Protocol and Registration

The study protocol was registered at the PROSPERO database (CRD42015022042). The methodology followed was based on PRISMA guidelines.¹¹

Eligibility Criteria and Study Selection

Articles that evaluated the periodontal side effects obtained from RME compared with SME were included. The controlled vocabulary (MeSH terms) and the entry terms in the search strategy were defined based on the PICOS format (Table 1) as follows:

• Population (P): patients in late mixed/early permanent dentition, with mild to moderate maxillary atresia and dental crowding, using orthodontic and/or orthopedic appliances for maxillary arch expansion.

• Intervention (I): RME to correct skeletal transverse maxillary discrepancies and transverse orthodontic problems.

• Comparison (C): SME.

• Outcome (O): periodontal side effects such as bone loss, gingival recession, and plaque index.

• Study design (S): randomized or nonrandomized human clinical trials.

Table 1 Electronic Database and Search Strategy (March 29, 2018, to April 13, 2018)

View Fullscreen

The exclusion criteria were the use of reverse facemask treatment, surgical and/or extraction cases, syndromic patients, and the use of only one expansion protocol.

Information Sources

The electronic databases PubMed (MEDLINE), Cochrane Library, Scopus, Web of Science, Virtual Health Library, Google Scholar, and OpenGrey were used for the search. A hand search was also carried out in the reference lists of the articles selected for any reference that could have been missed during the electronic database searches. The related articles tool in the PubMed database was checked for each article included.

The search strategy was based on English MeSH terms, using all reports containing the combination of controlled vocabulary and entry terms relating to these words and adapted for the syntax rules of each bibliographic database (Table 1). No language or publication date restrictions were applied. Searches were performed between March 29 and April 13, 2018. All relevant titles were saved in a reference manager (EndNote basic, 2016, Thomson Reuters), and duplicate hits were removed.

Study Selection

Two examiners (Dr Bastos and Dr Blagitz) performed the searches independently to identify relevant and eligible studies. Initially, articles were selected by title and abstract according to the previously described search strategy. If the article appeared to meet the inclusion criteria in the title and/or in the abstract, the full-text version of the article was retrieved for further assessment and data extraction.

Data Collection Process

The following information was extracted from the articles included (Table 2): details of the participants, intervention, evaluation, and authors' conclusion.

Table 2 Description of Studies Included

View Fullscreen

Results of data extraction, independently performed by two researchers, were compared. If discrepancies were unresolved, a third evaluator was consulted (Dr Aragón).

Risk of Bias in Individual Studies

Risk of bias assessment was performed by two independent reviewers (Dr Bastos and Dr Blagitz) using the Cochrane Collaboration's tool for assessing risk of bias in randomized controlled trials (RCTs)¹² and the ROBINS-I tool for assessing risk of bias in nonrandomized studies of interventions.¹³ The Cochrane Collaboration's tool was used to perform quality assessment of one article included. The assessment criteria contained seven items. For each bias domain, a judgment score was given following the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions 5.1.0 (http://handbook.cochrane.org).

The judgment involved recording “yes” for low risk of bias, “no” for high risk of bias, and “unclear” in the case of no, insufficient, or uncertain information about the bias involved in the domain. Of the seven bias domains, six were considered key domains, excepting blinding of participants and personnel, because the participant always knew which treatment was used, even considering similar appliances, since the activation protocol was not the same. Studies were classified as having a “low” risk of bias if there was adequate judgment of the key domains. If any of the key domains were not met, the study was classified as having a “high” risk of bias. Finally, if there was lack of information in a key domain of the study, it was judged as “unclear,” and the reviewers tried to contact the paper's authors to obtain more information and allow a definitive judgment of “yes” or “no.”

The ROBINS-I tool is made of seven domains through which bias might be introduced. Each domain was scored as low, moderate, serious, or critical risk of bias. If there was no clear indication that the study was at serious or critical risk of bias, and there was a lack of information in one or more domains, the study was judged as having “no information.” The overall risk of bias of the study was scored as serious if serious risk of bias was scored in at least one domain. If not, the study received a moderate risk of bias in the overall evaluation. If all evaluations of the study were classified as low risk of bias, then low risk was the final evaluation of the study included.

During the period of study, if disagreements were unresolved between the reviewers, a third evaluator was consulted (Dr Aragón).

Summary Measures

Measurements of continuous data were in millimeters^5,8,10 and degrees,⁸ and categorical data and scores¹⁰ were collected for some selected clinical indices,¹⁰ cone-beam computerized tomography images,⁸ and periodontal and clinical examinations.⁵

Synthesis of Results

Data collected were synthetized in a descriptive table. A meta-analysis was planned if there was relative homogeneity of the data and the methods for obtaining it, for each selected article.

Evaluation of the Level of Evidence

The level of evidence was calculated using the Grading of Recommendations, Assessment, Development and Evaluation Pro software (GRADEpro Guideline Development Tool, available online at gradepro.org.). It grades the quality of evidence in four levels: very low, low, moderate, and high. “High quality” suggests that the true effect lies close to the estimate of the effect. “Very low quality” suggests that there is very little confidence in the effect estimate, and the estimate reported can be substantially different from what was measured. This tool considers five aspects for rating the quality of evidence.¹⁴

Study Selection

After the database screening, 620 articles were selected: 109 from PubMed, 334 from BVS, 25 from the Cochrane Library, 35 from Web of Science, one from OpenGrey, five from Google Scholar, and 111 from Scopus. After the duplicate studies were removed, 388 studies were identified. After title screening, nine studies remained for further careful examination of the abstracts. The full texts of these nine studies were assessed to check if they were eligible. Among them, six were excluded. The reasons for exclusions were as follows: expert opinion (n = 1),¹⁵ the use of only one expansion protocol, not comparing periodontal side effects of RME with SME, and/or reporting dental effects, but not related to the periodontium and/or corrective orthodontic treatment (n = 5).^6,16–19 Therefore, three articles were finally included and selected for qualitative analysis of risk of bias. The flow diagram (Figure 1) illustrates the results of the search, summarizing the process of identifying, including, and excluding studies.

View Figure

• Download as Powerpoint
• Download Figure

Study Characteristics

The characteristics of the three studies selected are listed in Table 2. Among them, there was an RCT,⁸ a nonrandomized pilot study,¹⁰ and a non-RCT.⁵

Regarding the type of orthodontic appliance, two studies used a Haas expander, changing the activation protocol to differentiate RME from SME.^8,10 The other study used a Haas expander for RME and a quad-helix appliance for SME.⁵

Concerning the measures used, one study evaluated the level of marginal alveolar bone by clinical examination with a millimeter probe, bone insertion, probing depths, width of keratinized gingiva, furcation involvement, and height of the bone crest.⁵ The second study used the same method and evaluated plaque index, papillary bleeding index, and probing depth.¹⁰ The third study, using computed tomography images, evaluated the alveolar height, bone thickness, displacement, and tooth inclination.⁸

The number of patients included in these studies ranged from 20 to 61 participants. The age range of the participants included in the clinical trials was similar, between 6.3 years and 13.2 years.

When evaluating the three articles, one⁸ reported sample loss and the reasons for it, with a loss of 39.9% for RME and 48.4% for SME. In one study,⁵ there was evaluation of patients' medical records, seeking information on age at the start of treatment and at the time of evaluation, premolar extraction, and facial type, among others.

Risk of Bias Within Studies

Assessment of risk of bias of the studies included in this systematic review was performed using the Cochrane Collaboration's tool in RCTs (Figure 2) and the ROBINS-I tool in nonrandomized studies of interventions (Table 3).

View Figure

• Download as Powerpoint
• Download Figure

Table 3 Risk of Bias According to the Cochrane Collaboration's Tool and ROBINS-I

View Fullscreen

One full-text study reported the method of randomization employed and how the allocation concealment was performed. This study⁸ was considered as having a “low” risk of bias according to the Cochrane Collaboration's tool, even for blinding of outcome assessment, because it was clear that the periodontal evaluation was performed by a blind operator.

In the other two articles,^5,10 the randomization method used for sample selection was not declared, and this information was confirmed after contact by email with the authors.^8,10 Because of this, the risk of bias followed the protocol recommended by the ROBINS-I tool. One article was scored as having a low risk of bias.¹⁰ For the other one, as information on bias in the measurement of outcomes was not clear, the overall classification was moderate risk of bias.⁵

The articles presented appropriate statistical tests for the data analyzed, and they did not have any other data that could lead to an increase in the risk of bias.

Results of Individual Studies

Two studies did not find significant differences in periodontal changes between the two expansion treatments,^5,10 and one reported increased periodontal bone loss with changes in height and thickness for slow expansion.⁸

A summary of the description of the studies included and evaluated is presented in Table 2.

Synthesis of Results

Because of the heterogeneity of the periodontal evaluation methods employed and the units of measurement (continuous and categorical data) used in the related articles found, the conduct of a meta-analysis was not justified and would not have allowed meaningful comparisons. Only simple and descriptive comparisons are reported.

Assessment of the Quality of Evidence

The GRADE evidence profile table is described in Table 4. For the outcome alveolar bone level (height and thickness) and tooth displacement and inclination, the GRADE quality of evidence was graded as moderate because of “serious” limitations in imprecision, because a narrative synthesis was conducted, because of the impossibility of performing meta-analysis, and because the estimates are not precise. In all other outcomes, the GRADE quality of evidence was judged as very low due to “serious” limitations in imprecision and risk of bias, since the information on bias in the measurement of outcomes was not clear in these observational studies.

Table 4 GRADE Evidence Profile Table: Should RME Versus SME Be Used for Maxillary Atresia?

View Fullscreen

Summary of Evidence

Statistical differences were found when comparing the measurements and the periodontal indices obtained in the RME and SME groups for the main variables in the studies selected.^5,8,10 However, none could be considered clinically relevant, since there was no definitive conclusion about which type of expansion was appropriate regarding the periodontal aspects.

Three articles were included in this systematic review, and methodological issues were identified. Regarding classification of the articles and their score for the Cochrane Collaboration's tool, one study was rated as having a low risk of bias for all key areas, mainly because there was an adequate method of randomization and allocation.⁸ The other two articles were scored as having low and moderate risk of bias according to the ROBINS-I tool.^5,10

The quality of evidence of clinical outcomes was also graded using the GRADE tool. The evidence scored for the alveolar bone level outcome, which included alveolar bone height and thickness, was moderate for the RCT study and very low for the observational study, since it was not possible to conduct a meta-analysis and narrative synthesis was conducted, and the estimates are not precise. The imprecision evaluation was classified as “serious.” A moderate evidence was scored for tooth displacement and inclination, again because of a “serious” imprecision. For potential periodontal irritation effects, there was a “serious” evidence level for imprecision and the risk of bias outcome, since the information on bias in the measurement of outcomes was not clear in the ROBINS-I tool for one article.⁵

Heterogeneity was observed for the periodontal evaluation methods and units of measurement used (continuous and categorical data). In the study by Mummolo et al.,¹⁰ clinical periodontal indices were used to assess gingival and periodontal health. Statistical differences occurred between rapid and slow maxillary groups when considering plaque index and papillary bleeding index, suggesting that the palatal expander, as well as fixed orthodontic appliances in adolescents, caused irritation side effects in the periodontium.²³ Greenbaum and Zachrisson⁵ measured periodontal parameters, such as alveolar bone height, attachment level, probing depth, and width of keratinized gingiva in millimeters, comparing rapid and slow maxillary groups with a control group not undergoing expansion. They concluded that, even with statistically significant differences for most of the variables included, with more damage resulting from the rapid protocol, the differences were not substantial enough to warrant a better treatment for maxillary atresia correction. Finally, Brunetto et al.⁸ took cone-beam computerized tomography images to evaluate, quantitatively, periodontal aspects related to the expansion protocol, such as buccal displacement of the maxillary first permanent molars, which was greater for the slow protocol; inclination movement of anchorage teeth, which was more intense for the rapid protocol; and bone loss, detected in both groups but with a more significant result for slow expansion. The authors concluded that the activation frequency of the palatal screw might have influenced dental and periodontal aspects and that there were differences in dental movements and periodontal side effects between RME and SME. The inclination movements of supporting teeth in RME were in agreement with another cone-beam computerized tomography study,⁹ which showed reduced buccal bone plate and increased lingual bone plate thickness of supporting teeth, as well as bone dehiscence on the buccal aspect of anchorage teeth. All of these features were demonstrative of tooth inclination movement.

The main conclusion of the studies selected was that it is necessary to standardize the methods of measurement as well as to randomize the sample to obtain a sound conclusion. This would make studies more clinically effective and able to help in decision making about which type of expansion is less harmful to the periodontal tissues.

Limitations

One article included in this systematic review⁸ had a 39.9% sample loss in the RME group and 48.4% sample loss for the slow protocol. Even considering high values of sample loss, the power of analysis of variance was calculated, and a value of 76% was obtained for intergroup comparison. Therefore, the article was rated as having a low risk of bias in this domain.

Another factor to be considered is the use of the same appliance, with a different activation protocol, in two studies.^8,10 Perhaps using a such a bulky appliance for a slow expansion protocol was excessive, since plaque accumulation and difficulties in oral hygiene can contribute to further damage of periodontal tissues and supporting teeth. However, this gave a more reliable finding of the approach itself.

Besides the type of expansion and the appliance selected, which teeth are banded must be considered as an important variable, since the process of deciduous tooth exfoliation can affect the interpretation of results. It is known that when expansion treatment is performed in the deciduous or mixed dentition, with bands on the deciduous molars, the result is better because the orthopedic effect is more favorable at early ages.²⁴ However, deciduous canines and molars, the anchorage teeth, can be involved in periodontal terms. In this instance, the permanent teeth would erupt in an area with new bone, reestablishing the alveolar area.⁹

In this systematic review, two studies used an expander appliance anchored to the first permanent molars.^5,8 The third study used a Haas expander anchored to the second deciduous molars with bands and bonded to the deciduous canines using acrylic resin.¹⁰ However, contradictorily, one of the periodontal indices, pocket probing depth, was measured at the first permanent molar.

Other important variables to be evaluated in more powerful, future research, that could affect the results are age at the initiation of treatment and at examination, extraction cases, sagittal movement of the maxillary first molar during treatment, the length of time in bands and retention, and the duration of the postretention period.⁵ Thus, a representative random sample, standardization of the appliance and anchoring teeth, age range, and a consistent method of evaluation are necessary for future studies to provide more sound clinical evidence.

Clinical Considerations

The lack of sufficient evidence makes it impossible to indicate which type of expansion protocol should be preferred. Perhaps the choice depends on the main desired effect (orthopedic or orthodontic) rather than on the appliance that would be less harmful to the periodontal tissue.