Static body postural misalignment in individuals with temporomandibular disorders: a systematic review

BACKGROUND: The association between body postural changes and temporomandibular disorders (TMD) has been widely discussed in the literature, however, there is little evidence to support this association. OBJECTIVES: The aim of the present study was to conduct a systematic review to assess the evidence concerning the association between static body postural misalignment and TMD. METHOD: A search was conducted in the PubMed/Medline, Embase, Lilacs, Scielo, Cochrane, and Scopus databases including studies published in English between 1950 and March 2012. Cross-sectional, cohort, case control, and survey studies that assessed body posture in TMD patients were selected. Two reviewers performed each step independently. A methodological checklist was used to evaluate the quality of the selected articles. RESULTS: Twenty studies were analyzed for their methodological quality. Only one study was classified as a moderate quality study and two were classified as strong quality studies. Among all studies considered, only 12 included craniocervical postural assessment, 2 included assessment of craniocervical and shoulder postures,, and 6 included global assessment of body posture. CONCLUSION: There is strong evidence of craniocervical postural changes in myogenous TMD, moderate evidence of cervical postural misalignment in arthrogenous TMD, and no evidence of absence of craniocervical postural misalignment in mixed TMD patients or of global body postural misalignment in patients with TMD. It is important to note the poor methodological quality of the studies, particularly those regarding global body postural misalignment in TMD patients.


Introduction
Temporomandibular Disorder (TMD) is a set of disorders characterized by signs and symptoms involving the temporomadibular joints and mastication muscles, as well as related structures 1 . There is evidence that its etiology is multifactorial and include psychological, biomechanical, and neurophysiological factors [2][3][4] .
Craniocervical posture is only one of the body segments that must be considered for postural assessment, specifically because adjacent postural compensations are expected in other segments considering that muscle chains are interconnected 23,24 .
Three systematic reviews regarding the theme were found in the literature 20,25,26 , however, the reviews by Olivo et al. 20 and Rocha et al. 26 only considered studies related to craniocervical posture and TMD, and the review by Perinetti and Contardo 25 did not include studies on craniocervical posture. Moreover, this review 25 classified, in the same list, 1 studies regarding stabilometry (i.e. postural balance assessment) and static posture. Therefore, there was no systematic review available in the present literature involving body postural alterations (either segmentary or global) in individuals with TMD. Given the great interest in the theme and the poor methodological quality of the studies about body postural misalignment and the postural assessment methods employed in these studies 20,25 , it was important to carry out a study that analyzed real evidence of associations between static postural changes and TMD in order to guide better controlled studies in the future.
The confirmation of the evidence of the association between craniocervical or body postural misalignment and TMD may help to determine the predisposing and/or perpetuating factors in the development of TMD and guide new and well designed research to confirm this association. Moreover, some studies have demonstrated the relief of TMD symptoms after treatment involving postural reeducation 27,28 .
It was expected that the findings of this systematic review would demonstrate whether the evidence available was sufficient to indicate an association between body postural misalignment and TMD and/ or subtypes. Thus, the aim of this study was to review the literature available on the main databases (i.e. PubMed/Medline, Embase, Lilacs, Scielo, Cochrane, and Scopus) about body postural misalignment in patients with TMD and subtypes.

Method Data sources
In order to find studies examining the relationship between static body posture and TMD, bibliographical surveys were performed in the following databases: PubMed/Medline, Embase, Lilacs, Scielo, Cochrane, and Scopus. PRISMA 29 (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines were followed.
The search comprised only studies in English Searches were performed by the same researcher. The limits of databases were selected when the option was available. In the Embase and Pubmed databases, the limits followed were: Published: 1966 to March 2012, quick limits: humans, only in English, article in press.

Eligibility criteria
Types of Studies. i) cohort/case-control studies; and ii) cross-sectional and survey studies. Publications such as case reports, case series, reviews, and opinion articles were excluded. As the main objective of this study was to verify the possible association between TMD and body postural changes, randomized controlled clinical trials were excluded, since these studies are used to verify the effectiveness of an intervention and, therefore, not adequate to verify relationships between variables.
Participants. Inclusion was restricted to studies using human participants who (i) were between 7 and 60 years of age; (ii) had been diagnosed with TMD; (iii) had not previously had TMJ surgery; (iv) had no history of trauma or fracture in the TMJ or craniomandibular system; and, (v) had no other serious comorbid conditions (e.g. cancer, rheumatic disease, neurological problems).
Types of Outcome Measures. The following methods of body postural assessment were considered: body landmarks, visual inspection, pictures or radiographs.

Data collection
The reviewers analyzed all studies initially selected by the title or abstract for the inclusion/ exclusion criteria. The published studies had to provide enough information to meet the inclusion criteria and not be eliminated by the exclusion criteria. In order for studies to be evaluated at the next level (critical appraisal), the study had to meet all of the inclusion criteria. When the reviewers disagreed on whether a study met a criterion, rating forms (form containing the Critical Appraisal completed by each reviewer - Table 1) were compared, and the criterion was discussed until a consensus was reached.
As recommended by PRISMA 29 , the studies were selected by the title, abstract, and full text. Two independent reviewers screened the abstracts of the publications found in the databases.

Quality evaluation
In order to document the internal and external validity of the studies, a modified quality evaluation instrument was applied 20,30 . This tool considered: 1-study design, 2-control of confounding variables, 3-subjects' agreement to participate, 4-sample size calculation, 5-validity/reliability of outcomes measurements, 6-blinding, 7-external validity, and 8 -statistical analysis (Table 1). Two independent reviewers evaluated the studies based on specific determined criteria. If there was inadequate information in the published papers to allow evaluation of the criteria, the authors of the studies were contacted to clarify study design and specific characteristics of the study. If the authors did not reply, the studies were evaluated with the information available.
Each evaluated study item was then given a grade of strong (S), moderate (M) or weak (W) in each category. The rating system was based on a similar procedure 20,31 . Critical appraisal was completed independently by the two reviewers, and their results were compared. Data were extracted from each article without blinding of the authors. Finally, every study was graded depending on the following criteria (Table 1): • STRONG -Strong for items: 2, 4, 5, 6, 7, and 8 or Moderate or Strong for items 1 and 3; • MODERATE -Moderate for the following items: 2, 4, 5, 6, 7, and 8 and Weak or Moderate for items 1 and 3; • WEAK -Weak for at least one of the items: 2, 4, 5, 6, 7, and 8. The results have clinical relevance 1 0 Patients are representative of the population / where screened / age / comorbidities / severity 1 0 Observed aspects were clarified in the conclusion and discussion

Statistical analysis
The kappa coefficient test was used to verify the agreement between both reviewers before the consensus stage in the analysis of studies. Results were obtained using the weighted kappa coefficient and analyzed using SPSS version 17, and the agreement was classified as follows: K<0.20 (poor), 0.21 to 0.40 (weak), 0.41 to 0.60 (moderate), 0.61 to 0.80 (good), 0.81 to 1.0 (excellent).

Results
The selection included 1067 studies (271 in Pubmed, 3 in Scielo, 703 in Scopus, 33 in Lilacs, and 57 in Embase) considering duplicates/triplicates. After the removal of duplicates among different databases, 393 studies remained. After comparison for the existence of duplicates in the same database, 348 studies remained. The studies were screened again by verifying the title, and only 36 studies were selected.
After analysis of the abstracts, all 20 studies were read once in full and five studies were excluded adopting the criteria previously defined. The studies were excluded because they consisted of: i) nonexperimental studies 46,47 ; ii) a study involving therapeutic intervention 27 ; iii) a study involving static postural assessment 48 ; and 4) a study with inappropriate sample eligibility criteria 49 .
The agreement between both reviewers for the final classification of the 20 studies obtained Interrater Kappa of 0.90 (Confidence Interval 95%: 0.73-1), demonstrating an excellent level of agreement between them.

Quality criteria score
Considering the criteria for assessment of methodological quality, only three studies were classified as moderate 51 or strong 19,21 .

Strengths and weaknesses
Lee et al. 50 -1995 The relationship between forward head posture and temporomandibular disorders.

Sample size, posture method assessment, and examiner blinding
Sample size was calculated in only three studies 19,21,49 (Table 2). Of the studies that analyzed only craniocervical posture, six studies described the use of assessment by radiographic analyses 10,13,15,18,49,51 , six studies used the photographic method 6,11,12,19,21,50 , and two described the use of the both radiographic and photographic methods 8,22 (Tables 3 and 4).
Only one of the studies included in this review mentioned the validity of the measures employed for postural assessment 22 , however the reference that certified the method validity was probably incorrect 54 . The authors did not answer the e-mail to clarify this possible error.
Of the six studies using global body posture, the standardization for posture analysis and analysis method was appropriately described in three studies 7,14,16 . The photogrammetry method was used by two studies 7,14 and a previously described method combining photographic and visual inspection was used in one study 16 (Table 4).
Among the six studies that assessed pelvic posture, four studies 7,9,16,17 verified pelvic misalignments in the frontal plane 7 , iliac crest 9 , muscle chain 16 , and posterior rotation 17 (Tables 3 and 4). Spinal misalignments were identified by two of the five studies that included this topic in the postural assessment 5,9,14,16,17 : greater thoracic kyphosis and lumbar hyperlordosis 9 and kyphosis straightening and lumbar hyperlordosis 17 (Table 4). However, of the studies that were classified as moderate or strong quality, Armijo-Olivo et al. 19,21 reported greater head extension and D'Áttilio et al. 51 observed cervical spine straightening.

Postural changes in TMD subtypes
Of the five studies that included a group of patients with myogenous TMD 5,8,19,21,22 , two found body posture misalignments (head extension) in the TMD group in relation to the control group or mixed TMD group 19,21 . Both studies were classified as strong according to the adopted quality criteria applied (Tables 3 and 4).
Concerning arthrogenous TMD, four studies verified body posture changes in the TMD group in relation to the control group or another TMD group 12,17,51,52 , and three did not report craniocervical postural changes 8,11,15 . Only the study by D'Áttilio et al. 51 was classified as moderate quality. The authors reported cervical spine straightening (Tables 3 and 4).
Among the studies that included a group of mixed TMD patients in relation to a control group or another TMD group [6][7][8][9][10]16,18,19,21,22,50 , seven reported body posture alterations 6,7,9,10,16,18,50 . Only two studies 19,21 were classified as strong quality and they did not report body posture alterations for the mixed TMD group (Tables 3 and 4), however in both studies this group had to have a diagnosis of myiogenous TMD according to the RDC/TMD but not a diagnosis of arthrogenous TMD according to these criteria, only signs and symptoms.

Discussion
The purpose of this systematic review was to identify the level of scientific evidence for the association between TMD and body and/or craniocervical posture misalignment. The quality criteria adopted for review of the studies have been described in previous studies 20 and the agreement between the reviewers for the methodological classification of the studies was high (kappa: 0.91), demonstrating that the review process was considered reliable.
This systematic review considered global body posture misalignment. Regarding the three systematic reviews on the subject, two of them considered craniocervical posture only 20,26 and the other presented records of static posture that were analyzed together with records of balance -static posturography 25 . Moreover, these authors 25 disregard studies about craniocervical posture. Postural assessments aimed at finding postural deviations are routinely made by physical therapists to analyze body segments in the static position and do not include the assessment of oscillations that must be considered as balance assessment.
A significant number of the studies found in the literature and included in this review (n=14) considered only the assessment of the head segment 6,8,[10][11][12][13]15,18,19,21,22,[50][51][52] . This aspect is probably related to the fact that it is easier to perform the procedure in the craniocervical segment, since the individual does not need to be evaluated in bathing clothes, and moreover because the radiographic procedure commonly employed in dentistry only considers the head and cervical spine, and it does not enable the analysis of global body posture. On the other hand, this aspect disregards posture assessment as a whole and it is possible that head changes are related to distal changes, since the connection between the muscles through the muscular chains would facilitate the emergence of postural compensation in other body segments 23 .

Main findings and TMD subtypes
This review demonstrated that there is evidence for craniocervical postural change (i.e. head extension) in patients with myogenous TMD in relation to controls.
Of the five studies that included a group of patients with myogenous TMD 5,8,19,21,22 , two studies were classified as strong according to the quality criteria employed and verified only craniocervical posture changes in TMD in relation to a control group or a mixed TMD group 19,21 .
Considering body posture misalignment in arthrogenous TMD, only the study of D'Attilio et al. 51 was classified as moderate according to the criterion quality adopted. Therefore, it was observed that there was moderate evidence and risk of bias for the presence of cervical posture misalignment (i.e.cervical spine straightening) in patients with arthrogenous TMD, diagnosed by MRI, in relation to a control group.
Considering studies involving patients with mixed TMD, only two studies 19,21 obtained a strong classification according to the quality criteria adopted and they did not report body postural misalignment for the mixed TMD group. One of the reasons for the absence of evidence of body postural misalignment in mixed TMD patients compared to myogenous and arthrogenous patients could be related to the sample selection adopted 19,21 . The patients should have a diagnosis of myogenous TMD according to the RDC/TMD associated with signs and symptoms of arthrogenic TMD. In this way, all of the patients must have a diagnosis of myogenous TMD, but not of arthrogenous TMD. It is possible that the "mixed TMD group" could not fill the criteria for an arthrogenous TMD diagnosis, since signs and symptoms of arthrogenous complaints have commonly been observed in the population 55 . Hence, there is no evidence that patients with mixed TMD (i.e. with myogenous TMD diagnosis and signs and symptoms of arthrogenous TMD) did not have body or craniocervical misalignment in relation to individuals without TMD or myogenous TMD. D'Atillio et al. 51 demonstrated cervical spine straightening in arthrogenous TMD patients and received a moderate evidence level classification. However, D'Atillio et al. 51 used radiographic analysis to assess cervical spine misalignment and Armijo-Olivo et al. 19,21 verified only head and cervical/head posture. In this way, it is possible that in patients with arthrogenous TMD, cervical spine misalignment could be more common, and in patients with myogenous TMD disorders, head posture misalignment could be more common. It could explain the absence of body posture misalignment for mixed TMD group described by Armijo-Olivo et al. 19,21 .
However, all of these theories are speculative and the attention should focus on the need for future studies to include a large sample size, control the diagnostic criteria for mixed and arthrogenous groups, and consider not only photographic records but also radiographic procedures to analyze the cervical spine more specifically. Two studies assessed body posture by both photography and radiography 8,22 , however the major flaw of these papers was their limited sample size. Armijo-Olivo et al. 19 described a minimum of 50 subjects (α=0.05, β=0.20, power=80%, and effect size of 0.5) to assess posture by photographic records.
Global body postural misalignment in the group of TMD patients was verified in four studies 7,9,16,17 . All studies obtained a weak classification. Aspects such as absence of blinding of the examiner 7,17 , failure in sample eligibility criterion 9,16 , and poorly described or undescribed reliability of the method 5,16,17 were some of the characteristics that did not support the evidence of possible global body postural changes in arthrogenous, myogenous or mixed TMD groups in relation to a control group.
As contribution for future publications, the authors recommend effect size and power analysis, a more controlled design, appropriate description of reliability/validity of the measures (specifically for global body postural assessment), blinding of the examiners, random sampling, and, eligibility criteria of patients with control of subtypes of TMD according to well stablished criteria.

Conclusion
The main contributions of the present review are the following: there is evidence and low risk of bias that patients with myogenous TMD have craniocervical postural misalignment. For the arthrogenous TMD group, moderate evidence for cervical spine alterations was observed. Moreover, there was no evidence in the literature for the absence of craniocervical posture misalignment in mixed TMD patients and for global body posture misalignment in TMD. The poor methodological quality of the studies considered in this revision, especifically for body postural misalignment could be the explanation for the weak evidence observed.