Sample size calculation
The sample size was calculated with the G* Power v3.1 software for Mac (Heinrich Heine, Universität Düsseldorf), selecting the Wilcoxon-Mann Whitney test from the T-test family, and defining an alpha error of 0.05, a beta power of 0.8, and an N2/N1 ratio of 1. A total of 7 samples per group were indicated as the ideal size to detect significant differences. Considering the risk of fracture of the file, the sample size was increased by 20%. A total of 9 MB2 canals were used for each group.
After the approval of the protocol by the Institutional Research Ethics Committee (protocol: 4,716,078, CAAE: 46099021.2.0000.5417), 36 extracted first and second maxillary molars were selected. Digital periapical radiographs were used to measure the angle and curvature radius of mesiobuccal roots, according to the method proposed by Gu et al. . Roots with curvatures between 10° and 20° and curvature radii between 3 and 5 were selected. Then, the specimens were scanned with a micro-CT system (SkyScan 1174; Bruker-micro-TC, Kontich, Belgium) to select mesiobuccal roots containing MB2 canals with Vertucci type IV configuration.
Micro-ct Scanning And Sample Division
The 36 molars were scanned with micro-CT using the following parameters: voxel size of 19 µm, 50 kV, 800 µA, and 360° rotation around the vertical axis with a rotation step of 0.8° at a resolution of 1024x1304. The obtained images were reconstructed with the NRecon v1.6.9 software (Bruker-micro-TC) and saved in BMP format. Then, the MB2 canals were anatomically paired by calculating the canal volume, diameter, and length using CTAn v1.12 (Bruker-micro-CT). Data were analyzed by Kruskal-Wallis and Dunn tests for sample pairing; no statistical difference was found among the groups (P > 0.05). The specimens were randomly divided in 4 groups (n = 9), according with the kinematics used for glide path preparation:
Continuous Rotation (CR)
continuous rotation at 350 rpm and torque of 1-Ncm;
150° clockwise (CW) and 30° counterclockwise (CCW) reciprocation at 400 rpm (REC 30°/150°);
90° CW and 30° CCW reciprocation at 500 rpm;
Optimum Glide Path 90°
Optimum Glide Path (OGP) motion at 300 rpm and 90° CW, 90° CCW, 90°CW, and then 120° CCW motions.
Experimental Glide Path Preparation
All experimental procedures were performed by a single previously trained endodontist. After coronal access, the MB2 canal was located with an ultrasonic tip (Helse Dental Technology, Santa Rosa do Viterbo, São Paulo, Brazil) under an Operating Microscope (DF Vasconcelos, Brazil) with a 6x magnification. Then, a no. 8 or no. 10 K file was inserted with low resistance only to determine the correct insertion angle of the canals. Canal irrigation was performed with 2.5% sodium hypochlorite solution (NaOCl) with Navitip gauge 30 needle (Ultradent Products, Inc, South Jordan, UT, USA).
Each tooth was mounted on a specific device (IM do Brazil, São Paulo, SP, Brazil) that simulated the alveolar socket, and the glide path was prepared according to group allocation using the Prodesign Logic 15.03 file. For the continuous rotation and reciprocation groups (30°/150° and 30°/90°) the endodontic motor E-Connect S (MK Life, Porto Alegre, Brazil) was used and for the OGP 90° group, the Tri Auto ZX2 motor (J. Morita MFG, Kyoto, Japan) was used.
The glide paths were prepared as described by De-Deus et al. . The instruments were inserted with a smooth back-and-forth motion with amplitude of about 2 mm until the full working length (FWL) was reached. After three repetitions, the file was cleaned and the canal irrigated with 2 mL of 2.5%sodium hypochlorite. This step was repeated until the file reached the FWL. After three trials, the file was held in position and the apical foramen was examined under a dental operating microscope (DF Vasconcellos SA, Valença, Brazil) at 40x magnification. The canals in which the FWL was reached were recorded, as well as the number of fractured files. The time required for the procedure was recorded with a stopwatch (in seconds), excluding the time used for irrigation. In addition, the canals where the file failed to reach the FWL were radiographed using digital imaging (Microimagem, São Paulo, Brazil) and the distance of the file tip to the canal terminus was measured. Each file was used in 3 canals and then cleaned in saline solution for 3 minutes in an ultrasonic cleaning device (Gnatus, Ribeirão Preto, São Paulo, Brazil).
After glide path preparation, the specimens were scanned using the same parameters as the first scan to evaluate the canal volume before and after the procedure. Reconstructed images from before and after the procedure were geometrically superimposed, and the data were compared with the DataViewer software v1.5.2 (Bruker-microCT, Kontich, Belgium). The analysis included the binarization of the root canals and measurement of the volume (mm3) and surface area (mm2) of the full canals using CTAn v.1.14.4 . The total canal volume was measured from the root canal orifice to 1 mm short of the apical foramen. Also, the last 4 mm of the apical portion was evaluated. All values were calculated by subtracting the scores for the treated canals from the untreated ones and then converting the values into percentages.
Torsional Fatigue Test
The mean values of torque and maximum angular distortion to failure were determined for the files used in glide path preparation and compared with those of a new file. A total of 8 new files were used.
The torsion test used in this study was described in previous studies [22, 25] and was based on the International Organization for Standardization ISO 3630-1 (1992) using a torsion machine (Analog, Belo Horizonte, Brazil). The file was fixed by the first 3 mm of the tip with a brass chuck and the engine speed was set at 2 rpm clockwise for all groups. The maximum torsional force and angular rotation to failure were measured.
The data were subjected to the Kolmogorov-Smirnov normality test. ANOVA and Tukey tests were used for between-group comparison of procedure duration, number of files at FWL, number of fractured files, and torsion data. Kruskal-Wallis and Dunn tests were used to compare the distance between the file tip and root apex in the specimens where FWL was not reached. The Wilcoxon test was used for within-group comparison of canal volume and the Kruskal-Wallis and Dunn tests were used for between-group comparison. The GraphPad Prism 8 software was used and a significance level of 5% was applied.