location.

Data Synthesis and Statistical Analysis

According to tooth- and cavity-related aspects of NCCLs, the data analyzed were dichotomized, as shown in Table 3. The extracted data were analyzed using the R statistical language R Studio (version 3.4.4; Studio Team, Boston, MA, USA).

Table 3 Dichotomy of Results According to the Evaluation Criteria of the Studies

View Fullscreen

Differences observed between the groups were expressed as pooled relative risk (RR), with 95% confidence interval (CI). The retention rate for each tooth- and cavity-related aspect of NCCLs was evaluated with an intention-to-treat protocol. When not reported in the article, the retention rate was calculated according to intention-to-treat. The factor effect on the defined outcome measurement was calculated from the study data using the random-effects model. Statistical heterogeneity of the treatment effects among studies was assessed using the Cochran Q test and the inconsistency I² test, in which values greater than 50% were considered indicative of substantial heterogeneity.⁴⁷ No subgroup analyses were performed.

RESULTS

Study Selection

The search in the databases led to 6738 articles (Figure 1). After the removal of duplicates, 3715 results remained. By screening title and abstract, articles not related to the topic of this systematic review were excluded. Thus, 235 articles were assessed to verify their eligibility. Among them, 211 were excluded for the following reasons: 1) the study did not report factors related to NCCLs; 2) the study did not compare the factors in the results; 3) they were conference meeting abstracts without full-text publication; 4) the study tested resin composite and other material, but the factors related to NCCLs were not presented for each material; 5) the study reported on lesions with the same characteristics; and 6) a study with longer-term follow-up was selected.

View Figure

• Download as Powerpoint
• Download Figure

Risk of Bias Within Studies

The results of the risk of bias assessment of the 24 included studies are presented in Figures 2 and 3. All the judgments in a cross-tabulation of the study by entry are presented in Figure 2. All the items in four studies were judged as low risk of bias.^33,37,41,46 Figure 3 illustrates the proportion of studies with each of the judgments for each entry. The risk of bias of selective reporting was judged as low, whereas blinding of outcome assessment was judged as unclear. The risk of bias of random sequence generation, allocation concealment, and incomplete outcome data were judged as high; however, the proportion was so small that it would not seriously weaken confidence in the results.

View Figure

• Download as Powerpoint
• Download Figure

Meta-analysis and Summary of Findings

For the meta-analysis, studies were grouped according to the factors related to NCCLs (Figures 4 and 5). All studies for each tooth- and cavity-related aspect were included, despite their risk of bias. The impact of this decision was evaluated in a sensitivity analysis, where only studies with low risk of bias were included. No change in the overall significance was shown (data not shown).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Arch Distribution

For arch distribution (maxillary vs mandibular), 14 studies were included.^{15-20,26,31,34-36,38,41,42} The overall results were RR, 1.01; 95% CI, 0.98-1.05; I² = 23%, suggesting that the arch distribution of the NCCL does not affect the success rate of the resin composite restoration (Figure 4A).

Tooth Location

The forest plot of the meta-analysis for tooth location can be seen in Figure 4B. Eleven studies were included.^{15-17,19,20,26,27,31,35,41,42} The overall results were RR, 1.08; 95% CI, 1.00-1.16. The heterogeneity was high (I² = 82%, p<0.001). Therefore, anterior tooth location favors the retention rates of resin restoration of NCCLs by a factor of 1.08.

Wear Facets

For wear facets, only two studies were included.^26,31 The results were RR, 0.91; 95% CI, 0.83-0.099; I² = 0% (Figure 4C). The presence of wear facets is a risk factor for the retention rate of resin composite restorations.

Dentin Sclerosis

For dentin sclerosis (with vs without), 11 studies were included.^{17,31-33,37,39,40,43-46} The overall results were RR, 0.99; 95% CI, 0.93-1.05 (Figure 4D). The heterogeneity was 60%, suggesting dentin sclerosis does not affect the success rate of the resin composite restoration of NCCLs.

Shape

The forest plot of the meta-analysis for shape can be seen in Figure 5A. Three studies were included.^31,39,46 The overall results were RR, 1.03; 95% CI, 0.91-1.18. The heterogeneity was 51% (p=0.13), suggesting the shape of NCCL does not affect the success rate of resin composite restorations.

Size

For the size of lesions, two studies by the same author were included.^17,45 The size of NCCL does not affect the retention rate of the composite restorations (RR, 0.97; 95% CI, 0.88-1.08; I² = 0%; p=0.44; Figure 5B).

Depth

For the depth (shallow/moderate vs deep) of NCCL, seven studies were included,^{17,18,31,32,39,45,46} and this characteristic does not seem to affect the retention rate of composite restorations (RR, 0.98; 95% CI, 0.92-1.04; I² = 0%; Figure 5C).

Occluso-gingival Distance

Two studies^31,46 reported the occluso-gingival distance of lesions and its influence on the retention rate of restorations (Figure 5D). The results of the meta-analysis for this characteristic showed no effect on the retention rate of restorations (RR, 1.03; 95% CI, 0.93-1.13; I² = 0%).

Margin Location

For margin location (above/aligned vs intrasulcus), only two studies were included.^32,33 High heterogeneity was detected (I²=83%; p=0.02). With an RR, 4.14; 95% CI, 0.17-99.76, the margin location of the NCCL did not influence the retention rate of the resin composite restorations (Figure 5E).

Grading the Body of Evidence

Table 4 displays a summary of the various aspects used to rate the quality of the evidence according to the GRADE working group.⁵²

Table 4 Summary of Finding and Quality of the Evidence

View Fullscreen

The indirect evidence was responsible for downgrading the quality of the evidence by one level for all factors related to NCCLs. For some characteristics (wear facets, shape, size, and occluso-gingival distance), the strength of evidence was also downgraded one level due to imprecision (the optimal information size criterion was not met). Moreover, the inconsistency in the data due to high and nonexplained heterogeneity was responsible for downgrading the results for tooth location, dentin sclerosis, shape, depth, and margin location.

DISCUSSION

To the best of our knowledge, this is the first systematic review to synthesize the influence of tooth and cavity-related aspects of NCCLs on the retention rate of resin composite restorations.

The results of the review suggest that the location of the tooth in the dental arch and the presence of wear facets interfere with the retention rate of resin restorations in NCCLs. In contrast, other aspects such as dentin sclerosis, shape, size, depth, occluso-gingival distance, and margin location of the cavity showed no influence on the retention rate. We set out to identify the best clinical evidence available to answer the focused question and performed an extensive search with careful quality assessment. However, as we could not find primary studies that answered our specific research question, we drew indirect conclusions by using randomized clinical trials testing different adhesive approaches for the resin restoration of NCCLs.

This type of lesion may be present in different shapes and sizes and may affect any tooth (anterior or posterior).^1,2,8,54-56 Moreover, they are considered a significant restorative challenge because they are in general nonretentive and the cervical area concentrates high stresses caused by masticatory forces.^10,57-60

Several clinical trials have attempted to study the longevity of resin restoration of NCCLs. Participants were recruited with different types of NCCLs, and the researchers usually reported the characteristics of the lesions that would be restored. However, they infrequently related restorative success with the aspect of the lesion. In our search strategy, we found 91 studies that reported the characteristics of the lesions but that did not correlate the success rate with the observed characteristics. Although this would be only a complementary analysis, without statistical power in individual studies, grouping the results in a systematic review such as the current one could provide important findings.

We found only 24 manuscripts that related some aspects with the success rate of the restorations. Although we found reports for nine different aspects, not all included articles reported all of them, making some aspects meta-analyzed with minimal data. In addition, the variability in reporting the characteristics made it difficult to group the data, and, as a result, the information was dichotomized to allow the meta-analysis. Some aspects already had a dichotomous nature, such as arch distribution (maxillary vs mandibular), tooth location (anterior vs posterior), wear facets (with vs without). However, other characteristics had more than two data possibilities, and for that, we created cutoff points to dichotomize and group the studies (Table 3).

A critical point to be considered in this systematic review is that the findings were derived from indirect evidence. The primary objective of the included studies was to compare different adhesive strategies or different resin composites for NCCL restoration, and this may have contributed to the heterogeneity found in some meta-analyses. An important factor that influences retention of NCCL restoration is the kind of adhesive system (etch & rinse or self-etch) used. However, due to the small number of included studies, the results of this study could not be controlled for this confounder. This limits the external validity and generalizability of findings.

In the analysis of risk of bias using the Cochrane tool,⁴⁷ some studies were categorized as high risk of bias in the selection bias domains. Although this domain is fundamental for the analysis of the RCTs, for this study, we do not consider it to be of primary importance, because the groups compared in the meta-analysis were not those being randomized.

Some studies^8,60-62 showed a higher prevalence of NCCLs in maxillary teeth. Those teeth seem more prone to NCCLs, possibly due to their lingual inclination.⁸ In our study, 14 studies were included, totaling 1021 restored teeth in the maxilla vs 780 restored teeth in the mandible. Although the distribution in the arch influenced the development of NCCLs, this factor did not seem to affect the restorative success of these lesions.

Regarding the tooth location in the arch, posterior teeth are more likely to present NCCLs, possibly because occlusal and nonaxial forces are exerted on that group of teeth.⁸ The referred forces could also stress the adhesive area, decreasing the longevity of the adhesive restorations, because from this meta-analysis, we found an RR of 1.08, favoring the success rate of restoration in anterior teeth. However, the meta-analysis of these data showed high heterogeneity; thus, the results should be considered with caution. Then, the quality of the evidence was downgraded due to the inconsistency.

The presence of wear facets also influenced restorative success and is strongly related to the development of NCCLs.^26,57,63-65 In a three-dimensional finite element model, some authors observed in malocclusion scenarios that tensile stresses generated on the cervical areas were higher compared with stresses generated under normal occlusion conditions, those possibly being capable of producing NCCLs.⁷ Moreover, in patients with untreated malocclusion, restoration success might also be affected, because the stress in the cervical area could be higher than the adhesive resistance of the restorations, possibly leading to restoration displacement. Therefore, the treatment of these lesions should go beyond their restoration, because if the causes of lesions persist, premature failure of the restoration might ensue.⁶⁶ The meta-analysis of this characteristic included only two articles, and despite the low heterogeneity, the quality of the evidence was downgraded as the Optimal Information Size (OIS) criterion was not met, that is, only a very small sample supported the evidence.

Different classifications for the shape of the lesions have been reported, with some formats relating to more specific causes of lesions.^{2,3,8,54-56,67} The studies included in this meta-analysis used different methods to classify the format of the lesions, which may have contributed to the heterogeneity of the data. We dichotomized in wedge and saucer formats and used an internal angle of 90° as the division between the two formats. The wedge format may be more related to occlusal stress, and, considering that occlusal stress indirectly observed on the wear facets influenced restorative success, one might suppose similar observations would be detected for the present characteristics; however, the influence of cavity format was not observed to influence restoration success. The quality of this evidence was considered very low because, in addition to the indirectness of evidence and the inconsistency (due to high heterogeneity), the reduced number of articles included with a small sample size meant the OIS criterion was not met.

As the development of NCCLs tends to be a slow and chronic process, dentin sclerosis is commonly found on the surface of the lesions.^1,5,8 Secondary dentin, the occlusion of open dentinal tubules, pulpal retreat, and other tooth protective measures slowly take place in the presence of noxious stimuli, thereby minimizing clinical symptoms and maintaining pulpal integrity. As for the shape characteristic, the different criteria for and methods of measuring dentin sclerosis are fraught with subjectivity and susceptible to inter- and intraexaminer variation with regard to features such as discoloration, smoothness, and translucency. In restorative procedures, sclerotic dentin is a substrate that can lead to adhesion difficulties, because, with some types of adhesives, hypermineralization can prevent the formation of a hybrid layer.^10,68 In this meta-analysis, there was great variability in the types of adhesives studied, which might have contributed to the high heterogeneity and lack of influence of this characteristic on restorative success.

The characteristics of size, depth, and occluso-gingival distance are all related to the dimension of the lesion. To allow the grouping of studies, we dichotomized some points. All these characteristics were shown not to affect restorative success. Evidence was downgraded to low for the three characteristics.

Only two articles reported the margin location and related it to restorative success. The high heterogeneity found allowed a wide confidence interval, overlapping the no effect line of this characteristic in restorative success. Despite the greater difficulty in isolating the operative field for subgingival lesions, this seems not to affect restorative success.^66,69,70

To minimize bias, we followed the guidelines provided in the Cochrane Handbook for Systematic Reviews of Interventions.⁴⁷ However, our findings and interpretations are limited by the quality and quantity of the available indirect evidence on the effects of tooth- and cavity-related aspects of NCCLs on the success rate of resin restorations. We assessed the risk of bias of the included trials by using the published data, which ultimately may not reflect the actual situation and may lead to publication bias. Also, to minimize publication bias, we did not restrict our search to the English language. For three aspects, we were able to evaluate publication bias by inspecting funnel plots and the Egger analysis, and no problem was detected. However, publication bias could not be determined for other studied aspects, as there was an insufficient number of studies to allow this inspection through a funnel plot. This would reflect on any conclusions drawn from this review.

A knowledge of tooth- and cavity-related aspects of NCCLs that could affect restorative success could help clinicians understand the prognosis of restorations. For future research, investigators are encouraged to collect information from RCTs regarding the tooth- and cavity-related aspects of NCCLs and relate them to success rates. In addition, future RCTs should follow the CONSORT guidelines to improve evidence quality and transparency in reporting.⁷¹

CONCLUSION

The results of the meta-analysis suggest that NCCLs restored in anterior teeth are more likely to succeed than in posterior teeth. The presence of wear facets can reduce the retention rate of resin restoration in NCCLs. Arch distribution, dentin sclerosis, shape, size, depth, occluso-gingival distance, and margin location did not affect the restorative success of resin restorations. However, the quality of evidence of the outcomes ranged from moderate to very low.