Study design
The present study is a prospective randomized, triple-blind controlled trial. The study was conducted in full accordance with the World Medical Association’s Declaration of Helsinki, as revised in 2008. The study protocol was approved by the Committee of Research and Ethics of the Dental School of the National and Kapodistrian University of Athens, Greece and registered in a publicly accessible database (clinicaltrials.gov/NCT 04934813). Written informed consent was provided by all study participants. The CONSORT statement was followed for the reporting of the present work [16].
Participants
Subjects were recruited from the patients proceeding to the post-graduate clinic of Periodontology of the Dental School of the National and Kapodistrian University of Athens, in Greece.
All candidates required extraction of at least one non-restorable single-rooted tooth in the maxilla which would be replaced by an implant as part of their ongoing treatment plan. Further inclusion criteria were: systematically healthy patients, non-smokers, no use of antibiotics for the last 3 months. Patients were excluded when a bony wall loss of more than 50% was verified for the socket after the extraction.
Randomization, allocation concealment and blinding process.
Randomization was performed by an independent dentist (YB), using a computer software (randomizer.org). The results were concealed in envelopes, and opened by the therapist (IES) immediately after tooth extraction and only if the socket met the inclusion criteria. If a patient had 2 or more teeth extracted, the respective number of envelopes were opened. The patients were blinded regarding the applied bone graft while the periodontist who measured the alveolar ridge dimensions at re-entry (EO) and the data analyst (KD) were all blinded regarding the treatment performed.
Each individual was assigned to one of the following four groups:
Group A: cancellous bone allograft with granules size 0.2–1.6 mm (Phoenix® cancellous bone powder, TBF) was placed to the bone crest and sealed with a collagen sponge (Jason® fleece, Botiss).
Group B: cancellous bone allograft with granules size 0.2–1.6 mm (Phoenix® cancellous bone powder, TBF) was placed to the bone crest and sealed with an autogenous FGG (soft tissue punch).
Group C: demineralized bovine bone mineral xenograft (Cerabone® bone mineral, Botiss) was placed to the bone crest and sealed with an autogenous FGG (soft tissue punch).
Group D: the socket was sealed with an autogenous FGG (soft tissue punch), while no bone graft was used.
Tooth Extraction And Grafting Procedure
Teeth were extracted in an atraumatic manner, using periotomes, thin luxators and forceps in order to maintain the anatomy of the buccal and lingual bony plate. Then, the granulation tissue was thoroughly removed and the integrity of the bony walls was evaluated. Clinical measurements were performed by means of a surgical stent, as described below, and patients were randomly assigned in one of the aforementioned four treatment groups.
When bone graft was applied into the socket, the allograft or xenograft was first placed in sterile saline for at least 5–10 minutes, as indicated by the manufacturer. Then, the bone graft was condensed into the socket and covered with the corresponding socket barrier. In case of a FGG sealing, a graft of 2–3 mm thickness was harvested from the patient’s palate. Subsequently, the soft tissue margin of the socket was de-epithelialized using a diamond bur under copious water irrigation and the tissue punch was stabilized with 6–8 single interrupted monofilament non-resorbable sutures (Seralon® polyamide 6/0), as previously described [15]. In cases where a collagen sponge was used, the material was stabilized with a “figure 8” suture technique, using the same suture material (Seralon® polyamide 6/0) (Fig. 1, 2).
Patients were instructed to rinse twice a day for 2 weeks with 0.2% chlorhexidine (Chlorhexil® 0,2%, Intermed) and nonsteroidal anti-inflammatory drugs (ibuprofen 400 mg, tid) and antibiotics (amoxicillin 500 mg, qid) were prescribed for 5 days. Patients were reevaluated at 10–14 days, and had their sutures removed, and then were followed-up until complete soft tissue healing.
Four months after tooth extraction, a full thickness flap was elevated at the site of extraction and prior to implant placement, the alveolar ridge dimensions were measured.
Clinical Measurements Of The Alveolar Ridge Dimensions
The measurements of the alveolar ridge dimensions were evaluated immediately after tooth extraction and 4 months later. In order to ensure standardization of the clinical measurements, surgical stents were fabricated for each site as previously described[17, 18]. In brief, prior to tooth extraction, study casts of the patients were fabricated. Subsequently, the surgical stents were constructed using a 1.5 mm thick hard celluloid shells which adapted at least on 2 adjacent teeth to ensure stability during the measurements. Four holes were prepared at the occlusal aspect of the stent, 1 central-buccal, 1 mesial, 1 distal and 1 central-lingual above the post-extraction socket, and served as reference points in order to measure the vertical dimensional alterations after the extraction. In addition, 2 holes were created at the buccal and 2 at the lingual aspect of the alveolar ridge, 2 mm and 4 mm apical to the bone margin of the socket, in order to measure the horizontal dimensional changes of the alveolar ridge (Fig. 3).
The distance between the above mentioned reference points and the alveolar crest was noted with an endo file, which had a stopper on it, and measured with a digital caliper (Logilink digital caliper WZ0031). For the horizontal measurements, a modified digital caliper with sharp edges was used to measure the ridge width. The edges passed through the soft tissue (bone sounding), in order to avoid flap elevation. All distances were measured twice to ascertain the reproducibility of the process. The initial measurements were performed by a periodontist (IES), while the measurements at reentry were performed by a different periodontist (EO) who was blinded on group allocation of the patients. Both periodontists were calibrated prior to the initiation of the trial and the inter-rater agreement was assessed through Bland- Altman plot, with the respective limits of agreement (LoA). Mean difference between interrater measurements were at the level of -0.01 (95% Confidence Interval, CI: -0.08, 0.05), with LoA ranging from − 0.19 to 0.16. The Pitman’s test of difference in variance was non- significant (p = 0.29), as well.
Statistical analysis
Based on previous data [5, 19] and assuming that a difference of 2.0 mm could be considered clinically significant (with common SD = 1.5) and with an alpha error rate 5% and power = 0.8, we calculated that a sample size of 10 patients/group were necessary. Data distribution was checked for normality of residuals visually through qqplots and statistically through Shapiro- Wilk tests. As normality could not be confirmed, non- parametric descriptive statistics were used. For the regression model and due to the small sample size, standard errors were derived using the bootstrap method with 1,000 replications.
Likelihood ratio tests were applied between multilevel and original models to test evidence of clustering effects within the sample. As this was not confirmed, we proceeded with the original model. In essence, a multivariate linear regression model was fitted with derived standard errors (SE) through the bootstrap method with 1,000 replications, for the effect of intervention, age and tooth type on the six correlated outcomes, namely: average bone change in the vertical mesial region, bone change in the vertical buccal central, bone change in the vertical palatal central, average bone change in the vertical distal as well as bone change in the horizontal region at 2 mm and 4 mm, within the socket of extracted teeth.
Box-plots were constructed to allow for an assessment of alveolar bone changes, by intervention category and across examined regions of the socket.
The level of statistical significance was pre-specified at p < 0.05 (two- sided). Statistical analyses were performed with STATA version 15.1 software (Stata Corporation, College Station, Tex, USA).