Introduction
Dentofacial asymmetries are often observed in patients with juvenile idiopathic arthritis (JIA) and temporomandibular joint (TMJ) involvements. The aim of this split-face study was to associate types of radiologic TMJ abnormalities with the degree of dentofacial asymmetry in patients with unilateral TMJ involvements assessed with cone-beam computed tomography.
Methods
Forty-seven JIA patients and 19 nonarthritic control subjects were included in the study. Normal condylar radiologic cone-beam computed tomography appearance in at least 1 TMJ was the inclusion criterion for all patients with JIA. The contralateral TMJ was thereafter scored as either “normal,” “deformed,” or “erosive,” consistent with predefined criteria. Based on the bilateral radiologic TMJ appearances, 3 JIA groups were assigned: normal/normal, normal/deformed, and normal/erosive. The severity of the dentofacial asymmetry was compared between the JIA groups and control subjects. Dentofacial asymmetry was expressed as interside ratios and angular measurements.
Results
Eighty-seven percent of the JIA patients were being treated or had previously received treatment with a functional orthopedic appliance at the time of the cone-beam computed tomography. Significantly greater dentofacial asymmetries were observed in the 2 groups of JIA patients with unilateral condylar abnormalities (deformation or erosion) than in the other groups. A similar degree of dentofacial asymmetry was observed in JIA patients with bilateral normal TMJs and in the nonarthritic control group.
Conclusions
JIA patients with unilateral condylar abnormalities (deformation or erosion) exhibited significantly more severe dentofacial asymmetries than did the JIA patients without condylar abnormalities and the control subjects. We found the same degree of dentofacial asymmetry when dividing patients with condylar abnormalities into deformation and erosion groups.
Highlights
- •
Dentofacial asymmetries often occur in patients with juvenile idiopathic arthritis (JIA).
- •
Dentofacial development is affected more severely in patients with condylar abnormalities.
- •
Dysmorphic dentofacial development is not associated with a specific type of condylar abnormality.
- •
Combined condylar deformations and erosions were observed in one third of the TMJs in JIA patients.
The involvement of the temporomandibular joint (TMJ) is a common finding in patients with juvenile idiopathic arthritis (JIA); prevalence varies up to 96%. A clinical consequence of TMJ involvement can be dysmorphic mandibular development resulting in dentofacial asymmetry, occlusal instability, alterations in muscular activity, and suboptimal TMJ function, which, in turn, may lead to the development of orofacial symptoms and a compromised esthetic appearance. One primary treatment goal for patients with JIA is to prevent this unwanted dysmorphic dentofacial development. However, the underlying processes are not yet fully understood.
Historically, dysmorphic dentofacial alterations in JIA patients have been regarded as a consequence of arthritis-induced TMJ degeneration resulting in the loss of condylar vertical height, which is an arthritis-induced degradation similar to what occurs in adults with rheumatoid arthritis and TMJ involvement. For decades, extensive research has fueled debates on the implications of TMJ lesions and degeneration on mandibular development and growth. However, the contemporary dominant theory explains the dysmorphic dentofacial development in JIA patients as a consequence of condylar growth disturbances rather than arthritis-induced condylar damage alone. In a recent review, Peltomäki et al supported this perception and emphasized the need for documentation in future prospective longitudinal studies. They further noted that altered dentofacial development may be caused by a combination of inflammatory effects on the intra-articular growth site of the condylar cartilage and impaired masticatory function.
There is currently a need to elucidate the intra-articular processes that occur during dysmorphic dentofacial development in patients with JIA. This is important from a pathogenic perspective to advance our diagnostic and therapeutic understanding of this challenging condition. Combined imaging and radiologic TMJ scoring systems have recently been published exclusively for JIA. Components of these scoring systems aim to score the severity of arthritis-induced osseous changes. The scoring systems imply that the progression of the arthritis-induced condylar osseous changes proceed from minor condylar head deformation (flattening) to condylar destructions that progressively reduces normal mandibular growth. Both scoring systems grade erosive osseous changes higher than condylar deformation (flattening). With an isolated focus on condylar conditions, these scoring systems may be useful for radiologic TMJ assessment. However, no human studies have determined whether particular types of abnormal TMJ condylar osseous changes (eg, erosive changes) are associated with a greater extent of mandibular dysmorphic development than others (eg, condylar deformations). Notably, previous experimental studies found no relationship between the severity of condylar lesions and a reduction in mandibular growth and development in young growing rabbits with experimentally induced and histologically confirmed TMJ arthritis.
Previous research on TMJ abnormality and dentofacial development in JIA patients was primarily based on conventional radiologic techniques, and only a few studies used more advanced techniques. The introduction of cone-beam computed tomography (CBCT), a cost- and dose-effective 3-dimensional (3D) imaging modality, has enabled the radiologic examination of TMJ hard tissue pathologies; it is superior to conventional radiologic methods. The aim of this cross-sectional CBCT-based study was to associate radiologic TMJ abnormalities with the degree of dentofacial asymmetry in patients with unilateral TMJ involvements. A control group was used to compare asymmetries with the normal population.
Material and methods
Eighty-six consecutive JIA patients at the Section of Orthodontics, Aarhus University in Denmark, were eligible for inclusion in this study. All candidates had received a craniofacial CBCT scan between February 2011 and April 2014. Patients were included when they complied with the following inclusion criteria: a diagnosis of JIA according to the International League of Associations for Rheumatology criteria, and at least 1 TMJ with radiologic normal or healthy osseous appearance. The exclusion criteria were patients with previous craniofacial trauma or an underlying diagnosis of a syndrome or congenital birth defect involving the craniofacial area, or CBCT images of poor quality. The data on patient characteristics were collected from the medical hospital records. This study was approved by the Danish Health and Medicines Authorities (DOK2129859) and the Danish Data Protection Agency (2007-58-0010) and was conducted in accordance with the Helsinki Declaration.
Nineteen control subjects without JIA previously treated at the Section of Orthodontics at Aarhus University for other reasons were identified. All had received a full-face CBCT scan for orthodontic treatment. Control subject criteria were no previous or current diagnosis of temporomandibular dysfunction, a high-quality CBCT allowing optimal radiologic assessment, and an age comparable with that of the JIA patients. Although control subjects with a neutral dental occlusion (Class I) were preferred, full-face CBCT images were rarely taken in this group; we therefore accepted subjects with dental Class II subdivision malocclusions and impacted canines.
CBCT scans (5G; NewTom, Verona, Italy) were acquired in accordance with the manufacturer’s instructions. The image acquisition parameters included a field of view of 18 ×16 cm, a scanning time of approximately 18 seconds, and 3.6 seconds of active radiation (pulsed mode) with settings of 110 kV and 3 to 7 mA. All CBCT scans were reconstructed with a 0.30-mm isotropic voxel dimension.
The radiologic evaluation was carried out on the joint level, giving each subject 2 independently assessed TMJ scores (1 score for the right joint, 1 score for the left joint). The interpretation of the CBCT images was conducted using NNT Viewer software (version 4.6; NewTom). The type of mandibular condylar abnormality was independently scored by a specially trained maxillofacial radiologist (L.Z.A.) who was blinded to the patient diagnosis and the order in which the TMJs were presented (control and JIA subjects were mixed). Cropped CBCT images only including the TMJs were viewed in the axial, oblique coronal, and oblique sagittal planes. The radiologist was therefore also blinded to the presence of any dentofacial asymmetry. Before patient inclusion, 3 definitions of condylar scores were determined based on categorization of the condylar radiologic appearance ( Fig 1 ): (1) normal: normal shape with smooth and intact outline and surface (score 0); (2) deformed: marked flattening or other changes in shape with smooth and intact outline and surface (score A); and (3) erosive: disruption of outline or uneven surface due to cysts or erosion (score B).
The maxillofacial radiologist was asked to subgroup each mandibular condyle based on the radiologic appearances into 1 of 2 groups: normal condylar outline (score 0) or abnormal condylar outline (score A or B). When both deformation (score A) and erosive change (score B) were found in the same TMJ, the score was based on the most prominent radiologic feature. Each subject was thereafter categorized by the bilateral joint scores and assigned to 1 of 6 groups: JIA 0-0, JIA 0-A, JIA 0-B, JIA A-A, JIA A-B, and JIA B-B. In compliance with the inclusion criteria of this split-face study, only patients from the JIA 0-0, JIA 0-A, and JIA 0-B subgroups were included in our data analyses because a minimum of 1 joint in each patient was required to have a normal appearance. The control subjects were also subgrouped based on their bilateral joint scores. However, to represent the population-based variation in radiologic condylar appearance, all control subjects were accepted without considering their bilateral joint scores. Therefore, control subjects with radiologic abnormalities A-A or B-B were also included. Double assessments of 20 consecutive TMJs were conducted with a 2-week interval to assess the intraobserver agreement.
To assess the degree of dentofacial asymmetry, a 3D dentofacial analysis was conducted using Mimics software (version 16.0, interactive medical image control system; Materialise, Leuven, Belgium). The aim was to assess the intrapatient mandibular asymmetries based on outcome variables representing facial asymmetries between the left and right sides of the face. Information on sagittal, vertical, and transversal mandibular asymmetries was collected using 8 anatomic landmarks. The landmarks are defined and shown in Table I and Figure 2 .
| Definition | Abbreviation or color | |
|---|---|---|
| Landmarks | ||
| Condyle point | Midpoint on the superior surface of the condyle | Cor/Col |
| Gnathion | Lowest point on the lower border of the chin | Gn |
| Gonion (right/left) | Constructed by the bisection by the angle formed by the tangents to the lower and posterior borders of the mandible | Gor/Gol |
| Incisura point (right/left) | Lowest point in the concavity between processus coronoideus and processus condylaris in relation to the axial plane | Incr/Incl |
| Latero-orbital point (right/left) | At the zygomaticofrontal suture at the lateral aspect of the orbit wall | LOr/LOl |
| Mandibular first molar point (right/left) | Mesiobuccal cusp of the mandibular first molar | Minfr/Minfl |
| Nasion | Midpoint between the maxillary-nasal-frontal right and left junction | N |
| Sella turcica | Center of the hypophyseal fossa | S |
| Planes | ||
| Axial plane | Through S, LOr, and LOl | Blue |
| Midsagittal plane | Through N and S and perpendicular to the axial plane | Dark blue |
| Coronal plane | Through S and perpendicular to the axial and midsagittal planes | Pink |
| Inc plane (right/left) | Through Inc (left/right) and perpendicular to the midsagittal and coronal planes; 1 plane for each side was created | Purple |
| Gonion plane | Plane through gonion (right/left) and gnathion | |
| Variables | ||
| Mandibular posterior height (right/left) | From Go to Co (mm) | |
| Condylar height (right/left) | From Co to Inc plane (mm) | |
| Ramus height (right/left) | Go to Inc plane (mm) | |
| Axial plane-Go (right/left) | Axial plane to gonion (mm) | |
| Axial plane-Minf (right/left) | Axial plane to mandibular molar mesiobuccal cusp tip (mm) | |
| Mandibular length right/left) | Gn to Go (mm) | |
| Jaw angle (right/left) | Angle between the line connecting Co and Go and the line connecting Go and Gn (°) | |
| Transversal width at gonion level (right/left) | Distance from the midsagittal plane to gonion (mm) | |
| Gonion plane angle | Angle between the gonial and midsagittal planes (°) measured to the affected side/group 0-0 to the smallest side |
For the analysis, 2 main reference planes were constructed: axial plane, through S and LOr and LOl, and midsagittal plane: through N and S, and perpendicular to the axial plane (abbreviations are given in Table I ).
The additional planes necessary for the analysis were constructed with reference to these 2 planes. The points and additional planes are defined in Table I and shown in Figure 2 , A and B .
A total of 9 predefined outcome variables were assessed ( Table I ; Fig 2 , C and D ). In the vertical direction, 5 bilateral length variables of interest were defined: (1) total mandibular height (from Co to Go), (2) condylar height (from Co to Inc plane), (3) ramus height (from Go to Inc plane), (4) distance from the axial plane to the Go, and (5) distance from the axial plane to Minf. One variable was defined in the sagittal direction: the lengths of the mandible’s left and right sides (from Go to Gn). In the transversal direction, 1 variable was defined: transversally, the distance from the midsagittal plane to the left and right gonions. Additionally, 2 angular measurements were defined: (1) the jaw angle, left and right sides (angle between the line connecting Co and Go and the line connecting Go and Gn), and (2) the gonial plane angle (angle formed between a plane passing through the right and left gonions (and gnathion and its angle to) and the midsagittal plane measured to the affected side. In group 0-0, the angle was measured to the shortest side. For all bilateral outcome variables, a ratio of asymmetry was calculated to obtain a measure of intrapatient dentofacial asymmetry. The ratio was calculated by dividing the value of the affected side (TMJ score A or B) by the value of the nonaffected side (TMJ score 0). In group 0-0, the ratio was calculated using the smallest value divided by the largest value.
Statistical analysis
The data concerning the 9 outcome variables were tested for a normal distribution by visual inspections of Q-Q plots. Intergroup differences in dentofacial asymmetry were assessed using analysis of variance (ANOVA) tests with independent Student t tests serving as post-ANOVA tests. Post-ANOVA testing was only performed in outcome variables where a statistically significant difference was observed in the primary ANOVA test. The significance level was adjusted in accordance with the Bonferroni correction method to avoid a type 1 error caused by multiple testing. In the primary ANOVA tests, the significance level was lowered with the number of tests conducted (0.05/9 → P <0.006). In the post-ANOVA tests, the significance level was adjusted based on the number of t tests conducted in each outcome variable (0.05/6 → P <0.008).
Intrarater agreement of the radiologic condylar TMJ scores was assessed by kappa statistics based on duplicate assessments 2 weeks apart. The error of the method of the 9 outcome variables describing craniofacial asymmetry was also evaluated based on duplicate assessments 2 weeks apart using scatterplots and correlation coefficients. For each of the 9 outcome variables, the error of measurement was assessed based on Bland-Altman plots. The limits of agreements derived from the Bland-Altman plots were used to define the smallest detectable difference for each outcome variable. The smallest detectable difference was defined as the minimal amount of change in the interside ratios and angles needed to overcome the measurement error in each outcome variable.
Results
In the JIA groups, 47 of the 86 examined patients met the inclusion criteria. All patients had inflammatory involvements in other joints in addition to the TMJs. Due to similar bilaterally abnormal radiologic TMJ assessment scores (JIA A-A, JIA B-B, and JIA A-B), 36 patients (41%) were excluded from the study. Three JIA patients were excluded due to low CBCT quality ( Fig 3 ). The remaining 47 JIA patients were grouped into 3 subgroups based on the radiologic findings: JIA 0-0 (n = 17), JIA 0-A (n = 20), and JIA 0-B (n = 10). The mean ages of the patients in the subgroups were 13.2 years (SD, 2.6) in the JIA 0-0 group, 12.8 years (SD, 2.7) in the JIA 0-A group, and 11.4 years (SD, 2.5) in the JIA 0-B group ( Table II ). Comparable distributions of disease duration of TMJ arthritis were observed among the JIA groups ( Table II ). Nine of the 47 patients (19%) had radiologic findings of both scores A and B in the same joint. Forty-one of the 47 JIA patients were currently being treated or had previously been treated with a functional orthopedic appliance and the method has previously been described by Stoustrup et al. The distributions of patients receiving treatment with functional orthopedic appliances were JIA 0-0, 13/17, 77%; JIA 0-A, 18/20, 90%; and JIA 0-B, 10/10, 100%. No significant intergroup difference was observed in the duration of treatment with functional orthopedic appliances at the time of the CBCT.
