To evaluate the stability after orthodontic treatment between two types of lower fixed retainers: those bonded onto all anterior teeth or those bonded only onto the canines.
The following electronic databases were consulted: PubMed, Scopus, Web of Science, Cochrane Library, Lilacs, OpenGrey, ClinicalTrials, and Google Scholar. No restriction of language or year were applied. After selection of studies, risk-of-bias evaluation and qualitative synthesis of the included studies were performed using The Cochrane Collaboration's tool for randomized studies and the “Risk of Bias in Non-randomized Studies of Interventions” (ROBINS-I) tool for nonrandomized studies, and a summary of the overall strength of evidence was presented using the “Grading of recommendations, assessment, development and evaluation” tool.
Among the 180 studies retrieved from the searches, five were included in this review. Three of them showed a low risk of bias, while two presented a high risk of bias. With regard to stability, two studies reported better stability for retainers bonded to all six teeth, while the other three showed no difference. The retainer bonded to all teeth presented a higher breakage rate in one study.
Stability seems better with lower fixed retainers bonded on all anterior teeth. The breakage rate may not change according to the bonding. However, studies with greater methodological soundness are necessary to reach a more reliable conclusion.
After orthodontic movement, teeth tend to return to the direction from which they were originally moved1 as a result of traction of the elastic fibers of the gingiva and the imbalance among forces between the lips and tongue.2 After finishing treatment, time is needed for the reorganization of alveolar bone and periodontium.3 Orthodontic retainers are ideally suited to maintain tooth alignment after treatment.4
In the lower arch, fixed retainers in the anterior segment are a valid option for managing the significant relapse rate in this area.5 Recent evidence6,7 indicated that retainers are essential for long-term stability. However, there are still questions regarding the efficacy of the different types of retainers available.8
The 3 × 3 fixed retainer, bonded using the direct technique, is a commonly used type of retainer in the lower arch. This is likely attributable to the accessibility and cost-effectiveness of the technique.9 This type of retainer can be bonded either to the lingual surface of the lower canines only or to all six of the lower teeth, on the lingual surface of the incisors and canines.4,9 In addition to stability, another concern associated with fixed retainers has been the possibility of increased levels of dental plaque and calculus as a result of poor dental hygiene due to the difficulty of brushing and flossing the area around the retainer.10
Several systematic reviews11,,–14 and original studies15,16 investigated the benefits and damage associated with fixed and removable retainers. Those studies focused on stability and clinical performance of both types of retainers, considering the different materials and design of the appliances.11,,,,–16
Despite the fact that the influence of design and bonding technique for lower fixed retainers was discussed in original studies,15,16 no systematic review has addressed this question. A previous systematic review11 evaluated the difference between fixed orthodontic retainers bonded to all teeth and those bonded only to the canines, but only the periodontal condition and bonding failures were presented; results regarding the stability of dental alignment were not reported. This systematic review aimed to evaluate the difference in stability between lower fixed retainers bonded on all six anterior teeth (lower incisors and canines) and those bonded only to the canines. Additionally, breakage rates were also evaluated.
MATERIALS AND METHODS
Protocol and Registration
The protocol of this systematic review was registered in the PROSPERO database (International Prospective Register of Systematic Reviews–PROSPERO; http://www.crd.york.ac.uk/PROSPERO) under the code CRD42016050719. The Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA)17 guidelines were followed in this review.
The PICO/PECO strategy was applied. Prospective and retrospective studies performed in adults and adolescents (P), using lower fixed retainers bonded only on canines (I/E) compared to retainers bonded to all anterior teeth (lower incisors and canines), were evaluated with regard to stability (O). Animal studies, technical articles, case reports, literature reviews, and noncontrolled studies were excluded.
Search Strategy and Study Selection
Searches in the following databases were performed through June 2019: PubMed, Scopus, Web of Science, The Cochrane Library, LILACS, and ClinicalTrials. The gray literature was consulted through OpenGrey and Google Scholar. No language or year restriction was applied. The predefined search strategies presented a combination of Mesh and free terms related to orthodontic retainers and were adapted to each database (Appendix 1).
After searches were conducted, the results were imported into a reference manager software (EndNote web, Clarivate Analytics, Philadelphia, Pa). Duplicated results were excluded by automatic and manual assessment.
The selection process was performed in two phases. In the first phase, the title and abstracts that did not follow the established eligibility criteria were excluded. In the second phase, articles remaining from phase I were assessed by full text. Among the selected studies resulting from this process, reference lists were also evaluated to retrieve new articles following the eligibility criteria. All steps of the selection process were conducted independently by two reviewers (ALCSB and LBM) and checked by a third reviewer (SMAM) in case of disagreement.
Data Extraction and Risk of Bias
The results extracted from included articles were evaluated qualitatively. The country, year of publication, study design, sample characteristics, methods of evaluation, results, and statistical analysis were obtained from the included studies. In the event of an absence of information among the articles, the authors were contacted by e-mail. In an attempt to contact authors, one e-mail was sent once every week for five consecutive weeks for each study.
The risk of bias was assessed through two tools: The Cochrane Collaboration's tool for assessing the risk of bias18 applied in randomized controlled studies and the “Risk of Bias in Non-randomized Studies of Interventions” (ROBINS-I) tool19 in nonrandomized studies. In The Cochrane Collaboration's tool for assessing risk of bias18 tool, seven domains were included: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias. The performance bias and the detection bias were not considered in this evaluation. For each domain, the risk of bias was judged as “low risk,” “high risk,” or “unclear risk.”
The ROBINS-I tool19 was used in nonrandomized studies. This checklist presents three main evaluation domains: preintervention, during intervention, and postintervention. After the individualization of the main criteria, the risk of bias was assessed for each domain and classified as “low,” “moderate,” “serious,” “critical,” or “no information.”
Level of Evidence
A summary of the overall strength of evidence was presented using the “Grading of recommendations, assessment, development and evaluation” (GRADE) tool.20 Included studies were evaluated according to their design, study quality, consistency, and directness. Evaluation of stability and frequency of breakage were performed.
Selection and Characteristics of Included Studies
A total of 180 citations were retrieved from databases. After exclusion of 63 duplicated results, 117 title/abstracts were evaluated. Among those, 111 studies were excluded: two were opinion articles, two were systematic reviews that did not assess the bonding of retainers, 12 studies did not evaluate stability of retainers, 32 did not compare the two types of bonding among retainers, and 63 studies did not assess orthodontic lower retainers (Appendix 2). Six studies were reviewed by full text,21,,,,–26 and one was excluded because of the absence of stability evaluation.26 Five studies were included in this review21,,,–25 and subjected to qualitative and risk-of-bias assessment (Figure 1; Table 1). Among the included articles, two were randomized trials21,22 and three were nonrandomized clinical trials23,–25 (Table 2).
A meta-analysis was not possible to achieve as part of this systematic review as a result of methodological heterogeneity. The included studies used a different wire thickness, so comparisons among them were not feasible.
Results from Individual Studies
Stormann and Ehmer22 and Al-Nimri et al.23 reported better stability for the retainer bonded on all lower anterior teeth. The other three studies21,24,25 showed no differences between the two types of retainers in the frequency of breakage or in the stability of treatment.
Two of the studies included21,22 were randomized clinical trials with a follow-up period ranging from 222 to 321 years. The sample sizes ranged from 4921 to 9822 patients, and only adolescents and young adults were included. The other articles included in this review23,–25 were nonrandomized studies. The sample size varied from 62 to 69 patients, and the follow-up time was between 1 and 9 years. The average patient age was from 12 to 25 years among the studies.23,–25
Risk of Bias
For the Artun et al.21 study, there was an unclear risk of bias on the domains random sequence generation (selection bias) and allocation concealment (selection bias). The authors were contacted to clarify the randomization process but they were unable to provide further information (Appendix 3). For the domains of incomplete outcome data (attrition bias) and other bias there was a high risk due to a reported dropout of patients during follow up and the absence of the primary evaluator during follow up. The domains blinding of participants and personnel (performance bias) and blinding of outcome assessment (detection bias) were not evaluated because of the nature of the intervention (Figure 2).
The study from Stormann and Ehmer22 presented a low risk of bias for all domains considered. The article thoroughly described the randomization process, reported all the results, and seemed to be free from other sources of bias (Figure 2).
Level of Evidence
The GRADE evaluation highlights the results reported by the included studies, suggesting that there was better stability when lower retainers bonded on all anterior teeth were used compared to retainers bonded only to canines. A higher rate of breakage was reported for retainers bonded on all teeth; however, the strength of these results should be carefully considered because of the risk of bias and type of the included studies (Table 3).
Summary of Evidence
Among the five studies included in this review, differences were found between the performance of lower fixed retainers bonded to all lower anterior teeth and those bonded to canines only. Stormann and Ehmer22 and Al-Nimri et al.23 reported lower relapse rates when lower retainers were bonded to all teeth (canines and incisors). Artun et al.,21 Steinnes et al.,24 and Schutz-Frazon et al.25 reported no difference between the two types of retainers evaluated. Regarding the risk of bias, two studies were classified as high risk21,25 and the other studies as low risk.22,–24
The better results that were reported for retainers bonded to all lower anterior teeth may be attributed to the reduced protrusive forces produced by the tongue during the retention phase. The lower retainers bonded only on the canines could result in protrusion of the incisors that were not bonded.29
The stability of tooth position during fixed orthodontic retention can also be influenced by other factors, such as breakage of the retainer,26,30 tooth rotation,31,32 increased intercanine distance,32 the diameter of retainer wire,26,32 and the follow-up duration.6,7 The occurrence of breakage during follow up between the two retainer designs was found to be different in only one study.22 That study reported a high rate of breakage for retainers bonded on all anterior teeth.22 Even though the breakage of fixed retainers was previously associated with poor stability,26,30 in this specific study,22 no relapse on orthodontic treatment was reported.
Another factor previously shown to influence the stability of orthodontic treatment was the presence of tooth rotations corrected during treatment.31,32 A high Irregularity Index (80%) was associated with relapse of tooth rotations in two included studies.22,23 The return of rotations seemed to be related to the amount of rotation before treatment.31,32 An increase in intercanine distance during treatment has also been shown to be a cause of relapse of the crowding of anterior lower teeth.32 In this review, no changes in intercanine distance were detected in the studies.22,23
The wire thickness of the retainers used was different in the various studies included in this review.21,,,–25 Canine-to-canine retainers can be made with stainless-steel twisted archwires or plain stainless-steel archwires.9 The twisted archwires have been described as very elastic and highly resilient,33 allowing for physiological movement of the teeth, in addition to having a design that offers great mechanical retention of the material.5 One study21 included in this review suggested that twisted archwires were more effective in tooth alignment maintenance. This result was in agreement with the findings of another study,31 which reported less adverse effects31 as compared to those associated with retainers made of plain stainless-steel bonded only on canines.34 It is suggested that the use of twisted archwires is a safer alternative to prevent unexpected movements during the retention phase.34
The variation in follow-up time may have been a factor contributing to the differences found between the studies. Artun et al.,21 Steinnes et al.,24 and Schutz-Frazon et al.25 observed that there was no difference in stability between the two types of retainers, those bonded to all teeth or those bonded only to canines, at 3, 8, and 9 years of follow up, respectively. However, all of the included studies21,,,–25 identified that the retainers bonded to all teeth resulted in adequate stability, comparing for 3, 2, 1, 8, or 9 years, respectively. In the long term, this stability seemed to remain regardless of the type of retainer. There was evidence that restraint stability was effective in the long term6,7 and that the highest relapse rate occurred during the first 2 years.6
Regarding the sample size, none of the studies21,,,–25 performed a power calculation to determine the appropriate sample size. The absence of sample calculations was previously reported as a common characteristic in orthodontic journals.35 The use of nonstatistical factors to assess sample size may compromise the validity of the results and the conclusions found.36 However, in the studies included, the samples used appeared to be large enough.
Among the randomized trials included in this review, a high risk of bias was detected in one study21 and a low risk was found22 in another study. Among the three nonrandomized studies, two were classified as low risk23,24 and one25 as high risk.
Randomization in controlled clinical studies is an essential step ensuring reliability of the sample allocation and, consequently, influencing the validity of the results.37 Thus, even with a low risk of bias from the evaluation of a suitable tool for nonrandomized studies,19 two studies included in this review23,25 were still less robust due to the absence of a randomized design.
In the fields of random sequence generation and allocation concealment, an uncertain risk was identified in one study21 due to the absence of precise information regarding the method of randomization. An attempt to contact the authors was conducted, but the author was not able to clarify the method used. It is essential to perform adequate allocation concealment to reduce selection bias for confounding factors and to improve the internal validity of the study and influence the clinical outcome.37
Regarding the field of incomplete outcome data, the study by Artun et al.21 was classified as high risk of bias since patients were abandoned as a result of bonding failure, causing an uneven distribution among genders, age, and gingival state. The loss of these patients may have impaired evaluation of the final results since it can generate failures in the homogeneity of the evaluated groups. In the other risks of bias domain in the same study it was identified that the principal investigator was not available for all contention maintenance and routine clinical practice and that absence of the primary evaluator may have represented a systematic error within the sample.21 A failure of the executed measurements assessment may have implications in the dimensioning of the size of the sample examined, generating a lack of reliability of the obtained data.35
Overall, the studies have suggested there was better stability and a higher rate of breakage in lower retainers bonded on all anterior teeth, with a certainty varying from very low to moderate. Even though the results suggested better outcomes from retainers bonded to all lower anterior teeth, the clinical decision made for individual patients may involve subjective factors, such as the professional training of the provider and the patient's compliance/acceptance. Inclusion of these variables in further studies may help in developing a more reliable conclusion.
The current evidence may point to better stability of dental alignment when lower fixed orthodontic retainers are bonded to all anterior teeth.
However, as uncertainty is moderate, more reliable research on the topic is needed to clarify some of the contradictions among the studies included.
APPENDIX 3: All attempts to contact any authors and their fate
Resident, Department of Orthodontics, Brazilian Dental Association, Belém, Brazil.
PhD student, School of Dentistry, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Canada.
Professor, Department of Orthodontics, Federal University of Pará, Belém, Brazil.
Professor, Department of Pediatric Dentistry and Orthodontics, Federal University of Rio de Janeiro, Brazil.
Professor, Department of Orthodontics, Brazilian Dental Association, Belém, Brazil.