Objective
The aim of this work was to evaluate the predictive value of clinical symptoms for magnetic resonance imaging (MRI) findings of temporomandibular joint (TMJ) effusion.
Study design
Sixty-one patients with TMJ pain were assessed by means of a standardized clinical examination and MRI. A calibrated investigator evaluated the presence of 8 clinical indicators (predictors) of TMJ effusion (outcome variable). A logistic regression analysis was performed to detect significant associations between clinical symptoms and MRI findings of TMJ effusion. The accuracy of the final logit to predict effusion was compared with that derived from univariate analysis.
Results
A clinical examination based upon the assessment of pain in the TMJ with lateral palpation, with posterior palpation, during motion, and during maximum assisted opening, and the presence of click and crepitus sounds has an accuracy of 78.7 % to predict TMJ effusion. Among the single clinical symptoms, the most reliable predictor of TMJ effusion is the presence of pain with lateral palpation (accuracy 76.2 % ; K = .525).
Conclusion
The use of a multiple regression approach demonstrated that an extensive clinical assessment which considers 6 main indicators consents to predict accurately the presence of MRI TMJ effusion. Based on these findings, MRI assessment should be reserved for patients in which the exact disc-condyle relationship needs to be evaluated.
