Three-Dimensional Evaluation Effects of Microimplant-Assisted Rapid Palatal Expansion on the Upper Airway Volume: A Systematic Review and Meta-Analysis

Microimplant-assisted rapid palatal expansion is increasingly used clinically; however, the effect on the upper airway volume in patients with maxillary transverse deficiency has not been thoroughly evaluated yet. The following electronic databases were searched up to August 2022: Medline via Ovid, Scopus, Embase, Web of Science, Cochrane Library, Google Scholar, and ProQuest. The reference lists of related articles were also reviewed by manual search. The Revised Cochrane Risk of Bias Tool for randomized trials (ROB2) and the Risk of Bias in non-randomized Studies of Interventions (ROBINS-I) tool were used to evaluate the risks of bias of the included studies. The mean differences (MD) and 95% confidence intervals (CI) of changes in nasal cavity and upper airway volume were analyzed using a random-effects model, and subgroup and sensitivity analyses were also performed. Two reviewers independently completed the process of screening studies, extracting data, and assessing the quality of studies. In total, twenty-one studies met the inclusion criteria. After assessing the full texts, only thirteen studies were included, with nine studies selected for quantitative synthesis. Oropharynx volume increased significantly after immediate expansion (WMD: 3156.84; 95% CI: 83.63, 6230.06); however, there was no significant change in nasal volume (WMD: 2527.23; 95% CI: −92.53, 5147.00) and nasopharynx volume (WMD: 1138.29; 95% CI: −52.04, 2328.61). After retention a period, significant increases were found in nasal volume (WMD: 3646.27; 95% CI: 1082.77, 6209.77) and nasopharynx volume (WMD: 1021.10; 95% CI: 597.11, 1445.08). However, there was no significant change after retention in oropharynx volume (WMD: 789.26; 95% CI: −171.25, 1749.76), palatopharynx volume (WMD: 795.13; 95% CI: −583.97, 2174.22), glossopharynx volume (WMD: 184.50; 95% CI: −1745.97, 2114.96), and hypopharynx volume (WMD: 39.85; 95% CI: −809.77, 889.46). MARPE appears to be linked with long-term increases in nasal and nasopharyngeal volume. However, high-quality clinical trials are required to further verify the effects of MARPE treatment on the upper airway.


Introduction
Maxillary transverse deficiency is a common malocclusion characterized by dental crowding, narrow nasal cavity, high palatal vault, and unilateral or bilateral posterior crossbite [1,2]. Compared with normal individuals, patients with maxillary constriction often have narrow airways [3]. It is known that the decrease in the amount of air passing through the nasal cavity into the lungs can affect craniofacial growth and development as well as overall health [4]. Moreover, some studies have reported that maxillary transverse deficiency can be a possible cause of obstructive sleep apnea (OSA) [5,6]. 2

of 21
Rapid palatal expansion (RPE) is often used to correct the narrow maxilla by separating the midpalatal suture. However, because the skeletal resistance of the midpalatal suture gradually increases with age [7] and some side effects may occur after RPE treatment, such as anchored teeth root resorption [8], dehiscence, and fenestration of the buccal cortex [9], traditional rapid palatal expansion in mature patients is still questionable. Surgicallyassisted rapid palatal expansion (SARPE) has been proposed to achieve skeletal expansion in mature patients with transverse maxillary deficiencies [10]. However, complications and the high cost of surgery disincline some patients from choosing this treatment [11]. Microimplant-assisted rapid palatal expansion (MARPE) was suggested as an alternative to SARPE [12]. Orthodontic micro-implants serve as a skeletal anchorage for RPE, which not only produce more skeletal expansion while reducing adverse dentoalveolar effects, but also can reduce surgical injury [13].
At present, there are many meta-analyses and systematic reviews on the two-dimensional width and three-dimensional volume changes of the upper airway after rapid palatal expansion [14][15][16][17][18][19]. However, only two studies synthesize on the changes of upper airway dimension in MARPE. Krüsi et al. [20] only described the width of the nasal cavity after MARPE. Abu Arqub et al. [21] included only three studies to describe the short-term changes of the upper airway dimension after MARPE but without quantitative analysis, which did not provide solid evidence of the relationship between airway changes and MARPE treatment. Hence, a thorough systematic evaluation of the clinical evidence related to the short-term and long-term changes of the upper airway volume after MARPE is needed to better understand the effects of MARPE on the dimension and function of the upper airway and to determine if the therapy is helpful for patients with airway stenosis.

Protocol and Registration
We used the Cochrane Handbook to perform the review [22], and the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guideline to report our results [23]. The protocol was registered on the PROSPERO database (number CRD42020198286).

Eligibility Criteria
We developed the following inclusion criteria using the principles of population, intervention, comparison, outcome, and study design (PICOS): Participants (P): patients with narrow maxilla needed maxillary expansion treatment. We have no age restrictions for patients who are included. Intervention (I): micro-implant-assisted rapid palatal expansion. Comparison (C): age-and sex-matched patients treated with RPE or SARPE, patients without maxillary expansion treatment, or comparison of the same patients before and after MARPE. Outcome (O): changes in upper airway volume, assessed by CT or CBCT. Study design (S): randomized controlled trials or non-randomized studies. The exclusion criteria included: (1) case reports, animal studies, and reviews. (2) Studies that applied another treatment or auxiliary surgery during micro-implant-assisted rapid palatal expansion. (3) Studies that included patients with craniofacial abnormalities (cleft lip and /or palate). (4) Studies without using CBCT or CT to measure the upper airway volume.

Information Sources and Search Strategy
The published literature was searched in the electronic databases Medline via Ovid, Cochrane Library, Embase, Web of Science, and Scopus, and the databases were searched up to August 2022.The reference lists of related articles were also reviewed to find any probable articles that may be missed during the electronic database searches. No restrictions on the year of publication or language. For grey literature, Google Scholar and ProQuest were searched. The search was performed independently by two authors. The search strategy for each database was shown in Table S1.

Study Selection
Two authors performed the selection process independently. The authors reviewed the titles and abstracts of retrieved articles. When no abstract was available or the abstract do not contain sufficient information, the full text of articles that met the inclusion criteria would be obtained to review. Authors would be contacted if additional information was needed. The reference lists of related studies were also reviewed in the same way to find other articles that met the eligibility criteria. Any discrepancy between the two authors was settled by discussion with a third author.

Data Items and Collection
Two authors used the data extraction form to extract the data independently. Details of the included studies were collected, including authors, year of publication, and study design. Information about the study samples, including the number of participants, gender, age, sample inclusion criteria, control group setting, as well as measure method of the upper airway volume, were recorded. Additionally, the authors retrieved the type of expander and expansion protocol, timing, and retention details from all included studies, and upper airway assessment methods, parameters used for CBCT or CT, and software used for image reconstruction also were recorded. The follow-up points were defined as: T0, before expansion; T1, immediately after expansion; T2, three months retention after expansion; and T3, six months retention after expansion. The volumetric changes in any region of the upper airway and upper airway boundary used in studies were recorded. Anatomically, the upper airway is divided into nasal cavity, paranasal sinuses, nasopharynx, oropharynx, and hypopharynx [24].

Risk of Bias in Individual Studies
Two authors evaluated the risk of bias for randomized studies using the Revised Cochrane Risk of Bias (ROB2) tool [25]. For non-randomized studies, the Risk of Bias in Non-randomized Studies-of Interventions (ROBINS-I) tool [26] was used to assess the quality of the included studies. Any disagreements between the authors were resolved through conversation with a third author.

Summary Measures and Approach to Synthesis
Changes in upper airway volume in each segment were regarded as the primary outcome. This systematic review only included studies measuring the upper airway volume with CBCT or CT, which minimized the differences among different studies. Stata MP 16.0 (Stata Corp, College Station, TX, USA) software was used for statistical analysis. All the indicators included in this meta-analysis were continuous variables. The mean differences (MD) and the associated 95% confidence intervals (CI) were calculated for all meta-analyses.
The heterogeneity test was conducted on the effect values of independent studies. If the heterogeneity was significant, the random effects model was used to combine the effect value according to DerSimonian and Laird [27]. Subgroup analysis was used in the meta-analysis to identify possible sources of heterogeneity. Sensitivity analyses were planned for the meta-analyses to assess their robustness. Egger's test would be used to assess publication bias if enough trials were included in this meta-analysis. A two-tailed value p = 0.05 was considered significant for hypothesis testing, but a two-tailed value p = 0.10 was used for heterogeneity testing and publication bias [28].

Risk of Bias Assessment across Studies
The overall quality of evidence was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) [29]. Two authors performed this analysis independently. Disagreements are resolved in discussion with the third author.

Study Selection
In total, 1649 articles were initially identified by means of searching databases (Table S2), and one additional study was retrieved through other sources. In total, 1218 articles were evaluated after excluding duplicates. By reviewing titles and abstracts, 1197 articles were excluded. Thus, the remaining 21 articles were eligible for the full-text assessment. Eight articles were excluded with reasons after reviewing the full texts (Table S3), and 13 articles were eventually included for the qualitative analysis. From these, a total of nine studies were included in the quantitative synthesis, as four studies reported incomplete data. The flow diagram of the study selection process is presented in Figure 1.

Risk of Bias Assessment across Studies
The overall quality of evidence was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) [29]. Two authors performed this analysis independently. Disagreements are resolved in discussion with the third author.

Study Selection
In total, 1649 articles were initially identified by means of searching databases (Table  S2), and one additional study was retrieved through other sources. In total, 1218 articles were evaluated after excluding duplicates. By reviewing titles and abstracts, 1197 articles were excluded. Thus, the remaining 21 articles were eligible for the full-text assessment. Eight articles were excluded with reasons after reviewing the full texts (Table S3), and 13 articles were eventually included for the qualitative analysis. From these, a total of nine studies were included in the quantitative synthesis, as four studies reported incomplete data. The flow diagram of the study selection process is presented in Figure 1.

Study Characteristics
Tables 1-3, S4 and S5 present the main information from the included studies in the present systematic review. In terms of study design, three studies were prospective clinical studies [30][31][32] and nine studies [33][34][35][36][37][38][39][40][41] were retrospective studies. The remaining was a randomized controlled trial [42]. All studies had a total of 244 participants, 117 women and 61 men. One study only included male patients [31] and three did not report a male-female ratio [32,38,39]. Patients ranged in age from 8 to 35 years, although one study did not report any age information [41]. Six studies had control groups [31,[36][37][38]41,42], but all of them compared different expanders. For the inclusion criteria, all patients had maxillary transverse deficiency, and the inclusion criteria of one study [32] was OSA patients who had obstructive sleep apnea with maxillary transverse deficiency. A maxillary skeletal expander with two micro-implants on each side in a mid-palatal suture to provide support was applied in seven studies [30,32,35,36,[39][40][41]. One study applied the palatal bracket implant anchorage arch expander [34]. One study used modified conventional four-banded hyrax expander [33]. Two studies used the bone-borne rapid maxillary expander with one microimplant on each side of the posterior dental region in palatine to provide support [37,38], and two studies used the hybrid hyrax expander with one implant on each side of the anterior dental region in palatine to provide support [31,42]. All studies reported the expansion protocol; however, the expansion protocol varied. The most common is 0.5 mm/day. Other studies applied 0.2 mm [33], 0.64 mm [36], or 0.25 mm [32] per day, and 0.26 mm every other day [35]. Five studies reported only the extent or amount of eventual expansion without specifying expansion protocol [31,[38][39][40][41]. CT images were used to measure upper airway volume in only one of the studies [31], and CBCT was used in the rest. The scanning parameter settings used in the included studies were different, but they all included FOV, voxel size, scan time, KV, and mA. Software packages for 3D image reconstruction included Dolphin software [34,35,[37][38][39][40]42], Mimics software [31,32], Invivo software [32], OnDe-mand3D [30,33,36], and Amira software [41]. All studies recorded data measurements at time points. Changes in the upper airway volume were measured in seven studies immediately after expansion (T1) [30,31,33,36,39,41], four studies at three months after expansion (T2) [34,35,37,40], and two studies at six months after expansion (T3) [32,38,42]. The included studies measured volumetric changes in any region of the upper airway, including the nasal cavity, nasopharynx, oropharynx, hypopharynx, and maxillary sinus. However, the delineation of upper airway boundaries varied among studies.

Risk of Bias in Individual Studies and across the Studies
The risk of bias assessment results for randomized and non-randomized studies were shown in Fgures 2, 3 and S1. In total, there were 13 studies, of which was a randomized controlled trial [42] and the other 12 were non-randomized studies [30][31][32][33][34][35][36][37][38][39][40][41]. The ROBINS-I tool was used to assess the 12 non-randomized studies. The outcome of the assessment revealed a serious risk of bias, mostly due to bias in confounding, participant selection, classification of the intervention, and measurement of outcomes. The randomized controlled trial was assessed using ROB2 tool and the result showed some concerns. Of the five assessment domains, the domain in bias due to randomization showed some concerns, and the remaining four domains showed low risk of bias. The GRADE rating of the quality of evidence for the meta-analysis was presented in Table S6. For quality rating, nasal cavity volume after retention were low, and the rest of outcomes was moderate. The main reason for lowering the quality of the evidence was with serious risk of bias, and inconsistency also occurred in some outcomes.

Risk of Bias in Individual Studies and across the Studies
The risk of bias assessment results for randomized and non-randomized studies were shown in Figures 2 and 3, and Figure S1. In total, there were 13 studies, of which was a randomized controlled trial [42] and the other 12 were non-randomized studies [30][31][32][33][34][35][36][37][38][39][40][41] The ROBINS-I tool was used to assess the 12 non-randomized studies. The outcome of the assessment revealed a serious risk of bias, mostly due to bias in confounding, participan selection, classification of the intervention, and measurement of outcomes. The randomized controlled trial was assessed using ROB2 tool and the result showed some concerns Of the five assessment domains, the domain in bias due to randomization showed some concerns, and the remaining four domains showed low risk of bias. The GRADE rating of the quality of evidence for the meta-analysis was presented in Table S6. For quality rating nasal cavity volume after retention were low, and the rest of outcomes was moderate. The main reason for lowering the quality of the evidence was with serious risk of bias, and inconsistency also occurred in some outcomes.

Risk of Bias in Individual Studies and across the Studies
The risk of bias assessment results for randomized and non-randomized studies were shown in Figures 2 and 3, and Figure S1. In total, there were 13 studies, of which was a randomized controlled trial [42] and the other 12 were non-randomized studies [30][31][32][33][34][35][36][37][38][39][40][41]. The ROBINS-I tool was used to assess the 12 non-randomized studies. The outcome of the assessment revealed a serious risk of bias, mostly due to bias in confounding, participant selection, classification of the intervention, and measurement of outcomes. The randomized controlled trial was assessed using ROB2 tool and the result showed some concerns. Of the five assessment domains, the domain in bias due to randomization showed some concerns, and the remaining four domains showed low risk of bias. The GRADE rating of the quality of evidence for the meta-analysis was presented in Table S6. For quality rating, nasal cavity volume after retention were low, and the rest of outcomes was moderate. The main reason for lowering the quality of the evidence was with serious risk of bias, and inconsistency also occurred in some outcomes.

Maxillary Sinus Volume
No study reported the changes of the maxillary sinus volume after expansion. However, the maxillary sinus volume was observed in two studies [37,42], one of which showed the total maxillary sinus volume increased by 10.0% after six months retention [42], the other showed that there was a 2.1% change in the left maxillary sinus volume and 5.2% change in the right maxillary sinus volume (Table 3).

Additional Analyses
Each outcome that included more than three studies was assessed for publication bias, with a total of six outcomes analyzed. According to Egger's test, no obvious publication bias was found (Table S8). Sensitivity analysis showed that the results were robust after one study was omitted (Table S9). However, due to the small number and the characteristics of the included studies, other sensitivity analyses could not be robustly performed.

Summary of Evidence
Cranial and maxillofacial malformations affecting upper airway volume have been proven to be one of the important causes of obstructive sleep apnea (OSA) [43]. Transverse deficiency of the maxilla is known as a contributor to the development of OSA [44]. Rapid maxillary expansion (RPE) is a common method to correct maxillary transverse deficiency in adolescents. However, in late adolescents and young adults with partial or complete ossification of the mid-palatal suture, MARPE is used to achieve more skeletal expansion [45]. MARPE not only opens the maxillary palatal suture, but also affects the upper airway to varying degrees. Brunetto et al. [46] conducted a study on 16 adult patients with OSA and showed that 6 months after receiving MARPE treatment, the drowsiness, oxygen saturation, and snoring duration were improved. Kim et al. [32] performed micro-implant-assisted palatal expansion on 14 children with OSA and found that MARPE can effectively treat OSA patients. By expanding the nasal and maxillary complex, the airway patency of the nose and pharynx can be enhanced, and pharynx collapse and the nasal airway can be improved eventually. However, due to the lack of research in this area, there are still many controversies about its effect and mechanism on the upper airway.
Traditionally, the upper airway was measured by 2D imaging (lateral cranial radiographs); however, measuring three-dimensional volume from 2D imaging has considerable limitations [47]. 3D imaging techniques (magnetic resonance imaging, computed tomography, and cone-beam computed tomography) allow airway analysis to be visualized and is more reliable [48,49]. All studies included in this systematic review adopted CBCT or CT 3D imaging technology, which can more intuitively and accurately reflect the changes of upper airway volume after MARPE treatment. Therefore, the purpose of this meta-analysis was to comprehensively analyze the studies on the changes of upper airway volume after micro-implant-assisted rapid palatal expansion for patients with maxillary transverse deficiency by CT and CBCT, so as to provide intentional clinical evidence for clinicians.
For nasal volume, there was no significant change immediately after expansion, which may be due to only two studies being included to conduct a meta-analysis. Moschik [36] mainly focused on MARPE regarding the movement pattern of the lateral wall of the nasal cavity. Although the study showed that the nasal volume increased after MARPE, it did not report data completely, so it was not included in the quantitative analysis. Significant changes in nasal cavity volume were observed after retention of a period of time, consistent with changes in nasal cavity volume after traditional rapid maxillary expansion [15] and surgically-assisted rapid maxillary expansion [50]. For total effect values, the nasopharyngeal volume also did not change significantly immediately after expansion, which may be related to the small number of studies included. Kim et al. [33] found no significant change in nasopharyngeal volume after MARPE, possibly because the oropharynx was included in Kim's study.
The oropharynx consists of the glossopharynx and the palatopharynx, so the volume of the oropharynx is closely related to the volume of the palatopharynx and the glossopharynx. Atia et al. [31] reported significant volume changes in the oropharynx compared to other studies. In Atia's study, all patients were male, and the upper airway space associated with the base of the tongue was subtracted to eliminate the change in airway volume caused by the change in tongue position, which may have resulted in a slight difference. Since some oropharyngeal anatomical structures, including the hyoid bone, tongue, and soft palates, are movable, the possibility of changes in the size and position of these structures due to the influence of gravity should be considered [51]. Tongue position could influence the dimensions of the oropharynx at the time of acquisition of the examination. The lack of control over the soft tissues related to breathing movements and swallowing and tongue positioning can be confounding factors when measuring the oropharynx volume using CBCT or CT, which can cause errors in interpreting the volume of this region.
Posture is considered to be an important determinant of upper airway size [52,53]. A study comparing changes in upper airway morphology in supine and upright positions found that there was indeed a difference in airway morphology between the two positions: when supine, the airway became significantly smaller, and its resistance increased [54]. In this systematic review, three studies reported supine position [35,39,40,42], two studies [32,34] reported upright position, and the remaining studies did not report patient posture, which may lead to some degree of heterogeneity. On lateral cephalograms, changing head position from natural head position (NHP) to twenty degrees increased the pharyngeal cross-sectional airway dimension and an increase of ten degrees of craniocervical inclination would lead to the pharyngeal airway space accordingly increased by approximately 4 mm [55,56]. Therefore, an alteration in head position can also influence the measurement of upper airway space. Most of the included studies used control measures, such as cranial localization, to minimize the measurement bias.
Furthermore, sagittal and vertical skeletal pattern can influence the pharyngeal airway dimension. In patients with ANB greater than 4 • , the airway was significantly narrower, and the nasopharyngeal and oropharyngeal airway were also negatively correlated with SN-MP angle [57]. Only two studies [35,42] attempted to examine the effect of vertical craniofacial patterns on the upper airway. Different skeletal patterns have different effects on the airway, which can lead to differences between individual outcomes. Further investigations with appropriately grouping patients are needed to more accurately assess upper airway changes after MARPE and provide convincing evidence about this research question.

Strengths and Limitations
This systematic review used the GRADE approach to assess of the quality of evidence and all steps of qualitative synthesis were performed in accordance with the PRISMA statement [23]. Besides, any literature screening, data extraction, and data analysis was carried out by the two authors and any disagreement was resolved by the third reviewer. All of this increases the accuracy and credibility of the results. Moreover, to our knowledge, this is also the first meta-analysis to comprehensively analyze changes in upper airway volume after MARPE treatment. The limitations of this systematic review are as follows: First, the sample size of quantitative synthesis in the meta-analysis is relatively small. The results may not provide strong evidence of a relationship between airway changes and MARPE treatment. Secondly, the studies included many observational studies, but only one randomized controlled trial, so the quality of the meta-analysis was low. In addition, there were differences in adopting expanders, patient's position during CBCT or CT photography, the measurement methods and indicators selected, and the retention time after palatal expansion, which may lead to certain clinical heterogeneity among the results of the studies and affect the results of the meta-analysis to some extent. Finally, most existing studies do not have a control group to reduce the confounding effects of normal growth. However, the maximum retention time included in this study was 6 months. If the upper airway of children aged 6-15 increased at a rate of 0.032 cm 3 / year [58], the effect of growth on upper airway volume would be small. Due to the limitations of this systematic review, the reliability of the meta-analysis results was reduced to some extent. It is suggested to carry out more high-quality, large-sample clinical trials in the future, strictly formulate research plans, unify measurement methods and indicators as far as possible, and obtain more scientific and objective data to provide more high-quality evidence for clinical research.

Conclusions
MARPE treatment might cause a long-term significant increase in nasal cavity volume and nasopharynx volume, while no significant changes were found in the oropharynx, palatopharynx, glossopharynx, hypopharynx, and maxillary sinus volume. However, given the limited number of existing studies and the problem of different degrees of heterogeneity, these results should be considered with caution. Well-designed and conducted randomized controlled studies are required to further explore this issue.