2.1

Ethical aspects

This study was approved by the Local Ethics Committee (protocol 210/2011). It was registered at clinicaltrials.gov (NCT01947192), and followed the CONSORT guidelines . Prior to the participation in the study, all participants signed a written informed consent.

2.2

Study design

This study is a 36-month follow-up of a prospective randomized clinical trial, designed as split-mouth and triple-blind (operators, patients and clinical evaluator) . The teeth with noncarious cervical lesions (NCCL) were randomly divided into two treatment groups: test group (application of CHX 2%) or control group (application of a placebo solution), both applied after etching and before the adhesive system. Ten operators (undergraduate students from the last year of the School of Dentistry, Federal University of Pelotas) placed the restorations in 2011 and 2012 supervised by two researches (AFM and MSC).

2.3

Operators’ training

The operators were previously trained according theoretical and practical aspects to ensure the standardization of the clinical procedures and minimize variations inherent to different operators. During the theoretical phase, lectures were given and pre-clinical activities performing class V restorations were carried out. During the practical phase, each operator performed the adhesive and restorative procedures in volunteers (10% of the sample) not included in this study.

2.4

Sample size

Taking into account a 92.3% retention percentage after 36 months to restorations of NCCL placed with the adhesive system Adper Single Bond 2 , the calculation of the size of the sample was based on a 20% difference in retention rates between groups at a significance level of 5%, with a power of 80%, resulting in a sample of 35 patients in a split-mouth design. Considering the dropout rate throughout the trial period, the need for 40 patients was considered.

2.5

Patient’s recruitment

Subjects were recruited by the examination of patients under treatment at the School of Dentistry and by advertisements posters set in School. All those who needed dental treatment of NCCL and fulfilled the inclusion and exclusion criteria ( Table 1 ) were invited to participate. Dental treatments were carried out in a dental clinic specially organized for the treatment of NCCL at the School of Dentistry, Federal University of Pelotas, to obtain the sample size required. All dental needs of the subjects enrolled in this study, except for prosthetic rehabilitation and orthodontic treatment, were provided.

2.6

NCCL screening

The selection of the NCCL was performed using a mouth mirror, an explorer and a periodontal millimetre probe by two researches. Reasons for treatment were cervical tooth sensitivity, aesthetic complaints and/or prevention of further tooth damage. The depth of NCCL was measured by placing the probe in its deepest part, while the height was calculated by the distance between the most coronal to the most apical point of the cavity margins. The degrees of dentinal sclerosis were identified using a rating scale of 1 up to 4 . The sensitivity test was measured by compressed air for 3 s at a distance of 2–3 cm, while the adjacent teeth were protected with the fingers. Tooth vitality was tested using an Endo-frost cold spray on the tooth. No attempt was made to determine the aetiology of cervical lesions.

The teeth to be restored had a normal occlusal relationship with natural dentition and at least one adjacent tooth contact; they were vital and had a positive response to cold. Cavosurface angle were not bevelled and no retentive grooves were performed.

2.7

Study groups

All volunteers who met the eligibility criteria were randomly divided into experimental or control groups, where the 2% CHX (experimental) or a placebo solution (control) were applied to the dentine after acid etching procedure.

2.8

Randomization and blinding procedures

A person not directly involved in the study performed the randomization process using a computer programme (Microsoft Excel, 2010). A table with random numbers was used to allocate the NCCL in each treatment group. The treatment (experimental or control) was allocated regarding the tooth group (incisors, canines and premolars), where the first tooth restored of each tooth-group was raffled for one treatment, while the next tooth from the same tooth-group was automatically assigned to the other treatment, according to the split-mouth design. Thus, after being randomly assigned, each patient received the same number of restorations of both groups. Each operator performed the same number of restorations for both groups.

Individual opaque sealed envelopes were used to conceal the randomization sequence according the treatment groups, which were coded as Treatment A or Treatment B. This condition enabled the blinding of operators and patients because the clinical procedure was the same for both treatment groups.

2.9

Clinical procedures

The clinical protocol (the same for experimental and control groups) was printed and posted in each dental unit; thus, the operator was able to easily review it. There was no making bevel at cavosuperficial angle or any cavity preparation before restorative procedures. Prophylaxis was done and the tooth shade was selected using a shade guide (VitaPan Classic, Vita Zahnfabrik, Bad Sackingen, Germany) before bonding procedures. Where necessary, local anaesthesia was taken. All procedures were carried out using relative isolation method with labial retractors, cotton rolls, saliva aspirator and gingival retraction cord (Cord # 000 Ultrapak, Ultradent, South Jordan, UT, USA) into the gingival sulcus.

The etching was performed with 35% phosphoric acid (Adper Scotchbond Etchant, 3M ESPE) applied for 20 s in enamel and 15 s in dentine, followed by rinsing with air/water spray for 30 s and removal of moisture excess from cavity using absorbent paper. Afterwards, (1) for the experimental group: 2% CHX was applied (manipulated solution; Intended Use Pharmacy, Pelotas, Brazil) under scrubbing action for 60 s using a disposable applicator and the solution excess was removed. For the control group (2): a placebo solution (solution similar to that used for experimental group, but without CHX) was applied on the cavity at the same way as for experimental group. The application of a two-step etch-and-rinse adhesive system (Single Bond 2 3M ESPE, St. Paul, MN, USA) was performed according the manufacture’s instructions and light-cured for 10 s using LED light-curing unit (Radii-Call; SDI, Bayswater, VI, Australia), intensity of 800 mW/cm ² . The NCCL were restored with a direct restorative nanocomposite resin (Filtek Z350, 3M ESPE, Irvine, CA, USA) applied in at least two increments (no more than 2 mm thick), using a selected instrument (Hu-Friedy, Chicago, IL, USA). Each composite increment was light cured for 20 s.

Final contouring and polishing of the restorations were performed using a fine diamond burs and ultra-thin (KG Sorensen, Barueri, SP, Brazil) under refrigeration and low speed, floppy disks (Sof-Lex Pop-On, 3M ESPE), polishing paste (Diamond Excel, Dental Products FGM, Joinville, Brazil) and rubber tips (Enhance; Dentsply Caulk, Milford, DE, USA).

2.10

Clinical assessment

Criteria approved by the FDI World Dental Federation were used for clinical evaluation of restorations . The primary clinical outcome was the retention of the restoration, considering as failure the complete loss of restoration. Secondary outcomes included the criteria as follows: (1) marginal adaptation, (2) marginal staining, (3) the colour of the surface, (4) post-operative sensitivity, (5) surface brightness, (6) translucency and colour, (7) fracture, (8) anatomical shape and (9) preserving the vitality and integrity of teeth. Each criteria was expressed in five scores, three for acceptable and two for non-acceptable (repair or replacement).

An experienced, trained and calibrated examiner conducted the clinical evaluations. The examiner was blind to the intervention and was not involved in allocations or in restorative procedures. A Web-based tool ( www.ecalib.info ) and clinical setting evaluation were used for training and calibration of the examiner. The clinical intra-examiner calibration was carried out with 30 NCCL restorations, which were re-examined 15 days later. A pre-evaluation intra-agreement of at least 90% was obtained.

2.11

Recalls

At each recall, the examiner evaluated the restorations blindly. Revaluations were made at baseline (1 week) and for periods of 6 months, 12 months, and 24 months after 36 months. Considering that some patients took more time to attend to the recalls the time of follow up extended for more than 36 months in some cases.

Telephone contact was made with the patients to recall them for the periodic assessment of the restorations. Letter was sent to the home address identified in the records for those who were not touched by phone call after several attempts. In cases where there was no return via phone contact, letters or e-mail, a home visit was made and the researchers (MSC and MF) carried out the examination of the patients in their own homes. Individuals, whose contact was not possible, were considered losses.

2.12

Statistical analysis

A researcher not directly involved in the study performed the statistical analysis of the data. The analysis was carried using Stata 11.0 Statistic Programme (Stata Corp LP, College Station, TX, USA), considering 80% statistic power and 5% significance level. Data were subjected to survival analysis using the Kaplan–Meier method. The log-rank test was used to evaluate the existence of differences between the survival curves. Crude Cox regression models with shared frailty were used to verify the association between treatment and the risk of failure over time, estimating the Hazard Ratios (HR) and 95% confidence intervals.