Establishing better biological models to understand occlusion. I: TM joint anatomic relationships.

Belief in and rejection of a relationship of occlusion and temporomandibular joint (TMJ) condyle-fossa position with normal and abnormal function are still contentious issues. Clinical opinions can be strong, but support in most published data (mostly univariate) is problematic. Distribution overlap, low sensitivity and specificity are a common basis to reject any useful prediction value. Notwithstanding, a relationship of form with function is a basic tenet of biology. These are multifactor problems, but the questions mostly have not been analysed as such. This review moves the question forward by focusing on TM joint anatomic organisation as the multifactor system it is expected to be in a closed system like a synovial joint. Multifactor analysis allows the data to speak for itself and reduces bias. Classification tree analysis revealed useful prediction values and usable clinical models which are illustrated, backed up by stepwise logistic regression. Explained variance, R(2), predicting normals from pooled TMJ patients was 32·6%, sensitivity 67·9%, specificity 85·7%; 37% versus disc displacement with reduction; and 28·8% versus disc displacement without reduction. Significant osseous organisational differences between TM joints with clicking and locking suggest that this is not necessarily a single disease continuum. However, a subset of joints with clicking contained characteristics of joints with locking that might contribute to symptom progression versus resistance. Moderately strong models confirm there is a relationship between TMJ osseous organisation and function, but it should not be overstated. More than one model of normals and of TM derangement organisation is revealed. The implications to clinical decision-making are discussed.