Effectiveness of occlusal splint therapy in the management of temporomandibular disorders: network meta-analysis of randomized controlled trials

Abstract

A network meta-analysis (NMA) of randomized controlled trials (RCTs) was performed to assess the effectiveness of various types of occlusal splint in the management of temporomandibular disorders (TMDs) and to rank them according to their effectiveness. An electronic search was undertaken to identify RCTs published until August 2019. Predictor variables were control, non-occluding splint, hard stabilization splint (HSS), soft stabilization splint (SSS), prefabricated splint, mini-anterior splint, anterior repositioning splint (ARS), and counselling therapy (CT) with or without HSS. Outcome variables were pain improvement, post-treatment pain intensity, improvement in mouth opening, and disappearance of temporomandibular joint (TMJ) sounds. Forty-eight RCTs were included. There was a significant decrease in post-treatment pain intensity in arthrogenous TMDs after ARS (low quality evidence), CT + HSS (moderate quality evidence), mini-anterior splints (very low quality evidence), and HSS alone (low quality evidence), when compared to the control. There was a significant decrease in post-treatment pain intensity in myogenous TMDs with mini-anterior splints (very low quality evidence), SSS (very low quality evidence), CT alone (moderate quality evidence), CT + HSS (moderate quality evidence), and HSS alone (moderate quality evidence), when compared to control. ARS and CT were superior in decreasing TMJ clicking than control and HSS alone. The three highest-ranked treatments for post-treatment pain reduction in arthrogenous TMDs were ARS (92%, very low quality evidence), CT + HSS (67.3%, low quality evidence), and HSS alone (52.9%, moderate quality evidence). For myogenous TMDs, they were mini-anterior splints (86.8%, low quality evidence), CT + HSS (61.2%, very low quality evidence), and HSS alone (59.7%, moderate quality evidence). Based on this NMA of 48 RCTs, there is moderate to very low quality evidence confirming the effectiveness of occlusal splint therapy in the treatment of TMDs. Multimodal therapy consisting of CT + HSS may produce the maximum improvement for TMD patients.

Temporomandibular disorders (TMDs) are classified into three categories based on their origin: myogenous, arthrogenous, and mixed . Several predictable treatment modalities for TMDs of any origin have been documented, such as occlusal splints, counselling therapy, physiotherapy, oral or injectable pharmacotherapy, and arthrocentesis or arthroscopy .

A variety of occlusal splints for the treatment of TMDs have been reported in the literature. The most widely used splints are stabilization splints (Tanner appliance, Fox appliance, Michigan splint, or centric relation appliance), anterior repositioning splint, and anterior bite splint.

Several published systematic reviews have shown the efficacy of occlusal splints in the treatment of TMDs . However, none of these systematic reviews has specifically covered randomized controlled clinical trials (RCTs) comparing the effectiveness of different occlusal splints versus control, non-occluding splints, or any of the other treatment modalities for myogenous, arthrogenous, or mixed TMDs. Furthermore, the effect of splint-wearing time and the total duration of occlusal splint therapy on the outcome of the treatment has not been investigated using a meta-analysis of RCTs with the GRADE system to rate the confidence of the evidence.

There are currently no published RCTs comparing the following different occlusal splints with or without counselling therapy and self-management in the management of TMDs: (1) full hard stabilization splint alone versus counselling in combination with a hard stabilization splint for patients with myogenous and arthrogenous TMDs; (2) non-occluding splints versus control, counselling with or without a hard stabilization splint, anterior repositioning splint, prefabricated splint, or nociceptive trigeminal inhibition tension suppression system (NTI-tss) in patients with mainly arthrogenous TMDs; (3) anterior repositioning splint versus full soft stabilization splint or a combination of full hard stabilization splint with counselling in patients with mainly arthrogenous TMDs; (4) mini-anterior splints such as the NTI-tss or midline anterior stop device versus the prefabricated splint or counselling plus full hard stabilization splint for patients with myogenous TMDs; (5) full soft stabilization splint alone versus NTI-tss, prefabricated splint, counselling with and without full hard stabilization splint, or full hard stabilization splint alone in patients with mainly myogenous TMDs. Therefore, a network meta-analysis (NMA) of RCTs was conducted to make comparisons among the different occlusal splints and counselling with and without a hard stabilization splint in order to rank the ideal and most effective occlusal splint in reducing signs and symptoms of TMDs.

The following hypotheses were considered in this analysis: (1) There would be no difference between a flat hard stabilization splint and other occlusal splints or counselling with or without a hard stabilization splint in the treatment of the signs and symptoms of TMDs. (2) Only patients with myogenous TMDs would significantly benefit from a hard stabilization splint compared to patients with arthrogenous or mixed TMDs. (3) There is no relationship between the duration of hard stabilization splint wearing and the type of TMD based on origin or the duration of follow-up with regard to splint efficacy.

The specific aims of this study were (1) to compare and rank the full hard stabilization splint, full soft stabilization splint, non-occluding splint, mini-anterior splint, prefabricated splint, anterior repositioning splint, and counselling therapy with and without a hard stabilization splint in the management of myogenous, arthrogenous, and mixed TMDs, with respect to pain reduction, mouth opening, and temporomandibular joint (TMJ) clicking; (2) to identify the effect of splint wearing time on its efficacy in pain reduction for patients with TMDs; (3) to assess the association between the duration of follow-up and the effectiveness of the hard stabilization splint.

Materials and methods

Protocol and registration

This study was accomplished according to the PRISMA extension statement for NMA ( Supplementary Material File 1). This study was registered in the PROSPERO database (CRD42018109352) .

Focused question

The following clinical research questions were established: (1) Does occlusal splint therapy treat TMDs? (2) What is the most effective oral occlusal splint for reducing pain intensity and TMJ clicking, and improving mouth opening for patients with arthrogenous and myogenous TMDs? (3) Does the pattern of hard stabilization splint wearing time have an impact on its efficacy in the treatment of TMDs?

Search strategy

All pertinent articles published between 1977 and March 2019 were identified through an electronic search of three major databases using the PICOTS criteria ( Supplementary Material File 2).

Inclusion criteria

The PICOTS criteria were applied, as outlined below.

‘P’ (population): adult patients with pain due to myogenous, arthrogenous, or mixed TMDs (Ia and Ib). The diagnosis had to be based on the Research Diagnostic Criteria for Temporomandibular Disorders (RDC/TMD) protocol, or a clear clinical diagnosis including signs and symptoms of TMDs.

‘I’ (intervention): different treatments for TMDs affecting the muscle, joint, or both using one of the following treatment modalities: (1) anterior repositioning splint (including maxillary or mandibular full coverage occlusal splints with an anterior ramp); (2) partial coverage splint such as an anterior midline stop device or the NTI-tss (including prefabricated or custom-made hard splints covering the two maxillary or mandibular central incisors); (3) prefabricated splint (including those covering the edges of the incisors and canines and with a palatal extension of approximately 1 cm); (4) non-occluding splint (including passive non-occluding splints); (5) full-coverage soft or resilience stabilization splint (including full maxillary or mandibular coverage soft stabilization splints); (6) control/no treatment (including patients who did not receive any treatment or those on a waiting list for treatment); (7) counselling therapy and self-management according to the definitions of the behaviour change technique taxonomy (version 1) (including basic elements of cognitive-behavioural therapy such as education, relaxation techniques, home physiotherapy (muscles exercises and joint mobilization), and avoidance of parafunctional habits); (8) counselling therapy plus hard stabilization splint (including any form of counselling therapy plus a full hard stabilization splint).

‘C’ (comparator): flat stabilization splint (including full hard maxillary or mandibular stabilization splints such as the Tanner appliance, Fox appliance, Michigan splint, or centric relation appliance).

‘O’ (outcomes): the primary outcome was pain intensity according to self-reported data (dichotomous data) or assessed clinically (continuous data) via visual analogue scale (VAS), numerical pain scale, or pain severity scale. Secondary outcomes were masticatory muscle tenderness (via algometer) and maximum mouth opening (MMO) in millimetres.

‘T’ (time): all included studies had to have followed up the patients for at least 1 month after treatment.

‘S’ (study design): RCTs comparing any occlusal splint to other treatments for patients with TMDs.

Exclusion criteria

The following studies were excluded: (1) non-randomized controlled clinical studies, (2) retrospective studies, (3) trials that did not investigate the outcomes of interest, (4) RCTs that did not report the required data as the mean and standard deviation values, as required to perform a meta-analysis, (5) RCTs that studied myofascial pain without clearly implicating the masticatory muscles.

Data extraction

Data were extracted separately by two researchers (R.F. and M.A.) using a specific form to summarize the following details: authors, study design, subgroup diagnosis, TMD diagnostic criteria used, age of the patients, male to female ratio, treatment groups (number), duration/frequency of treatment, outcomes investigated, and follow-up time.

Risk of bias

Two authors (R.F. and M.A.) investigated the risk of bias of included trials independently using the modified version of the Cochrane tool . Any dispute between the two authors was resolved by a third reviewer (E.A.).

Data synthesis

For continuous data, the post-treatment value was used to compute the standardized mean difference (SMD). For dichotomous data, the risk ratio (RR) was analysed using the number of patients reporting an improvement in TMJ pain and associated masticatory muscles at a post-treatment time. All NMAs were conducted using a frequentist framework via random-effects model in Stata Release 13, 2013 (StataCorp LLC, College Station, TX, USA) and the mvmeta command .

The loop-specific approach using the ifplot command in Stata and ‘design-by-treatment’ model using the mvmeta command were performed to evaluate the assumption of consistency at the local and global levels. Additionally, the authors assumed a common heterogeneity estimate within each loop . The ranking probabilities for all treatments at each possible rank for each group were analysed using the surface under the cumulative ranking (SUCRA) curve . SUCRA can also be presented as a percentage of treatment that can be ranked first without uncertainty. The rank-heat plot to visualize and present the treatment hierarchy across the multiple outcomes of interest was produced . Subgroup analyses were performed according to (1) the origin of the TMD (myogenous, arthrogenous, or mixed); (2) the duration of follow-up, either short-term (≤6 months) or intermediate-term (>6 months); (3) the wearing time of the splints, either at night only or 24 hours a day.

Certainty of the evidence

The GRADE (Grading of Recommendations Assessment, Development and Evaluations) approach to meta-analysis was used to assess the certainty of the NMA effect estimates for all outcomes of interest . In the GRADE system, RCTs begin as high quality evidence, but may be rated down due to limitations in the study design, inconsistency, imprecision, indirectness, and publication bias. The summary of confidence for the present evidence was estimated using the GRADEpro Guideline Development Tool .

Results

Outcome of the literature search

Of a total of 600 reports identified in all databases and two additional articles retrieved through the manual search, only 48 RCTs met the inclusion criteria and were included in the NMA . Fig. 1 illustrates the process of article evaluation for inclusion in the systematic review and meta-analysis.

Fig. 1
PRISMA flow diagram.

Presentation and summary of network geometry

With regard to the dichotomous data, 16 RCTs assessed the improvement in pain in 744 patients who received nine different treatments for TMDs of mainly arthrogenous origin (10 RCTs on arthrogenous TMDs; 6 RCTs on mixed TMDs) and 19 RCTs measured the improvement in pain in 946 patients who received eight different treatments for TMDs of mainly myogenous origin (13 RCTs on myogenous TMDs; 6 RCTs on mixed TMDs).

Regarding the continuous data, 16 RCTs evaluated post-treatment pain intensity in 929 patients who received eight different treatments for TMDs of mainly arthrogenous origin (9 RCTs on arthrogenous TMDs; 7 RCTs on mixed TMDs) and 18 RCTs measured post-treatment pain intensity in 1129 patients who received seven different treatments for TMDs of mainly myogenous origin (11 RCTs on myogenous TMDs; 7 RCTs on mixed TMDs) ( Fig. 2 ).

Fig. 2
Network geometry for the outcome of pain improvement (dichotomous data) and post-treatment pain intensity (continuous data) for arthrogenous and myogenous temporomandibular disorders. (Abbreviations: TMD, temporomandibular disorder; SSS, soft stabilization splint; HSS, hard stabilization splint; CT&SM, counselling therapy and self-management; NTI-tss, nociceptive trigeminal inhibition tension suppression system; ARS, anterior repositioning splint.)

Features of included trials

A full description of the trials, patient age and sex distribution, and how the treatments were conducted in all groups is given in Supplementary Material File 3.

Risk of bias

Twenty-five RCTs had an unclear risk of bias , , , , 13 RCTs had a low risk of bias , , , , and 10 RCTs had a high risk of bias , . Allocation concealment was adequate in 27 RCTs , , , . Assessment of outcome assessors showed that 24 RCTs were assessed by a blinded assessor , , , four RCTs were not blinded , and 20 RCTs did not report any information about blinding of assessors , , , . Nine RCTs had an attrition bias , , , 17 RCTs had no attrition bias , , , , , and 22 RCTs did not report any information about dropouts , , , , ( Supplementary Material File 4).

Results of individual studies

Online Supplementary Material File 5 summarizes the details of the outcomes. For dichotomous data, the number of patients reporting an improvement in pain and TMJ clicking and the total number of patients are stated. For continuous data, the mean, standard deviation, and sample size for the outcome of pain intensity and MMO are reported.

Synthesis of results—Results of the outcome variables

Pain improvement (dichotomous data): number of patients reporting pain improvement in RCTs including patients with TMDs of mainly arthrogenous origin (risk ratio)

Sixteen RCTs evaluated changes in pain reduction in 744 patients who received different treatments for TMDs of mainly arthrogenous origin (10 RCTs on arthrogenous TMDs; 6 RCTs on mixed TMDs) , , , , , , , , , , , , , , , . The follow-up time ranged from 1 to 12 months. There was no significant difference in pain improvement between the hard stabilization splint and the other treatments ( Fig. 3 ).

Fig. 3
Network meta-analysis forest plot for pain improvement (dichotomous data): the number of patients reporting pain improvement in RCTs including patients with TMDs of mainly arthrogenous origin. (Abbreviations: RCT, randomized controlled trial; TMD, temporomandibular disorder; RR, risk ratio; CI, confidence interval; PrI, prediction interval; NTI-tss, nociceptive trigeminal inhibition tension suppression system.)

Pain improvement (dichotomous data): number of patients reporting pain improvement in RCTs including patients with TMDs of mainly myogenous origin (risk ratio)

Nineteen RCTs assessed pain intensity in 946 patients who received different treatments for TMDs of mainly myogenous origin (13 RCTs on myogenous TMDs; 6 RCTs on mixed TMDs) , , , , . The follow-up time ranged from 1 month to 12 months.

There was a significant pain reduction after hard stabilization splint compared to control (RR 0.46, 95% confidence interval (CI) 0.26–0.80; low quality evidence) and non-occluding splints (RR 0.58, 95% CI 0.41–0.83; moderate quality evidence) ( Fig. 4 ).

Fig. 4
Network meta-analysis forest plot for pain improvement (dichotomous data): the number of patients reporting pain improvement in RCTs including patients with TMDs of mainly myogenous origin. (Abbreviations: RCT, randomized controlled trial; TMD, temporomandibular disorder; RR, risk ratio; CI, confidence interval; PrI, prediction interval; NTI-tss, nociceptive trigeminal inhibition tension suppression system.)

A significant pain reduction was noted with the use of hard stabilization splints (RR 2.18, 95% CI 1.25–3.8; moderate quality evidence), prefabricated splints (RR 2.24, 95% CI 1.12–4.49; very low quality evidence), and NTI-tss splints (RR 0.2.41, 95% CI 1.26–4.60; low quality evidence) when compared to control.

Post-treatment pain intensity (continuous data): RCTs including patients with TMDs of mainly arthrogenous origin (SMD)

Sixteen RCTs evaluated pain intensity in 929 patients who received different treatments for TMDs of mainly arthrogenous origin (9 RCTs on arthrogenous TMDs; 7 RCTs on mixed TMDs) , , , . The follow-up time ranged from to 1 to 12 months post-treatment.

There was a significant decrease in post-treatment pain intensity following hard stabilization splint (SMD −0.74, 95% CI −1.38 to −0.11; low quality evidence), anterior repositioning splint (SMD −1.18, 95% CI −1.90 to −0.47; low quality evidence), counselling therapy plus hard stabilization splint (SMD −0.78, 95% CI −1.42 to −0.14; moderate quality evidence), and mini-anterior splints (SMD −1.02, 95% CI −1.87 to −0.17; very low quality evidence) when compared to control ( Fig. 5 ).

Fig. 5
Network meta-analysis net league for post-treatment pain intensity (continuous data); RCTs including patients with TMDs of mainly arthrogenous origin (SMD, standardized mean difference).
An SMD of less than 0 favours the treatments in the column; an SMD of more than 0 favours the treatments in the row. Numbers in bold represent statistically significant results. Comparisons between treatments should be read from left to right and the estimate is in the cell in common between the column-defining treatment and the row-defining treatment. (Abbreviations: S, stabilization; CT, counselling therapy; R, repositioning.)

Post-treatment pain intensity (continuous data): RCTs including patients with TMDs of mainly myogenous origin (SMD)

Eighteen RCTs measured pain intensity in 1129 patients who received different treatments for TMDs of mainly myogenous origin (11 RCTs on myogenous TMDs; 7 RCTs on mixed TMDs) , , , , . The follow-up time ranged from to 1 to 12 months.

There was a significant decrease in post-treatment pain intensity following hard stabilization splint (SMD −1.25, 95% CI −1.69 to −0.80; moderate quality evidence), NTI-tss (SMD −1.49, 95% CI −2.19 to −0.79; very low quality evidence), soft stabilization splint (SMD −1.23, 95% CI −1.86 to −0.61; very low quality evidence), counselling therapy (SMD −1.04, 95% CI −1.55 to −0.52; moderate quality evidence), and counselling therapy plus hard stabilization splint (SMD −1.18, 95% CI −1.72 to −0.64; moderate quality evidence) when compared to control ( Fig. 6 ).

Fig. 6
Network meta-analysis net league for post-treatment pain intensity (continuous data); RCTs including patients with TMDs of mainly myogenous origin (SMD, standardized mean difference).
An SMD of less than 0 favours the treatments in the column; an SMD of more than 0 favours the treatments in the row. Numbers in bold represent statistically significant results. Comparisons between treatments should be read from left to right and the estimate is in the cell in common between the column-defining treatment and the row-defining treatment. (Abbreviations: S, stabilization; CT, counselling therapy.)

Post-treatment MMO: RCTs including patients with arthrogenous and myogenous TMDs (SMD)

Twelve RCTs measured the improvement in MMO in 491 patients who received different treatments for TMDs of arthrogenous and myogenous origin (8 RCTs on arthrogenous TMDs; 2 RCTs on myogenous TMDs; 2 RCTs on mixed TMDs) , , . The predictor variables were control group, counselling therapy, counselling therapy plus hard stabilization splint, and partial coverage splints such as NTI-tss and prefabricated splints. The follow-up time ranged from 1 to 12 months. There was no significant difference for any comparison ( Fig. 7 ).

Fig. 7
Network meta-analysis forest plot for post-treatment MMO (continuous data): RCTs including patients with TMDs of mainly mixed origin. (Abbreviations: MMO, maximum mouth opening; RCT, randomized controlled trial; TMD, temporomandibular disorder; SMD, standardized mean difference; CI, confidence interval; PrI, prediction interval; NTI-tss, nociceptive trigeminal inhibition tension suppression system.)

Improvement in TMJ clicking: number of patients reporting the disappearance of TMJ clicking in RCTs including patients with mainly arthrogenous TMDs (risk ratio)

Thirteen RCTs recorded improvements in TMJ clicking in 789 patients who received treatment for TMDs of mainly arthrogenous origin (2 RCTs on myogenous TMDs; 2 RCTs on mixed TMDs; 9 RCTs on arthrogenous TMDs) , , , .

There was significantly decreased TMJ clicking after anterior repositioning splint (RR 4.19, 95% CI 2.06–8.50; moderate quality evidence) and counselling therapy (RR 3.10, 95% CI 1.23–7.87; very low quality evidence) when compared to hard stabilization splint ( Fig. 8 ).

Fig. 8
Network meta-analysis forest plot for improvement in TMJ clicking: the number of patients reporting the disappearance of TMJ clicking in RCTs including patients with mainly arthrogenous TMDs. (Abbreviations: TMJ, temporomandibular joint; RCT, randomized controlled trial; TMD, temporomandibular disorder; CI, confidence interval; PrI, prediction interval; NTI-tss, nociceptive trigeminal inhibition tension suppression system.)

There was a significant decrease in TMJ sounds following anterior repositioning splint (RR 4.54, 95% CI 2.17–9.50; moderate quality evidence) and counselling therapy (RR 3.37, 95% CI 1.33–8.52; moderate quality evidence) versus control.

Synthesis of results—Treatment ranking

Pain reduction for patients with TMDs of mainly arthrogenous origin

From the dichotomous data, the most effective treatment to reduce pain in patients with TMDs of arthrogenous origin at follow-up ranging from 1 to 12 months was anterior repositioning splint (86.5%, very low quality evidence), followed by counselling therapy and self-management plus hard stabilization splint (75.6%, very low quality evidence), NTI-tss (58%, very low quality evidence), soft stabilization splint (56.3%, very low quality evidence), prefabricated splint (53.3%, very low quality evidence), hard stabilization splint (47.5%, moderate quality evidence), counselling therapy and self-management (40.8%, low quality evidence), non-occluding splints (32.4%, low quality evidence), and control (17.5%, very low quality evidence) ( Fig. 9 and Supplementary Material File 6).

Aug 10, 2020 | Posted by in Oral and Maxillofacial Surgery | Comments Off on Effectiveness of occlusal splint therapy in the management of temporomandibular disorders: network meta-analysis of randomized controlled trials

VIDEdental - Online dental courses

Get VIDEdental app for watching clinical videos