Towards an optimal therapy strategy for myogenous TMD, physiotherapy compared with occlusal splint therapy in an RCT with therapy-and-patient-specific treatment durations

Like in a traditional RCT, stringent inclusion and exclusion criteria were applied to select myogenous TMD patients. In contrast to previous RCT studies on myogenous TMD patients [14, 15, 18, 33, 34], the entire treatment procedure using TDC was similar to one in general clinical care. This procedure included counselling, a therapy-and-patient-specific duration of treatment, and possibly intermediate adjustment of an occlusal splint. Furthermore, patients whose therapy was successful were allowed to apply splint wearing or physiotherapy (home exercise) respectively when they would feel a need. The influence on pain intensity of the standardized reassurance and counselling which preceded both types of therapy, will first be discussed, followed by the influence of treatment on pain intensity. The post-treatment TDC-values will be discussed thereafter.

The intensity of the predominant pain from the masticatory system decreased on average 32% between the initial visit (100%) and the start of treatment with a CD of 16.6 scale-%. The value of Cohen’s d, on average 0.80, reflects a moderate/large effect of counselling on pain intensity. Pain intensity was constant in a pain diary, during 2 weeks before the start of treatment [23]. Hence, the decrease of pain intensity occurred shortly after counselling rather than late in the waiting period of 4.4 weeks.

Regarding other studies, in a heterogeneous sample of TMD patients [35], pain intensity decreased similarly by 34% (CD: 22.8 scale-%), with an effect size d of 1.13 (mean and SD values of d derived from a Figure), following counselling and a waiting period of 2 weeks. A similar decrease of 15.4% (CD 10.1 scale %, d: 0.73) of pain intensity was observed 1 month following initial visit and solely counselling, during maximum unassisted jaw opening in myogenous TMD patients whose jaw opening was limited (RDC/TMD: class 1b) [36].

Because of a general association with pain and the decrease of pain intensity following counselling, the signs and symptoms of one patient assigned to physiotherapy decreased below the threshold following the initial visit, so that this patient became a dropout. Hence, for patients with a low level of signs and symptoms at the initial visit, treatment may be restricted to counselling and instructing simple stretch and auto-massage techniques. A future application of the TDC-procedure from the initial visit will be of interest to determine objectively for which patients counselling and education will be sufficient to attain functional status.

Because counselling at the initial visit causes a fast decrease in pain intensity which stabilizes, a further decrease must be due to the applied therapy, rather than to a natural course of the disorder. Cohen’s d was 0.86 and 1.39 for physiotherapy and splint therapy respectively. Thus the effect on pain intensity of these therapies which were controlled by the TDC-procedure is large (d > 0.80). These d-values and the CD-values (21.8–27.6 scale-%) are similar to those from previous studies on mainly myogenous TMD patients without overt psychosocial factors, regardless of the type of pain of which the intensity was considered. In these previous studies [14, 15, 33, 34, 37], the effectiveness of either physiotherapy or splint therapy has been examined using a constant treatment duration between 5 and 10 weeks, and treatment was preceded or combined with counselling including reassurance of the patients. Following splint therapy, CD of maximal pain was 23.2 scale-% (final evaluation after 5 weeks, active splint group) [33] and 28.8 scale-% (12 months) [14, 15] and d was 0.99 and 1.48 respectively. Following physiotherapy, CD of present pain was 22.5 scale-% for intra-oral physiotherapy alone and 34.9 scale-% for a combination of such physiotherapy and education (12 months) [34]. In another study on physiotherapy (6 weeks, combined groups) [37], CD was 25.4 scale-% for maximal pain and d was 1.35. The number of treatment visits (10–18) in the previous studies on physiotherapy was similar to the ones in the present study (mean 11.5 visits, SD 2.0). The number of treatment visits for splint therapy was five in the study of Raphael and Marbach [33] which is similar to the mean one in the present study (4.4 visits), and at least two visits in the study of Ekberg et al. [15]. The number of visits from the previous studies on splint therapy does not include additional visits for a withdrawal of wearing the splint when splint therapy became successful. Treatment durations of 5–10 weeks in these studies are therefore a lower limit of a realistic duration.

The abovementioned findings show that when splint therapy is preceded or combined with counselling including reassuring the patients, this therapy has a large effectiveness which is similar to that of physiotherapy from separate studies. However, without such counselling, there was no effect of splint therapy on pain intensity [18]. Following 3 months of splint therapy (6 visits with only splint control), CD for intensity of spontaneous muscle pain was −2.8 scale-% and Cohen’s effect size (ES: ratio between difference in means and baseline SD) was merely −0.14 (negative sign: slight worsening). An effect of solely splint therapy was also lacking for pain intensity during gum chewing [18], i.e. CD was 3.4 scale-% and ES was 0.11. The effect of solely counselling and education which was repeated 6 times during 3 months was moderate for intensity of spontaneous muscle pain [18], i.e. CD was 11.2 scale-% and ES was 0.58 and small for pain intensity during gum chewing (CD: 7.0 scale-%, ES: 0.30). Thus all findings suggest that a synergy occurs between counselling and splint therapy, i.e. absence of counselling blocks the potential effect of splint therapy and/or splint therapy enhances the effect of counselling. Because the effect of a one-time counselling at the initial visit was stabilized in the present study before splint therapy was started, blocking of the effect of splint therapy is probably involved anyhow.

The small/moderate effect of solely counselling during 3 months was reflected in a relatively large VAS-score of intensity of spontaneous muscle pain (30 mm on a 100 mm VAS) or pain following chewing (38 mm) at the end of the program [18]. Such end levels of pain intensity, which are even higher than the mean ones for patients whose treatment was unsuccessful in the present study (Fig. 2), are much higher than that of patients with a successful treatment, who have attained functional status (mean pain score < 7 mm; Fig. 2). Hence, solely counselling and education will, in general, not be sufficiently effective to attain functional status within a reasonable time. This conclusion is reinforced by findings from another study on myogenous TMD patients (RDC/TMD: class 1b) [34]. Although monthly counselling continuously decreased the intensity of pain at maximum unassisted mouth opening, this intensity was still large following 3 months (43 mm) and 12 months (27 mm). The large effectiveness which was observed for physiotherapy in the present study might in part be due to a repeated education procedure which is inherently part of this type of physiotherapy.

In the present study, Cohen’s d of pain intensity was significantly larger for splint therapy (d = 1.39) than for physiotherapy (d = 0.86), for the entire therapy groups. However, the inter-therapy d-values were similar for the sub-samples of patients whose treatment was successful (d = 2.02–2.07) or unsuccessful (d = 0.39–0.73). The larger overall d-value for splint therapy reflects a value of success rate for splint therapy which, although not significant, was slightly larger than that for physiotherapy (Table 2). Thus slightly more patients with a successful treatment and therefore a larger d-value, contribute to the overall d-value of splint therapy.

Advantages of considering the post-treatment TDC outcome as a measure of effectiveness are an involvement of scores from various signs and symptoms of myogenous TMD rather than from solely pain intensity, and therefore also more statistical power to detect possible inter-therapy differences. The similar mean post-treatment TDC-values for physiotherapy and splint therapy (−0.512– −0.575, Table 3) indicate a similar effectiveness and a large effect size (TDC < −0.379) for both types of therapy. The post-treatment TDC-values were also similar between both therapies, in subgroups of patients for which treatment was successful or unsuccessful. The post-treatment TDC-values for patients whose treatment was unsuccessful (−0.200– −0.208) which were larger (less negative) than others, differed nevertheless significantly from zero (TDC = 0 means no change). Thus even in patients whose treatment was unsuccessful, their signs and symptoms improved significantly and similarly for both therapies. The success rate based on post-treatment TDC-values from individual patients is also similar between both therapies, i.e. 73–83% in the short-term and 51–60% in the long-term (Table 2).

Patients with mainly basic signs and symptoms of myogenous TMD were selected in the present study. This selection is reflected in mean baseline values of various psycho-social factors which, in general, correspond to a low degree of involvement [23]. Significant differences did, in general, not occur between these baseline values for patients whose treatment became successful or unsuccessful respectively. Only internal locus of control differed for physiotherapy, i.e. patients whose physiotherapy became successful believed more that own health is controlled by their own behaviour than patients whose physiotherapy became unsuccessful. This outcome may be related to the degree of the patients’ active participation during physiotherapy. The difference in mean score level was however small, i.e. 9.2% with respect to the mutual mean. Baseline values of age or clinical variables from the present study did also not influence a patient’s responsiveness to a therapy. Regardless of the type of therapy, a modest use of OTC-medication (mean: 3–7% of the times of scoring) is not specifically related to any of the two treatment outcomes (successful/unsuccessful). Although larger patient groups are needed for more power in testing frequencies, this finding suggests that such a modest use may have only a small influence on success rate.

Despite a selection of ‘basic’ myogenous TMD patients, the success rates in the short-term (73–83%) were similar to that reported in text books for TMD in general [1, 38, 39]. The success rates in the long-term (51–60%) were such that a large fraction of the patients, on average 44%, still needed a subsequent treatment, regardless of the type of initial therapy. In contrast to previous studies, profoundly confounding factors which might hamper an improvement in central pain mechanisms, were excluded in the present study. A patient’s treatment may therefore become unsuccessful merely because the patient’s speed of recovery is intrinsically low. Post-treatment TDC-values indicate that even for patients whose treatment is unsuccessful, their TMD signs and symptoms have improved significantly. This improvement is also reflected as a significant decrease in pain intensity alone. The decrease in pain intensity does not continue beyond the end of treatment, as although not significant, the mean pain intensity was slightly larger at the first post-treatment visit than at the end of treatment about 4 weeks earlier (Fig. 2). For patients whose initial therapy is unsuccessful, stepped-care with a subsequent therapy is therefore necessary to attain a further decrease in pain intensity and other signs and symptoms beyond a critically low residual level.

Our suggestions for an optimal treatment strategy refer to myogenous TMD which is not confounded by overt psychological problems, or severe sleep bruxism (SB) as diagnosed instrumentally [40]. A multi-therapy modality is more appropriate for patients with major psychological problems [20]. An occlusal appliance may still be preferred for active SB [40], because, apart from reducing SB, further extreme tooth wear is prevented.

In order to assess the outcome of stepped-care from the present study, a single telephone survey was carried out at the end of the project, 1–4 years following the end of a trajectory of 1–2 therapies [21]. This survey, using the anamnestic questionnaire with eight items from the TDC-procedure (assessing pain intensity by a 0–10 point score), revealed that 93% of all patients had no need for treatment anymore. This success rate for stepped-care corresponds with the mean of the range of overall success rates (91–95%) in the stepped-care model using the basic success rates of physiotherapy and splint therapy, and a value of the modulation factor m which is larger than one (m = 1.5; Table 5). This finding suggests that, regardless of the sequence of physiotherapy and splint therapy, the initial treatment triggers an improvement that, although insufficient for patients whose initial treatment becomes unsuccessful, enhances the success rate of the subsequent treatment.

Physiotherapy may be preferred as initial therapy over occlusal splint therapy in stepped-care of myogenous TMD, for four reasons. First, with a similar success rate and effectiveness, the duration of physiotherapy is, on average, 10.4 weeks shorter. Hence, patients whose initial physiotherapy is unsuccessful can continue earlier with subsequent treatment thus lowering the risk on a sustained chronicity of pain.

Second, the stepped-care model shows that small and even moderately large inter-therapy differences in basic success rate are not important. The overall success rate of stepped-care does not depend on the therapy sequence, irrespective of any inter-therapy difference in basic success rate, if the success rate of a subsequent therapy is not influenced by the preceding one (m = 1). For a wide range of m-values around m = 1 (0 < m < 2), the difference in overall success rate between the two possible trajectories is smaller than the inter-therapy difference in basic success rate. Hence, the overall success rate of stepped-care hardly depends on therapy sequence. For basic myogenous TMD, an unsuccessful initial treatment most likely enhances the success rate of the subsequent treatment (m ≈ 1.5). The largest overall success rate of stepped-care is then even attained by choosing the therapy with the lower basic success rate as the initial therapy, hence physiotherapy.

Third, the same therapeutic techniques of physiotherapy can be applied to decrease pain or pain-related function impairment of other body parts than facial ones, like the neck and shoulders. Pain was not restricted to facial areas in 69% of the patients (Table 1). A broader application of physiotherapy might even increase its basic success rate. Fourth, at least in the Netherlands, the costs of a trajectory that starts with physiotherapy, are lower.

Physiotherapy might have a disadvantage, i.e. the number of visits is on average 7.1 larger than for splint therapy. If availability of a physiotherapist is limited, for example for patients living in rural areas, or if patients have limitations of transportation and/or time, splint therapy might still be preferred as initial treatment. However, technology-assisted interventions, including internet-based ones [41, 42], may diminish the disadvantage of physiotherapy. In-person treatment may, at least in part, be replaced by internet-assisted treatment.

This study includes some limitations. First, patients with mainly basic signs and symptoms of solely myogenous TMD were selected to facilitate straightforward conclusions on therapy outcome. The prevalence of solely myogenous TMD is low in samples of patients who are referred to a specialized clinic, for example, 4.5% of all TMD patients in a tertiary clinic [43], and 9.0% in the present study where patients were mainly (85%) referred to a clinic for Special Dental Care. However, this prevalence is much higher in a more general clinic, i.e. 39.5% in an outpatient physical therapy practice (myogenous TMD without arthralgia) [44], and 47.8% in a university dental clinic [45]. The prevalence is also higher, 36.1%, in a community sample of young women (age 19–23 years) [46]. Hence, the results of the present study are representative for about 40–48% of all TMD patients, i.e. patients who have an age range of 18–65 years (median at about 32 years), and a need of treatment for solely myogenous TMD. This conclusion is reinforced by the finding that the patients from the present study are largely comparable to those with myogenous TMD from previous studies in which selection was less amended (for a detailed discussion, see ref [23]).

Second, patient groups with placebo treatments or a long waiting period were not included for ethical reasons. Using a placebo group in non-pharmacological studies may be theoretically, methodologically, practically, and ethically unsound [47]. Regarding a placebo effect of occlusal splint therapy, the effectiveness of a splint which does not cover the teeth but only the palate (palatal splint) is either similar [12], or corresponds in part to the effectiveness of an occlusal splint [15, 33]. However, occlusal splint therapy and physiotherapy have in common an extensive non-noxious mechanical stimulation of peri-oral and/or intra-oral tissues. A palatal splint also causes mechanical stimulation of intra-oral tissues. Non-noxious mechanical stimulation, hence stimulation of afferent A-β fibres, may relieve pain by modulation of central pain mechanisms [48, 49]. In order to avoid any stimulation of tissues, a sham laser treatment may be an appropriate placebo treatment. Recording the patients’ expectation on improvement will then be useful for assessing the role of this factor in effectiveness of various treatment modalities. Recording this expectation will also solve the third limitation of the present study.

Although a non-treatment group was not included in the present study, two waiting periods of about 4 weeks were present in both treatment modalities, before and after treatment. These waiting periods were long enough to show, in addition with a 2-week-pain diary, that the effect of one-time counselling at the initial visit (all patients) stabilized as well as the effect of an unsuccessful treatment. An initial treatment was stepped up for ethical reasons, when it was unsuccessful at a post-treatment occasion of evaluation. Because the decrease of pain intensity was halted following the end of an unsuccessful initial treatment, it is unlikely that the success rate and the last TDC-value would have altered by a spontaneous improvement if a follow-up of 1 year were completed. In a study using a waiting list of 1 year [34], intensity of various types of pain did not change in myogenous TMD patients.

Bias by regression to the mean is avoided by using TDC-values as outcome parameter [19]. TDC is therefore useful in future research to examine changes of all pronounced signs and symptoms during all phases, thus from the initial visit up to the end of a possible second treatment of a trajectory.