Implant Volume Loss, Misfit, Screw Loosening, and Stress In Custom Titanium and Zirconia Abutments

The aim of this study was to verify the effect of the implant volume loss, vertical misfit between abutment and prosthetic platform, prosthetic screw loosening torque, and screw stress distribution in titanium and zirconia abutments. Ten CAD/CAM system custom abutments of each material were milled and attached to the titanium implants. The implant volume loss was evaluated by microtomography, the vertical misfit with optical microscopy, and digital torque wrench measured the prosthetic screw loosening. All experimental analyses were performed before and after mechanical cycle (1,000,000 cycles, 100 N/2 Hz). Virtual models of the structures were created for finite element analysis, and the stress on the screw obtained with von Mises procedure. Data were analyzed using an independent t-test, two-way ANOVA for repeated measures, and Tukey’s HSD test ( α =0.05). There was no significant difference in the implant volume loss for the two abutment materials (p=0.662). Titanium abutments provided higher loosening torque values after mechanical cycling (p<0.001). Lesser marginal misfit was obtained with titanium abutments before and after mechanical cycling (p<0.001). The stress distribution on the screw was similar between abutment materials. In conclusion, CAD/CAM custom titanium abutment reduced the marginal misfit and increased the torque maintenance of prosthetic screws when compared to CAD/CAM custom zirconia abutment.


Introduction
Despite the recognized success of dental implants, mechanical problems associated with implant-supported single crowns are still a challenge to the dentist, mainly in relation to loosening and fracture of the prosthetic screws (1,2). For the single crowns, the cemented technique is generally an adequate esthetic solution to cover the screw access hole, especially for tilted implants (3).
The maintenance of prosthetic screw torque is extremely important, since the retention does not influence the fracture resistance but has a relevant effect on the failure pattern of the specimens. However, CAD/CAM milled lithium-disilicate crowns seem to be a preserving factor for dental implants (4). Vestibular fractures are restricted to ceramic abutments and lesser gingival discoloration with zirconia abutments; but there is no differences in patient's esthetic satisfaction between ceramic and metal abutments. For the anterior region, implant with internal connection and custom metal abutment showed the least mechanical complications (5).
For aesthetic reasons, ceramic abutments were associated to implant-supported rehabilitations in the early 1990s to avoid the pigmentation caused by titanium abutments due to interaction between wear and corrosion occurred when masticatory forces combine with the saliva in the oral cavity (6). CAD/CAM manufacturing method improves management of the subgingival depth of the crown/abutment interface and enhances the esthetic of the restoration (7).
The best way to obtain a suitable relation between the gingival depth and adjacent dental structure is with custom abutment and subgingival margin between 0.5 to 1 mm, considered important factor in preventing inflammation caused by cement residue and food debris (8). In addition, custom abutments are significantly stronger than the trade abutments but its fit is less precise. However, the mechanical strength and the adjustment level were not different for both abutments and showed clinically acceptable limits (9).
With the CAD/CAM system is possible to mill metal blocks or pre-sintered ceramic zirconia blocks. The zirconia abutment must be milled with larger dimensions before the sintering process, fact that may cause dimensional changes in the abutment. On the other hand, the abutment fitting problems generally favor the loosening of the prosthetic screw or possible fracture (4).
Previous study showed that the CAD/CAM abutment appear to show a fit and marginal adaptation level comparable to those trade abutment for most of the systems evaluated, while zirconia abutment revealed a good adjustment level in the interface with dental implants (10). In addition, the copy-milled zirconia abutment does not fit as accurately as trade abutment titanium; however, the less precise fit of the custom zirconia abutment did not affect the ability to maintain the applied torque (11). On the other hand, the fracture resistance of titanium and zirconia abutments with internal connection under simulating cyclic masticatory load showed that 1-piece zirconia abutments exhibited a significantly lower fracture resistance than titanium abutments (12).
Therefore, there is concern that zirconia may damage the implant surface when in oral function. Previous study showed greater wear on the zirconia-titanium internal hexagon interface when compared to titaniumtitanium interface (13). In contrast, a three-dimensional finite element analysis showed that the zirconia implant generated the lowest stress in cortical bone, and the zirconia abutment resulted in lower von Mises and compressive stresses than the titanium abutment in implant and cortical bone (14).
Although internal connections are the most used due to greater stability, the external hexagon implants are used for decades and have still large application today in totally implant-supported prosthesis. Given these considerations, it would be appropriate to evaluate the marginal fit between abutment and implant, the screw stability, the implant volume reduction, and stress at screw in implants with external hexagon connections submitted to mechanical cycling. The study hypotheses were that titanium and zirconia abutments would not show relevant differences on: 1) passivity of the abutment, 2) maintenance of the initial torque, 3) implant volume loss, and 4) concentration level of stress on the screw.

Material and Methods
Twenty external hexagon implants with a regular platform (4.1 mm in diameter) and dimensions of 11 mm in length and 3.75 mm in diameter were used (Titamax; Neodent, Curitiba, PR, Brazil). The implants were randomly divided into two groups, according to the titanium or zirconia abutment material (n=10). Each implant was fixed in epoxy-based resin block (F160 A+B; Axson, São Paulo, SP, Brazil) with aid of a parallelometer.
A trade abutment (UCLA; Neodent) was used to make a matrix in low shrink acrylic resin (Pattern Resin LS; GC South America, Sao Paulo, SP, Brazil) with shape and dimensions of a first mandibular molar. The crown was prepared with an occlusal thickness of 1.5 mm and axial walls of 1.2 mm. An acrylic resin matrix was scanned with CAD/CAM system (Neodent Digital) and used for the abutments preparation (n=10/group) in commercially pure titanium or a polycrystalline tetragonal zirconia stabilized by yttrium of first generation (3Y-TZP) (Fig. 1). After model images definition, the abutments were manufactured with CAD/CAM system (Neodent Digital). The zirconia abutments were sinterized at 1500 o C for 2 h.

Implant Volume Loss
The implant volume (mm 3 ) was evaluated before the abutment installation on the platform. A microtomograph device (SKYSCAN 1176: High-resolution in-vivo micro-ct; Bruker, Kartuizersweg, Belgium) was used for a better resolution and construction of the implant models. Images with a power of 90 Kv and a spatial resolution of 9 µm (maximum resolution allowed by the equipment) served to build the 3D implant models and to evaluate the quantitative analysis. The same analysis process for the volume of the implant before installation were used after the mechanical cycle.

Screw Loosening Torque
The values for the initial torque and the tightening torque of the prosthetic screw (Neotorque; Neodent) were applied with torque wrench with precision digital (TQ8800; Lutron, Taipei, Taiwan). The initial torque was the recommended by the manufacturer (32 N.cm) and the loosening torque measurement was performed 10 min after the initial torque. The screw was replaced and the torque processing repeated after each measurement cycle. In the same condition, the screw loosening torque was assessed after the mechanical cycling.

Vertical Misfit Measurement
The initial misfit level of each specimen was measured at the buccal, mesial, lingual, and distal points of the abutment-implant interface with linear optical microscope (UHL VMM 100 BT; UK) with 1.0 μm accuracy equipped with digital camera (KC-512NT; Kodo BR Electronics, Sao Paulo, SP) and analyzer unit (QC 220-HH; Quadra-Check 200, Metronics Inc., Canada). The same researcher performed three evaluations on each point, and the average obtained was the value considered for each point. The average of the four points represents the final misfit value for each specimen. The misfit value after mechanical cycling was assessed with the same measurement method. All passive fits were evaluated at magnification of 120 times.

Mechanical Cycling
Specimens were submitted to the mechanical cycler (ER -1100 Mechanical Fatigue Simulator; Erios, Sao Paulo, Brazil) with 1,000,000 cycles under compressive axial load (100 N/2 Hz frequency) exerted on the occlusal region of the crown immersed in distilled water.

Scanning Electron Microscopy
Before-and after-cycling qualitative analyzes were performed with scanning electron microscopy (JSM-5600LV; JEOL, Tokyo, Japan) to verify the possible changes or wear on implant platform.

Finite Element Analysis
A three-dimensional (3D) model composed by crown, prosthetic screw, implant and resin base was assembled. Three-dimensional geometries and assemblies were developed with computer-aided design (CAD) software (SolidWorks 2010, Concord, MA, USA). The 3D models were imported to computer-aided-engineering software (ANSYS Workbench 11; Ansys Inc., Pittsburg, PA, USA) for numerical analysis. Meshes composed with 10-nodes tetrahedrons, checked for element quality were refined in the interest areas.
A 100 N standard axial loading was applied to occlusal surface or at 45 degrees on the implant axis. Model movement was restricted at 6 degrees of freedom in the bottom region and lateral surface of the nodes of the acrylic resin cylinder base. Before axial occlusal loading, a pre-load of 20 N was applied at prosthetic screw neck (18,19). Similar von Mises stresses values were assigned for implant and prosthetic screw.

Statistical Analysis
Data were submitted to Kolmogorov-Smirnov and Shapiro-Wilk tests to identify data normality, and transformed in logarithmic scale when necessary to achieve normality. The influence of abutment type on the implant volume loss was verified with independent t-test. Two-way ANOVA for repeated measures and Tukey HSD test assessed the influence of abutment type and mechanical cycling on the misfit and loosening torque values. P value <0.05 was considered statistically significant (IBM SPSS Statistics for Windows, v.21.0, IBM Corp., Armonk, NY, USA).

Implant Volume Loss
The independent t-test showed no significant difference when comparing the implant volume loss for the two groups (t=0.444; df=18; p=0.662). Mean (Standard deviation) for implant volume loss was 0.161 (0.20) mm3 for titanium, and 0.135 (0.16) mm3 for zirconia. The qualitative analysis performed with SEM showed no relevant differences for the implant platforms in relation to wear or damage of the structures (Fig. 2).

Screw Loosening Torque
ANOVA showed that the factors abutment material (p=0.398) and mechanical cycling (p=0.099) were not significant for screw loosening torque, although the factors interaction was significant (p<0.001). Zirconia showed a significantly higher screw loosening torque (24.8 N.cm) when compared to the titanium (17.9 N.cm) (p=0.001) for the before cycling time. After cycling, the zirconia abutment showed lower loosening torque (13.8 N.cm) when compared to titanium abutment (24.8 N.cm) (p<0.001).
Statistical differences were observed for screw loosening torque in both materials when the intra-group comparison was performed (Table 1). Zirconia showed significantly lower value for screw loosening torque after cycling when compared to the initial torque (p<0.001). For the titanium, higher values of screw loosening torque were observed after cycling (p=0.004).

Vertical Misfit Measurement
ANOVA showed significant influence for the abutment material factor (p<0.001), but not for vertical misfit after mechanical cycling (p=0.209). Interaction between the factors was significant (p<0.001). Table 2 showed that titanium abutment presented lower misfit for before-(p<0.001) and after-cycling (p<0.001). When comparing the passivity for the same material before and after cycling, the titanium abutment showed reduced misfit after cyclic loading (p<0.001), while the zirconia abutment presented an increase in the same cycling conditions (p =0.034).

Finite Element Analysis
There was no relevant difference in the stress concentration for screws between the titanium and zirconia abutments when axial load (Fig. 3) or oblique load (Fig.  4) was applied. The axial load promoted greatest stress concentration at the junction between head and neck of the screw, whereas the oblique load concentrated stress on the central region of the screw.

Discussion
This study verified the mechanical behavior of custom titanium and zirconia abutments manufactured with the CAD/CAM system. According to the results, the first hypothesis that there would be no difference in the abutments passivity was not accepted. The titanium fit was better than the adaptation of the zirconia at the before and after cycling times, showing a significant improvement for titanium when compared to zirconia material (Table 1).
These results are in agreement with previous study showing that the abutment adaptation to the implant was close, the amount of contact larger for assembly with premachined abutment and laboratory modified premachined abutment than for those with cast abutments (18). However, all screw types displayed some decrease in preload with repeated tightening, regardless of abutment type and insertion torque (19).
In addition, the implant-titanium abutment connection showed significantly better fit than the implant-zirconia abutment configurations with gap values three to seven times larger than those for titanium abutment (20). Therefore, titanium CAD/CAM abutment showed similar adaptation with different prefabricated abutments for most of the systems evaluated, and design differences between the abutment connections affected the fit of internal components of the implant-abutment connections (10).
Zirconia abutment showed value of fracture strength lesser as conventional titanium abutment; but it promoted good adjustment in the interface with dental implant,   excellent biocompatibility and good aesthetical appearance (21). On the other hand, the copy-milled zirconia abutment did not fit as accurately as prefabricated titanium abutment. However, the less precise fit of customized zirconia abutment did not negatively affect their ability to maintain the applied torque (11). The comparison of the abutment misfit in each material in relation to cycling time showed significant statistical difference for the titanium with lower misfit after mechanical cycle. Significant statistical difference occurred also for the zirconia; however, the misfit level was greater in both evaluation times when compared to titanium ( Table 2). This fact allows assuming that the higher elasticity modulus of the zirconia would negatively affect the implant passivity during the mechanical load application on the screw.
Probably, the better fit obtained for the titanium is due to greater malleability of the metallic alloy when compared to zirconia rigidity. Consequently, the zirconia rigidity would promote greater misfit level, decreasing the force value employed in the initial torque. This assumption would also related to other variables of the study, by example, the initial torque applied and the torque maintenance resulting in lower misfit after application of the screw load.
On the other hand, titanium implants showed higher wear at the interface after cyclic loading when connected to one-piece zirconia abutments compared to titanium abutments. However, the clinical relevance was not clear, since the damages in the internal implant connection could result in prosthetic failures or the need for implant removal (13).
The different CAD/CAM systems used in the manufacture of the abutment could also influence these mechanical differences. However, the CAD/CAM abutments appear to have a fit comparable to the prefabricated abutments for most of the different evaluated systems. On the other hand, the design differences of the abutment connections affected the adjustment level of the internal components between abutment and implant (10).
There were statistically significant difference concerning the cycling time, with better value for zirconia in the precycling time, and greater value for the titanium in the post-cycling time ( Table 2). Since the maintenance of the initial torque was different for the titanium in relation to the zirconia the second hypothesis was not confirmed.
These findings highlight the assumption that the material stiffness influences the passivity maintenance and pre-torque load level. These findings highlight the assumption that the material stiffness influences the passivity maintenance and pre-torque load level. Thus, screw preload maintaining is an important factor to maintain the fixation of implant-supported rehabilitations for long term. Another interesting factor shown in previous study was that there is always a loss of the initial torque value, whatever the evaluation condition. The wet condition showed higher torque and better preload values, suggesting that the abutment screw must be lubricated with saliva to avoid further loosening (1), whereas the zirconia and titanium abutments exhibited similar survival times, and same technical, biological and esthetical outcomes after three years in use (22).
Over the years, implants with external connections screwed or cemented for single prosthetic rehabilitations were frequently analyzed by different methods in relation to the stability level of the screw torque (13,23,24). The screw loosening occurs by reduction or complete loss of the initial torque, which depends of several factors including improper torque, lack of adaptation, large rotational freedom, and failure in the screw when re-tightened, favoring the wear and the friction between the components (24).
In addition, the aging by thermo cycling and chewing simulation with static load applied at 30 degrees angle to the palatal face showed that bending moment for different types of zirconia abutments changes with different implant-abutment connections when compared to titanium abutments with internal connection (25).
Due to the lack of a clear conclusion in the literature about possible damage caused by titanium or zirconia abutments in implants of external connections, the present study also verified a reduction in the volume of the implant platform. The third hypothesis was confirmed since there was no statistically significant difference between the abutment materials in relation to implant volume loss. Although the levels of hardness and stiffness showed higher values for the zirconia, it is possible to assume that the functional load was not sufficient to promote different wear patterns in relation to volume loss of the implant.
Although the abutment materials are different in the current study, the axial force was similarly exerted in both abutment types, and distributed on the screws and implant platforms in similar mode. Since the concentration of stress at screws was not different between abutments, the fourth hypothesis was accepted. This fact appears to confirm the result similarity in relation to stress supported by the screws under axial force (Fig. 3).
However, the oblique load promoted greater deformation on the titanium abutment probably due to lesser elastic modulus, increasing the stress transfer to the prosthetic screw (larger tension on the screw neck) and smaller increase for the von Misses maximum stress value on the prosthetic screw body (Fig. 4). By analogy, previous study showed that stiffer frameworks promoted higher stress concentration; however, these frameworks caused lower tension on the porcelain veneer, retention screw, and peri-implant bone tissue (23).
The results of the current study suggest that titanium abutment used with external hexagon implant placed in posterior region might cause less biomechanical problem related to the screw loosening or fracture when compared to zirconia abutment, since its mechanical properties are considered superior to those of zirconia. However, it is claimed that the zirconia material can be used especially as abutment (26). In addition, internal tapered connection implants assessed with strain gauges showed that the screw-retained or cement-retained do not always exhibit different levels of stress (27).
Future clinical long-term follow-ups are necessary to confirm the current in vitro results. The effect of the connection and abutment on the fatigue strength and assembly change is important for the maintaining of the interface, which wear is not always similar between titanium and zirconia materials (28). In addition, the type of abutment material and the connection design may affected the fit and the sealing capability of the different abutments (29).
Possible study limitations were considered, such as lack of simulation of the mandibular dynamics and absence of prosthetic crown, variables that should be subjected to further investigations.
According to the results of the study, the following conclusions can be drawn: 1. The passivity of the custom titanium abutments adapted to external connections was higher than for zirconia, with lower misfit in the before-and after-cycling times; 2. Titanium abutments showed better results than zirconia in maintaining of the screw torque in the aftercycling time; 3. There was no statistically significant difference for implant volume reduction in relation to the different abutment materials; 4. The screw stress magnitude was not influenced by the abutment material.