Results
The study selection process is described in the flowchart in Figure 1. Initial screening retrieved 44 studies in MEDLINE|PubMed, 82 in Web of Science, 10 in Cochrane Library, 38 in Embase, 150 in Scopus and two in LILACS|VHL. Searching other sources and manual-searching did not add any studies to this overview. After excluding duplicate records, 314 articles were analyzed in the title and abstract reading. Seven articles were selected to read the full text and apply the exclusion criteria. Of these, a systematic review was excluded as it included patients with disc displacement with reduction in the analysis. The level of agreement between reviewers was almost perfect (κ ≥ 0.81) for all databases consulted.
Six articles were included in this overview, five systematic reviews with meta-analyses7-10,12 and a qualitative review.11 Two articles included only randomized controlled trials (RCTs),9,11 two articles included RCTs and quasi-randomized clinical trials (q-RCTs)8-10,12 and one article included RCTs, q-RCTs and case series in the analyses.7 Only three articles described the male: female proportion of the sample, with females being predominant in the three articles.8,9,11 As for the diagnosis of DDwoR, different criteria were reported in the articles: Wilkes staging,8,10 American Association of Orofacial Pain (AAOP),10 Research diagnostic criteria for temporomandibular disorders (RDC/TMD) 10, Diagnostic criteria for temporomandibular disorders (DC/TMD),11,12 and Magnetic Resonance Imaging (MRI).9,11 One study did not describe the diagnostic criteria for participants.7 No secondary studies described the duration of TMD or whether previous treatment for this condition was performed (Table 2).
This overview showed that systematic reviews have confirmed methodological heterogeneity in the clinical studies. Furthermore, the meta-analyses showed statistical heterogeneity.7-10,12In general, the studies investigated the following comparisons: comparison between different non-invasive modalities;8-10 non-invasive versus minimally invasive modalities;10,12 non-invasive versus invasive modalities;8-10 minimally invasive versus invasive modalities;7,10 and comparison between invasive modalities.7-10 Outcomes reported in secondary studies were: improvement assessed by the patient7; reduction of pain8; frequency, intensity and duration of pain events9; symptom remission9; mandibular function; TMJ pain intensity10-12; and maximum mouth opening (MMO)7,8,10-12(Table 2).
Follow-up time for patients varied between studies between 10 minutes (only one primary study) and 60 months. Most of the articles included assessed the risk of bias within studies using the Cochrane’s tool; none of the primary studies had a low risk of bias in all domains.8-10,12 One article used the PEDro scale and noted that only 5:10 primary studies showed high methodological quality with scores ranging from 7 to 8 points.11 One article did not report the risk of bias analysis of the included studies (Table 2).7
Tables 3 and 4 (Table 3 and 4) present descriptive data and results of pairwise comparisons of different treatment modalities for DDwoR reported by secondary studies included in this overview.
Comparisons such as “rehabilitation” vs “medical management”,8 “physical medicine” vs “palliative care”,8 “physical medicine” vs “controls”,8-10 “palliative care” vs “physiotherapy” vs “no treatment”,9 revealed the absence of significant difference in the improvement of pain8-10 and/or mandibular function.8 Comparing “mandibular manipulation” vs “no treatment”, pain improvement and MMO were significantly greater in patients who received mandibular manipulation.10 “Jaw exercises” vs “education” showed no significant difference in pain and MMO improvement.10 Comparing “self-management” (self-exercises + self care/medication) vs “no treatment”, “self-care” represented a significant improvement in pain and MMO compared to “no treatment”.10
Comparing ”self-care” (self-care + self-care/non-steroidal anti-inflammatory drugs) vs ”splint”: although there was a greater reduction in pain intensity in the ”self-care” group in the short term, the difference was not statistically significant.10For MMO, ”self-care” was more effective than ”splint”.10 Comparing ”immobilization” vs ”information and pain medication as needed”, pain reduction was less frequent in the ”immobilization” group in the long-term.10
The number of patients with ≥ 50 % pain reduction was significantly higher in the ”splint” group compared to the ”transcutaneous electrical nerve stimulation (TENS)” group; however, there was no statistically significant difference between both interventions for MMO.10
”Combination therapy” (splint plus exercise + self-care/medication/education) compared to ”Education” (only) or ”self-care” (self-care/medication/education ± self-care) showed no significant difference in pain improvement and MMO.10”Manual therapy and exercise interventions” vs ”splints” showed no significant difference for pain improvement; however, the improvement in MMO was significantly greater for ”manual therapy and exercise interventions”.11
There was no statistically significant difference in the effects of ”combination of splint plus jaw exercises” compared to ”splint” for short-term pain. For MMO, however, there was a statistically significant difference in favor of combined treatment.10
Comparing ”active pulsed electromagnetic fields (PEMF)” vs ”placebo”, there was no significant difference in pain improvement and MMO. ”Active iontophoresis” (by dexamethasone + lidocaine) showed short-term improvement in pain and MMO compared with the ”placebo” group, but the differences were not statistically significant.10
In the ”arthrocentesis” vs ”arthrography only” and ”arthrocentesis” vs ”auriculotemporal nerve block” comparisons, there was no significant difference in the effects for pain and MMO. ”Arthrocentesis” showed significant short-term pain improvement compared to ”combination therapy” (splint plus self-care/self-exercise); there was no significant difference between interventions for MMO10. Another review observed that arthrocentesis showed a significant difference in terms of increased mouth opening and reduced pain level in relation to the stabilization splint12.
In the “rehabilitation” vs “arthroscopic surgeryand “rehabilitation” vs “arthroplasty” comparisons, there was no significant difference in the effects for pain and MMO.8 There was no significant difference in pain control comparing ”arthroscopic surgery” vs ”self-management” (self-care/medication/education), ”arthroscopic surgery” vs ”combination of splint plus exercises” [+ self-care/medication/education + cognitive behavioral therapy (CBT)] and ”open surgery” vs ”self-management” (self-care/medication/education). Comparing ”open surgery” vs ”combination of splint plus exercises” (+ self-care/medication/education + CBT), ”open surgery” demonstrated a significant improvement in pain control in the short term, but not in the long term.10
There was no “improvement” assessed by the patient,7pain reduction and MMO10 improvement when comparing ”arthrocentesis” vs ”arthroscopy”. Comparing ”open surgery” vs ”arthroscopy”, one of the secondary studies did not report a significant difference for the ”improvement assessed by the patient” (meta-analysis through indirect comparisons).7 Meta-analyses from the other study showed an overall improvement in open surgery and a significant reduction in long-term pain intensity. Sensitivity analysis excluding a primary study without imaging diagnosis showed no significant difference between surgical procedures.10
Figures 2 and 3 show the results of the methodological analysis of the articles included, according to AMSTAR 2. These data suggest methodological weaknesses in the studies, added to the limitations of the primary studies related to heterogeneity. In the overall confidence rating, all articles included had a critically low confidence level. According to the OQAQ methodological quality analysis (figure 4), only three articles had a positive response for all domains.10-12 In the quality analysis using the CASP, only two articles showed a positive response for all domains.10,12 Regarding the assessment of the risk of bias of the systematic reviews, after answering the focused questions (Figure 5), none study presented a low risk of bias in all domains of the instrument (Figure 6).