Peri-implantitis is an inflammatory disease which affects the tissues surrounding dental implants, characterized clinically by increased probing pocket depth (PPD), bleeding on probing (BOP)/suppuration (SUP) and marginal bone loss (BL) beyond initial bone remodeling (1, 2). The disease progresses rapidly, following a non-linear accelerating pattern (3) that could lead to implant loss (4).

Different approaches for treating peri-implantitis have been proposed including non-surgical and surgical procedures (5). However, implants affected by peri-implantitis are considered to have doubtful prognosis (6) as despite the therapy, 30–50% of treated patients do not achieve complete disease resolution (7). These outcomes seem to be influenced by several factors, such as the implant surface (8), initial BL (9), baseline PPD (10), SUP at baseline (11) level of compliance, plaque control, diagnosis of severe periodontitis (12) and tobacco consumption (13).

The initial BL is related to the magnitude of the lesion. Hence, bone defects can be differentiated according to the BL pattern (intrabony/suprabony), the number of residual bone walls (dehiscence, 2-, 3-walls or circumferential) and the severity of the bone defect (relative BL related to the total implant length or the intra-surgical defect depth in millimeters) (14–16).

When BL progresses, a significant portion of the implant surface becomes exposed to the oral cavity, making it susceptible to biofilm contamination. Removal of the biofilm can be particularly challenging, especially on implants with rough or modified surfaces (11, 13). In these cases, peri-implant surgery allows better access to the implant surface in order to achieve optimal decontamination (17). However, depending on the amount of initial BL and its morphology, the instrumentation of the implant surface is hindered, with the apical part of the implant becoming less accessible and more difficult to decontaminate (18). Moreover, while suprabony defects allow better instrumentation, intrabony defects could hamper the possibilities for thorough decontamination (19, 20) (Figure 1).

When treating deep defects by resective approaches, the apical repositioning of the flap can be impossible due to the remaining vestibulum height (11). On the other hand, it is not likely that deep bone defects remain self-containing with preservation of the bone walls, which could impact the prognosis for regenerative approaches (21). In addition, BL can result in unfavorable configurations of hard and soft tissues related to the neighboring teeth, that could also hinder optimal cleaning (13). As recently supported by Monje et al. (2023), it can also cause proximal loss of periodontal support for these teeth (22). Finally, after the surgical treatment, a reduced bone level has also been associated with recurrence or progression of the disease (23).

Implant removal can be the most viable option in some severe peri-implantitis cases where the surgical approach has low predictability (24). However, there is no universally accepted threshold beyond which implant preservation becomes impossible (25, 26).

In consequence, the marginal bone level prior to the surgery becomes a key factor in the decision-making process. The aim of this narrative review is to analyze the scientific evidence related to the initial BL in peri-implantitis as prognostic factor prior to surgical treatment.

Bone remodeling is a physiological process that occurs during the first months of function after prosthesis delivery (2). Historically, different thresholds were proposed to clinically distinguish the physiological bone remodeling from the pathological bone loss (27). Consequently, these differences had affected the reported prevalence of the disease (28). It is accepted that bone remodeling should be limited to 2 mm after prosthesis delivery, and that beyond this threshold, it should be considered as pathological (2). However, these criteria are older and nowadays we should aim to prevent excess remodeling (29).

The marginal bone level is a primary diagnostic factor for peri-implantitis, with a radiographic sensitivity threshold for detecting bone changes of 1.0 mm (30). When initial data is not available to assess the limit of bone remodeling, a 3-mm BL threshold has been proposed for the diagnosis of peri-implantitis in conjunction with the clinical measurements (1, 2).

The long-cone parallel intraoral radiographic projection is recommended for peri-implant bone evaluation which allows to assess crestal bone changes longitudinally (27). The evaluation of marginal BL on radiographs has shown a positive linear correlation between mesial and distal marginal bone levels and the BL evaluated clinically during surgery (30, 31); however, the radiographic measurements resulted frequently in underestimation of the real bone loss assessed after flap elevation (mean 0.7 and 0.6 mm for mesial and distal, respectively) (30).

Although this radiographic technique is considered the gold-standard, it presents an important limitation related to the lack of information on the vestibular and lingual bone walls. When analyzing the residual walls, the morphology of the bone defect was correlated to the buccolingual crestal width, and 4-walls defects were found in broader crests (31). It is important to consider that the vertical component of the BL at buccal and lingual aspects was found to be statistically deeper at 4-walls defects than at 2-walls defects (31). Other authors have analyzed three dimensional assessment of peri-implant BL, however, scatter and artefacts may play a role in reducing the quality of the CBCT imaging, limiting the indication of this radiographic method for the evaluation of peri-implant defects (32).

Based on the literature, multiple classifications have been developed to categorize peri-implant BL (14–16, 33). While Schwarz et al, included only the type of bone loss (suprabony/infrabony) and the residual bone walls (16), other studies proposed the classification of peri-implantitis based on the severity of the BL as follows: Early/Slight, Moderate or Advanced peri-implantitis (BL <25%; 25–50%; >50% of the implant length, respectively) (14, 15, 33). Moreover, the position of the implant in relation to the alveolar crest has recently been included in the classification (BL due to implant malposition or ridge defects). This tridimensional consideration may have prognostic value on determining the best choice of the surgical approach, as for example, regenerative approaches can be limited if implants are placed outside the bone housing (33).

Several studies (7–9, 11–13, 34, 35) have analyzed the negative impact of the initial marginal bone levels on the surgical peri-implant outcomes in terms of healing, stabilization or implant failure. All these studies reported an inverse association between the initial severity of the disease and the therapeutic effectiveness. However, the degree of severity is not consistently reported across the studies, and it is classified heterogeneously. Table 1 summarizes the main findings of these studies.

The baseline BL prior to treatment can influence the choice of the peri-implant surgical approach (24) and the treatment outcome (13). Several studies have shown low success rates when initial BL is advanced (8, 9, 13). BL is inversely correlated to the effectiveness of therapy: the highest the baseline BL, the lowest the therapeutic success rate, independently of the surgical approach (resective or regenerative) (7–9, 13). Moreover, it has been shown that each additional 1-mm of BL prior to the surgical treatment, increases the risk of future implant failure by 65% (39).

The classification of BL used in the studies is heterogenous, not allowing to establish a cutoff point for explantation. The results of the analyzed studies depend on the reference point for the statistics and conclusions should be interpreted carefully. Thus, when the reference point was stated at BL < 40, the statistic threshold for implant failure was BL > 60% (8); however, when the reference was < 25%, the risk for failure increased statistically even for BL 25%–50% (9). Similar results occurred when the reference point was 33% of the implant length, with higher risk for BL up to the middle third (12). The differences between studies can be in part due to the disparities on “advanced BL” definition. Thus, it is important to report consistently the severity of the lesion, keeping the term “advanced” for BL > 50% (14, 15, 33).

When BL is restricted to one third of the implant length, non-surgical therapy may, in some cases, resolve the disease without further need of surgical treatment (12); however, surgical approaches in those situations, have shown better outcomes with success rates up to 70% (13, 34, 37). More factors other than apical BL extent such as defect configuration or soft tissue quality may influence the therapeutic choice and outcome (9, 31, 39).

It is recognized that for the treatment of the majority of advanced peri-implant lesions, surgical approaches are often needed (40). However, when the marginal BL reaches the apical region of the implant, the prognosis becomes hopeless and explantation should be considered (6). In fact, the stabilization of implants with BL > two thirds of the length is unpredictable (8).

From a clinical point of view, the diagnosis of peri-implantitis when baseline data is not available requires a PD ≥ 6 mm + BOP/SUP + marginal BL ≥ 3 mm (1). Considering a standard implant (10 mm length), BL would be almost at the middle third. When the implant BL affects the middle third of the implant (33%–66%), the decision between peri-implant surgery or explantation becomes more sensitive. When BL is in the middle third, studies have shown worst prognosis compared to those presenting BL < 25%–33% of the length (9, 12), however another study did not found significant differences between BL < 60% and BL < 40% (8). Thresholds of BL > 50%–60% of the implant appear to be aligned with the clinical reality. Furthermore, a study analyzing at which BL level implants affected by peri-implantitis were explanted, concluded that the practitioners performed explantations when implants presented a mean BL of 66.5% (25).

In this context, BL exceeding 50% of the implant length has been proposed as threshold for explantation (4, 24). However, the stabilization of implants with advanced BL is also achievable, thus explantation based only on bone levels could result in the loss of opportunity for some patients to maintain their implants (Figure 2). Moreover, although the success rates of implants placed in early failed implants sites are high (96% compared to 98% for implants placed for the first time), the new implants had significantly higher risk of failure (41). Hence, “rescue therapy” can be performed in advanced cases, depending also on patient desires (24) as the surgical treatment performed in implants with BL > 50% has shown 21% success rate (9), thus avoiding the sequela from implant explantation (42). In those cases a thorough examination of the risk indicators is important for establishing a pretherapeutic prognosis, such as hard and soft tissue deficiencies, tobacco consumption, medical condition, implant surface or cleanability of the prosthetic restoration (6, 8, 13). In fact, smoking is associated to advanced BL (13), residual pockets after surgical peri-implant treatment, and is considered as a prognostic indicator in treatment outcome (13). Thus, smoking cessation is advised in both pre- and post-surgical care (10). In addition, technical factors that may influence our treatment choice, such as the risk of potential damage of the neighboring anatomical structures (maxillary sinus or inferior alveolar nerve) or the prosthetic implication of the implant, should be considered.

Further research for development and validation of composite models for peri-implant prognosis is needed. These models should combine several risk factors to allow the practitioner to decide the best option according to the risk classification. In our knowledge, only one model is nowadays available based on Nobel Biocare implants (43). This model considers that BL of more than one third of implant length (>33%) and combined with factors, such as history of periodontitis, early disease development (<4 years of function) and implant length >13 mm, as having unfavorable prognosis. Unfortunately, this model cannot distinguish finer degrees of BL, highlighting the need for further research to support practitioners in making evidence-based decisions regarding explantation or peri-implant surgery.

In conclusion, several limiting factors pertain to the early diagnosis of peri-implantitis and its accuracy, such as the low sensitivity of intra-oral radiographs for detecting bone changes of approximately 1.0 mm (30), the low sensitivity of peri-implantitis classification for early-stage detection of the disease in the absence of baseline data (44), the rapid progression of the disease (3), and the prognostic significance of initial marginal BL (13). This situation highlights the importance of baseline documentation following implant placement for the early diagnosis and treatment of peri-implantitis.

BL is a key pretherapeutic prognostic factor for peri-implant surgical treatment outcome, as treatment effectiveness is inversely correlated with initial bone levels. In advanced cases, where BL exceeds 50–60% the success rates decrease significantly, thus the decision for explantation should be considered.