INTRODUCTION

One of the goals of orthodontic treatment is creating a balanced and harmonious facial appearance. One aspect of this harmony is craniofacial symmetry. However, minor asymmetry is a desirable variation of the craniofacial structures because we perceive these little inconsistencies as esthetically pleasing.1,2 Although facial asymmetry exists in individuals with normal facial appearance,3–11 there is no consensus concerning its degree, side prevalence, or localization. A larger left side has been reported by several authors,3–6 while others have reported a larger right side.2, 7–9 Asymmetry of the craniofacial complex may be greater in childhood and adolescence because of relative growth imbalances between the right and left sides.10 Mandibular asymmetry may fluctuate in magnitude and side prevalence with increasing age.10,12 Facial asymmetry in children and adolescents might be related to gender.8 In the adult population, no gender-associated differences in craniofacial asymmetry have been reported.6,9

Orthodontists generally use visual analysis to determine facial balance for their patients.13 In the treatment-planning process, additional tests or imaging are ordered if the asymmetry appears clinically significant or if patients have conditions previously associated with asymmetry. For instance, temporomandibular joint (TMJ) pathology, such as degenerative joint disease, has been recognized as an important factor in growth disturbances, including mandibular deficiency and open bite.14,15 Occlusal instability, vertical facial asymmetry, and chin deviation to the affected side are other clinical signs associated with TMJ pathology.16

Internal derangement (ID), ie, disc displacement of the TMJ, is a common intra-articular disorder characterized by an abnormal relationship of the articular disc relative to the mandibular condyle, fossa, and articular eminence. Almost 80% of patients with temporomandibular joint disorders (TMD) have a form of internal derangement.17–20 TMJ internal derangement has been found in 94% of the pediatric TMD population.21 When magnetic resonance imaging (MRI) has been used in asymptomatic orthodontic juvenile patients, disc displacement has been discovered in 5%22 to 11.8% of patients.23

Juvenile TMD patients may have a smaller overall mandibular length, shorter posterior facial height, shorter ramus and corpus, larger gonial angle, and steeper mandibular plane than adults.24,25 Several authors have suggested that both TMJ internal derangement and degenerative joint disease could be the main causes for mild to moderate facial asymmetry because of mandibular growth deficit in a growing child or adolescent.16,26–30

The degree to which TMJ disorder can affect facial growth might depend on the time of onset and the duration of the condition.31 The severity of the disorder and the side it involves are other important questions to consider.24

The objective of this cross-sectional study was to explore whether TMJ internal derangement is associated with greater-than-average craniofacial asymmetry in a sample of growing female patients.

MATERIALS AND METHODS

Cephalometric analysis

All radiographs were traced twice by 1 observer on transparent acetate paper using a 0.3-mm 3H Pentel 120 A3DX (0.3 mm) A313 Tokyo, Japan pencil. The choice of landmarks was based on previously published PA cephalometric reproducibility studies.32 Landmarks were registered manually. Six bilateral skeletal, 1 dental bilateral, 2 skeletal midline, and 2 dental midline points were identified. Orbita tangent (OT) represented the horizontal reference plane. The vertical reference plane, ie, the facial midline (FM), was constructed as a line passing through the midpoint between the intersections of the greater wing of the sphenoid bone and the orbital margin (GWSO), perpendicular to the OT (Figure 1). Deviations from the midline were also measured as the perpendicular distance from the midpoints of the facial skeleton to the facial midline. The landmarks and reference planes used are defined in Table 1.

View Figure

• Download as Powerpoint
• Download Figure

TABLE 1. Cephalometric Landmarks, Reference Planes, and Asymmetry Measurements

View Fullscreen

Analysis of asymmetry

Seven paired horizontal and 7 paired vertical variables, as well as ramus height on the right (R) and left (L) sides, were obtained for each tracing. Differences between the right and left sides for each measurement were used to calculate asymmetry according to the formula (R − L)/(R + L) × 200. Deviations of the midline landmarks from the facial midline provided 4 asymmetry variables, which were assigned a positive value if skewed to the right and a negative value if skewed to the left. Nineteen asymmetry variables were collected for each patient.

MRI analysis

Quantitative measures of disc displacement and disc length for each joint were obtained from consecutive sagittal MRI slices (produced from 3-mm–thick volume slices, 4 to 6 slices per joint) using the procedure described by Nebbe et al.33 Each MRI slice was traced by 1 investigator (BN) to determine the relationship between the osseous articular structures and the articular disc. As a first step in the MRI analysis, Frankfort horizontal was transferred from the corresponding lateral cephalometric tracing onto the MRI tracing. A reference plane was then drawn as a line that intersected the patient's Frankfort horizontal at a 50° angle through a point 10 mm anterior to the maximum height of the articular fossa. (Figure 2) The condylar load point (CLP) was determined as the shortest distance between the condyle and the articular eminence along a line perpendicular to the reference plane. Three landmarks were identified on the contours of the articular disc: anterior band, posterior band, and midpoint of the disc (Figure 3). From each landmark, perpendicular lines were erected to the constructed reference plane. Disc displacement was measured as the distance from the midpoint of the disc to the CLP along the reference plane. Negative values for disc displacement implied that the midpoint of the disc was posterior to the condylar loading point and represented a variation of normal disc position. Disc length was measured as the distance from the anterior to the posterior band of the disc along the reference plane. A negative correlation between disc displacement and disc length has been described.33 To produce a positive correlation between these 2 scores, each measurement for disc length was subtracted from a calculated value for normal disc length of 10 mm. The values thus obtained ranged from negative values, representing disc length greater than 10 mm (and therefore normal), to positive values describing shortening of the disc of increasing severity. In this way, several variables measuring disc displacement and disc length were obtained per joint. Principal components analysis (SPSS for Windows, SPSS Inc, Chicago, Ill) was used to integrate the measurements from each slice into 2 single weighted scores for each joint, 1 for disc displacement and 1 for disc length. The same statistical procedure was used to synthesize disc displacement and disc length scores of each patient into 1 total score for TMJ ID for the right (TMD ID R) and for the left (TMD ID L).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

RESULTS

For the purposes of this study, acceptable margins of error were set at 1% for paired asymmetry measures and at 1 mm for midline asymmetry measures. Errors of most paired asymmetry measures and of all 4 midline measures did not exceed acceptable limits. The mean of the standard error for midline measurements ranged from 0.30 mm for the anterior nasal spine (ANS) to 0.72 mm for the menton (Me). Variables that had more than 1% measurement error were zygomatic vertical (ZyV%), mastoid horizontal (MaH%), and jugal horizontal (JH%), and these were considered with caution in further analysis.

Asymmetry in patients without TMJ ID was calculated as presented in Table 2. Classifications of patients according to cluster analyses are listed in Tables 3 and 4. Forty-seven percent of patients had some form of TMJ ID. Twenty-nine percent had unilateral TMJ ID (16% involving the right TMJ and 13% involving the left TMJ), whereas 19% had a bilateral TMJ ID. Tables 5 and 6 contain the one way ANOVA outcome when means of craniofacial asymmetry measures were compared between cluster groups.

TABLE 2. Range of Asymmetry in Adolescent Females Free of Temporomandibular Joint Internal Derangement*

View Fullscreen

TABLE 3. Cluster Groups Formed by Combining TMJ ID Scores*

View Fullscreen

TABLE 5. ANOVA Comparisons of Amount of Asymmetry (Absolute Values of Asymmetry Measurements) for Cluster Groups I, II, and III

View Fullscreen

The results obtained with multiple linear regression are presented in Table 7. The purpose of this analysis was to determine which asymmetry measurements might be significantly associated with TMJ ID and therefore used to explain TMJ ID.

TABLE 7. Multiple Linear Regression Analysis Results. Asymmetry Measurements Significant in Explaining the Variability of the Dependent Variables R, L, (R − L), and (R + L)

View Fullscreen

DISCUSSION

The results revealed that for most craniofacial regions, the amount of asymmetry did not differ significantly between females with normal TMJs and those with either unilateral or bilateral TMJ ID (Tables 5 and 6). These findings were expected for upper and middle craniofacial structures. The deviation of the anterior nasal spine from the facial midline (ANS − FM) in patients with unilateral TMJ ID was 0.6 mm greater for females with normal joints, which was not clinically significant. We expected differences in the lower facial regions, especially between patients with unilateral TMJ ID vs bilateral TMJ ID or bilaterally normal TMJs. The amount of vertical asymmetry in the region of antegonion was significantly different between the first 3 cluster categories. However, contrary to expectations, a greater amount of asymmetry was found in patients with bilateral TMJ ID (4.96%) compared with females with bilaterally normal TMJs (2.36%), whereas females with unilateral TMJ ID did not differ significantly from the other 2 groups (3.59%). Asymmetry appeared to progress from group I to group III. Time differences in the onset of TMJ ID in patients with bilateral TMJ ID, on one side first and then on the other, could explain this finding. We had no knowledge about the exact time of onset or the duration of the changes within the TMJs. Bilateral TMJ ID likely develops over a longer time period than does unilateral TMJ ID; therefore, bilateral TMJ ID can have a greater influence on facial development. In children and adolescents with juvenile rheumatoid arthritis (JRA), reduced ramus and total face height consistently support the idea that mandibular growth fails to proceed at a normal rate.36–38 Greater variance of ramus height asymmetry and overall mandibular length asymmetry has been demonstrated in JRA patients as compared with healthy children, whereas transverse mandibular dimensions were similar to controls.38,39 In adult patients with JRA, mandibular asymmetry is infrequent.40 These studies imply that the time period between 7 years and 17 years may be a transitional stage from unilateral to bilateral growth retardation. A similar concept could apply to internal derangement of the TMJs. Available literature suggests that mild-to-moderate asymmetry of unknown etiology tends to improve with growth.10,12

Cluster groups IV, V, and VI allowed us to compare both the magnitude and the direction of asymmetry in relation to TMJ ID asymmetry. The one measurement significantly different for cluster groups IV, V, and VI was ramus height (Table 6). In this instance, patients with TMJ ID dominant on the right side had a shorter right ramus, as expressed with the negative mean value of ramus percentage (−2.452%), compared with patients who had dominant left ID (0.2148%) and those with similar left and right TMJ disc status (−0.6536%). Although the direction confirmed a trend, the magnitude of asymmetry did not appear to be clinically significant. In an experimental study on rabbits, it has been shown that TMJ ID can significantly affect further mandibular growth, specifically ramus height.30 However, a shorter ramus on the disc displacement side can be partially compensated for by growth at the base of the mandible, so that the overall height of the mandible (ramus and body) may not be reduced.30

TABLE 6. ANOVA Comparisons of Means of Asymmetry for Craniofacial Measurements Between Cluster Groups IV, V, and VI

View Fullscreen

Our patients who were free of TMJ ID showed dominance of the left side of the face (Table 2). We assume that a pre-existing left side dominance could help explain why ramus height was more asymmetric in cluster group V (unilateral right TMD ID), whereas in cluster group VI (unilateral left TMD ID) the amount or direction of asymmetry did not differ significantly (Table 7). Preexisting “normal” asymmetry can either camouflage or exaggerate any reduction or lack of condylar growth in the TMJ. For instance, in a patient with left unilateral TMJ ID, or a bilateral TMJ ID that started on the left side, a preexisting left-side asymmetry will be neutralized by the reduced growth on the left side, and therefore may not result in an increase of asymmetry. The opposite situation, a right unilateral (or bilateral TMJ ID that has started on the right side) and a preexisting left-side asymmetry will result in even more pronounced asymmetry with a dominating left side as the right side lags in growth.

This study has several limitations. The sample analyzed in this study was selected in an attempt to study a spectrum of TMJ disc status, ranging from normal to disc displacement with deformation. The prevalence of ID identified in this study cannot be applied to the general female preorthodontic population. The results of this study are not directly comparable with previous investigations. The only other work that assessed the relationship between TMD and facial symmetry used patient complaints and clinical signs as well as transcranial radiography to assess joint status.24 This study used objective MRI measurements to characterize the status of the TMJ articular disc and precisely measure internal derangement. The accuracy of MRI in diagnosing internal derangement ranges from 73% up to 90% or higher.41,42

We used multiple linear regression analysis to evaluate which regions of the craniofacial skeleton exhibit asymmetry that might explain the variability of TMJ ID. The results of the multiple regression analysis (Table 7) showed that the highest R² value was obtained for the (R − L) variable (0.20). This finding indicates that asymmetry of ramus height and deviation of ANS and the lower dental midline (InI) explained 20% of the difference in severity of TMJ ID between the right and the left sides. The low R² value for (R + L) (4%) could mean that the sum of structural changes within both TMJs was not associated with a greater amount of asymmetry. The group (R + L) contained patients with both unilateral and bilateral TMJ ID. The discrepancy between these two R² values suggests that asymmetry of the craniofacial structures could be associated with “asymmetry” of TMJ ID, regardless of the severity of changes in the TMJs.

Overall, vertical asymmetry showed a statistically significant difference with regard to TMJ ID as opposed to transverse asymmetry. This is consistent with previous studies that have shown reduced vertical posterior facial and ramus height in adolescent TMD patients24 and in adults43 on lateral cephalograms.

This was a cross-sectional investigation, and the time of onset of TMJ ID was not considered. Patients with TMJ ID should be reevaluated later, and the rate of change in TMJ ID severity should be compared with the rate of change in craniofacial asymmetry. The amount of asymmetry that was statistically significant between patients in different cluster groups did not appear to be of a magnitude that is clinically important. Follow-up investigation of adolescents with TMJ ID documented by MRI is required to establish the long-term relationship between TMJ ID and craniofacial asymmetry.

SUMMARY AND CONCLUSIONS

In this study, we investigated the association between TMJ internal derangement (disc displacement) and craniofacial asymmetry in 80 adolescent females prior to any orthodontic treatment. We found that females with bilateral TMJ ID have greater vertical mandibular asymmetry than do females with unilateral TMJ ID or females with normal TMJs. Other craniofacial regions did not exhibit significantly different asymmetry. These results indicate that a longitudinal evaluation is required to study whether further growth of female patients with disc displacement may result in clinically significant facial asymmetry.