The aim of this systematic review was to assess the prevalence of adverse effects associated with lingual and buccal fixed orthodontic techniques.
Two authors searched the PubMed, EMBASE, Cochrane Library, and LILACS databases up to October 2014. Agreement between the authors was quantified by the Cohen kappa statistic. The following variables were analyzed: pain, caries, eating and speech difficulties, and oral hygiene. The Newcastle-Ottawa scale was used to assess risk of bias in nonrandomized studies, and the Cochrane Collaboration’s tool for assessing risk of bias was used for randomized controlled trials.
Eight articles were included in this systematic review. Meta-analysis showed a statistically greater risk of pain of the tongue (odds ratio [OR], 28.32; 95% confidence interval [95% CI], 8.60-93.28; P <0.001), cheeks (OR, 0.087; 95% CI, 0.036-0.213; P <0.0010), and lips (OR, 0.13; 95% CI, 0.04-0.39; P <0.001), as well as for the variables of speech difficulties (OR, 9.39; 95% CI, 3.78-23.33; P <0.001) and oral hygiene (OR, 3.49; 95% CI, 1.02-11.95; P = 0.047) with lingual orthodontics. However, no statistical difference was found with respect to eating difficulties (OR, 3.74; 95% CI, 0.86-16.28; P = 0.079) and caries (OR, 1.15; 95% CI, 0.17-7.69; P = 0.814 [ Streptococcus mutans ] and OR, 0.67; 95% CI, 0.20-2.23; P = 0.515 [ Lactobacillus ]).
This systematic review suggests that patients wearing lingual appliances have more pain, speech difficulties, and problems in maintaining adequate oral hygiene, although no differences for eating and caries risk were identified. Further prospective studies involving larger sample sizes and longer follow-up periods are needed to confirm these results.
Adverse effects of lingual and buccal appliances were assessed in a meta-analysis.
Lingual appliance patients had more pain and speech difficulties and poorer oral hygiene.
No differences in eating and caries risk were identified.
Larger sample sizes and longer follow-up periods are needed to confirm these results.
Malocclusion has a significant impact on smile esthetics and is increasingly believed to have an impact on sociopsychological well-being and quality of life. Although lingual orthodontics were introduced over 30 years ago, in recent years the demand for and the provision of lingual orthodontic treatments have increased among patients seeking improved esthetics. This trend reflects an acceptance that invisible and tooth-colored orthodontic appliances are preferable in terms of appearance to more traditional metal appliances.
Lingual orthodontic appliances unquestionably have esthetic advantages over conventional buccal appliances. Moreover, it has been claimed that lingual appliances may have a reduced risk of caries, perhaps related to the pooling of saliva on the lingual aspect of the mandibular teeth in particular, but greater difficulties for maintaining oral hygiene and heightened pain experience; the latter may impact negatively on the objectives and outcomes of the orthodontic treatment. A retrospective study involving 111 patients showed that between 57% and 76% of those treated with lingual orthodontics experienced pain in the tongue, eating and speech difficulties, and problems in maintaining adequate oral hygiene. From the esthetic perspective, it is undeniable that lingual orthodontics offer better results. However, it remains unclear as to which orthodontic technique (lingual or buccal) has fewer side effects in terms of pain, caries, eating and speech difficulties, and levels of oral hygiene.
This systematic review of the scientific literature was directed at answering the following question: Do lingual orthodontics cause more adverse effects than do buccal orthodontics?
Material and methods
This study was conducted and reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA).
On the basis of patient, intervention, comparison, and outcome (PICO), the question that guided this review was: Are lingual fixed orthodontic appliances associated with greater or fewer adverse effects than buccal fixed appliances?
Search strategy for the identification of studies
The following search strategy was used for the identification of studies. Two authors (F.A.A. and J.A.A.) searched electronic databases (PubMed, EMBASE, Cochrane Library, and LILACS) up to October 2014 with these key words or combinations: lingual orthodontics, lingual orthodontics treatment, lingual bracket, labial or buccal orthodontics, labial or buccal orthodontics treatment, labial or buccal bracket. The same 2 authors read the titles and abstracts of all the studies without blinding to the names of the authors or the publication dates. They reviewed the full-text articles of the potentially relevant titles and abstracts against the inclusion criteria. The search was completed with a review of the references cited in the selected articles to identify additional studies not found in the initial search. This was supplemented with a manual search of leading orthodontic journals.
Study screening criteria
Before the study, a screening protocol was developed, and the following inclusion criteria were established.
Studies comparing healthy patients with malocclusions requiring orthodontic treatment, in which one group of subjects received treatment with lingual fixed appliances and the other group received treatment with the buccal technique.
Articles analyzing the following adverse effects: pain, caries, eating difficulties, speech difficulties, and deficient oral hygiene.
All randomized controlled trials (RCTs) and controlled clinical trials (CCTs) were included for the study of adverse effects. We included studies involving either single or dual arch appliances. Patients with systemic diseases that could have influenced the orthodontic treatment and in vitro and animal studies were excluded.
The authors of the publications were contacted for data clarification when necessary. There were no restrictions in terms of year of publication or language. In cases involving more than 1 publication with the same group of patients and the same follow-up period, we included only the study that came closest to the objectives of our review or which involved the largest sample. A translation was arranged for 1 article in Portuguese. All the articles selected in the electronic and manual searches were evaluated independently by the first and second authors in accordance with the established inclusion criteria. Any disagreements between the authors were resolved by consensus or by consulting another author.
Assessment of risk of bias of the studies
Two authors (F.A.A. and J.A.A.) independently assessed the risk of bias of the included RCTs using the Cochrane Collaboration’s tool for assessing risk of bias guided by the Cochrane Handbook for Systematic Reviews of Interventions (version 5.1.0). Six specific domains were assessed: sequence generation, allocation sequence concealment, blinding, incomplete outcome data, selective outcome reporting, and other sources of bias. The overall risk of bias in each study was assessed using the following judgments: low, moderate, or high. The risk of bias was evaluated as follows: low risk of bias when any bias was unlikely to seriously affect the results (if 5 or more domains were recorded with “yes”), moderate risk of bias when any bias could question the results (if 3 or 4 domains were recorded with “yes”), or high risk of bias when any bias could seriously affect the results (if 3 or fewer domains were recorded with “yes”).
The quality of the included nonrandomized studies was independently assessed by the same authors using the Newcastle-Ottawa Scale. This scale comprises 8 items covering 3 dimensions: selection (adequacy and representativeness of the subjects, selection and definition of controls), comparability, and exposure (ascertainment of exposure, discrimination between subjects and controls, and nonresponse rate). The maximum score was 9 points; high quality was defined as 6 points or more, and low quality was defined as less than 6 points. Any disagreements between the authors were resolved by consensus or by consulting the last author (J.C.) of this study, and the level of agreement between the 2 reviewing authors was assessed with the Cohen kappa statistic.
This meta-analysis was based on the inverse-variance method of DerSimonian and Laird, with the odds ratio (OR) as the measure of effect. The estimates for a random effects model were obtained. The OR estimates are accompanied by the corresponding 95% confidence interval (95% CI) and P value of the null effect contrast of the factor “type of orthodontic treatment” (OR = 1) in the meta-analysis. A graphic representation using forest plots was made for the OR and, in parallel, for the natural logarithm of the OR (which renders the estimates symmetrical around 0 and favors normal adjustment). Because of variations in the presentation of the results from the primary studies, transformation of the standardized mean differences into odds ratios using the formula of Chinn was undertaken when necessary to allow for comparison of the results as part of the meta-analysis. The statistical analysis was performed using the statistical package for Microsoft Windows (version 17.0; SPSS, Chicago, Ill) and R (version 3.0.2; R Foundation for Statistical Computing, Vienna, Austria).
Assessment of heterogeneity
Statistical heterogeneity and the possible causes for it were assessed by visual inspection of the forest plots, the Q statistic, and the corresponding P value, and by means of the I 2 index. An I 2 value of 0% indicates no observed heterogeneity, whereas larger values reflect increasing heterogeneity (with 25% indicating low, 50% moderate, and 75% high heterogeneity). Forest plots involve a weighted compilation of all the effect sizes reported by each study and also provide an indication of heterogeneity between studies.
Assessment of publication bias
Publication bias was initially to be evaluated through visual inspection of funnel plot asymmetry. In addition, publication bias was to be assessed statistically using the test of Begg and Mazumdar (rank correlation method). Funnel plots and Begg and Mazumdar rank correlation test were chosen to detect publication bias if the number of included studies exceeded ten.
Selection and description of the included studies
The electronic search procedures and excluded articles (along with the reason for exclusion) are outlined in Figure 1 . A total of 8 articles were included in the systematic review, and a meta-analysis was made of 6 studies. Two publications lacked adequate follow-up periods (2-4 weeks). From the study of Khattab et al, we excluded the quantitative analysis of the variable pain because the authors failed to specify its location. The demographic data (numbers of patients, mean ages, and sexes), types of study, and information about the bracket systems used in each study are summarized in Table I . Of the 8 articles included in our systematic review, five were CCTs, and three were RCTs. The 6 studies in the meta-analysis included a total of 131 patients who received buccal orthodontic treatment and 131 patients treated with lingual brackets.
|Authors and design||Demographic data||Type of brackets|
|Authors||Study design||Number of patients (B/L)||Sex (M/F)||Age (y) buccal/lingual||Buccal||Lingual|
|Caniklioglu and Öztürk||CCT||60 (30/30)||21/39||17/18.3||Roth||7th generation; Ormco, Orange, Calif|
|Galvão et al||CCT||31 (20/11)||NA||25/21||NA||NA|
|Van der Veen et al||RCT||28 (14/14)||NA||15.3 ± 1.2||Orthos; Ormco, Glendora, Calif||Lingualtechnik, Bad Essen, Germany|
|Wu et al||CCT||60 (30/30)||22/38||20.3/21.6||Mini-Diamond; Ormco. Orange, Calif||Incognito, 3M Unitek, Bad Essen, Germany|
|Wu et al||CCT||60 (30/30)||22/38||20.3/21.6||Mini-Diamond; Ormco. Orange, Calif||Incognito; 3M Unitek, Bad Essen, Germany|
|Shalish et al||CCT||47 (28/19)||18/29||18-60||GAC and Ormco||Incognito|
|Khattab et al||RCT||34 (17/17)||13/21||21.8/20.6||Roth Mini-Master series; American Orthodontics, Sheboygan, Wis||Stealth; American Orthodontics, Sheboygan, Wis|
|Lombardo et al||RCT||20 (10/10)||5/15||19.3 ± 3.6/22.3 ± 3.2||Roth; American Orthodontics, Sheboygan, Wis||STb lingual bracket; Ormco, Glendora, Calif|
Risk of bias/quality assessment of primary studies
The risk of bias of the randomized controlled trials included in the systematic review is summarized in Table II . Of the 3 RCTs, 1 was considered to have low risk of bias; 1 had a moderate risk of bias, and 1 was judged at high risk of bias. The physical placement of brackets precludes blinding of both the operator and the participant, but the variable of blinding was considered to have a low risk in all studies, since the results are unlikely to have been influenced by a lack of blinding. With regard to the 5 nonrandomized studies, 4 were considered to be of high quality, with a Newcastle-Ottawa Scale score of 6 or more. One study was regarded as being of low quality, with a Newcastle-Ottawa Scale score of less than 6 ( Table III ). The agreement between the reviewers for risk of bias assessment, based on the kappa statistic, was 0.90.
|Study||Sequence generation||Allocation concealment||Blinding||Incomplete outcome data||Selective outcome reporting||Other sources of bias||Risk of bias|
|van der Veen et al||Yes||Uncertain||Yes||Yes||Yes||Yes||Low|
|Khattab et al||Yes||Uncertain||Yes||No||Yes||No||Moderate|
|Lombardo et al||Yes||No||Yes||Uncertain||No||No||High|
|Quality criteria||Selection (4)||Comparability (2)||Exposure (3)||Total (9)|
|Is case definition adequate? (1)||Representativeness of the subjects (1)||Selection of controls (1)||Definition of controls (1)||Comparability on basis of design or analysis (2)||Ascertainment of exposure (1)||Same method of ascertainment
for subjects and controls (1)
|Nonresponse rate (1)|
|Caniklioglu and Öztürk||●||●||○||●||● ●||●||●||○||7|
|Galvão et al||●||●||○||●||○ ○||●||●||○||5|
|Wu et al||●||●||○||●||● ●||●||●||○||7|
|Wu et al||●||●||○||●||● ●||●||●||○||7|
|Shalish et al||●||●||○||●||○ ●||●||●||○||6|
Description of the variables
The following variables were analyzed in this systematic review.
Pain. Five studies were included. Caniklioglu and Öztürk evaluated discomfort in the tongue, cheeks, and lips after 3 months, using a 12-item questionnaire. Galvão et al evaluated discomfort in the tongue, cheeks, gums, and lips after 1 month using a 15-item questionnaire. Wu et al evaluated pain in the tongue, cheeks, lips, gums, mandible, and face after 1 week and 1 and 3 months, with the measurements based on a visual analog scale scored from 0 to 10. Shalish et al analyzed pain after 2 weeks, also using a visual analog scale scored from 0 to 10, although without specifying the location of the pain. Khattab et al assessed soft tissue irritation 1 and 3 months after cementing of the brackets, based on a 5-item questionnaire. This study likewise failed to specify the location of the pain. The results of the meta-analysis related to pain in the tongue (OR, 28.32; 95% CI, 8.60-93.28; P <0.001), cheeks (OR, 0.087; 95% CI, 0.036-0.213; P <0.001), and lips (OR, 0.13, 95% CI, 0.04-0.39; P <0.001) are shown in Figure 2 .