The study protocol was reviewed and approved by the ethics review board of the XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX (2013/17). A signed informed consent form was obtained from each participant and their parents. The sample size of this prospective clinical study was calculated to be at least 26 individuals for 98.6% significance level and 95% reliability. Individuals with the following conditions were sought for inclusion in the study:
1. Individuals who will continue to the retention phase after fixed orthodontic treatment,
2. Presence of all mandibular incisor and canine teeth,
3. Good oral hygiene,
4. No caries,
5. No fractures,
6. Healthy periodontal condition,
7. No restorations,
8. No previous bonded retainer,
9. No traumatic parafunctional habits such as bruxism.
The average age of the 86 participants (72 female and 14 male) was 17 years 4 months (11-34 years). After the orthodontic appliances were removed, a mandibular alginate impression was taken, and a plaster study model was constructed. The lingual retainers, which were made of six-stranded stainless steel wire 0.0215 inches in diameter (American Orthodontics, Washington. Avenue, Sheboygan, USA), first of all were bent on the study models by a single orthodontic technician.
To avoid inter-examiner variation, the bonding procedures were conducted by the same operator (XX). Before bonding, non-fluoridated pumice was used for 20 s to polish the teeth. A split-mouth design was used. Thus, individual differences such as age, sex, enamel, and salivary structure and chewing function between the study and control groups are eliminated. The mouth was divided in half and a randomly alternating contralateral bonding pattern was used to make sure that the enamel treatment was equally distributed between right and left [13]. The direction of acid application in patients was determined by drawing cards using the simple randomization method. The flow chart of this study was shown in Figure 1.
In the study group, the SEP (3M Unitek, Monrovia, California) was used according to the manufacturer’s instructions, namely it was administered to the lingual surfaces of the teeth and rubbed for 3 s. Then a gentle burst of dry air was delivered to thin the primer. In the control group, the lingual surfaces of the teeth were etched using 37% phosphoric etchant liquid gel (3M Espe, St Paul, Minnesota, USA) for 30 s, followed by rinsing and drying. Next, the primer (Transbond XT Primer; 3M Unitek, California) was applied in a thin and uniform coat. Then the adhesive resin (Transbond LR Light Cure Adhesive Paste; 3M Unitek, California) was administered to the lingual surface of the anterior teeth and the lingual retainer was placed in position. The adhesive resin was polymerized from two directions for a total of 20 s using a visible-light curing unit (Hilux 200, Benlioglu Dental Inc., Ankara, Turkey) with an output power of 600 mW/cm2.
The lingual retainer’s surface was examined for smoothness, the contact points and gingival areas for surplus adhesive.
The patients were evaluated again after 1, 3, 6, and 12 months. Then they were checked after 24 months. Bond failures were recorded by the same researcher (XX). The amount of adhesive left on the tooth was ascertained visually using the adhesive remnant index (ARI) [14].
The statistical analysis was performed using the Statistical Package for the Social Sciences (version 12.0, SPSS Inc., Chicago, Illinois, USA). The bond failure rate over 24 months was established for both bonding procedures and only the first failures were used for the statistical analysis. The failure rates were compared using the chi-square test (P < 0.05). Chi-square analysis was also used to ascertain the differences in ARI scores between the bonding procedures (P < 0.05). The retainers’ survival rates were estimated by Kaplan–Meier test. Their survival distributions regarding the bonding procedure were compared using the log-rank test (P < 0.05).