Eligibility criteria Consecutive informed and consenting healthy subjects (ASA 1-2) of both genders with no medical history of systemic disease and presenting a diagnosis of symptomatic or asymptomatic apical periodontitis in correspondence of endodontically treated teeth with medium (2-5 mm3) and large (≥5 mm3) periapical lesions were enrolled. The diagnosis was performed clinically and radiographically through periapical radiographs and small field of view (FOV) Cone-Beam Computed Tomography (CBCT). The lesions volume and diameters were evaluated using the dedicated software Mimics 24.0 (Materialise, Belgium). Only periapical defects with one or two residual cortical walls were considered, while bone cavity designs with no cortical walls were excluded. The clinical indications for retrograde endodontic retreatment were reported for all cases: the teeth presented inadequate root canal treatment with periapical radiolucency and impossible access through orthograde retreatment due to the presence of prosthetic restorations, posts, canal blocks and ledges. Moreover, the teeth presented an adequate coronal seal and absence of pathological periodontal records.
The subjects affected by diabetes mellitus, hypertension and hepatic or renal diseases were excluded from the study. Similarly, exclusion criteria included bleeding disorders, pregnancy and antiplatelet or anticoagulant therapy. The teeth that were considered unrestorable and the elements affected by fractures or extensive periodontal defects were also dismissed. A written consensus reporting information about the surgical procedure and the clinical and radiographical follow-up was supplied.
Sample size calculation The sample size calculation was estimated basing on clinical outcomes. In particular, basing on the VAS score reduction as primary endpoint, a study power of 80%, with a significance level of 5%, and a standard deviation of 6.24 was calculated and 10 subjects per set group were needed.
Randomization and allocation The patients were blind about the randomization that was performed by assigning a code. Double blinding was not possible due to the use of different hemostatic agents during the surgical procedure.
Treatment procedure The procedure was completed by one single experienced surgeon in a sterile environment using an operating microscope (OPMI Pro Ergo, Carl Zeiss, Germany). After the administration of local anesthesia with articaine 1:100,000 (Septodont, France) and 10 minutes later lidocaine 2% with epinephrine 1:50,000 (Dentsply Sirona, USA), a full thickness mucoperiosteal flap was elevated. The debridement of the periradicular lesion was obtained using surgical curettes (Hu-Friedy) and the bone crypt was completed with a dedicated bur (#557 SL Lindemann; Brasseler USA, Savannah, GA). After curettage, the surgeon and two independent assistants evaluated the hemorrhage degree and the visibility of the surgical field. Afterwards, in group 1 (n=15) the surgical glue (IfaBond, Dipromed, Italy) was applied for the hemorrhage control in the bone crypt with a needle 23G, while in group 2 (n=19) the ferric sulfate 15.5% (Astringedent) was positioned in the surgical site through compression with a cotton pellet. The qualitative analysis of bleeding score and the visibility of the surgical site after the use of the hemostatic agent was evaluated.
The apical resection of 2-3 mm with a 0º-10º bevel angle was achieved with a surgical bur (#557 SL Lindemann; Brasseler USA, Savannah, GA). Afterwards, a solution of 2% methylene blue was used to stain the apical portion of the root to visualize anatomies, isthmuses and fractures. In both groups the preparation of the apical 3 mm was achieved following the long axis of the root with a piezoelectric device (MiniPiezon Device, EMS, Switzerland) and dedicated tips (ProUltra Surgical Tips, Dentsply Sirona, USA). Root end filling was completed with superEBA sealer (SuperSeal, OGNA, Italy). After irrigation with saline solution and the removal of the hemostatic agent, the wound site was filled with blood clot, and the bleeding recovery was measured using the hemorrhage control score and comparing data with the previous records. Afterwards, the flap was sutured with a 5-0 resorbable suture (Vicryl suture, Ethicon, Switzerland). A post-surgical protocol was presented to the patients. The patients were advised to follow a specific diet, and to dismiss hard physical exercise and smoking. Mouth rinses with chlorexidine digluconate 0.2% mouthwash for seven days were prescribed. A pharmacological prescription including optional analgesics (ibuprofen 600 mg), and no antibiotics was given in any case. All patients were booked for suture removal after five days.
Outcomes According to a previous study, the hemorrhage control and the surgical field view were evaluated assigning a dedicated score. The hemorrhage control and the surgical field view evaluation was performed two times: after the curettage of the bone lesion before the use of the hemostatic agents, and after the hemostasis. The bleeding recovery was evaluated after the removal of the hemostatic agent.
The patients' quality of life (QoL) questionnaire Oral Health Impact Profile-14 (OHIP-14) assessed the impact of the surgical procedure on the signs and symptoms affecting the oral and general functions. The questionnaire evaluated different domains: functional limitations, physical pain, psychological discomfort, physical disability, psychological disability, social disability and handicap. A 5-point Likert-like scale (1 = never, 2 = hardly ever, 3 = occasionally, 4 = fairly often, 5 = very often) was introduced. Pain assessment was conducted using the Numerical Rating Scale (NRS). In this scale a 20-point difference between groups is considered clinically significant. The subjects were asked to record the amount of pain they felt every day after the surgical procedure. Moreover, patients were asked to fill out the frequency of the analgesic intake from immediately after surgery up to 1 week. All subjects were requested to fill out the questionnaire and the NRS scale and to record possible symptoms during the first seven days after surgery. To ensure the correct recording of these forms, they were contacted two times in the first week to check on their status and to answer any possible dilemma regarding the filling of the documents.
Statistical analysis A blinded statistical analysis was performed. A chi-square-test was used for the analysis of the hemorrhage control, while the Fisher t-test was utilized for the quality of life and analgesic intake. Inter-examiner reliability was assessed using Cohen's Kappa statistic. Moreover, the Mann-Whitney U test was used for pain assessment on postoperative days. The statistically significant level was settled with a P value \< .05. The software SPSS Statistics v.23.0 (IBM Corp, Armonk, NY) was used for the analysis.
