Background
External apical root resorption (EARR) is a reduction in root structure involving the apices, and it is a common phenomenon of orthodontic treatment in the modern world [
1]. Most resorption is clinically insignificant, but severe root resorption threatens tooth longevity and causes tooth mobility or loss [
2]. With improvements in orthodontic techniques and increased patient expectations, orthodontists need to be aware of EARR [
3].
The prevalence of EARR is high, and the factors affecting it are complex and multiple, including internal and external factors. Internal factors are patient factors that include genetics, age at the start of treatment, gender, nutrition, root morphology, alveolar bone density, type of malocclusion, and so on [
4‐
6]. External factors are primarily caused by orthodontic treatment, such as the type of appliance, treatment technique, continuous or intermittent force, force magnitude and direction, duration of the applied force, premolar extractions, tooth distance and root movement are risk factors for EARR [
7‐
9]. The causes and mechanisms of resorption are not completely clear.
Different aspects of tooth resorption, including prevalence and degree, were investigated using conventional radiographs. Conventional radiographs include periapical film, panoramic radiograph, and lateral cephalometric images. Image distortion and magnification are common characteristics of panoramic radiography, also known as non-positioning radiographs, and this technique imprecisely measures cephalometric points [
10,
11]. The disadvantages of this approach include confounded images caused by superimposed anatomic structures and a lack of right- and left-side information [
12].
However, root resorption occurs 3-dimensionally, and 2D images cannot detect root resorption on lingual or buccal surfaces nor measure the volume of root loss. Therefore, quantification of treatment outcome using 2D images raised some criticism because of its reliability.
CBCT is an effective imaging method for the diagnosis of orthodontic root resorption using a 1: 1 ratio for reconstruction with no amplification error [
13]. CBCT clearly shows the root structure, which results in more accurate qualitative judgment of orthodontic root resorption [
14,
15]. CBCT images enhance cross-section research in three dimensions because the images may be observed at any angle using 3D reconstruction. Therefore, studies on RR using CBCT demonstrated improved accuracy and sensitivity in comparison with those using 2D data [
16]. CBCT data contain equal image information of the right and left sides with no interference due to image overlap. Wang demonstrated that CBCT accurately measured tooth and root resorption volumes, and it was a more accurate and reliable 3D measuring method for EARR investigation [
17]. Another significant advantage of CBCT in root resorption studies is that it could be used in vivo, compared with Micro CT.
Weltman et al. [
4] and Roscoe et al. [
18] systematically reviewed root resorption associated with orthodontic treatment based on 2D images, but they did not do quantitative synthesis. There are no systematic reviews of root resorption associated with orthodontic treatment based on CBCT data, which is a more accurate and scientific method [
19]. Therefore, it is necessary to integrate the data and conclusions of these trials. The purpose of this article is to report the results of a rigorous systematic review of the scientific literature relating to EARR in patients with fixed orthodontic appliances using the most accurate imaging information, CBCT.
Methods
This meta-analysis was performed in accordance with the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) checklist and PRISMA harm checklist items.
Types of studies
Study design: Randomized and non-randomized controlled trials, clinical trials, and prospective and retrospective reports were included. Longitudinal studies that observed root changes at different time points of treatment (before and after orthodontic treatment) were included. Self-controlled studies were included. Case reports, case series studies, descriptive studies, opinion articles and reviews were not included.
Types of participants
We included studies of orthodontic patients with no restrictions in the characteristics of occlusion, age or gender, and with available pre-and post-operative CBCT data. Patients with periodontitis were excluded. Pregnant patients and patients with systemic diseases, syndromes, pathologies, or history of root resorption were excluded.
Types of interventions
For comprehensive orthodontic treatment, patients in permanent dentition with fixed appliances, such as brackets and bands were included. Patients with different wire techniques and orthognathic surgery patients with pre-operative orthodontics with extraction treatment (bicuspid extraction on the upper and/or lower arch) or non-extraction were also included. Patients with local orthodontic treatment or stage treatment were excluded.
Types of outcome measures
Primary outcomes: Root resorption was evaluated using CBCT after orthodontic treatment. The primary outcomes were tooth/root length and tooth/root volume.
The PICOS format and null hypothesis are shown in Table
1.
Table 1
PICOS format and null hypothesis
Population | Patients with orthodontics |
Intervention | Comprehensive orthodontics; not local orthodontics |
Comparison | Before and after treatment |
Outcome | Root resorption evaluated as tooth/root length and volume assessed using radiographic imaging CBCT |
Null hypotheses | There is no difference in the incidence and severity of root resorption before and after comprehensive orthodontic treatment. |
Search methods for study identification
For the identification of studies to include or consider for this review, we developed detailed search strategies for each database searched. These strategies were based on the search strategy developed for MEDLINE but revised appropriately for each database. We searched the following databases: the CNKI database, the Cochrane Library, Web of Science, PubMed and EMBASE (to October 2017). We used no language or date restrictions in the searches of the electronic databases. The key words used to screen the databases are shown in Table
2. Citations of the remaining studies were examined to identify publications not located in the MEDLINE database. We contacted the authors of randomized controlled trials to identify any unpublished trials.
CNKI | Subject = root resorption AND Subject = orthodontic AND Subject = CBCT (accurate match) | 103 |
Cochrane | Root resorption: ti, ab kw and Orthodontics: ti, ab, kw and cone beam computed tomography: ti, ab, kw | 8 |
Web of Science | TS = ((root resorption) AND orthodontics AND (CBCT OR (Cone Beam Computed Tomography))) | 34 |
PubMed | (Cone Beam Computed Tomography))) (root resorption) AND (root resorption) AND (orthodontics OR orthodontic) AND (CBCT OR (Cone Beam Computed Tomography)) | 132 |
EMBASE | ‘tooth disease’ AND ‘orthodontics’ AND ‘cone beam computed tomography’ AND [1–1-1966]/sd NOT [30–9-2017]/sd AND [1966–2017]/py | 178 |
Data collection and analysis
Methodological quality assessment
Two independent authors (Deng and Sun) assessed the quality of each study included in the meta-analysis using Methodological Index for Non-randomized Studies (MINORS). Evaluations were compared, and any inconsistencies between the review authors were discussed and resolved. For the self-controlled studies, the MINORS scores ranged from 9 to 15 out of a possible score of 16 (Table
3). There were no clear and consistent inclusion criteria for the included studies, but they were identified as moderate scientific evidence considering their prospective properties and the consecutive inclusion of participants.
Table 3
Methodological index for non-randomized studies (MINORS)
Statistical analysis
Two authors independently screened the eligible studies, assessed the risk of bias in the trials and extracted data. The following outcomes of interest were recorded: tooth/root length and tooth/root volume. We calculated the mean differences (MD) with 95% confidence intervals (CI) for continuous data and risk ratios (RR) with 95% CI for dichotomous outcomes. Heterogeneity was tested using Cochrane’s Q-test. I
2
> 50% was defined to indicate significant heterogeneity (I
2
-value superior to 25, 50 and 75% corresponding to low, medium and high heterogeneity, respectively). Meta-analyses were performed using Review Manager 5.3 (Nordic Cochrane Centre, Copenhagen, Denmark). If the studies used similar participants and similar interventions, the fixed-effect model was used; if there was potential heterogeneity among studies, we preferred to use random-effects models. If sufficient data were available, we performed the following subgroup analyses: position of tooth; different intervention; extraction group and non-extraction group. If there were insufficient clinical trials for specific interventions or insufficient data for extraction, we qualitatively described the results. We then evaluated the influence of each subgroup on heterogeneity using forest plot analysis. Publication bias was assessed using funnel plot analysis.
Discussion
Summary of main results
Evidence suggests that orthodontic treatment causes an increased incidence and severity of apical root resorption. Tooth length and root volume were reduced after orthodontic intervention.
Overall completeness and applicability of evidence
Numerous CBCT studies investigated root resorption before and after orthodontic treatment, but quantitative data of the extraction were only available in 12 articles for a meta-analysis. We chose two major indicators, tooth/root length and root volume, to reflect root resorption. Other indicators that reflect changes were not included, and numerical integration and meta-analysis were not performed. Quantitative analysis of other indicators may be performed using numerical analyses to improve the data on this issue.
Quality of the evidence
Twelve articles were included in this meta-analysis, and all of the studies were non-randomized controlled trials. Randomized controls cannot be used because it is ethically impossible to perform CBCT screenings in patients who have not undergone orthodontic treatment. All studies were self-controlled, which reflects the impact of intervention on these patients more accurately. All patients were diagnosed and measured using CBCT, which excluded the shortcomings of magnification and distortion in 2D image data. All data sources were evaluated more accurately and reproducibly by CBCT, which is the most accurate method of obtaining data in vivo. Methodological evaluation scores from 9 to 15, and 4 studies received scores that were above or equal to 13, which were considered as high quality. One study was below 10 and was considered low quality. The remaining studies were of medium quality. The overall quality of the selected literature was good. There were 3 prospective trials, and the remaining trials were retrospective experiments. Prospective experimental evidence will provide more adequate data.
Potential biases in the review process
The different orthodontic technologies, such as straight wire and edge wise, and corticotomy, incompleteness of some of the reports and lack of quality control in some trials may have contributed to bias in study assessments. We made every attempt to limit bias in the review process by ensuring a comprehensive search for potentially eligible studies. Time limitations prevented the search of additional databases and sources, which may have identified additional published and unpublished studies. There may also be publication bias because of the lack of publication of negative results. We strictly controlled the inclusion of exclusion criteria. There were only two prospective experiments, while others were retrospective studies. There may be biases in sample selection and dropout. The sample size of all studies was relatively small, and only one of the studies calculated the sample size [
27].
The age distribution did not completely cover the age of patients with orthodontic treatment, but it covered the span of treatment for most patients. However, there were few samples for adolescent studies, and many patients of orthodontic treatment are adolescents in clinical practice. Therefore, whether the conclusions of this meta-analysis are well suited for adolescent samples must be further studied.
Primary outcome tooth length
Tooth position
Total tooth root length was reduced after orthodontic treatment, and the included studies exhibited a small publication bias. The effect value indicated that the maxillary lateral incisors were the most absorbed, followed by the maxillary central incisors, mandibular anterior teeth and upper canines. Nanekrungsan et al. [
8] found that the maxillary lateral incisors were primarily reduced after treatment. Yu et al. [
30] found that the maxillary lateral incisors exhibited greater resorption than maxillary central incisors and canines using CBCT. Kennedy et al. [
31] also found that the maxillary lateral incisors were more prone to root resorption than the central incisors. Pejicic et al. [
32] found that lateral incisors were primarily affected, and mean values ranged from 0.5 mm to 3 mm, which is consistent with the conclusions of our study. Previous 2D studies [
3,
33,
34] found that maxillary central incisors were the most affected teeth. Sameshima et al. [
11] demonstrated that the absorption order was the upper central incisors, the upper lateral incisors, the lower central incisors, and the lower lateral incisors. Jung et al. [
35] found that the maxillary central incisors were the most resorbed, with 27% undergoing greater than 1 mm of root resorption. Inaccuracies caused by the magnification and overlap of 2D data, the different types of patient malocclusion, or the different treatment methods may explain the differences in absorption of the anterior teeth.
A meta-analysis of Segal et al. [
36] was based on 2D data and demonstrated a strong correlation between root resorption and apical displacement. The mean resorption of upper central incisors was 1.421 ± 0.448 mm, which was slightly higher than the present study (MD = 0.84,CI[0.56,1.12]). Two-dimensional data may over-estimated root resorption.
The present meta-analysis revealed that the root resorption of the upper central incisors and upper lateral incisors were similar, and it was difficult to determine which tooth was the most affected. The sample of 3D studies was quite small. Therefore, larger 3D sample sizes and more clinical trial evidence are required to supplement and confirm these conclusions.
The heterogeneity of root resorption of the mandibular anterior teeth was high partially because two studies measured the root length and other studies measured tooth length. Removal of the data from control group of Wang et al. [
21] reduced the heterogeneity to zero, which suggested that this data was the source of heterogeneity. This group was pre-operative orthodontic of orthognathic surgery patients, which resulted in greater root resorption (1.55 ± 0.66 mm). Experimental group was also pre-operatively compensated, but corticotomy was added. A previous study demonstrated that corticotomy reduced the duration of treatment and resulted in lower root resorption compared to traditional methods. Therefore, the value of experimental group may be reduced.
The key of orthodontics prior to orthognathic surgery lies in the anterior teeth for compensation [
37]. The root much more easily touches the alveolar bone, which increases the risk of root resorption [
38]. Some scholars confirmed that maxillary root resorption is most likely to occur in the patients with bony malocclusion because of the maximum distance of tooth movement, especially the incisors [
39,
40]. Therefore, the data suggest that orthopedic orthodontic treatment of skeletal malocclusion using simple orthodontics will cause more root resorption, regardless of cortical osteotomy intervention.
Primary outcome tooth volume
Root resorption is not just a two-dimensional change in length, it actually occurs in three dimensions, including on the buccal-lingual and mesial-distal sides. Therefore, the volume index is more realistic than the length index, which reflects the root resorption accurately. All three studies [
20,
26,
29] focused on the maxillary and mandibular incisors and were consistent with no heterogeneity in the length index.
Factors affecting root resorption
Clinical significance
It is necessary to inform patients of the risk of root resorption prior to orthodontic treatment. Significant shortening of the root length will lead to inappropriate crown-root ratios and adjacency to the periodontal tissue. Apical resorption greater than 3 mm is equivalent to 1 mm of bone loss, which will accelerate the periodontal disease process. The treatment program must be re-evaluated if serious root resorption is found. Future experiments should be based on a reasonable sample size and the implementation of the entire orthodontic process to ensure the effect value.
The above meta-analysis demonstrated that the average value of root resorption was approximately 1 mm, and the upper limit did not exceed 2 mm. Sharpe’s grading standards indicated that the first grade was slightly blunt (1–2 mm) [
51]. The results were grade 1. Malmgren’s rating standard indicates that level 1 is irregular root contour and level 2 is root resorption apically amounting to less than 2 mm (minor resorption) [
40]. The most serious case did not exceed 2 degrees, and it was a mild absorption.
Conclusions
The following conclusions are drawn based on the existing research:
1.
Evidence suggests that orthodontic treatment increased the incidence and severity of apical root resorption. Tooth length and root volume were reduced after orthodontic intervention, but these changes were in a clinically acceptable range.
2.
Different tooth positions exhibited different degrees of absorption, and the sequence of RR from heaviest to lightest was maxillary lateral incisors, maxillary central incisors, mandibular anterior teeth, and maxillary canines.
3.
Tooth extraction may result in more root resorption than non-extraction.
4.
Most of the patients measured using CBCT exhibited root resorption within the clinically acceptable range. The RR value of CBCT was lower than the 2D data.
However, more methods are needed to provide more reliable evidence for clinical trials based on CBCT data. Experimental studies on treatment, the distance and angle of tooth movement, and correction techniques may provide more clinical guidance for orthodontic treatment of RR reduction.