CAD/CAM Milled Glass Fiber Posts: Adaptation and Mechanical Behavior in Flared Root Canals

The rehabilitation of endodontically treated teeth requires an intraradicular post cementation in cases where considerable coronal destruction is found.^1,2  Different materials are used to produce the posts,²  and glass fiber-based posts have been widely used in clinical practice. The esthetic properties and mechanical behavior of these materials closely resemble those of dental hard tissues, contributing to their high acceptance and success for prosthodontic treatments.^3–6  The failure rate for such treatments ranges from 7% to 11% after up to 10 years.^3,4,7  Most of the failures are related to root fracture and loss of post retention.^2,7,8 

The clinical success of fiber posts depends on dental geometry, the position of the tooth in the arch, adaptation, cementation technique, residual coronal structure, and ferrule presence; an effective bonding is related to successful restorative treatments.^9–11  When glass fiber posts are used in these conditions,lower stress in the post-cement interface and a lower elastic modulus reduces the fracture risk.^12–13  The success of this technique may be impaired by the presence of flared root canals, where extensive loss of tissue is found^14–16  and the adaptation of posts is challenging.

In these cases, prefabricated glass fiber posts cannot match the canals’ size and shape, and for this reason, a considerable amount of cement is needed for cementation.^16–17  A thick cement layer may be related to less retention due to impaired polymerization over the root canal,^18,19  reduced mechanical strength in the cement layer,^16,20–23  and increased formation of voids and gaps.²⁴  Custom-made posts may be produced by relining with composite resins to increase adaptation and reduce the cement layer’s thickness.^17,25,26  In these cases, different interfaces are created, as several materials are used.^27,28  The manual adaptation may be time consuming and technique sensitive for clinicians, impairing its applicability and the long-term success of this procedure.

The computer-aided design/computer-aided manufacture (CAD/CAM) process may be applied to the production of individually, anatomically fitted, and monolithic glass fiber posts.^29,30  The possibility of scanning root canals or models allows the design of customized structures that match the shape and the size of the root canal system and may be installed with a thin layer of cement with better adaptation.^31,32  CAD/CAM milled glass fiber blocks are used to produce a monolithic post-and-core that avoids the use of different materials and the presence of multiple interfaces on the cemented structure.^29,33  As CAD/CAM manufacturing becomes more accessible and is increasingly used in dental applications, it is important to understand the ability of these methods to design and build posts for flared root canals. While several advantages have been observed for this manufacturing technique, few studies have attempted to understand monolithic glass fiber post-and-core behavior.^17,29,31  The effect of these posts on roots with extensive loss of tissue diameter should be investigated for the application of these structures in the clinical scenario. This study aims to evaluate the cementation and mechanical behavior of flared root canals restored with CAD/CAM milled glass fiber post-and-core systems.

Sixty-six extracted, single-rooted permanent upper and lower human canines were used. Inclusion criteria were the absence of calcifications, cavities, cracks, and previous endodontic treatment. Teeth were measured and included when the distance between the cementoenamel junction and the apex was at least 15 mm. All teeth were stored in distilled water at 4°C for no more than six months.

The selected teeth were decoronated 2 mm above the cementoenamel junction using a double-sided diamond disc (KG Sorensen Ltda, Cotia, SP, Brazil) with low speed and under water cooling. All root canals were instrumented with reciprocating files (V-file TDK Blister, Eurodonto, Curitiba, PR, Brazil), under NaOCl 2.5% irrigation. The root canals were then filled with 1 ml of EDTA (Odahcam, Dentsply, São Paulo, SP, Brazil) for 3 minutes. Root canal obturation was performed using a lateral condensation technique with gutta-percha (Dentsply) and a resin-based endodontic sealer (AH Plus, Dentsply, Konstanz, Germany). All specimens were stored for seven days in distilled water at 37°C before further preparation.

Before the post-cementation procedure, gutta-percha removal was performed with a stainless steel bur (Exacto #3, Angelus Indústria de Produtos Odontológicos S/A, Londrina, PR, Brazil), and 4 mm of filling material remained in the apical third. The root canals were enlarged with a sequence of Largo burs, washed with 5 ml of distilled water, and dried with paper points (Dentsply, São Paulo, SP, Brazil). The enlargement was performed to guarantee that dentin walls presented a 1 mm thickness to simulate flared root canals, to standardize the samples, and to minimize the anatomical differences between teeth. The prepared teeth were divided into three groups according to the different glass fiber posts used for reconstruction: prefabricated glass fiber post (G_PF), composite resin relined glass fiber post (G_REL)_, and CAD/CAM milled glass fiber post (G_MILLED). The preparation of posts for the different groups is summarized in Table 1. All specimens were water stored for 24 hours at 37°C.

For the G_PF group, posts (Exacto #3, Angelus) were cleaned with 70% ethanol. Silane (Angelus) was actively applied for 60 seconds and air-dried for 30 seconds. A self-etching resin cement (U200, 3M Oral Care, São Paulo, SP, Brazil) was mixed and dispensed into the post space with a syringe (Centrix, DFL Indústria e Comércio S.A., Rio de Janeiro, RJ, Brazil) and an endodontic tip. The post was inserted and polymerized for 40 seconds from the buccal and palatine faces with 1500 mW/cm² from a calibrated LED light-curing unit (Radii-cal, SDI Indústria e Comércio Ltda., São Paulo, SP, Brazil).

For the G_REL group, an adhesive resin (Adper Scotchbond Multi-Purpose, 3M Oral Care) layer was applied around the post and light-cured for 1 minute. For the relining, a composite resin (Filtek Z350, 3M Oral Care) was used directly around the post, inserted into the root canal, and light-cured for 5 seconds. The post was removed and light-cured for 60 seconds. After the relining, the post space was cleaned with distilled water and dried with paper cones. The post was washed with 70% ethanol, air dried, and cemented, as described in Table 1.

The G_MILLED group was prepared as described in Figure 1. The post space was filled with acrylic resin (Duralay-Reliance Dental, Alsip, IL, USA) for root canal modeling. The acrylic resin posts were scanned with a digital scanner (Shining3D DS-EX, Shining 3D Tech, Hangzhou, China) and computer-aided designed (CAD) posts were obtained with the root dimensions. Glass fiber discs (Fiber Cad - Post & Core, Angelus Indústria de Produtos Odontológicos S/A, Londrina, PR, Brazil) were used for milling on an industrial computer-aided machine (CAM) (DM5 Tecnodrill Indústria de Máquinas Ltda, Novo Hamburgo, RS, Brazil). The CAD/CAM milled posts were prepared and cemented as described in Table 1.

After the cementation, five specimens per group were scanned by x-ray microcomputed tomography (inspeXio SMX-90CT Plus, Shimadzu, Kyoto, Japan). The X-ray tube was operated at 80 kv and 120 mA with a voxel size of 0.013 mm. Reconstruction was performed by inspeXio SMX-90CT software (Shimadzu) and resulted in a stack of 541 images. The Dicom files were used for tridimensional reconstructions with 3D Slicer software (https://slicer.org).³⁴  The quantification of the volume of voids and gaps was performed in 220 two-D images that were standardized in the samples to analyze the cement layer. The quantification was performed with imaging software (ImageJ, NIH, Bethesda, MD, USA) by the segmentation of the 2D image stack with a grayscale threshold selected between 0–140. The empty spaces in the cement layer were selected with an automatic selection tool, and the area was calculated on each image. The volume of empty spaces in the cement layer was calculated based on the measured area and the thickness of each slice in the 2D stack. All analyses were performed with the same threshold.

A single operator performed all measurements. Failure pattern was classified as adhesive, between dentin and cement interface (Ac/d) and between cement and post interface (Ac/p) or cohesive in dentin.

For the load-to-fracture analysis, ten samples per group were submitted to a core reconstruction. For G_PF and G_REL groups, the coronal portion of the posts was etched with phosphoric acid 37% (Scotchbond etchant, 3M Oral Care) for 30 seconds in enamel and 15 seconds in dentin. A commercial primer (Adper Scotchbond Multi-Purpose, 3M Oral Care) was applied to the dentin. Solvent evaporation was performed, and an adhesive layer was applied. After light-curing, the core was built with a composite resin (BISCORE; Bisco Inc, Schaumburg, IL, USA) with a 5 mm height. Polymerization was performed for 60 seconds on four sites equally distributed in vestibular, bucal, distal, and mesial regions of the core. For G_MILLED, the core was modeled with acrylic resin (Reliance Dental Manufacturing LLC; Alsip, IL, USA) in the moment of root canal modelling. The dimensions of the core were the same used for G_PF and G_REL. The scanning and milling were performed at the same time, and thus a monolithic post-and-core structure was obtained.

The restored root canals were embedded in acrylic resin, and simulation of the periodontal ligament was performed with silicone rubber. Specimens (n=10) were then fixed in a metallic device, and a load was applied to the palatal surface at a 45° angle in a testing machine (Servopulser, Shimadzu Servohydraulic Fatigue Testing System, Kyoto, Japan) with a crosshead speed of 1 mm/min until fracture. Mode of fracture was recorded and classified as repairable in the tooth where debonding, core fractures, or oblique fractures in the cervical third were identified, or non-repairable in the tooth where vertical fractures or oblique fractures in middle and apical third were observed.

A descriptive analysis was performed for microCT images. Data were submitted to equal variance analysis, and the normality of the data was assessed by Shapiro-Wilk. The volume of voids and gaps on microCT, as well as the load-to-fracture values, were compared by one-way ANOVA and Tukey. Repeated measures were used as dependent samples. Each tooth was sliced in nine samples, and thus a dependent analysis was performed for this factor. For the push-out test, two-way repeated-measures ANOVA and Tukey were used. The significance level was set at 0.05.

MicroCT images (Figure 2) were analyzed, and the presence of voids and gaps was evaluated in the cement layer of all specimens tested. Reduced volume of voids and gaps was observed for G_MILLED when compared to G_REL and G_PF (Figure 2; p<0.05). The presence of voids and gaps was observed mainly in the cervical third for all groups, while for G_PF (Figure 2A) and G_REL (Figure 2B), voids were observed in the apical third. The representative slices of microCT analysis are shown in Figure 3. The cement layer thickness and the presence of voids and gaps are shown for the three groups. Fiber alignment is similar for G_PF (Figure 3A) and G_REL (Figure 3B), while a different pattern is observed for G_MILLED (Figure 3C).

For the push-out bond strength analysis, the obtained values ranged from 4.80 ± 2.15 MPa to 2.71 ± 1.81 MPa on G_REL and G_PF, respectively. No statistically significant difference was found between different thirds in any of the groups (p=0.103). There was a statistically significant difference between groups (p=0.023), and there was no interaction between the two factors (p=0.424). G_MILLED showed no statistically significant difference from G_REL and G_PF (Table 2; p>0.05). The failure modes are shown in Table 2. Adhesive failure was found on at least 58% of the samples. G_REL group presented a higher percentage of adhesive failure (69.66%), while G_MILLED presented the lowest percentage of adhesive failure (58.9%).

The load-to-fracture results are shown in Table 3. The average load-to-fracture on G_REL was 738.13 ± 151.19 N. G_MILLED values were statistically significantly lower than G_REL (544.83 ± 135.80 N; p<0.05). G_PF results were not statistically significantly different than G_REL and G_MILLED (p>0.05). The failure mode analysis showed increased core fractures (reparable) for G_MILLED. Non-reparable fractures were observed for G_REL (1) and G_MILLED (3).

The rehabilitation of teeth with extensive loss of hard tissues is a challenge in clinical practice,⁷  and different techniques have been studied for the reinforcement of flared root canals.²⁶  In the present study, CAD/CAM milled glass fiber posts were used to restore flared root canals in vitro. A monolithic system was produced, and when compared to prefabricated and relined glass fiber posts, the tested posts presented a lower amount of voids and gaps, with comparable push-out bond strength values. The load-to-fracture was lower as compared to the other groups.

The internal adaptation of the post influences the cement needed for the procedure.^20,28  The formation of voids and gaps is found due to air entrapment during the cement manipulation, and it is known that a thin layer of cement is related to the formation of fewer voids and gaps.¹⁷  When flared root canals were considered, the amount of cement used on prefabricated posts was obviously increased, and the space between the root canal shape and the fiber post may reduce the amount of cement and the formation of these voids and gaps.^17,33  Although the custom-build posts (G_REL and G_MILLED) are known to use a small amount of cement, the volume of empty spaces on the cementation region was higher for G_REL when compared to G_MILLED (Figure 2; p<0.05). This finding may be attributed to the space found on the apical third, because the shape of the post in this region is not as closely adjusted to the conical shape of the root in the G_PF and G_REL as it is for G_MILLED (Figure 2). The CAD/CAM milled was able to reproduce the apical portion’s conicity, leading to a reduced amount of cement and a reduced volume of voids in this third (Figure 2 and Figure 3). This shows that the accuracy of the scanning and milling process was guaranteed by the chosen method. It is important to highlight that an industrial milling machine was used in the present study, and thus the results may not be directly translated for the production of posts in dental milling machines. Despite this limitation, the benefits in post adaptation may be further explored for other CAD/CAM systems such as the ones designed exclusively for dental applications. The relationship between the presence of voids and gaps and the bond strength has been studied in different posts, and conflicting results have been found. While some reports related lower bond strength to void-containing structures,^22,28  other studies did not find a statistically significant difference.^23,24  Although the effect of voids and gaps is not always present, a thin and uniform cement layer is desired for stability of bonding and the mechanical interlocking of the cemented post.²⁸ 

The cement layer may play an important role in push-out bond strength. Because cement is known to present lower mechanical strength than glass fiber posts, the higher the amount of cement in the root canal, the lower the force needed for dislodgement of the root slices. The lower volume of voids found for G_MILLED may explain these results because of the CAD-controlled cementation layer and the amount of cement used for cementation. Polymerization was performed for 60 seconds on 4 sites equally distributed in vestibular, bucal, distal and mesial regions of the core.⁷  The push-out bond strength results were not affected by the tooth third in each group, showing that the cementation procedure was effective and that polymerization was achieved along the full extent of the root, excluding the possible effect of a low degree of conversion in the bond strength results.^18,35  Adhesive failures between the cement and the dentin were the most common failure mode found for all the groups, showing that dislodgement is more prone to happen in this interface and that the mechanical properties of luting agents were maintained.

The mechanical behavior of glass fiber posts may be influenced by the production method, and this is directly related to the orientation of the fibers, the adaptation of the post, and the stress dissipation on the post-cement-dentin interface.^5,29,36  The milling process is shown to result in decreased resistance for glass fiber posts alone.²⁹  In contrast, previous studies found comparable results for load-to-fracture when cemented prefabricated and milled posts were tested.^30,32,33,37  In the present study, static mechanical testing was performed on cemented posts, and the G_MILLED group showed the lowest load-to-fracture values (Table 3), without a statistically significant difference from G_PF. The fibers’ direction is a key factor for adequate stress dissipation, and milling the glass fiber blocks with vertical fiber orientation is required to guarantee adequate mechanical properties.²⁹  The alignment of fibers during milling must be obtained by the adequate positioning of the glass fiber block, avoiding different fiber orientation in the post. The representative slices in Figure 3 show that fiber orientation was kept parallel to the root’s long axis in all tested groups. The analysis of failure modes in the G_MILLED group shows that most of the samples showed chipping failures on the core, and that fracture occurred parallel to the fiber orientation, showing that milling was performed in the right direction (Table 3). Although lower values were found, the core chipping indicates easily reparable fractures, but nonreparable failures were also found. Some studies suggest that stiff materials with high load-to-fracture values may result in increased catastrophic failure.⁶  The possibility of repair of the fractured post and core structures is desirable for tooth maintenance, and the impact of the nonreparable failures found in the CAD/CAM milled group must be further evaluated.

The CAD/CAM milling process emerges as a machinable custom-build option to increase the adaptation of glass fiber posts on flared root canal systems. The application of a laboratory CAD/CAM system may contribute to the translation of these results to clinical practice, as access to this technology may be easier for dentists. CAD/CAM is known to present a high startup cost and the need for training, potentially limiting its application in clinical practice, especially for chair-side CAD/CAM.^38,39  Laboratory systems may be more accessible, and, in the results of this initial analysis, were shown to result in well-adapted posts for flared root canals. Using the findings of the present study, it is possible to design further analysis to compare the accuracy of different CAD/CAM systems in the design and fabrication of these posts. The production of a one-body monolithic system avoids the need for in-office adaptation of post and core production and prevents the formation of a multiple-interface system in the post system. The CAD/CAM manufacturing used in this study offered adequate adaptation with reduced cement layer and a lower amount of voids and gaps, evidenced by the high bond strength obtained in the G_MILLED group compared to the group with prefabricated posts. This adaptation is essential, especially for flared root canal systems, and may represent an alternative for reliable rehabilitation with lower technique sensitivity.

CAD/CAM milled glass fiber post-and-cores presented an enhanced adaptation of glass fiber posts to flared root canal systems. Their results were comparable in bond strength but lower in load-to-fracture when compared to relined posts.

The authors would like to thank CAPES “Coordenação de Aperfeiçoamento dePessoal de Nível Superior - Brasil -Finance Code 001”.