Risk factors associated with temporomandibular joint

sounds in children 6 to 12 years of age

Stephen D. Keeling, DDS, MS," Susan McGorray, PhD, b Timothy T. Wheeler, DDS, PhD, ~ and
Gregory J. King, DMD, DMSc ~
Gainesville, Fla.

The relationship between temporomandibular joint (TMJ) sounds and a person's dental and skeletal
characteristics is poorly understood. In this study, data were obtained from 3428 grade
schoolchildren (mean age = 9.0 years, SD = 0.8, range 6 to 12 years), without a history of
orthodontic treatment. Each child had been examined independently by one of six orthodontists to
assess: TMJ sounds (none, click, crepitus), gender, age, race (white/black), skeletal relationships
(convexity, maxillary, and mandibular positions), malocclusion (molar class, overjet, overbite, anterior
crowding, posterior crossbite), maximum opening, chin trauma (none, cut, scar), and history of lower
facial trauma. Temporomandibular joint sounds were present in 344 children (10.0% of the sample);
276 (8.1%) had an isolated unilateral sound, 254 (7.4%) had unilateral clicking, 50 (1.5%) had
bilateral clicking, 22 (0.6%) had unilateral crepitus, and 11 (0.3%) had bilateral crepitus. Univariate
analyses compared children with and without sounds for each variable; logistic regression analyses
examined the relationship between groups of variables and TMJ sounds. The prevalence of TMJ
sounds was associated with examiner (• = 23.4, df = 5, p < 0.001); increased prevalence of TMJ
sounds occurred in children with maxillary anterior crowding (t = 2.8, p < 0.006), mandibular
anterior crowding (t = 3.0, p < 0.002), and increased maximum opening (t = 4.7, p < 0.001). In
contrast to other reports on children, the prevalence of joint sounds was not associated with age,
race, gender, or molar class. After accounting for examiner differences and date of school
examination, we concluded that those grade schoolchildren with larger maximum opening, increased
anterior crowding, and deeper overbites had an increased risk for having a TMJ sound. (AM J
ORTHOD DENTOFACORTHOP 1994;105:279-87.)

T h e term temporomandibular joint disor-
ders (TMD) describes a condition characterized by pain
in the preauricular area, the temporomandibular joint
(TMJ) or the muscles of mastication, by a limitation of
the range of mandibular motion, and by the presence
of joint sounds during jaw function. I In addition, pain
on movement and deviation on opening have been con-
sidered signs of TMD. 2
Temporomandibular joint sounds have been de-
scribed as clicking, popping, crepitus, and grating and
a r e the most prevalent of all the signs of TMDP '4 Pa-
tients seeking care for TMD exhibit joint sounds sig-
nificantly more frequently than do controls. 4
The prevalence of TMJ sounds in children deter-
mined by palpation/unaided listening has been reported
From the University of Florida.
Supported by NIH-NIDR Grant DE 8715.
'Department of Orthodontics. College of Dentistry.
bDivision of Biostatistics, Department of Statistics, College of Liberal Arts
and Sciences.
Copyright 9 1994 by the American Association of Orthodontists.
0889-5406194153.00 + 0.10 8/1/41966
to range from 2.7% 5 to 26.6%. 6 In studies that used a
stethoscope, prevalence rates range from 0% 7 to
35.8%. 8 In an adult population, TMJ sound prevalence
has been described as dependent on the method used
with rates increasing to 78% when a stethoscope was
used and to 94% when phonographic recordings were
made. 9
Studies that have examined the association between
TMJ sounds and gender, age, and various descriptors
of morphologic/functional malocclusion are reported in
Table I. In children, gender differences have been re-
ported in only two populations, 1~ with girls having
more frequent clicking than boys. The majority of stud-
ies have not found gender differences in TMJ sound
rates in children.
Several cross-sectional studies have described in-
creasing prevalences of TMJ sounds with increasing
a g e . 7'8'11"16 Interestingly, this association with age has
not been confirmed in longitudinal studies in chil-
dren. '~ In several longitudinal studies, the presence
of TMJ sounds has been described as intermittent and
episodic, with the nature of the sound changing in char-

279
280 Keeling et al. American Journal of Orthodontics and Dentofacial Orthopedics
March 1994

Table !. Factors associated with clinically determined TMJ sounds in untreated children

Sample* Age Gender . Occlusion shiftlinterference Trauma Others

Saskatchewan, Canada ~3't4 Yes No {Class II molar} No {Muscle tenderness}

{Overjet > 7 mm)
{Buccal erossbite}
Northern Sweden t~ No {Yes}~ No No No {Opening <50 mm}
{Lateral pterygoidten-
derness}
Groningen, Netherlands 39 No No No {Permanent> deciduous}
{Deviationopen/close}
Malmo, Sweden ~9'n No No {Interferences in No
RP}
Varberg, Sweden (cross- Yes Yes No {Unilateral con- {Ocelusal Wear}
sectional data) u'n tact in RP} {Maximum9 opening}
{Orthodontic treatment}
{Number of teeth}
Varberg, Sweden (longi- No No Lateral deviation
tudinal data) ~7'1g'35 RCP-IP
Vertical deviation
RCP-IP
Tel Aviv, Israel ~ Yes No No No Occlusal wear
Jaw locking
TMJ tenderness
Cardiff, Wales n No No No
Ghent, Belgium ~ {Yes} No No No
University of Illinois 15 {Yes} No No No No {Large interincisal angle}
University of Florid# No

*Sample location with references noted.

{}Univariate association only. Variable did not enter multivariate model analysis.
1Not all age groups.
RP, Retruded contact position; RCP-IP, retruded contact position and intercuspal position.

acter and the sound disappearing over time, with Or
without treatment. 1~176
In the individual case, it is difficult toidentify the
i:ause o f a clinically determined TMJ sound. 2~ Clicking
sounds have been associated with uncoordinated muscle
function, irregularities of the joint compartmerits,22 disk
displacement with 23and without 2~reduction; crepitation
has been associated with degenerative arthritis. ~ Joint
sounds also have been associated with arthrosis of the
articular surfaces, 26 subluxation, 27 fluid c~ivitation, 2~
sudden movement of ligaments, 29 and lack of synovial
fluid. 3~
A relation between morphologic malocclusion and
signs and symptoms of dysfunction in children has been
suggested. 3~33 However, these telations are most often
between components of occlusion and an index of dys-
function. ~ Only one data set, the Saskatchewan sample
of 6- to 17-year-old children, 13.14demonstrated a sig-
nificant relation between morphologic malocclusion and
TMJ sounds: increased prevalence of TMJ sounds oc-
curred with Class II molar relation (versus Class I and
III), overjets greater than 7 mm, and buccal crossbites.
Other studies have not shown a relation between TMJ
sounds and occlusal descriptors (see Table I). Interest-
ingly, the Saskatchewan data set, when treated with
multivariate methods to model TMJ sound presence
using logistic regression, showed that only age re-
mained as a predictor of TMJ sound presence; the oc-
clusal descriptors did not add significantly to the
model)4
A relation between lateral deviation of the mandible
between retruded contact and intercuspal position and
Unilateral contact in retruded contact position and TMJ
Sounds in children has been shown) 1.12.35.36However,
the majority of Studies have shown no relation between
TMJ sounds and functional malocclusions (see Table
I). In those studies thathave, the relationships have
been univariate and have not held up under multivariate
statistical analysis. ~2
In addition, increased prevalence of TMJ sounds in
children has been associated in a univariate fashion with
muscle tenderness, l~ reduced mouth opening,37"3s
deviation in the open/close cycle, 39 occlusal wear, s.".12
increased maximum opening, orthodontic treatment,
American Journal of Orthodontics and Dentofacial Orthopedics
Volume 105, No. 3
and increased number of teeth, "'~2 jaw locking and
TMJ tenderness, s and a large interincisal angle. ~5
Facial trauma has not been associated with TMJ
sounds, i0o15.36-39
Collection of data on clinical signs of TMD in pop-
ulation studies by multiple examiners is further com-
plicated by the poor reliability of the measures2 '2~176
Reported symptoms of dysfunction have been shown
to have poor consistency,4~ the capability of question-
naires to detect clinical signs of dysfunction (click,
crepitus) has been shown to be poor. 42.43
The available evidence does not strongly support
the suggestion that TMJ sounds are associated with
morphologic or functional malocclusion. However, be-
cause the relationship between sounds and morphologic
malocclusion were demonstrated in one of the largest
samples reported (albeit, in only univariate anal-
yses) ~3"z4and many studies may be limited in their abil-
ity to detect these relationships because of their small
sample sizes ( < 5 0 0 ) and the poor reliability of clinical
measures, these associations need further examination.
The purpose of this study was to examine whether de-
mographic and malocclusion features of children would
be predictive of temporomandibular joint sounds.
METHODS AND MATERIALS
Subjects
The 4393 third and fourth grade students attending the
21 public elementary schools located in Alachua County, Fla.,
during the academic years 1990 and 1991 were targeted for
inclusion. Those students who had health permission screen-
ing forms on file with the schoo! system and were present on
the day of the school screening were examined.
Examination form
Examinations were performed in a room separated from
the classroom (typically, a media center or vacant classroom)
with the student standing erect in front of a seated examiner.
Each examiner used a handheld or headheld light, a millimeter
ruler, gloves, and tongue depressor for check retraction. Only
one of the six examining orthodontists examined each child;
all examiners did not go to each school. The mean number
of children seen by an examiner was 569.8, with a range of
369 to 1008. Information from six specific areas was recorded
on a standardized examination form:
1. Demographics. Data included name, address, phone
number, date of birth, sex, race, and history of ortho-
dontic therapy.
2. Skeletal relationships. The facial profile was assessed
visually as Class I (orthognathic), II (convex), or III
(concave). The anterior/posterior positions of the
maxilla and mandible were assessed as (1) rctro-
gnathic, (2) orthognathie, or (3) prognathic.
3. Trauma. History of trauma to the incisor teeth, lips
Keeling et al. 281
or chin was recorded as (0) no or (1) yes. Chin trauma
was scored as (0) none, (1) cut/bruise, or (2) scar.
4. Morphologic malocclusion. The following parameters
were assessed with the child in centrie (habitual) oc-
clusion: Overbite was scored in units of 1/3, from
(0) open bite to (4) greater than 100% overbite. Over-
jet was measured in millimeters using the central in-
cisors. Molar relationship was scored for right and
left sides in 1/4 cusp increments from (0) greater than
full cusp Class II to (1) greater than full cusp Class
III. Presence/absence of teeth, both deciduous and
permanent, in the oral cavity was recorded. Anterior
crowding/spacing from canine to canine, in each
arch, was scored as none, slight (>0 to --<3 mm),
moderate (>3 to -<6 mm), or excessive (>6 mm) on
a scale from (0) excessive space to (6) excessive
crowding.
5. Mandibular function. Joint sounds were scored as (0)
none, (1) clicking, or (2) crepitus by using light finger
pressure laterally over the TMJ area, while the sub-
jects opened and closed their mouths. Maximum ver-
tical openh,g of mandible was measured in millimeters
between the incisal edges of the central incisors while
the subjects opened their mouths as wide as possible.
6. Soft tissue relations, hzterlabial gap was measured in
millimeters from the inferior portion of the upper lip,
at the midline, to the superior portion of the lower
lip. The subject was instructed to lick his/her lips,
swallow, and relax before this measurement. Incisor
exposure was determined at the maxillary central in-
cisors with a millimeter ruler with the lips at rest.
Training of examiners
The six examining orthodontists met to design and to
review the screening forms and examination procedures. Be-
fore the initiation of data collection, all orthodontists partic-
ipated in calibration sessions, and each examined the same
39 elementary school students. Differences between ortho-
dontists were observed for several variables, and results of
this comparison were thoroughly discussed. Kappa statistics"
have been calculated to assess interexaminer reliability re-
garding presence or absence of TMJ sounds (Table II). These
statistics, comparing pairs of orthodontists, ranged from
- 0 . 1 0 to 0.66, with a median value of 0.36. It has been
suggested Kappa values less than zero be interpreted as show-
ing poor agreement, 0.00 to 0.20 slight, 0.21 to 0.40 fair,
0.41 to 0.60 moderate, 0.61 to 0.80 substantial, and 0.81 to
1.00 perfect. "~ Thus these data suggest, based on the median
value, fair agreement among examining orthodontists for de-
termining the presence or absence of TMJ sounds in a
person.
Statistical analysis
Descriptive statistics and graphic methods were used to
examine the data and characterize the population. Univariate
analysis compared students with TMJ sound(s) to those with-
out a TMJ sound for demographic, malocclusion, and screen-
282 Keelim, et al.
Table Ii. Kappa statistics for assessing interjudge agreement for presence of a TMJ sound (N = 39 for
each pair)
J, e I a I " I C
A 0.22 0.64
B -0.07
C 0.64
D 0.29
E - 0.03
ing-related variables. Chi-square tests, t tests, and Wilcoxon
rank sum tests were used where appropriate to test for dif-
ferences between the TMJ sound and no TMJ sound groups.46
A p value of less than 0.05 was considered statistically sig-
nificant. All available data for each variable were used.
Logistic regression47 was used to evaluate the joint or
multivariate effect of the explanatory variables with regard
to the presence of TMJ sound(s). The primary goal of this
analysis was to develop a parsimonious model for the prob-
ability of having a TMJ sound, by selecting an important
subset of the explanatory variables. This yielded information
regarding the relative risk associated with each variable, si-
multaneously taking into account other variables related to
TMJ sound. Only students with complete information for all
variables under consideration were used in this analysis. A
forward selection procedure, with the significance level re-
quired for variable entry set at 0.05, was used to develop the
model. A priori, it was decided to include the five variables
distinguishing the six orthodontists if any one of the variables
was selected. Goodness-of-fit of this model was evaluated
with a Hosmer-Lemeshow test.'~ These procedures were per-
formed on a Unix workstation with SAS~9 and S-pluss~ sta-
tistical software.
RESULTS
A total of 3742 out of 4393 third and fourth grade
public elementary school children in Alachua County,
Fla., were examined between Jan. 2, 1990, and Feb.
20, 1991; these students represented 85.2% of the tar-
geted population. After the exclusion of students with
missing TMJ sound data (n = 12) and those with a
history of previous orthodontic treatment (n = 302), a
sample of 3428 students remained (78% of the targeted
population). The mean age of this sample was
9.0 +_ 0.8 years (range: 6 to 12 years), with a
52.2/47.8 male/female ratio. The racial makeup was
59.8% white, 34.1% black, and 6.1% other (including
Oriental and Spanish).
Temporomandibular joint sounds were recorded in
344 children (10.0% of the total sample), with clicking
in 304 (8.9%), crepitus in 33 (1.0%), and click/crepitus
in 7 children (0.2%). In children with only clicking,
254 (7.4%) had a unilateral click, whereas 50 (1.5%)
had bilateral clicking. Unilateral crepitus was observed
in 22 (0.6%), bilateral crepitus in 11 (0.3%).
American Journal of Orthodontics and Dentofacial Orthopedics
March 1994
I O I I
0.46 0.29 0.37
-0.10 0.36 -0.04
0.47 0.66
0.37
Univariate results assessing associations between
categorical factors and TMJ sounds are presented in
Table III. As shown, the prevalence of TMJ sounds was
associated with examiner, with rates ranging from
6.25% to 13.78%. Statistically significant associations
were not detected between TMJ sounds and gender,
race, history of facial trauma, chin trauma, profile clas-
sification, maxilla classification, mandible classifica-
tion, posterior crossbite, or anterior crossbite.
Comparisons of ordinal and continuous variables
for those with and without TMJ sounds are presented
in Table IV. Significant differences between these two
groups were detected with regard to maximum opening
capacity and anterior crowding. Maximum opening was
greater in children with TMJ sounds. In addition, max-
illary anterior crowding and mandibular anterior crowd-
ing were greater in those with TMJ sounds. The TMJ
sound and no sound groups of children did not differ
with respect to age, overjet, interlabial gap, incisor
exposure, screening date, overbite, or molar classifi-
cation.
The relation between presence of TMJ sounds and
molar relation was further explored by examining right
and left molar Class combinations, both symmetries and
asymmetries (Table V). No differences in TMJ sounds
among the various molar class combinations were de-
tected, with the use of logistic regression t'o evaluate
the joint effects of left and right molar class combi-
nations (Chi-square for covariates indicating molar
class groupings = 2.452, 4 df, p = 0.65).
Logistic regression analysis was used to identify a
set of variables that, taken together, produced a parsi-
monious model of the probabilky of having a TMJ
sound. This multivariate approach allowed the simul-
taneous estimation of relative risk of TMJ sounds for
each of a set of factors. Interactions between risk factors
were explored and quantified. The final model was
based on forward selection criteria, univariate results,
and a priori m6del building decisions.
The explanatory variables in the final model, along
with their relative risk estimates and 95% confidence
intervals, are presented in Table VI. Variables repre-
senting each orthodontist and screening date were in-
 American Journal of Orthodontics and Dentofacial Orthopedics Keeling et al. 283
Volume 105, No. 3

Table Ill. Association between categorical variables and TMJ sounds

Clickl crepitus [ Chi square test
Variable Group n percent I (df)

Gender Female 1634 10.28 0.142 (1)

Male 1789 9.89 p = 0.707
Race White 2046 10.41 1.578 (2)
Black 1167 9.25 p = 0.454
Other 209 11.48
Orthodontist A 542 9.41 23.404 (5)
B 384 6.25 p < 0.001
C 369 7.86
D 434 7.60
E 682 13.78
F 1008 11.31
History of trauma No 2051 10.63 1.579
Yes 1375 9.31 p = 0.209
Chin trauma None 2937 10.15 0.259 (2)
Cut/bruise 45 8.89 p = 0.878
Scar 433 9.47
Profile Class I 2036 9.18 5.001 (2)
Class II 1332 11.56 p = 0.082
Class 11I 50 10.00
Maxilla Retrognathic 20 20.00 2.207 (2)
Orthognathie 3091 10.09 p = 0.332
Prognathic 309 9.71
Mandible Retrognathic 1204 11.63 4.909 (2)
Orthognathic 2083 9.22 p = 0.086
Prognathic 132 10.61
Posterior No 3155 10.21 0.980 (1)
Crossbite Yes 265 8.30 p = 0.322
Anterior No 3275 10.08 0.047 (1)
Crossbite Yes 147 9.52 p = 0.827

cluded in the model to account for these factors; the
other variables were maximum opening, anterior
crowding/spacing scores, and overbite classification.
From this analysis we conclude that increased risk for
having a TMJ sound is associated with larger maximum
opening, increased anterior crowding, and deeper
overbites. The Hosmer-Lemeshow goodness-of-fit test
was performed on this model, and a lack of fit was not
indicated (Chi-square = 7.184, p = 0.52).
Table VII illustrates the joint effect of two of the
selected variables, maximum opening and anterior
c r o w d i n g . Maximum opening and anterior crowd-
ing/spacing were used as continuous and ordinal vari-
ables, respectively, in the model, but have been dis-
played in categories in Table VII for illustrative pur-
poses. The percent of students with TMJ sounds ranged
from 2.0% for the group with variables associated with
the lowest level of sounds to 18.9% for those students
with explanatory variables associated with the highest
levels of TMJ sound.
DISCUSSION
Alachua County, Fla., 3rd and 4th grade school-
children were screened to identify subjects for a pro-
spective clinical trial evaluating early Class II ortho-
dontic treatments. The primary variables of interest
were molar relation, overjet, overbite, and numbers of
permanent teeth, as inclusion/exclusion criteria for the
clinical trial were based on these measures. The addi-
tional data were collected to characterize the popula-
tion. No attempt was made to exclude any subjects; the
response rate of 85% appeared to adequately represent
the population.
The reported prevalence rate of palpable TMJ
sounds (10.0%) in Alachua County grade schoolchil-
dren most closely resembled that reported in the Sas-
katchewan population) 3 Our data confirm earlier re-
ports that crepitus is an infrequent finding (1%) in chil-
dren,'2""37"39"s~'52and that unilateral sounds are far more
common than bilateral. 51"52
In contrast to other reports, the prevalence of
joint sounds was not associated with age or gender
in our sample of schoolchildren. Cross-sectional
studies 8'n'n'~4 showing increased prevalence of TMJ
sounds with age have included older children (ages 13
to 17 years) than those in this report. With the exception
of two Swedish populations of older children (ages 15,
17, and 18 years) where the girls had greater prevalence
 ,, ,, fLe'et;net
" al. American Journal of Orthodontics and Dentofacial Orthopedics
284 March 1994

Table IV. Comparison of ordinal and continuous variables for those with and without TMJ sounds

Variable sounds n I Mean SD Median

I Range
Test*
p value

Age no 3029 9.47 0.77 9.43 6.95-12.61 0.449

yes 343 9.50 0.75 9.54 7.73-12.30
Overjet (mm) no 3080 3.54 1.95 3 -4-12 0.874
yes 346 3.52 1.87 3 -2-12
Interlabialgap (mm) no 2339 1.25 2.00 0 0-12 0.824
yes 255 1.28 2.14 0 0-15
Incisor exposure (mm) no 2340 1.62 1.80 0 0-10 O.15l
yes 255 1.79 1.82 2 0-10
Maximum opening (mm) no 3077 47.21 5.81 46 25-70 0.001
yes 346 48.77 6.10 49 32-68
Screening date~ no 3079 260.74 82.20 295 1-415 0.088
yes 346 253.12 84.74 288 113-347
Overbite2 no 3063 1.58 0.88 2 0-4 O.164
yes 344 1.65 0.86 2 0-4
Molar class3, left no 3057 4.34 1.22 5 0-10 0.941
yes 342 4.31 1.30 5 0- tO
Molar class3, right no 3060 4.21 1.27 5 0- l 0 0.46 l
yes 344 4.15 1.37 5 0-I0
Anterior crowding4, upper no 3008 2.92 0.99 3 0-6 0.006
yes 332 3. l0 1.05 3 0-6
Anterior crowding4, lower no 3040 3.38 0.90 3 0-6 0.002
yes 335 3.54 0.95 3 0-6

*Two sample t tests were used for variablesage through maximumopening;Wilcoxonrank sum tests were used for the rest of the variables.
~Numberof days from I/1190 to screeningdate.
2Overbite coded from 0 to 4 with 0 = none, 1 = > 0 to -- 33%, ..., 4 = >100%.
3Molar class coded from 0 to l0 with 0 = >FC II, 5 = Class l, 10 = >FC III.
4Crowding coded from 0 to 6 with 0 = >6 mm space, 3 = none, 6 = >6 mm crowding.

of TMJ sounds than boys, 1~ most reports have not
reported gender differences in children. Ninety-seven
percent of ot/r sample was between 6 and I l years of
age, and thus most likely prepubertal. The findings of
no age or gender differences in prepubertal children
(this article and others), plus the previously mentioned
reports of age and gender differences in older children
(presumably postpubertal), suggest that changes during
the time of puberty may be linked to the appearance of
TMJ sounds. These changes might involve hormonal
changes or occlusal interferences during maturation of
the occlusion. Unfortunately, there are no studies re-
porting on longitudinal signs of T M D during the peak
pubertal years (l I to 16 years).
With the sole exception of the Saskatchewan
sample '3a4 of 6- to 17-year-old children which detected
univariate relations between TMJ sounds and Class II
molar, overjet > 7 mm, and buccal crossbites, all other
previous reports examining children have not shown
that morphologic malocclusion is associated with TMJ
sounds. However, the data in this report support the
concept that certain morphologic malocclusions (ante-
rior crowding, deep overbite) a r e associated with TMJ
sounds in children. The relation between anterior
crowding and TMJ sounds was unexpected and is dif-
ficult to explain. Of the articles reviewed, only one
reported a relation between anterior crowding and TM
dysfunction?3 In this study, TMJ sound data were a
component of the dysfunction scores, but were not eval-
uated separately. Our logistic regression model indi-
cated that increased maximum opening was associated
with increased TMJ sound prevalence. The overbite
variable, which did not have a univariate relation to
TMJ sounds, would probably not have entered the
model predicting TMJ sounds had the maximum open-
ing measures taken into consideration the amount of
anterior overbite.
It has been speculated that an association between
TMD and craniofacial structure exists. 3'33 However,
scant data appear in the literature to examine these
relationships. A review of the literature uncovered only
one report examining craniofacial structure and TMJ
s o u n d s Y These authors reported no relation between
measures of facial divergence and TMJ sound presence.
We did not assess vertical facial proportions directly
(long, short, average anterior facial height), but did
assess lip incompetence and incisor display at rest.
These measures of vertical facial features were not re-
AmericanJournal of Orthodonticsand DentofaclalOrthopedics Keeling et al. 285
Volume 105. No. 3

Table V. Symmetries and asymmetries in molar Table Vl. Logistic regression model for TMJ
classification and TMJ sound prevalence* sounds (yes, no)
Right Relative
Parameter [ risk 95% confidence
0,1,2,3 4,5,6 7,8,9,10 Variable estimate estimate* interval**
L~ ClassH Classl Classlll
Orthodontist t 1 -.0.50 0.61 0.42-0.89
0,1,2,3 701671 12/132 0/1 2 - 0.92 0.40 0.25-0.65
Class II 10.4% 9.1% 0.0% 3 - 0.23 0.80 0.59-1.08
4,5,6 36/288 218/2239 2/15 4 -0.99 0.37 0.22-0.61
Class I 12.5% 9.7% 13.3% 5 - 0.72 0.49 0.31-0.78
7,8,9,10 0/1 2/22 2/28 Screening date s 0.00 1.00 0.99-1.00
Class III 0.0% 9.1% 7.1% Maximum opening 0.05 1.06 1.03-1.08
Anterior crowding 3 0.26 1.30 1.14-1.48
*Logistic regression analysis indicated no significant effect of various Overbite 0.15 1.16 1.00-1.34
molar class combinations (Chi-square = 2.453, 4 df, p = 0.65).
*Value >1 corresponds to increased risk of sounds.
**For relative risk estimate.
1Compared to orthodontist no. 6.
lated to TMJ sound presence in children. In addition, 2Number of days from Jan. !, 1990, to screening date.
measures of facial convexity and maxillary and man- 3Maximum code of maxillary or mandibular crowding used.
dibular horizontal positions were not associated in either
a univariate or multivariate way with the presence of
TMJ sounds. Table VII. Anterior crowding, maximum
Fewer TMJ sounds were detected as the screening opening, and TMJ sound prevalence
progressed during the year. This trend also occurred in
Anterior crowdinglspacing
assess!ng incisor trauma) 4 Perhaps the orthodontists
tired of the task or developed more stringent criteria for Maximum 0,1,2 J 4,5,6
what constituted a TMJ sound. These data suggest that opening Space 3 J Crowding Total
there are challenges in maintaining consistency between
25-42 1/49 16/255 27/369 441673
and within examiners in assessing clinical outcome 2.0% 6.3% 7.3% 6.5%
measures during long-term clinical investigations. 43-46 2/56 27/390 54/451 831897
The results show variations among examiner 3.6% 6.9% 12.0% 9,3%
subgroups in prevalence of children with TMJ sounds. 47-51 8/81 39/417 43/396 901894
9.9% 9.4% 10.9% 10.1%
These differences might have resulted from differences
52-70 13/95 47/407 411217 101/739
in each examiner's population subset, differences be- 13.7% 11.6% 18.9% 13.7%
tween examiners in what constituted a TMJ sound, or TOTAL 24/281 129/1469 165/1453 318/3203
random misclassifieations. To support the first, data 8.5% 8.8% I 1.4% 9.9%
indicated that the orthodontists did evaluate different
groups of students With regard to several variables (gen-
de.r, age, overjet).* In support of the latter two, others
have also show n only moderate .interjudge reliability in pies of pers0ns seeking care and those not seeking care,
judging TMJ sounds, z~176Such differences in interpre 7 as well as the risk factors involved and the natural
tation could account for large differences in reported history of these disorders, are largely lacking: 6 al-
prevalences in the literature s-s and suggest that collec- though some recent reports appear promising in their
tion of prevalence data on TMJ sounds remains prob- approach. 4'4~ The prevalence of TMJ sounds appears
lematic. to decline rather dramatically after the fourth decade of
Although the signs and symptoms of TMD are pres- life: 8
ent in 28% to 86% of the adult population, whereas Temporomandibular disorder has been described as
0nly 3% to 7% of the population seek care for a tem- a condition characterized by pain, limitation in range
poromandibular joint or facial pain condition, it is not of motion, and TMJ sounds. One report has associated
at all clear which signs and symptoms of TMD are increase sound prevalence with reduced maximum
important in terms of illness.SS Adequate data describing opening (<50 mm) in 17-year-old adolescents? s How-
the incidence and prevalence in randomly selected sam- ever, our data support the finding of Egermark-Eriksson
et al. H,~2of a positive relation between maximum open-
*Original research data available on request. ing and TMJ sounds in children. These findings caution
286 Keelinget al.
that diagnostic schemes developed largely from adult
populations m a y not have utility in children.
We have e x a m i n e d o n l y the presence or absence of
a n y T M J sound and have not attempted to discriminate
a m o n g types o f T M J sounds. Crepitus was an infrequent
finding and afforded too small a sample size for anal-
ysis. Certain j o i n t sounds have been characterized, such
as crepitus associated with degenerative j o i n t disease
and reciprocal click with an anteriorly displaced
disk. :3"z5 However, there are a n u m b e r o f sounds that
have not been characterized and m a y be falsely inter-
preted as reciprocal click or crepitus. 2' In short, the
ability to differentiate a m o n g types o f sounds in the
clinical setting with regard to presence or absence of
specific disorder has not been shown. 26
It is not possible from these data to determine
whether early orthodontic treatment to establish "good"
occlusion is an efficacious approach to reduce the risk
o f developing T M dysfunction. F u n c t i o n a l parameter
data are currently being collected as o u t c o m e measures
in an o n g o i n g prospective clinical trial evaluating the
effectiveness, costs, and benefits o f early orthodontic
therapy, b y using a stratified r a n d o m a s s i g n m e n t of
children to control, headgear, and b i o n a t o r groups.
In c o n c l u s i o n , after accounting for e x a m i n e r dif-
ferences and date o f school e x a m i n a t i o n , the data in-
dicated that grade schoolchildren with larger m a x i m u m
o p e n i n g , increased anterior crowding, and deeper
overbite had an increased risk for h a v i n g a T M J sound.
We thank the schoolchildren of Alachua County, Fla.,
who participated, the County school administrators, espe-
cially Dr. Mel Lucas, and the classroom teachers who pro-
vided access and help during the screenings. In addition, we
thank Ms. Sarah Garrigues-Jones, Dr. Sal Cabassa, Dr. Rich-
ard Hocevar, Dr. Michael Kania, Dr. Debra Sappington, and
Mr. Nirander Nangia for their assistance during the screenings
and data management tasks.
REFERENCES
I. Laskin D, Greenfield W, Gale E, ct al. The president's confer-
ence on the examination, diagnosis, and management of tem-
poromandibular disorders. Chicago: American Dental Associa-
tion, 1983.
2. ttelkimo M. Epidemiological surveys of dysfunction of the mas-
ticatory system. Oral Sci Rev 1976;7:54-69.
3. Wabeke KB, Hansson TL, Hoogstraten J, van der Kuy P. Tem-
poromandibular joint clicking: a literature overview. J Cranio-
mandib Disord Facial Oral Pain 1989;3:163-73.
4. Dworkin SF, Huggins KH, LeResche L, et al. Epidemiology of
signs and symptoms in temporomandibular disorders: clinical
signs in cases and controls. J Am Dent Assoc 1990;120:273-81.
5. Mohlin B, Pilley JR, Shaw WC. A survey of craniomandibular
disorders in 1000 12-ycar-olds. Study design and baseline data
in follow-up study. Eur J Orthod 1991;13:111-23.
6. Hirata RH, Heft MW, Hernandez B, King GJ. Longitudinal study
of signs of temporomandibular disorders (TMD) in orthodonti-
American Journal of Orthodontics and Dentofacial Orthopedics
March 1994
cally treated and nontreated groups. AM J ORTtIODDENTOFAc
OR'rHOP 1992;101:35-40.
7. de Boever JA, van den Berghe L. Longitudinal study of func-
tional conditions in the masticatory system in Flemish children.
Community Dent Oral Epidemlol 1987;15:100-3.
8. Gazit E, Lieberman M, Eini R, et al. Prevalence of mandibular
dysfunction in 10-13 year old Israeli schoolchildren. J Oral Re-
hab 1984;11:307-17.
9. Pollman L. Sounds produced by the mandibular joint in young
men. A mass examination. J Maxillofac Surg 1980;8:155-7.
10. Wanman A, Agerberg G. Temporomandibular joint sounds in
adolescents: a longitudinal study. Oral Surg Oral Med Oral Pathol
1990;69:2-9.
1I. Egermark-Eriksson I, Carlsson GE, Ingervall B. Prevalence of
mandibular dysfunction and orofacial parafunction in 7-, 11- and
15-year-old Swedish children. Eur J Orthod 1981;3:163-72.
12. Egermark-Eriksson I, Ingervall B, Carlsson GE. The dependence
of mandibular dysfunction in children on functional and mor-
phologic malocclusion. AM J ORTHOO1983;83:i87-94.
13. Brandt D. Temporomandibular disorders and their association
with morphologic malocclusion in children. In: Carlson DS,
McNamara JA Jr, Ribbens KA, eds. Developmental aspects of
temporomandibular joint disorders. Mon~raph 16. Craniofacial
Growth Series. Ann Arbor: Center for Human Growth and De-
velopment, The University of Michigan, 1985:279-98.
14. Riolo ML, Brandt D, TenHave TR. Associations between oc-
clusal characteristics and signs and symptoms of TMJ dysfunc-
tion in children and young adults. AM J ORTI~ODDENTOFAC
ORTttOP 1987;92:467-77.
15. Runge ME, Sadowsky C, Sakols El, BeGole EA. The relation-
ship between temporomandibular joint sounds and malocclusion.
AM J ORTttOD DENTOFACORTIIOP 1989;96:36-42.
16. Ogura T, Takanubu M, Ohno H, Sumi K, Hatada K. An epi-
demiological study of TMJ dysfunction syndrome in adolescents.
J Pedodonties 1985;10:22-35.
17. Magnusson T, Egermark-Eriksson I, Carlsson GE. Four-year
longitudinal study of mandibular dysfunction in children. Com-
munity Dent Oral Epidemiol 1985;13:17-20.
18. Magnusson T, Egermark-Eriksson I, Carlsson GE. Five-year lon-
gitudinal study of signs and symptoms of mandibular dysfunction
in adolescents. J Craniomandib Pract 1986;4:338-44.
19. Nilner M. Functional disturbances and diseases in the stomato-
gnathic system among 7- to 18-year olds. J Craniomand Pract
1985;3:359-67.
20. Dworkin SF, LeResche L, DeRouen T. Reliability of clinical
measurement in temporomandibular disorders. Clin J Pain
1988;4:89-99.
21. Widmer CG. Temporomandibularjoint sounds: a critique of tech-
niques for recording and analysis. J Craniomandib Disord Facial
Oral Pain 1989;3:213-7.
22. Nanthaviroj S, Onmell K-A, Randow K, Oberg T. Clicking and
temporary "locking" in the temporomandibular joint. Dento-
maxillofae RadioI 1976;5:33-8.
23. lsberg-Holm AM, Westesson P-L. Movement of the disc and
condyle in temporomandibular joints with clicking. An arthro-
graphic and cincradiographic study on autopsy specimens. Acta
Odontol Scand 1982;40:153-66.
24. Miller TL, Katzbcrg RW, Tallents RH, Bessette RW, Hay~awa
K. Temporomandibular joint clicking with nonreducing anterior
displacement of the meniscus. Radiology 1985;154:12I-4.
25. Kopp S. Clinical findings in temporomandibular joint osteo-
arthrosis. Scand J Dent Res 1977;85:434-43.
26. Widmalm S-E, Westesson P-L, Brooks SL, Hatala MP, Paesani
American Journal of Orthodontics and Dentofacial Orthopedics
Vohune 105, No. 3
D. Temporomandibular joint sounds: correlation to joint structure
in fresh autopsy specimens. A.~t J ORTHOD DENTOFACORTHOP
1992;101:60-9.
27. Oster C, Katzberg RW, Tallents RH, et al. Characterization of
temporomandibular joint sounds. Oral Surg Oral Med Oral Pa-
thol 1984;58:10-6.
28. Unsworth A, Dowson D, Wright V. 'Cracking joints'--a bioen-
gineering study of cavitation in the metacarpophalangeal joint.
Ann Rheum Dis 1971;30:348-58.
29. Watt DM. Temporomandibular joint sounds. J Dent 1980;8:119-
27.
30. Gay T, Bertolami CN. The acoustical characteristics of the nor-
mal temporomandibular joint. J Dent Res 1988;67:56-60.
31. Williamson EH. Temporomandibular dysfunction in pretreat-
mcnt adolescent patients. AM J ORTHOD 1977;72:429-33.
32. Moyers RE. The development of occlusion and temporoman-
dibular joint disorders. In: Carlson DS, McNamara JA Jr, Rib-
bens KA, eds. Developmental aspects of temporomandibular
joint disorders. Monograph 16. Craniofacial Growth Series. Ann
Arbor: Center for Human Growth and Development, The Uni-
versity of Michigan, 1985:49-70.
33. Egermark-Eriksson I, Carlsson GE, Magnusson T, Thilander B.
A longitudinal study on malocclusion in relation to signs and
symptoms of craniomandibular disorders in children and ado-
lescents. Eur J Orthod 1990;12:399-407.
34. Helkimo M. Studies on function and dysfunction of the masti-
catory system. II. Index for anamnestic and clinical dysfunction
and occlusal state. Swed Dent J 1974;67:i01-21.
35. Egermark-Eriksson I, Carl~son GE, Magnusson T. A long-term
epidemiologic study of the relationship between occlusal factors
and mandibular dysfunction in children and adolescents. J Dent
Res 1987;66:67-71.
936. Nilner M. Epidemiology of functional disturbances and diseases
in the stomatognathic system. A cross-sectional study of 7-18-
year-olds, from an urban district. Swed Dent J 1983;Suppl 17:1-
44.
37. Wanman A, Agerberg G. Mandibular dysfunction in adolescents.
I. Prevalence of symptoms. Acta Odontol Scand 1986;44:55-
62.
38. Wanman A, Agerberg G. Relationship between signs and symp-
toms of mandibular dysfunction in adolescents. Community Dent
Oral Epidemiol 1986;14:225-30.
39. Meng HP, Dibbets JMH, van der Weele LTh, Boering G. Symp-
toms of temporomandibular joint dysfunction and predisposing
factors. J Prosthet Dent 1987;57:215-22.
40. Dworkin SF, LeResche L, De Rouen T, Von Korff M. Assessing
clinical signs of temporomandibular disorders: reliability of clin-
ical examiners. J Prosthet Dent 1990;63:574-9.
41. Heikinheimo K, Salmi K, Myllamiemi S, Kirveskari P. Symp-
toms of craniomandibular disorder in a sample of Finnish ado-
lescents at the ages of 12 and 15. Eur J Orthod 1989;I 1:325-31.
42. Riolo ML, Tenttave TR, Braudt D. Clinical validity of the re-
Keeling et al. 287
lationship between TF,IJ signs and symptoms in children and
youth. J Dent Children 1988;Mar.-April:110-3.
43. Nielsen L, Terp S. Screening for functional disorders of the
masticatory system among teenagers. Community Dent Oral Ep-
idemiol 1990;18:281-7.
44. Cohen J. A coefficient of agreement for nominal scales. Edue
Psych Meas 1960;20:37-46.
45. Landis JR, Koch GG. The measurement of observer agreement
for categorical data. Biometrics 1977;33:159-74.
46. Rosner B. Fundamentals of biostatistics. Boston: Duxbury Press,
1982.
47. Matthews DE, Farewell V. Using and understanding medical
statistics. Basel: S. Karger, 1985.
48. Hosmer DW, Lemeshow S. Goodness-of-fit tests for the multiple
logistic regression model. Communications in Statistics
1980;A9:1043-69.
49. SAS/STAT user's guide. Ver. 6, 4th ed., Vol. 1-2. Cary, North
Carolina: SAS Institute, 1990.
50. S-Plus user's manual. Seattle: Statistical Sciences, 1990.
51. Nilner M, Lassing S. Prevalence of functional disturbances and
diseases of the stomatognathic system in 7-14 year olds. Swed
Dent J 1981;5:173-87.
52. Heikinheimo K, Salmi K, Myllamiemi S, Kirveskari P. A lon-
gitudinal study of occlusal interferences and signs of cranio-
mandibular disorder at the ages of 12 and 15 years. Eur J Orthod
1990;12:190-7.
53. Lieberman MA, Gazit E, Fuchs C, Lilos P. Mandibular dys-
function in 10-18 year old school children as related to mor-
phological malocclusion. J Oral Rehab 1985;12:209-14.
54. Keeling SD, McGorray SP, Kania MJ, Wheeler TT, Hocevar
RA, King GJ. Risk factors associated with incisor trauma in
3396 Alachua County grade school children. Oral Epid Comm
Dent [in press].
55. Rugh JD, Solberg WK. Oral health status in the United States:
temporomandibular disorders. J Dent Educ 1985;49:398-406.
56. National Institutes of Health. The integrated approach to the
management of pain. National Institutes of Health Consensus
Development Conference Statement 6(3), 1986.
57. Dworkin SF, LeResche L, Von Korff MR. Diagnostic studies of
temporomandibular disorders: challenges from an epidemiologic
perspective. Anesthet Prog 1990;37:147-54.
58. Gross A, Gale EN. A prevalence study of the clinical signs
associated with mandibular dysfunction. J Am Dent Assoc
1983;107:932-6.
Reprint requests to:
Dr. Stephen D. Keeling
Department of Orthodontics
College of Dentistry
Box 100444, JHMHC
University of Florida
Gainesville, FL 32610