Variables That Influence Changes in Temporomandibular Joint Structure.

BACKGROUND: There are no objective measures which forecast longitudinal changes in the temporomandibular joint (TMJ). OBJECTIVE: To identify variables that predicted longitudinal changes in TMJ structure. METHODS: Over an average of 8 years, baseline and follow-up computed tomography and magnetic resonance imaging data were reviewed by calibrated radiologists. Baseline to follow-up changes in TMJ structure were grouped according to A = no changes, B = got better and C = got worse. For the current investigation, Groups A and B were combined (AB). Baseline measures included ramus length; occlusal plane angle and anteroposterior position; and condyle rectilinear area (major axis × minor axis lengths), aspect ratio (major axis length/minor axis length) and axial angle. These variables were combined with DC/TMD Axis II physical symptoms scores (PHQ-15) and an abbreviated 7-question oral behaviour checklist score (7QBC) to determine an Anatomic-Psychophysiologic Score (APS). Mixed-effects logistic modelling was used to estimate the contribution of baseline variables in APS that were significantly different between groups. RESULTS: From 401 adult subjects who provided longitudinal data, TMJs from females (426) and males (90) had baseline images with sufficient field-of-view for cephalometric measurements. TMJs were assigned to groups AB (n = 303) and C (n = 213). Significant group differences (p < 0.001) in APS were found for the combination of sex, ramus length, condyle rectilinear area and axial angle, occlusal plane anteroposterior position and 7QBC. Model accuracy was good (sensitivity, 0.85; specificity, 0.80; aROC = 0.91). CONCLUSIONS: The combination of sex and cephalometric and behaviour measures was found to be predictive of longitudinal changes in the TMJ.