This randomized, single-blind clinical trial was conducted according to the PRIRATE 2020 (Preferred Reporting Items for Randomised Trials in Endodontics 2020) Guidelines. The ethics committee approval was obtained from the local university clinical research ethical committee with the decision numbered B.30.2.ATA.0.01.00/1 and registered in the clinical trials database (www.clinicaltrials.gov) (NCT06307678).

The sample size was calculated using the G*Power program (Franz Faul, University of Kiel, Germany) based on data from a previous study that evaluated mediator levels in periapical lesions [17]. It was determined that at least 20 patients should be included per group (α = 0.05, power = 0.95, effect size = 1.0612). To enhance statistical power and account for possible dropouts, each group was determined to consist of 30 patients. All the participants fully understood the purpose of the study and signed informed consent forms.

The patients ranged in age from 18 to 59 years. Incisor, canine, and premolar teeth that had previously undergone root canal treatment, exhibited no pain or swelling, showed negative responses to palpation and percussion, and had a periapical index (PAI) score of > 2 with a diagnosis of asymptomatic apical periodontitis were included in the study. Periapical radiographs taken from different angles were used to assess the periapical status and root canal anatomy. To prevent difficulties in preparation, obturation, and restoration in molar teeth and to standardize treatments, only teeth with a single root canal were included in the study.

Patients classified as ASA II or higher, pregnant women, and those with a Schilder canal curvature exceeding 25° or internal/externsl root resorption were excluded. Additionally, patients with generalized periodontitis, or periodontal pockets greater than 3 mm, as well as those who had used NSAIDs within the past 24 h or antibiotics within the last 3 months, were not included.

The randomization procedure was carried out using the online platform www.randomizer.org, which generates random numbers for assigning participants to experimental groups. Patient allocation followed the sequential order determined by this tool. The randomization process and its implementation were managed by the secretary.

The teeth were isolated with a rubber dam, the crown and surrounding structures were disinfected with 30% H2O2 for 30 s and 2.5% NaOCl was applied for the same period. Subsequently, 5% sodium thiosulfate was used to deactivate the effect of the NaOCl. Traditional access openings were prepared following complete caries excavation and removal of any previous restoration. The filling material from the cervical third was removed using Gates Glidden burs. The working length was determined using an electronic apex locator (Propex Pixi, Dentsply Sirona, Ballaigues, Switzerland) and a 15 K file (Mani Inc.; Utsunomiya, Tochigi, Japan). The length was established as soon as the “over” signal was observed on the apex locator, and the procedure was performed 0.5 mm short of this length. The working length was then confirmed using radiography.

Root canal shaping was completed using R25 and R50 Reciproc files (VDW, Munich, Germany) at the working lengths to finalize the canal preparation. The R25 file (VDW) was used to remove the root canal filling material, while the R50 file (VDW) was used to complete the root canal shaping. After each file was used, the root canals were irrigated with 2 ml of 1% NaOCl. Before obtaining the samples, the canals were irrigated with 5 ml each of 1% NaOCl, neutralized with 0.5% sodium thiosulphate, and then irrigated with distilled water, with each step lasting for 1 min. A sterile paper point (size #20; Dentsply Maillefer, Ballaigues, Switzerland) was placed into the root canals beyond the 2 mm root apex and were kept in this area for 1 min.

The procedure was repeated with three paper cones. Paper cones were cut 4 mm from the tip and the samples was stored in Eppendorf tubes containing a phosphate buffer saline at − 80 °C.

Afterward, the root canals were dried with paper points, and two groups were planned according to the intracanal treatment group. Calcium hydroxide based intracanal dressing (Calcicur) or tricalcium silicate based intracanal dressing (Bio-C Temp) was placed in the canal, 1 mm shorter than the working length using the applicator tip.

The canal orifices were temporarily closed using a sterile teflon pallet and the coronal cavities were restored with Cavit G (3M ESPE, Seefeld, Germany). Radiographs were taken to confirm the presence of the dressing placed in the canal (Fig. 1).

Periapical radiographs were taken in patients to assess the placement of intracanal dressing material. A Calcium silicate-based intracanal dressing, B calcium hydroxide-based intracanal dressing

Seven days later, the root canals were accessed aseptically under rubber dam isolation using the disinfection protocol as previously described. The medication was mechanically removed using a master apical file and irrigated with 5 ml of distilled water. Then, the root canals then filled with 5 ml of 17% EDTA (Endo-Solution; Cerkamed, Wojciech, Poland) and the solution was activated using the EndoActivator (Dentsply Tulsa Dental Specialties, Tulsa, OK). Finally, the canals were irrigated again with 5 ml of saline solution. The final samples were collected from the apical tissue as previously described and stored at − 80 °C. The root canals were filled with cold lateral condensation technique using guta percha cones and sealer (Sealapex, Sybron Kerr, Brea, CA).

The RANKL, OPG, TNF-α, TGF-β, and PGE2 levels in the samples taken from the interstitial fluid were measured on the same day to avoid inter-day variation. The serum PGE2, RANKL, OPG, TNF-α, and TGF-β levels were measured through the ELISA method using a human PGE2 ELISA kit (Sunred, China), a human RANKL ELISA kit (Sunred, China), a human OPG ELISA kit (Sunred, China), a human TNF-α ELISA kit (Sunred, China), and a human TGF-β ELISA kit (Sunred, China) according to the manufacturer’s instructions. Eppendorf tubes stored at − 80 °C were thawed gradually 1 day before the study. The samples collected on the study day were vortexed, the tubes were placed in the centrifuge, and the paper points were collected at the bottom. Briefly, the procedure applied for measurement is as follows: in 96-well microplates coated with specific monoclonal antibodies against human PGE2, RANKL, OPG, TNF-α, and TGF-β, serum and standard solutions obtained using serial dilutions at decreasing concentrations were added. The PGE2, RANKL, OPG, TNF-α, and TGF-β molecules in the samples were bound to these coated antibodies. The unbound molecules were removed by washing. A second antibody specific for PGE2, RANKL, OPG, TNF-α, and TGF-β and labelled with biotin was added to the wells. After another wash, the peroxidase enzyme bound with streptavidin was added. The peroxidase enzyme in this complex bound with avidin oxidized the 3,3′ 5,5'-tetra-methyl benzidine added to the medium and caused a color change in direct proportion to the concentration of PGE2, RANKL, OPG, TNF-α, and TGF-β in the samples. Acid was then added to each well to stop the reaction. The absorbance values of each well were measured with a spectrophotometer at a wavelength of 450 nm. The PGE2, RANKL, OPG, and TNF-α concentrations in each sample were calculated in ng/L and the TGF-β concentration in ng/ml from the absorbance–concentration graph using standards prepared at decreasing concentrations, and their levels were evaluated using an ELISA reader.

The data were analyzed using IBM SPSS v23 and IBM AMOS v24 (SPPS Inc, Chicago, IL, USA). Compliance with normal distribution was analyzed using the Shapiro–Wilk test and the multivariate normality assumption. The Mann–Whitney U test was used to compare non-normally distributed data, and independent two-sample t test was used to compare normally distributed data. In the comparison of non-normally distributed data over time within groups, the Wilcoxon test was used, while the paired two-sample t test was used to compare normally distributed data.

The Fisher–Freeman–Halton test and Yates correction were used to compare the categorical variables by groups. Path analysis was used to analyze the effect of the independent variables on the percentage changes, and the maximum likelihood method was used as the calculation method. The results of the analyses were presented as the mean ± standard deviation, and the median (minimum–maximum) was presented for the quantitative data. The significance level was set at 5% (p < 0.05).