This study aims to evaluate the levels of cytokines, chemokines and growth factors in peri-implant sulcular fluid (PISF) during healing and osseointegration at osteotomy sites prepared either with piezosurgery (PS) or drills (D). Fourteen patients having bilateral partial edentulism in the posterior maxilla were enrolled and 38 osteotomies were prepared. Implants were placed with one-stage surgery. Insertion torque, early healing index, probing depth and modified gingival and plaque indices and crestal bone (CB) loss were measured. PISF was collected from 4 sites from each implant at weeks 2, 4, 8, 12 and 24. PISF samples were analysed by a 30-Plex immunoassay. Effect of time and osteotomy method on molecules employed Brunner-Langer method.

Surgical procedures Fourteen patients received 38 bone-level implants (4.1 mm diameter and 8 mm, 10 mm or 12 mm in length, Biodenta®, Bone Level Implant, Biodenta Swiss AG, Switzerland) Osteotomies were prepared with drills (drill group, control, n=19) on one side and with PS tips (piezosurgery group; test, n=19) on the contralateral side in a single session. Surgical, pre- and post-operative procedures were carried out as was previously described. Briefly, right side was always the first operated site where the osteotomies were prepared with one of the randomly selected methods. Toss of a coin at the beginning of the surgery by an independent examiner designated the random allocation and modality to be used on the right side of the patient. Left side received the other osteotomy modality. First a midcrestal incision was made and full-thickness flap was raised. Osteotomies were prepared in the drill group by marking the appropriate point with a trispade drill and then advancing the 2.0 mm pilot drill to the planned depth followed by 2.5 mm, 2.8 mm and 3.5 mm drills, respectively. In PS group osteotomies were prepared with PS device (Piezonmaster®, EMS SA, Switzerland) and its relevant tips (Swiss Instruments Surgery, Implant System, Switzerland). Initial tip with a 1.15 mm diameter was used along the predetermined depth to create a pilot osteotomy. Then the osteotomy was widened to a final diameter of 3.5 mm by using 1.95 mm, 2.5 mm, 2.8 mm, 3.05 mm and 3.3 mm tips, respectively. Intermediate and final diameters, depth and direction of the osteotomies were controlled in both groups with drill try-ins, which also function as paralleling pins. Bone taps or crestal drills were not used for final contouring in both groups. Then 4.1 mm-diameter implants were placed equicrestally by a handpiece at a speed of 15 rpm in both groups and insertion torque was recorded. Following transfer abutment removal, 4.0 mm diameter straight healing abutments were connected for non-submerged healing and flaps were stabilized with 5.0 polypropylene interrupted sutures. Patients were instructed to rinse with 0.2% chlorhexidine gluconate for 2 weeks and to abstain from brushing the surgery site for this period and not to chew on the healing abutments. They were prescribed 200 mg ibuprofen t.i.d for 1 week. The sutures were removed at 2nd week following the surgery.

Clinical and radiological procedures A single examiner performed clinical measurements. Modified gingival (MGI) and plaque indices (MPI) were taken on weeks 2, 4, 8, 12 and 24 from 4 points around each implant with a plastic probe (UNC 12 Colorvue probe, Hu-Friedy, Chicago, IL). Probing depth (PD) was measured on weeks 12 and 24 following surgery with the same probe type (Figure 3). Repeatability of the examiner for PD measurements was κw=0.88. Flap closure and its continuity were evaluated on days 7 and 14 by early healing index (EHI), which had been originally described for postoperative monitoring of regenerative procedures for intrabony defects.

Crestal bone level measurements were performed as previously described. In brief, radiographic images were obtained by cone-beam computerized tomography (CBCT) (Kodak 9000 3D, Practice Works, Inc., Atlanta, USA) on the day of surgery and at week 24. Standardized periapical radiographs were obtained at week 12 using a photostimulable phosphor plate with position holders (Rinn XCP, Dentsply International) and the long-cone paralleling technique. Images were digitalized by a photostimulable phosphor plate scanner (Digora® Optime, Soredex, USA).

CB levels on radiographic images were measured with a Java-based software (Image-J 3.0, NIH, Bethesda, USA) by a masked and calibrated examiner (GPT; Cronbach's alpha=0.99). Implant shoulder (IS), first bone to implant contact (fBIC), implant abutment interface and apex of the implant were used as reference points. Mean of triple measurements rounded to the nearest 0.01 mm were used. CB loss was recorded by measuring the IS-fBIC distance on periapical radiograms at week 12 and on CBCT sections at week 24.

Biochemical procedures Postoperative PISF samples were obtained from 4 aspects of implants on weeks 2, 4, 8, 12 and 24. Sites were isolated by cotton rolls and visible supramucosal plaque was removed from healing abutment surfaces with a fiber carbon curette before sampling. Following gentle air-drying, paper strips (Periopaper, ProFlow, Amityville, NY, USA) were inserted 1 mm into the crevice and left in place for 30 s. Care was taken to avoid mechanical injury. The PISF volume absorbed on each strip was then determined by means of an electronic impedance device (Periotron 8000, ProFlow, Inc., Amityville, NY, USA), and all four were pooled into a sterile polypropylene tube which was previously coded to ensure masking of the laboratory technician and kept at -40C until analysed. The readings from the Periotron 8000 were converted to volume (µl) by reference to the standard curve. The collected PISF samples were eluted in 450 µl phosphate buffer saline (PBS, pH 7.2) in the presence of EDTA-free protease inhibitor cocktail ( Roche Applied Science, Rotkreuz, Switzerland) and centrifuged at 2000 x g for 15 min, at 4C. The levels of the molecules under investigation in the eluted PISF samples were determined by the cytokine human magnetic 30-Plex panel (Novex®, ThermoFisher Scientific, Waltham, MA, USA) consisting of cytokines (G-CSF, GM-CSF, IFNα, IFNγ, IL1β, IL1RA, IL2, IL2R, IL4, IL5, IL6, IL7, IL8, IL10, IL12 (p40/p70), IL13, IL15, IL17, TNFα), chemokines (Eotaxin, CXCL10, MCP1, MIG, MIP1α, MIP1β, RANTES) and growth factors (EGF, FGF-basic, HGF, VEGF), on the Luminex®200 platform. Bead fluorescence readings were done by Luminex®200 and analyzed using a software (xPONENT®, ThermoFisher Scientific, Waltham, MA, USA).

Data analysis A statistician who was blinded to the groups performed data analysis using non-parametrical methods. Implants were used as the unit of analysis. Clinical and radiological parameters served as primary outcome variables.The secondary outcome variable was selected as cytokine, chemokine and growth factor levels. Both primary and secondary outcomes were tested with Brunner and Langer method LDF2 model using a software (R software, version 3.3.1, package: nparLD, R Foundation for Statistical Computing, Vienna, Austria; r-project.org). The following hypothesis was tested: "Changes in PISF cytokine, chemokine and growth factor levels are dependent on preparation method of osteotomy (piezosurgery vs drilling) and time after surgery".

Week 2 values of RANTES in study groups were compared by Mann-Whitney U test. RANTES values of groups at weeks 4, 8, 12 and 24 which were calculated as the difference from week 2 values were compared with Bonferroni corrected Mann-Whitney U test. EHI scores were compared with McNemar-Bowker chi-square test. Examiner calibration was assessed by weighted kappa and intraclass correlation coefficient methods for PD and radiological CB loss measurements, respectively with a statistical software (SPSS 20.0, SPSS for Windows, SPSS Inc., Chicago, USA).

Significance level was set at 5% for all analyses. Required sample size was calculated using a software28 (G*Power 3.1, version 3.1.9.2) estimating a power of 80%, p-value of 5% in study groups for one-tailed test of matched pairs. Sample size calculation analysis suggested at least 18 implants for both groups.

Inclusion Criteria:

• Bilateral partial edentulism in posterior maxilla
• >8 mm available bone height (distance between bone crest and maxillary sinus) and ≥7 mm bone width
• At least 2 mm buccal keratinized mucosa width and 3 mm mucosa thickness.

Exclusion Criteria:

• Diseases and conditions or medications which may negatively influence biological dynamics of bone and wound healing
• Ridge deficiencies requiring additional augmentation
• Indication of sinus lifting with crestal or lateral approach
• Endodontic or periodontal lesions neighbouring the edentulous sites were anatomic exclusion criteria