Impact of bone defect morphology on the outcome of reconstructive treatment of peri-implantitis

Data from a single-blinded prospective longitudinal human randomized clinical trial [13] was used to evaluate the association between defect configuration and clinical evidence of healing. The study randomization process has been described previously [13].

The Institutional Review Board at Lund, Sweden, approved the study (id nr: 89: 2007). All study participants signed informed consent. Thirty-nine consenting individuals with 74 dental implants with a diagnosis of peri-implantitis demonstrating ≥ 2 bone wall lesions at the surgical intervention were enrolled in the present study between 2007 and 2010 at the Uppsala Käkkirurgiska Centrum, Sweden.

The distributions of implants included were the following: Brånemark, Nobel Biocare dental implants (74.1%) (Nobel Biocare Services AG, Kloten, Switzerland), Astra Tech implant system (13.7%) (Astra Tech AB, Mölndal, Sweden), Straumann implants (7.4%) (Institute Straumann AG, Basel, Switzerland), ImplaMed implants (2.8%) (Sterngold-ImplaMed™, Attleboro, MA, USA), and non-identifiable dental implant types (0.4%).

An update of the medical and dental histories was made. Before entering the study, participants with periodontitis were treated such that no pockets > 5 mm were present at the time of surgical intervention. The study participants also underwent a preparatory routine treatment phase, including mechanical debridement of teeth and implants. Hand instruments, and/or ultrasonic devices as designed either for teeth or dental implants were used. Before treatment, all study participants were instructed in oral hygiene measures. The oral hygiene instructions were reinforced at the recall visits as deemed necessary. No surgical intervention for study purpose was performed before the re-assurance of good patient motivation and compliance had been established.

Standardized intra-oral radiographs of implants were obtained using an Eggen holder and long cone equipped dental X-ray unit. Pre- and postoperative radiographs presenting the study implants were digitalized, coded, and evaluated using a computer program (OsiriX Imaging software 3.9 for MAC OS 10.6, Osirix Foundation, Geneva, Switzerland). The distance equal to three implant threads (known for each implant system in the analysis) was measured and used for the calibration of images. The mesial and distal bone level distances were measured from the implant platform to the most apical point of contact to implant. Radiographs were taken at baseline and 12 months. One calibrated examiner (GRP) who was unaware of the study group/procedures assessed the radiographs. Radiographs were studied both as black and white images and by using the CLUT (colour lookup tables) option. The most coronal confluent aggregation of bone or bone-like material was used to define the coronal bone level. Black and white images versus CLUT images were switched on and off to select the position of bone levels for the assessments. Single strands or islets of bone/radiopaque material were not considered.

One experienced examiner (UL) performed all clinical examinations. The examiner was unaware of treatment group allocation.

Before treatment, the following baseline recordings were performed:

The following clinical data were collected at baseline and 12 months after therapy:

Following the administration of local anaesthetics, a sulcular incision was made around the neck of the implant abutments, and full-thickness flaps were raised at the buccal and lingual surfaces to access peri-implant defects. After removing all granulomatous tissues and carefully cleaning the implants from mineralized calculus, the implant surfaces were cleaned with hydrogen peroxide (3%) for 1 min, followed by profuse rinsing with saline. Assessments of defect characteristics, including the extent of bone loss/vertical defects from the implant platform to the most apical bone defect, were made. The extent of bone loss/vertical defects from the implant platform to the most apical bone defect and the distance from the implant platform to the most coronal part of the bone was measured (in millimetre) at the mesial, buccal, distal, and lingual surfaces around the implant. The number of bone walls was assessed. These measurements were used to calculate the defect depth and to classify the defect as a 2, 3, or 4 wall defect.

The distance from the affected implant to a neighbouring tooth or implant was measured (in millimetre) as well as the width of the alveolar crest mesially and distally of the affected implant. Depending on the assigned treatment, the defect was then either filled by autogenous bone (AB) obtained from the ramus region using a bone scraper (Safescraper® TWIST; Biomet3i Inc., Palm Beach, FL, USA) or bovine-derived xenograft (BDX) (Geistlich Pharma, Wolhusen, Switzerland). A resorbable membrane (OsseoGuard®; Biomet3i Inc., Palm Beach, FL, USA) was used to cover the bone or bone substitute. The flaps were sutured using 4.0 sutures (Ethicon vicryl polyglactin, Johnson & Johnson, San Angelo, TX, USA) allowing non-submerged wound healing.

Postoperative antibiotics (Azithromycin®; Sandoz A/S, Copenhagen Denmark; 2× 250 mg day 1 and 1× 250 mg days 2–4) were prescribed to all study participants. During the first 6 weeks after surgery, all study participants rinsed with 0.1% chlorhexidine (Hexident, Meda AB, Stockholm, Sweden). During the first 3 days, they also received anti-inflammatory and analgesic medications (Ibuprofen 400 mg × 3 days; Ibumetin, Nycomed AB Stockholm, Sweden).

Defects were defined at the time of surgical intervention as ;(1) 2-wall defects with one of the following affected bone wall combinations a mesial-buccal, buccal-distal, mesial-lingual, distal-lingual, or a mesial and a distal defect; (2) three-wall defects requiring one surface without a bone wall, but with vertical angulated bone loss towards three implant surfaces; and (3) four-wall defect includes vertical angulated defects towards the implant at all four surfaces. The four-wall defect can also be described as a saucer-like defect.

Repeated measurements of bone levels on radiographs were made from 15 implants and assessed both at the mesial and the distal surfaces using baseline and year one images. The differences in bone levels between baseline and year one were studied and compared between the two sets of measurements. Analysis by paired t test identified a 0.1-mm (SD ± 0.4) bone level change difference between the two sets of measurements (95% CI 0.2 to 0.1, p = 0.36). The ICC coefficient was 0.97 (95% CI 0.95 to 0.99, p < 0.001).

The mesial and distal crestal width was significantly greater at 4-wall defects than at two wall defects (mean diff 1.8 mm, SE ± 0.4, 95% CI 0.7, 2.8, p < 0.001). No differences were found between 2- and 3-wall defects or between 3- and 4-wall defects. The relationships between crestal width and defect characteristics are illustrated (Fig. 1).

Statistical analyses failed to demonstrate plaque score differences, bleeding score differences, or differences in suppuration by defect configuration. Statistical analyses also were unable to show differences for the distances to both the mesial and the distal aspects of implant fixtures in relation to adjacent tooth/or another implant. There were no differences in the presence of attached buccal mucosa by defect type.

Probing pocket depths at all implants varied between 4 and 11 mm. At 2-, 3-, and 4-wall defects, PPD values ≥ 6 mm were found at 65.4%, 66.7%, and 100% of implants, respectively. Analyses by one-way ANOVA (Bonferroni post hoc tests) identified that mesial and distal PPD values at 4-wall defects were deeper than at 2-wall defects (p = 0.01 and p = 0.03, respectively). No differences were found between any of the comparisons of baseline PPD values made between 3- and 4-wall defects. The clinical assessments identified that defect depth was more significant at 4-wall defects both at buccal and lingual aspects than at 2-wall defects (mean diff 1.5 mm, SE ± 0.3, 95% CI 0.8, 2.3, p < 0.001 and mean diff 1.8 mm, SE ± 0.5, 95% CI 0.7, 3.7, p = 0.002). Clinical assessments of the defects are illustrated (Fig. 2).

Analysis by paired samples t test (equal variance not assumed failed to identify differences of bone defect depth when mesial (mean diff 0.2 mm ( SD ± 1.8, 95% CI − 0.2, 0.6, p = 0.34), distal (mean diff 0.2 mm (SD ± 1.7, 95% CI − 0.2, 0.6, p = 0.34), clinical, and radiographic assessments were studied (Fig. 3).

Analyses by Pearson’s correlation failed to demonstrate that changes in defect fill were correlated with changes in BOP, changes in plaque scores, or changes in PPD.

When all cases with 2-wall defects were included (groups merged), analyses by paired samples t test (equal variance not assumed) failed to demonstrated changes over time for plaque level changes, PPD changes, or radiographic evidence of defect fill. When all cases with 3-wall defects were included (groups merged), analyses by paired samples t test identified improvements in BOP scores (p < 0.001) and probing depths (mean diff 1.5, SE ± 0.3, 95% CI 1.5, 2.8, p < 0.001). The analyses failed to demonstrate differences in plaque scores or evidence of defect changes between baseline and year 1. At 4-wall defects (groups merged), the analysis identified radiographic evidence of defect fill (mean diff 1.2 mm, ± SE 0.5, 95% CI 0.1, 2.4, p = 0.05). Over time, probing depth reductions (mean diff 2.4, SE ± 0.5, 95% CI 1.3, 3.5, p < 0.001) and BOP % reductions were significant (mean diff 34.6, SE ± 10.0, 95% CI 12.8, 56.5, p < 0.01). No differences in plaque scores could be verified. The distributions of defect fill (millimetre), probing depth change (millimetre), change in BOP (%), and plaque index change (%) between baseline and year one by defect type and treatment are presented in Fig. 4a–d.

Analyses by Pearson’s correlation between baseline implant defect depth and the extent of radiographic evidence of defect fill were significantly related at both mesial (corr. coeff 0.32, p = 0.02) and distal aspects (corr. coeff 0.46, p = 0.001). More defect fill was found at deeper defects.