Peri-implantitis: A Comprehensive Overview of Systematic Reviews

Dental implants have become widely used in restoring the fully or partially edentulous patient. They have become a predictable alternative to fixed and removable partial dentures and were often the treatment of choice.^1,2  High implant survival rates of 92.8%–97.1% over a follow-up period of 10 years indicated that dental implants were a valid treatment option for the dental rehabilitation of the partially and fully edentulous patient.^3,4  However, despite its high survival rates, dental implants were prone to biological complications like peri-implantitis.⁵  Peri-implantitis was described as a destructive inflammatory lesion affecting hard and soft tissues of the osseointegrated implant causing bone loss and peri-implant pocketing.⁶  Peri-implantitis can be asymptomatic, showing only signs of bleeding on probing, attachment loss, and bone loss. Or peri-implantitis can manifest clinical signs of increasing probing depths, suppuration, draining sinus, and peri-implant mucosal swelling or recession.⁷  If peri-implantitis was not detected early and treated, the bony destruction could extend the whole lengthen of the implant, resulting in loss of implant stability.⁷  Thus, early peri-implantitis detection and effective treatment is crucial in a practice that focuses on implant rehabilitation of the edentulous patient.

Some studies indicated that patients, who have lost 1 implant due to peri-implantitis, were more prone to implant failure.^8,9  Patients with periodontal disease seemed to experience more implant loss due to peri-implantitis than periodontally healthy patients.^10,11  Patients who smoke were also at risk for peri-implantitis, but non-smoking patients can develop peri-implantitis, and not all smoking patients develop peri-implantitis.^12,13  Radiographically, patients with periodontitis and smokers have also reported significantly more marginal bone loss around their implants.¹⁴  Thus, these factors predisposing peri-implantitis should be closely examined when treatment planning the dental patient for implants.

The aim of this comprehensive review was to provide a systematically derived overview of systematic reviews pertaining to different aspects of peri-implantitis that will help the clinician understand and manage peri-implantitis in their practice.

• What is the prevalence, incidence, or risk of peri-implantitis in periodontal health and disease?

• What factors are associated with peri-implantitis?

• What treatment intervention is most effective in treating peri-implantitis?

A systematic search was conducted of PubMed, Embase, Web of Science, Cochrane library, and Google Scholar for systematic reviews and meta-analyses of peri-implantitis published from October 1989 until October 2016. The keywords used for the search were ‘“peri-implantitis” OR “peri-implant disease”' AND ‘“systematic review” OR “meta-analysis.”' Gray literature was also searched on Google Scholar using advance search to find articles with the word “peri-implantitis” and then again for “peri-implant”. Both searches were done with at least 1 of the words used: “systematic review” or “meta-analysis”. In addition, hand-searching was conducted on the reference list of selected meta-analyses and systematic reviews.

• The review must be identified as a meta-analysis or a systemic review in the abstract or title.

• All definitions of peri-implantitis included were specified as one of the following: (1) the consensus definition agreed upon in the 1st European Workshop on Periodontology,¹⁵  (2) the presence of inflammation in the peri-implant mucosa, as indicated by bleeding and/or pus on probing, with loss of supporting bone,⁶  (3) a continuous marginal bone loss beyond biological bone remodeling or more than 2 mm; and with signs of inflammation like purulence, bleeding on probing, and more than 6 mm probing pocket depth, (4) an incidence of probing pocket depth ≥5 mm with bleeding on probing and or suppuration and radiographic signs of bone loss of ≥2.5 mm or bone loss extending ≥ the first 3 threads,¹⁶  (5) peri-implant probing depth >5 mm with bleeding on probing, or (6) peri-implant crestal bone loss at osseointegrated dental implants in conjunction with inflammation of peri-implant mucosa.¹⁷ 

• The focused questions or review objectives must pertain to peri-implantitis in humans.

• If peri-implant mucositis was included in the review, only the peri-implantitis data was included.

• Only systematic reviews or meta-analyses that reviewed 5 or more studies pertaining to peri-implantitis were included.

• Based on the AMSTAR (A MeaSurement Tool to Assess systematic Reviews) checklist,¹⁸  only studies scoring >3 were included.

• Reviews including animal studies were excluded.

• Marginal bone loss in the absence of inflammation or marginal bone loss with no mention of peri-implantitis or gingival condition were excluded.

• Peri-implant mucositis were excluded.

• Comments, editorials, posters, and critical reviews of systematic reviews were excluded.

Two reviewers (MT and JC) independently screened the title and abstract to exclude articles that clearly were not systematic reviews or meta-analyses pertaining to peri-implantitis. The inclusion and exclusion criteria previously described were independently applied by the reviewers (MT and JC) while analyzing the full-text for inclusion. Disagreements were resolved by discussion with a third reviewer (JBS).

One reviewer (MT) extracted the data using a previously pilot tested data extraction form, and 2 other reviewers (BEB and JC) independently checked the extraction data for precision and entirety. Disagreements were resolved through discussion.

The methodological quality of a systematic review can be evaluated using the AMSTAR tool.¹⁸  AMSTAR has been specifically developed to overcome the shortcomings of previous measurement tools that were lengthy and complicated to use.¹⁸  AMSTAR consists of 11 questions; each question is given a score of 1 if the criteria is satisfied, or a score of 0 if the criteria is not met, unclear, or not applicable.¹⁹  The sum of the scores from each question results in an overall score reflecting the review quality.¹⁹  Although controversial, AMSTAR characterized systematic review quality at 3 levels: 8 to 11 for high quality, 4 to 7 for medium quality, and 0 to 3 for low quality.¹⁹  The AMSTAR tool was used to assess the quality of the selected systematic reviews. The scoring used the AMSTAR checklist¹⁸  and was performed by 2 reviewers (MT and BEB). Disagreements were resolved by discussion with a third reviewer (JC). Reviews scoring 3 or less were excluded in this overview.

The search yielded 351 reviews in PubMed, 161 in Embase, 165 in Web of Science, 20 in Cochrane Library, and 108 in Google Scholar. After the initial abstract and title screening, 59 reviews were selected from PubMed, 39 from Embase, 54 from Web of Science, 19 from the Cochrane Library, 64 from Google Scholar, and 14 from hand searching of the reference list of the selected systematic reviews or meta-analyses. The duplicates were eliminated and a total of 83 reviews remained for full-text analysis. After full-text analysis, 50 were eliminated, resulting in 33 selected for data extraction (Figure 1).

Of the 33 articles selected,^20–52  8 reviewed prevalence, incidence, or risk of peri-implantitis,^{20,23,24,37,43,47,48,52}  2 reviewed diagnostic findings,^25,29  3 reviewed microbial findings,^40–42  2 reviewed the effects of systemic disease,^50,51  and 18 reviewed treatment interventions.^*

A total of 8 reviews reported on the prevalence, incidence, and risk of peri-implantitis (Table 2). Six of the 8 reviews reported on the prevalence or incidence of peri-implantitis in patients with chronic periodontitis or a history of periodontitis.^{20,23,37,43,48,52}  Three out of 8 reviews reported on prevalence or incidence of peri-implantitis in smokers and non-smokers.^20,23,47  However, the conclusion drawn from these systematic reviews was based on significant heterogeneity among most of the studies reviewed.

Based on a computed overall summary estimates, the frequency of patients with peri-implantitis was 18.8%, and the frequency of implants with peri-implantitis was 9.6%.²⁰  Another review reported that the prevalence of peri-implantitis ranged from 1%–47% with an estimated weighted mean prevalence of 22%.²⁴  Figure 2 showed differences in reported prevalence ranges on a patient level compared to on an implant level, in the same patient population.^20,23,24,37  Peri-implantitis was less likely to occur during the first 5 years of implant function; implants affected by peri-implantitis ranged from 0–3.4%.²³  After an observation period of 10 years, implants affected by peri-implantitis ranged from 10.7%–47.2%.²³  Thus, after the first 5 years, peri-implantitis was a frequently observed problem. With patients who were enrolled in supportive maintenance programs, the frequency of patients with peri-implantitis decreased to 14.3%.²⁰  A minimum implant recall interval of 5 to 6 months was suggested for a significant positive impact on the incidence of peri-implantitis.³⁷ 

A majority of the reviews reported significantly lower occurrence of peri-implantitis in non-periodontitis patients compared with periodontitis patients.^{20,23,37,43,48}  Implants placed in patients with a history of treated periodontitis ^43,48,52  reported wider ranges and higher percentages of peri-implantitis prevalence (Figure 3). Patients with a history of periodontitis also have a higher incidence of marginal bone loss around implants and peri-implantitis compared to non-periodontitis patients.⁴³  A higher incidence of peri-implantitis was also observed in generalized aggressive periodontitis at 26% compared to non-periodontitis patients at 10%.⁴⁸  Patients with residual pockets have more implant sites with peri-implantitis when compared to patients without residual pockets.⁵² 

A higher prevalence of peri-implantitis was reported in smokers.^20,23  An implant-based analysis revealed significantly greater risk of peri-implantitis in smokers compared to non-smokers.⁴⁷  However, the patient-based analysis conducted by the same systematic review did not find significant difference in peri-implantitis risk.⁴⁷ 

Two systematic reviews^25,29  reported on the effects of peri-implantitis on the levels of specific proinflammatory or anti-inflammatory cytokines (Table 3). There were higher levels of proinflammatory cytokines in the peri-implant crevicular fluid of implants with peri-implantitis than in healthy implants.^25,29  The studies included in both systematic reviews were heterogeneous regarding the diagnosis of peri-implantitis. IL-1β release and TNF-α release was significantly higher in peri-implantitis compared to healthy peri-implant mucosa.^25,29  However, the IL-1β levels in peri-implantitis was not statistically significant when compared to peri-implant mucositis.²⁹  Increased levels of IL-1β and TNF-α in peri-implant crevicular fluid from sites with peri-implantitis have been related to increased gingival index, probing depth, bleeding on probing, and bone loss.²⁹  Other cytokines like IL-4, IL-6, IL-8, IL-10, IL-12, and IL-17 have also been investigated for a link to peri-implantitis. These proinflammatory or anti-inflammatory cytokines associated with peri-implantitis increased with peri-implant establishment and progression.²⁹ 

Three systematic reviews^40–42  reported on the microbial findings in peri-implantitis (Table 4). The microbiologic profile of peri-implantitis is different from periodontitis and can be complex and variable.⁴²  It consists of aggressive and resistant microorganisms and may include opportunistic microorganisms, gram-negative anaerobic pathogens, gram-positive nonsaccharolytic anaerobic rods, and Epstein–Barr virus. Although conflicting results have been reported, the following microorganisms were found to be more prevalent in peri-implantitis^40,41  than in peri-implant health: Aggregatibacter actinomycetemcomitans, Porphyromonas gingivalis, Prevotella intermedia, and Treponema denticola, human herpesvirus 4 and 5, Epstein–Barr 1, and human cytomegalovirus 2.⁴⁰  In addition, microorganisms such as Tannerella forsythia, Porphyromonas gingivalis, Treponema socranskii, Staphylococcus aureus, Staphylococcus anaerobius, Staphylococcus intermedius, and Streptococcus mitis were also found comprising 30% of the total microbiota at peri-implantitis sites.⁴⁰  Peri-implantitis sites have higher mean colony-forming units in peri-implantitis sites compared with healthy sites.⁴²  The reported active periodontal pathogens are not limited to periodontopathic bacteria, and can include opportunistic bacteria like Staphylococcus aureus, Staphylococcus intermedius, Streptococcus mitis, and Haemophilus influenzae.⁴² 

Two systematic reviews^50,51  reported on the effects of systemic diseases on peri-implantitis (Table 5). Patients with diabetes were at a higher risk of peri-implantitis.⁵¹  The gingival index, probing depths, and bone loss were higher in poorly controlled compared to well-controlled diabetic peri-implantitis patients.⁵⁰  However, conflicting results were reported for type 2 diabetes.⁵⁰ 

Patients with cardiovascular disease were also at a higher risk of peri-implantitis.⁵⁰  In addition, patients with peri-implantitis were found to have a 3 times greater chance of harboring Epstein–Barr virus.⁵⁰  However, for patients with rheumatoid arthritis, statistical analysis demonstrated no associations.⁵⁰ 

A total of 18 reviews^†

Surgical interventions focused on flap elevation, implant surface treatment, and detoxification, with or without the use of an anti-microbial agent, and with or without the use of membranes or grafting materials. The surgical treatments included (1) open-flap debridement with plastic or carbon curettes, ultrasonic scaler, rotating instruments, air powder, or soft laser treatment; (2) resective peri-implant surgery and implantoplasty; and (3) guided bone regeneration techniques with or without different types of membranes (synthetic membranes, resorbable bovine or porcine collagen) in combination with or without bone substitutes (demineralized freeze dried bone alone or in combination with growth factors, autogenous bone, hydroxyapatite, xenografts, and algae-derived calcium carbonate).

Various adjunctive therapies may improve the efficacy of conventional peri-implantitis treatment.⁴⁶  Debridement together with antibiotics resulted in the greatest probing depth reduction compared to debridement only.²⁸  At a short-term follow-up of 12 months, mechanical debridement and minocycline appeared to improve treatment outcomes of peri-implantitis when compared to debridement and chlorohexidine.^35,38  The use of erbium: yttrium–aluminum–garnet (Er:YAG) laser and carbon dioxide (CO²) lasers can improve short-term implant clinical parameters up to 6 months.^35,39  Er:YAG laser treatment may also result in greater reduction in bleeding on probing (BOP) scores compared with submucosal debridement with adjunctive submucosal irrigation with chlorhexidine.^35,38  Implantoplasty or lasers might provide equivalent effects when compared to other commonly used methods for surface decontamination.²¹  In addition, the use of submucosal glycine powder air polishing may greatly reduce BOP scores compared to submucosal irrigation with chlorhexidine digluconate and debridement; and produced similar clinical outcomes compared with Er:YAG laser treatment.³⁸  Network meta-analysis of other non-surgical approaches in peri-implantitis treatment showed that single or combined non-surgical interventions also resulted in greater probing depth reduction than debridement alone.²⁸ 

In short-term follow-ups, surgical interventions reduced probing depth by 30%–50% of the initial probing depth.^21,32  Although regenerative procedures can achieve a mean of 2–2.41 mm radiographic bone fill,^21,32,33,44  and can improve clinical parameters of peri-implant tissues,^22,44  the use of a guided bone regeneration protocol with membrane and bone graft does not seem to be predictable in treatment of peri-implantitis.^21,22,33,45  When all surgical and non-surgical approaches were pooled together, surgical approaches showed greater improvements in probing depth and clinical attachment levels. However, when the surgical and non-surgical approaches were analyzed separately, the difference between the approaches were not statistically significant.²⁷ 

Successful treatment outcomes of peri-implantitis were described as post-treatment implants with a mean probing depth of less than 5 mm and no progressing bone loss. At 12 months' follow-up, Heitz-Mayfield et al³¹  reported successful overall peri-implantitis treatment outcomes for different combinations of adjunctive treatments for surgical and non-surgical interventions at 76%–100% of patients, and at 75%–93% of implants.

This overview had only included systematic reviews of medium to high quality, and the majority of the systematic reviews reviewed in this overview were of high quality in terms of the conduct of the systematic review according to the AMSTAR rating¹⁹  (Table 1). However, the conclusions derived from most of the peri-implantitis systematic reviews needed to be interpreted with caution as stated by the individual systematic reviews included in the overview. In general, the included systematic reviews had the following limitations inherent in their selected studies: variation of the study designs, different implant systems used, and varying duration of follow-up periods, as well as the lack of standardization in reported outcomes at participant and implant levels. Other limitations were from the inability to control co-existing confounding factors in the pre-existing studies, and from restricting the search to English, as studies published in other languages were overlooked.

Furthermore, the definition used for peri-implantitis was different across studies and all variations of peri-implantitis definitions were included in this overview. The following are the different definitions of peri-implantitis used by the selected systematic reviews: (1) the consensus definition agreed upon in the 1st European Workshop on Periodontology,¹⁵  (2) the presence of inflammation in the peri-implant mucosa, as indicated by bleeding and or pus on probing, with loss of supporting bone,⁶  (3) a continuous marginal bone loss beyond biological bone remodeling or more than 2 mm; and with signs of inflammation like purulence, bleeding on probing, and more than 6 mm probing pocket depth, (4) an incidence of probing pocket depth ≥5 mm with bleeding on probing and or suppuration and radiographic signs of bone loss of ≥2.5 mm or bone loss extending ≥ the first 3 threads,¹⁶  (5) peri-implant probing depth >5 mm with bleeding on probing, (6) peri-implant crestal bone loss at osseointegrated dental implants in conjunction with inflammation of peri-implant mucosa.¹⁷ 

Another limitation to the systematic reviews was due to a lack of standardized assessment tools; there was a wide variation of unvalidated tools used to assess the quality of reporting of the selected studies; some were modified from validated assessment tools to evaluate the quality of non-randomized studies. The following were the quality assessment tools or criteria used by the selected systematic reviews of this overview to assess the quality of their included studies: (1) the tool derived from the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement,⁵³  (2) the tool developed by den Hartog et al,⁵⁴  (3) the Newcastle–Ottawa scale⁵⁵  or an adaptation of it, (4) the tools modified from the randomized controlled trial checklist of the Cochrane Center⁵⁶  and or the Consolidated Standards of Reporting Trials (CONSORT) statement,⁵⁷  (5) the Risk of Bias tool from the Cochrane Collaboration,⁵⁶  (6) the tool from the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach,⁵⁸  (7) the modified criteria proposed by Esposito et al⁵⁹  and Roccuzzo et al,⁶⁰  (8) tool by Khan et al,⁶¹  and (10) the criteria from the Agency for Healthcare Research and Quality.⁶²  The quality of the studies included in most of the systematic reviews were of high, moderate, or unclear risk of bias. Most studies were limited by an inadequate protection from bias leading to an insufficient quality to assess the validity of the trial.

The limitation of the peri-implantitis prevalence reviews (Figure 2) were that the studies reviewed did not factor in patients' age, systemic status, and history of periodontal status. This may account for the wide ranges reported for the % patients and % implants affected by peri-implantitis. In Figure 3, when the history of periodontitis was factored into the peri-implantitis prevalence. Lower ranges were reported in patients with healthy periodontium compared to patients with a history of periodontitis, although other patient factors were not taken into consideration.

The limitations of the microbial findings for peri-implantitis were the heterogeneity of the studies included in the selected reviews, and the differences in sensitivity and specificity of the different microbial identification methods used. The different microbial identification methods used included culture techniques, darkfield microscopy, DNA–DNA checkerboard hybridization technique, DNA probe analysis, 16S rRNA gene sequencing technique, and polymerase chain reaction (PCR) techniques. The different detection techniques used in the studies prevent comparisons of microbial data across studies. Furthermore, most studies report the microbial data as frequency of detection and may not provide enough information to ascertain the extent of microbial involvement. In addition, there are also differences in the mode of sample collection; some studies use paper points whereas other studies use curettes.

The microbiologic analysis of the peri-implantitis sites can be divided into 2 types: studies that tested for target pathogens and studies that evaluated the entire microbiome.⁴²  Metagenomic and metatranscriptomic techniques that analyze the entire microbiome included 16S pyrosequencing⁶³  and use of the 16S gene clone library⁶⁴  to test for a wide range of microorganisms. Many of the studies included in the selected systematic reviews only tested for target pathogens. Studies that only test for target pathogens will lack the data pertaining to the overall microbial composition in peri-implantitis. Thus, without more studies with a comprehensive analysis of the phylogenetic and taxonomic bacterial diversity that exist in the peri-implantitis sites, the conclusions drawn in the systematic reviews on the microbial findings will be limited.

The conclusions derived from most of the systematic review pertaining to peri-implantitis treatment also needed to be interpreted with caution. This was because the number of included studies for each surgical or non-surgical procedure was too low to enable strong statistical analysis. Furthermore, only some of the included systematic reviews comprised studies that compared treatment effects of different approaches. The included studies have various degrees of heterogeneity in study design, case selection, and treatment. Since no methodology was established as the gold standard for the treatment of peri-implantitis, the majority of studies were designed as a comparison between 2 completely different types of intervention rather than between a recognized control. This reduced the clinical implications even in the higher quality studies. Also, in a systematic review of treatment effectiveness, it is unacceptable that the definition of peri-implantitis was not standardized across studies. In addition, a high risk of bias can result in an exaggeration of treatment effect, and coupled with a low level of trials reporting, this can lead to a significant overestimation of intervention efficacy.

In view of the limitations of the included systematic reviews, the outcome of this overview suggested the following:

• (1)

   There was a higher occurrence of peri-implantitis after 5 years of implant function.

• (2)

   There was a higher occurrence of peri-implantitis in patients with aggressive periodontitis, chronic periodontitis or a history of periodontitis compared to non-periodontitis patients.

• (3)

   There was a higher occurrence of peri-implantitis in smokers compared to non-smokers.

• (4)

   IL-1β release and TNF-α release was significantly higher in peri-implantitis compared to healthy peri-implant mucosa.

• (5)

   The microbiologic profile of peri-implantitis is different from periodontitis and may include A actinomycetemcomitans, P gingivalis, P intermedia, T forsythia, T denticola, T socranskii, S aureus, S anaerobius, S intermedius, S mitis, human herpesvirus 4 and 5, Epstein–Barr 1, and human cytomegalovirus 2. The microorganisms active in peri-implantitis are not limited to only periodontopathic pathogens and may involve some opportunistic pathogens.

• (6)

   Patients with uncontrolled diabetes and cardiovascular disease have a higher risk of peri-implantitis, but there was no association between rheumatoid arthritis and the risk of peri-implantitis.

• (7)

   Any other single or combined non-surgical interventions were better in peri-implantitis treatment than debridement alone.

• (8)

   Surgical treatment of peri-implantitis can reduce probing depths.

• (9)

   Guided bone regeneration can be unpredictable in peri-implantitis treatment.

• (10) 

   Different combination of adjunctive treatments for surgical and non-surgical interventions can produce successful peri-implantitis treatment outcomes.

• (11) 

  There was no strong evidence to suggest the most effective treatment intervention for peri-implantitis.

• (12) 

  Postimplant maintenance may be necessary to reduce the occurrence of peri-implantitis in high-risk patients.

• (13) 

  More randomized controlled trials using standardized definitions for peri-implantitis were needed for all forms of peri-implantitis treatment interventions.

Abbreviations

AMSTAR

A MeaSurement Tool to Assess systematic Reviews

BOP

bleeding on probing

CCT

nonrandomized controlled trials

CONSORT

Consolidated Standards of Reporting Trials

DFDBA

demineralized freeze-dried bone allograft

Er,CR:YSGG

erbium, chromium: yttrium–scandium–gallium–garnet

Er:YAG

erbium: yttrium–aluminum–garnet

FES

fully edentulous subjects

GBR

guided bone regeneration

GRADE

Grading of Recommendations Assessment, Development, and Evaluation

IL

interleukin

IL-1β

interleukin-1 beta

NRP

non-residual pocket

PCR

polymerase chain reaction

PD

probing depth

PES

partially edentulous subjects

PICF

peri-implant crevicular fluid

PPD

probing pocket depth

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RCT

randomized controlled trials

RP

residual pocket

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology

TNF-α

tumor necrosis factor alpha

MT, JC, BEB, and JBS declare that they have no competing interests with regard to the content of the manuscript.