Reconstruction of Peri-implant Osseous Defects: A Randomized Trial

Scientific rationale for study:

There is a paucity of data for the effectiveness of reconstructive procedures in the treatment of peri-implantitis. Recently, titanium brushes have been introduced for implant surface debridement and porous titanium granules (PTG) as an osteoconductive bone substitute.

Study design The study was designed as a prospective multicentre, multi-national, randomized, parallel group-controlled clinical trial of 6 months duration.

All participating investigators attended training and calibration meetings to discuss and optimize patient accrual and retention as well as data management. Preliminary clinical cases were discussed to standardize case selection, the measurement techniques and surgical procedures. On-site rules, for the compilation of the data collection sheets for appropriate oversight were frequently reassured by a study monitor to ensure the validity of the data.

Open flap debridement and surface decontamination with titanium brush (n = 30/Control), or additional insertion of PTG (n=33/Test) was performed for reconstruction of 3- or 4-wall intra-osseous peri-implant defects with at least 3 mm defect depth and circumferential (≥270°) morphology. Patients were given amoxicillin and metronidazole for 8 days. Radiographic and clinical assessments were performed at baseline and 6 months.

Study population Study subjects were recruited consecutively from patients treated by dental specialists in periodontology or implantology in Germany (University of Bonn), the Netherlands (Amsterdam), Italy (Rome University), Spain (University Madrid) and Sweden (Kristianstad University). Each participant signed an informed consent in accordance with the Helsinki Declaration of 1975 as revised in 2008. The Ethical Committee for human subject trials in each institution approved the study protocol individually. Seventy informed and consenting patients above the age of 18 years of age with a diagnosis of peri-implantitis were enrolled in the study and surgery was performed.

Pre-surgical treatment and evaluation Clinical screening included medical history, physical examination and admission criteria. All necessary endodontic and periodontal treatments were finished as evaluated with recording of pocket probing depth (PPD), inflammation (BoP) and plaque at least 1 month prior to the peri-implant surgical procedure and to entry into the study. Potential prosthetic factors to the aetiology of the disease process were resolved prior to study treatment.

Patients who met all criteria for inclusion, verified at surgery, underwent baseline measurements and investigational procedures.

Clinical measurements Baseline measurements at the included implant site were performed on the same day as the surgical procedure. All probing measurements were obtained with a pressure (0.20-0.25N) sensitive probe (Click-Probe®, Kerr, Switzerland) to the nearest mm at 6 sites per tooth/implant (mesio-buccal, buccal, disto-buccal, disto-palatal, palatal, mesio-palatal).

At baseline and 6-months visits, the following recordings were taken:

• Probing pocket depths (PPD)
• Bleeding on probing (BoP)
• Suppuration (PUS)
• Plaque BoP and PUS at the affected implant sites were assessed within 30 seconds following probing of the pocket/implant mucosa.

Intraoperative measurements during surgery

• Intra-osseous defect component (mm) at the deepest point.
• Circumference of the osseous defect (degrees), 3 or 4 wall intra-osseous component.
• Defect width (mm).

Radiographic measurements Intra-oral peri-apical radiographs were obtained using a reproducible technique with standard Eggen holders and long cone equipped dental X-ray units. All radiographs presenting study implants were digitalized, coded and evaluated by a computer program (MATLAB® Vers. R2013b software for MAC OS 10.9; Mathworks, Natick, MA, USA).

Radiographs were studied by changing parameters in black and white modus as well as in colour look-up tables. Measurements from a well defined reference point at the coronal part of the implant body taken at baseline and 6-months visits were: vertical defect depth and width, marginal bone level and horizontal bone level. Based on these measurements, changes in vertical defect depth, marginal bone level, % defect fill and % defect resolution from baseline to 6 month were calculated.

The most coronal confluent aggregation of bone or graft material was used to define bone levels. Single isles of bone or bone-like material were not considered.

Implant length and width or known dimensions of implant-threads were used as reference for calibration of measurements. The implant abutment interface served as well defined landmark reference for follow-up recordings. Radiographic evaluations were performed by a physicist trained on sample images on a LCD-display with a resolution of 2560 x 1600 and a 32-Bit colour pixel depth and two periodontists.

Sample size calculation/Power analysis The calculation of patients to be treated (sample size) was based on a previous proof-of-concept single-centre RCT (Wohlfahrt et al. 2012) and the primary objective to detect a true mean difference of at least 2 mm between test and control treatment for radiographic defect fill after 6 months. With a level of significance of alpha = 0.05 in a two-sided hypothesis and a power of 90 %, the number of patients needed was 48. Assuming a dropout rate of about 30% the total number of patients required was 60.

Randomization, Allocation Concealment and Blinding Patients were randomly assigned to treatment modality by using a randomization scheme prepared prior to study initiation and to treatment in blocks of 6. The patients were allocated to either reconstruction of the defect with PTG (Test group) or closure of the flap after implant debridement (OFD/Control). Documentation of treatment allocation for each patient was placed in separate, sealed opaque envelopes that were opened and revealed to the surgeon after debridement of the defect and implant surface was finished. Clinical examiners remained blinded to the treatment assigned.

Interventions Pre-Surgical Phase All patients had a full periodontal examination with probing of pocket depth (PPD) and registration of full mouth bleeding and plaque scores. Pre-surgical interventions included oral hygiene instructions to the individual needs of the patient and a complete non-surgical and surgical periodontal treatment.

Surgical Procedure A non-submerged surgical technique was used for both experimental and control sites. Following administration of local anaesthesia flap elevation procedure included an intra-crevicular incision around the implant. Full thickness mucoperiosteal flaps were raised on the buccal and lingual aspect to gain access to the complete peri-implant defect and to the implant surface. The extent of this incision was determined by supra-crestal incisions extending mesial and distal of the implant site. Vertical incisions into the vestibule at a distance of at least one tooth/implant from the implant were performed as necessary for adequate access. Granulation tissue was removed using titanium curettes (HuFriedy®, Chicago, IL, USA) and the exposed implant surfaces were cleaned mechanically by using a rotary titanium brush (Tigran PeriBrush™, Tigran Technologies, Malmö, Sweden) and decontaminated chemically with 3% H2O2 for 1 minute followed by rinsing with saline for 60 seconds (2 x 20 ml).

Following treatment allocation in accordance to randomization, titanium granules (Tigran™, Tigran Technologies, Malmö, Sweden) were applied into the osseous defects of the test sites. After insertion of the granules, excess material was carefully removed. Flaps were then repositioned and sutured back into position using monofilament non-resorbable sutures.

Peri-operative protocol Patients were prescribed a combination of Amoxicillin 500 mg 3/day for 8 days and Metronidazole 400 mg 2/day for 8 days, starting one day prior to surgery. Patients were then instructed to rinse twice daily with chlorhexidine mouth rinse (0.2%) for 1 month. The patients used brushes as usual in other areas of the mouth. Anti-inflammatory and analgesic therapy was prescribed (Ibuprofen 3 x 600 mg/day) during the first two days and according to the individual needs thereafter.

The sutures were removed after 7 to 14 days and patients were instructed in the use of soft toothbrushes and soft interdental brushes (super soft - Gentle/Implant Care TePe, Malmö, Sweden) in the surgical area.

Patients were recalled at 6 weeks, 3, 6 months after surgery for professional oral hygiene procedures with individually based supra-gingival debridement and hygiene instructions as needed.

Early withdrawal criteria:

• non healing infections
• substantial exfoliation of graft material (rejection)
• local intolerance to graft material
• recurrence of active peri-implantitis at the test site
• loosening of implant
• poor patient cooperation, not returning for control visits

Statistical Analysis Statistical analysis of primary efficacy endpoint was performed using stratified Wilcoxon test, using centre as stratification variable.

Treatment intergroup comparisons of secondary efficacy endpoints were based on least squares means obtained from the ANCOVA model. Least-square means for each treatment group and differences in least-squares means between treatment groups are presented along with associated 95% Confidence Intervals (CI) as well as p-values for differences within treatment groups.

All analysis was performed using SAS version 9.2 (SAS Institute Inc., Cary, NC, USA).

The statistical hypotheses for the Primary Objective was:

H0: Δ bone level Test - Δ bone level Control = 0 H1: Δ bone level Test - Δ bone level Control ≠ 0

For percent changes the statistical hypotheses was based on:

• Defect resolution=(Vertical defect baseline- Vertical defect 6months)/( Vertical defect baseline) x100
• Defect Fill = (Marginal bone level baseline- Marginal bone level 6months )/( Vertical defect baseline) x100

If the p-value from this analysis fell below 0.05 in both mesial and distal measurements it was concluded that there was a statistically significant difference in average change between the two treatment groups.