Introduction

Non-carious cervical lesions (NCCLs) are defects of the tooth tissue at the cervical region unrelated to caries. The aetiology of NCCLs is multifactorial, mainly including erosion caused by acids, toothbrush abrasion, and abfraction caused by occlusal loading [1, 2]. The mean prevalence of NCCLs was 46.7%, with a strong tendency to increase with age [3, 4]. As the ageing population continues to grow, NCCLs will become increasingly common. NCCLs should be restored when hypersensitivity, aesthetic problems, or obvious defects are engendered [5]; otherwise, they could result in caries, pulp diseases, or even tooth fracture. However, challenges always exist in restorative treatments, including saucer or wedge-shaped NCCLs that have no retention form, sclerotic dentin on the surface that is unfavourable for adhesion, and a cervical margin that is usually located subgingivally, which obstructs filling access and makes the control of moisture difficult [6,7,8].

Previous studies have reported on restorative materials and techniques for improving the clinical performance of cervical restorations [5, 8,9,10]. The retention rate of NCCLs in tooth-coloured restorations has been shown to decrease significantly over time, with an average retention rate of 85.4% after 3 years and 56.5% after 10 years [11]. Composite resins have been recognized as the first choice for the restoration of NCCLs because of their excellent aesthetic properties, improved adhesive capacity, and mechanical properties [5, 12]. However, due to the wide variety in the market, it is still controversial which type of composite resin is the most suitable material for restoring NCCLs.

Flowable composite resins introduced in the late 1990s have been popular in recent years due to their easy handling and good rheological properties [13]. Flowable composites combined with a transparent matrix have been proposed to efficiently restore NCCLs [9]. Considering the flexure in the cervical area caused by occlusal loading, it is recommended to use a flowable composite to restore NCCLs, as it is assumed to flex with the tooth because of its low elastic modulus [12]. However, no clinical trials have detected significant differences in clinical performance between flowable composites and conventional paste-type composites for NCCLs [12, 14, 15]. This may be related to the fact that flowable composites are usually produced by reducing the filler content of the composite resin formulation[16]. In vitro studies have demonstrated that the lower the filler content is, the lower the mechanical properties and the higher the polymerization shrinkage of composite resins [17,18,19]. Besides, the lower the filler content is, the easier it is for the composite surface to be affected by ageing [20]. One 2-year clinical study showed that the retention rates and marginal adaptation of low-filler loading flowable composites (filler content 38 vol%) were significantly inferior to those of conventional paste-type composites (filler content 60 vol%) in NCCLs [21].

Recently, flowable composites with a high filler content have been introduced. In addition to maintaining a low viscosity, these flowable composites have a filler content and mechanical properties comparable to those of conventional paste-type composite resins [13]. In vitro experiments have shown that these new materials exhibit lower polymerization shrinkage, higher flexural properties, and higher wear resistance than conventional paste-type composites [22, 23]. Ikeda et al. found that the highly filled flowable composite showed a similar marginal integrity and wall adaptation to the conventional paste-type composite in 2-mm-deep cavities [24]. In 2-year randomized controlled trials, the clinical performances of highly filled flowable composites were similar to that of conventional paste-type composites in class I and class II restorations [25, 26]. However, limited data exist concerning the applicability of highly filled flowable composites for NCCLs.

Clearfil Majesty ES Flow (ES, Kuraray Noritake Dental, Tokyo, Japan; filler content: 75 wt% and 62 vol%) is a novel flowable composite developed using special submicron fillers to make it highly filled compared to the conventional paste-type composite Majesty (MJ, Kuraray Noritake Dental, Tokyo, Japan; filler content: 78 wt% and 69 vol%).Thus, the purpose of this randomized, controlled, split-mouth clinical trial was to provide evidence about the performance of a highly filled flowable composite in NCCLs. The null hypothesis was that no difference would be found between the 3-year clinical performance of a highly filled flowable composite and that of a conventional paste-type composite in NCCL restorations.

Materials and methods

Study design

This study was a prospective, double-blind (volunteers and examiners), split-mouth randomized controlled clinical trial with a 3-year follow-up period. The research protocol was approved by the Institutional Review Board of Beijing Stomatological Hospital, Capital Medical University, before the enrolment of the subjects and was conducted in accordance with the 1964 Helsinki Declaration and its later amendments. This trial was reported according to CONSORT checklists and registered at the Chinese Clinical Trial Registry (ChiCTR1900028484). Written informed consent was obtained from all the participants prior to the study.

Sample size calculation

The sample size was calculated using the two-sided test formula for the paired design of counting data with an error probability of α = 5% and β = 10% (power of 0.90). The mean retention rate of cervical restorations was reported to be 90% after 3 years [27]. Thus, the minimum sample size was determined to be 35 pairs to detect a difference of 10% between ES and MJ. Considering a possible visit loss of 20%, the sample size was increased to 42 pairs. Each pair included 2 teeth; thus, the total number (n) was 84 teeth.

Participants and lesion selection

Participants were recruited and treated at Beijing Stomatological Hospital from August 2015 to July 2016.The inclusion criteria were as follows: patients aged 18 to 65 years in good general health, with acceptable oral hygiene, and presenting at least two vital teeth that were in occlusion and had NCCLs in need of restoration. The exclusion criteria were as follows: patients with a history of allergies to dental products, rampant caries, DMFT≥12, poor oral hygiene, severe periodontal disease, orthodontic appliances, and severe bruxism.

Thirty participants who met the inclusion criteria were recruited and examined to determine whether they met any of the exclusion criteria. Finally, twenty-seven subjects were enrolled in the study, and 42 pairs of teeth were included. For each subject, two teeth in different quadrants with similar positions were preferentially matched into one pair. If bleeding upon probing of the gingiva was positive, periodontal scaling was performed at least 1 week before restoration. For each pair, teeth with the smaller tooth number were filled first.

Randomization and allocation concealment

The randomization method of this trial adopted a matching design randomized grouping scheme. Prior to the start of the trial, a statistician used RandA 1.0 software to generate a random allocation table and concealed it using the envelope method. The main researcher numbered each tooth according to the treatment sequence. Just before the restorative procedure began, an assistant who was not involved in this study directly opened the sealed envelope to reveal the allocation corresponding to the tooth number.

Restorative procedures

All restorations were performed by a single experienced and trained specialist. NCCLs were cleaned with a suspension of pumice and water. The surface of the lesions was roughened slightly and intermittently with a pear-shaped, coarse diamond bur (EX-41, MANI, Japan) at high speed under water cooling. No bevels or retentive grooves were made. A shade selection guide was used to determine the proper shade. The operative field was isolated with cheek retractors, cotton rolls, and a saliva suction device. Before restoration, a gingival retraction cord (#000 or #00, Ultrapak Cord, Ultradent, South Jordan, UT, USA) was inserted into the gingival sulcus to separate and expose the lesion.

Information and instructions (provided by the manufacturer) for all materials used in this study are shown in Table 1. The two-step self-etching adhesive Clearfil SE Bond (Kuraray Noritake Dental, Tokyo, Japan) was used according to the manufacturer’s instructions. Primer was applied with a disposable brush tip to the surface of the entire lesion for 20 s and then dried with gentle air flow. The adhesive was applied in frictional mode and gently blown evenly with mild airflow. Then, the adhesive was light-cured for 10 s at 800 mW/cm2 with an LED light-curing unit (MiniLED PEF004, Satelec, Merignac, France).

Table 1 Information of materials used in the study
Full size table

A transparent cervical matrix (Hawe Transparent Cervical Matrices, KaVo Kerr, Bioggio, Switzerland) was chosen according to the size of the lesion. For shallow lesions, both composites were placed within a single increment and contoured by a transparent cervical matrix. Then, the excess composite was removed, and the remaining composite was light-cured through the matrix. During this process, the gingival side of the transparent matrix was inserted into the gingival sulcus and firmly attached to the neck of the tooth to ensure a smooth and continuous margin at the gingival wall. For deep lesions, both composites were placed using an incremental technique. The gingival wall was restored first with the help of the matrix as described above. Then, the matrix was removed, and the occlusal wall was restored and contoured with a hand instrument. Each increment of 2 mm was cured for 20 s.

After restoration, trimming and finishing were performed using a fine-grained diamond tip (TC-21F, MANI, Japan) and 12-fluted conical carbide burs (FG7613, BluWhite Carbide Burs, KaVo Kerr, Ontario, Canada). Finally, polishing was performed with silicone tips (OneGloss, Shofu, Kyoto, Japan) and silicon-impregnated brushes (Occlubrush, KaVo Kerr, Bioggio, Switzerland).

Clinical examination

A clinical examination was performed at baseline (BL), 1 year, 2 years, and 3 years by one examiner who was an experienced specialist and was blinded to all assessments. The examiner was trained and calibrated together with the operator before the trial; they observed 10 photographs that were representative of each score for each criterion and evaluated 20 patients with cervical restorations. An intra- and inter-evaluation agreement of 90% was required before the assessment. The FDI (World Dental Federation) criteria and scoring system were used for assessing the restorations (Table 2) [28]. The following criteria were selected for evaluation: surface lustre (A1), surface staining (A2a), marginal staining (A2b), aesthetic anatomical form (A4), fracture of the material and retention (B5), marginal adaptation (B6), and occurrence of caries (C12).

Table 2 Category and score descriptions for the clinical evaluation according to FDI criteria
Full size table

Statistical analysis

The two types of composites were compared at every assessment to detect differences between their clinical performances. A paired chi-squared test was used to compare the clinically acceptable rate (FDI score = 1, 2, 3) for each category, and the Wilcoxon signed-rank test was used to analyse the difference in FDI score distribution between the two types of composites.

Changes in clinical performance over time within groups for all criteria were analysed by the Friedman test, and multiple comparisons between the periods of evaluation within groups were conducted using the Wilcoxon signed-rank test. All tests were performed at a significance level of α = 0.05 using the SPSS statistical software package (version 20.0).

Results

Base information

A total of 84 teeth with NCCLs from 27 subjects were filled with MJ and ES in this clinical trial. The distribution of subjects and NCCLs are summarized and shown in Tables 3 and 4, respectively. The subjects were mainly between 31 and 60 years old. NCCLs were mainly localized in the premolars (59.5%) and evenly distributed in the upper and lower dentition. Figure 1 represents the study flow chart. The recall rates at BL and 1, 2, and 3 years were 100%, 95.2%, 92.9%, and 88.1%, respectively. One subject with 4 restorations was lost at 1 year of follow-up, one subject with 2 restorations was lost at 2 years of follow-up, and one subject with 4 restorations was lost at 3 years of follow-up due to moving to another city or work. At the 3-year assessment, one failure (loss of restoration) occurred in both groups; thus, 72 restorations were evaluated.

Table 3 Distribution of the restorations according to the teeth location of NCCLS
Full size table
Table 4 Distribution of participants and restorations according to gender and age
Full size table
Fig. 1
figure 1

CNSORT flow chart of the study

Full size image

Overall analysis

Table 5 presents the evaluation results of the restorations according to FDI criteria from BL to the 3-year assessment.

Table 5 Evaluation results (numbers and percentage) according to FDI criteria for restorations at baseline (BL) and at the 1-year, 2-year, and 3-year recall
Full size table

Acceptable rate

No significant differences in the acceptable scores (FDI scores 1–3) were found between MJ and ES restorations for any evaluation category at any assessment point (p > 0.05). At BL and the 1-year assessment, all pairs of restorations were found to be clinically acceptable (100%). At the 2-year assessment, the acceptable rate of MJ restorations in aesthetic anatomical form and marginal adaptation decreased slightly (97.4% and 97.5%, respectively), lasting until the 3-year assessment (97.2% for both). At the 3-year assessment, 94.4% of the marginal staining was acceptable.

Retention rate

From BL to the 2-year assessment, all restorations were clinically acceptable (100%) without debonding. At the 3-year assessment, one restoration was lost in each group. Thus, the cumulative retention rate of both materials declined to 97.3%.

Recurrence of caries

None of the restorations showed secondary caries.

Aesthetic properties

The aesthetic property scores, including surface lustre, surface staining, marginal staining, and aesthetic anatomical form, increased over time within each group (p < 0.01). At the 1-year assessment, the surface lustre of ES was significantly better than that of MJ (p < 0.05), and no differences in the other aesthetic properties were observed. At the 2-year and 3-year assessments, significant differences were found between ES and MJ restorations for marginal staining (p < 0.05) in favour of ES, and no differences were observed among the other aesthetic properties.

Functional properties

With respect to the fracture of materials and retention, no significant differences were found between ES and MJ restorations at any examination time point. At the 3-year assessment, one restoration was lost in each group, and it was a mandibular premolar from different subjects. Regarding changes over time, significant differences were found within each group (p < 0.01).

Regarding marginal adaptation, performance declined over time within each group (p < 0.01). The scores of ES were better than those of MJ at all assessment points, and the results were significantly different (p < 0.05).

Discussion

Randomized controlled trials are one of the most convincing types of evidence-based studies in dentistry [29]. The current study presents a prospective, randomized, double-blind, split-mouth clinical trial for evaluating the clinical performance of a highly filled flowable composite in NCCLs, and it was designed and reported in strict accordance with CONSORT guidelines. In previous studies, most clinical trials used United States Public Health Service (USPHS) criteria and had a 2-year follow-up on average [30]. However, in our study, the FDI criteria were used to assess restorations because they are more precise for evaluating NCCLs and “marginal adaptation” [31]. Furthermore, our follow-up time was 3 years.

From BL to the 3-year assessment, the clinical acceptance of the highly filled flowable composite was similar to that of the conventional paste-type composite for all criteria in NCCLs. Therefore, the null hypothesis was accepted. In the present study, no difference was found in the surface lustre (p = 1.00), surface staining (p = 1.00), marginal staining (p = 1.00), aesthetic anatomical form (p = 1.00), fracture of material and retention (p > 0.05), marginal adaptation (p = 1.00), or recurrence of caries (p = 1.00) between materials. Both materials showed a decrease in perfect scores for all criteria over time except for secondary caries. These results are consistent with recent clinical studies evaluating the clinical performance of flowable composites in cervical restorations [14, 32, 33] and class I [25] and class II restorations [26] for up to 3 years.

In the present study, one restoration from each group was lost at the 3-year assessment, and the retention rate of both materials was 97.4%. This result was in accordance with two 3-year clinical trials in which the retention rate of flowable composite resin in NCCLs was 94.0% and 95.8%, respectively [14, 34]. Few studies exist about the clinical performance of highly filled flowable composites in NCCLs. One study published in 2010 reported that the retention rate for highly filled flowable composite resin (filler content: 80.2% by weight) with a one-step self-etching adhesive was only 54.0% after 24 months, and the retention rate for nanohybrid composite resin was 60.0% in NCCLs [32]. This difference might be related to the bonding strategy.

A two-step self-etching adhesive, Clearfil SE Bond, was used in this study. Clearfil SE Bond showed the best bonding performance in NCCLs, and it is regarded as the gold standard material for dentin bonding systems [27]. The results of an in vitro study showed that the tensile strength of the SE bond to enamel and dentin was equivalent to that of the etch-and-rinse system [35]. Clearfil SE Bond showed better long-term retention rates than a two-step etch-and-rinse system in NCCLs [36,37,38]. This excellent bonding performance may be related to the presence of the functional monomer 10-MDP, which bonds chemically with hydroxyapatite (HAP) through its phosphate groups, providing a more effective bond and more stability in water than other monomers [38, 39].

On the other hand, the dentin surface of NCCLs was roughened, which may be one of the reasons for the high retention rate in this study [27]. Generally, hypermineralized sclerotic dentin occurs on the surface of NCCLs [6, 7]. The surface is smooth, and the dentinal tubules are sealed by minerals, which block the penetration of adhesive and affect the adequate establishment of a hybrid layer [7, 10]. In vitro experiments have confirmed that the bond strength of non-carious sclerotic dentine is significantly lower than that of sound dentine [40, 41]. Removing the hypermineralized surface layers with a bur improves micromechanical retention in sclerotic dentin and the retention rates of composite resin restorations in NCCLs [7, 10]. Similar high success percentages were observed in previous studies of NCCLs that roughened the surface during tooth preparation [14, 34]. Moreover, regardless of the presence or absence of sclerotic dentin, Van Dijken reported that roughening the tooth surface prior to adhesive applications improved the retention rates of restorations in NCCLs [36].

Although both composites demonstrated highly acceptable marginal adaptation from BL to 3 years (ES: 100% at each assessment; MJ: 100%, 100%, 97.4%, and 97.2% at each assessment, respectively), the marginal adaptation of ES was significantly better than that of MJ (p < 0.05) at each assessment in this study. In a systematic review that analysed eight clinical trials, the marginal adaptation of flowable composites may have been better, but the evidence for this conclusion was not sufficient [15]. To our knowledge, this was the first clinical observation of flowable composite resin with a high filler content showing better marginal adaptation, which may be related to its low viscosity. According to the manufacturer’s information, except for the viscosity, most properties of the two tested materials are similar, including the polymerization shrinkage (MJ: 1.9 vol%, ES: 1.9 vol%) and elastic modulus (MJ: 10 GPa, ES: 9.6 GPa). ES is a low-viscosity restorative material that differs from paste-type MJ by having a patented filler and less viscous resin content (Table 1). Despite the similar filler contents of MJ and ES, the filler of ES is a patented filler that is surface-treated with a proprietary silane coupling agent (MUS) to enable effective silanization to occur and make it highly filled and easy to polish, according to the manufacturer. Furthermore, unlike the Bis-GMA contained in MJ, the main monomer of ES is TEGDMA, a small molecule with no aromatic cycle or hydroxyl radicals, which helps to reduce the viscosity [42]. The lower the composite viscosity is, the higher its wettability is, and the larger the free surface formation, thus representing a smaller restriction to shrinkage and resulting in the reduction of shrinkage stress [19]. Two studies reported a negative correlation between viscous flow and gap formation in vitro [43, 44]. In another in vitro study, flowable composites were confirmed to reduce the cervical microleakage of class II restorations, and a significant correlation was found between its viscosity and microleakage [45]. In the initial stage of light-curing, polymerization shrinkage of the resin composite is counteracted by plastic flow. The higher the plastic flow is, the longer the resin composite can counteract shrinkage and bear gap formation and the smaller the gap formed [43]. Thus, with the same polymerization shrinkage, a flowable composite would result in less gap formation than a paste-type resin composite, which can also explain the better marginal adaptation outcome of ES in this study.

In addition, the better marginal adaptation outcome of flowable composite resin may be related to the application of a transparent cervical matrix. The gingival margin of NCCLs is well known to be difficult to fill perfectly, especially when it is near or under the gingiva. In previous clinical studies, retraction cords or rubber dams were usually introduced to avoid contamination, and hand instruments were sometimes used to contour composite resin before curing [14]; however, detailed restorative techniques on the gingival margin have rarely been introduced. To the best of our knowledge, only one 20-year retrospective evaluation of compomer restorations in NCCLs introduced the usage of a transparent cervical matrix during the restoration procedure [46]. In this study, a transparent cervical matrix was firmly attached to the tooth after composites were placed or injected into the gingival wall to assist in isolation from the gingiva and shaping. Because a low-viscosity composite can flow into irregularities as well as the margins, it may be more adaptable than a high-viscosity composite with fewer marginal defects. Similar to the experience shared by Giachetti [8] and Perez [9], using a matrix properly can improve gingival marginal continuity, provide good cervical contour, and reduce the finishing and polishing steps.

For both test materials, a significant increase in the number of scores of 2 and 3 was observed over time regarding the aesthetic properties of surface lustre, surface staining, marginal staining, and aesthetic anatomical form. These deteriorations observed over time might be related to the effects of complex oral environmental factors, such as physical stresses, chewing and dietary habits, and changes in temperature and pH. In an in vitro study, with exposure to acids and alcohol, surface degradation phenomena (matrix resin decomposition and fillers falling out) were observed in composite resin [20]. When immersed in a beverage for 1 week, the colour changes of the resin-based materials were clinically visible to varying degrees [47]. At the 2-year and 3-year follow-ups, ES showed significantly less marginal staining than MJ (p < 0.05). At the 3-year assessment, 58.3% of the ES restorations had a score of 2, whereas 63.8% of the MJ restorations had a score of 3. Marginal staining is usually caused by the accumulation of stains at the marginal steps or crevices in NCCLs [14]. As is often described in the literature, marginal discolouration is correlated with marginal adaptation [38, 48], which may be the reason for less marginal staining in ES restorations. However, these superficial and acceptable discolourations can be easily removed by repolishing, which has been recommended to be performed at each regular check-up.

In the present study, the restorations were observed for only 3 years. Further evaluations are needed to obtain the long-term performance of highly filled flowable composites.

Conclusions

This 3-year clinical trial provided the following conclusions.

  1. 1.

    Both highly filled flowable composites and conventional paste-type composites showed high retention rates and similar acceptable clinical performance for the restoration of NCCLs.

  2. 2.

    Obvious changes in aesthetic and functional criteria were observed over time in both composite resins. However, most changes were clinically acceptable.

  3. 3.

    The highly filled flowable composite had a better marginal adaptation score than the conventional paste-type composite.