Aetiology, microbiology and therapy of periapical lesions around oral implants: a retrospective
analysis
For figures, tables and references we refer the reader to the original paper.
A periapical lesion around a dental implant, also called retrograde peri-implantitis, is defined as a
clinically symptomatic lesion at the apex of an implant that develops shortly after fixture installation,
while the coronal portion of the implant achieves a normal bone to implant interface (Quirynen et al.
2003). These clinical symptoms can include pain, tenderness, swelling and/or the presence of a
fistula. An incidence of 0.26% up to 1.86% has been reported in literature (Reiser & Nevins 1995,
Quirynen et al. 2005). In addition, the knowledge on the incidence, aetiological factors and treatment
of periapical lesions around implants is scarce.
Different aetiological factors have been suggested to play a role in the emergence of retrograde periimplantitis such as excessive heating of the bone during the preparation of the osteotomy (McAllister
et al. 1992, Piattelli et al. 1998a), failed endodontic and/or apicoectomy procedures (Ayangco &
Sheridan 2001), placement of the implant in proximity of an existing infection (Reiser & Nevins 1995,
Shaffer et al. 1998), excessive tightening of the implant with compression of the bone during implant
placement (Piattelli et al. 1998b) and the presence of pre-existing microbial pathology (McAllister et
al. 1992, Sussman & Moss 1993, el Askary et al. 1999). One case report even mentions the
occurrence of a lesion after the implant was placed, penetrating the nasal cavity (Silva et al. 2010). All
these factors cause what can be called an active lesion.
Besides these, there are also inactive lesions, when radiographic findings are not associated with
clinical symptoms. For example an asymptomatic radiolucency can be seen at the radiography when
the drilling depth during the preparation of the surgical site exceeded the length of the implant that
is placed or when an implant is placed next to an already existing radiolucency, caused by scar tissue
(McAllister et al. 1992, Reiser & Nevins 1995).
At present, periapical lesions around dental implants are considered to have a multifactorial
aetiology, combining one or more of the above. Therefore, there is no consensus regarding the
therapy. While most authors agree that the apex of the implant should be surgically exposed, there is
less agreement about how the therapy should be continued. Some propose a curettage and
detoxification of the implant surface using different solutions (Balshi et al. 1994, Bretz et al. 1997,
Esposito et al. 1999), while others combine this with a guided bone regeneration (GBR) procedure
using a bone substitute material and a bioresorbable membrane (Quirynen et al. 2005). Even more
aggressive approaches have been proposed such as an intra-oral apicoectomy procedure (Balshi et al.
2007, Dahlin et al. 2009).
Even though most authors suggest that there is a microbiological factor in the aetiology of periapical
lesions around dental implants, this aspect has been hardly investigated. In a recent review, it was
reported that in a total of 32 implants with retrograde peri-implantitis that were histologically
evaluated, bacteria were detected in only one case. Three other studies showed aseptic bone
necrosis (Romanos et al. 2011).
Therefore, the aim of this study was threefold. As a pre-existing microbial pathology and the
presence of an existing infection is suggested to play a role in the aetiology of periapical lesions, the
first aim of this study was to evaluate retrospectively whether an endodontic pathology or apical
radiolucency at the tooth prior to extraction (at the implant site) or at the adjacent teeth has an
influence on the emergence of a periapical lesion around a dental implant. Secondly, a retrospective
analysis of the different treatments performed at our department was included to evaluate their
outcome. The last aim of this study was to determine which kind of bacteria were present in
periapical lesion around implants.
Materials and Methods
Endodontic pathology
All single tooth implants, placed in the upper and lower jaw at the Department of Periodontology
(University Hospital, Catholic University Leuven) between January 2004 and December 2009, were
included in this retrospective case-control study. If a tooth showed a periapical lesion at the moment
of extraction, a thorough curettage was standard performed. Adjacent teeth, that not needed to be
extracted, were (re)treated endodontically. Implant placement was always delayed until full healing
of the future implant site.
The periapical status of the tooth at the implant site and the adjacent teeth prior to implant
placement were evaluated primarily on digital intra-oral radiographs (Minray®; Soredex, Tuusula,
Finland). If such radiograph was not available, an extra-oral panoramic radiograph was used to judge
the periapical condition (Cranex Tome®; Soredex, Tuusula, Finland). No cone beam CT's were
evaluated. The analysis of the periapical implant lesions was performed by two calibrated,
independent periodontologists. Results were re-evaluated when there was an inter-examiner
difference. The periapical status of the tooth at the implant site and the adjacent teeth prior to
extraction was explored and identified as:
a.no endodontic treatment and no periapical lesion
b.a periapical lesion at the root combined with or without an endodontic treatment, and
c.an endodontic treatment without clear signs of a periapical lesion.
For implants with a radiological follow-up of at least 6 months after implant placement, the implant
itself was evaluated for the presence of an apical pathology visible via intra-oral radiographs. All
radiographs available in the patient files, taken after implant installation until August 2010 were
subjected to such evaluation. If no periapical radiological information was available about the tooth
at the implant site, or the adjacent teeth the implant was excluded from the analysis. A lesion was
radiologically defined as a radiolucent area at the apical part of the teeth or implants.
Treatment and survival
All periapical lesions around dental implants, observed and treated at our department since January
2000, were examined retrospectively. Included were: implants placed for single tooth replacement,
partial and full fixed restorations, or as support for overdentures in both the upper and lower jaw.
For each implant, the kind of treatment performed and the implant survival (no disintegration,
absence of mobility, implant still in function without signs of infection or pain), were registered.
Microbiology
In 21 cases, whenever possible, a microbial sample of the periapical lesion was taken at the moment
of treatment. These samples were analysed for enterococcus species, Aggregatibacter
actinomycetemcomitans, Campylobacter rectus, Fusobacterium nucleatum, Prevotella intermedia
and Porphyromonas gingivalis. Furthermore, also the total count of aerobic and anaerobic bacteria
was explored.
The samples were taken with sterile paper points (Roeko®; Roeko, Langenau, Germany), which were
inserted into the area for 20 s during the moment of treatment and were dispersed in 2 ml of
reduced transport fluid (RTF). Each sample was homogenized by vortexing for 30 s and processed
within 12 h by culturing (for details, see Quirynen et al. 1999).
Results
Endodontic pathology
Between January 2004 and December 2009, a total of 230 implants in the upper jaw and 88 in the
lower jaw were placed in 291 patients for single tooth replacement. Two hundred and sixty-five
patients received a single implant, 25 patients 2 and 1 patient 3 implants. Of the 318 implants placed,
245 (77%) were Brånemark TiUnite implants (Nobel Biocare, Gothenburg, Sweden), 67 (21%) were
OsseoSpeed (Astra Tech Dental, Mölndal, Sweden), and 6 (2%) were SLActive (Straumann, Basel,
Switzerland).
For 70 of these implants, no information about periapical condition of the extracted tooth could be
retrieved. For four implants no information was available about the periapical condition on the
mesial and for eight on the distal adjacent tooth. The kappa-value for the inter-examiner reliability
was 0.9 (p < 0.001).
In total, 248 implants (182 upper jaw / 66 lower jaw) placed in 232 patients were subjected to
analysis [185 TiUnite (75%), 58 OsseoSpeed (23%), and 5 SLActive implants (2%)]. For 124 and 38 of
the implants in upper and lower jaw, respectively, intra-oral digital radiographs were available, the
others, (58 in upper, 28 in lower jaw) were analysed on panoramic radiographs.
Most of the periapical lesions were identified at a control visit between implant insertion and
abutment connection (thus during the submerged healing, n = 36), at abutment connection (n = 11),
or soon afterwards (n = 5 within first year, n = 7 afterwards).
The relationship of the endodontic status of the tooth prior to extraction and periapical pathology
around inserted implant is summarized in Table 1. If the tooth showed no signs of a periapical lesion
and had no endodontic treatment, the incidence of an apical pathology was 2.1%. On the other hand,
if an endodontic treatment or a periapical lesion at the apex of the tooth was present at the moment
of extraction, a periapical lesion could be found around the implant in 8.2% and 13.6% of the cases
respectively (Table 1). The Odds ratio for an apical pathology around an implant was 7.2 (p < 0.001)
for a tooth with an endodontic history, compared with a pathology free tooth. Figure 1 shows the
radiological findings of a tooth prior to extraction and the periapical status of the implant that was
subsequently placed.
The influence of the adjacent teeth, in case the extracted tooth showed neither a pathology nor an
endodontic treatment, can be seen in Table 2. If the neighbouring teeth, either on the mesial or the
distal side, showed no signs of pathology or did not receive endodontic treatment, only 1.2% of the
implants presented with a periapcial lesion. When an endodontic treatment was performed in the
past, but no signs of periapical pathology could be detected, no periapical lesions around the
implants could be found. However, if there were signs of periapical pathology at the neighbouring
teeth, 25% of the implants also showed a periapical lesion (OR 8.0, p = 0.01).
Table 2. Percentage distribution of periapical lesions along implants per condition of neighbouring
teeth, for sites where a tooth had been extracted without endodontic pathology or treatment
Since 2000, 59 implants (41 upper / 18 lower jaw) treated at our department for a periapical lesion
could be retrieved. Different treatment options had been chosen: explantation of the affected
implant (17/59), curettage of the defect and application of a bone substitute (Bio-Oss, Geistlich,
Schlieren, Switzerland) in the defect (14/59), curettage and administration of systemic antibiotics
(10/59), simply curettage (11/59), no treatment (2/59), only systemic antibiotics without curettage
(2/59), curettage with the usage of a barrier membrane without application of a bone substitute
(2/59), and curettage and application of autogenous bone chips (1/59). All curettages were
conducted via an explorative flap. Of the 42 non-explanted implants, 9 implants were lost during
follow-up, all during the first 4 years of loading. The cumulative survival rate for an implant showing a
periapical lesion is 46.0% (Table 3). When the explanted implants were excluded, the cumulative
survival rate reached 73.2% after 10 years. The decision whether to remove an affected implant or to
treat it, was mainly based on the extent of the lesion and the part of osseointegration that was left at
the moment of treatment. Whenever an implant showed very few bone to implant contact (mostly
only two or three threats coronally remaining), meaning that a very large lesion was present, it was
decided to remove this implant. Other cases were treated. Figure 2 shows an implant before and
after treatment with a bone substitute. A clearcut selection of “best treatment strategy” could not be
found. The marginal bone level changes along the surviving implants remained stable (26/33
implants coronal to 2nd thread, 7/33 implants coronal to third thread).
The implant on position 23 shows an active periapical implant lesion. During surgery the defect was
visualized and a thorough curettage was performed. Hereafter, a bone substitute was used to fill the
defect. After 6 months of healing, there was no recurrence of the lesion.
Table 3. Cumulative survival rate (%) of an implant showing a periapical lesion, with a maximum
follow-up time of 126 months
Microbiology
A total of 21 lesions were sampled. Bacteria were found in nearly all sites, but only in 9/21 in
concentrations ≥log 4. The proportion of anaerobic species was always higher when compared with
aerobic species. P. gingivalis and P. intermedia were detected in reasonable concentrations at six and
four sites respectively. The other specific species tested (A. actinomycetemcomitans, C. rectus and F.
nucleatum, Enterococci) never reached the threshold level for identification (see Table 4).
Table 4. Microbiological data from 21 sampled periapical lesions
Discussion
A route for implant surface contamination at insertion can be the site into which the implant is
placed. An endodontic infection of an adjacent tooth is frequently mentioned in the literature to be a
causal factor in the process of developing a periapical lesion (Shaffer et al. 1998, Sussman 1998). Our
observations are in agreement with these findings. If an active endodontic lesion is present on an
adjacent tooth, which can be verified true conventional radiographs, indeed a much higher frequency
of affected implants was observed. When the adjacent teeth did not suffer from endodontic
pathology or when there was an endodontic treatment present, but without any signs of periapical
pathology, only in 1% of the cases the implant showed a periapical lesion. This percentage is much
higher (25%, OR 8.0), when an adjacent tooth showed clear signs of endodontic pathology, meaning
that an infection from a neighbouring tooth can spread to the apex of an implant, and thus can be
the cause of an infection in this area. This observation can be of great importance in the treatment
protocol of the implant patient. Teeth showing an endodontic pathology should therefore be treated
prior to implant placement and the latter should best be postponed until the radiological sign of a
periapical lesion at the adjacent teeth disappeared.
In this study, two dimensional radiographs were used for the analysis. This type of radiographs does
not always show the correct size of an intra-bony defect. Only when the junctional area is involved,
these kinds of defects can be identified. Therefore, it might be possible that some periapical
pathologies are not recognized on two dimensional radiographs, which is a limiting factor in the
diagnosis. A cone beam CT can be used to overcome this limitation (Van Assche et al. 2009).
A residual lesion of the extracted tooth might be another explanation of the contamination of the
implant surface. This can include granulomas, residual apical cysts, root remnants,… (McAllister et al.
1992, Ayangco & Sheridan 2001, Jalbout & Tarnow 2001). This study clearly indicates that the dental
history of the implant site is of remarkable importance. At sites where a tooth with a periapical lesion
had been extracted, the odds ratio for an apical lesion along the inserted implant is very high (7.23)
Even if no periapical lesion is present (i.e. visible on conventional radiographs) at the moment of
extraction, but there is an endodontic treatment, the risk of implant pathology is higher. In a cadaver
study, Green et al. (1997) reported that even though endodontically treated teeth may be clinically
classified as successful on the basis of conventional radiographs, up to 26% still exhibits histological
signs of inflammation. Therefore, a thorough curettage of the future implant site, even with bur,
should be recommended at the moment of extraction. In the past, we always tried to hermetically
close the wound after tooth extraction. Perhaps this approach facilitated the encapsulation of
remaining pathology. As we changed this approach, in combination with a thorough curettage
including the use of burs, the incidence of periapical lesions around implants reduced dramatically.
Today, there is no consensus on which kind of therapy is superior to another for the treatment of
periapical lesions around oral implants. Already in 1992, McAllister and co-workers introduced the
concept of a surgical flap approach with curettage of the bony defect and granulation tissue,
detoxification of the exposed implant surface and possibly a GBR procedure, but long-term outcome
data are lacking. In this study, 15 implants were treated using a GBR procedure. After 36 months, 11
were still in function without clinical or radiological signs of inflammation, thus 26.7% of the implants
was lost during follow-up.
Another group suggested an apicoectomy on the affected implants (Balshi et al. 2007). They treated
39 implants in 35 patients with this intra-oral apicoectomy procedure. Thirty-eight of the 39 implants
(97.4%) remained stable and continued in function with no further complication up to 5 years.
This could suggest that an apicoectomy procedure should be preferred above a GBR procedure. The
latter is in agreement with other small case series (Bretz et al. 1997, Ayangco & Sheridan 2001,
Quirynen et al. 2003, Bousdras et al. 2006, Dahlin et al. 2009, Rosendahl et al. 2009).
Another group, however, published a case series where they did not remove the apex of the implant,
but combined a thorough curettage with saline and chlorhexidine rinsing, followed by defect fill,
using a tetracycline paste, and wound closure. A follow-up time up to 36 months showed no adverse
events and a good clinical and radiological healing (Zhou et al. 2012).
Another recent case report described a treatment without a surgical intervention (Chang et al. 2011).
A patient showing periapical lesions around two implants was first treated with systemic antibiotics
(amoxicillin) and paracetamol. When the symptoms did not improve, the medications were changed
to prednisolone, augmentin and mefenamic acid. After a follow-up period of 2 years, there were no
signs of recurrence. Another case report also mentioned a non-surgical approach, using systemic
antibiotics. The lesion slowly resolved after 9 months (Waasdrop & Reynolds 2010). In this study, two
cases were treated with only systemic antibiotics. In both cases amoxicillin was the medication of
choice. After 3 years, both implants were still in function without signs of recurrence. These data,
with low number of cases, are, however, not enough to conclude that a treatment with systemic
antibiotics, without surgical intervention, should be preferred.
Another treatment method for these kind of lesions could be the use of laser therapy, such as Er:YAG
laser or CO2 laser (Schwarz et al. 2006, Romanos & Nentwig 2008), or the use of photodynamic
therapy (Takasaki et al. 2009). However, there are no scientific data available for the use of these
methods in the treatment of peri-apical implant lesions, only in the treatment of peri-implant
disease.
Most authors agree that there is a role of bacteria in the development of the periapical
lesion.Rokadiya & Malden (2008) described a case were a mandibular implant showed swelling and
pain 5 weeks after the installation due to a periapical lesion. After removal of the implant, a
microbial analysis of the apical tissue was performed showing the presence of Staphylococcus
aureus. A more recent case report found actinomycotic colonies in the granulation tissue
surrounding a periapical lesion around an implant (Sun et al. 2012).
In this study, most lesions were found to be colonized by bacteria with higher counts of anaerobic
bacteria, compared with aerobic bacteria. This is logical as the lesions are located inside the bone,
were anaerobic bacteria can survive more easy. The most prominent species (from the group of
periodontal pathogens) was P. gingivalis. The latter might be taken into consideration when
antibiotics have to be selected.
In this study, the microbial samples are all from rough surfaced implants. One can however question
if the type of implant surface is an important factor in bacterial adhesion and thus if these findings
could be compared with machined implant surfaces. A recent literature review addressed this
question (Renvert et al. 2011). The review revealed that very little data are provided on how implant
surfaces influence peri-implant disease. They concluded that there is no evidence that implant
surface characteristics can have a significant effect on peri-implant disease. It is important to note
however, that this study has very few to no data specific for periapical lesions.
Conclusion
Different aetiological factors have been suggested to play a role in the occurrence of a periapical
lesion around a dental implant: excessive heating of the bone during the osteotomy, failed
endodontic and/or apicoectomy procedures, placement of the implant in proximity of an existing
infection, excessive tightening of the implant with compression of the bone during implant
placement, and the presence of pre-existing microbial pathology. From this study one can conclude
that the endodontic status of the extracted tooth has a major influence on the occurrence of
periapical lesions. When an endodontic pathology is present at the moment of extraction, it is
approximately seven times more likely that a lesion will develop around the implant. Also, if
periapical pathology is present at the adjacent teeth, the risk of developing a lesion at the apex of the
implant is increased. A thorough curettage at the time of tooth extraction seems crucial.
There are different treatment approaches, with explantation as the most aggressive one, but an
effective one to resolve the pathology. Furthermore, curettage of the defect in combination with a
bone substitute shows to be an effective but not impeccable technique, an apicoectomy procedure
seems more favourable. The microbiological part of this study indicates that bacteria are present in
these periapical lesions, even though their active role is still unclear.