TEMPOROMANDIBULAR DISORDER

TOP ABSTRACT TEMPOROMANDIBULAR DISORDER SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Dao and LeResche⁶ found that 8 to 15 percent of women had chronic TMD, while the prevalence among men was 3 to 10 percent. Lipton and colleagues² also found a sizable difference in the prevalence of jaw or facial pain for men and women. Few studies, however, have assessed distinctions between patients with acute and chronic jaw pain. Rieder and colleagues⁷ and Loeser⁸ summarized some of the many methodological flaws within the TMD prevalence literature, such as limiting data collection to a specific age group or one sex, as well as selection bias. Furthermore, researchers differ in how they assess TMD, particularly in the acute phases.^3,9

TMD is characterized by a combination of physical, functional and psychosocial factors.^3,9–11 Although the etiology is not clearly understood, the cause has been attributed to many factors, including physical structures in the mouth, musculature and psychosocial functioning.^1,11–13 Traditionally, clenching and grinding have been the most agreed-upon cause of TMD, typically resulting from psychological stress.¹⁴ Not only has stress been identified as a cause of TMD, but it has been identified as being important in its maintenance.

Many patients who seek treatment for their pain are confronted with frustrations unique to the multifactorial origin of this disorder.

Oakley and colleagues¹⁵ discussed the need to consider psychosocial factors as a contributing factor in patients with TMD who do not respond to more traditional medical and dental approaches to pain management. In addition to stress, the clenching and grinding response is believed to be related to poor muscle discrimination, unconscious bracing of the orofacial musculature or both.^11,14 Many patients with TMD experience pain in the head and neck distinct from that in the muscles in the masticatory region.¹

Many patients who seek treatment for their pain are confronted with frustrations unique to the multifactorial origin of this disorder. In light of the fact that medical, dental and psychosocial factors have been implicated as causes of TMD, there are three sources of insurance that may be tapped into to address the costs of assessment and treatment (that is, medical [physical and mental health] and dental). As a result, patients often find their carriers passing on responsibility and liability to another carrier. Unfortunately, this leaves many patients without coverage for a costly disorder.

In the mid-1990s, Fricton and Schiffman¹⁶ surveyed the literature and found few cost data regarding TMD pain outside of estimates from the 1970s, which indicated that 40 percent of the total cost of treatment for chronic pain was attributed to craniofacial pain, including TMD. More recently, Brotman¹⁷ found that managed care treatment costs for a patient with orofacial pain often ranged from $12,000 to $20,000 annually. These factors combine to highlight the need to identify patients who are at risk of developing chronic TMD conditions.

Grzesiak¹⁸ provided an overview of acute versus chronic TMD pain, and the rationale for distinguishing the two, in order to develop predictive models that allow for unique and more effective treatment interventions. Early identification and conservative treatment may pre-empt costly, more invasive treatments, lost time from work and the social repercussions of chronic pain and disability.

To predict the progression from acute to chronic pain, Garofalo and colleagues¹⁹ devised an algorithm to distinguish risk factors in patients with acute jaw pain who progress to develop chronic pain from those in patients who do not. They found that Axis I-Group 1 disorders, Axis II graded chronic pain status, non-specific physical symptoms and Characteristic Pain Intensity, or CPI, all assessed using the Research Diagnostic Criteria, or RDC,²⁰ as well as sex, were significant risk factors. Their model correctly classified 77 percent of patients; however, the study was limited by a small sample size and the sole reliance on RDC data.

Epker and colleagues²¹ subsequently refined the original algorithm, based on a larger sample size and additional variables, and found that a logistic regression relying on the CPI and presence or absence of myofacial pain correctly classified 91 percent of the subjects in their study.

This latter algorithm formed the basis of our study in which we classified patients with acute jaw pain who were at risk of developing chronic jaw pain. Our goal was to further characterize patients who are at high risk, or HR, or low risk, or LR, of progressing from acute to chronic pain. Using a biopsychosocial approach to jaw pain, we assessed subjects using both functional and psychosocial measures.

Functional measures included a brief jaw pain evaluation based on the RDC and a chewing performance evaluation. We chose these measures because the tests can be administered easily in a dental or medical office; moreover, these evaluations generally are already being done in some capacity via palpation or questioning (for example, "Are you limited in your ability to chew?"). Our aim was to strengthen the methodology behind these well-used data collection strategies, so that health care providers may make increased use of such data in clinical settings.

We chose psychosocial measures in this evaluation to assess current states of anxiety and depression, as well as personality styles, to help identify people who are at risk of progressing to chronic TMD. Although we do not anticipate that the health care provider will assess patients early on using our extensive psychosocial battery, our aim was to distill the resulting characteristics into an easily identifiable profile that the clinician might observe readily in a clinical setting.

We chose psychosocial measures in this evaluation to assess current states of anxiety and depression, as well as personality styles, to help identify people who are at risk of progressing to chronic TMD.

Our study aimed to assess a wide array of functional and psychosocial measures that provide a snapshot of the patient at risk of developing chronic pain. By arming themselves with this information, practitioners then may make appropriate early referrals to adjunctive care, such as cognitive-behavioral intervention, and prevent the progression from acute to chronic pain that is costly to treat.

	SUBJECTS AND METHODS

TOP ABSTRACT TEMPOROMANDIBULAR DISORDER SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Subjects. Seventy-four patients with complaints of jaw pain or facial discomfort of less than six months’ duration participated in the initial assessment phase of this study. The sample was composed of 60 women (81 percent) and 14 men (19 percent); the mean age was 37.5 years, with a range from 18.00 to 66.25 years. General dentists and oral surgeons in the Dallas/Fort Worth metropolitan area referred patients to the TMD Clinical Research Project at the University of Texas Southwestern Medical Center at Dallas. In addition, we placed fliers at local universities and advertisements in local newspapers to recruit subjects.

Inclusion criteria for the study included the following: adults between the ages of 18 and 70 years who had acute jaw or facial pain (defined as being present for less than six months). Potential subjects were excluded if they had a comorbid pain-exacerbating physical condition (such as cancer or fibromyalgia) or a history of jaw pain.

Procedure. Clinical psychology research personnel (psychologists and masters-level counselors), including two of us (A.W., L.W.), reviewed the purpose and procedures with subjects before obtaining informed consent. All subjects then completed the following self-reported measures:

– a general information questionnaire;

– the Beck Depression Inventory-II, or BDI-II,²² a measure of depression;

– the West Haven-Yale Multidimensional Pain Inventory, or MPI,²³ a measure of pain intensity, related life interference and the ability to manage pain;

– the CPI,²⁰ a measure of pain;

– the Schedule for Nonadaptive and Adaptive Personality, or SNAP,²⁴ a measure of dimensions of personality;

– the Ways of Coping–Revised, or WOC,²⁵ questionnaire, a measure of coping.

The research personnel then interviewed all subjects using the Structured Clinical Interview (SCID I and SCID II),^26,27 which is based on the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, or DSM-IV,²⁸ to determine DSM-IV Axis I clinical disorders and Axis II personality disorders.

Physical examination. On completion of the self-reported measures and the structured interview, subjects received physical examinations according to the RDC examination form.²⁰ All subjects also completed a chewing performance evaluation in which they chewed an artificial test food substance (CutterSil Putty PLUS, Heraeus Kulzer, Armonk, N.Y.) and then expectorated the substance.²⁹ One of us (P.B.) trained the research personnel to administer this evaluation. Buschang and colleagues²⁹ have described the methodology for evaluating the chewing performance.

Based on the study findings of Epker and colleagues,²¹ we assessed only for the presence or absence of myofacial pain. This determination was based on the administration of the Axis I-Group 1a of the RDC examination form, which consists of palpation of 20 muscle sites involved in the diagnosis of myofacial pain, as well as on the subjects’ responses to question number 3 on the RDC history questionnaire (that is, "Have you had pain in the face, jaw, temple, in front of the ear, or in the ear in the past month?").²⁰ An oral surgeon (E.E.) knowledgeable in the RDC trained and periodically recalibrated the clinical research personnel.

The assessment took approximately 2.5 hours. All subjects were paid $70 for participating in the study. Of the 78 subjects screened for the study, 74 were eligible to participate. On completion of the initial assessment, we classified subjects as HR or LR based on the Epker and colleagues²¹ algorithm. We slightly modified the cutoff scores for the predictive algorithm to adjust for sample size. Of the 74 subjects, 52 (70 percent) were classified as HR and 22 (30 percent) were classified as LR.

	RESULTS

TOP ABSTRACT TEMPOROMANDIBULAR DISORDER SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Demographic characteristics. Table 1 presents the demographic characteristics of the HR and LR groups. We identified no significant differences between these groups at baseline.

Cluster analysis of the Axis II personality disorders also revealed that subjects in the HR group were nearly four times as likely as subjects in the LR group to have a cluster C diagnosis (that is, avoidant personality disorder, dependent personality disorder or obsessive-compulsive personality disorder, or OCPD). Most notably, subjects in the HR group were more than three times as likely to have OCPD as were subjects in the LR group.

Functional measures. All subjects completed a chewing performance evaluation, which involved chewing an artificial test food substance, then expectorating it (one subject in the LR group was excluded because of missing data). We analyzed the resulting material according to the average time taken to chew the substance 20 times, the broadness of the resulting particles and the median particle size. We found no differences between the groups on any of these measures.

Predictor variables differentiating HR and LR groups. A primary intent of this study was to perform an in-depth biopsychosocial evaluation of patients classified as being at high risk or low risk of developing chronic jaw or facial pain, based on the predictive algorithm discussed above.²¹ Therefore, we considered it fruitful to examine the current variables in light of their predictive capacity for identifying patients at risk of developing chronic pain. We selected variables for inclusion in the logistic regression equation on the basis of statistical differences that emerged from the baseline analyses. Items considered for inclusion in the regression model, based on their ability to distinguish statistically between the HR and LR groups, were the following:

– CPI total score;

– BDI total score;

– presence of total and specific Axis I and Axis II pathology (that is, GAD, somatization, OCPD, any one of the anxiety disorders or somatoform disorders, and cluster C);

– avoidance on the WOC;

– coping style on the MPI.

To prevent criterion contamination (that is, reliance on a predictor that we already knew predicted group membership, thus contaminating the potential predictive power of new variables), we did not use the CPI in the logistic regression model, because we relied on it as a predictor in determining the original criterion measure (HR versus LR).²¹ In addition, we used the presence of an Axis I or Axis II diagnosis instead of specific DSM-IV diagnostic categories because the predictor variables would otherwise have been redundant. (For example, a diagnosis of major depressive disorder would enter into the regression analysis twice, as a specific diagnosis and as an Axis I diagnosis, thus confounding the predictive power of the algorithm.)

In the regression analysis, we used the presence or absence of adaptive coping styles and interpersonally distressed coping styles from the MPI instead of overall coping style, because none of the patients in the LR group was classified as being a dysfunctional coper. In creating a stable logistic regression model, a rule of thumb is to have 15 to 20 subjects per each predictor variable. For our model, only the HR group had dysfunctional copers (from the MPI) and subjects with an anxiety disorder (from the SCID I). Therefore, we entered the remaining variables into a binary logistic regression (enter method) procedure to determine the array of variables that could best predict membership in the HR versus LR category.

This procedure resulted in a six-factor solution that predicted membership in the HR or LR group with 77 percent accuracy, and with a sensitivity and specificity of 82 percent and 63 percent, respectively. This model (Table 4) correctly classified 87 percent of the patients in the HR group and 55 percent of the patients in the LR group. These six factors were the BDI total score, scores on the WOC Avoidance scale, presence or absence of MPI adaptive coping style, presence or absence of MPI interpersonally distressed coping style, and presence of Axis I or Axis II pathology.

	DISCUSSION

TOP ABSTRACT TEMPOROMANDIBULAR DISORDER SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

Epker and colleagues²¹ developed a two-variable predictive model, based on a six-month follow-up evaluation of patients with acute TMD. They found that their model correctly classified 91 percent of patients who met the criteria for chronic TMD at this six-month evaluation. Our study built upon this model by using it to classify patients early on into the HR or LR group. We then further revised the algorithm by adding biopsychosocial variables to provide insight into further distinctions between HR and LR patients with TMD.

As reviewed earlier, Epker and colleagues²¹ found that the CPI and presence or absence of myofacial pain had statistically significant practical utility in predicting progression from acute to chronic pain. The primary focus of our study was to use this algorithm in a sample of patients with acute pain, in an effort to further characterize those who were at risk of developing chronic problems.

Measures of pain and depression. Our in-depth biopsychosocial evaluation of patients in the HR and LR groups revealed significant differences between the groups in regard to self-reported measures of pain and depression. On both the CPI and BDI, the HR group scored significantly higher than did the LR group. Measures of coping also significantly differentiated the groups; we found that patients in the HR group coped in a significantly more dysfunctional manner.

Axis I and Axis II pathology. In regard to Axis I and Axis II pathology, the HR and LR groups differed significantly. Specifically, patients in the HR group had significantly higher Axis I and Axis II rates of overall psychopathology. Patients in the HR group also had significantly more anxiety disorders and somatoform disorders than did patients in the LR group.

Personality disorders. In addition, we found that personality functioning differed greatly between the groups. The HR group had significantly more Axis II personality diagnoses. There are three main clusters of personality disorders: A, B and C. We found no significant differences between the groups in regard to clusters A (paranoid, schizoid, schizotypal) and B (antisocial, narcissistic, histrionic, borderline); however, the HR group had significantly more cluster C diagnoses. Cluster C diagnoses consist of the following: avoidant personality disorder, dependent personality disorder and OCPD. Key traits of this cluster center on anxiety and fearfulness.

Within this cluster, the HR group had a significantly greater prevalence of OCPD. Key features of OCPD are "preoccupation with orderliness, perfectionism, and mental and interpersonal control, at the expense of flexibility, openness, and efficiency."^28(p660) We will continue to evaluate the implications of the high rate of OCPD in our ongoing research study.

At present, we hypothesize that the presence of OCPD among patients with acute jaw pain who are at high risk of developing chronic pain is the result of their tendency to seemingly bite back emotions, particularly of an unpleasant nature, and to maintain a façade of control and socially acceptable behavioral responses. Physically, this biting back of emotions may manifest itself in grinding the teeth, clenching the teeth or tensing the jaw muscles. These patients do not appear to outwardly display emotions that might alleviate pent-up anger and frustrations and allow for a reduction of psychophysiological responses (that is, increased cortisol levels, muscle tension and pain).

Need for early assessment. These findings highlight the need for early assessment of both physical and psychosocial variables to identify patients with acute pain who may progress to experience chronic pain that is more costly to treat. Such early identification will enable clinicians to refer patients for adjunctive care (such as cognitive-behavioral therapy provided by a behavioral pain-management psychologist and a physical therapist trained to work with patients in acute and chronic pain). This treatment should be aimed at all etiologic factors (that is, physical, behavioral and psychological).

Our prediction model, consisting of six factors (excluding the CPI), correctly classified 57 (77 percent) of 74 patients in this study. The model was limited somewhat by low power, caused by the relatively small number of patients at this early stage in our four-year research study. Despite this, one of the predictors, the presence or absence of an Axis I SCID diagnosis, was quite robust. Subjects who received an Axis I diagnosis (for example, somatoform disorders, anxiety disorders) were more than 6.5 times as likely to be in the HR group than were those who did not receive such a diagnosis. Overall, among all subjects, 20 (27 percent) had an anxiety disorder (HR = 2 0 subjects [38.5 percent]; LR = 0 subjects), 30 (40.5 percent) had a somatoform disorder (HR = 27 subjects [51.9 percent]; LR = three subjects [13.6 percent]) and 29 (39.2 percent) had an affective disorder (HR = 24 subjects [46.2 percent]; LR = five subjects [22.7 percent]).

In addition, we should note that, although high scores on the BDI significantly differentiated the HR and LR groups when assessed as individual predictors (univariate analysis), high BDI scores no longer served as robust predictors when we considered all other significant predictors in the logistic regression. The presence of an Axis I diagnosis—and specifically an anxiety disorder—eclipsed the predictive value of the BDI when they were considered together.

This demonstrates that patients with a more anxious component to their illness tend toward being at high risk, whereas those with a depressed component are more likely to be at low risk. A primary difference between the two groups is the increased number of somatic symptoms, such as clenching and grinding, among patients with anxiety disorders, as well as an increased tendency to deny the intensity of their feelings.

In contrast, patients who report experiencing symptoms of depression may have more awareness of and insight into their discomfort and, consequently, may be more likely to be at decreased risk of developing chronic pain. Because of potentially increased awareness of their emotional distress, these people may be more likely to seek treatment for their physical or emotional discomfort or find an alternative means of lessening their discomfort themselves (such as reducing stress levels in their lives).

Several studies have examined the contribution of awareness and insight as they affect the response to treatment and compliance.^30–35 Although these studies largely have been limited to mood and psychotic disorders, findings of increased insight and improved treatment response among patients with depression compared with findings among patients with mania or psychosis are informative. These findings provide some evidence in support of the theory that patients with heightened anxiety levels or excessive, purposeless, poorly directed energy may have difficulty recognizing the need for, and focusing on, treatment aimed at reducing anxiety.

In fact, some studies have assessed the role of attention and memory and have found some evidence of reduced abilities among patients with anxiety.^36–39 In future research, we will further examine the role of anxiety and depression in the maintenance of symptoms, response to treatment and progression to chronic pain.

Functional measure. It is important to note that the functional measure, chewing an artificial food substance, did not significantly differentiate between the groups. This is not unexpected, because all patients in this study had significant complaints of pain, regardless of which group they were placed into based on the risk algorithm. This underscores the importance of moving beyond purely physical measures and assessments to more fully assess these potentially HR patients in the early stages of pain.

As stated above, a major distinction between patients in the HR and LR groups was the presence of an anxious component. Subjects who had more insight into the probable stress-related nature of their illness were less likely to develop chronic pain. Consequently, health care providers who assess patients’ perceptions of the meaning of their pain may be able to more effectively and accurately predict the pain progression than those who rely on self-reports of pain intensity alone.

The original algorithm by Epker and colleagues²¹ centered on self-reports of pain as well as palpation as the primary predictive determinants of risk status. Our research adds an understanding of the contributions of the patient’s psychosocial functioning. Based on our research thus far, the snapshot of a patient at high risk of developing chronic pain includes self-reports of significant or extreme pain, presence of anxiety and a rigid, uptight manner.

When health care providers see such patients, they should consider referring them for adjunctive care to address the multiple factors that contribute to the onset and maintenance of pain. Further research, aimed at developing treatment interventions for these high-risk patients, is imperative. Toward that end, we plan to follow up these patients and monitor their long-term outcomes after they participate in an early-intervention cognitive-behavioral treatment program we have developed.

	CONCLUSION

TOP ABSTRACT TEMPOROMANDIBULAR DISORDER SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Our study has built on previous research that found self-reports of pain and physical examination findings to be predictive of the progression from acute to chronic pain. By adding predictive biopsychosocial variables to the algorithm, we have provided health care providers with a more complete snapshot of patients who are at risk of developing chronic TMD. Early intervention of a biopsychosocial nature then may be initiated to prevent the costly and time-consuming progression from acute to chronic pain.

View larger version (144K): [in this window] [in a new window]   Dr. Wright is an assistant professor, Department of Psychiatry and Department of Anesthesiology and Pain Management, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas, Texas 75390-9044, e-mail "Anna.wright{at}utsouthwestern.edu". Address reprint requests to Dr. Wright.

View larger version (131K): [in this window] [in a new window]   Dr. Gatchel is a professor, Department of Psychiatry and Department of Anesthesiology and Pain Management, University of Texas Southwestern Medical Center at Dallas.

View larger version (143K): [in this window] [in a new window]   Ms. Wildenstein is a research associate, Department of Psychiatry, University of Texas Southwestern Medical Center at Dallas.

View larger version (138K): [in this window] [in a new window]   Dr. Riggs is in private practice limited to orofacial pain/TMD and is an associate professor, Department of Restorative Sciences, Baylor College of Dentistry, The Texas A&M University System Health Science Center, Dallas.

View larger version (131K): [in this window] [in a new window]   Dr. Buschang is a professor, Department of Orthodontic Research, Baylor College of Dentistry, the Texas A&M University System Health Science Center, Dallas.

View larger version (116K): [in this window] [in a new window]   Dr. Ellis is a professor, Department of Oral and Maxillofacial Surgery, University of Texas Southwestern Medical Center at Dallas.