Effect of dentin roughening and type of composite material on the restoration of non-carious cervical lesions: an in vivo study with 18 months of follow-up

Study design

This prospective, double-blind (that is, with blinding of both patients and calibrated examiners), split-mouth, randomized clinical trial was conducted in the Department of Conservative Dentistry and Endodontics, in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines. The protocol for the present study was reviewed and approved by the Institutional Ethics Committee of JCD Dental College (JCDV/DC/19/1740). The clinical trial was registered in the clinical registry (www.ctri.nic.in) under the registration number CTRI/2021/08/035863. All participants provided written informed consent after being fully informed about the nature, objectives, risks, benefits, and alternatives of the clinical procedure. This included the option to withdraw from the trial at any stage without repercussions.

Patient selection

Patients were randomly selected from the regular outpatient clinic of the Department of Conservative Dentistry and Endodontics. Evaluations were performed using a mouth mirror, an explorer, and a periodontal probe. Two trained postgraduate students screened the patients based on the following criteria.

Regarding inclusion criteria, the study included patients who provided informed consent, were in good health, were over 18 years old, and had a minimum of 4 NCCLs (in 4 different teeth) that required restoration. Additionally, these patients had to have acceptable oral hygiene, as indicated by a score of 0.7 to 1.6 on the Oral Hygiene Index-Simplified, and at least 20 teeth in occlusion. The lesions in question needed to be non-retentive, to be deeper than 2 mm, and to involve both the enamel and dentin of vital teeth without any mobility.

For exclusion criteria, this study did not include patients who had grossly carious teeth or teeth that served as abutments for fixed or removable prostheses. Additionally, individuals with extremely poor oral hygiene, severe or chronic periodontitis, or parafunctional habits such as heavy bruxism were excluded. Those with a history of periodontal surgery, who were currently undergoing orthodontic treatment, or who did not provide informed consent were also excluded from the study.

In total, 550 patients were analyzed. Of these, 36 patients were selected in accordance with the inclusion criteria.

Sample size determination

The sample size for the current study was determined based on data derived from a previous study conducted by Loguercio et al. [6]. The data were analyzed using G*Power (version 3.1.9.7; University of Düsseldorf, Düsseldorf, Germany). To maintain the study’s power at 80% and the level of significance at 5%, a minimum sample size of 32 patients was calculated as adequate. Considering a potential dropout rate of 10%, the total sample size for the study was set at 36.

Randomization

Randomization was conducted on an intra-individual basis. That is, for each patient, 4 teeth were selected for the study. The type of restoration (BFF composite or NHC) and procedure (dentin roughening) to be performed were determined randomly. This randomization was carried out using the chit method. The groups were indicated on 4 cards, and for each tooth in every patient, a single card was drawn. The treatment was then performed according to the group indicated on the card. Both the patients and the 2 evaluators were kept unaware of the roughening procedure and the type of composite material assigned to each tooth. However, operator blinding was not possible, as all procedures were performed by a single operator. The 4 groups into which the teeth were randomly divided were as follows:

Clinical procedure

Follow-up

Final polishing was completed 1 week after restoration. The baseline analysis of these restorations was performed and recorded using FDI criteria [24], as assessed by 2 calibrated examiners. Patients were subsequently recalled at intervals of 6, 12, and 18 months for further analysis based on the FDI criteria (Figure 1).

Figure 1

Clinical photographs of a patient having cervical lesions in #23, #24, #25, and #34 (FDI notation) during study period. Representative photographs taken (A) preoperatively and (B) at baseline (immediately postoperative), (C) 6 months, (D) 12 months, and (E) 18 months postoperatively.

Examiner calibration

Two experienced examiners, each with over 5 years of experience, were calibrated using images of NCCL restorations that were unrelated to this study. The agreement between examiners was evaluated using the kappa statistic. Following calibration, both examiners evaluated the study restorations at baseline and at each subsequent follow-up point. At all times, the examiners were blinded to the study group to which each tooth belonged. The scores for each restoration were compared at the conclusion of the examination session. In instances of disagreement, a consensus score was reached by the examiners and recorded as the final score.

Statistical analysis

The data were analyzed using SPSS version 21 (IBM Corp., Armonk, NY, USA). All variables were categorical; as such, they were summarized using frequency and mean (standard deviation). Given that all types of test interventions were administered to all study participants, the Friedman test was employed to compare the 4 groups. Inferential statistics were also conducted using the Friedman test. The threshold for statistical significance was established at 0.05.

Effect of restorative material

In group I and group II of this study, dentin roughening was not performed; the only difference between those groups was the restorative material used. Only 2 in vivo studies, conducted by Vildósola et al. [19] and Canali et al. [20], have compared bulk-fill composites and NHCs in the restoration of NCCLs. These researchers found the performance of both BFF composite and NHC to be clinically acceptable, with no significant differences in the examined parameters. This finding aligns with the results of our study. However, our results contradict the in vitro study by Correia et al. [21], which found reduced gap formation with BFF composites. In our study, the criteria related to gap formation, such as marginal staining, marginal adaptation, and caries, did not significantly differ between the 2 groups. This discrepancy in results can be attributed to the different design (in vitro), methodology, and assessment technique (stereomicroscope with ×50 magnification) used by Correia et al. [21]. Moreover, the cavity margins in the study by Correia et al. [21] were located exclusively in enamel.

Kemp-Scholte and Davidson [17], as well as Ichim et al. [18], have underscored the importance of a low modulus of elasticity in class V restorations. However, in the present study, no significant difference was observed in any criteria, including retention rate and marginal integrity, between the BFF composite and NHC groups (with or without dentin roughening), despite the differing elastic modulus of the NHC (20.4 GPa) and BFF (6-7 GPa) groups. These findings align with those reported in studies by Peumans et al. [25], van Dijken and Pallesen [26], and Kubo et al. [27], in which no significant differences were observed in the clinical performance of composite materials with varying stiffness values (moduli of elasticity). One potential factor that could account for the lack of difference in the clinical performance of these materials is the selection criteria used in our study. Although studies [2829] have demonstrated a positive correlation between parafunctional habits and the presence of NCCLs, we excluded patients with parafunctional habits. As a result, a relatively large number of non-abfraction NCCLs were present in our study.

Effect of dentin roughening

Meta-analyses conducted by Heintze et al. [3], Rocha et al. [10], and Mahn et al. [11] have recommended dentin roughening to enhance the durability of restorations. However, in our observations, dentin roughening did not significantly affect the outcome, regardless of whether the restorative material was the same (as in group I versus III and group II versus IV) or different (as in group III versus IV, group I versus IV, and group II versus III). Both Mahn et al. [11] and Rocha et al. [10] based their conclusions on a variety of heterogeneous studies that employed different assessment criteria, adhesive systems, and techniques. Furthermore, Heintze et al. [3] did not analyze studies based on clinical parameters such as dentin/enamel preparation, and some of their included studies even involved class III restorations. In the present study, most of the included cases exhibited mild sclerosis (grade 1 or 2). Therefore, the roughening of dentin may not have produced the anticipated results.

Interestingly, even the significantly greater number of cases of sclerotic dentin (grade 3) in group I did not appear to influence the results. These findings are consistent with the studies conducted by van Dijken [78], which indicated no significant effect of dentin roughening. This lack of effect could be attributed to the fact that dentin roughening removes only the superficial layers of hyper-mineralized dentin, leaving the deeper layers untouched. Furthermore, dentin roughening may produce a smear layer that could potentially jeopardize the formation of the hybrid layer, as it is resistant to acid dissolution [30]. Other potential obstacles to bonding in NCCLs, such as a partially mineralized surface bacterial layer and intra-tubular mineral casts, are unaffected by dentin roughening [313233].

Role of universal adhesive

The use of universal adhesive may also have influenced the outcomes of the present study. Loguercio et al. [6] recently concluded that restorations implemented with universal adhesive demonstrated comparable clinical performance, irrespective of the preparation of the dental substrate of NCCLs. Apparently, newer single-component dentin adhesive systems, including universal adhesives, have enhanced the bond strength to dentin to a degree sufficient to counterbalance other factors, such as dentin modification or differences in the elastic moduli of composite materials. Numerous studies employing single-component systems or universal adhesives have reported negligible or no retention loss [252627343536].

We utilized a universal adhesive in self-etch mode with the aim of reducing the number of clinical steps involved in restoration, thereby decreasing the likelihood of clinical errors. Research has indicated that self-etch mode is superior to total etch [37] or comparable to selective etch in terms of clinical performance [38]; other findings have indicated no significant effect of bonding strategy [39]. Notably, however, conflicting results are present in the literature. For instance, de Paris Matos et al. [40] found that the etch-and-rinse strategy was superior to the self-etch strategy.

Fracture and retention

The most important parameter for assessing NCCL restorations is the retention rate, since if the restorations are lost, no other variables can be evaluated [6]. In our study, no statistically significant difference was observed across all groups in fracture and retention at the 18-month mark. The loss of retention can be attributed to a variety of factors, including polymerization shrinkage, thermal changes, oral hygiene, patient age, the restorative material used, the location of the restoration, and occlusal forces [33]. Furthermore, the bonding agent and restorative materials may undergo hydrolytic degradation. In our study, the total loss of retention was 7% (10 of 144 restorations) over an 18-month period. This is higher than the rates reported by the studies of Vildósola et al. [19] and Canali et al. [20], which were 0% and 2.3%, respectively. The longer follow-up period in our study could explain the higher retention loss observed. While the studies by Vildósola et al. [19] and Canali et al. [20] had follow-up periods of 6 and 12 months respectively, our study had a follow-up period of 18 months. Various studies have demonstrated that retention loss increases when restorations are evaluated over longer follow-up periods [1138]. Another potential reason for the higher failure rate in the present study could be the technical aspects related to bonding and the operator’s experience, as our study was conducted by a postgraduate student.

In the present study, we employed the FDI criteria for evaluation, as in other comparable studies [619]. While many previous studies, and even some contemporary ones, continue to use the United States Public Health Service (USPHS) or modified USPHS criteria, the FDI criteria offer a broader categorization with a greater number of evaluated parameters. This makes these standards more sensitive in detecting potential differences in the various clinical characteristics of a restoration [820212540].

In the present study, an 18-month follow-up period was utilized. The follow-up duration in comparable studies has typically been shorter, ranging from 6 months to 1 year [1920]. However, our follow-up period could have been extended further, as composite restorations are anticipated to last for a longer period (years) within the oral cavity, and signs of deterioration may take time to manifest. Due to the constraints of the postgraduate period, however, the follow-up period was limited to 18 months in this study.

To maintain uniformity in clinical and other conditions across patients in the study, a modified split-mouth design was employed. This design required at least 4 restorations (1 from each group) to be performed for each patient. This approach helped to ensure that patient-related factors such as occlusal forces, salivary flow, and oral hygiene remained consistent across restorations. Consequently, the likelihood of patient and oral environment-related factors influencing the study outcome was reduced. Furthermore, any loss to follow-up did not result in unbalanced data, as 1 restoration was then omitted from each group. Notably, however, even a single loss to follow-up led to the loss of 4 restorations, thereby diminishing the statistical power.

This study was limited by a small sample size, primarily due to the challenge of locating patients with at least 4 NCCLs. One week prior to the procedure, and again immediately before its commencement, the teeth were cleaned with a pumice slurry. This was done to remove plaque and cleanse the tooth, a practice consistent with other studies [37]. The possibility of some degree of dentin roughening during this procedure cannot be entirely discounted. Additionally, most cases included in the study exhibited mild sclerosis. Patients with severe bruxism and parafunctional habits were not included, resulting in a predominance of non-abfraction cases. The attrition rate (loss to follow-up) in this study was high, a circumstance that was unavoidable due to the lockdown conditions imposed by the 2019 coronavirus disease pandemic.

Future clinical studies that incorporate larger sample sizes, extended follow-up periods, and the use of varied adhesive systems are necessary to derive meaningful clinical conclusions. It may be intriguing to conduct a specific study on sclerotic NCCLs or abfraction lesions.