RDC/TMD assessment

Patients were selected to receive the treatment protocol on the basis of a TMJ osteoarthritis diagnosis according to the RDC/TMD 1.0 guidelines, provided the following signs and symptoms were present: arthralgia (TMJ pain with lateral and/or posterior palpation plus anamnestical reporting of TMJ pain during maximum voluntary mouth opening and/or maximum assisted mouth opening and/or lateral excursions); crepitus sounds; radiological signs of TMJ bone structures abnormalities, such as erosions, sclerosis, flattening, osteophytes, showed by means of imaging techniques (e.g. magnetic resonance imaging).

The RDC/TMD axis II, contains a questionnaire which enables an assessment of chronic pain severity according to the Graded Chronic Pain Scale (GCPS) as well as an evaluation of depression and somatization scores according to the dedicated Symptoms Checklist-90-R scales.

The GCPS was originally developed by Von Korff et al. Its validity has been tested in a large population survey, and the prognostic value has been tested in a 3-year follow-up study in large samples of primary care pain patients, also including TMD pain patients. The GCPS is composed of six items assessed on a 10-point scale, and one item on the number of disability days due to facial pain. These items are suitable for self-reported use and, even though the characteristics of the scale enable measuring pain dysfunction as a continuous variable, the authors have provided hierarchical criteria to grade pain dysfunction into ordinal categories. The scoring criteria are simple to use, and allow categorizing pain patients into five levels of chronic pain grades (0, no disability; 1, low disability, low pain intensity; 2, low disability, high pain intensity; 3, high disability, moderately limiting; 4, high disability, severely limiting).

As for depression and somatization levels, the RDC/TMD axis II enables their assessment by means of the depression and somatization scales of the Symptom Checklist 90-R (SCL-90-R), an instrument originally developed by Derogatis. The choice to include the SCL-90-R depression and somatization scales (SCL-DEP, SCL-SOM; abbreviated to DEP and SOM in the text below) in the RDC/TMD axis for psychosocial assessment provides a contemporary evaluation of concurrent depressive and non-specific physical symptoms. 31 items were included in axis II, belonging either to the Depression and Vegetative Symptom Scale or to the Somatization Scale, which is used here to evaluate the presence of non-specific physical symptoms, plus seven additional items added to the Depression and Vegetative Symptom Scale. The mean scale score is calculated by summing up the scores of the single items. This makes it possible to rate patients as having normal, moderate or severe levels of impairment in the depression and non-specific physical symptoms scales. On the DEP scale, scores below 0.535 were considered normal, between 0.535 and 1.105 indicated moderate depression, and above 1.105 the presence of severe ongoing depressive disorder. On the SOM scale, including the pain items, scores lower than 0.5 were considered normal, values between 0.5 and 1 indicated moderate somatization, and above 1 severe somatization.

All the clinical RDC/TMD assessments were performed using the standard, internationally accepted Italian version of the RDC/TMD instrument available since 2002 on the RDC/TMD consortium website. The RDC/TMD axis I assessments for the inclusion of the patients in the treatment protocol were performed by the two same trained examiners (DM, LGN), who took part in several previous studies using the RDC/TMD. The same two investigators performed the interventions, while two other clinicians (ADG, CC) recorded pain at chewing using VAS scores (see below) and were responsible for axis II data collection and elaboration. In order to ensure single-blindness of the study design, all axis II and VAS scores were recorded separately on dedicated sheets not containing any information on the patients.

Treatment protocol

The treatment protocol adopted in this study provided five weekly arthrocentesis plus single-needle hyaluronic acid injection. The single-needle injection technique employs the same reference points as used in arthroscopic examination (lateral cantus-tragus) and is a modified version of the classical two needles technique for TMJ arthrocentesis. The skin surface is disinfected with povidone iodine. Local anaesthesia is achieved with mepivacaine 2% (Carbocaine, Sanofi Winthroph, NY, USA). The anaesthetic is first injected into the joint cavity, relaxing this virtual space. Subsequently, the needle is withdrawn gently to the skin surface, thus anaesthetizing the soft tissues over the joint, too. A single 19 G needle is then placed within the intra-articular space and physiological saline is injected under pressure with the patient in the open mouth position. After the injection, the patient is asked to close the mouth and the fluid is removed with the same injection needle, thus acting as both entry and exit point for the liquid. The injection–ejection process must be performed for up to 10 repetitions (for a total amount of about 40 cm ³ ), thus allowing elimination of catabolytes present in the synovial fluid. Once arthrocentesis is completed, 1 cm ³ of hyaluronic acid is injected into the joint in 3 s and the needle is removed. This technique has been shown to be as effective as the two-needle approach, and is currently adopted as the standard of reference at the authors’ clinic thanks to its reduced invasiveness.

Statistical design

The statistical approach was designed to answer the clinical research question: do axis II psychosocial findings predict treatment effectiveness at 6 months in patients undergoing the five weekly arthrocenteses plus HA injections? To test the null hypothesis that knowing the axis II scores makes no difference for predicting therapeutic effectiveness, an outcome variable for treatment effectiveness (i.e. pain levels at chewing) was assumed as the dependent variable and the axis II scores were assumed as predictors.

Scores in the clinical parameter ‘pain on chewing’, assessed by means of a VAS from 0 to 10, with the extremes being ‘no pain’ and ‘pain as bad as the patient ever experienced’ respectively, were assessed by the same operator at the time of the diagnosis (baseline) and at the time of the 6 month follow-up. Percentage changes in VAS values at the end of the observation period with respect to baseline were assessed to create an outcome variable that was not dependent on the absolute magnitude of VAS scores. This variable was adopted as the dependent variable to be identified in a linear regression model by the following predictors (independent variables) gathered with RDC/TMD axis II psychosocial assessment : depression levels, as based on the SCL-90 depression scale scores; non-specific physical symptoms (somatization, scores (pain items excluded), as based on the SCL-90 non-specific symptoms scale scores); characteristic pain intensity score; disability score; disability points; pain-related impairment, as based on the GCPS scores.

A single variable regression analysis was performed to test the existence of a correlation between the outcome variable and any of the predictors. A cut-off value of p < 0.10 was set to screen variables which were related to the treatment effectiveness, and a multiple variable linear regression analysis was performed to identify the significant associations between the predictors (independent variables: depression levels; somatization levels; characteristic pain intensity score; disability score; disability points; pain-related impairment) and the outcome (dependent variable: percentage change in pain at chewing levels) by including only those variables which were significant at p < 0.10 at the single variable analysis. Selection was made among the potential predictors of positive outcome using a backward stepwise selection method. R square ( R ² ) was obtained as an estimation of the variance explained by a summation of the significant psychosocial factors, and represents the amount that the independent variables can differentiate the dependent variable. R ² represents a numerical expression of the dependent variable’s (percentage changes in VAS pain at chewing scores) variance accounted for by the model of the significant predictors. All statistical procedures were assessed with the Statistical Package for the Social Sciences (SPSS 19.0, SPSS Inc., Chicago, IL, USA).