2. Materials and Methods
To carry out the present study, the clinical records of the database of a Dental Clinic in Valencia (M-A Clinic) were analyzed and those that met the following inclusion criteria were selected: lateral teleradiography of the skull and frontal teleradiography of the skull in the anteroposterior projection of healthy patients aged 5 to 18 years with transverse problems. Cases in which the radiographs showed dental absences not typical of age or in which lateral or frontal teleradiography of the skull did not allow for correct location of the points either due to lack of cooperation of the patient or an incorrect technique were excluded.
All cases were re-evaluated by 3 orthodontic specialists with more than 10 years of experience in the field of orthodontics with the aim of classifying patients into 3 groups: Group 1 (growth subjects suffering from a transverse maxillary deficit), Group 2 (subjects without transverse problems) and Group 3 (cases in which the orthodontic specialist did not have a clear diagnosis, leaving the decision of whether to carry out a skeletal expansion treatment of the upper jaw based on other variables such as the position of the upper canines, the amount of crowding present in the arch, or a family history of maxillary compression).
The final sample of the study was constituted of those cases where at least 2 of the 3 experts had classified the case within the same category, obtaining a final sample of 97 cases—62 cases in Group 1 and 35 cases in Group 2—with an overall mean age of 9.8 ± 2.6 years and a range between 5.5 and 17.8 years and a median of 9.3 years.
An analysis of the cases was carried out where the following data was collected:
- -
Facial axis, mandibular plane, ANB and Wits.
- -
Maxillary and mandibular width: frontal teleradiograph of the skull was measured using the Kodak Dental Imaging Software 6.12.11.0 computer program, converting the automatically generated magnification to a 1:1 image ratio, with the aim of standardizing the magnification produced. Cranial width, orbital width, width between the zygomatic bones, maxillary width, and mandibular width were obtained (
Figure 1).
- -
Difference between maxillary and mandibular width, difference between orbital and maxillary width, difference between zygomatic and maxillary width, difference between maxillary and bigonial width and difference between maxillary and biantegonial width.
- -
Chronological age.
- -
Biological age, according to the Demirjian method [
14].
Statistical Analysis
Statistical analysis was performed using the SPSS version 15.0 program, with a significance level of 5% (α = 0.05) and a confidence value of 95%, where homogeneity was assessed using the t-test for independent samples and with normality fit (confirmed by the Kolmogorov–Smirnov test). Descriptive and inferential analysis of the data collected was performed using the chi-squared test and a binary logistic regression model was estimated to explain/predict the diagnostic group from each of the cephalometric variables. The odds ratio (OD) of the unadjusted association and 95% confidence intervals were provided. Next, a model was made, also logistic, but in this case multiple, with the set of independent variables. The objective of this analysis was to identify the optimal subgroup of variables (determining as optimal variables those that could best explain the presence of maxillary compression) and to consider that any other does not increase the explanatory capacity of the model.
Data analysis was performed using simple models and multivariate models: sensitivity, specificity, and the predictive value of positive and negative tests. The ROC curve (Receiver Operating Characteristic) was also obtained for the diagnostic rule, thus evaluating its discriminatory capacity by estimating the AUC (Area Under the Curve). The typical error was also provided, with a confidence interval at 95% of the AUC and a contrast test AUC of 0.5.
Finally, the model was re-estimated on an estimation sample corresponding to 60% of the cases of the global sample and it was evaluated on the remaining validation sample. Validity parameters were recomputed to compare the potential loss of efficacy and method error calculation was performed.
4. Discussion
Prior to planning and designing this study, the need to carry out this diagnostic study was raised, for which a search of the literature was performed to establish the measurement methods that had been used on frontal teleradiographs for the diagnosis of maxillary compression deformity. Regarding the results obtained, the lack of analysis of this type of image was corroborated, especially given the multitude of existing methods for assessing the problems that occur in the anteroposterior plane, as stated by Athanasiou et al. [
15].
Regarding the design of a logistic regression model, no similar study was found in the literature, based on the analysis of frontal teleradiographs of the skull, with the aim of performing a multivariate model to obtain an easily applicable diagnostic tool in the daily clinic.
The first to describe a method to quantify the transverse problem at the maxillary level was Ricketts (1981) through the study of frontal teleradiographs of the skull. Although it is true that the introduction of this image projection was much earlier (almost 50 years earlier), it was not until 1968 that it began to be used in studies on the growth and development of human beings. The introduction of skeletal and/or dentoalveolar expanders of the upper jaw expanded new horizons in orthodontic practice, forcing the need to establish specific standards capable of quantifying the discrepancy between maxillary and mandibular width [
16].
When analyzing the evaluation of the methods used, all the radiographs were obtained from the same office and were measured with the same computer program. The analysis obtained a high degree of reproducibility, so the digital measurement and location of the cephalometric points used turned out to be an adequate and reliable procedure.
If the most frequent dentoalveolar characteristics and traits that can be found both in patients with a transverse growth deficit of the nasomaxillary complex and in patients with normal transverse dimensions are analyzed, a greater number of studies and data in the analyzed literature are determined. Yavuz et al. [
9] found that the cranial width is a practically constant value in development from 6 years of age, with slightly more development remaining in the facial, nasal, maxillary and mandibular width. These same authors agree with Graber et al. [
17] in the sequence of development, according to which the transverse dimensions are much more stable and stop before the anteroposterior and vertical ones.
The study found that one of the variables most closely related to the presence of superior skeletal compression was cranial width and its difference in relation to the maxilla. If it is taken into account that it is one of the most stable factors from the age of 6, it can be affirmed that it is definitively related to the diagnosis of the deformity, regardless of the age of the subject.
When analyzing the data obtained in the group of cases from the lateral projections, it is observed that the only statistically significant measure in relation to the healthy group is the Wits, which shows a tendency to a negative result according to the mean obtained (−0.15 ± 2.79), which shows a tendency of the transverse growth deficit of the maxilla and skeletal class I borderline towards class III.
Within the reviewed literature, only one study was found that analyzed the radiological characteristics at the frontal level in class III in the geographical environment in which the present investigation was carried out [
18]. These same authors highlight the absence of more cross-sectional studies on this skeletal malocclusion as the result of the extensive application of the treatment protocol for this type of deformity, consisting of the use of a skeletal expansion of the upper jaw prior to the anterior traction of the maxilla, either through extraoral appliances (a face mask) or through the use of mini plates and/or mini screws. It is then stated that the need to carry out a cross-sectional study of this type of case has been relegated to the background for years. In conclusion, since the effectiveness of anterior traction improves with expansion, it is not important to study whether expansion is necessary or not due to an underlying transverse problem, because regardless of the result obtained in the analysis in the frontal plane, the maxilla should be expanded for a more successful result.
Performing the first study based on a posteroanterior skull radiograph, Franchi and Baccetti [
18] found that the maxillary width was 3.8 mm smaller in class III cases when compared to skeletal class I. This same procedure was also performed for the skeletal class II group, where a difference in maxillary width also appeared, but of a smaller dimension than in the previous case (2.5 mm smaller). If these results are compared with those obtained, a concordance is observed, since the Wits value together with the standard deviation falls within the normal values of skeletal class I as well as class III and light class II [
19].
In addition to the study already mentioned [
19], two others analyze the transverse characteristics at the radiological level of class II and class I cases. Alarashi et al. [
20] described the characteristics in the frontal plane on a teleradiography of the skull in cases of class I and II first division (in which a proinclination of the upper incisors is characteristic) without an apparent transverse problem. The need to assess the transverse malocclusion in an X-ray is highlighted, especially at greater discrepancies at the anteroposterior level. These authors found statistically significant differences in maxillary width and upper intermolar width, these being narrower in class II/1 cases than in class I cases. These results were in turn associated, although not statistically significantly, with greater vertical growth of the upper jaw, more typical of dolichofacial patterns. A certain contraction of the nasal width was also shown, but it was not statistically relevant. However, no differences in mandibular width were shown when comparing class I and skeletal class II cases. When comparing these data with the results obtained, it can be observed how subjects who present a greater transverse defect of the maxillofacial complex showed higher maximum values in terms of anteroposterior retrusion of the upper jaw (ANS), normally associated with class III, and minimum values of mandibular position (SNB) more extreme than the controls, associated in this case with class II.
Lux et al. [
21] also analyzed the transverse cephalometric characteristics of classes I and II, including in this case both a subgroup of class II/1 and class II/2. They found that class II cases with maxillary incisor proclination had narrower jaws than any other group. No statistically significant differences could be found in mandibular size between the groups (C I, C II/1 and C II/2). This coincides with the results of the research carried out, according to which the mandibular width is not a significant datum to take into account when assessing the presence of a transverse growth deficit of the upper jaw. However, in the design of the study, a great difference is found with respect to the one developed in the present work and the two found in the literature, since the frontal teleradiographs were taken with an inclination of 35° with respect to the Frankfurt plane, so they are not comparable.
In addition to what has been said so far, it can be observed how the analyzed literature shows that any skeletal class can be associated with a transverse growth deficit of the maxilla, both in extreme cases of class III and cases of mandibular retrusion and class II.
This absence of a relationship between the skeletal compression of the upper jaw and a clear skeletal class is also shown in the present study, as there is no statistically significant relationship between the case group sample and the result of the ANB variable, despite the fact that there is a tendency toward negative values of the Wits index.
This is a key point to highlight since when designing the study and selecting which variables should be analyzed, two assessments were included for each of the characteristics studied. Thus, on the one hand, the skeletal class was defined by the Wits and ANB values and, on the other hand, the facial growth pattern was defined by the facial axis and the mandibular plane. In this way, it was possible to verify or rule out a possible statistically significant result. This same phenomenon was verified again with the analysis of facial growth. In the statistical study of independent variables, there was a certain statistical tendency of the mandibular plane to be greater in subjects from the compression group (36.2° compared to 33.8° on average in the healthy group), which could have shown a certain tendency toward vertical growth of subjects with compression. However, when checking this suggested relationship in light of the results obtained with the second variable analyzed on vertical growth, it was observed that the angle of the facial axis did not show any statistical relationship between the groups. Therefore, such an association was ruled out as statistically significant but left a statistical suspicion between the more vertical patterns and the presence of a skeletal compression deficit of the maxilla.
This requirement of double assessment of the skeletal class and the facial pattern made it difficult to classify the cases. Initially, the design was planned with the idea of dividing the sample into three subgroups according to facial pattern and skeletal class. However, in most cases, one of the values classified the subject in one group and the second variable determined the classification of the same subject in another group. Thus, for example, for a specific patient, the ANB could be within the class I parameters, while the Wits belonged to the class III set. For this reason, it was finally decided to consider only the absolute values of the variables as well as the measures derived from them (mean, median, maximum and minimum).
In the literature, based on the analysis of the cross-sectional characteristics based on diagnostic models, it is detected that patients who show a class II/1 dental relationship manifest a smaller dimension of the dental width of the upper arch [
22].
Coinciding with these same authors, Bichara et al. [
23] described how the subjects with this dental malocclusion showed this relative compression from the first stages of primary dentition until the dental replacement was completed, validating the premise already stated that there can be no relationship between the age of the patient and the probability of the transverse problem of the maxilla, as the results obtained could appear a priori; on the contrary, the deficient width of the maxilla not only does not resolve spontaneously but also worsens over time. DaSilva et al. [
24] agree with these same authors when studying the dental characteristics of dental class II, thus recommending an assessment of the need to perform an expansion of the upper arch prior to sagittal correction. They did not establish whether this expansion must be skeletal or dental, since they did not analyze radiographic records and, therefore, did not take into account the width of the skeletal bases.
Sayin and Turkkahraman [
25] designed a study using plaster models to assess not only tooth width but also dentoalveolar width, measuring the greatest width of the model at the vestibular level in class II first division subjects and class I subjects. They concluded that there is indeed a tendency to compress the upper dental arch in the case of subjects with distocclusion, but that these differences dissipate when measured at the level of the dentoalveolar process (as they call it). Despite coinciding with the results of the study carried out (in which a statistical relationship between skeletal class and the presence of maxillary compression could not be established), this study method raises great doubts because it is not corroborated a frontal projection of the skull that makes it possible to demonstrate the findings that were observed.
The only one of the studies analyzed that makes an exclusive reference to the dental characteristics of class II/2 cases [
26] shows that this type of malocclusion presents more similar characteristics with respect to classes I than to classes II/1, with broader dental arches, without apparent constriction. The slight compression that appears at the level of the lower arch stands out exclusively, probably due to the effect caused by the overbite typical of this type of malocclusion, by totally blocking the lower dental arch (this reduction in the dental diameter being more pronounced at the level of the canines).
Therefore, it can be concluded that neither the predominant growth pattern in the patient nor the skeletal class shows a statistically significant relationship with the presentation of the transverse growth deficit of the maxilla, and it is not possible to identify a specific pattern.
Regarding the analysis of the relationship between the gender variable and the presence of a transverse growth deficit of the maxilla, no statistically significant relationship is found in the literature regarding this premise. In our results, a concordance is detected with respect to this condition, since, although women accounted for a higher percentage of patients than men in the case group, these data were not statistically significant.
Finally, if the results obtained through the individual analysis of the linear variables and the multivariate analysis carried out on the relationship established between the maxilla and the rest of the linear measurements obtained in the frontal teleradiograph of the skull are evaluated, it is observed that at first Ricketts (1981) exclusively related the width of the maxilla with the width of the mandible, regardless of the rest of the cranial morphology, with the only differentiation between sex, thus attending to the different development times experienced by men and women respectively. If only the studies obtained that can be compared to this method are taken into account, the study by Lux et al. should be discarded. [
20] since the imaging technique differs enormously in relation to the other three articles studied, which placed the Frankfurt plane parallel to the ground.
Therefore, only the studies by Franchi and Baccetti [
18] and Nanda et al. [
27] were comparable. None of the studies tried to extract a diagnostic method for this type of dentofacial deformity; the former extracted prediction tables for the normal growth of the nasomaxillary complex, based on absolute values without establishing proportional relationships, which are of much more interest, taking into account the wide differences that may exist in the growth and development of a sample group of the same age as a whole, but whose cranial complexion characteristics differ radically. The second group of authors determined that class III and class II cases show a certain tendency towards maxillary compression, especially those of class III. This fact could not be correctly contrasted in the present study since the sample size was insufficient and lacked clear criteria that would allow dividing the cohort into three groups according to skeletal class. However, certain minimum and maximum values were observed in the maxillary compression group, so it is possible to detect a certain tendency to associate the most extreme anteroposterior discrepancies with the transverse growth deficit of the maxilla.