2.1

Participants

Twenty-eight patients were consecutively enrolled in 5 private dental offices situated in the Paris region in France.

2.1.1

Inclusion and exclusion criteria

Patients had to be affiliated to a social welfare organization, be 18 or older, have an acceptable oral hygiene level and present at least two cervical erosions to be restored on two different teeth. The lesion had to be non-carious, more than 1 mm deep and involve both the enamel and dentin of a vital incisor, canine or premolar without mobility. All patients had to sign a written consent before being included. Patients with a compromised medical history, a severe or active periodontal or carious disease and heavy bruxism or a traumatic occlusion were excluded from the study.

Exclusion criteria after enrolment consisted of the modification of the restoration by another dentist and of the loss or fracture of the tooth for reasons not connected with the restoration.

2.1.2

Settings and locations where the data were collected

The data were collected in Toulouse (France) at the dental university by the study’s clinical research associate (Vergnes JN).

2.1.3

Eligibility criteria for centers and those performing the interventions

To participate in the study, dentists had to be part of the Paris Practice-based Research network, to have at least 3 years of experience, a relationship with the dental university and basic clinical research knowledge. Dentists also had to work in a dental practice and have no plan to move in the next 2 years.

2.2

Interventions: restorative procedure

2.2.1

Precise details of the interventions intended for each group and how and when they were actually administered

Operative procedures were performed by five specially instructed and experienced dentists in their own private practice.

The new adhesive had been sent to them in June 2007 so that they could get used to it. The protocol and the Case Report Form (CRF) were written in details during the summer. They were sent to the operators so that they could evaluate the protocol’s feasibility. A meeting took place in September, and operators’ remarks were taken into account. The protocol was finalized and sent to the five dentists. Inclusion started on September 5th in 2007 and ended on November 27th.

2.2.2

Details of how the interventions were standardized

The interventions were standardized by a detailed protocol, which is briefly summarized below.

Local anesthesia was applied if needed to prevent patient discomfort during intervention. A dry operative field was obtained using cotton rolls and aspiration by a dental assistant. If necessary, a retraction cord was used.

A preliminary cleaning of the tooth surface with pumice and water aimed to remove the salivary pellicle and any remaining dental plaque. Then sclerotic dentin and/or discolored tooth tissue was removed with a red diamond bur, and a short enamel bevel (1–2 mm) was prepared.

The various materials’ composition and application procedure are detailed in Table 1 .

In the control group, the adhesive was applied strictly following a self-etch approach on both enamel and dentin according to manufacturer’s instructions: after rinsing and moderate drying, the adhesive was applied for 20 s by brushing the cavity surface. An air spray was then applied on the adhesive from a distance for 5 s and nearer for another 5 s. Polymerization was obtained with a light output not less than 550 mW/cm ² .

In the experimental group, the application of the same adhesive followed the same protocol, but the adhesive was applied after selective 40% phosphoric acid-etching of the enamel cavity margins, followed by rinsing and gentle air drying.

The cervical tooth anatomy was restored by applying a first increment of flowable resin composite (Estelite Flow Quick/Tokuyama), followed by a second increment of micro-hybrid resin composite (Estelite Sigma/Tokuyama). Both increments were light-polymerized for 20 s.

Two increments were used in order to reduce polymerization shrinkage effects and achieve effective setting with light-curing . Flowable resin was used first since the authors consider its manipulation easier (gaps might appear with the spatula’s withdrawal when using hybrid composite) . Adherence between both increments is excellent due to compatibility between the different materials and to their copolymerization .

Finally, restorations were sculpted and polished using fine diamond burs, rubber points and flexible discs.

2.2.3

Details of how adherence of care providers was assessed or enhanced

Emails were sent every month by the main investigator (J.P. Attal) to motivate the participating dentists. Ambiguous inclusions were resolved either on the telephone or by email.

2.3

Outcomes: evaluation criteria and procedure

Restorations were evaluated at baseline, 6 months, 1 year and 2 years of clinical service by the operator who had completed them, using simplified binary FDI criteria . Operators were trained to evaluate these criteria before the trial started. A meeting took place in September 2007 and statement of the criteria in the Case Report Form was slightly modified according to operators’ remarks.

The primary clinical endpoint was restoration retention. Secondary endpoints included: marginal adaptation (0: no defect clinically detectable or minor marginal gap or fracture removable by polishing; 1: localized or generalized gap resulting in exposure of dentin or base, repair is necessary), postoperative (hyper-)sensitivity and tooth vitality, marginal staining at the enamel, marginal staining at the cement, surface staining and minor defects (they concerned restoration margins: the evaluator did not consider there was a real defect but, for example “the probe hooked to a location at the restoration’s margin”; 0: no minor defect; 1: minor defect).

Major defects were defined as a defect in at least one of the following properties: sensitivity, surface staining, marginal adaptation and marginal staining at the enamel or the dentin.

2.4

Objectives

The purpose of this paired-tooth design trial was to evaluate the effectiveness of the new one-step self-etch adhesive Bond Force to restore cervical non-carious lesions in dental practices and to test the hypothesis that this adhesive is equally effective with or without beforehand selective etching of enamel after a 2-year follow-up period. The authors’ assumption was that selective enamel acid-etching would enhance clinical performances of the adhesive and that effectiveness of cervical restorations in dental practices would be lower than that observed in previous studies in university dental school settings because of the time constraint. In terms of secondary endpoints, the authors assumed an absence of difference between both groups for marginal staining at the dentin and surface staining since only marginal enamel treatment differed. They expected differences in favor of the experimental group for minor defects, marginal adaptation and marginal staining at the enamel.

2.5

Sample size

We assumed an odds ratio of 4 in favor of the experimental group and 30% of discordant pairs. Sample size was thus calculated to be 67 restorations (34 patients) in order to detect a two-sided difference with a 5% alpha risk and a power of 80%. Clustering by patient was addressed since sample size was determined using a McNemar test, but clustering by operator was not addressed because intraclass correlation was hard to determine given the absence of former data. This report is the 2-year final analysis that was requested by the industrial firm supporting the study.

2.6

Randomization

A different pre-set random table mentioning 10 pairs of experimental and control procedures was used by each dentist to allocate the procedure to each new patient.

2.7

Implementation

The study’s clinical research associate generated the allocation sequence.

The five dentists participating in the study enrolled patients who met the inclusion criteria and assigned their teeth to one of the groups. The tooth with the lowest tooth number received the treatment mentioned first, while the one with the highest number received the treatment mentioned second.

Only one pair of restorations was placed in each patient.

2.8

Blinding

Patients were blinded to group assignment. Operators were not blinded to group assignment when administering interventions. Each dentist carried out all evaluations concerning the restorations made using the pre-determined set of criteria. All parameters were recorded using a standardized paper case report form (CRF). The evaluation paper had to be sent after each observation to the study’s clinical research associate, so that evaluators were blinded to group assignment during follow-up recalls.

2.9

Statistical analysis

The analysis was by intention to treat. The two treatment protocols were mutually compared per restoration pair.

Statistical analysis compared the proportions for the primary endpoint between the experimental and the control group using the McNemar test at a significance level of 5% ( p < 0.05). Since secondary endpoints were multiple, an inflation factor was applied to the traditional 5% alpha risk: a simple Bonferroni correction allowed us to conclude for secondary endpoints with a p -value below 1%.

A hierarchical regression was then used to address clustering by patient and care provider. Statistical analysis was performed using Stata software (version 10).

2.1

Participants

Twenty-eight patients were consecutively enrolled in 5 private dental offices situated in the Paris region in France.

2.1.1

Inclusion and exclusion criteria

Patients had to be affiliated to a social welfare organization, be 18 or older, have an acceptable oral hygiene level and present at least two cervical erosions to be restored on two different teeth. The lesion had to be non-carious, more than 1 mm deep and involve both the enamel and dentin of a vital incisor, canine or premolar without mobility. All patients had to sign a written consent before being included. Patients with a compromised medical history, a severe or active periodontal or carious disease and heavy bruxism or a traumatic occlusion were excluded from the study.

Exclusion criteria after enrolment consisted of the modification of the restoration by another dentist and of the loss or fracture of the tooth for reasons not connected with the restoration.

2.1.2

Settings and locations where the data were collected

The data were collected in Toulouse (France) at the dental university by the study’s clinical research associate (Vergnes JN).

2.1.3

Eligibility criteria for centers and those performing the interventions

To participate in the study, dentists had to be part of the Paris Practice-based Research network, to have at least 3 years of experience, a relationship with the dental university and basic clinical research knowledge. Dentists also had to work in a dental practice and have no plan to move in the next 2 years.

2.2

Interventions: restorative procedure

2.2.1

Precise details of the interventions intended for each group and how and when they were actually administered

Operative procedures were performed by five specially instructed and experienced dentists in their own private practice.

The new adhesive had been sent to them in June 2007 so that they could get used to it. The protocol and the Case Report Form (CRF) were written in details during the summer. They were sent to the operators so that they could evaluate the protocol’s feasibility. A meeting took place in September, and operators’ remarks were taken into account. The protocol was finalized and sent to the five dentists. Inclusion started on September 5th in 2007 and ended on November 27th.

2.2.2

Details of how the interventions were standardized

The interventions were standardized by a detailed protocol, which is briefly summarized below.

Local anesthesia was applied if needed to prevent patient discomfort during intervention. A dry operative field was obtained using cotton rolls and aspiration by a dental assistant. If necessary, a retraction cord was used.

A preliminary cleaning of the tooth surface with pumice and water aimed to remove the salivary pellicle and any remaining dental plaque. Then sclerotic dentin and/or discolored tooth tissue was removed with a red diamond bur, and a short enamel bevel (1–2 mm) was prepared.

The various materials’ composition and application procedure are detailed in Table 1 .

In the control group, the adhesive was applied strictly following a self-etch approach on both enamel and dentin according to manufacturer’s instructions: after rinsing and moderate drying, the adhesive was applied for 20 s by brushing the cavity surface. An air spray was then applied on the adhesive from a distance for 5 s and nearer for another 5 s. Polymerization was obtained with a light output not less than 550 mW/cm ² .

In the experimental group, the application of the same adhesive followed the same protocol, but the adhesive was applied after selective 40% phosphoric acid-etching of the enamel cavity margins, followed by rinsing and gentle air drying.

The cervical tooth anatomy was restored by applying a first increment of flowable resin composite (Estelite Flow Quick/Tokuyama), followed by a second increment of micro-hybrid resin composite (Estelite Sigma/Tokuyama). Both increments were light-polymerized for 20 s.

Two increments were used in order to reduce polymerization shrinkage effects and achieve effective setting with light-curing . Flowable resin was used first since the authors consider its manipulation easier (gaps might appear with the spatula’s withdrawal when using hybrid composite) . Adherence between both increments is excellent due to compatibility between the different materials and to their copolymerization .

Finally, restorations were sculpted and polished using fine diamond burs, rubber points and flexible discs.

2.2.3

Details of how adherence of care providers was assessed or enhanced

Emails were sent every month by the main investigator (J.P. Attal) to motivate the participating dentists. Ambiguous inclusions were resolved either on the telephone or by email.

2.3

Outcomes: evaluation criteria and procedure

Restorations were evaluated at baseline, 6 months, 1 year and 2 years of clinical service by the operator who had completed them, using simplified binary FDI criteria . Operators were trained to evaluate these criteria before the trial started. A meeting took place in September 2007 and statement of the criteria in the Case Report Form was slightly modified according to operators’ remarks.

The primary clinical endpoint was restoration retention. Secondary endpoints included: marginal adaptation (0: no defect clinically detectable or minor marginal gap or fracture removable by polishing; 1: localized or generalized gap resulting in exposure of dentin or base, repair is necessary), postoperative (hyper-)sensitivity and tooth vitality, marginal staining at the enamel, marginal staining at the cement, surface staining and minor defects (they concerned restoration margins: the evaluator did not consider there was a real defect but, for example “the probe hooked to a location at the restoration’s margin”; 0: no minor defect; 1: minor defect).

Major defects were defined as a defect in at least one of the following properties: sensitivity, surface staining, marginal adaptation and marginal staining at the enamel or the dentin.

2.4

Objectives

The purpose of this paired-tooth design trial was to evaluate the effectiveness of the new one-step self-etch adhesive Bond Force to restore cervical non-carious lesions in dental practices and to test the hypothesis that this adhesive is equally effective with or without beforehand selective etching of enamel after a 2-year follow-up period. The authors’ assumption was that selective enamel acid-etching would enhance clinical performances of the adhesive and that effectiveness of cervical restorations in dental practices would be lower than that observed in previous studies in university dental school settings because of the time constraint. In terms of secondary endpoints, the authors assumed an absence of difference between both groups for marginal staining at the dentin and surface staining since only marginal enamel treatment differed. They expected differences in favor of the experimental group for minor defects, marginal adaptation and marginal staining at the enamel.

2.5

Sample size

We assumed an odds ratio of 4 in favor of the experimental group and 30% of discordant pairs. Sample size was thus calculated to be 67 restorations (34 patients) in order to detect a two-sided difference with a 5% alpha risk and a power of 80%. Clustering by patient was addressed since sample size was determined using a McNemar test, but clustering by operator was not addressed because intraclass correlation was hard to determine given the absence of former data. This report is the 2-year final analysis that was requested by the industrial firm supporting the study.

2.6

Randomization

A different pre-set random table mentioning 10 pairs of experimental and control procedures was used by each dentist to allocate the procedure to each new patient.

2.7

Implementation

The study’s clinical research associate generated the allocation sequence.

The five dentists participating in the study enrolled patients who met the inclusion criteria and assigned their teeth to one of the groups. The tooth with the lowest tooth number received the treatment mentioned first, while the one with the highest number received the treatment mentioned second.

Only one pair of restorations was placed in each patient.

2.8

Blinding

Patients were blinded to group assignment. Operators were not blinded to group assignment when administering interventions. Each dentist carried out all evaluations concerning the restorations made using the pre-determined set of criteria. All parameters were recorded using a standardized paper case report form (CRF). The evaluation paper had to be sent after each observation to the study’s clinical research associate, so that evaluators were blinded to group assignment during follow-up recalls.

2.9

Statistical analysis

The analysis was by intention to treat. The two treatment protocols were mutually compared per restoration pair.

Statistical analysis compared the proportions for the primary endpoint between the experimental and the control group using the McNemar test at a significance level of 5% ( p < 0.05). Since secondary endpoints were multiple, an inflation factor was applied to the traditional 5% alpha risk: a simple Bonferroni correction allowed us to conclude for secondary endpoints with a p -value below 1%.

A hierarchical regression was then used to address clustering by patient and care provider. Statistical analysis was performed using Stata software (version 10).