Introduction

Individuals with unilateral cleft lip and palate (CLP) generally present a transverse maxillary deficiency that calls for prompt correction. Usually, the treatment option offered to these patients is the rapid maxillary expansion (RME) [1]. Despite being widely discussed in the literature and existing some consensus about its importance, little is known about the potential adverse effects of this procedure on both dental and supporting structures.

In this sense, there is a strong concern about the possible development of external apical root resorption (EARR) following RME, but the studies on the impact of this procedure on root development of partially formed teeth have been neglected. This is a very pertinent matter considering that the treatment of transverse maxillary discrepancies is commonly performed in young patients, in whom the root is still immature.

Some scientific speculation and the general belief of the orthodontic community itself end up perpetuating the hypothesis that high magnitude forces on teeth with immature roots would lead to the interruption of the root development [2]. This theory has also been suggested by other authors, who attributed deleterious effects such as root shortening and dilacerations to orthodontic forces in mixed and permanent dentition of young individuals [3]. However, some authors claim that the treatment of partially formed roots would result in little or no resorption, despite using inadequate methods to assess root development [4, 5].

Moreover, one cannot extrapolate results of studies conducted on a non-CLP sample as standard for CLP patients. Many variations are observed in cleft subjects, syndromic or not, such as, supernumerary teeth, missing teeth, shape and size defects, enamel hypoplasia, among others [6,7,8]. These findings are an invitation to a thorough and independent investigation on subjects affected by this deformity.

To our knowledge, no previous work has been published about EARR in cleft patients undergoing orthodontic treatment. Likewise, data was not found on root development of CLP subjects submitted to both orthopaedic and orthodontic forces. The aim of this prospective clinical cohort study was to answer the following question: Are there any length changes in the mature and immature roots of the first permanent maxillary molars of cleft lip and palate subjects following rapid maxillary expansion (RME)?

Materials and methods

The present study has been approved by the Institutional Review Board of the Pontifical Catholic University of Minas Gerais. The sample of this cohort study consisted of 30 subjects (20 boys, 10 girls) with unilateral cleft lip and palate. The inclusion criteria were the presence of maxillary atresia, age between 8 and 15, erupted first permanent molars, no previous orthodontic treatment, no associated syndrome or congenital anomalies, and no alveolar graft. The average age was 10.7 years, ranging from 8 to 14.8.

The sample was divided into three RME groups with different types of appliances: Hyrax (n = 10), iMini (n = 10) and Fan-type (n = 10). The allocation method was based on the type of maxillary atresia presented by subjects. The subjects presenting an anterior and posterior transverse maxillary deficiency were treated by the Hyrax expander. In cases where only the anterior region was compromised, both iMini and Fan-type were indicated.

The Hyrax group received a tooth-borne appliance with a jackscrew (Leone Orthodontics and Implantology, Firenze, Italy) located in the mid portion of the arch, soldered to molar bands that were cemented to the first permanent molars and bonded with light-cured composite to the primary molars or premolars (Fig. 1a). The iMini group received a tooth-borne appliance with a mini-hyrax screw (Dynaflex, Saint Ann, Mo) located in the anterior part of the arch, with a posterior arm extension on each side, soldered to first molar bands and bonded with light-cured composite to the primary molars or premolars (Fig. 1b). The Fan-type group, on the other hand, received a tooth tissue-borne expander with a hinge located in the posterior portion and a screw (Orthodontics Morelli, Sorocaba, Brazil) in its middle portion. This appliance bears an acrylic pad that stays in close contact with the palatal mucosa and is also anchored by bands on the first permanent molars (Fig. 1c).

Fig. 1
figure 1

a Hyrax expander. b iMini expander. c Fan-type expander

Full size image

The activation protocol of the expanders was 2/4 of a turn daily until an overcorrection of the maxillary atresia was reached. Once at this point, screws were tied off with a ligature wire. The inactive appliance was kept in place for retention purposes for 3 months.

Registration method

Participants underwent two cone beam CT scans (CBCT). Acquisitions were obtained before treatment (T1) and 3 months after the expansion was completed (T2). A few aspects were observed during the acquisition of CT images for standardization purposes and to improve image quality, namely, removal of the expander, natural head position and maximum intercuspation. The CBCT scans were performed by the same technician, all with the iCAT scanner (Imaging Sciences International, Hatfield, PA), obeying the following acquisition parameters: 120 kV, 8 mA, scanning time of 40 s, and 0.3-mm isotropic voxel.

The use of CBCT scans is now part of our routine orthodontic documentation at the Craniofacial Deformities Center, substituting all other X-rays required for the initial orthodontic diagnosis and treatment planning. The second exam was necessary to proceed with the surgical planning for the secondary or tertiary alveolar graft.

Measurement methods

Measurements were executed by an independent examiner who was blinded to the type of treatment and patient’s identity. Blinding was set up by a second investigator who randomly coded DICOM files from 60 scans (30 T1, 30 T2), so that the examiner did not have access to patient’s name and naturally to the pre- and post-treatment stages.

Images were imported to and analysed in the Dolphin Imaging software (version 11.7; Dolphin Imaging & Management Solutions, Chatsworth, CA). At first, a simplified staging of mesiobuccal, distobuccal and palatal roots of the permanent maxillary first molars (direct anchorage to the expander) was performed. They were classified as the stage of root formation in open or closed apex. This approach allowed us to identify the roots with and without growth potential.

The second stage of measurements consisted of a linear evaluation of the root length of 360 roots (180 T1, 180 T2), as follows: mesiobuccal root, shortest linear distance between the tip of the mesiobuccal cusp and the apex of the mesiobuccal root; distobuccal root, shortest linear distance between the tip of distobuccal cusp and the apex of distobuccal root; and palatal root, shortest linear distance between the tip of the mesiolingual cusp and the palatal root apex (Fig. 2).

Fig. 2
figure 2

Linear measurements of the permanent maxillary first molar. a Mesiobuccal root. b Distobuccal root. c Palatal root

Full size image

A three-dimensional marking of the points was made using multi-planar reconstructions of parasagittal and paracoronal orthogonal slices allowing for a correct identification of cusp tips and the root apexes (Fig. 3). The apical marking at the open roots was done in the most prominent root wall.

Fig. 3
figure 3

Measurement of the palatal root on a multi-planar reconstruction

Full size image

Statistical methods

In order to evaluate intraobserver agreement, all measurements were assessed twice in 20 CBCT scans randomly selected in a 2-week interval. Repeatability was determined by the intraclass correlation coefficient (ICC).

The data were evaluated by Kolmogorov–Smirnov and Levene tests to check the normality and homogeneity, respectively. Initially, a multilevel linear regression analysis was performed considering three levels: root length assessment made before and after treatment (first level), tooth (second level) and participant (third level). This analysis was carried out to recognize and quantify differences between T1 and T2. The analyses were firstly performed in the whole sample and then a stratified analysis was conducted considering the root stages (open or closed apex).

Following these comparisons, a second multilevel linear regression analysis was performed to evaluate the association between cleft side, age, gender, type of expander and stage of root formation with changes observed in root length after treatment. The coefficients and standard errors were calculated on this approach and the p values were derived from the maximum-likelihood test. The analyses were carried out using the statistical software Stata 13.0 (StataCorp LP, College Station, TX) and the level of significance was fixed at 5%.

Results

The repeatability of the variables was very high where the ICC values ranged from 0.996 to 0.998 for all root lengths. Out of 180 roots examined before treatment, 68 presented open apexes, and 112 presented closed apexes. The average root length of open and closed apex roots was calculated, and the age and gender distribution for each group are presented on Table 1. Considering a minimally important difference among groups of 20%, a statistical power of 88.7% was observed.

Table 1 Age, gender, screw activation (mm), intermolar distance changes (mm) and root length (mm) distribution for all three groups
Full size table

As the data related to the amount of screw activation did not present homogeneous variance, Kruskal–Wallis was used to verify differences among the groups. There was no statistically significant difference (p = 0.07) in the amount of millimetres of activation among the three groups. Also, there was no difference in the changes in intermolar distance between the iMini and Hyrax groups, but both presented greater changes than the Fan-type group (p < 0.05) (Table 1).

According to the multilevel linear regression analysis (Table 2), the open apex palatal roots (n = 26) showed a statistically significant increase in root length after treatment (p < 0.001). The same trend was observed for the mesiobuccal (n = 21) and distobuccal (n = 21) open apex roots. The average increase was around 0.5 mm for the palatal and mesiobuccal roots and 0.4 mm for the distobuccal roots. In contrast, there were no significant changes in root length of the roots presenting closed apex before treatment (p > 0.05).

Table 2 Multilevel linear regression analysis of the difference between the root length (mm) before and after treatment considering all three roots and its development stages
Full size table

When the roots of hemiarches with and without cleft were compared (Table 3), there were no statistically significant differences in changes in root lengths after treatment (p > 0.05). The same was observed to the gender variable, which expressed no significant effect (p > 0.05). Regarding the type of expander used, the iMini and Fan-type groups did not differ from the Hyrax group (p > 0.05).

Table 3 Multilevel linear regression analysis with tooth length as dependent variable
Full size table

Patient’s initial age had an inversely proportional and significant impact on the post-treatment root length increase (p < 0.001). Similarly, root development stage significantly influenced changes in root length (p < 0.01). The roots with open apex exhibited approximately 0.5 mm larger changes if compared to those with apical closure.

Analysis of multiple models was performed to assess the association between root length changes and the root development stages, adjusting for types of expanders and for cleft side, but did not reveal alterations in the response pattern.

Discussion

One of the most precise indications for using CBCT in orthodontics is diagnosis and treatment plan of patients with craniofacial anomalies [9]. The use of CT scans in the assessment of dental structures provides precise linear measurements without distortions and overlapping, and allows direct visualization of the palatal roots [10]. Conversely, conventional X-rays tend to either over- or underestimate the amount of loss on root structures. Panoramic X-rays may overestimate EARR by 20% or more [11]. Periapical radiographs offer lower sensitivity and specificity when compared to CBCT and tend to underestimate EARR [12]. Therefore, the accuracy of this tomographic imaging technique, with less than 1% of relative error, makes it the current gold standard for in vivo linear measurements [13, 14].

The method of measuring the roots with CBCT adopted in this study was similar to those already reported in the literature, as an attempt to facilitate the comparison of our findings [10, 15]. The points of interest were marked on multi-planar reconstructions, which are more precise for linear measurements in hard tissues than on any other volume or surface rendering method [16]. The use of 0.3 mm isotropic voxel appears to be the most appropriate when it is aimed at a satisfactory accuracy in the detection of root resorption, with the lowest possible radiation, according to the ALARA principle (as low as reasonably achiavable) [17, 18].

The choice for the linear measurement method of the roots rather than its volumetric quantification was based on two aspects. The first concerns the need for manual three-dimensional object segmentation, which may incorporates gaps in the standardization process, especially if the scanning is performed with orthodontic appliances in place. This condition, despite maintaining the repeatability of the method within acceptable standards, greatly affects its reproducibility [19]. The second aspect lies in the accuracy of volumetric measurement methods. Its detection capability is limited to defects larger than 3.47 mm3, even if using a highly accurate CBCT scanner, with 0.125 or 0.2 mm voxel [19]. As the sample of the present study was scanned with a 0.3 mm voxel, the choice for this method would generate inconsistent results. Still, the authors believe that volume measurement should become part of the routine in orthodontic investigations as technology evolves, by incorporating new data on the assessments of the root structure. Linear assessment methods, however, should not be replaced since they facilitate clinical and immediate interpretation.

In attempting to understand the effects of RME on root structures of cleft subjects, some questions came into play. Is root resorption a factual adverse effect of this procedure? Does RME interrupt root development in young patients? Is there any difference on root response depending on the types of appliances used?

Despite the wide variations on the amount of root resorption among individuals, given the fact that the susceptibility threshold is established by genetic determinants [20], it is known that high magnitude forces lead to increased stresses on the periodontal ligament and higher amounts of root damage [20,21,22]. A few studies have been carried out with focus in evaluating the external root resorption after RME. Measurements were done in extracted first premolars that had been used as anchorage to the expander [23,24,25]. Resorption occurred in the buccal surface and apex of these teeth. At the same time, an attempt to repair the resorbed surface by the cellular cementum had been initiated [23,24,25]. The repair was noticed in the third month post-treatment, and it was still possible to identify active repair process 1 year after the expansion [26].

In studies where volumetric root changes after RME in non-CLP patients have been assessed, significant root resorption was found [2, 27]. This finding was also seen in an investigation conducted with a linear measurement method similar to the one adopted in this study [28]. The average EARR for the three roots of the permanent maxillary first molar was approximately 0.5 mm [28]. The present study, however, showed no significant root resorption, even when the analysis would only consider the closed apex roots since treatment started.

It is our belief that the dissident results may be due to the age of the population studied. The average age of our sample was 10.7 years, while in other studies it ranged from 12.7 to 13 years [2, 27, 28]. Nevertheless, our main hypothesis for this difference lies in the presence of the cleft palate, what would allegedly decrease the overall resistance of the maxillary complex, favouring the preservation of the expander’s anchoring teeth. As far as we are concerned, there have been no other papers investigating root resorption in CLP patients.

Regarding the effect of different types of expanders on the root structure, a study showed no difference between Hyrax (tooth-borne) and Haas (tooth tissue-borne) appliances [27]. Our study also found no differences between the Hyrax and the other two appliances used.

Although the root resorption topic is frequently addressed by scientific studies, little is known about the effects of orthodontic forces in the development of immature roots. Speculations are that orthodontic forces, especially the heavy ones, would lead to either the cessation of root development or to an abnormal development [2, 3]. This hypothesis or myth might have been perpetuated as truth among orthodontists today.

Some studies have sought to investigate the effects of light forces of fixed orthodontic appliances on immature roots [4, 29, 30]. These evaluations were performed based on panoramic and periapical radiographs and concluded that root growth was uneventful, without compromising its final length. Furthermore, no root resorption was observed in developing teeth [30]. It is important to emphasize that the root shape remained preserved and dilacerations free, quite differently from what some histological analysis on the deformation of Hertwig’s epithelial root sheath would propose [5, 30].

Although our study applied orthopaedic forces, potentially more harmful, our sample showed no root growth impairment. The continued growth of the roots with open apex was approximately 0.5 mm in 3 months post-treatment. According to the literature, normal growth rates for the roots of first permanent molar in control samples without CLP are approximately 6 μm/day [31,32,33]. This amount, over a period of 3 months, would correspond to 0.54 mm, quite consistent with the present findings. For ethical reasons, it was not possible to proceed with a third CT scan to check the final root length of the subjects.

Attention should be paid to the fact that studies comparing dental development of patients with and without cleft lip and palate reveal a delay in the root formation of the latter [34, 35]. However, literature is still contradictory vis-à-vis to the difference in tooth development between the affected side as compared to the normal side. This study found no differences between the sides neither in the amount of resorption nor in the amount of root growth of teeth with closed and open apexes, respectively.

The authors believe that the dental papilla of developing tooth may have a protective role in root growth, preventing apical resorption during and after treatment. Other authors seem to agree with this assumption, based on the fact that they present data of normal root length at the end of treatment for initially immature roots [4, 29]. In this sense, early orthodontic intervention appears to be welcome with regard to root health, given the greater biological tolerance at younger ages.

Further long-term studies should be carried out to investigate both resorption and root development processes of CLP subjects. It is recommended that quantitative measurement methods should be deployed rather than the root resorption staging ones, what would facilitate comparison between different studies and generate more consistent data for statistical analysis.

Conclusions

Considering the effects of RME on the roots of cleft subjects under different development stages, the following set of conclusions can be drawn:

  • The RME orthopaedic forces do not interrupt root formation process and the growth rate seems to continue within normal values.

  • RME does not lead to significant external apical root resorption.

  • No differences were observed among the three types of appliances regarding adverse effects on the root structure.