This article was reported according to CONSORT (Consolidated Standards of Reporting Trial)17 guidelines, and the checklist is available as a supplementary file.
Trial Design and Ethical Aspects
This is a two-sided equity, parallel arms, one-to-one allocation ratio, single-blinded (examiner), controlled randomized clinical trial.This RCT is nested to a caries diagnostic RCT intitled CARies DEtection in Children 1 (CARDEC 1)18.
It was conducted in dental office setting with children who sought dental care at the School of Dentistry, University of São Paulo, Brazil. This study was approved by the local research ethics committee (protocol #864.396) and registered on 15/10/2014 on the ClinicalTrials.gov platform (NCT02274142). All parents or legal guardians signed the informed consent form.
Initially children from three to six years of age would be included, however, to cover the largest number of children included in CARDEC 1, we increased this age group to three to ten years old. In addition, we have performed an Intention to Treat (ITT) and subgroup analysis that were not declared on trial registry.
Sample Size
The sample size estimation was performed on the Power and Sample Size website (http://powerandsamplesize.com/). A two-tailed hypothesis was considered. We considered parameters from a systematic review19, which reported an average survival rate of 78% after two years of follow-up (mean for occlusoproximal and occlusal restorations after 2 years of follow-up). A clinically important difference of 15% on survival rate between ENC and H/M groups was considered. We added 20% for possible losses to follow up, and 20% for the cluster effect, as the same child could have more than one tooth included in the study. Thus, 116 teeth were needed per group, reaching a minimal sample size of 232 teeth. A significance level of 5% and a power of 80% were used for calculation.
Eligibility Criteria
Healthy children aged 3 to 10 years, who had sought dental treatment in the University of São Paulo and participating in CARDEC 01 study18 were assessed. Only children presenting dentin caries lesion in primary molars detected clinically as a cavitation or radiographically as a dentine radiolucency in the bitewing radiograph in occlusal and/or occlusoproximal surfaces were eligible to participate20. However, when any signs or symptoms of irreversible pulp inflammation or pulp necrosis were detected clinically (nocturnal pain, fistula, abscess, pulp exposure, pathological mobility) or radiographically (radiolucency into the pulp, furcal bone radiolucency or pathological root resorption) the tooth was excluded.
Children with severe behavioral problems and those whose parents or guardians refused to sign the informed consent form were excluded.
Randomization, Allocation concealment, and implementation
A sequence of random numbers, stratified according to caries experience and in blocks of four, was generated using a Random Allocation Software 2.021, and these numbers were packed in opaque sealed envelopes by an external member of the research team who did not participate in the operative stages of the study to guarantee the allocation concealment.
Children with low experience were considered those who presented dmfs less than or equal to 3, and children with high experience those whose dmfs were higher than 320.
The randomization implementation was made by two team members who did not participate in the research's operative phase, as they were responsible for the treatment plan design. Thus, the operators received the treatment plan to be performed, avoiding selection bias.
The randomization unit was the tooth, so each child could be able to contribute with more than one tooth to the study, and therefore could have received different restorative material in different teeth.
Blinding
The restorations were performed with GIC in hand-mixed (H/M) and encapsulated (ENC) versions. Thus, there was no possibility of blinding participants and operators. Only the outcome assessors were blinded regarding groups.
Interventions
All treatments were performed by trained general dentists and graduate students in Pediatric Dentistry. After clinical and radiographic evaluation and being accepted for the participation in the study, children were randomized the following treatments groups:
-
Hand-mixed Group: restorations performed with GIC Fuji IX Gold Label® (GC Europe NV, Leuven, Belgium), in the hand-mixed version with manual dosage and handling.
-
Encapsulated Group: restorations performed using GIC Equia Fill® (Easy/Quick/Unique/Intelligent/Aesthetic) – GC Europe NV, Leuven, Belgium, in the pre-dosed encapsulated version and mechanical manipulation.
Restorative Procedures
The procedures were performed without the use of local anesthesia. High speed rotary burs were used for enamel removal to gain access in case of non-cavitated dentin lesions detected radiographically.
Selective caries removal was performed in occlusal and occlusoproximal lesions for both groups. The caries dentin was removed using hand instruments appropriate for the cavity size. Cavity conditioner (polyacrylic acid – GC cavity conditioner) was applied for 15 seconds using a wet cotton pellet. Rinsing was performed using a sequence of three wet cotton pellets followed by three dry cotton pellets. The GIC was mixed and applied according to the following groups:
H/M:
The GIC hand-mixed was handled in a paper block with a plastic spatula (GC Corporation, Japan) by two trained operators following the manufacturer’s recommendations. The GIC was inserted into the cavity with #1 spatula. Press finger was performed with a gloved finger coated with petroleum jelly for 10 seconds. After the initial material setting (from 3 to 5 minutes), the occlusion was checked with carbon paper and adjusted when necessary. Finally, restoration protection was performed with petroleum jelly.
ENC:
The GIC (Equia Fill - GC Corporation, Japan) was activated, following the manufacturer's recommendations, and taken to the mixer by a team member, other than the operator of the restorative procedure. The encapsulated material was inserted directly from the capsule using a capsule applier (Riva Applicator – SDI Limited®, Australia). Press finger was performed with a gloved finger coated with petroleum jelly for 10 seconds. After the initial material setting (from 3 to 5 minutes), the occlusion was checked with carbon paper and adjusted when necessary. Restoration protection was performed with petroleum jelly.
For all occlusoproximal cavities, metal matrixes and wooden wedges were used. All subjects were instructed not to eat for one hour and received instructions on sugar consumption and oral hygiene for caries control.
Outcome Evaluation
The restorations were assessed clinically continuously, with a minimum of 4 months and a maximum of 8 months between evaluations, up to 24 months by two trained, calibrated, and blinded examiners (D.P.R. and L.B.C), following the Frencken and Holmgren (2001) criteria for occlusal restorations and Roeleveld et al. (2006) criteria for occlusoproximal restorations and the additional files 1 and 2, respectively, shows this in more detail. Kappa test was performed to evaluate the level of inter-examiner agreement.
We considered as success for occlusal restorations the scores 0, 1 e 7, and for occlusoproximal, 00 and 10, which indicate the presence of good restoration, or only minor defect, with no repair needed. If other minor or major restoration failure was noted, the repair of the restoration was performed. In case of bulk fracture, the tooth received a new restoration.
For survival analysis, the reintervention was not considered. If the tooth presented a failure during the evaluation period, it was considered as a restoration failure for survival analysis.
All participants received full dental treatment, except orthodontic appliances. Parents could bring the child if any treatment need was detected between the pre-determined assessments.
Outcomes
This study's primary outcome was restorations' survival performed with hand-mixed and encapsulated GIC in occlusal and occlusoproximal cavities in primary molars after 24 months of follow-up. The occlusal restorations were evaluated according to the criteria proposed by Frencken and Holmgren (2001). For these restorations we considered successful those restorations that did not present an operative need, i.e., scores 0, 1 and 7. For occlusoproximal restorations the criteria used was the one proposed by Roeleveld et al. (2006), where the criteria considered successful were 00 and 10. Minor or major failures, according to the mentioned criteria, were considered as failure for the survival analysis. As a secondary outcome, we aimed to evaluate the cost-effectiveness of both GIC, considering the longevity of restorations. The secondary outcome will be published separately elsewhere.
Statistical Analysis
The analyses were performed in Stata 13 (StataCorp, USA). Kaplan-Meier's analysis shows the survival of the restorations over the 24 months of follow-up. Participants evaluated at least once during the study were included in the analysis.
To evaluate the primary outcome (restoration survival), we performed an intention to treat (ITT) analysis, considering the success and failures at 24 months of follow-up.
To evaluate the association between restoration survival and independent variables such as surface (occlusal or occlusoproximal), caries experience (dmfs ≤ 3 or > 3), type of molar (first or second molar), sex (male or female), age and arch (upper or lower), Cox Regression with shared frailty was used. Initially, the analysis was performed in a univariate model. Independent variables reaching a p-value < 0.20 (cavities and tooth type) fitted in the adjusted model. The final model only included variables showing p ≤ 0.05. As only the independent variable surface reached this p-value, we have conducted a subgroup analysis, considering the survival of occlusal and occlusoproximal restorations. Hazard ratios (HR) and relative risk (RR) were calculated with 95% confidence interval (CI). The significance level was set at 5%.