Influence of apico‐coronal positioning of tissue‐level implants on marginal bone stability during supracrestal tissue height establishment: A multi‐center prospective study

Abstract Introduction Supracrestal tissue height establishment is a crucial factor influencing peri‐implant marginal bone modifications prior to prosthesis delivery. If mucosal thickness is insufficient, peri‐implant marginal bone resorption occurs to allow appropriate supracrestal tissue height formation. This study evaluates if marginal bone resorption occurring around tissue‐level implants before prosthetic loading could be compensated by adapting apico‐coronal positioning to mucosal thickness. Methods Patients requiring placement of one single implant in the posterior mandible were treated with tissue‐level implants with a 3‐mm high transmucosal machined component and moderately rough implant body. Based upon vertical mucosal thickness measured after buccal flap reflection, implants were placed with the treated part: (group 1) 2 mm below crestal level in presence of thin mucosa (<2.5 mm); (group 2) 1 mm below the crestal level in presence of medium mucosa (2.5–3.5 mm); (group 3) at equicrestal level in presence of thick mucosa (>3.5 mm). Results Forty‐nine implants, placed in 49 patients were included in final analysis (group 1: 18 implants; group 2: 16 implants; group 3: 15 implants). Mean marginal bone resorption after 5 months of healing was 0.66 ± 0.49 mm, 0.32 ± 0.41 mm, and 0.22 ± 0.52 mm in groups 1, 2, and 3, respectively. Inter‐group analysis highlighted significant differences between the three groups after ANOVA test (p = 0.025). However, adaptation of apico‐coronal implant positioning in relation to mucosal thickness, allowed to avoid early exposure of the treated surface in 100%, 93.7%, and 53.3% of the implants in groups 1, 2, and 3, respectively. Conclusion During supracrestal tissue height formation, tissue‐level implants inserted adapting apico‐coronal positioning in relation to mucosal thickness exhibited greater marginal bone resorption at sites with thin mucosa than at sites with medium or thick mucosa. However, anticipating supracrestal tissue height establishment by adapting apico‐coronal implant positioning in relation to mucosal thickness may effectively prevent unwanted exposure of treated implant surface.

thickness, allowed to avoid early exposure of the treated surface in 100%, 93.7%, and 53.3% of the implants in groups 1, 2, and 3, respectively.
Conclusion: During supracrestal tissue height formation, tissue-level implants inserted adapting apico-coronal positioning in relation to mucosal thickness exhibited greater marginal bone resorption at sites with thin mucosa than at sites with medium or thick mucosa. However, anticipating supracrestal tissue height establishment by adapting apico-coronal implant positioning in relation to mucosal thickness may effectively prevent unwanted exposure of treated implant surface. What is known • Early marginal bone loss >0.5 mm during the first year of function represents a risk factor for future peri-implantitis development.
• Supracrestal tissue height establishment is a crucial factor influencing peri-implant marginal bone modifications prior to prosthetic loading.
• If mucosal thickness is insufficient, peri-implant marginal bone resorption occurs to allow appropriate supracrestal tissue height formation.

What this study adds
• Anticipating supracrestal tissue height establishment by adapting apico-coronal positioning of tissue-level implants in relation to mucosal thickness may effectively prevent early unwanted exposure of treated implant surface.

| INTRODUCTION
Stability of marginal bone levels has always been considered fundamental to evaluate long-term dental implant efficacy. Radiographic marginal bone loss up to 1.5-2 mm during the first year of function and a maximum of 0.2 mm annually thereafter, is a traditionally accepted criterion to define implant success. [1][2][3][4][5][6] Early marginal bone loss (EMBL) is a non-infective peri-implant crestal bone remodeling process occurring within the first year of function. 7 EMBL has a multi-factorial etiology, being influenced by various surgical and prosthetic factors including insufficient crestal width, [8][9][10][11] surgical trauma, 12 supracrestal tissue height formation, 13,14 microbial colonization of implant-abutment microgap, 15,16 presence of horizontal implant-abutment mismatch ("platform-switching"), [17][18][19] number of abutment connections/ disconnections, 20 prosthetic abutment height, [21][22][23][24][25][26] implant-abutment connection design and mechanical stability and adaptive response to occlusal loading. 7 Recent studies highlighted the importance of limiting EMBL to improve dental implant clinical outcomes. Galindo-Moreno and colleagues observed that EMBL >0.44 mm after 6 months of prosthetic loading is a strong predictor of >2 mm of marginal bone loss at 18-month follow-up. 27 In a recent 10-year prospective study, Windael and colleagues demonstrated that implants with EMBL ≥0.5 mm showed 5.43 times higher odds of future peri-implantitis development than implants with EMBL <0.5 mm during the first year of function. 28 Supracrestal tissue height formation is a principal factor influencing peri-implant marginal bone adaptation processes prior to prosthesis delivery. When the implant becomes exposed to the oral cavity, soft tissues establish a "cuff-like" barrier sealing the trans-epithelial component of the fixture. Pre-clinical studies by Abrahamsson and colleagues 29 and Berglundh and Lindhe 30 suggested that a minimum mucosal thickness is required to establish correct epithelialconnective tissue attachment. If mucosal thickness is insufficient, peri-implant marginal bone resorption occurs to allow appropriate supracrestal tissue height formation. Further clinical studies by Linkevicius and colleagues confirmed these findings in humans, suggesting that mucosal thickness is a significant influencing factor on periimplant marginal bone stability. 31,32 Subcrestal implant positioning was originally proposed as a clinical strategy to compensate possible reduction of peri-implant marginal bone levels during the first year of function. 33 However, studies performed in dogs using two-piece implants showed that the more subcrestal the micro-gap position, the greater the marginal bone loss. 13,34 Even if deeper implant insertion does not limit marginal bone resorption, 13,35 adaptation of apico-coronal implant placement may prevent colonization of the treated implant surface by the oral bacterial biofilm. In this regard, Linkevicius and colleagues distinguished marginal bone resorption occurring around subcrestal implants into two different components: bone remodeling (bone resorption occurring above the implant neck) and bone loss (bone resorption exposing implant neck and/or the underlying implant surface). 36 Unlike two-piece implants which present a microgap at crestal bone level, tissue-level implants have no gap in this region. 37 The absence of this gap seems to influence supracrestal tissue height formation around tissue-level implants. 38 However, only limited evidence exists of the relationship between apico-coronal positioning of tissuelevel implants, EMBL and vertical mucosal thickness.
The present multi-center prospective study aims to evaluate if EMBL occurring around tissue-level dental implants before prosthesis delivery could be compensated by adapting apico-coronal positioning to mucosal thickness. All the clinical centers participated to a calibration meeting prior to the study to discuss study protocol and standardize collection of experimental parameters in order to obtain acceptable inter-examiner consistency. Before implant placement, patients received oral hygiene instruction and professional deplaquing 1 week before surgery. After 5 months of healing, patients were referred to prosthodontists for subsequent rehabilitation.

| Radiographic measurements
Digital radiographs, customized with patient-specific bite jigs, were taken using a long-cone paralleling technique with a Rinn-type film holder at implant placement (baseline, T0), 3 months after implant placement (T1), and 5 months after implant placement, immediately before impression taking (T2) (Figure 3). All radiographs were performed using the same X-ray generator technology (FOCUS, KaVo), set to the same parameters (60 kV, 7 mA).
Peri-implant bone levels (PBLs) were calculated on each radiograph as the linear measurement of the distance between two points: the most coronal point of the implant platform and the most coronal bone-to-implant contact. Measurements were corrected referring to the known height and diameter of each implant. The vertical distance F I G U R E 2 Implants were placed based upon vertical mucosal thickness measured after buccal flap reflection: group 1 (thin mucosa <2.5 mm) with transmucosal portion 2 mm below crestal level; group 2 (medium mucosa between 2.5 and 3.5 mm) with transmucosal portion 1 mm below crestal level; group 3 (thick mucosa >3.  Gomez-Roman and Launer. 40 Examiner calibration was performed by measuring PBL on a sample of 10 radiographs not included in the study. Cohen's k coefficient for intra-examiner and inter-examiner agreement was 87.2% and 81.6%, respectively, for linear measurements within ±0.1 mm.

| Statistical analysis
Statistical analysis was performed by means of Primer of Biostatistics (6th Ed.) software. 41 The patient was considered as the statistical unit (one implant per patient). Considering three treatment group comparisons, a sample of 15 patients from each group was required to detect significant differences (confidence level 5% with statistical power of 80%), with an expected difference in MBL of 0.3 mm (±0.25 mm). 42 Data for descriptive statistics were expressed as mean ± standard deviation, and were analyzed using the one-way ANOVA test. Simple linear regression was used to analyze trends. Overall analysis for coincidence was performed to compare regression lines. 24 The null hypothesis (no difference in MBL among groups) was rejected for a critical significance level of p < 0.05.  ΔPBL variations over time are reported in Figure 5 and Table 2. Intra-group comparisons showed no significant differences after ANOVA test between mean T1 ΔPBL and mean T2 ΔPBL in groups 2 and 3, whilst in group 1 the difference was only marginally significant (p = 0.075). Inter-group analysis highlighted significant differences for mean T2 ΔPBL between the three groups after ANOVA test (p = 0.025).

| Radiographic measurements
The ΔPBL trend over time was also analyzed using simple linear regression and analysis of the coincidence of the regression lines (Table 3). The regression lines showed a significant direct relationship between ΔPBL and time (p < 0.05) for groups 1 and 2, but only a marginally significant correlation (p = 0.0871) for group 3. No significant differences were demonstrated comparing the intercepts of the three groups, whilst the slope inclination halved from group 1 to group 2 (p = 0.04), and became one third in group 3 with no significant difference between groups 2 and 3 (p = 0.55).
Overall analysis of the coincidence between groups 1 and 2 linear regressions ( Table 3)    This finding suggests that, in presence of thick peri-implant mucosa (>3.5 mm), very slight peri-implant bone resorption occurs during supracrestal tissue height establishment, in accordance with numerous previous clinical studies. 31 Notes: Data expressed in mm (mean ± standard deviation). p2 = probability after one-way ANOVA test (2 groups). p3 = probability after one-way ANOVA test (3 groups). Blue color indicates significant value (p < 0.05); red color indicates non-significant value (p > 0.05). In presence of medium peri-implant mucosa (between 2.5 and 3.5 mm), mean marginal bone resorption was slightly greater than in presence of thick mucosa, but followed the same trend over time with previous animal 29,30 and clinical studies. 31,32,48,49 However, two of the latter clinical studies 31,48  The method used to measure mucosal thickness is potentially questionable due to the deformability of soft tissue. However, similar methods have been used previously in numerous studies 25