J Periodont Res 2003; 38; 318–323 Copyright Ó Blackwell Munksgaard Ltd

Printed in the UK. All rights reserved

JOURNAL OF PERIODONTAL RESEARCH
ISSN 0022-3484

J. Slots1, J.J. Kamma2, C. Sugar3

The herpesvirus– 1
University of Southern California, School of
Dentistry, Los Angeles, CA, USA, 2Private
practice, Athens, Greece, and 3Department of

Porphyromonas gingivalis– Information and Operations Management,

Marshall School of Business, University of
Southern California, Los Angeles, CA, USA

periodontitis axis
Slots J, Kamma JJ, Sugar C. The herpesvirus–Porphyromonas gingivalis–perio-
dontitis axis. J Periodont Res 2003; 38; 318–323. Ó Blackwell Munksgaard, 2003

Objectives and background: Members of the herpesvirus family have accumulated

considerable support for a role in severe types of periodontitis. This study aimed to
examine whether human cytomegalovirus (HCMV), Epstein–Barr virus type 1
(EBV-1) or herpes simplex virus (HSV) together with the major periodontopathic
bacterium Porphyromonas gingivalis might interact in the pathogenesis of perio-
dontal breakdown.
Methods: Sixteen subjects each contributed paper point samples from two pro-
gressing and two stable periodontitis lesions, as determined by ongoing loss of
probing attachment. Polymerase chain reaction methodology was used to identify
subgingival herpesviruses, P. gingivalis and other bacterial pathogens. Chi-
squared tests and multivariate logistic regression were employed to identify sta-
tistical associations between herpesviruses, periodontopathic bacteria and clinical
variables.

Results: HCMV and HSV were both significant predictors of the presence of
subgingival P. gingivalis. In turn, P. gingivalis was positively associated with
periodontitis active disease, probing attachment level, probing pocket depth,
gingival bleeding upon probing and patient age. EBV-1 was not linked to
P. gingivalis, although the virus was predictive of periodontitis active disease. The Jørgen Slots, DDS, DMD, PhD, MS, MBA,
University of Southern California, School of
periodontitis disease risk associated with herpesvirus–P. gingivalis combinations Dentistry, MC-0641, Los Angeles,
depended on both site-specific and subject-specific factors. CA 90089–0641, USA
e-mail: jslots@usc.edu
Conclusion: The present data of aggressive periodontitis implicate HCMV, HSV Key words: cytomegalovirus; herpes simplex
and P. gingivalis as either cofactors in its etiology or triggers of relapses. Further virus; Porphyromonas gingivalis; alveolar bone;
studies are needed to determine the spectrum of periodontopathogenicity of periodontal disease; pathogenesis
herpesviruses and effective management of these viruses in periodontal sites. Accepted for publication September 5, 2002

Despite considerable research efforts,
several aspects of the etiology and
pathogenesis of periodontitis remain
unclear, in part due to difficulty in
diagnosis, variability in the clinical
course of the disease and insufficient
knowledge about the infectious
disease process. One of the more
intractable questions in periodontology
concerns the events that trigger con-
version from gingivitis to periodonti-
tis or from stable to disease-active
periodontitis. Attempts to delineate
the etio-pathogenesis of periodontitis
have been hampered by the reality
that the clinical diagnosis of perio-
dontitis includes a multiplicity of
disorders with different subsets of
pathogenic determinants variably
interacting with each other and the
host defense. Another complicating
factor is that the infectious agents in
periodontitis may only give rise to
destructive disease in a subset of
genetically or immunologically sus-
ceptible individuals.
It is generally accepted that perio-
dontal breakdown in most individuals
is closely related to a relatively small
group of bacteria, most notably
Porphyromonas gingivalis (1). P. gingi-
valis occurs in elevated proportions in
progressive periodontal lesions (2) and
possesses numerous virulence factors
pertinent to periodontitis (3). Further-
more, a growing body of evidence
suggests that periodontal infection
with human cytomegalovirus (HCMV)
and Epstein–Barr virus type 1 (EBV-1),

both herpesviruses, are risk factors for
periodontitis (4, 5). HCMV and EBV-1
DNA sequences are frequently present
in gingival tissue and gingival crevice
fluid of periodontitis lesions, and
accelerated periodontal breakdown is
more common in periodontal sites
exposed to multiple herpesvirus infec-
tions or to HCMV active infection
than in those that show no such
exposures (5). Also, herpesviruses have
attracted interest because their ability
to cause latent infections that period-
ically reactivate has some similarity
with the relapsing-remitting course of
periodontitis. The display of tissue
tropism of herpesvirus infections is
also in accordance with the segmental
pattern of breakdown in most perio-
dontitis patients.
Slots and Contreras (6) proposed that
various aggressive types of periodontal
disease develop as a result of a series of
interactions among herpesviruses, bac-
teria and host immune reactions. Pre-
sumably, herpesviruses residing in
inflammatory cells enter gingival tissue
with the development of gingivitis.
Subsequent herpesvirus reactivation
then diminishes the resistance of perio-
dontal tissue, leading to overgrowth of
pathogenic bacteria and release of
cytokines and chemokines from macr-
ophages and other host cells. Consistent
with this hypothesis, we have demon-
strated an association among peri-
odontal HCMV, the putative
periodontal pathogen Dialister pneu-
mosintes, and alveolar bone loss (7). In
the present study, we examined whether
periodontal HCMV, EBV-1 or herpes
simplex virus (HSV) were linked to the
presence of subgingival P. gingivalis
and destructive periodontal disease.
Materials and methods
The patients, clinical examination
procedures, virological and bacterial
identification methods and statistical
techniques used in this study have been
described previously (7, 8). Also the
major clinical and microbiological
findings are presented elsewhere (8).
Briefly, 16 adult subjects from Greece,
age 33.1 ± 2.6 years, with aggressive
periodontitis were studied. Each of the
16 patients contributed subgingival
paper-point samples from two perio-
dontitis lesions of 5.9 ± 0.8 mm aver-
age probing depths that had
experienced at least 2 mm loss of clin-
ical attachment during the preceding
3–6 months, and from two periodon-
titis lesions of 5.2 ± 1.0 mm average
probing depths that had remained
stable during the same time period.
Infectious agent identifications were
performed without knowledge of the
clinical status of the sample sites.
HCMV, EBV-1 and HSV were
identified using nested polymerase
chain reaction (PCR) and major
periodontopathic bacteria including
P. gingivalis, D. pneumosintes, Bacter-
oides forsythus and Actinobacillus
actinomycetemcomitans using PCR
amplification of signature sequences of
bacterial 16S rRNA genes. Pearson
chi-squared tests of independence,
confirmed by Fisher exact tests, were
used to determine whether any of the
four test bacteria and any of the three
test viruses were associated with each
other. Logistic regression was em-
ployed to determine whether individual
viruses, combinations of viruses, or
other covariates were, on their own,
associated with periodontal presence of
P. gingivalis. Finally, stepwise logistic
regression was used to identify the best
multivariate model for predicting the
presence of subgingival P. gingivalis
and periodontitis active disease with
various clinical variables. Because of
multiple study sites per individual, it
was important statistically to identify a
possible subject effect. By accounting
for subject specific variability in logis-
tic modeling techniques, that normally
assume independence for all response
measurements, the power to detect the
Table 1. Statistical relationship between periodontal herpesviruses and clinical variablesa
Item
Periodontitis disease activity
Probing attachment level
Probing pocket depth
Average % alveolar bone loss
Total number of teeth
Suppuration
Gender (female)
a
Gingival bleeding upon probing, cigarette smoking and patient age were not significantly
associated with any herpesvirus tested.
*P-value.
Herpesvirus–P. gingivalis interaction 319
significance of various factors in pro-
moting the presence of P. gingivalis
could be increased. However, the
simplest approach to including a sub-
ject effect, adding dummy variables for
each of the patients, greatly increases
the number of parameters required to
fit the model and may lead to insta-
bility and biases in the estimates of
some of the regression coefficients,
especially in studies having relatively
few observations. For these reasons,
the results of our logistic regression
models are presented both with and
without subject effects and compared
for consistency.
Results
Interactions among the three test her-
pesviruses and among the four test
bacteria were studied using chi-square
tests of independence. The presence of
HCMV was associated with the pres-
ence of HSV (P ¼ 0.01), and the
presence of EBV-1 was associated with
the presence of HSV (P ¼ 0.04). No
significant relationship was found
between HCMV and EBV-1. Also, no
evidence of significant relationships
among the four test bacteria was
detected.
Table 1 shows associations between
the individual herpesviruses and clin-
ical variables. All three study viruses
were positively associated with probing
pocket depth. In addition, HCMV was
associated with periodontitis disease
activity as determined by change in
probing attachment level, with average
percentage of alveolar bone loss of all
teeth and with female gender. EBV-1
was linked with probing attachment
level and with total number of teeth,
HCMV EBV-1 HSV
0.0003* 0.08 0.02
0.12 0.04 0.002
0.04 0.03 0.02
0.03 0.39 0.08
0.47 0.03 0.94
0.26 0.03 0.32
0.04 0.11 0.49
320 Slots et al.

Table 2. Statistical relationship between periodontal herpesviruses and P. gingivalis

Chi-squared Logistic Logistic regression, Logistic regression Logistic regression with

tests, regression, P-value for with subject effect, subject effect, P-value for
Viruses P-value overall P-value individual variables overall P-value individual variables

HCMV 0.07 0.07 0.08 < 0.0001 0.01

EBV-1 0.52 0.52 0.52 0.0007 0.15
HSV 0.04 0.04 0.05 0.0002 0.03
HCMV + EBV-1 NA 0.18 0.09 (HCMV) 0.0001 0.02 (HCMV)
0.74 (EBV-1) 0.43 (EBV-1)
HCMV + HSV NA 0.05 0.21 (HCMV) < 0.0001 0.04 (HCMV)
0.12 (HSV) 0.14 (HSV)
EBV-1 + HSV NA 0.11 0.89 (EBV-1) 0.0003 0.96 (EBV-1)
0.06 (HSV) 0.07 (HSV)
HCMV + EBV-1 + HSV NA 0.11 0.21 (HCMV) < 0.0001 0.18 (HCMV)
0.996 (EBV-1) 0.77 (EBV-1)
0.12 (HSV) 0.04 (HSV)

NA: not applicable.

and HSV with periodontitis disease
activity and probing attachment level.
Multiple study sites in each subject
enabled us to assess whether the dis-
tribution of P. gingivalis among sub-
gingival sites of each subject occurred
at random or was subject dependent. A
preliminary chi-squared test of inde-
pendence provided evidence of a sub-
ject relationship with a P-value of 0.02.
However, because of the large number
of patients relative to the number of
measurements per subject, which may
make the assumption of a chi-squared
distribution unreliable, a simulation
study was performed to estimate the
actual null distribution of the chi-
squared statistic under an assumption
of independence. The simulation study
indicated a subject dependency with a
similar level of significance (P ¼ 0.02).
The suggested subject effect pointed to
the desirability of including subject-
specific indicators in the overall logistic
model.
First, logistic regression analysis was
employed to determine whether any of
the herpesviruses, either singly or in
combination, were predictive of the
presence of P. gingivalis. As seen in
Table 2, when testing was done with-
out subject effect, HSV was mildly
significant and HCMV was borderline
significant in being associated with the
presence of subgingival P. gingivalis.
However, when including subject
effect, HCMV appeared just as signifi-
cant as HSV, if not more so, and each ence for categorical variables and
of the two viruses appeared much more logistic regression for continuous vari-
significant than when the subject effects ables, the strongest relationships with
were excluded. In models including P. gingivalis were found with probing
combinations of viruses, whether attachment level, periodontitis disease
HCMV or HSV remained significant activity, patient age, gingival bleeding
depended on whether or not subject upon probing and probing pocket
effects were included, but in no case depth. No significant univariate
was it beneficial to include more than association with P. gingivalis was
one of the two viruses as a predictor. found for average percentage of
This was not surprising given our ear- alveolar bone loss, presence of sup-
lier finding that the presence of HCMV puration, cigarette smoking, patient
was highly correlated with that of age, gender or total number of teeth.
HSV. EBV-1 was a poor predictor in Stepwise logistic regression was then
all models, suggesting that the virus is employed to identify the best multiva-
not predictive of the presence of sub- riate models for predicting the presence
gingival P. gingivalis. of subgingival P. gingivalis. Initial
Univariate analyses were performed testing was performed without using
to delineate clinical variables that were dummy variables for subject effects.
associated with the presence of sub- This allowed us to include covariates
gingival P. gingivalis (Table 3). Using such as patient age, gender, cigarette
Pearson chi-squared test of independ- smoking, and total number of teeth,
Table 3. Statistical relationship between subgingival P. gingivalis and clinical variablesa
P-value of Overall
Pearson P-value of Individual variable
chi-squared tests logistic P-value for
Covariates of independence regression logistic regression
Periodontitis disease activity 0.006 0.005 0.007
Probing attachment level NA 0.003 0.007
Probing pocket depth NA 0.03 0.04
Gingival bleeding upon probing 0.03 0.03 0.03
Patient age NA 0.02 0.03
a
Subgingival P. gingivalis was not significantly associated with average percentage of alveolar
bone loss, suppuration, cigarette smoking, patient age, gender or total number of teeth.
NA: not applicable.

which are constant within patients and
would therefore be confounded with a
subject effect. Periodontitis disease
status, average percentage of alveolar
bone loss and probing attachment level
were not included in these analyses
since they are essentially outcomes
rather than predictors. The analysis
resulted in a final model containing the
variables HSV (P ¼ 0.01), age
(P ¼ 0.008) and cigarette smoking
(P ¼ 0.05) with an overall P-value of
0.001. Replacing HSV with HCMV
gave a similar result, which is in
agreement with the high correlation
between the two viruses. Next we fit
models including dummy variables for
the individual patients in an attempt to
account for correlations among the
multiple site measurements. Possible
predictors included presence of the
viruses, probing pocket depth, pres-
ence of suppuration and gingival
bleeding upon probing. Not included
were periodontitis disease status and
probing attachment level, which are
outcomes, and the other covariates
because they were confounded with the
dummy variables. The analysis resulted
in a final model consisting of subject
effects plus just HCMV (overall
P-value ¼ 0.0001, effect P-value for
HCMV ¼ 0.01). When excluding
HCMV as a possible predictor, the two
best models were subject effects plus
HSV alone (overall P-value ¼ 0.0002,
effect P-value for HSV ¼ 0.03) and
subject effects plus just probing
attachment level (overall P-value ¼
0.0001, effect P-value for probing
attachment level ¼ 0.03), though the
latter is really an outcome, not a pre-
dictor. Since HCMV and HSV were
both useful predictors of the presence
of subgingival P. gingivalis and since
they were highly correlated with each
other, it was not worthwhile to include
both viruses in the final model. In all
three of these models the overall
P-value was lower than in the models
without subject effects, although the
analyses both with and without subject
effect were consistent in their basic
conclusions. However, because the co-
efficients of the dummy variables were
unstable, some of the estimates may
have been biased and the exact P-val-
ues must be interpreted with caution.
Herpesvirus–P. gingivalis interaction
We finally tested multivariable
models having periodontitis disease
activity status as the response. When
considering all infectious agents
as possible predictors, the final
model included P. gingivalis (P ¼
0.005), D. pneumosintes (P ¼ 0.006),
A. actinomycetemcomitans (P ¼ 0.007),
HCMV (P ¼ 0.008) and EBV-1
(P ¼ 0.02). When considering
P. gingivalis as the sole bacteriological
predictor, the final model comprised
HCMV (P ¼ 0.001) and gingival
bleeding upon probing (P ¼ 0.005).
Apparently, gingival bleeding upon
probing substituted for test bacteria in
terms of disease relationship.
Discussion
The present study demonstrated signi-
ficant associations among HCMV,
P. gingivalis and progressive perio-
dontitis. Even though each subject
provided relatively few measurements,
the fact that all statistical analyses
showed significant associations among
HCMV–P. gingivalis–active periodon-
titis supports the notion that HCMV
and P. gingivalis serve as cofactors in
periodontal breakdown. Contreras
et al. (9) also found HCMV to be as-
sociated with P. gingivalis and severe
periodontitis.
In contrast to previous findings in
US patients (9), we showed a signifi-
cant link between HSV and P. gingi-
valis or periodontitis. The prevalence
of herpesviruses varies in different
geographical areas of the world and
among different socio-economical
groups (10, 11), and Greek and US
patients may differ with respect to type
and frequency of herpesvirus perio-
dontal infections. In any event, the
present findings pointed to the poten-
tial importance of HSV in the patho-
genesis of some periodontitis lesions.
This study revealed no significant
relationship between EBV-1 and
P. gingivalis. However, it is worth
noting that while EBV-1 was not pre-
dictive of the presence of subgingival
P. gingivalis, the virus was correlated
with periodontitis disease activity,
suggesting that EBV-1 has a promo-
tional role in periodontitis, probably
through mechanisms involving perio-
321
dontopathic species other than P. gin-
givalis. Contreras et al. (9) reported
periodontal EBV-1 to be linked to
P. gingivalis and severe periodontitis
but did not resolve the question of
possible interactions among infecting
herpesviruses.
In the study of Contreras et al. (9),
HCMV periodontal presence was
associated with severe adult periodon-
titis with an odds ratio of 4.7 and
constellations of P. gingivalis and
other periodontal bacteria were asso-
ciated with severe adult periodontitis
with odds ratios of 2.6–3.2. That a
synergistic periodontopathogenic rela-
tionship might exist between HCMV
and P. gingivalis was implied by
Michalowicz et al. (12), who demon-
strated an odds ratio as high as 51.4 for
the association of the combined sub-
gingival occurrence of HCMV and
P. gingivalis with juvenile periodontitis
but individual odds of only 4.6 and 7.8
for the association of, respectively,
HCMV and P. gingivalis with the dis-
ease. A cooperative pathogenic inter-
action between a murine strain of
CMV and P. gingivalis has also been
observed in experimental mice (13).
When infected with P. gingivalis prior
to infection with CMV, the mice ex-
hibited higher mortality rates than mice
infected with P. gingivalis subsequent
to CMV infection, suggesting preexist-
ing P. gingivalis infection increased the
pathogenicity of CMV (13). Most like-
ly, the interaction between HCMV and
P. gingivalis is bi-directional, with
HCMV reactivation suppressing host
defenses and permitting overgrowth of
P. gingivalis, and with P. gingivalis
enzymes and other inflammatory-
inducing products having the potential
to activate periodontal HCMV (the vi-
cious circle concept). Available know-
ledge seems to support the concept that
periodontitis occurs more frequently
and progresses more rapidly in perio-
dontal sites that harbor a combination
of HCMV and P. gingivalis.
A variety of mechanisms have been
suggested by which HCMV may con-
tribute to periodontal breakdown (14).
Herpesviruses show tropism for cells of
the immune system and HCMV resides
in periodontal macrophages and
T-lymphocytes (15), possibly causing
322 Slots et al.
alterations in immune functions of the
periodontium. Activation of macroph-
ages and T-lymphocytes leads to
increased production of the pro-
inflammatory cytokines interleukin-1b
and tumor necrosis factor-a as well as
prostaglandin E2 (14), which have each
been associated with destructive peri-
odontal disease (16). Furthermore,
Ongradi et al. (17) found that phago-
cytic and bactericidal capacities of
periodontal neutrophils, cells of key
importance in the periodontal defense
(18), were significantly impaired in
subjects who carried herpesviruses in
oral lymphocytes and epithelial cells,
as compared to virus-free persons.
HCMV has also devised numerous
mechanisms of escaping immune sur-
veillance and so can remain latent in
the host. The HCMV genome encodes
genes involved in down-regulating
surface expression of HLA class I
molecules, it sequesters CC chemok-
ines, induces Fc receptors, interferes
with induction of HLA class II anti-
gens, and can inhibit natural killer cell
activity (19). Taken together, HCMV-
mediated alterations of immune func-
tions may at times lead to a state of
immunosuppression or may contribute
to immunopathologic reactions in the
periodontium.
The chi-squared tests of independ-
ence and the simulation analysis sug-
gested that the distribution of
P. gingivalis among subgingival sites
did not occur at random but was sub-
ject dependent, meaning the response
values (presence or absence of P. gin-
givalis) were correlated within patients.
Also, the fact that including subject
effect markedly decreased P-values of
all models argued for the presence of a
considerable subject effect in the
HCMV–P. gingivalis–active periodon-
titis interrelationship, although exact
significance levels need to be inter-
preted with caution due to the small
sample size and instability in some of
the parameter estimates. In compar-
ison, D. pneumosintes, another puta-
tive periodontopathic bacterium,
revealed no evidence of subject specif-
icity in its intraoral distribution (7). A
subject effect could be due to inherent
or environmentally determined differ-
ences in susceptibility to P. gingivalis
among individuals, or to the possibility
that once a patient has a single infected
site other sites are more likely to
become infected. The observed subject
effect may mean that changes in certain
host factors are capable of enhancing
the periodontopathogenicity of
HCMV or P. gingivalis. Indeed, fac-
tors that impair the host defense have
the potential to reactivate latent
HCMV infections and cause increased
viral pathogenicity (6). Also, an
impaired host may not be able to
defend against P. gingivalis or other
periodontopathic bacteria as effectively
as a normal host (3).
One of the greatest challenges in
confirming or refuting a role for
HCMV or other herpesviruses in
human periodontitis is their ubiquitous
nature and the relatively rare occur-
rence of progressive periodontitis. This
dilemma is apparent not only for per-
iodontitis but also for the expanding
spectrum of human diseases with
which HCMV has been associated
(20). We hypothesize that periodontal
breakdown has a polymicrobial caus-
ation and depends upon the simulta-
neous occurrence of a number of
infectious disease events, including at
least (i) herpesvirus presence at perio-
dontal sites, (ii) reactivation of latent
periodontal herpesviruses, (iii) inad-
equate antiviral cytotoxic T-lympho-
cyte response, (iv) presence of specific
pathogenic bacteria, and (v) insuffi-
cient levels of protective antibacterial
antibodies. Presumably, the patho-
genic determinants of periodontitis
cooperate with each other in destruc-
tive constellations relatively in-
frequently and primarily during
periods of impaired host defense. Also,
the pathogenic determinants have to
interact for a sufficiently long period of
time to produce periodontal break-
down. Reactivation of herpesviruses at
periodontal sites may constitute a
particularly important pathogenic
event (6). Of interest, various factors
that have shown to enhance the risk
for periodontitis also have the poten-
tial to activate herpesviruses (6). The
ability of tobacco products to interact
with and possibly reactivate perio-
dontal herpesviruses (21) may partly
explain the increased risk for
periodontitis from tobacco usage (22).
The present data also pointed to a re-
lationship between smoking habits and
the occurrence of subgingival P. gin-
givalis, which provides another poss-
ible explanation for the observed link
between tobacco smoking and perio-
dontitis. However, due to the lack of
an appropriate animal model, insight
into the pathogenesis of periodontal
infections has primarily been derived
from examination of periodontal
patients at different disease stages and
not from a dynamic model for a
detailed step-wise dissection of the
infectious and molecular determinants
of disease progression. Most likely, the
definitive proof of a causal role of
HCMV or other herpesviruses in
human periodontal disease will have to
await the development of effective
antiherpesviral vaccines.
In conclusion, the present findings
support the concept that HCMV and
probably HSV are involved in the etio-
pathogenesis of some types of aggres-
sive periodontitis. They also suggest
that the two herpesviruses cooperate
with P. gingivalis in the destruction of
periodontal tissues. Further studies are
needed to investigate the mechanisms
by which a combined herpesvirus–
P. gingivalis periodontal infection may
lead to periodontitis, including an
analysis of the involvement of cytok-
ines and other inflammatory mediators
released from mammalian cells in
response to herpesvirus infections of
the periodontium.
References
1. Slots J. Actinobacillus actinomycetemcom-
itans and Porphyromonas gingivalis in
periodontal disease: introduction. Period-
ontol 2000 1999;20:7–13.
2. Slots J, Ting M. Actinobacillus actin-
omycetemcomitans and Porphyromonas
gingivalis in human periodontal disease:
occurrence and treatment. Periodontol
2000 1999;20:82–121.
3. Holt SC, Kesavalu L, Walker S, Genco
CA. Virulence factors of Porphyromonas
gingivalis. Periodontol 2000 1999;20:168–
238.
4. Slots J. Interactions between herpesviruses
and bacteria in human periodontal dis-
ease. In: Brogden KA, Guthmiller JM,
eds. Polymicrobial diseases. Washington,
DC: ASM Press, 2002:317–331.

5. Kamma JJ, Slots J. Herpesviral–bacterial
interactions in aggressive periodontitis.
J Clin Periodontol 2003; (in press).
6. Slots J, Contreras A. Herpesviruses: a
unifying causative factor in periodontitis?
Oral Microbiol Immunol 2000;15:277–
280.
7. Slots J, Sugar C, Kamma JJ. Cytomega-
lovirus periodontal presence is associated
with subgingival Dialister pneumosintes
and alveolar bone loss. Oral Microbiol
Immunol 2002;17:369–374.
8. Kamma JJ, Contreras A, Slots J. Herpes
viruses and periodontopathic bacteria in
early-onset periodontitis. J Clin Period-
ontol 2001;28:879–885.
9. Contreras A, Umeda M, Chen C, Bakker
I, Morrison JL, Slots J. Relationship
between herpesviruses and adult perio-
dontitis and periodontopathic bacteria.
J Periodontol 1999;70:478–484.
10. de Jong MD, Galasso GJ, Gazzard B et al.
Summary of the II International Sympo-
sium on Cytomegalovirus. Antiviral Res
1998;39:141–162.
11. Glaser SL, Lin RJ, Stewart SL et al.
Epstein–Barr virus-associated Hodgkin’s
Herpesvirus–P. gingivalis interaction
disease: epidemiologic characteristics
in international data. Int J Cancer
1997;70:375–382.
12. Michalowicz BS, Ronderos M, Camara-
Silva R, Contreras A, Slots J. Human
herpesviruses and Porphyromonas gingi-
valis are associated with juvenile perio-
dontitis. J Periodontol 2000;71:981–988.
13. Stern J, Shai E, Halabi A, Houri-Haddad
Y, Shapira L, Palmon A. Interaction
between murine cytomegalovirus and a
periodontal pathogen Porphyromonas
gingivalis in a mouse model. J Dent Res
2001;80:1314 (Abstract no. 46).
14. Contreras A, Slots J. Herpesviruses in
human periodontal disease. J Periodont
Res 2000;35:3–16.
15. Contreras A, Zadeh HH, Nowzari H,
Slots J. Herpesvirus infection of inflam-
matory cells in human periodontitis. Oral
Microbiol Immunol 1999;14:206–212.
16. Page RC, Offenbacher S, Schroeder HE,
Seymour GJ, Kornman KS. Advances in
the pathogenesis of periodontitis: sum-
mary of developments, clinical implica-
tions and future directions. Periodontol
2000 1997;14:216–248.
323
17. Ongradi J, Sallay K, Kulcsar G. The
decreased antibacterial activity of oral
polymorphonuclear leukocytes coincides
with the occurrence of virus-carrying oral
lymphocytes and epithelial cells. Folia
Microbiol (Praha) 1987;32:438–447.
18. Hart TC, Shapira L, Van Dyke TE.
Neutrophil defects as risk factors for
periodontal diseases. J Periodontol
1994;65:521–529.
19. Michelson S. Human cytomegalovirus
escape from immune detection. Intervi-
rology 1999;42:301–307.
20. Griffiths PD. The 2001 Garrod lecture.
The treatment of cytomegalovirus
infection. J Antimicrob Chemother
2002;49:243–253.
21. Park NH, Herbosa EG, Sapp JP. Effect of
tar condensate from smoking tobacco and
water-extract of snuff on the oral mucosa
of mice with latent herpes simplex virus.
Arch Oral Biol 1987;32:47–53.
22. Albandar JM. Global risk factors and risk
indicators for periodontal diseases. Peri-
odontol 2000 2002;29:177–206.