0150_IPC_AAP_553281 4/3/01 1:18 PM Page 485

J Periodontol • April 2001

The Effect of Occlusal Discrepancies

on Periodontitis. I. Relationship of
Initial Occlusal Discrepancies
to Initial Clinical Parameters
Martha E. Nunn* and Stephen K. Harrel†

Background: A causal relationship between occlusal dis-

crepancies and periodontal disease has been postulated in the
past. However, animal studies and clinical studies have not been
able to clearly demonstrate or rule out this potential relationship.
Methods: The records from a private practice limited to peri-
odontics were reviewed to find patients who had complete peri-
odontal examination records, including occlusal analysis, that

T
he role of trauma from occlusion
were recorded at least 1 year apart. Patients who fit these cri- in the progression of periodontal
teria were divided into a group who had none of the recom- disease is controversial. Trauma
mended treatment (untreated n = 30), those that had only non- from occlusion was once thought to be
surgical treatment (partially treated n = 18), and a control group one of the primary etiologic agents of
that had complete all recommended treatment (surgically treated periodontal disease.1,2 This theory was
n = 41). The data for each tooth of each patient, including based on the clinical observation that
occlusal status, were placed in a database and analyzed using periodontally involved teeth often showed
the generalized estimating equations (GEE) method to test for evidence of occlusal wear and that they
associations between initial occlusal discrepancies and various appeared to be undergoing occlusal
initial clinical parameters while adjusting for significant con- trauma at the time that periodontal dis-
founders. ease was diagnosed. As research began
Results: Teeth with initial occlusal discrepancies were found to demonstrate that the initiation and pro-
to have significantly deeper initial probing depths (P <0.0001), gression of periodontal disease was
significantly worse prognoses (P <0.0001), and significantly worse directly related to microbiologic factors,
mobility than teeth without initial occlusal discrepancies. In addi- the role of occlusion as an etiologic agent
tion, this association between initial occlusal discrepancies and came under question.3 Glickman and
initial periodontal condition was found to hold for various sub- Smulow postulated that instead of initi-
sets considered as well, including posterior teeth only and when ating periodontal disease, trauma from
only patients with good oral hygiene were considered. occlusion worked in concert with bacte-
Conclusions: This study indicates that there is a strong asso- ria to cause a more rapid progression of
ciation between initial occlusal discrepancies and various clin- periodontal destruction. They termed this
ical parameters indicative of periodontal disease. Based on relationship between trauma from occlu-
adjustments made for other known risk factors for periodontal sion and bacterial plaque as co-destruc-
disease, such as smoking, poor oral hygiene, etc., this study tive factors.4-6 The concept of occlusion
provides some evidence that occlusal discrepancy is an inde- as a co-destructive factor was questioned
pendent risk factor contributing to periodontal disease. J Peri- by Waerhaug.7,8 Based on observations
odontol 2001;72:485-494. of extracted teeth and the shape of
KEY WORDS osseous defects, he concluded that the
primary etiology of periodontal disease
Dental occlusion; malocclusion; periodontal diseases/
was bacterial plaque and that there was
etiology; follow-up studies.
no conclusive evidence that occlusal
trauma played a role in the progression
* Previously, Department of Public Health Sciences, Baylor College of Dentistry, Dallas, TX; of periodontal disease.
currently, Department of Health Policy and Health Services Research, Goldman School of
Dental Medicine, Boston University, Boston, MA. In an attempt to address this contro-
† Baylor College of Dentistry; private practice, Dallas, TX. versy, several animal studies were carried
out. The role of occlusion as an inde-

485
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 486
Relationship of Occlusal Discrepancies to Periodontal Disease
pendent factor and in combination with various other
factors were studied in beagle dogs9-17 and squirrel
monkeys. 18-26 Both groups found histologic evidence
that trauma from occlusion in the presence of good
plaque control caused bone loss but not necessarily
loss of attachment. In the beagle dog model, there was
evidence that, in the presence of both occlusal trauma
and bacterial plaque, there was both bone loss and
attachment loss. However, in the squirrel monkey
model, attachment loss was not detected despite the
presence of both occlusal trauma and bacterial plaque.
These animal studies seem to clarify that occlusal
trauma does not cause attachment loss when plaque
is not present. However, some question remains as to
whether there may be co-destruction when occlusal
trauma and bacterial plaque are both present at the
same time. Interpreting the results of these animal
studies is further complicated by the fact that in both
studies, due to the animal models used, it was neces-
sary to use mechanical methods to simulate occlusal
trauma. It can be argued that the cap-and-ball splint
used in the beagle dog model and the elastic tooth
separators used in the squirrel monkey model do not
adequately mimic the occlusal trauma observed in
humans.
There have been only a small number of clinical
studies that evaluated the relationship between occlusal
factors and the clinical manifestations of periodontal
disease. It has been consistently reported that the pres-
ence of occlusal discrepancies are not associated with
increased periodontal destruction.27-29 Two of these
studies did find a relationship between mobility and
increased probing depths, attachment loss, and dimin-
ished bone support.28,29 Another study also reported
that mobility was related to periodontal degeneration.30
This study found that greater attachment loss was pre-
sent on mobile teeth with furcation involvement as
compared to teeth with furcation involvement with no
mobility. A study that reviewed the results achieved
from periodontal therapy reported that patients who
had received an occlusal adjustment as part of their
periodontal therapy had a statistically greater gain in
attachment level than patients who had not received
an occlusal adjustment.31
A comprehensive literature review on the role of
occlusal trauma in periodontal disease is included in
the 1996 World Workshop in Periodontics.32 The pro-
ceedings of this workshop point out that there has been
very little recent research on the role of occlusion in
periodontal destruction and there have been no
prospective controlled studies on the role of occlusion
in the progression of untreated periodontal disease.
The proceedings also observed that ethical consid-
erations make such a prospective controlled clinical
trial unacceptable and, due to this, it is unlikely that
such a study will be performed. Another exhaustive
486
Volume 72 • Number 4
review of the literature in this area is contained in the
Occlusal Trauma section of the International Work-
shop for a Classification of Periodontal Diseases and
Conditions.33
With ethical considerations precluding the possibil-
ity of performing a controlled clinical trial evaluating
the role of occlusion in untreated periodontal disease
in humans, other types of studies are necessary to
assess the relationship of occlusal discrepancies and
periodontal disease. Taking into account these ethical
considerations, a retrospective epidemiological study
was conducted to investigate the relationship of
occlusal trauma to the severity of periodontal disease
as reflected in commonly measured clinical parame-
ters and to investigate possible effects of occlusal treat-
ment on the progression of periodontal disease. In this
paper, the relationship of initial occlusal discrepancies
to initial clinical parameters is explored.
MATERIALS AND METHODS
The data for this study were obtained from the clini-
cal records of a private periodontal practice. All avail-
able records from 24 years of practice were searched
for patients who fit the following criteria. All patients
had to be seen for a complete periodontal examina-
tion with data recorded for each tooth. These data
consisted of at least 6 sites of probing depths mea-
sured with a non-automated Michigan type probe,
bifurcation involvement (Glickman) measured with a
Nabor’s bifurcation probe, measurement of the width
of keratinized gingiva, measurement of mobility
(Miller), and analysis of occlusal relationships. All
patients had non-surgical and surgical periodontal
treatment recommended at their initial appointment as
part of their comprehensive treatment plan and failed
to complete all of the recommended periodontal treat-
ment. Additionally, all patients had to have a second
examination at least 12 months after the initial exam-
ination that included recording of another complete set
of data that duplicated the data recorded at the first
examination. All patients for whom this information
was available were included in this study. All exami-
nations and data collection were performed by the
same examiner.
Occlusal analysis included notation of initial con-
tact, discrepancies between initial contact in a retruded
position (centric relation), and maximum intercuspa-
tion (centric occlusion), and working and balancing
contacts in lateral and protrusive movements. Initial
contact was determined by having the patient close in
a retruded position until they detected an initial con-
tact. Lateral and protrusive contacts were determined
by having the patient move into lateral excursions from
a maximal intercuspation position. Contact points were
visualized and confirmed by the use of articulating rib-
bon.
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 487
J Periodontol • April 2001
The patients who fit these criteria were then divided
into 2 groups. An untreated group consisted of patients
who had none of the recommended periodontal treat-
ment performed between the 2 examinations, and a
partially treated group consisted of patients who had
completed the non-surgical portions of their treatment
but had not completed the recommended surgical
treatment. For comparison, a control group was also
formed, consisting of patients who had completed all
of the recommended periodontal treatment at least 12
months prior to the final examination. The control
group was formed by including the first 41 patients
who were seen during routinely scheduled periodontal
maintenance visits and who fit the criteria specified for
inclusion in the control group.
All patients who met the specified criteria for these
3 groups were entered into a database which included
the following patient information: age, smoking status
(smoker or non-smoker), presence or absence of a
medical condition such as diabetes or medication
known to negatively effect the periodontium (negative
health history), gender, oral hygiene (good, fair, poor),
compliance with treatment recommendations (com-
pliant, partially compliant, and non-compliant), and
the date of each examination where complete clinical
records were recorded. The following information was
recorded for each tooth for each visit: prognosis (good,
fair, poor, hopeless), probing depth (PD) in millimeters,
bifurcation involvement (Glickman class I, II, or III),
the presence of occlusal discrepancies (premature con-
tact with a vertical slide greater than or equal to 1 mm
or balancing contact in lateral movement), presence
or absence of a mucogingival defect, and mobility
(Miller 1, 2, or 3). The treatment performed for each
tooth was recorded as a yes or no response for the
following categories: root planing, occlusal adjustment,
osseous surgery, osseous regenerative procedure, and
soft tissue grafting
Prognosis for each tooth was assigned based on the
projected treatment outcome. A tooth with a good
prognosis was projected to be retained as a functional
unit with little or no treatment. A tooth with a fair prog-
nosis was projected to be retained as a functional unit
after treatment was completed. Teeth with a good or
fair prognosis were expected to have probing depths
of 2 to 4 mm following treatment. A tooth with a poor
prognosis was projected to be lost within 1 to 2 years
following treatment. A tooth with a hopeless progno-
sis was projected to be extracted during the course of
treatment. A diagnosis of fair to poor was given to
those teeth where the treatment outcome was in ques-
tion and where probing depths were projected to be ≥5
mm after treatment.34
The database was designed so that the data could
be evaluated for the effect of presenting factors, non-
treatment, partial treatment, and complete treatment
Nunn, Harrel
on the progression and/or resolution of periodontal
disease. The current paper presents the relationship
between the presence of occlusal discrepancies and
the presenting findings recorded at the initial evalua-
tion appointment.
Statistical Methods
Summary statistics or frequencies were computed for
initial patient characteristics, including gender, health
history, smoking status, oral hygiene status, and age
with patients classified according to occlusal status:
those patients with some initial occlusal discrepancies
and those patients without initial occlusal discrepan-
cies. Possible associations between initial occlusal sta-
tus and initial patient characteristics were tested using
chi-squared tests of independence for categorical
patient characteristics (such as gender, health history,
smoking status, parafunctional habit, etc.) and inde-
pendent sample t tests for continuous patient charac-
teristics (such as age). Summary statistics or fre-
quencies were also computed for initial clinical
parameters, including initial probing depth, initial prog-
nosis, initial mobility, and initial furcation involvement,
for teeth classified according to occlusal status (no
occlusal discrepancy versus occlusal discrepancy).
Because of the lack of independence of teeth within
each patient’s mouth, comparisons of each initial clin-
ical parameter by initial occlusal status were made
using the generalized estimating equations (GEE)
method while assuming an exchangeable working cor-
relation structure. The method of GEE is used in place
of traditional analysis of variance or regression analy-
sis when there is a lack of independence among obser-
vations, as is the case with tooth-level data collected
for this study.
In order to more fully evaluate the relationship of
initial occlusal status to initial probing depth, a multi-
ple regression model using GEE was constructed in
order to adjust for potential confounders such as age,
gender, health history, smoking status, parafunctional
habit, and oral hygiene status. A confounder is any
variable which may be associated with the outcome of
interest and may also be associated with the variable
under investigation, which in this case, is initial occlusal
status. Similarly, a multiple regression model using GEE
was constructed for evaluating the relationship of ini-
tial occlusal status to initial prognosis while account-
ing for potential confounders. Adjusted means and con-
fidence intervals were obtained for both initial probing
depth and initial prognosis by initial occlusal status
while adjusting for statistically significant confounders
in the multiple GEE regression models.
All statistical analyses were conducted using a sta-
tistical software program.‡
‡ Version 8.0, SAS Institute, Inc., Cary, NC.
487
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 488

Relationship of Occlusal Discrepancies to Periodontal Disease Volume 72 • Number 4

RESULTS ancy versus occlusal discrepancy). Associations

Exploratory Analysis
Data were collected retrospectively on 89 patients who
had sought consultation and/or treatment for moder-
ate to severe chronic adult periodontitis in the private
practice of one periodontist. Periodontal surgery was
indicated and recommended to all 89 patients in the
study. However, through self-selection, only 41 patients
completed all treatment recommended (control group),
another 18 patients consented to some non-surgical
treatment (partially treated group), and 30 patients
refused any treatment whatsoever (untreated group).
Those patients refusing treatment voluntarily returned
to the office at a future date and were re-evaluated. Of
the 71 patients who were treated either fully or partially,
26 received some form of occlusal adjustment, 17 out
of 41 (39%) fully-treated patients received occlusal
treatment, and 9 out of 18 (50%) partially-treated
patients, in order to correct occlusal discrepancies and
to alleviate potential occlusal trauma. In addition, there
were 30 patients who had occlusal discrepancies who
were not treated for this condition. Of these 30 patients,
5 were in the partially treated group (non-surgical treat-
ment), and 25 in the untreated group.
Table 1 shows the distribution of patient character-
istics by initial occlusal status (no occlusal discrep-
between patient characteristics and initial occlusal sta-
tus were tested using chi-squared tests of indepen-
dence. No statistically significant association between
gender and initial occlusal status (P = 0.26), health
history and initial occlusal status (P = 0.62), smoking
status and initial occlusal status (P = 0.35), oral
hygiene status and initial occlusal status (P = 0.64), or
parafunctional habit and initial occlusal status (P =
0.19) was found. When average age was calculated
according to initial occlusal status, and the mean age
of subjects with initial occlusal discrepancies was com-
pared to the mean age of subjects without initial
occlusal discrepancies using an independent samples
t test, it was found that subjects without an initial
occlusal discrepancy were significantly older than sub-
jects with an occlusal discrepancy (P = 0.005).
Table 2 shows statistics for initial clinical parame-
ters by initial occlusal status (no occlusal discrepancy
versus occlusal discrepancy with data collected on
each tooth). Associations between initial clinical pa-
rameters in Table 2 and initial occlusal status (no
Table 2.
Initial Clinical Parameters Individually by
Initial Occlusal Status

Table 1. No Occlusal Occlusal

Discrepancy Discrepancy P
Patient Characteristics by Initial Occlusal
Status Initial Probing Depth
N 1991 156
Mean (±SD) 4.77 (±1.31) 5.53 (±1.51)
No Occlusal Occlusal
Median 5.0 5.0
Discrepancy Discrepancy Range 2.0 to 9.0 3.0 to 9.0 <0.0001
Gender Initial Prognosis
Female 20 (61%) 27 (48%) N 1993 307
Male 13 (39%) 29 (52%) Good 896 (45%) 71 (23%)
Health Fair 1012 (51%) 198 (65%)
No Negative History 30 (91%) 49 (87%) Fair to Poor 41 (2%) 20 (7%)
Negative Health History 3 (9%) 77 (13%) Poor 36 (2%) 14 (5%)
Hopeless 8 (<1%) 4 (1%) <0.0001
Smoking Status
Nonsmoker 21 (64%) 30 (54%) Initial Mobility
Smoker 12 (36%) 26 (46%) N 1894 281
0 1467 (77%) 192 (68%)
Oral Hygiene 1 382 (20%) 74 (26%)
Satisfactory 22 (67%) 40 (71%) 2 34 (2%) 12 (4%)
Unsatisfactory 11 (33%) 16 (29%) 3 11 (1%) 3 (1%) 0.0316
Parafunctional Habit Initial Furcation
No Bruxism 30 (91%) 45 (80%) N 461 200
Bruxism 3 (9%) 11 (20%) 0 192 (42%) 80 (40%)
1 184 (40%) 72 (36%)
Age
2 64 (14%) 42 (21%)
Mean (±SD) 58.2 (±12.7) 50.4 (±11.6)
3 21 (5%) 7 (4%) 0.5508
Median 58.8 50.3
Range 35.9 to 80.5 24.9 to 88.1 P values based on simple GEE regression models using an exchangeable
working correlation matrix.

488
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 489

J Periodontol • April 2001 Nunn, Harrel

occlusal discrepancy versus occlusal discrepancy) were
tested using simple GEE regression models with an
exchangeable working correlation matrix. Teeth with
initial occlusal discrepancies were found to have sig-
nificantly deeper initial probing depths (P <0.0001),
significantly worse initial prognoses (P <0.0001), and
significantly greater mobility (P = 0.0316) than teeth
without initial occlusal discrepancies. No significant
differences in initial bifurcation involvement were
detected between teeth with initial occlusal discrep-
ancies and teeth without occlusal discrepancies.
Statistical Analysis With GEE
To further investigate the association of initial occlusal
status to these initial clinical parameters, multiple GEE
regression models were fit to include the following
potential confounders: age, smoking status (non-
smoker versus smoker), health history (no adverse
health history versus some adverse health history),
oral hygiene level (satisfactory versus unsatisfactory),
parafunctional habit, and gender. Confounding vari-
ables were retained in the GEE regression model only
when the P value was <0.10 for that particular variable.
Table 3A shows the final multiple GEE regression
model for evaluating the relationship of initial occlusal
status to initial probing depth. Smoking status, gender,
and oral hygiene were all found to be significant con-
Table 3A.
GEE Multiple Regression for Relationship of
Initial Occlusal Status to Initial Probing
Depth With Adjustment for Confounders
Parameter
Regression Parameter Estimate Standard Error
Intercept 4.39 0.11
Occlusal discrepancy 0.99 0.08
founders for predicting initial probing depth. The mul-
tiple GEE regression model shows that smokers have
significantly deeper initial PD than non-smokers, males
have significantly deeper initial PD than females, and
patients with unsatisfactory oral hygiene have signifi-
cantly deeper initial PD than patients with satisfactory
oral hygiene. In addition, teeth with initial occlusal dis-
crepancies have deeper initial PD than teeth without
occlusal discrepancies, even while adjusting for sig-
nificant confounders (Table 3B). On average, teeth
with an initial occlusal discrepancy will have approx-
imately 1 mm greater PD when compared to teeth
without an initial occlusal discrepancy, while adjust-
ing for significant confounders, such as smoking, gen-
der, and oral hygiene status.
In order to evaluate the relationship of initial prog-
nosis to initial occlusal status, initial prognosis was
scored from 1 for an initial prognosis of “good” to 4
for an initial prognosis of “hopeless.” Interim prog-
noses were assigned interim numerical values, such
as 2.5 for “fair-to-poor.” The final multiple GEE regres-
sion model is shown in Table 4A. The only significant
confounder for predicting initial prognosis was smok-
ing status. Specifically, teeth in smokers were found to
have a significantly worse prognosis than those in non-
smokers. Again, teeth with occlusal discrepancies were
found to have significantly worse initial prognoses than
teeth without occlusal discrepancies (Table 4B). On
average, teeth without occlusal discrepancies have a
“fair-to-good” initial prognosis (mean initial prognosis
Table 4A.
GEE Multiple Regression for Relationship of
Initial Occlusal Status to Initial Prognosis
P With Adjustment for Confounders
<0.0001
Parameter
<0.0001 Regression Parameter Estimate Standard Error P

Smoker 0.34 0.15 0.0245 Intercept 1.52 0.036 <0.0001

Male 0.30 0.14 0.0364 Occlusal discrepancy 0.35 0.038 <0.0001

Unsatisfactory oral hygiene 0.30 0.16 0.0672 Smoker 0.16 0.062 0.0097

Table 3B. Table 4B.

Average Initial Probing Depth by Initial Initial Prognosis by Initial Occlusal Status
Occlusal Status Adjusted for Significant Adjusted for Significant Confounders
Confounders
Adjusted Initial
Initial Occlusal Status Adjusted Initial PD 95% CI* Initial Occlusal Status Prognosis 95% CI*

No occlusal discrepancy 4.83 4.66, 4.99 No occlusal discrepancy 1.59 1.53, 1.65

Occlusal discrepancy 5.82 5.60, 6.03 Occlusal discrepancy 1.94 1.86, 2.02
* Confidence intervals. * Confidence intervals.

489
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 490
Relationship of Occlusal Discrepancies to Periodontal Disease
 1.5), whereas teeth with occlusal discrepancies have,
on average, just a “fair” initial prognosis (mean initial
prognosis  2.0).
When confounders were considered for inclusion in
the GEE regression model for predicting initial mobil-
ity from initial occlusal status, no terms were found to
add anything to the model beyond initial occlusal sta-
tus. The average initial mobility for teeth with initial
occlusal discrepancies was found to be 0.38 (95% CI:
0.28 to 0.49) while the average initial mobility for teeth
without initial occlusal discrepancies was found to be
0.27 (95% CI: 0.21 to 0.34). In order to investigate
the possible association of occlusal discrepancies to
mobility further, the mobility of teeth were reclassified
as either “no initial mobility” or “initial mobility.” A
GEE logistic regression model was fit to determine the
relationship of initial occlusal status to the presence of
any initial mobility. In addition, all potential confounders
were considered for inclusion in the GEE logistic
regression model for predicting the probability of the
presence of mobility initially. The only predictor that
approached significance was initial occlusal status as
shown in Table 5. The odds of a tooth exhibiting mobil-
ity initially was about 1.2 times as great among teeth
with initial occlusal discrepancies compared to teeth
without initial occlusal discrepancies, although this
association failed to achieve the 0.05 level of signifi-
cance (P = 0.056).
No significant association between initial occlusal
status and initial furcation involvement was found with
any of the multiple GEE regression models. The fur-
cation involvement of teeth was also reclassified into
“no initial furcation involvement” and “initial furcation
involvement.” GEE logistic regression models were
again fit as in the case of mobility, but no significant
association between the presence of initial furcation
involvement and initial occlusal discrepancies was
found, although the estimation of the odds for the pres-
ence of an initial furcation involvement of a tooth with
an initial occlusal discrepancy was about 1.04 times
Table 5.
GEE Logistic Regression for Relationship
of Initial Occlusal Status to Initial Mobility
and Initial Furcation Involvement
Clinical Parameter Odds Ratio 95% CI*
Mobility
No occlusal discrepancy 1.00
Occlusal discrepancy 1.19 1.00, 1.43 0.056
Furcation Involvement
No occlusal discrepancy 1.00
Occlusal discrepancy 1.04 0.87, 1.26 0.644
* Confidence intervals.
490
Volume 72 • Number 4
that of a tooth without an initial occlusal discrepancy.
No confounders added significantly to the model for
predicting the presence of initial furcation involvement
in the GEE logistic model. However, age was a signif-
icant predictor of the severity of initial furcation involve-
ment when considered for inclusion in the multiple
GEE linear regression model with all levels of furca-
tion involvement included in the response.
To further investigate the relationship of initial
occlusal discrepancies to initial clinical parameters, all
of the previous GEE analyses were repeated on a sub-
set of posterior teeth alone since the risk for occlusal
discrepancies often occurs among posterior teeth. Sim-
ilar results were found for the association of initial
occlusal discrepancies to initial clinical parameters
with the presence of initial occlusal discrepancies con-
tributing significantly to increased initial probing depths
among posterior teeth (P <0.0001), to worse initial
prognoses among posterior teeth (P <0.0001), and to
increased initial mobility among posterior teeth (P =
0.0363).
Again, a multiple GEE regression model was con-
structed to fully investigate the relationship of initial
occlusal discrepancies to initial probing depth while
adjusting for significant confounders among posterior
teeth alone (Tables 6A and 6B). As in the analysis of
all teeth, smoking, poor oral hygiene, and male gen-
der were all significant factors in increased initial PD
among posterior teeth.
A multiple GEE regression model was constructed
to fully investigate the relationship of initial occlusal dis-
crepancies to initial prognosis while adjusting for sig-
nificant confounders among posterior teeth alone
(Tables 7A and 7B). As indicated previously, higher
prognosis scores are indicative of worse prognoses so
that Table 7B demonstrates that posterior teeth with ini-
tial occlusal discrepancies have significantly worse
prognoses with an average initial prognosis of 2.0 than
posterior teeth without initial occlusal discrepancies
that have an average initial prognosis of 1.8. This mul-
tiple GEE regression model for predicting initial prog-
nosis for posterior teeth was slightly different from the
model for all teeth with poor oral hygiene also included
in the model since it met the 0.10 level of significance
criterion for inclusion in the model. The presence of ini-
tial occlusal discrepancies, poor oral hygiene, and
smoking were all found to contribute significantly to a
P worse initial prognosis among posterior teeth.
Since previous studies have indicated that the
impact of occlusal trauma on the periodontium may
be linked with oral hygiene, separate GEE regression
models were constructed for only patients with good
oral hygiene in order to determine the effect of initial
occlusal discrepancies on initial PD and initial prog-
nosis. As in the original analysis, all potential con-
founders were considered for inclusion in the model.
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 491

J Periodontol • April 2001 Nunn, Harrel

Table 6A. Table 7A.

GEE Multiple Regression for Relationship of GEE Multiple Regression for Relationship of
Initial Occlusal Status to Initial Probing Initial Occlusal Status to Initial Prognosis
Depth with Adjustment for Confounders with Adjustment for Confounders Among
Among Posterior Teeth Only Posterior Teeth Only

Parameter Parameter
Regression Parameter Estimate Standard Error P Regression Parameter Estimate Standard Error P

Intercept 4.82 0.11 <0.0001 Intercept 1.65 0.04 <0.0001

Occlusal discrepancy 0.65 0.10 <0.0001 Occlusal discrepancy 0.22 0.04 <0.0001

Smoker 0.35 0.15 0.0203 Smoker 0.17 0.06 0.0045

Male 0.31 0.15 0.0342 Unsatisfactory oral hygiene 0.11 0.06 0.0608

Unsatisfactory oral hygiene 0.33 0.16 0.0361

Table 7B.
Table 6B. Average Initial Prognosis by Initial
Average Initial Probing Depth by Initial Occlusal Status Adjusted for Significant
Occlusal Status Adjusted for Significant Confounders Among Posterior Teeth Only
Confounders Among Posterior Teeth Only
Adjusted Initial
Adjusted Initial Initial Occlusal Status Probing Depth 95% CI*
Initial Occlusal Status Probing Depth 95% CI* No occlusal discrepancy 1.78 1.72, 1.84
No occlusal discrepancy 5.28 5.12, 5.44 Occlusal discrepancy 2.01 1.92, 2.09
Occlusal discrepancy 5.93 5.71, 6.15 * Confidence intervals.

* Confidence intervals.

occlusal discrepancies among patients with good oral

However, the only significant predictor of initial PD
among patients with good oral hygiene was found to
be initial occlusal discrepancies. In particular, teeth
with initial occlusal discrepancies among patients with
good oral hygiene had an average initial PD of 5.86
mm (95% CI: 5.52 to 6.20) while teeth without initial
occlusal discrepancies among patients with good oral
hygiene had an average initial PD of 4.98 mm (95%
CI: 4.70 to 5.26). Significant confounders were also
considered for inclusion in the model for the predic-
tion of initial prognosis of teeth in patients with good
oral hygiene. Again, only the presence of occlusal dis-
crepancies was found to be a significant predictor of
the initial prognosis of teeth in patients with good oral
hygiene. Specifically, teeth with initial occlusal dis-
crepancies among patients with good oral hygiene had
an average initial prognosis of 1.96 (95% CI: 1.83 to
2.89), while teeth without initial occlusal discrepan-
cies among patients with good oral hygiene had an
average initial prognosis of 1.65 (95% CI: 1.53 to 1.77).
Hence, on average, teeth without initial occlusal dis-
crepancies among patients with good oral hygiene
have a “fair-to-good” prognosis whereas teeth with
hygiene only have a “fair” prognosis, on average.
DISCUSSION
Although many studies have been conducted to inves-
tigate the potential relationship of occlusal trauma to
the severity and/or progression of periodontal disease,
only one other study28 is known in which data were col-
lected on a tooth level in order to assess the effect of
occlusal discrepancies on each tooth. This is an impor-
tant consideration since the problems that each patient
encounters with occlusal discrepancies can vary quite
widely. Some patients may have only isolated teeth
that are exposed to occlusal trauma while other
patients may experience more widespread problems
with occlusion. Hence, by collecting data on the
occlusal status of each tooth and relating that condi-
tion to the clinical parameters associated with that
tooth, it is possible to more accurately explore the
association between occlusal discrepancies and the
severity of periodontal disease than by looking at
occlusion on a patient level as has been explored in
previous studies.
When comparing patients with occlusal discrepan-

491
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 492
Relationship of Occlusal Discrepancies to Periodontal Disease
cies to patients without occlusal discrepancies, the
only statistically significant difference in any patient
characteristic was age. In particular, patients without
occlusal discrepancies were significantly older than
patients with occlusal discrepancies. The reason for
this finding is unknown. Possible explanations could
include reduction of occlusal discrepancies by attri-
tion of the contact points, loss of the teeth that are
experiencing initial or balancing contacts, or move-
ment of teeth due to occlusal pressure and/or peri-
odontal disease.
Investigation of the relationship of initial occlusal
status to initial clinical parameters revealed that teeth
with occlusal discrepancies had significantly deeper
probing depths, were significantly more likely to have
increased mobility, and had significantly worse prog-
noses than teeth without occlusal discrepancies. This
is different from the results of the study by Jin and
Cao,29 who found no increase in probing depth in the
presence of occlusal discrepancies. The reasons for
this difference in results is not clear, although at least
2 potential reasons are evident. The current study has
a larger number of patients (n = 89 versus n = 32) and
a larger number of teeth (n = 2,147 versus n = 410)
than the earlier study so that the current study cer-
tainly has more power for detecting differences in prob-
ing depth. In addition, it is unclear what statistical
methods Jin and Cao used to analyze their data. They
mention using t tests to make their comparisons, but
this type of testing would not be valid unless the data
were somehow reduced down to a patient level since
data collected for each patient is certainly correlated.
This may have had a serious impact on their study
since reducing the data down to a patient level with-
out making use of the more sophisticated form of
analysis possible with GEE would most likely result in
an even more serious reduction in the power of the
study than that introduced by the smaller sample size.
Because of the much larger sample size involved and
the sophisticated statistical techniques used in the pre-
sent study, it is felt that the current study may be more
representative of the conditions found in the general
population and may have much greater power to detect
differences in the initial clinical condition of teeth with
occlusal discrepancies versus teeth without occlusal
discrepancies.
No statistically significant association was found
between initial furcation involvement and initial occlusal
status. The association of initial occlusal discrepan-
cies to increased mobility was to be expected, since
teeth exposed to occlusal trauma often exhibit mobil-
ity. The association of initial occlusal discrepancies to
significantly worse prognoses may be related to the
deeper probing depths, so that this finding may not be
completely independent of the probing depth results.
With the relationship of initial clinical discrepancies to
492
Volume 72 • Number 4
both initial probing depth and initial mobility, it was
unexpected that no association between initial occlusal
status and initial furcation involvement was found. The
lack of association between initial furcation involve-
ment and initial occlusal status may be related to the
high proportion of molars with initial furcation involve-
ment, with almost 60% of molars having some initial
furcation involvement. In addition, only 661 molars
were available, so that analysis of furcation involvement
afforded less statistical power for detection of differ-
ences than the other clinical parameters collected.
Since previous studies have indicated that the role
of occlusal trauma in the etiology of periodontal dis-
ease is strongly related to other factors such as oral
hygiene, a multiple regression model using GEE was
constructed to take into account other factors besides
occlusal discrepancy that might explain the differences
in initial probing depths and initial prognoses that were
observed in this study. Even when smoking status and
oral hygiene were taken into account, initial occlusal
discrepancy was still shown to be strongly associated
with deeper initial probing depths. In fact, initial
occlusal discrepancy was a stronger predictor of ini-
tial probing depth than any other factor considered.
Overall, teeth with initial occlusal discrepancies
exhibited probing depths that were about 1 mm deeper
than teeth without initial occlusal discrepancies. In
addition, when a GEE multiple regression model was
constructed for predicting initial prognosis, initial
occlusal discrepancy was also a very strong predictor
of initial prognosis, even when both smoking status
and oral hygiene status were included in the model.
The results of these multiple GEE regression models
for initial probing depth and initial prognosis were fur-
ther strengthened by finding similar results when only
posterior teeth were considered since posterior teeth
are frequently the teeth that most often suffer from
occlusal trauma. It was also an interesting and some-
what unexpected finding to discover that the presence
of initial occlusal discrepancies among patients with
good oral hygiene was the only statistically significant
predictor of initial probing depth and initial prognosis.
This is of particular interest since it raises questions
about previous findings in animal models and indicates
that more research should be conducted to fully under-
stand the relationship of oral hygiene and occlusion
to the progression of periodontal disease as well as
any possible negative synergistic effects that these fac-
tors may have on the periodontium.
Another interesting finding was the failure of a his-
tory of a parafunctional habit to be significantly asso-
ciated with initial probing depth, initial prognosis, ini-
tial mobility, or initial furcation involvement as
previously reported by McGuire and Nunn.35 However,
there are some important differences; they investigated
the association of untreated parafunctional habits to
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 493
J Periodontol • April 2001
tooth loss over time, whereas the current study only
looks at the association of parafunctional habits to ini-
tial clinical parameters when the patient was initially
evaluated. In addition, the McGuire and Nunn study
covered a long period of time, so that it was more likely
to detect and treat parafunctional habits. In this study,
patients in the untreated and partially treated groups
were not followed as closely or for as long a period,
and many simply denied having a problem. In this study,
we only had 14 individuals that admitted to having a
parafunctional habit, which may indicate that we did
not have many subjects with parafunctional habits or
that some of the subjects were reluctant to admit that
they might have a problem. Further long-term studies
would be useful in addressing this issue more fully.
The results of this study would seem to raise ques-
tions concerning the true role of occlusal trauma in
the etiology of periodontal disease. Although previous
studies27-31 have indicated that occlusion may play
only a secondary function in the initiation and pro-
gression of periodontitis, none of them evaluated
occlusal trauma on a tooth level basis as we have done
in this study. Therefore, it is possible that previous
studies have underestimated the impact of occlusion
on the periodontium because of the way in which
occlusal trauma and/or occlusal treatment were mea-
sured. This is a particularly important point because
the statistical tools for conducting a tooth level analy-
sis have only become generally available in recent
years so that early studies did not have the means for
appropriately testing the relationship of occlusal trauma
to the initiation and progression of periodontal disease.
There are limitations to this study that must also be
considered. For instance, data on occlusal discrepan-
cies were collected as a dichotomous variable so that
the degree of discrepancy was not taken into account.
In addition, data were not collected for some initial
clinical variables, such as bone loss, clinical attachment
level, crown-to-root ratio, etc., which would have pro-
vided more compelling evidence of a relationship of
occlusal discrepancies to the severity of periodontal
disease. However, it should be pointed out that these
types of data are not routinely collected outside of
prospective clinical trials and that the current data
came from the records of a private periodontal prac-
tice. In addition, because one clinician collected these
data, there is a potential source of bias in measure-
ment. It should be noted, however, that any bias in
measurement by having a single clinician conducting
the study should occur in the same direction, so that
observed differences in clinical measures between teeth
with occlusal discrepancies and teeth without occlusal
discrepancies cannot be discounted.
Further studies on a tooth level basis should be con-
ducted to fully evaluate the role of occlusion in the
progression of periodontal disease. In particular, future
Nunn, Harrel
studies should attempt to measure the amount of
occlusal discrepancy on each tooth in order to more
finely delineate the associations noted in this paper.
REFERENCES
1. Stillman PR. The management of pyorrhea. Dent Cosmos,
1917;59:405-414.
2. Box HK. Experimental traumatogenic occlusion in sheep.
Oral Health 1935;29:9-15.
3. Löe H, Theilade E, Jensen S. Experimental gingivitis in
man. J Periodontol 1965;36:177-187.
4. Glickman I, Smulow JB. Alterations in the pathway of
gingival inflammation into the underlying tissues induced
by excessive occlusal forces. J Periodontol 1962;33:7-
13.
5. Glickman I, Smulow JB. Effect of excessive occlusal
forces upon the pathway of gingival inflammation in
humans. J Periodontol 1965;36:141-147.
6. Glickman I, Smulow JB. Further observations on the
effects of trauma from occlusion. J Periodontol 1967;38:
280-293.
7. Waerhaug J. The infrabony pocket and its relationship
to trauma from occlusion and subgingival plaque. J Peri-
odontol 1979;50:355-365.
8. Waerhaug J. The angular bone defect and its rela-
tionship to trauma from occlusion and down growth
of the subgingival plaque. J Clin Periodontol 1976;3:
110-122.
9. Svanberg G, Lindhe J. Experimental tooth hypermobil-
ity in the dog. Odontol Revy 1973;24:269-282.
10. Svanberg G, Lindhe J. Vascular reactions in the peri-
odontal ligament incident to trauma from occlusion. J
Clin Periodontol 1974;1:58-69.
11. Svanberg G. Influence from trauma from occlusion on
the periodontium of dogs with normal or inflamed gin-
giva. Odontol Revy 1974;25:165-178.
12. Lindhe J, Svanberg G. Influence of trauma from occlu-
sion on the progression of experimental periodontitis in
the beagle dog. J Clin Periodontol 1974;1:3-14.
13. Lindhe J, Ericsson I. Influence of trauma from occlusion
on reduced but healthy periodontal tissues in dogs. J
Clin Periodontol 1976;3:110-122.
14. Nyman S, Lindhe J, Ericsson I. The effect of progres-
sive tooth mobility on destructive periodontitis in the
dog. J Clin Periodontol 1978;5:213-225.
15. Ericsson I, Lindhe J. Effect of long standing jiggling on
experimental marginal periodontitis in the beagle dog.
J Clin Periodontol 1982;9:497-503.
16. Lindhe J, Ericsson I. The effect of elimination of jiggling
forces on periodontally exposed teeth in the dog. J Peri-
odontol 1982;53:562-567.
17. Ericsson I, Lindhe J. Lack of significance of increased
tooth mobility in experimental periodontitis. J Periodontol
1984;65:447-452.
18. Polson AM. Trauma and progression of marginal peri-
odontitis in squirrel monkeys. II. Co-destructive factors
of periodontitis and mechanically produced injury. J Peri-
odont Res 1974;9:108-113.
19. Polson A, Meitner S, Zander H. Trauma and progression
of marginal periodontitis in squirrel monkeys. III. Adap-
tation of interproximal alveolar bone to repetitive injury.
J Periodont Res 1976;11:179-289.
20. Polson A, Meitner S, Zander H. Trauma and progression
of marginal periodontitis in squirrel monkeys. IV.
Reversibility of bone loss due to trauma alone and
trauma superimposed upon periodontitis. J Periodont
493
0150_IPC_AAP_553281 4/3/01 1:18 PM Page 494

Relationship of Occlusal Discrepancies to Periodontal Disease Volume 72 • Number 4

Res 1976;11:290-297. 33. Hallmon W. Occlusal trauma: Effect and impact on the
21. Kantor M, Polson A, Zander H. Alveolar bone regener- periodontium. Ann Periodontol 1999;4:102-107.
ation after the removal of inflammatory and traumatic 34. Harrel S, Nunn M, Belling C. Long-term results of a min-
factors. J Periodontol 1976;47:687-685. imally invasive surgical approach for bone grafting. J
22. Polson A, Kantor M, Zander H. Periodontal repair after Periodontol 1999;70:1558-1563.
reduction of inflammation. J Periodont Res 1979;14:520- 35. McGuire MK, Nunn ME. Prognosis versus actual out-
525. come. III. The effectiveness of clinical parameters in
23. Perrier M, Polson A. The effect of progressive and accurately predicting tooth survival. J Periodontol 1996;
increasing tooth hypermobility on reduced but healthy 67;666-674.
periodontal supporting tissue. J Periodontol 1982;53:
152-157. Send reprint requests to: Dr. Stephen K. Harrel, 10246 Mid-
24. Polson A, Zander H. Effects of periodontal trauma on way Road, Suite 101, Dallas, TX 75229. Fax: 214/350-6383;
intrabony pockets. J Periodontol 1983;54:586-591. e-mail: skh1@airmail.net
25. Polson A, Adams R, Zander H. Osseous repair in the
presence of active tooth hypermobility. J Clin Periodontol Accepted for publication October 27, 2000.
1983;10:370-379
26. Polson A. The relative importance of plaque and occlu-
sion in periodontal disease. J Clin Periodontol 1986;13:
923-927.
27. Shefter G, McFall W. Occlusal relations and periodontal
status in human adults. J Periodontol 1984;55:368-374.
28. Pihlstrom B, Anderson K, Aeppli D, Schaffer E. Associ-
ation between signs of trauma from occlusion and peri-
odontitis. J Periodontol 1986;57:1-6.
29. Jin L, Cao C. Clinical diagnosis of trauma from occlu-
sion and its relation with severity of periodontitis. J Clin
Periodontol 1992;19:92-97.
30. Wang H, Burgett F, Shyr Y, Ramfjord S. The influence
of molar furcation involvement and mobility on future
clinical periodontal attachment loss. J Periodontol
1994;65:25-29.
31. Burgett F, Ramfjord S, Nissle R, et al. A randomized trial
of occlusal adjustment in the treatment of periodontitis
patients. J Clin Periodontol 1992;19:381-387.
32. Gher ME. Non-surgical pocket therapy: Dental occlu-
sion. Ann Periodontol 1996;1:567-580.

494