Orthognathic Surgery as a

Tr e a t m e n t f o r
Te m p o rro
omandibular Disorders
M. Franklin Dolwick, DMD, PhDa,*,
Charles G. Widmer, DDS, MSb

KEYWORDS
 Orthognathic surgery  Temporomandibular disorders  Skeletal malocclusion

KEY POINTS
 The impact of orthognathic surgery on the signs and symptoms of temporomandibular disorders
(TMDs) has been unclear.
 Many studies have not evaluated single jaw surgeries; instead, TMD outcomes assessments were
the result of a mixture of osteotomies combined with preorthodontic and postorthodontic therapy.
 Most clinical studies on the effects of orthognathic surgery on TMD signs and symptoms did not
include a control group and, when included, most control groups were not matched on age and sex.
 The best evidence in the current literature supports the concept that orthognathic surgery does not
increase the overall frequency of TMD signs and symptoms at a follow-up of 2 years or more.
 However, correction of a retrognathic mandible with a counterclockwise rotation increased masti-
catory muscle myalgia and, combined with a 7 mm advancement, elicited an increase in myalgia
and TMJ arthralgia.

Orthognathic Surgery as a
Tr e a t m e n t f o r
INTRODUCTION
Temporomandibular disorders (TMDs) are muscu-
address skeletal malocclusions in patients with
TMDs.10 The rationale for pursuing an early surgi-

Te m p o r o m a n d i b u l a r D i s o r d e r s
loskeletal disorders involving the temporomandib-
ular joint (TMJ), masticatory muscles, or both.
cal approach has been the clinical impression of
success in reducing signs and symptoms of
TMDs while correcting the skeletal malocclusion.
Treatments for TMDs that have moderate evi- Clinical impressions of success are prone to
dence for pain reduction efficacy include pharma- different biases that can influence the perception
cologic therapies,1–3 physical medicine,4 of clinicians. For many years, these clinical impres-
behavioral therapies, 5,6 and occlusal appliance sions have driven the course of various treatments
therapies.7,8 These treatment approaches are for TMDs, with the conclusion that the specified
considered reversible compared with irreversible intervention was the cause for success in reducing
treatments, which include occlusal equilibration, or eliminating TMD pain and dysfunction.11,12 In the
mandibular repositioning, orthodontics, and case of orthognathic surgery, some surgeons
orthognathic surgery.9 In recent years, there have consider their treatment to be successful in perma-
been reports describing early irreversible interven- nently resolving both a skeletal abnormality as well
tions such as the bilateral sagittal split osteotomy as a musculoskeletal disorder, particularly one
oralmaxsurgery.theclinics.com

or the intraoral vertical ramus osteotomy to associated with pain.13 To address this question,

Disclosure: The authors have nothing to disclose.

a
Department of Oral and Maxillofacial Surgery, University of Florida College of Dentistry, PO Box 100416,
Gainesville, FL 32610-0416, USA; b Division of Facial Pain, Department of Orthodontics, University of Florida
College of Dentistry, PO Box 100444, Gainesville, FL 32610-0444, USA
* Corresponding author.
E-mail address: [email protected]

Oral Maxillofacial Surg Clin N Am - (2018) -–-

https://doi.org/10.1016/j.coms.2018.04.007
1042-3699/18/Ó 2018 Published by Elsevier Inc.
2 Dolwick & Widmer

a few publications have pursued a meta-analysis of Therefore, it is paramount to consider only the
clinical studies based on well-defined inclusion and best evidence available to determine the effects,
exclusion criteria.14,15 However, there were few if any, that orthognathic surgery may have on
well-controlled studies available in the literature different skeletal malocclusion types.
that addressed a specific skeletal malocclusion Retrospective studies using chart reviews also
and allowed an adequate assessment of the have been shown to be limited in the comprehen-
outcome of orthognathic surgery on TMD. sive review of TMD signs and symptoms compared
The purpose of this article is to provide an up- with prospective studies.49 Multiple signs and
date on the efficacy of orthognathic surgery as a symptoms are frequently not included in chart
treatment of TMD. This update was accomplished documentation and the lack of these data may
by assessing peer-reviewed, published studies of result in a low estimation of TMD prevalence.
orthognathic surgery procedures that were per- Thus, chart reviews should be considered a limited
formed in the absence and presence of TMD signs and potentially biased source of data to assess
and symptoms. This topic has been a focus by a signs and symptoms of TMD in a clinical sample.
few previous investigators who provided an In many studies, the patient samples contained
overview of their study design and target study a mixture of skeletal malocclusion types and the
populations. However, this review evaluates the proportion of subtypes was usually reported.
methodology that was used to minimize observer However, the TMD outcomes associated with a
bias (if any) and to provide an updated evidence- specific malocclusion were commonly not re-
based assessment. It targets 3 types of skeletal ported. Instead, the entire surgical sample was
malocclusions commonly addressed by orthog- evaluated for efficacy in reducing (or increasing)
nathic surgery: class II, class III, and anterior TMD signs and symptoms. In a few studies, there
open bite. In addition, the surgical approach is was a focus on a single malocclusion (class II,
considered as an independent variable when the class III, or anterior open bite) rather than a group
results for each type of surgery are reported in of malocclusion types and this allowed a better
the publication. Individual case studies (n 5 5 or assessment of the effects of the surgery on masti-
less) and meta-analyses were not included in the catory musculoskeletal pain and dysfunction.
literature that was reviewed. These studies are discussed in greater detail later.
A few studies focused on the technique of
EXPERIMENTAL DESIGN OF CLINICAL STUDIES mandibular stabilization after surgery and also
evaluated TMD signs and symptoms.28,45 One
Clinical studies that evaluated TMD signs and well-controlled study that randomized patients
symptoms after orthognathic surgery were evalu- into nonrigid or rigid fixation of the mandible after
ated using the criteria listed in the headers of a bilateral sagittal split osteotomy (BSSO) found
Table 1. These criteria were identified as important no statistically significant difference in TMD signs
to document the quality of the experimental design and symptoms after a 2-year follow-up.28 This
and appropriateness of the statistical analyses study was well designed, with randomization of a
used in the study. Fulfillment of these criteria relatively homogenous sample of orthognathic pa-
generally ranked the study into an upper level tients into 2 mandibular fixation groups after the
and was viewed as contributory to the evidence patients were screened based on specific inclu-
that did or did not show efficacy of the orthog- sion and exclusion criteria. In addition, this study
nathic surgery as a treatment of TMDs. used well-defined and validated TMD assay mea-
Many of the reviewed studies were retrospective sures and calibrated examiners. The only limitation
and commonly were consecutive cases seen at an with this study, and this was acknowledged by the
educational institution. Retrospective studies, investigators, was the inability to minimize bias by
although convenient, have some advantages but blinding the examiners to the type of mandibular
also many limitations in providing an accurate fixation. The study had sufficient statistical power
assessment of treatment efficacy (Table 2). to determine differences between the two groups,
Assessing the unbiased, accurate pain outcome and none was found, so it is doubtful that the lack
measures of TMDs in orthognathic surgery pa- of blinded examiners had an impact on the
tients to determine the effect of this treatment re- study outcome measures, particularly for TMD
quires special attention to the experimental assessment.
design. It has been elegantly shown in 1 well- Random assignment of a homogenous group of
controlled clinical trial on migraine headaches orthognathic surgery patients into an experimental
that pain can be influenced by multiple factors, and a control group is an optimal experimental
including patient expectation,48 a factor that would design but is not feasible in a clinical study of patients
be associated with any treatment approach. who are having orthognathic surgery to correct their
Table 1
Class II, class III, and anterior open-bite skeletal malocclusions: temporomandibular disorder signs and symptoms before and after orthognathic surgery

Well-Defined
Calibrated/ Inclusion/
No. of Surgical Approach Randomized Blinded Exclusion Postsurgical
Patients (% Combined Study/Control Examiners Criteria Follow-up
Reference Type of Study (F:M) Ortho/Surgery) Group (Yes/No) (Yes/No) (Yes/No) TMD Assessment Duration Comment
Upton et al, 16 Retrospective Class II: 46 Ortho/surgery No/no No/no No Questionnaire Not reported Primary outcome
1984 case series (36:10) (75%) or variable: TMJ
Res. design Class III: 39 surgery only pain/dysfunction
assessment: 9 (22:17) Descriptive
Ant. open statistics reporte
bite: Inferential statist
14 (9:5) tests not calculat
for presurgical a
postsurgical outc
Karabouta & Retrospective Class II: 46 Surgery only: No/no No/no No Physical examination: 6 mo Primary outcome
Martis,17 1985 case series Class III: 161 BSSO TMJ pain, myofascial variable: TMJ
Res. design Ant. open pain, TMJ sounds, pain/dysfunction
assessment: 17 bite: 45 restricted jaw Descriptive statis
movement reported. Inferen
statistical tests no
calculated for
presurgical and
postsurgical
outcomes
Timmis et al,18 Prospective Class II: 25 Ortho/surgery: No/no No/no No Physical examination: NR: 6–36 mo Primary outcome
1986 cohort (12 R; BSSO TMJ sounds, R: 6–12 mo variable: TMJ
Res. design 13 NR) myofascial pain, pain/dysfunction
assessment: 20 Class III: 3 deviation with incidence in R
jaw opening and NR fixation.
No significant
difference
between
preoperative
and postoperativ
muscle or TMJ pa
incidence. Individ
skeletal malocclu
groups not statis
tested
(continued on n
Table 1
(continued )

Well-Defined
Calibrated/ Inclusion/
No. of Surgical Approach Randomized Blinded Exclusion Postsurgical
Patients (% Combined Study/Control Examiners Criteria Follow-up
Reference Type of Study (F:M) Ortho/Surgery) Group (Yes/No) (Yes/No) (Yes/No) TMD Assessment Duration Comment
Magnusson Retrospective Class II: 3 Ortho/surgery No/no No/no No Helkimo index Variable: Primary outcome var
et al,19 case series Class III: 12 (65%) or surgery 1–2.5 y pain/dysfunction.
1986 Ant. open only: BSSO, Descriptive statisti
Res. design bite: 1 segmental reported. Ai signif
assessment: 17 osteotomy of reduced postsurgi
mandible, compared with pr
Le Fort I, assessments. Indivi
or combination skeletal malocclusi
groups not statisti
tested
Kerstens et al, 20 Retrospective Class II: 338 Ortho/surgery No/no No/no No Questionnaire and Clinical Primary outcome va
1989 case series Class III: 142 (91%) or surgery physical examination examination: TMJ pain/dysfunc
Res. design only: BSSO (TMJ sounds, limited 1y Descriptive statist
assessment: 17  Le Fort I mandibular movement, Questionnaire: reported. No stat
osteotomy palpable muscle 1.4–4.7 y difference betwee
tenderness) preoperative and
postoperative TM
or symptoms or
questionnaire sym
Smith et al,21 Prospective Class II: 22 Ortho/surgery No/no No/no No Modified Helkimo 6–7 mo Primary outcome va
1992 cohort (14:8) (100%): index TMJ pain/dysfunc
Res. design BSSO (R) Descriptive statist
assessment: 21 reported. Inferen
statistical tests no
calculated for pre
and postsurgical o
Athanasiou & Retrospective Class III: 36 Ortho/surgery No/no No/Yes No Modified Helkimo 6 mo Primary outcome va
Melsen,22 1992 case series (25:11) (100%): index TMJ pain/dysfunc
Res. design IVRO Clinical examination incidence in R an
assessment: 24 NR: 18 (TMJ sounds, muscle NR fixation. Maxi
No fixation: 18 and joint palpation, interincisal openi
mandibular was significantly
movement, ROM) reduced by 5.4 m
Muscle pain frequ
and lateral excurs
movements did n
change. No statist
difference betwe
TMD frequency
in men and wome
De Clercq et al, 23 Retrospective Class II: 196 Surgery only: No/no No/no Yes Physical examination Minimum 6 mo Primary outcome
1995 case series (150:46) BSSO (R) (muscle tenderness, variable: TMJ
Res. design  Le Fort I joint sounds, pain, pain/dysfunction.
assessment: 15 osteotomy limitation of Descriptive statist
movement) reported. Statistic
decrease of TMD
from 27%
(preoperative)
to 18% (postoper
for all patients an
from 30%
(preoperative)
to 1.8% (postope
for the normal/lo
angle deficiency
group. No statisti
difference for TM
in the high angle
deficiency group
Feinerman & Retrospective 66 NR (21:11) Surgery only: No/no No/no No Questionnaire Variable: 2–10 y Primary outcome
24
Piecuch, 1995 case series R (21:13) BSSO  Le Fort I and physical variable: TMJ
Res. design Class II and osteotomy examination pain/dysfunction
assessment: 14 class III (34 R, 32 NR) (TMJ sounds, limited in R and NR fixati
patient jaw movement, Descriptive statist
distribution palpable muscle reported. Specific
not reported or joint pain) outcomes not rep
class II and class III
Statistical differen
observed betwee
NR fixation for pa
muscle tendernes
(increased in NR)
clicking (increased
Onizawa et al, 25 Retrospective 30 (20:10) Surgery only: No/yes No/no No Questionnaire 6 mo Statistically significa
1995 case series class II: 10 BSSO  Le Fort I Control group: and physical decreases in maxi
Res. design Class III: 17 osteotomy 30 dental examination mandibular open
assessment: 17 students (TMJ sounds, protrusion betwe
(11 F:19 M) limited jaw presurgical and
movement, postsurgical asses
palpable muscle for class II and cla
or joint pain) patients. No statis
significant differe
between patients
controls for TMJ s
mandibular devia
palpable muscle
tenderness
(continued on n
Table 1
(continued )

Well-Defined
Calibrated/ Inclusion/
No. of Surgical Approach Randomized Blinded Exclusion Postsurgical
Patients (% Combined Study/Control Examiners Criteria Follow-up
Reference Type of Study (F:M) Ortho/Surgery) Group (Yes/No) (Yes/No) (Yes/No) TMD Assessment Duration Comment
Rodrigues-Garcia Prospective Class II: 124 Ortho/surgery Yes/no Yes/no Yes Craniomandibular 2y MI had small, statis
et al, 26 1998 cohort (92:32) (100%): BSSO index significant impro
Res. design after surgery. Op
assessment: 38 clicking incidenc
significantly redu
whereas crepitus
significantly incr
Panula et al, 27 Retrospective 60 (49:11) Ortho/surgery No/yes No/no No Helkimo index Variable: Specific TMD outco
2000 case series Class II: 49 (100%): BSSO Control group: 20–44 mo reported for skel
Res. design Class III: 11  Le Fort I 20 patients postsurgery malocclusion sub
assessment: 21 osteotomy who decided (aggregate of cla
not to have and ant. open bi
ortho/surgery Statistically signi
(16 F:4 M) decrease of TMJ
palpable muscle
pain, and headac
found. No statist
significant differ
between ortho/s
group and contr
at start of study.
statistically signif
differences for co
group assessmen
presurgery and
postsurgery time
Nemeth et al, 28 Prospective 64 R (47:17) Ortho/surgery Yes/no Yes/no Yes CM 2y No statistically sign
2000 cohort 63 NR (48:15) (100%): BSSO difference for ov
Res. design Class II: 127 MI, DI, or joint so
assessment: 38 (95:32) were found betw
NR fixation
Egermark, et al, 29 Retrospective 52 (34:18) Ortho/surgery No/no No/no No Modified Helkimo 2.2–9.5 y No description of
2000 case series (100%): BSSO index malocclusion typ
Res. design  Le Fort I significant differ
assessment: 14 osteotomy TMD signs and sy
between presurg
postsurgery outc
Kobayashi Retrospective 145 (99:46) Ortho/surgery No/no No/no No Physical examination 1y Primary outcome m
et al,30 case series Class III (100%): BSSO (TMJ pain, sounds, NR and R fixation
2000 malocclusion (40 R, 105 NR) movements and and TMJ pain/dys
Res. design  anterior limitations) No statistically sig
assessment: 17 open bite difference was fo
relapse between
fixation. No signi
differences were
TMJ pain and dy
between fixation
Westermark Retrospective 1516 (958:558) Surgery only: No/no No/no No Questionnaire 2y Questionnaire requ
et al,31 case series Class II: 526 BSSO (R and NR) (TMJ noise, pain, accurate memory
2001 Class III: 580 IVRO headaches, bruxism) presurgical condi
Res. design Ant. open Clinical examination Physical examina
assessment: 19 bite: 396 (only at 2-y follow-up) not include speci
3 fingers: normal measurements. S
2–3 fingers: reduced significant differe
1–2 fingers: severely preoperative sym
reduced free prognathic p
developing TMD
compared with sy
free retrognathic
developing TMD
Aghabeigi Retrospective Ant. open bite: Ortho/surgery No/no No/no No Clinical examination >1 y No statistically signi
et al,32 case series 83 (2:1 ratio) (100%): Le Fort I included mandibular difference betwe
2001  BSSO or IVRO ROM and deviation, presurgical and
Res. design osteotomy TMJ sounds, and postsurgical TMD
assessment: 17 TMJ and masticatory assessments. Post
muscle pain. pain was significa
Questionnaire associated with a
completed by abnormal psycho
42% of patients profile (higher st
included TMJ trait anxiety) and
scale, SCL-90, STAI, (female)
and VAS for overall
satisfaction
(continued on n
Table 1
(continued )

Well-Defined
Calibrated/ Inclusion/
No. of Surgical Approach Randomized Blinded Exclusion Postsurgical
Patients (% Combined Study/Control Examiners Criteria Follow-up
Reference Type of Study (F:M) Ortho/Surgery) Group (Yes/No) (Yes/No) (Yes/No) TMD Assessment Duration Comment
Dervis & Tuncer, 33 Prospective 50 NR (29:21) Surgery only: No/yes No/no Yes Helkimo index 2y No statistically sign
2002 cohort Unknown BSSO  Le Fort I Control group: differences betw
Res. design number osteotomy 50 healthy patients and con
assessment: 34 of patients subjects before surgery. N
in class I, II, without significant differ
and III TMD found among cl
subgroups (28 F:22 M) III presurgical an
postsurgical TMD
symptoms for ea
technique (beca
small N). After 2
postoperative, s
decreases found
palpation scores
scores in patient
statistically signi
difference was f
47% of women
improved subjec
symptoms comp
18% of men
Ueki et al,34 Retrospective Class III: 43 Ortho/surgery: No/no No/no No TMJ sounds and pain 6 mo Primary outcome m
2002 case series (100%) with mandibular TMJ pain/dysfun
Res. design BSSO  Le Fort I opening condylar positio
assessment: 19 osteotomy (23); disc displacemen
IVRO (20) statistically signi
 Le Fort I differences in TM
osteotomy or pain between
procedures. IVRO
associated with
normalized disc
Wolford et al, 35 Retrospective 25 (23:2) Surgery only: No/no No/no Yes Pain VAS; physical Variable: TMJ pain VAS increa
2003 case series Class II: 24 BSSO 1 Le Fort I examination 12–81 mo significantly betw
Res. design osteotomy presurgical and
assessment: 19 postsurgical meas
Maximum mandib
vertical ROM was
significantly decre
TMJ sound chang
not statistically ev
and reported
Pahkala & Retrospective 72 (49:23) Ortho/surgery: No/no No/no No Modified Helkimo Variable: mean TMJ clicking, TMJ pa
Heino, 36 case series Class II: 46 (100%) index; AAOP TMD 1.9 y (SD, 0.5) headaches were
2004 Class III: 14 BSSO (R) subgroups significantly decre
Res. design Ant. open (includes all skele
assessment: 21 bite: 4 malocclusion grou
Myogenous and
myogenous/arthr
groups had signif
reduction in symp
Arthrogenous gro
no significant pre
and postoperative
differences in sym
Borstlap et al, 37 Prospective Class II: 222 Ortho/surgery: No/no No/no Yes Questionnaire 24 mo TMJ pain was signifi
2004 cohort (169:53) (100%) and physical decreased after 2
Res. design BSSO (R) examination whereas clicking
assessment: 31 (TMJ sounds, unchanged comp
limited jaw presurgical assess
movement, Maximum mandib
palpable opening was not
joint pain) significantly diffe
whereas lateral a
protrusive movem
a statistically sign
decrease
Kallela et al, 38 Retrospective Class II: 40 Ortho/surgery: No/no No/no No Helkimo index Variable: 1–5 y TMJ clicking, crepitu
2005 case series (29:11) (100%) muscular pain we
Res. design BSSO (R) statistically reduc
assessment: 27 postoperatively co
with preoperative
assessment
(continued on n
Table 1
(continued )

Well-Defined
Calibrated/ Inclusion/
No. of Surgical Approach Randomized Blinded Exclusion Postsurgical
Patients (% Combined Study/Control Examiners Criteria Follow-up
Reference Type of Study (F:M) Ortho/Surgery) Group (Yes/No) (Yes/No) (Yes/No) TMD Assessment Duration Comment
Farella et al, 39 Prospective Class III: Ortho/surgery: No/no No/no Yes Axis I of RDC 12 mo No statistically sign
2007 cohort 14 (9:5) (100%) for TMD and differences in TM
Res. design BSSO (R) masticatory lateral or protrus
assessment: 33 1 Le Fort I muscle PPT movement, or m
osteotomy thresholds assessment. A sig
reduction in max
interincisal open
found ( 5 mm)
Valle-Corotti Prospective Class III: 25 Ortho/surgery: No/yes No/no No Questionnaire >1y No statistically sign
et al, 40 cohort (100%) Control group: and physical differences were
2007 Unknown surgery 25 patients examination between experim
Res. design only (TMJ sounds, control groups b
assessment: 20 treated by limited jaw presurgery and
orthodontics movement, postsurgery or
palpable joint preorthodontic a
and muscle pain) postorthodontic
Frey et al, 41 2008 Prospective 127 (95:32) Ortho/surgery Yes/no Yes/no Yes CMI 2y Patients who had a
Res. design cohort Class II: 127 (100%): BSSO counterclockwise
assessment: 40 Study compared mandibular rotat
amount of significantly mor
mandibular symptoms. Patien
advancement and had a mandibula
counterclockwise advancement >7
rotation counterclockwise
had significantly
muscle and TMJ
than shorter adv
and/or clockwise
Dujoncquoy Retrospective 57 (35:22) Ortho/surgery: No/no No/no Yes Questionnaire Variable: Descriptive statistics
et al,42 2010 case series (100%) 6–30 mo reported. Inferen
Res. design BSSO (R) statistical analyse
assessment: 13  Le Fort I not calculated for
osteotomy presurgical and
postsurgical outco
Data were depen
recall of TMD sign
symptoms. Result
included all skele
malocclusion grou
assessment of imp
each skeletal mal
Abrahamsson Retrospective 98 (60:38) Ortho/surgery: No/yes Yes/yes Yes Questionnaire 36 mo Class III patients had
et al,43 2013 case series Class II: 27 (100%) Control group: and RDC for significant decrea
Res. design Class III: 58 BSSO, IVRO, 38 healthy TMD physical myofascial pain a
assessment: 37 Ant. open Le Fort I, age-matched examination arthralgia postsur
bite: 13 maxillary and sex-matched (sounds, limited Class II patients h
segmental subjects movement, significant differe
osteotomies (R) without palpable TMD signs or sym
TMD muscle pain) between presurgi
postsurgical asses
No significant dif
in mandibular mo
Significant sex dif
in myofascial pain
frequency at base
all groups (F, 32%
12%) but not pos
Statistically signif
decreases in frequ
were found for m
pain, disc displace
and arthralgia be
presurgery and
postsurgery for a
malocclusion grou
Significant differe
between patients
controls were fou
myofascial pain a
arthralgia for all
(continued on n
Table 1
(continued )

Well-Defined
Calibrated/ Inclusion/
No. of Surgical Approach Randomized Blinded Exclusion Postsurgical
Patients (% Combined Study/Control Examiners Criteria Follow-up
Reference Type of Study (F:M) Ortho/Surgery) Group (Yes/No) (Yes/No) (Yes/No) TMD Assessment Duration Comment
Togashi et al, 44 Retrospective 170 (133:37) Ortho/surgery: No/no No/no No Physical examination 1y Significant differen
2013 case series Class II: 20 (100%) (sounds, limited for TMD signs an
Res. design (17:3) BSSO (R) movement, symptoms among
assessment: 19 Class III: 131  Le Fort I palpable malocclusion gro
(99:32) osteotomy TMJ pain) (prognathism, 17
retrognathism, 4
asymmetry, 58%)
surgery, no signif
differences in sig
symptoms of TMD
between subgrou
preoperative orth
no significant dif
of TMD signs and
symptoms compa
baseline for all
orthognathic sur
patients. Howeve
orthognathic sur
significant decrea
TMD signs and sy
were found for a
orthognathic sur
patients
Mladenovic Retrospective Class III: 40 Ortho/surgery: No/yes No/no No RDC for TMD 12 mo Only statistical com
et al,45 case series (25:15) (100%) Control group: was made postsu
2013 (25 R, 25 NR) 42 (17:25) compared with c
Res. design BSSO untreated Prevalence of my
assessment: 27 class III patients pain was similar
and women of co
group but was sig
higher in women
orthognathic sur
group, whereas T
was lower
Scolozzi et al, 46 Retrospective 219 (123:96) Ortho/surgery: No/no No/no No Helkimo index 1y TMJ clicking (Ai) wa
2015 case series Class II: 76 (100%) RDC for TMD predictive for TM
Res. design Class III: 51 BSSO (R) diagnostic 3.61) or MPD (OR
assessment: 27 Ant. open  Le Fort I classification TMJ clicking dete
bite: 42 osteotomy by Di, combined
Fort I and BSSO,
statistically signif
postoperative TM
determined by A
muscle tenderne
predictive of pos
TMD (Di). Maxim
mandibular open
mandibular prot
were significantl
decreased postsu
mandibular devi
increased
Yoon et al, 10 Retrospective 30 (11:4) Surgery only: No/yes No/no Yes Questionnaire 6 mo or more Both experimental
2015 case series Class III: 30 BSSO (R) 1 Control group: (subjective control groups h
Res. design Le Fort I 15 (8:7) changes in TMD skeletal malocclu
assessment: 29 osteotomy class III, symptoms) and TMD signs and sy
2 jaw surgery RDC for TMD before surgery.
and TMD clinical examination Experimental gro
treatment (ROM <35 mm, TMJ (surgical group)
sounds, TMJ pain) prior TMD treatm
had a statistically
significant reduc
TMJ noise and p
surgery. The cont
had treatment o
until signs/sympt
eliminated befor
and had a signifi
reduction in TMJ
not pain
(continued on n
 Table 1
(continued )

Well-Defined
Calibrated/ Inclusion/
No. of Surgical Approach Randomized Blinded Exclusion Postsurgical
Patients (% Combined Study/Control Examiners Criteria Follow-up
Reference Type of Study (F:M) Ortho/Surgery) Group (Yes/No) (Yes/No) (Yes/No) TMD Assessment Duration Comment
Sebastiani et al, 13 Prospective 54 Surgery only: No/no No/no Yes RDC for TMD; 6 mo Specific TMD outco
2016 cohort Class I: 4 BSSO (R) (TMJ sounds, not reported for
Res. design Class II: 17  Le Fort I muscle or TMJ pain, class I, II, and III
assessment: 25 Class III: 33 osteotomy mandibular mobility) patients. Inappro
statistical tests
(univariate tests)
were used when
comparing 3 or m
variables. Overall
significant decrea
sounds was repor
comparing preop
and postoperativ
assessments. No
statistically signif
changes were no
arthralgia, muscl
maximum openin
pain. A significan
difference was fo
between preoper
postoperative ma
mandibular open
without pain
Di Paolo et al, 47 Retrospective 76 Ortho/surgery: No/no No/no Yes RDC for TMD; 6–12 mo Statistical difference
2017 case series Malocclusion (100%) (TMJ sounds, presurgery and
Res. design types and BSSO (R) muscle or postsurgery outc
assessment: 17 frequency  Le Fort I TMJ pain, were not clearly
not reported osteotomy mandibular
mobility)

Abbreviations: AAOP, American Academy of Orofacial Pain; Ai, anamnestic index of Helkimo; Ant., anterior; BSSO, bilateral sagittal split osteotomy; CMI, craniomandibular index; Di, clinical dysfu
of Helkimo; DI, dysfunction index, a subindex of the CMI; F, female; IVRO: intraoral vertical ramus osteotomy; M, male; MI, muscle index, a subindex of the CMI; MPD, myofascial pain dysfunction; N
patients in study; NR, nonrigid fixation; OR, odds ratio; PPT, pressure pain threshold; R, rigid fixation; RDC, research diagnostic criteria; Res., research; ROM, range of motion; SCL-90, symptom chec
standard deviation; STAI, Spielberger state-trait anxiety inventory; TMD, temporomandibular disorders; VAS, visual analog scale.
 Orthognathic Surgery for TMD 15

Table 2
Advantages and limitations of a retrospective study for a pain condition

Advantages Limitations
 Has the potential to collect data from a large  Missing data are a limitation because there
number of patients usually is not a standardized set of data
 Retrospective chart reviews are relatively acquisition variables that are complete for all
inexpensive patients
 Can provide some limited information on the  Limited or no inclusion and exclusion criteria
characteristics of a study sample and their to identify a homogenous group of patients
response to treatment  Patients are not randomly selected (usually
 Can be valuable in rare pain conditions in referrals) and may not represent the general
which the number of patients is low and there population
is a paucity of treatment outcome assessment  Lack of a parallel control group matched for
 Can provide data to generate hypotheses for age/sex/ethnicity
a prospective pain study  Some retrospective studies require patients to
recall their pain conditions after months/years
and there is good evidence that memory of
pain is not accurate
 Lack of methods to calibrate a single examiner
or multiple examiners to minimize examiner
bias
 Lack of blinded examiners to minimize
examiner bias
 Descriptive statistical reporting is common but
inferential statistical testing is often not con-
ducted or is limited by an inadequate sample
size when stratification of patients is
performed (ie, sex, age, type of malocclusion,
type of surgery)

malocclusion. However, control groups were incor- (preorthodontics and presurgery) in a subgroup of
porated into some studies.10,25,27,33,40,43,45 They 30 patients in their study. However, this subgroup
consisted of non-patients with TMD such as dental was probably statistically underpowered because
students, orthodontic patients who did not want to of the small N (number of patients in study). Support
pursue orthognathic surgery, or untreated patients for this criticism is based on publication of a similar
with the same malocclusion as the experimental study that found statistically significant differences
orthognathic surgery patients. Some studies over time for the muscle index only after combining
matched the control patients by age and sex, the groups to increase the number of patients.41
whereas other studies did not incorporate this One consideration for future studies would be to
requirement in their control patients. Using un- have a parallel age-matched and sex-matched
matched controls is a major flaw in TMD studies control group similar to the orthognathic surgery
because sex and age are two known factors that group that was pursuing orthodontic therapy and
have an impact on the prevalence of TMD signs to continue to evaluate both groups over time using
and symptoms. the same assay time points. Using this strategy,
combined with control of experimental bias, it
INFLUENCE OF ORTHODONTIC TREATMENT would be possible to more accurately determine
the effect that orthognathic surgery may have on
Most patients in the reviewed studies (25 out of 34, TMD signs and symptoms.
or 74%) had orthodontic treatment before orthog-
nathic surgery, which is a potential confounding CONTROL OF EXPERIMENTAL BIAS
factor because it is possible that the signs and
symptoms of patients with TMD could be greater Experimental biases can appear in many different
at the orthodontic assessment time and could aspects of a clinical study and can be seen in
spontaneously be reduced during the treatment most of the reviewed studies. Well-designed ran-
period. This issue was examined by de Boever domized clinical trials require a rigorous attempt
and colleagues,50 who reported no statistically sig- to minimize various sources of bias, but this cannot
nificant differences between T1 and T2 time points be achieved in retrospective studies because of the
16 Dolwick & Widmer

lack of preplanned experimental design and data method to determine the relative impact of one pub-
acquisition. Examples of different experimental lished study relative to other published studies.
biases that were observed in the reviewed literature Therefore, a criteria-based assessment was devel-
are shown in Table 3. oped (Table 4) and applied to each clinical study
Another form of bias that has received more that was reviewed (reported in column 1 in
recent attention is the manipulation of acquired Table 1). The distribution of assessment scores
data, statistical analysis, and output bias.51 It is un- ranged from 9 to 40 and had a bimodal distribution
known whether any orthognathic surgery clinical (Fig. 1). The top 25% of scores formed the second
studies that were reviewed had such bias because peak in the distribution and is designated to the right
none were registered in ClinicalTrials.gov to allow a of the dotted line in the figure. For each skeletal
comparison of the proposed experimental design malocclusion subsequently discussed, emphasis is
and the reported outcomes measures in the publi- placed on the higher tier of studies and the effect
cations. However, registered neurology clinical tri- that orthognathic surgery had on TMD signs and
als were found to have 66% major discrepancies symptoms.
in the experimental design and statistical reporting
(eg, primary and secondary outcome variables MANDIBULAR RETROGNATHISM: EFFECTS OF
switched; timing of data collection; discrepancies ORTHOGNATHIC SURGERY ON INCIDENTAL
in reporting of probability values). TEMPOROMANDIBULAR DISORDER SIGNS
AND SYMPTOMS
ASSESSMENT OF BIAS MINIMIZATION IN
ORTHOGNATHIC SURGERY CLINICAL STUDIES Patients with class II skeletal malocclusions cor-
rected by a BSSO (Le Fort I osteotomy) were
Assessment of clinical studies that evaluated orthog- examined for changes in TMD signs and symptoms
nathic surgery for TMD management based on spe- in 23 publications (3 publications examined the
cific research design criteria would be an objective same data set) (see Table 1). Six of these

Table 3
Description and examples of bias in clinical studies

Experimental Bias Description Example

Sampling bias Sample does not reflect the Patient sample acquired from
characteristics of the general referrals to a practice/university or
population under study consecutive patients encountered
in practice
Selection bias Sample is not a homogenous group Studies that do not have specific
of patients with similar inclusion/exclusion criteria may
characteristics have a nonhomogenous group of
patients
Response bias TMD questionnaires are used to Questionnaires sent to patients are
determine previous and current returned by a motivated or
pain levels. Also, the questions interested subset and the
may depend on the patient’s characteristics of these patients may
memory of pain not represent the general population
Memory of pain has been shown not to
be accurate
Measurement bias Data acquisition lacks reliable and Examiners are not calibrated
valid measurement techniques regarding assessment of TMD
and this can bias the measures variables such as TMJ and muscle
pain, joint sounds, and
mandibular range of motion
Procedural bias Interaction and time spent with the Lack of a scripted set of questions
patient is not standardized and during the clinical interview
may bias the study
Outcome reporting bias Data are selectively omitted before No studies in the review of literature
statistical analysis and these are examined this issue. However,
reported in the publication approximately 66% of clinical
neurology publications had such
major discrepancies
 Orthognathic Surgery for TMD 17

publications were found to be in the top 25% of the

Table 4
Research design criteria for assessment of studies reviewed.26,28,33,37,41,43 After review of the
clinical studies top publications, the impact of orthognathic surgery
on TMD signs and symptoms in this group of pa-
Study Design Points tients was mixed at a 2-year follow-up assessment.
In 1 study, a counterclockwise rotation of the
Type of Study
mandible with at least a 7-mm advancement
Retrospective 1
caused an increase in muscle and TMJ pain,
Prospective 3 whereas a clockwise rotation had no effect.41 In 3
Randomized, prospective 5 other studies, there was no statistical difference be-
Control Group tween preoperative and postoperative assess-
None 0 ments.28,33,43 This lack of statistical difference can
Unmatched 2 be attributed to the small number of patients that
Age and sex matched 5 participated in the studies, resulting in an under-
powered ability to detect statistical differences. In
Calibrated/Blinded Examiners
addition, in 2 studies, overall muscle pain26 or
None 0 TMJ pain37 was reduced on postsurgical assess-
Calibrated or blinded 3 ment at 2 years. However, in 1 study, approximately
Calibrated and blinded 5 25% of patients had an increase in the muscle in-
Well-defined Inclusion/Exclusion Criteria dex score, demonstrating the variability in the
None 0 response to orthognathic surgery (see Figure 1 in
Inclusion or exclusion criteria 2 Rodrigues-Garcia and colleagues26). Based on
the current literature, a mandibular advancement
Both inclusion and exclusion criteria 4
greater than 7 mm combined with a counterclock-
Well-defined Skeletal Malocclusion Group
wise rotation seems to increase the risk of postop-
Not reported 0 erative masticatory muscle and TMJ pain, whereas
Mix of skeletal malocclusion groups 1 smaller advancements with a counterclockwise
Single skeletal malocclusion groups 5 rotation have a predominant effect on increasing
assessed muscle symptoms.
Well-defined Orthognathic Surgery Type
Not reported 0 MANDIBULAR PROGNATHISM: EFFECTS OF
Mix of maxillary/mandibular 1 ORTHOGNATHIC SURGERY ON INCIDENTAL
surgeries TEMPOROMANDIBULAR DISORDER SIGNS
Single maxillary or mandibular 5 AND SYMPTOMS
surgery assessed
Twenty-two publications were identified describing
TMD Assessment
correction of class III skeletal malocclusions and
Questionnaire 1
Clinical signs/symptoms 2
(variable)
Modified validated criteria 4
(ie, modified Helkimo index)
Validated criteria (RDC for TMD, 6
CMI, Helkimo index)
Follow-up Duration
Not reported 0
6 mo or less 2
1y 4
2 y or greater 6
Statistics Fig. 1. Histogram of assessment scores for rating
None 0 research design biases in clinical studies of orthog-
Descriptive statistics 2 nathic surgery and TMD. The vertical dotted line di-
vides the top 25% of studies (right side of dotted
Inferential: univariate 4 line) from the lower 75% of studies (left side of dotted
Inferential: multivariate 6 line). These top 25% represent the best of the reviewed
studies that were found in the current literature.
18 Dolwick & Widmer

evaluating TMD signs and symptoms (see Table 1). orthognathic surgery on preexisting TMD condi-
Of these publications, 4 were in the top 25% of all tions, or of their causing masticatory muscle or
studies reviewed.10,33,37,43 Two of the 4 publica- TMJ pain and dysfunction when none is present
tions found no overall difference in TMD signs and before surgery, is important for appropriate surgi-
symptoms between preoperative and postopera- cal planning and postsurgical management.
tive assessments at 1-year 39 and 2-year33 follow- Although there is a need for additional, well-
ups. Again, the lack of statistical difference could controlled studies to refine knowledge on this
be caused by the small number of patients exam- topic, there are a few studies that have provided
ined in each study. The other 2 studies found a sig- insight into potential TMD complications after
nificant decrease in the frequency of masticatory class II and class III skeletal malocclusion correc-
muscle pain43 and/or TMJ arthralgia10,43 at 6- tion. In studies that have minimized investigation
month or 3-year follow-up, but individual responses bias, class III correction seems to have an overall
varied.10 However, in contrast with class II skeletal minimal effect on the masticatory muscles and
malocclusion correction, there was a consistent TMJs in the follow-up time frame of 2 to 3 years.
overall reduction in TMD signs and symptoms. Class II correction seems to reduce the overall fre-
quency of TMD signs and symptoms with the
APERTOGNATHIA: EFFECTS OF exception of those cases that require greater
ORTHOGNATHIC SURGERY ON INCIDENTAL than 7 mm of mandibular advancement and a
TEMPOROMANDIBULAR DISORDER SIGNS counterclockwise rotation of the mandible or only
AND SYMPTOMS a counterclockwise rotation. The type of mandib-
ular fixation (ie, rigid or nonrigid fixation) also has
Although 9 studies were identified in the literature not been shown to elicit or increase TMD signs
review, none of the studies that evaluated only and symptoms after orthognathic surgery.
apertognathia were in the top 25% of studies. Studies of patients with preexisting TMD condi-
Most of the studies addressed anterior open-bite tions that were treated using orthognathic surgery
presentation in combination with correction of are limited in number. However, based on the
another skeletal malocclusion. Therefore, it was study outcomes for TMD in patients with class III
not possible to isolate the effects of correcting malocclusion treated by a BSSO and Le Fort I
apertognathia on TMD signs and symptoms. osteotomy, the frequency of TMD signs and symp-
toms was not increased and, when present, both
PRIMARY PATIENTS WITH TMJ pain and noise were significantly reduced.
TEMPOROMANDIBULAR DISORDER: EFFECTS However, individual responses varied. The lack
OF ORTHOGNATHIC SURGERY ON SIGNS AND of studies of other skeletal deformities, and the ef-
SYMPTOMS fect of orthognathic surgery on preexisting TMDs,
Only 1 study was identified that compared patients limits interpretation of orthognathic surgery pro-
with class III TMD treated using reversible treat- cedures as reliable techniques to reduce or elimi-
ment approaches before orthognathic surgery nate TMD signs and symptoms.
with patients with TMD treated only by orthog- This article clearly shows that most studies use
nathic surgery.10 This study had a small number orthodontic treatment before surgery and after
of patients per group (n 5 15). Self-reported symp- surgery to refine the occlusion. However, these
toms were evaluated by a questionnaire, whereas studies have not addressed the impact of ortho-
clinical signs and symptoms were assessed using dontics on TMD signs and symptoms. Most of
the research diagnostic criteria (RDC) for TMD the studies were underpowered statistically to
criteria. Using these data, the investigators determine this impact. Incorporation of an ortho-
concluded that treating the TMD signs and symp- dontic control group that is age and sex matched
toms before surgery had no effect on reducing and has the same timing of TMD assessment as
postoperative TMD signs and symptoms the orthodontic/surgical treatment group might
compared with orthognathic surgery alone. Both identify the possibility of the interaction of ortho-
groups of class III patients were reported to have dontic treatment and orthognathic surgery on
a reduction in signs and symptoms. TMDs. The number of patients that would be
required in each group to achieve sufficient statis-
DISCUSSION tical power to detect differences in TMD outcomes
between experimental and control groups seems
The concept of primary treatment of TMDs using to be greater than 125 based on the work of Frey
orthognathic surgery is not consistent with the cur- and colleagues.41
rent evidence-based approach to using reversible This article also emphasizes the need to mini-
therapies. However, knowledge of the impact of mize investigational biases observed in clinical
 Orthognathic Surgery for TMD 19

studies that have included orthognathic surgery/

TMD outcomes assessment. In our evaluation of
the literature, assessment criteria for rating clinical
studies were compiled and used to provide a
means of comparing each clinical study that was
reviewed. Most of the studies did not adequately
address various biases that are encountered in
clinical research. Thus, there is a need for addi-
tional well-designed studies to continue to explore
the potential impact of orthognathic surgery on
TMD signs and symptoms.

SUMMARY
The currently available evidence in the literature
does not support most orthognathic surgical inter-
ventions for correction of class II malocclusions,
class III malocclusions, or apertognathia as a
cause of postoperative temporomandibular disor-
ders. The exception seems to be a counterclock-
wise rotation of the mandible after a BSSO
procedure, which has been shown to be associ-
ated with an increase in masticatory muscle pain.
When this counterclockwise rotation is combined
with a mandibular advancement of greater than
7 mm, both masticatory myalgia and TMJ
arthralgia were found to increase in prevalence.
Patients who have a preexisting TMD and have a
class III skeletal malocclusion may experience an
overall reduction in the frequency of TMJ pain
and noise. However, individual responses are var-
iable and some patients have an increase in TMD
signs and symptoms. Therefore, more studies
incorporating appropriate sex-matched and age-
matched controls, controls for the effects of ortho-
dontics before orthognathic surgery, and
adequate sample sizes to allow for adequate sta-
tistical power to test for differences between pa-
tient and control groups are required to confirm
these earlier results.

REFERENCES
20 Dolwick & Widmer
Orthognathic Surgery for TMD 21