Introduction

The introduction of the biological concept of osseointegration [1] led to the development of endosseous root-formed implants, a revolutionary alternative in restorative dentistry. The highly successful use of osseointegrated dental implants to restore partial and full edentulism has been documented in a wealth of scientific literature [2–5]. According to available statistics, the number of dental implants installed annually around the world exceeds 2 million [6]. Just as the total population of the world is growing, and given that implant placement has become a routine treatment modality in daily clinical practice by specialists as well as by general practitioners, this number is expected to rise. The excellent success rates for dental implants presented in the literature lie in the fact that the criterion for failure has been implant loss, the terminal result of complete loss of osseointegration. Given the successful initial osseointegration, it is difficult to believe that implant loss is the result of a sudden burst without intermediate events. Thus, long-term biological complications around dental implants do occur and should be monitored by clinicians.

For more than three decades we have witnessed the emergence of two new diseases, peri-implant mucositis and peri-implantitis. These pathological conditions are considered to be major complications and draw significant attention in the literature. The term peri-implantitis was introduced to stress the infectious nature of the pathological conditions around peri-implant tissues [7, 8]. The consensus report from the 1st European Workshop on Periodontology [9] defined peri-implantitis as an irreversible inflammatory destructive reaction around implants in function, accompanied by loss of supporting bone. The 6th European Workshop on Periodontology presented modified definitions, not only to acknowledge that peri-implantitis is a treatable condition but also to include the collective term peri-implant disease for both entities, i.e., peri-implant mucositis and peri-implantitis [10]. Peri-implant mucositis was defined as a reversible inflammatory reaction in the soft tissues surrounding an implant in function, whereas peri-implantitis was defined as a more profound inflammatory lesion characterized by a deepened peri-implant pocket and loss of supporting bone around a functional implant. These two disorders are biofilm-mediated diseases developing in a fashion similar to the diseases occurring in the natural dentition. Mucositis corresponds to gingivitis and peri-implantitis is analogous to periodontitis [11•, 12].

Disease definition is quite different from a case definition. Despite the very clear and comprehensive disease definition, inconsistencies and confusion emerge when the terminology is applied in clinical practice. The variance in prevalence of peri-implantitis in studies available up to 2010 was the result of inconsistent case definitions [13, 14, 15•, 16, 17]. Koldsland et al. [15•] and Roos-Jansåker et al [17] demonstrated that the freedom to accept different thresholds of disease in the terminology had a direct impact on the prevalence of peri-implantitis.

This narrative systematic review encompasses the available literature since 2011, with the aims of identifying the core problem in defining peri-implant disease and elucidating how this affects the reports on prevalence in the studied populations.

Materials and Methods

All three authors performed an electronic search of the MEDLINE database via PubMed to find relevant articles using the following terms alone or in combination: peri-implant disease, peri-implant mucositis, peri-implantitis, periimplantitis, definition, prevalence, epidemiology. Articles were eligible if they reported data in the English language. In addition, a hand search was carried out to include reference lists as well as systematic reviews. While prospective longitudinal studies were eligible for the incidence of peri-implant disease, cross-sectional studies were considered for the assessment of prevalence. Focus was placed on recent studies, published in 2011 and later, so as to provide the reader with up-to-date information. All cross-sectional studies presenting a clear case definition for peri-implant disease and reporting prevalence data on a subject level are highlighted as being of importance (•) among the included studies. Studies since 2010 that are classical or exceptional in their methodology and concluding remarks were highlighted as being of major importance (••).

Results

The final results of the electronic search are summarized in Tables 1 and 2. Table 1 includes articles related to the focused question ‘Definition and prevalence of peri-implantitis’ and provides the aim of the study, the prevalence of mucositis and peri-implantitis (where available), as well as a comment on the major findings. Table 2 provides the most recent literature review articles related to the focused question.

Table 1 Included studies on case definition and the prevalence/incidence of peri-implant disease, presented in chronological order for the period 2011–2014 (N = 26)
Full size table
Table 2 Reviews and consensus reports 2011–2014 (N = 13)
Full size table

Case Definitions for Peri-Implant Disease

Case definitions for peri-implant mucositis and peri-implantitis are discussed separately. Note that studies included in the case definitions do not necessarily focus on prevalence, but could focus on risk factors [18, 19] or biological markers associated with pathogenesis [20]. The rationale for including such studies was to show the variance in case definitions in published studies.

  1. a)

    Peri-Implant Mucositis

    In most studies, the case definition for peri-implant mucositis is either not presented or is extracted directly from the terminology. Thus, mucositis was defined as visual inflammation and bleeding on probing (BoP) with no/slight bone loss [18, 20, 21•, 22, 23•, 24, 25•, 26, 27•, 28•, 29•]. However, the threshold for acceptable bone levels was further defined in some of the studies. In one, the marginal bone level was defined as the number of threads [25•]. Specifically, peri-implant mucositis was defined as presence of BoP and a bone level of <2 threads. A drawback of this study was that different implant systems were used and the variability in inter-thread distances cannot be neglected. In another study [27•], bone level was presented in millimeters and mucositis involved the presence of BoP, probing pocket depth (PPD) ≤5 mm, and a bone level of ≤2 mm. In other studies [28•, 29•], bone loss was assessed and mucositis was defined as the presence of BoP and a threshold of bone loss ≤0.5 mm.

  2. b)

    Peri-Implantitis

    In some studies, lack of clear case definitions was obvious since radiological assessments were made alone without any clinical registrations. In a cross-sectional study [24] and in two prospective cohort studies [30, 31], the threshold of ≥3 mm bone loss was considered to be associated with pathology without reporting on clinical parameters.

    In a few studies [32, 33], clinical registrations were used alone to assess biological complications, i.e., peri-implantitis. Whereas BoP was a prerequisite as a finding for these two studies, the threshold for PPD varied significantly, i.e., PPD >3 mm [32] and PPD >4 mm [33].

    In most studies, a combination of clinical and radiological measurements was used for case definition. Four studies assessed bone levels in the definition of peri-implantitis [23•, 25•, 27•, 34•]. Costa et al. [23•] and Dvorak et al. [34•] required BoP and/or suppuration (SUP) with PPD >5 mm. In neither of these two studies was the bone level specified. Mir-Mari et al. [25•] defined peri-implantitis as the presence of BoP and/or SUP and a bone level of ≥2 threads, whereas Marrone et al. [27•] defined it as PPD >5 mm, BoP, and bone level >2 mm. The rest of the studies [18–20, 21•, 22, 24, 26, 28•, 29•, 33, 35•, 36–38] assessed bone loss for the diagnosis of peri-implantitis. Casado et al. [18], Meyle et al. [19], and Rinke et al. [21•] did not specify a certain threshold for bone loss. Cecchinato et al. [28•, 29•] defined peri-implantitis as BoP, PPD ≥4 mm, and bone loss >0.5 mm. Charalampakis et al. [22] applied the criteria BoP/SUP, PPD ≥5 mm, and bone loss ≥1.8 mm after 1 year of the implants in function. This study also included the time that the implants had been in function in the case definition. The definition of Ravald et al. [37] required BoP/SUP and bone loss ≥2 mm. Stoker et al. [38] included cases of PPD >5 mm and bone loss ≥3 mm; however, no clinical signs of inflammation were recorded (i.e., BoP). Other studies, though reported in Table 1, do not present clear case definitions [39–43].

    A recent study proposed a simple classification of peri-implantitis based on the severity of the disease [44•]. Early peri-implantitis was defined as BoP/SUP, PPD ≥4 mm, and bone loss <25 % of the implant length. Moderate peri-implantitis would imply the presence of BoP/SUP, PPD ≥6 mm, and bone loss 25–50 % of the implant length. Advanced peri-implantitis would require BoP/SUP, PPD ≥8 mm, and bone loss >50 % of the implant length. Kadkhodazadeh and Amid have also attempted to classify a combined periodontal, endodontic, and peri-implant lesion [45].

Prevalence of Peri-Implant Mucositis and Peri-Implantitis

It is well-accepted that the prevalence/incidence of a disease should be based on subject-level analysis, as it is the subject and not the site (implant) suffering from the disease independently. Implants in the same patient cannot be independent from each other and data extracted at an implant level should be considered as clustered data with all the relevant limitations [46•]. With regard to the prevalence and incidence of peri-implant disease (Tables 1 and 2), cross-sectional and longitudinal studies should be included, respectively [47•]. All studies included on prevalence that are highlighted are of a cross-sectional design, demonstrate a clear case definition, and present patient-level results.

The various definitions of peri-implant disease including different clinical and radiological thresholds would exert a significant influence on the magnitude of the reported prevalence/incidence of peri-implantitis. Another factor that can affect the frequency of disease is the type of population included in the study (general or specific). Certain groups of patients might be of higher risk and the prevalence of peri-implant disease should be expected to increase in such individuals. A recent systematic review with meta-analysis [48•] concluded that the frequency of peri-implant mucositis was 63.4 % on the patient level and 30.7 % on the implant level. Corresponding figures for peri-implantitis were 18.8 and 9.6 %. For smokers, a group considered to be high risk, the frequency of peri-implantitis increased to 36.3 %. In another review of a narrative nature on the epidemiology of peri-implantitis, it was similarly stated that the prevalence of peri-implantitis over a 5- to 10-year period following implant placement is in the order of 10 % of implants and 20 % of patients [49•]. Below, we discuss the prevalence figures separately for the two forms of peri-implant disease.

  1. a)

    Peri-Implant Mucositis

    The prevalence of peri-implant mucositis at the patient level was reported in six studies and varied between 1.1 and 80 % [20, 21•, 22, 25•, 27•, 32].

  2. b)

    Implantitis

    The prevalence of peri-implantitis at the patient level was assessed in 12 studies and ranged from 1.4 to 53.3 % [18–20, 21•, 23•, 25•, 27•, 28•, 29•, 32, 34•, 36].

Discussion

The Problem with a Case Definition of Peri-Implant Disease

We have been facing a lack of consensus on a case definition of peri-implant disease. This is not surprising as we have encountered similar difficulties in the case of periodontitis. It has been shown that a major limiting factor in determining and comparing prevalence estimates of periodontitis across surveys is the lack of a universally accepted case definition for periodontitis [50, 51]. There is no agreement on the cut-off points for different levels of disease. Similar to periodontal disease [52], peri-implant disease does not seem to hold any diagnostic truth, in the sense that there is no natural basis for a clear distinction between different severity levels of disease.

An additional problem, inherent in the case of the dental implant, is that the position of the implant in the oral cavity is not biologically determined, implying that the acceptable bone level around an implant in function may not always be analogous to the 2 mm distance from the cementoenamel junction, the standard reference point in the case of the tooth. In order to assess bone level changes accurately, we need baseline radiological images. As reported in a recent consensus [53•], “in the absence of previous radiographic records, a threshold vertical distance of 2 mm from the expected marginal bone level following remodeling post-implant placement is recommended, provided peri-implant inflammation is evident”.

Another issue is the starting point for monitoring bone-level changes around an implant. Given that a dental implant, though osseointegrated with the host, is still a ‘foreign body’, some bone loss during the tissue homeostasis period for the first year of an implant being in function should be accepted. Most authors accept a marginal bone loss in the first year after implant placement of 1.5–2 mm [54–56]. As early as 1986, it was claimed that tissue stability is expected at 1 year after placement and that a continuous bone loss of more than 0.2 mm per year was undesirable [55]. It has been advocated that marginal bone loss around the neck of implants should be considered a remodeling phenomenon, not only some months after loading but continuously [56]. Thus, continuous bone loss was regarded as a natural bone remodeling process rather than a pathological condition. In a prospective study, it was demonstrated that bone remodeling continued for 6 months, after which no further changes were observed [57]. It should therefore be expected that the bone remodeling process occurs within a few months and terminates as soon as the biological width is established.

For the diagnosis of peri-implant mucositis, clinical features are sufficient. However, for peri-implantitis we not only need a combination of clinical and radiological features for a correct diagnosis, we additionally need evidence for progressive bone loss. There are cases where suboptimal bone levels are depicted on the x-rays in conjunction with BoP and PPD. However, this could be a result of suboptimal bone levels at the time of implant insertion or subcrestal implant placement for aesthetic reasons [58] and not a result of an infectious disease. In addition, certain implant designs create ‘pseudopockets’ by creating a distance from the implant shoulder to the mucosal margin of up to 5 mm [59].

A recent consensus statement [60••] required the presence of mucositis in conjunction with progressive crestal bone loss to diagnose peri-implantitis. However, universally accepted thresholds for bone loss and pocket depth are unavailable. The same consensus concluded that “the case definition of peri-implantitis remains unclear and varies substantially between studies”. This variance in the case definition of peri-implantitis results in the inclusion of heterogenous studies, which has a negative impact on disease epidemiology. All included studies in the current review report prevalence figures of peri-implant disease. There are only two available studies presenting data on the incidence of peri-implantitis [29•, 61•]. Cecchinato et al. [29•] reported the incidence of peri-implantitis over a 10-year period as being 12 % of patients, whereas an earlier 5-year follow-up study [61•] reported peri-implantitis at a level of 1 % according to the specific case definition used. To our knowledge, only one study has been published on the incidence of peri-implant mucositis to date in the literature [32]. This study demonstrated that with stringent maintenance, the incidence of peri-implant mucositis was less than 10 %.

Conclusions

Peri-implant disease is undoubtedly a reality in clinical practice. However it has some ‘gray zones’ with unanswered questions that require further investigation. Therefore, in some cases there is a current dilemma for both the patient and dental practitioner as to whether or not to take action regarding peri-implantitis.

A universally accepted case definition for peri-implantitis is mandatory for both clinicians and researchers. It is fundamental not only for clinicians to speak the same language when they define a peri-implantitis case but also for researchers in order to compare different studies and come up with robust conclusions in meta-analysis. This has not happened so far and needs to change. To reach a consensus on case definition would have two additional advantages. First, the true prevalence of peri-implantitis would be reported with a slighter discrepancy among studies, giving clear answers on whether peri-implantitis is a problem in clinical reality or not. In addition, homogenous figures on the successful outcome of peri-implantitis therapy would be achieved, as similar thresholds for the therapy goals would be used in the studies performed.

The case definition of peri-implantitis should be kept as simple as possible. We would assign a case as peri-implantitis on the conditions that (a) presence of BoP and/or suppuration is recorded and (b) progressive bone loss is clearly shown on radiographs. We recommend additional classification of the case as mild, moderate, or severe because this would have a direct impact on the treatment modalities to be used. We should bear in mind that PPD may not reflect disease severity in all cases and setting thresholds based on PPD may under-/overestimate the clinical problem. Having said this, bone loss in relation to the implant length is a better marker for characterizing the case as mild/moderate or severe peri-implantitis. Early diagnosis and correct classification is critical in order to stop future progressive bone loss by taking action accordingly. Peri-implantitis is unpredictable to treat and late therapeutic intervention may lead to implant loss.