You are on page 1of 19

JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No.

1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

The Clinical, Microbial, and Host Response


Characteristics of the Failing Implant
Jeanne M. Salcetti, DDS, MS/John D. Moriarty, DDS, MS/Lyndon F.Cooper , DDS,
PhD/Frances W. Smith, MS/John G. Collins , DMD/Sigmund S. Socransky, DDS/Steven
Offenbacher, DDS, PhD

The goal of this study was to provide new data regarding levels of inflammatory and growth factor
mediators and bacterial pathogens associated with failing implants, as compared to healthy
implants. Twenty-one patients with failing implant sites (group 1) and 8 patients with only healthy
implants (group 2) were included. Fifteen of the 21 failing implant patients (group 1) also
presented with at least one stable nondiseased implant. Plaque samples were examined, using DNA
oligonucleotide probes for 40 different microbes. Gingival crevicular fluid samples were collected
for the analyses of catabolic bone resorbing agonists prostaglandin E 2 (PGE2), interleukin-1ß
(IL-1ß) and IL-6 and anabolic bone-forming growth factors transforming growth factor ß (TGF-ß)
and platelet-derived growth factor (PDGF). Although positive trends were noted, there were no
significant differences in any of the microbial, inflammatory, or growth factors mediators
comparing failing to stable implants in group 1. This study found greater detection frequencies of
P. nigrescens, P. micros, F. nucleatum ss vincentii, and F. nucleatum ss nucleatum , as well as
significant elevations in GCF levels of PGE 2, IL-1ß, and PDGF in mouths with failing implant sites
as compared to mouths with healthy control implants. Risk appears to be primarily at a patient
level and secondarily at a site or implant level from a clinical, microbial (P. micros and P.
nigrescens), and biochemical (PGE2 and IL-1ß) perspective. Furthermore, the counts of P.
nigrescens and P. micros were found to correlate with concentrations of PGE2 at a site level.
(INT J ORAL MAXILLOFAC IMPLANTS 1997;12:32–42)
Key words: bacterial pathogen, failing implant, growth factor, inflammatory mediator, peri-implantitis

F or more than 30 years, the placement and restoration of endosseous dental implants has been shown to
be a successful treatment alternative for edentulous or partially edentulous patients.1-8 Despite the
long-term success of osseointegrated implants, clinicians report the occasional incidence of pathogenic
complications associated with the maintenance and retention of implants. Changes in clinical signs such
as soft tissue inflammation, bleeding on probing, suppuration, pain, increased probing depth, and
radiographic evidence of bone loss,9 as well as microbiologic alterations in the flora, are similar to those
signs associated with periodontal disease. The progressive loss of peri-implant bone, as well as soft tissue
inflammatory changes, is defined as peri-implantitis.10
The etiologic factors associated with peri-implantitis are reported to be bacterial infection and
biomechanical factors associated with occlusal overloading.1,3,11-13 Numerous studies have evaluated
the microbiota around implants associated with states of health and disease. The primary finding is that a
flora composed of gram-positive cocci and rods is seen associated with stable, nondiseased implants,
while a more gram-negative, anaerobic flora with high levels of spirochetes is associated with failing or
failed implants.12-23 The periodontopathogens commonly identified are P. intermedia and P. gingivalis.

Article Text 1
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

Others have reported findings of Actinobacillus actinomycetemcomitans (Aa) micros, Fusobacterium


spp, C. rectus and Pseudomonas aeruginosa. From these studies, evidence supports the concept that the
microbiota associated with stable and failing implants is similar to the microbiota of periodontally
healthy and diseased teeth, respectively. Less is known regarding whether the host response to this
bacterial challenge in peri-implantitis is similar to that of periodontitis.
Although extensive research has been done in the area of periodontal inflammatory mediators, few
have studied the biologic response of peri-implantitis. Several facts have now been established regarding
the role of the host immune response in the pathogenesis of periodontal diseases. Inflammatory
mediators, such as prostaglandin E2 (PGE2), interleukin-1ß (IL-1ß), and possibly interleukin-6 (IL-6)
produced by the chronic inflammatory cells of the periodontal tissues initiate pathways that stimulate
osteoclastic bone resorption.24 Offenbacher et al25 demonstrated that there was an increased level of
gingival crevicular fluid (GCF) PGE2 preceding an attachment loss episode in periodontitis patients who
had undergone initial therapy and subsequent monitoring. A maximum level of PGE 2 occurred at sites
undergoing attachment loss, and levels subsided following treatment. Honig et al26 found IL-1ß levels to
be higher in the gingivae of adult periodontitis patients than in clinically healthy control subjects. These
levels are also higher in active periodontitis sites than in inflamed stable sites.27
In the implant literature, very few studies have evaluated the presence and level of inflammatory
mediators associated with peri-implantitis. Jovanovic et al28 found elevated levels of PGE2 associated
with disease progression around implants, and Kao et al29 found gingival crevicular fluid IL-1ß levels of
diseased implants to be elevated three-fold as compared to clinically healthy sites. Recent advances in
understanding of the cellular events involved not only in the resorption process, but the repair and
regeneration process as well, indicate that polypeptide growth factors, such as transforming growth factor
ß (TGF-ß) and platelet-derived growth factors (PDGF) are key feedback inhibitors of bone resorption.
During bone resorption, these mediators are released from bone where they are deposited during bone
synthesis. The release of these anabolic peptides, in turn, tends to turn off bone resorption and stimulate
de novo bone formation.30 Thus, the normal turnover of bone in peri-implantitis may be in an unstable
state of equilibrium because of the relative imbalance of catabolic (ie, PGE2, IL-1ß) versus anabolic (ie,
TGF-ß and PDGF) local hormones.
The purpose of the present study was to characterize the clinical, microbial, and host response
associated with failing implants. The ultimate research goal is to further the understanding of the
pathogenesis of implant failure so as to possibly predict a patient or implant at risk for future failure, thus
providing an opportunity for therapeutic measures to slow down or stop the disease process.
Materials and Methods
Patient Selection. The patient population was obtained from referrals by the private practices of general
dentists, oral and maxillofacial surgeons, and periodontists. Some of these patients were screened and
evaluated outside of the university research center, in the private office of Dr Crawford A. Tatum, Jr,
Opilika, AL. Each patient signed an informed consent form that had been reviewed and approved by the
Committee for Use of Human Subjects in Investigations at the University of North Carolina School of
Dentistry. Patients in group 1 had one or more failing implants. Any mobile implants were considered
failed implants and were not considered in the study. The inclusion criteria for patients with failing
implants were as follows:

Article Text 2
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

1. Patient had at least one root-form implant that had been in function for at least 1 year.
2. Patient with one or more failing implants must have presented with evidence of peri-implant
radiolucency and/or vertical crestal bone loss greater than 2.0 mm after the first year of service.
3. Patients with one or more failing implants may have had at least one other stable implant that showed
no clinical signs of inflammation or radiographic evidence of crestal bone loss.
The exclusion criteria for the study were as follows:
1. Patient had taken any antibiotic within 6 weeks prior to the clinical examination and biologic
sampling.
2. Patient had taken any steroidal or nonsteroidal anti-inflammatory drug within 2 weeks prior to the
clinical examination and biologic sampling.
3. Patient had undergone any periodontal or peri-implant therapy within 21 days.
4. Patient has had a history of chronic illness, such as cardiovascular, hepatic, rheumatic fever, asthma,
diabetes, immune disorder, bleeding disorder, or any condition requiring prophylactic antibiotic
coverage.
5. Women who were pregnant or nursing.
Patients with one or more healthy implants (group 2) were also selected for inclusion into the study
as the control subjects. The same inclusion and exclusion criteria applied to these patients as well. Thus,
the three categories of patients seen were: (1) patients with only failing (diseased) implants (6 patients);
(2) patients with both failing and stable (nondiseased) implants (15 patients); and (3) patients with only
healthy implants and considered as the control subjects (8 subjects).
Clinical Measurements. The initial complete-mouth clinical measurements were for plaque and
redness (inflammation) by visual inspection. These scores were dicotomously evaluated (0 = absent, 1 =
present) at six sites per tooth and implant. Following these assessments, any supragingival plaque present
was removed around the implant where there was to be the collection of gingival crevicular fluid (GCF)
and subgingival plaque samples. Following the collection of the peri-implant GCF and the microbial
plaque sample, a complete-mouth recording of the probing depths was completed using the Florida Probe
(Florida Probe, Gainesville, FL) with a titanium insert. Clinical attachment levels were recorded with a
UNC15 probe from the cementoenamel junction on teeth and the suprastructure-abutment junction on
implants, where possible. Bleeding on probing was dicotomously recorded as present or absent. All
measurements and collection of GCF and microbial samples were completed by one investigator (JMS).
This investigator had been calibrated to a gold standard in the UNC Clinical Research Unit, as described
elsewhere.31
Gingival Crevicular Fluid Sampling. Gingival crevicular fluid samples were collected from failing,
stable, and control implants for analysis of the following mediators: PGE2; IL-1ß; IL-6; PDGF; and
TGF-ß. The tissues surrounding the implants were air dried, and supragingival plaque was removed.
Sterile periopaper strips were inserted into the sulcus surrounding the implant, and the peri-implant fluid
was collected. The volume was recorded on a Periotron 6000 (Periotron Flow, Amityville, NY; IDE
Intrastate Distributing, Newark, NJ). The strip was wrapped in aluminum foil and placed into a cryovial
for immediate storage in liquid nitrogen. Duplicate samples were collected for all mediators at each
implant at each site. Strips were consecutively placed around the circumference of the implant, so that

Article Text 3
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

one strip was placed facially, then lingually, then mesially, and finally distally until all samples were
collected. The order of mediator collection was as follows: PGE2; PDGF; IL-1ß; TGF-ß; and IL-6.
Sampling four sites twice provided first and second samples of GCF from each site, and these were each
assayed independently for the selected mediator for duplicate determinations. All samples were
transported and stored in liquid nitrogen vapor phase at –160ºC until analyses could be performed.
Microbial Sampling. Microbial sampling and analyses were performed following methods described
by Socransky et al.32 At each implant, a sterile plastic curet was gently inserted to the base of the
peri-implant sulcus, and the subgingival plaque was removed with one sweep of the curet. One sample
was collected for each implant to be included in the study. The sample was then placed into an individual
tube containing 0.1 mL of tromethamine EDTA (TE); 0.1 mL of 0.5-mol/L of sodium hydroxide was
added, and the tube was agitated. All samples were sent to Dr Sigmund Socransky for analysis by
checkerboard DNA–DNA hybridization technique to enumerate 40 organisms (including two
streptococcus).
Laboratory Procedures. All GCF and plaque analyses were performed by blinded technicians. All
samples were identified by accession numbers and merged with clinical data at the time of statistical
analysis.
Gingival Crevicular Fluid Analyses. Levels of PGE2 were determined by radioimmunoassay (RIA).
This procedure has been described previously by Offenbacher et al.25 Levels of IL-1ß, IL-6, PDGF, and
TGF-ß were determined by enzyme-linked immunosorbent assay (ELISA). An ELISA kit was purchased
for IL-1ß and IL-6 from Cistron Biotechnology (Pinebrook, NJ), and PDGF and TGF-ß kits were
purchased from R&D Systems (Minneapolis, MN); manufacturers’ directions were followed. The TGF-ß
samples were not acid treated because levels of free TGF-ß present in the GCF were sought. Gingival
crevicular fluid mediator concentrations are reported as the mean of duplicate samples assayed
independently for each mediator.
Microbial Analysis. The checkerboard DNA–DNA hybridization technique has been described by
Socransky et al.32 Plaque samples were probed for 40 microbes using 105 and 106 standards. A
representative checkerboard blot appeared, and the blot intensity for each organism was scored. Samples
were scored as 0, no signal detected; 1, equal to or less than 105 standard; 2, equal to approximately 105
standard; 3, equal to between 105 and 106 standard; 4, equal to approximately 106 standard; and 5, equal
to or greater than 106 standard. The DNA probes determined the presence of 40 different microbial
species (Table 1).
Data Analyses. Each variable was pooled for each implant category for each implant and then across
all three implant categories, based on case status: (1) healthy implants in healthy patients; (2) healthy
(referred to as stable) implants in patients with failing implants; and (3) failing implants. Mean values
were calculated for each of the mediator levels and clinical assessments, and detection frequency was
determined for each of the identified microbes. A P value was set at < .05 with the Bonferroni corrections
for multiple comparisons. A two-way analysis of variance was performed among and between each of the
categories.
Results
Patient Population. Twenty-one patients with failing implant sites (group 1) were included in the study.
Fifteen of these 21 patients presented with both a failing implant site, as well as at least one stable

Article Text 4
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

nondiseased implant site. There were eight patients who had only nondiseased implant sites (group 2).
These subjects were considered the control subjects. Of the 21 patients with failing implants in group 1,
14 were women (mean age 58.2 years, range 25 to 76), and 7 were men (mean age 71.1 years, range 61 to
77). Nine of the 14 women and 6 of the 7 men had additional implants that were considered stable. Of the
8 control implants in patients in group 2, 6 were in women (mean age 47.5 years, range 37 to 82) and 2
were in men (mean age 75 years, range 66 to 84).
Pooling all of the patients in the study, there was a total of 28 failing implant sites, 27 stable sites,
and 14 nondiseased, control sites. Failing and nondiseased implant sites were present in each of the four
quadrants of the mouth and therefore were not noted to be in only one arch. Although data were recorded
regarding the various root-form implant types and surfaces, there were no particular statistically
significant trends in the failing implant category.
Among the 15 patients with both stable and failing implants in group 1, three patients were
completely edentulous. In group 2, there were three completely edentulous patients as well. All of the
completely edentulous patients were restored with maxillary dentures and fixed detachable “high-water”
mandibular anterior frameworks. Any patient with mobile implants, ie, failed implants, were excluded
from the study. None of the clinical, microbial, or mediator values were significantly different for the
edentulous versus partially edentulous patients; therefore, these patients were not segregated for further
analyses.
Clinical Measurements. Table 2 presents the mean data for plaque, inflammation, and bleeding on
probing (BOP) scores of failing (group 1) versus control patients (group 2). Neither the mean plaque nor
BOP scores were significantly different; however, the mean inflammation scores of the failing sites (P =
.002) and stable sites (P = .019) in group 1 patients were significantly higher than the healthy control
sites in group 2.
The mean probing depths (PD) and clinical attachment levels (CAL) of failing (group 1) versus
control (group 2) patients are presented in Table 3. As noted, there was a statistically significant
difference in the mean PD of the group 1 complete-mouth sites (2.14 mm/site) and failing sites (3.79
mm/site) versus the group 2 complete-mouth sites (1.81 mm/site) and control sites (2.11 mm/site) at P =
.04 and P = .03, respectively. The mean CAL of the failing sites in group 1 (3.59 mm/site) was also
significantly different than the group 2 control sites (1.39 mm/site) at P = .003.
Gingival Crevicular Fluid Mediator Levels. Using commercially available assay techniques,
gingival crevicular fluid levels of IL-6 were detected in the group 2 control sites (59.7 ng/mL), and in
group 1 failing sites (48.4 ng/mL) and stable sites (44.8 ng/mL). However, these levels were not
significantly different among the three implant categories. The TGF-ß was not clinically detectable at the
sensitivity levels determined by the ELISA kit.
The mean levels of PGE2 were found to be significantly higher in the failing sites (74.2 ng/mL) and
stable sites (64.2 ng/mL) of group 1 patients versus the nondiseased control sites (36.1 ng/mL) of group 2
patients at P = .01 and P = .008, respectively (Fig 1). A similar trend was seen in the IL-1ß levels (Fig 2),
with the highest level among the failing sites (761.5 ng/mL) and stable sites (626.7 ng/mL) of group 1
patients versus the group 2 patients (255.4 ng/mL). These levels are also statistically significant at P =
.023 for the failing sites versus the control sites and P = .029 for the stable sites versus the control sites.
Figure 3 illustrates that the mean GCF PDGF levels were the highest as detected at the stable implant
sites (50.8 ng/mL). These GCF PDGF levels at both the stable and failing implant sites (33.3 ng/mL) in
group 1 patients were significantly higher than those of the nondiseased control sites (6.55 ng/mL) of

Article Text 5
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

group 2 at P = .01 and P = .041, respectively. There were no statistically significant within-mouth
differences in the mean PGE2, IL-1ß, or PDGF levels when the failing implant sites were compared to
the stable implant sites in group 1 patients.
Microbial Analyses. Of the 40 different microorganisms evaluated, based on detection frequency,
only four were found to be positively associated with failing implants versus nondiseased control
implants using a bivariate analysis (chi square or Fisher’S exact). These organisms were Prevotella
nigrescens, Peptostreptococcus micros, Fuso . nucleatum ss vincentii, and Fuso. nucleatum ss nucleatum
(Table 4). In no instance was there a significant difference between the frequency of detection or the
levels present comparing the flora of failing and stable sites within the same patient. Overall, 95.7% of
failing sites harbored either P. nigrescens or P. micros. F. nucleatum species may be pathogenic. It is
also seen as a commensal organism associated with P. nigrescens and P. micros . Our mediator data
would suggest that F. nucleatum is serving as a key commensal organism in this mixed infection, as will
be discussed.
Attendant risk is summarized in Table 5, with an overview of the crude (unadjusted) odds ratio,
overall risk, and kappa and P values for each of these organisms. For example, P. micros and P.
nigrescens would have a 26.7-fold and 13.7-fold increased chance, respectively, of being present in the
subgingival plaque of a failing implant site, and these are shown to be highly significant. There is also a
2- to 3-fold attendant risk that P. micros or P. nigrescens would be present in the flora of a failing
implant site versus a nondiseased control site.
Although not statistically significant, there was a tendency for a greater detection frequency of P.
gingivalis, B. forsythus, and T. denticola in the failing and stable implant sites of group 1 patients versus
the control patients (group 2). No levels of A. actinomycetemcomitans were detected in any of the
implant sites. Of the nondiseased control sites (group 2), there were also greater detection frequencies
noted with the Streptococcus species, especially S. gordonii and S. mitis, as well as P. intermedia , but
again these were not statistically significant.
When evaluating the flora of the six completely edentulous patients (three group 1 and three group 2
patients), there were no differences seen in the species or detection frequencies between or among the
completely edentulous patients and the partially edentulous patients. However, with the small number of
patients in these categories, differences cannot be excluded.
Microbial and Mediator Associations. The relationship between the four pathogens associated with
failing implants and the levels of mediators present in the local gingival crevicular fluid was examined.
There was no significant relationship seen between the four predominant organisms and the GCF
mediator levels of IL-1ß or IL-6. However, PGE 2 was highly correlated with the levels of P. nigrescens
and P. micros present when examined pairwise (Table 6). This suggests that F. nucleatum ss vincentii
and F. nucleatum ss nucleatum are commensal organisms and are associated with disease. However, the
quantitative relationship between the level of P. nigrescens and P. micros present and the level of GCF
PGE2 present suggests and is consistent with a causal relationship.

Discussion
In the present investigation, the clinical variables of inflammation, probing depths, and attachment levels
were shown to be significantly different with failing implants than with nondiseased control implants.
Microbially, we also found greater detection frequencies of two fairly uncommon species, P. nigrescens
and P. micros , in the subgingival plaque of patients with failing implants. This study has also shown

Article Text 6
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

increased gingival crevicular fluid levels of PGE2, IL-1ß, and PDGF associated with failing and stable
implant sites in group 1 patients compared to nondiseased control sites in group 2 patients.
Marginal tissue reactions at osseointegrated implant–tissue interfaces have been reviewed by Schou
et al.33 The general consensus appears to be that successfully osseointegrated implants exhibit low
amounts of plaque concomitant with the absence of marginal inflammation. However, plaque
accumulation may cause inflammatory reactions around the implants. Perspective may be gained by
analogy to gingivitis and periodontitis, in determining the reactions of host tissues to plaque on abutment
and prosthesis surfaces. A recent clinical study in humans by Pontoriero et al 34 on experimentally
induced peri-implant inflammation reaffirms the similar association between the accumulation of
bacterial plaque and the development of peri-implant mucositis, as established for the gingival units by
the experimental gingivitis model. In most of the implant studies reported, Plaque, Bleeding, and
Gingival indexes are used to evaluate the degree of health or disease present, as well as are
measurements of probing depth, attachment levels, mobility, and radiographic bone loss. The Plaque and
Gingival indexes, which are descriptive reflections of the patients’ compliance with their home care
regimen, may not correlate as well with osseous changes.
In the present study, there were no significant differences in the plaque or bleeding scores of control
patients or patients with failing implants. Therefore, this suggests that these clinical measures could not
be used as discriminators of a healthy or diseased peri-implant state. One clinical variable of significance
in this study was the marginal redness seen with the peri-implant tissues in patients with failing implants.
Unfortunately, however, marginal redness was seen around failing implant sites and within-mouth stable
implant sites, as well. This may be accounted for by the greater detection frequencies of the four
periodontopathogens seen in those sites. This suggests that redness may potentially identify patients at
risk but may not discriminate implants at risk. Given the lack of significant differences in clinical
measures of home care compliance (eg, plaque and BOP) in the patient groups, it seems reasonable to
assess patient response variables to identify high-risk patients.
As expected, other clinical variables found to be associated with failing implants in this study were
increased mean probing depths and attachment levels. Analogous to periodontitis, the loss of attachment
around an implant would be reflected clinically as deeper probing depths and increased levels of clinical
attachment. However, a recent report by Ericsson and Lindhe 35 questioned the significance of probing
around osseointegrated implants. In their experimental data, they observed that the resistance to probing
offered by the gingiva was greater than that offered by the peri-implant mucosa, and, consequently, the
probe penetration was more advanced at implants than at teeth. The tip of the probe terminated apically
of a laterally displaced junctional epithelium and close to the crest of the alveolar bone (0.2 mm).
Furthermore, probing depth measurements at clinically healthy gingivae failed to provoke bleeding,
whereas bleeding on probing was recorded from most of the nondiseased implant sites. They concluded
that the interaction of soft tissue and implants is less resistant to mechanical forces than the connective
tissue attachment to teeth, and that probing the peri-implant mucosa dislocates the mucosa and allows the
probe to penetrate beyond the apical termination of the junctional epithelium.
This hesitation in the diagnostic applicability of probe depth measurements is reiterated by Schou et
36
al. These authors state, “. . . the disparity between clinical and radiographic measurements indicates
that probing measurements around implants and teeth are not fully comparable, [and that] . . . the altered
probing depths of implants may be directly related to bone loss, whereas changes around teeth with
attached dentogingival fibers may rather be related to attachment loss and changes in inflammation.”

Article Text 7
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

Until long-term studies demonstrate the diagnostic applicability and significance of probing, careful
monitoring of probing depths and clinical attachment levels over time needs to be evaluated
concomitantly with radiographic examination. In the present study, the primary basis for patient selection
as a failing implant patient was the radiographic appearance of crestal bone loss generally greater than 2
mm. We did not include an analysis of occlusal loading, which is a recognized risk factor for implant
failure.37 Grieve et al31 have demonstrated that teeth under significant loading, secondary to orthodontic
forces, have significant increases in GCF levels of PGE2 and IL-1ß. These may serve to promote
inflammation, bacterial overgrowth, and attachment loss. This occlusal component may be a factor in the
failing implant data observed in the present study.
The microbial findings in this study highlight the significant detection of two rather unfamiliar
microbes associated with failing implants, P. nigrescens and P. micros , as well as the presence of F.
nucleatum ss vincentii and ss nucleatum, which are herein considered to be commensal organisms. As
reported earlier, the primary florae seen associated with failing or failed implants often contain high
numbers of gram-negative anaerobic rods and spirochetes. P. intermedia and Fusobabacterium spp,
which have been regularly found in elevated proportions in failing sites,15,38 are suspected periodontal
pathogens and are commonly associated with periodontitis lesions. Recently reported by Haffajee and
Socransky, 39 strains of P. intermedia expressing different phenotypic traits have been separated into two
species, P. intermedia and Prevotella nigrescens: “This distinction makes earlier studies of this species
difficult to interpret, since data from two different species may have been inadvertently pooled. However,
new studies that discriminate the species in subgingival plaque samples might strengthen the relationship
of one or both species to periodontal disease pathogenesis.” Until additional studies have been
completed, they rank P. nigrescens as a moderate contender in strength of the relationship to periodontal
disease. This has also been implicated in refractory periodontitis.40
P. micros is a gram-positive, anaerobic, small asaccharolytic coccus. Although this species has only
recently been implicated in destructive periodontal diseases, it has long been associated with mixed
anaerobic infections in the oral cavity and other parts of the body. 39 In the article by Haffajee and
Socransky, 39 it was noted that P. micros has been detected more frequently and in higher numbers at
sites of periodontal destruction, as compared with gingivitis or healthy sites, and was elevated in active
sites. With additional studies, this microorganism may become a key player in the microbial diagnosis
and pathogenesis of failing implants.
In our study, there were greater detection frequencies of four microbes—P. nigrescens, P. micros, F.
nucleatum ss vincentii and F. nucleatum ss nucleatum —present in patients who had both failing and
stable implants versus patients who had only nondiseased implants. The presence of these
periodontopathogens in the peri-implant subgingival flora of patients with both failing and
radiographically nondiseased stable implants are consistent with other studies that demonstrate the
overall site-to-site consistency of the florae within subjects. The fact that an organism, if present, is
generally located at many other sites affecting most teeth is consistent with intraoral spread or site-to-site
translocation of organisms. It is possible that translocation could potentially occur from surrounding
periodontally involved teeth to an implant site. The suspicion that teeth serve as an important source of
bacteria for the colonization of implants has recently been investigated by Mombelli et al.41 These
investigators evaluated the presence of suspected periodontal pathogens in the peri-implant microflora of
osseointegrated implants exposed 3 and 6 months to the oral environment of patients previously treated
for periodontal disease. The patients in this study showed a high peri-implant prevalence of anaerobic

Article Text 8
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

putative periodontal pathogens (P. intermedia, Fusobacterium, C. rectus, and P. gingivalis ) 3 to 6


months after exposure of the implants to the oral environment. The findings of this study, as well as
others,16,42,43 corroborate the concept that the dental microflorae appear to be an important source of
bacteria to the peri-implant microflorae in partially edentulous patients. Therefore, if antimicrobial
treatment is to be directed to the failing implant site, consideration should be given to the surrounding
periodontal tissues for the reduction or elimination of potential periodontopathogens.
The local host response to peri-implant bacterial infection is a relatively new area of research with
several pressing questions remaining unanswered. Is the host defense different in the peri-implant tissues
as a result of the lack of a periodontal ligament and inserting connective tissue fibers into cementum?
Does loss of attachment and bone loss occur more readily around an implant than around the natural
tooth because of these missing components of the attachment apparatus? Can mediators of this
inflammatory process provide early diagnostic or prognostic information as to the status of the implant?
Recent reports44,45 on epithelial implant interactions have demonstrated the shared expression of
surface receptors and production of inflammatory mediators by the junctional epithelium. The findings
from the present study further strengthen the belief that the local host response to this peri-implant
infection is biochemically similar to the response seen in periodontitis. It is also interesting to note that
the prevalent microbes detected in our study with failing implants are similar to the pathogens associated
with periodontal infections. The significant elevations of PGE2 and IL-1ß, in both failing implant sites
and at in-mouth stable implant sites, demonstrate that an increased local response is measurable on the
patient level as well as at local sites of inflammation. It is not surprising to see such a response, because
substances (eg, lipopolysaccharide or lipotechoic acid) from periodontopathic bacteria initiate and drive
the inflammatory response, and their continued presence is essential for maintenance of the
inflammation.45 However, endogenous molecules such as PGE 2 and IL-1ß mediate the inflammatory
process and play a major role in its amplification and perpetuation and in the ensuing tissue destruction.
Therefore, the significant correlation of elevated GCF levels of PGE2 with greater detection frequencies
of P. nigrescens and P. micros might provide additional diagnostic insight into the pathogenesis of
failing implants.
Overall, findings from the present study support the hypothesis that risk for developing infection
appears to be primarily at a patient level and secondarily at a site or implant level from a clinical
(redness), microbial (P. micro s and P. nigrescens ), and biochemical (PGE2 and IL-1ß) perspective. This
suggests that in patients with failing implants, not only is the failing implant at risk, but other implants
that may be considered healthy and stable at that time may also be at risk for future failure. This study
was of a cross-sectional nature and provides only a glimpse in time at the spectrum of peri-implant
pathogenesis. Longitudinal studies would provide a greater perspective as to the monitoring of disease
activity around implants and the usefulness of these disease markers to predict the potential susceptibility
of a nondiseased stable implant becoming a failing implant.
Conclusions
The following conclusions can be drawn from this study:
1. Four microorganisms—P. nigrescens, P. micros, F. nucleatum ss vincentii, and F. nucleatum ss
nucleatum—are associated in mouths with failing implants.
2. In contrast to previous studies demonstrating the similarity of pathogens in failed versus healthy
implants, this investigation places new emphasis on the potential role of P. nigrescens and P. micros ,

Article Text 9
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

an anaerobic gram-positive organism.


3. PGE2, IL-1ß, and PDGF are significantly elevated in mouths with failing implant sites as compared
to healthy implant control sites.
4. From a microbial and mediator perspective, risk is principally at a patient level and apparently not at
a site level.
5. Dual assessment of the microbes and mediators provides additional diagnostic insight into
peri-implant pathogenesis.
Acknowledgments
This research was awarded the 1993 Osseointegration Foundation/ITI Implant Research Grant. The
authors of this paper wish to acknowledge the significant contributions of Dr Crawford A. Tatum, Jr,
Opilika, AL; office of Dr Stephen B. Mackler, Dr Neil Lutins, and Dr Robert Knox, Greensboro, NC;
Beverly Bell, RDH; Betsey Warhurst; Karen Bullock, CDA; Morris Worley; and Bettye Jones. This
research was also supported in part by The Southern Academy of Periodontology and NIH
2P01DEO8917–0681, R01DE10519, and K15–DE00325.

Jeanne M. Salcetti

Postdoctoral Fellow, Department of Periodontics,


University of North Carolina at Chapel Hill, School
of Dentistry, Chapel Hill, North Carolina; Private
Practice, Colorado Springs, Colorado (July 1995).

John D. Moriarty

Associate Professor, Department of Periodontics,


University of North Carolina at Chapel Hill, School
of Dentistry, Chapel Hill, North Carolina.

Lyndon F.Cooper

Assistant Professor, Department of


Prosthodontics, Dental Research Center,
University of North Carolina at Chapel Hill, Chapel
Hill, North Carolina.

Frances W. Smith

Footnotes 10
Laboratory Research Specialist, Dental Research
Center, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina.

John G. Collins

Research Associate, Dental Research Center,


University of North Carolina at Chapel Hill, Chapel
Hill, North Carolina.

Sigmund S. Socransky

Senior Staff Member, Department of


Periodontology, Forsyth Dental Center, Boston,
Massachusetts.

Steven Offenbacher

Professor, Department of Periodontics, Dental


Research Center, University of North Carolina at
Chapel Hill, Chapel Hill, North Carolina.

FIGURES

Footnotes 11
Figure 1

Fig. 1 Mean GCF PGE2 levels at implant sites.

Figure 2

Fig. 2 Mean GCF IL-1ß levels at implant sites.

Figures 12
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

Figure 3

Fig. 3 Mean GCF PDGF levels at implant sites.

TABLES

Figures 13
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

Table 1

Tables 14
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

Table 2

Table 3

Tables 15
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

Table 4

Table 5

Tables 16
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42): The Clinical, Microbial, and Host Response Characteristics of the

Table 6

The Clinical, Microbial, and Host Response Characteristics of the Failing Imp

1. Adell R, Lekholm U, Rockler B, Brånemark P-I. A 15-year study of osseointegrated implants in the treatment
of the edentulous jaw. Int J Oral Surg 1981;10:387–416.
2. Cox JF, Zarb GA. The longitudinal clinical efficacy of osseointegrated dental implants: A 3-year report. Int J
Oral Maxillofac Implants 1987;2:91–100.
3. Albrektsson T. A multicenter report on osseointegrated oral implants. J Prosthet Dent 1988;60:75–84.
4. Bower RC, Radny NR, Wall CD, Henry PJ. Clinical and microscopic findings in edentulous patients 3 years
after incorporation of osseointegrated implant-supported bridgework. J Clin Periodontol 1989;16:580–587.
5. van Steenberghe D. A retrospective multicenter evaluation of the survival rate of osseointegrated fixtures
supporting fixed partial prostheses in the treatment of partial edentulism. J Prosthet Dent 1989;61:217–223.
6. Jemt T, Lekholm U, Adell R. Osseointegrated implants in the treatment of partially edentulous patients: A
preliminary study on 876 consecutively placed fixtures. Int J Oral Maxillofac Implants 1989;4:211–217.
7. Lekholm U, van Steenberghe D, Herrmann I, Bolender C, Folmer T, Gunne J, et al. Osseointegrated implants
in the treatment of partially edentulous jaws: A prospective 5-year multicenter study. Int J Oral Maxillofac
Implants 1994;9:627–635.
8. Quirynen M, Naert I, van Steenberghe D, Dekeyser C, Callens A. Periodontal aspects of osseointegrated
fixtures supporting a partial bridge. An up to 6-years retrospective study. J Clin Periodontol 1992;19:118–126.
9. Lekholm U, Adell R, Lindhe J, Brånemark P-I, Eriksson B, Rockler B, et al. Marginal tissue reactions at
osseointegrated titanium fixtures (II). A cross-sectional retrospective study. Int J Oral Maxillofac Surg
1986;15:53–61.
10. Meffert RM. Treatment of failing dental implants. Curr Opin Dent 1992;2:109–114.
11. Worthington P, Bolender CL, Taylor TD. The Swedish system of osseointegrated implants: Problems and
complications encountered during a 4-year trial period. Int J Oral Maxillofac Implants 1987;2:77–84.

References 17
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42):

12. Rosenberg ES, Torosian JP, Slots J. Microbial differences in two clinically distinct types of failures of
osseointegrated implants. Clin Oral Implants Res 1991;2:134–144.
13. Tonetti M, Schmid J. Pathogenesis of implant failures. Periodontology 2000 1994;4:127–138.
14. Lekholm U, Ericsson I, Adell R, Slots J. The condition of the soft tissues at tooth and fixture abutments
supporting fixed bridges. A microbiological and histological study. J Clin Periodontol 1986;13:558–562.
15. Mombelli A, Van Oosten MAC, Schürch E, Lang NP. The microbiota associated with successful or failing
osseointegrated titanium implants. Oral Microbiol Immunol 1987;2:145–151.
16. Apse P, Ellen RP, Overall CM, Zarb GA. Microbiota and crevicular fluid collagenase activity in the
osseointegrated dental implant sulcus: A comparison of sites in edentulous and partially edentulous patients. J
Periodontal Res 1989;24:96–105.
17. Becker W, Becker BE, Newman MG, Nyman S. Clinical and microbiologic findings that may contribute to
dental implant failure. Int J Oral Maxillofac Implants 1990;5:31–38.
18. Rams TE, Roberts TW, Tatum H, Keyes PH. The subgingival microbial flora associated with human dental
implants. J Prosthet Dent 1984;51:529–534.
19. Ong ES-M, Newman HN, Wilson M, Bulman. The occurrence of periodontitis-related microorganisms in
relation to titanium implants. J Periodontol 1992;63:200–205.
20. Leonhardt Å, Berglundh T, Ericsson I, Dahlén G. Putative periodontal pathogens on titanium implants and
teeth in experimental gingivitis and periodontitis in beagle dogs. Clin Oral Implants Res 1992;3:112–119.
21. Leonhardt Å, Adolfsson B, Lekholm U, Wickström M, Dahlén G. A longitudinal microbiological study on
osseointegrated titanium implants in partially edentulous patients. Clin Oral Implants Res 1993;4:113–120.
22. Smedberg JI, Svensäter G, Edwardsson S. The microflora adjacent to osseointegrated implants supporting
maxillary removable prostheses. Clin Oral Implants Res 1993;4:165–171.
23. George K, Zapiropoulos GG, Murat Y, Hubertus S, Nisengard RJ. Clinical and microbiological status of
osseointegrated implants. J Periodontol 1994;65:766–770.
24. Genco RJ. Host responses in periodontal diseases: Current concepts. J Periodontol 1992;63:338–355.
25. Offenbacher S, Odle BM, Van Dyke TE. The use of crevicular fluid prostaglandin E2 levels as a predictor of
periodontal attachment loss. J Periodontal Res 1986;21:101–112.
26. Honig J, Rordorf-Adam C, Seigmund C, Wiedemann W, Erard F. Interleukin-1 beta (IL-1ß) concentration in
gingival tissue from periodontitis patients. J Periodontal Res 1989;24:362–367.
27. Stashenko P, Fujiyoshi P, Obernesser MS, Prostak L, Haffajee AD, Socransky SS. Levels of interleukin 1ß in
tissue from sites of active periodontal disease. J Clin Periodontol 1991;18:548–554.
28. Jovanovic SA, James RA, Lessard G. Bacterial morphotypes and PGE2 levels from the perigingival site of
dental implants with intact and compromised bone support [abstract 1395]. J Dent Res 1988;67:287.
29. Kao R, Curtis D, Preble J, Finzen F, Richards D. Crevicular fluid analysis of diseased and healthy implants
[abstract 1463]. J Dent Res 1993;72:286.
30. Graves DT, Cochran DL. Periodontal regeneration with polypeptide growth factors. Curr Opin Periodontol
1994;2:178–186.
31. Grieve WG III, Johnson GK, Moore RN, Reinhart RA, Dubois LM. Prostaglandin E (PGE) and interleukin-1
beta (IL-1ß) levels in gingival crevicular fluid during human orthodontic tooth movement. Am J Orthod

References 18
JOMI on CD-ROM (1997 © Quintessence Pub. Co.), 1997 Vol. 12, No. 1 (32 - 42):

Dentofacial Orthop 1994;105:369–374.


32. Socransky SS, Smith C, Martin L, Paster BJ, Dewhirst FE, Levin AE. ‘Checkerboard’ DNA–DNA
hybridization. Biotechnics 1994;17:788–792.
33. Schou S, Holmstrup P, Hjørting-Hansen E, Lang NP. Plaque-induced marginal tissue reactions of
osseointegrated oral implants: A review of the literature. Clin Oral Implants Res 1992;3:149–161.
34. Pontoriero R, Tonelli MP, Carnevale G, Mombelli A, Nyman SR, Lang NP. Experimentally induced
peri-implant mucositis. A clinical study in humans. Clin Oral Implants Res 1994;5:254–259.
35. Ericsson I, Lindhe J. Probing depth at implants and teeth. An experimental study in the dog. J Clin Periodontol
1993;20:623–627.
36. Schou S, Holmstrup P, Stoltze K, Hjørting-Hansen E, Kornman KS. Ligature-induced marginal inflammation
around osseointegrated implants and ankylosed teeth. Clin Oral Implants Res 1993;4:12–22.
37. Mombelli A, Lang NP. Antimicrobial treatment of peri-implant infections. Clin Oral Implants Res
1992;3:162–168.
38. Saadoun AP. Microbial infections and occlusal overload: Causes of failure in osseointegrated implants. Pract
Periodontics Aesthet Dent 1993;5(6):11–20.
39. Haffajee AD, Socransky SS. Microbial etiological agents of destructive periodontal diseases. Periodontol 2000
1994;5:78–111.
40. Haffajee AD, Socransky SS, Dzink JL, Taubman MA, Ebersole JL. Clinical, microbiological and
immunological features of subjects with refractory periodontal diseases. J Clin Periodontol 1988;15:390–398.
41. Mombelli A, Marxer M, Gaberthüel T, Grunder U, Lang NP. The microbiota of osseointegrated implants in
patients with a history of periodontal disease. J Clin Periodontol 1995;22:124–130.
42. Leonhardt Å, Adolfsson B, Lekholm U, Wickström M, Dahlén G. A longitudinal microbiological study on
osseointegrated titanium implants in partially edentulous patients. Clin Oral Implants Res 1993;4:113–120.
43. Quirynen M, Listgarten MA. The distribution of bacterial morphotypes around natural teeth and titanium
implants ad modum Brånemark. Clin Oral Implants Res 1990;1:8–12.
44. Carmichael RP, McCulloch CAG, Zarb GA. Quantitative immunohistochemical analysis of keratins and
desmoplakins in human gingiva and peri-implant mucosa. J Dent Res 1991;70:899–905.

Schmid B, Spicher I, Schmid J, Chambers DA, Lang NP. Plasminogen activator in human gingival tissue
adjacent to dental implants. Clin Oral Implants Res 1992;3:85–89.
45. Page R. The role of inflammatory mediators in the pathogenesis of periodontal disease. J Periodontal Res
1991;26:230–242.

References 19

You might also like