We found four other randomly controlled trials of SS versus CS in the literature, all of which were carried out at specialist centers. When we compared these trials with our study, two showed similar results. Dao and colleagues² reported no difference between splint types, and Rubinoff and colleagues³ reported a minor difference. By contrast, in their two studies (one concentrating on pain of arthrogenous origin and the other concentrating on pain of myogenous origin), Ekberg and colleagues^4,5 found that SSs were significantly better than the CSs. The arthrogenous group was followed up after one year.⁶ Such follow-up is unusual, as most studies only report the results after initial treatment. In a systematic review of occlusal treatments for TMD, Forssell and colleagues⁷ found that only three of 18 randomly controlled trials had followed up subjects for more than six months.

As initial treatment may not guarantee long-term success, it is important that patients with TMD return for a follow-up visit after a reasonable length of time. In our trial, we considered such follow-up essential to determine whether initial success of splint treatment was not merely a short-lived response. Longer follow-up also gives patients time to reflect on their treatment management, and clinicians can determine whether splints still are being used. In long-term trials, clinical outcome measures can be repeated to supplement subjective scores, which provides a more robust assessment than follow-up visits that are restricted only to questionnaires or telephone calls.^8,9

The aims of our study were

– to determine if there had been any significant changes in outcome measures between the end of splint treatment and a one-year follow-up for subjects showing improvement after initial treatment;

– to investigate subjects’ perceptions of how effective the splint had been in treating jaw pain;

– to investigate subjects’ perceptions of how clicking had changed from the start of treatment;

– to ascertain the need for continued splint wear and its frequency.

	SUBJECTS, MATERIALS AND METHODS

TOP ABSTRACT SUBJECTS, MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
We published details of initial treatment in 2004.¹ In this article, we summarize the one-year follow-up part of our study. Five local ethics committees in the areas in which the dentists practiced granted ethical approval for our trial.

We originally enrolled 93 patients into the trial using a concealed randomization process in which dentists were blinded to the allocation until after the patient had been entered into the study. The trial’s inclusion criteria were

– subjects’ being 18 years or older;

– the presence of pain in temporomandibular joint (TMJ), muscles or both, plus one or more of the following: joint sounds (clicking or crepitus), a history of jaw locking or limitation of opening, muscle tenderness on palpation and TMJ tenderness on palpation;

– symptoms that had been present for more than four weeks;

– sufficient teeth to support and occlude against a mandibular splint.

We excluded patients who were dentists, those who were admitted as emergency patients and were in acute pain and dysfunction that required other forms of treatment, and those who were unable to cooperate with the trial requirements. We did not charge the subjects for treatment. The participating dentists attended a training day to ensure consistency of examination, splint fitting and splint adjusting.¹⁰

We diagnosed the subjects’ TMD according to criteria developed by the International Headache Society¹¹ and the American Academy of Craniomandibular Disorders¹² (now the American Academy of Orofacial Pain) (Table 1). The diagnoses for the 77 subjects who started treatment and for the 52 subjects who were followed up for one year are shown in Table 2. We made no attempt to tailor treatment according to diagnosis.

Of the 77 subjects who started treatment, 72 subjects completed treatment. Nine of these subjects were male and 63 were female, and the mean ages (± SD) were 35.1 ± 8.7 and 37.7 ± 11.9, respectively. The subjects’ ages ranged from 19 to 65 years. Table 3 summarizes the activity at each visit during three months of treatment with either an SS or CS. At visit 5, we identified subjects who were not responding to the CS (< 50 percent reduction in pain) and crossed them over to the SS group at visit 6 and gave them three additional visits. If they then responded to treatment (> 50 percent pain reduction), we scheduled them for a one-year follow-up visit. If they did not, we referred them to the dental hospital. On completion of active treatment, we instructed subjects to reduce progressively the amount of time they wore the splint from full-time to nighttime only to only as needed.

To avoid bias, we examined subjects clinically after they had completed a pain VAS and answered a questionnaire in the waiting room. We asked subjects to mark the VAS indicating the worst pain they has experienced over the previous week. The questionnaire asked the following questions:

– How effective do you think the splint has been in reducing jaw pain?

– If your jaw clicked at the start of treatment, how is the clicking now? If the clicking was painful at the start of treatment, how has this changed?

– Over the past six months, have you had to wear your splint? If so, how often did you wear it? When during the day did you wear it?

Our clinical outcome measures were

– VAS of worst pain experienced over the previous week;

– interincisal opening (maximum opening with pain);

– number of tender muscles (masticatory and cervical);

– aggregate joint tenderness (determined by giving a score of 1 for tenderness laterally, in the external auditory meatus or on movement and adding the scores for both TMJs for a maximum total of 6);

– clicking (presence or absence);

– headaches (number per week).

We transcribed the data into a statistical spreadsheet and a graphics package. We analyzed any differences between subject groups at baseline (including differences in the CS group between subjects who we crossed over to the SS group and those who we did not) using either analysis of variance (ANOVA) for continuous variables or a ² test for categorical ones. We could only make direct comparisons between groups up to six weeks, after which some subjects in the CS group had been crossed over to the SS group. We analyzed the difference between the end of active treatment and one-year follow-up among the splint groups using ANOVA or a ² test, taking into account subjects who had been referred for specialist treatment. We analyzed the proportions of subjects in each splint group who responded to the various questionnaire options using a ² test. The level of significance was P < .05.

	RESULTS

TOP ABSTRACT SUBJECTS, MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
At baseline, we found no significant differences between CS and SS groups for any of the outcomes measured. Neither were there any significant differences between subjects in the CS group who we crossed over to the SS group and those who we did not. Our short-term treatment results showed that there were no significant differences between SS and CS groups for any of the selected outcome measures at six weeks. In Figures 2 through 4, the apparent improvement of the CS group at nine weeks is due to the crossover of the nonresponders from the CS group to the SS group. Also, the response of the crossover group before nine weeks is shown as a dashed line.

View larger version (56K): [in this window] [in a new window]   Figure 2. Worst pain experienced over previous week. The lettered arrows relate to the treatment stage. The splint was fitted at A. Subjects in the control group who were not responding crossed over to the stabilizing group at B. Most subjects finished active treatment at C, but if they crossed over, they finished at D. This resulted in a staggered one-year recall for most subjects (E) and for crossover subjects (F). mm: Millimeters. VAS: Visual analog scale.

View larger version (58K): [in this window] [in a new window]   Figure 3. Number of tender muscles of 24 regions palpated: deep masseter, superficial masseter, anterior temporalis, posterior temporalis, temporalis insertion, medial pterygoid, lateral pterygoid, sternomastoid, mylohyoid/digastric, posterior intrinsic neck, trapezius, shoulder. The lettered arrows correspond to the treatment stages described in Figure 2.

View larger version (62K): [in this window] [in a new window]   Figure 4. Percentage of subjects with clicking temporomandibular joints on clinical examination. The lettered arrows correspond to the treatment stages described in Figure 2.

Of those subjects who were aware of clicking at one year, 73 percent considered their clicking to have been reduced, and 85 percent reported that their associated pain had been reduced or had gone away. Nevertheless, only 19 percent of subjects thought that their clicking had been eliminated. There were no significant differences between the three splint groups.

In relation to the effectiveness of the splint in reducing jaw pain, 81 percent of all recalled subjects rated it as either good or excellent (Table 5).

	DISCUSSION

TOP ABSTRACT SUBJECTS, MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Our one-year recall in combination with the results of initial treatment revealed a number of important issues in relation to the management of TMD in general dental practice. These include the importance of occlusion in TMD, difficulties in diagnosis, appropriateness of the SS, therapeutic effect of the CS and the need for referrals.

Importance of occlusion. While there are patients with TMD with apparent occlusal etiology who benefit from occlusal treatments when other conservative treatments fail, they are not typical of patients who are seen routinely in general dental practice. In our study, the performance of the CSs matched that of the SSs for all outcomes after six weeks of treatment. The one-year follow-up showed that subjects treated with a CS continued to do as well as those treated with an SS. Our one-year recall findings were influenced by the fact that we needed to move some subjects between groups during initial treatment and to refer a limited number of subjects who did not respond to treatment to a dental hospital (Figure 1). Nevertheless, after initial treatment, only six subjects, according to predefined criteria, needed occlusal adjustment.¹ Our criteria were that, on splint removal, the subject experienced the return of pain with evidence of occlusal interference, an awareness of an uncomfortable occlusion or both.

Some might argue that the need for continued splint wear indicates the need for irreversible occlusal treatment. We found, however, that at one year fewer than one in six subjects still wore their splints on a daily basis, with the majority wearing them only at night. We were reassured to find that more than four of five recalled subjects thought their splints were either good or excellent in reducing jaw pain, whether or not they were supplied with an occlusal surface. These results do not suggest that a secondary phase of irreversible occlusal treatment is either necessary or desirable for the majority of patients with TMD in general practice.

Difficulties in diagnosis. We encountered some diagnostic difficulties despite using established criteria,^11,12 especially in relating TMJ clicking to a diagnosis of disk displacement with reduction (DDR). We found that subjects were aware of more of their TMJ clicks than dentists observed. This would suggest that dentists were not especially sensitive in palpating what subjects perceived as clicking, that clicking was intermittent or both. For our study, we define a DDR as a clicking sound on both opening and closing. The research diagnostic criteria (RDC)¹³ have improved the reliability of diagnosing DDR somewhat by including the presence of clicking during excursions, eliminating the click during protrusion and specifying that the click must be present in at least two out of three repeated jaw movements.

The other advantage of the RDC is that, in addition to making a physical "axis I" diagnosis, they also provide an "axis II" diagnosis of pain-related disability and psychological status, which helps provide a measure of the severity of the condition. In our study, we were able to quantify severity only in terms of physical findings (for example, number of tender muscles, limitation of opening and clicking), pain intensity and headache experience. A measure of how the condition affected the subjects’ quality of life would have been helpful.

Appropriateness of the SS. Our concern over making a reliable diagnosis of DDR relates to reports on the relative ineffectiveness of treating DDR with a SS.^14,15 Nevertheless, after initial treatment, we found that only 25 percent of subjects diagnosed with DDR had been referred to a dental hospital as nonresponders. Furthermore, at one year, our subject questionnaires established that SSs were effective in reducing the severity and pain from clicking in the majority of subjects diagnosed DDR. However, both the SS and CS were ineffective in reducing the proportion of subjects with clicking detected by dentists, and patients need to be aware of this limitation. Further work may help clarify which splint design is best for patients with specific types of TMJ clicking.

After initial treatment, only 25 percent of subjects diagnosed with disk displacement with reduction had been referred to a dental hospital as nonresponders.

In the introduction to this article, we mentioned two studies that compared SS with CS that showed no significant difference in outcome^2,3 and another two studies that reported the SS to be superior for both arthralgia and myogenous TMD.^4,5 The reason for this difference is unclear, but it may involve multiple factors, rather than simply being related to the physical diagnosis. Such factors may include the psychosocial characteristics of the groups being treated, as well as the impact of TMD on their lives. Again, further work is needed. From a purely clinical standpoint, though we use SSs less than we did previously, we still find them useful for some of our patients, especially when we suspect an underlying occlusal problem and when patients do not show a response to simpler treatments.

Interestingly, similar proportions of subjects who wore SSs and CSs thought their splints were either good or excellent in reducing jaw pain. Nevertheless, subjects may incorrectly ascribe the reason for symptom improvement. As we discuss in the following section, the true effect of the splint is difficult to disentangle from other elements of treatment management (for example, the patient/dentist relationship).

Therapeutic effect of the CS. While the CS did not involve the occlusal surfaces, it would be incorrect to think of it merely as a placebo. The presence of the lingual flange may have had an influence on tongue position and oral perception. However, TMD in general can be self-limiting, and it is possible that some subjects could have shown improvement spontaneously rather than this improvement’s being solely attributable to the result of treatment. Alternatively, subjects in the CS group who showed little improvement during the initial treatment may have been slow responders rather than nonresponders. There is evidence that prolonged treatment of whatever type can be beneficial for some patients.¹⁶ Also, this same review¹⁶ and a separate meta-analysis of occlusal splints¹⁷ have shown that both placebo and active treatments are significantly better than no treatment at all. In our study, while it would have been ideal to have a nontreatment group with which to compare the SS and CS groups, the practicalities of doing so in general practice were prohibitive.

All of the dentists who participated in our study were aware that patient counseling⁹ and an empathic manner are important factors in influencing treatment outcome and had been instructed not to vary their approach to subjects in the CS group.

Need for referrals. As patients are more likely to be seen quickly in general dental practice and most appear to respond to relatively uncomplicated treatment, the GDP in-office treatment approach should be encouraged. In the United Kingdom, it would reduce hospital waiting lists and free up hospitals to deal with more intractable cases, diagnostic difficulties and patients who have not responded to primary treatment by their dentists. There still are, however, issues of training and suitable remuneration that need to be resolved before such a clinical pathway could be implemented.

When referring patients, dentists should consider, in addition to the physical diagnostic findings, other factors such as behavioral and psychosocial factors, which have been highlighted in the literature as critical with regard to the development of chronic pain.^19,20

In our instructions to dentists participating in our trial for referring subjects to the dental hospital, we gave them a pain reduction of less than 50 percent as a guide to define nonimprovement. Although this is pragmatic and widely used in medicine,^21–23 it is empirical and deserves further investigation.

Trial critique. Our 2004 trial and this one-year follow-up provide a detailed analysis of how symptoms of TMD respond to SSs provided in a general dental practice. Most clinical trials can be criticized, and ours is no exception. Possible criticisms of the treatment phase include the following:

– a trial design that ideally should have crossed over nonimproving subjects from the SS group to the CS group but was constrained from doing so ethically;

– diagnostic criteria that were limited;

– the fact that 21 subjects and two of the 11 trial dentists dropped out from the initial treatment, and that a further nine subjects were lost by the time of the one-year follow-up;

– that clinical assessments were not made blind to treatment allocation, as the practice setting made it difficult for follow-up examinations to be made by personnel other than the dentist providing treatment;

– at one year, we did not monitor whether subjects were receiving other TMD treatment, which could have been in the form of self-prescribed analgesics, other mainstream treatments or alternative treatments.¹⁸

	CONCLUSIONS

TOP ABSTRACT SUBJECTS, MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
At one year, we found that a good response to TMD treatment in general practice had been maintained for both clinically evaluated and subject-rated outcomes. More than four of five subjects considered their splint treatment to have been either good or excellent. The only outcome to return to pretreatment levels according to dentists’ evaluations was the percentage of subjects for whom clicking TMJs returned. Nevertheless, the majority of subjects considered that their clicking had reduced, and most clicks that were painful had become less so.

We recommend that dentists review their patients’ splint use at long-term follow-up visits because many patients will continue to wear the splints, some on a regular basis, though most just occasionally. Our study confirms that when they are interested and suitably trained, GDPs can manage the treatment of four of five patients with TMD with reversible treatment in their practices with only a small proportion of patients’ needing referral for specialist attention. Finally, though our work was carried out in the United Kingdom, we believe that there is nothing to suggest that there would different conclusions under similar circumstances in the United States.