Assessment of the Stress Transmitted to Dental Implants Connected to Screw-Retained Bars Using Different Casting Techniques

At the bone-implant interface, there is no flexibility as in natural teeth. As such, any tensile, compressive, or bending forces introduced into an implant-supported restoration by misfitting superstructures lacking a passive fit will probably result in problems, ranging from screw loosening to bone resorption.^1,2 

The importance of a superstructure passive fit to ensure a good treatment prognosis is a central theme in the dental implant restoration literature. Impression and master cast accuracy—the major determinants of fit—have been examined several times, with varying results.^1–8  The large number of different fabrication techniques in use (conventional lost-wax casting and cast-on abutment techniques, newly developed computer numerical control [CNC] milling, laser sintering) indicates that inherent problems remain and that there is a great deal of interest in improving results in this area.^4,9,10  Published studies on clinical techniques for adjusting an inaccurate fit caused by various working procedures—including soldering, adhesive bonding, luting and laser welding, spark erosion, and computer-aided design/computer-assisted manufacture (CAD/CAM) procedures—provide clear evidence that there is still no satisfactory practice-orientated technique available.^11–17 

Clinically, these fit inaccuracies result in complications, such as screw loosening and fracture of screws, implant abutments, superstructures and implants in the functional phase.^18,19  Intraoral measurement of stress levels during prosthesis fitting would be the ideal gold standard to judge the accuracy of implant-supported restorations; however, the analysis of an absolute passive fit of superstructures is not possible using conventional laboratory and clinical procedures since clinical fit evaluation methods usually do not detect inaccuracies below the level of visual acuity or the measurement capacity of the testing equipment.¹⁸ 

Several approaches have been proposed to enhance the fit of implant frameworks; however, there is still no consensus in the literature as to which method is more appropriate to measure the misfit of implant-supported superstructures. Thus, the aim of this study was to perform a measuring technique for comparative checking and localization of mechanical stress in bar restorations fabricated using the cast-on abutment or lost-wax casting techniques supported on four implants.

A total of 15 patients each had four implants placed in the interforaminal region of the mandible (34, 32, 42, 44 regions) and were selected for this study, which was conducted at the Department of Prosthodontics and Dental Materials of the University Hospital RWTH Aachen, in Germany. Identical implants (BEGO Semados, Bremen, Germany) with 13 mm length × 3.75 mm diameter were placed in all cases, and after three months, all implants showed successful osseointegration. The transfers (BEGO Semados) (Figure 1a), together with open custom impression trays, were used in the pick-up impression technique with a polyether impression material (Impregum, 3M ESPE, Seefeld, Germany) (Figure 1b). The analogues were screwed into the transfers using a torque wrench set at 10 Ncm, and the master models were fabricated according to the manufacturer's instructions.

An experienced dental technician fabricated a bar alloy (BIO SEMADOR H – BEGO, Bremen, Germany) with preformed Dolder bar patterns for each of the 15 cases using the cast-on abutment (Figure 2) or lost-wax casting techniques.

The main difference between these techniques is that the cast-on abutment has a machined metal interface with the implant, whereas the lost-wax technique has a full plastic burnout pattern that is cast into metal and then interfaces with the abutment. Both techniques utilize either wax or resin patterns between the abutments (depending on the situation) that are then invested and cast in dental alloy.^20,21 

After the 30 bars were prepared and finished, the fit accuracy was first visually checked intraorally according to the Sheffield's test criteria,²²  which evaluates the absolute passive fit of the implant bar on the master cast. Thereafter, the abutments were tightened to the analogues with a torque of 25 Ncm. The bar was then placed on the master cast and only one screw was used to attach the bar to the cast. With a gap-free passive fit before and after tightening the prosthetic screws, measurements were recorded on the master models.

Sheffield's tests are carried out before machining to determine passivity. To perform this test, the metal superstructure should be inserted over the supporting implants or abutments. At that point, the most distal retaining screw should be tightened, and the rest of the retaining screws should be kept out. If a gap appears between the remaining supporting implants or abutments and the metal superstructure, it indicates that the metal framework does not fit passively.^22,23 

To measure the axial distortion, the sensors were supported on implant extensions (Figure 3a) and were screwed into the laboratory analogues using a torque of 25 Ncm (Figure 3b). A mechanical torque wrench (BEGO Semados) was used according to the manufacturer's instructions for tightening the analogues and all other screw procedures. Figure 3c shows how the sensors were placed on implant extensions.

Six foil strain gauges (EA-06-031DE-350 type, Micro-Measurements Division, Vishay Precision Group, Wendell, NC) were symmetrically attached and spaced at 60° angles on the outer surfaces of each of the system implant extensions, making them sensors for axial distortion (Figure 4a). Another three foil strain gauges were used externally to complete the individual active foil strain gauges with the three-wire configuration for the Wheatstone bridge electrical circuit.

The measuring signals were amplified by the MGC/MC 30 data acquisition system (HBM, Darmstadt, Germany) and transferred to a computer for storage (Figure 4b and c). DIADAGO software (GFS, Aachen, Germany) was used to control the test procedures via the measurement system and for data acquisition and processing (Figure 4d).

Following sensor calibration, the bars were tightened and then loosened again on the sensors in a specific sequence (34 – 42 – 32 – 44), according to the manufacturer's instructions and as in the Sheffield's test. Each measurement was repeated and recorded five times. The resulting axial distortion of the individual implant abutments/extensions was measured and recorded for each screw procedure.

Repeated measures analysis of variance (ANOVA) with P < 0.05 was used for statistical analysis of the measurements with the SAS System software (SAS Institute, Inc, Cary, NC) at the Institute for Medical Statistics of the Medicine Faculty at the Rheinisch-Westfälische Technische Hochschule (RWTH), Aachen.

The results illustrated in Figure 5 show a comparison of the axial deformation between the cast-on abutment (Figure 5a) and lost-wax casting techniques (Figure 5b) in relation to the position after screw-retention of the bar only at position 34.

Figure 6 illustrates the mean axial distortion after tightening the screws of all four implants through the cast-on abutment technique (Figure 6a) and lost-wax casting technique (Figure 6b), with time sequence of the tightening and loosening of all abutment screws measured on sensor 34 and loosening of all abutment screws measured on sensor 34. The greatest axial distortion was at position 42 with the cast-on abutment technique, with a mean distortion of 450 μm/m. The lowest axial distortion occurred at position 44 with the lost-wax casting technique, with a mean distortion of 100 μm/m. The minimal differences between the axial distortion means do not indicate any significant differences between the various fabrication techniques (P = 0.2076). Analysis of the sensors axial distortion in relation to the implant position produced a significant difference (P < 0.0001).

Figure 7 illustrates examples of axial distortion in relation to the sequence of tightening (positions 34 – 42 – 32 – 44, respectively, represented by numbers 1–4, respectively, on the graphic) and loosening (positions 44 – 32 – 42 – 34, respectively, represented by numbers 5–8 on the graphic) of the retention screws for implant position 34 in the cast-on abutment (Figure 7a) and lost-wax casting techniques (Figure 7b).

Tightening the bar at position 34 results in axial distortion that alters significantly when the retention screws at the other contact points of the bar superstructure are tightened and loosened. There is a much wider distribution of measurements with the cast-on abutment technique than with the lost-wax casting technique. The greatest mean distortion measurements also differ depending on the fabrication techniques.

In a comparison of the global means of all measurements, axial distortion in the third quadrant with a mean of 133 μm/m is higher than in the fourth quadrant with 89 μm/m, though statistical analysis did not produce any significant differences in this comparison of right and left quadrants (P = 0.1410). However, significantly higher measurements were recorded in the axial distortion analysis of the distal sensors on the implants in the 34 and 44 regions than on the mesial positions in the 32 and 42 regions (P = 0.0481).

This study describes a method that records reproducible measurements of implant extension axial distortion following retention of implant-supported superstructures using strain gauges.

An overdenture supported by a bar framework on four implants placed in the interforaminal region is the standard restoration in implant prosthetic treatment of the edentulous mandible.²⁴  This type of restoration involving primary connection with bars can be fabricated using either the cast-on abutment or lost-wax casting technique.^10,25  Some basic differences between the two fabrication techniques were described by Carr and Master²¹  but only concerning single abutments. The authors also made recommendations for fabrication, finishing, and application of burnout plastic sleeves, including procedures for cast-on gold copings.²¹ 

Although the results were taken into account, as well as the manufacturer's instructions when making the bar restorations, this study showed no significant difference between the techniques based on the 15 patient models. The influence of non-ideal situations and the limited number of specimens, however, should always be taken into consideration.

In a comparison of the effect of manual screw tightening, mechanical torque wrenches, and electronic torque wrenches on prosthetic screw preloading as a source of error, Standlee et al²⁶  found deviations of less than 10%. Another study compared the type of torque transmission—that is, manual, mechanical, and electronic—and the least deviation was recorded with the mechanical torque wrench, less than 5%.²⁷ 

To minimize this source of error in the present study, all tests were completed using a mechanical torque wrench. The effect of the screw tightening sequence according to the Sheffield's test was also analyzed. In this case, there were significant differences among the individual screw procedures: each tightening, retightening, or loosening of a screw affected the other screw connections and altered the axial distortion at each individual implant extension along the whole bar system.

Watanabe et al²⁸  obtained corresponding results in their tests on one-piece cast superstructures on three implants compared with other fabrication techniques. Nissan et al²⁹  established that under ideal in vitro conditions, no statistically significant differences among the variables of tightening sequence, tightening force, and operators were found.

Some authors have attempted to define an acceptable level of fit between the metal superstructure and the osseointegrated dental implant.^30,31  In 1983, Branemark was the first to define passive fit, and he proposed that this should be at the level of 10 μm to enable bone maturation and remodeling in response to occlusal loads.³¹  In 1991, Jemt³⁰  defined passive fit as the level that did not cause any long-term clinical complications and suggested that misfits smaller than 150 μm would be acceptable.

It has been documented in the literature that the amplitude of movement of osseointegrated implant is limited to 50–150 μm, allowed by the resilience of bone.³²  This rigidity in the connection means that any stress or tension arising from the prosthesis fixation will be transmitted directly to the implant components or to the bone-implant interface.^9,33  Clinical consideration of the cortical and cancellous bone intrinsic elasticity may allow reduction of the gap if the bar is inserted under compression, but this is not recommended since undesirable stress onto the bone-implant interface may cause screw loosening, implant fracture, or loss of osseointegration.^34,35  Regarding these concerns, the fit accuracy of screw-retained prostheses must be maximized.^36–38 

Nevertheless, several studies indicated the presence of some biologic tolerance of dental implants to certain levels of misfit.^39–41  However, it is difficult to determine these states due to limitations of these investigations, including ethical principles of clinical human research.³⁵ 

The aim of this study was not only to record the total number of axial inaccuracies but also to localize inaccurate areas. The measurements at the posterior positions produced significantly greater distortion than at the anterior implant positions; however, comparison of the quadrant sides established no significant difference.

Apart from considering these mechanical aspects, there is also the question of the clinical impact and significance of the measured inaccuracies.³³  Despite the fact that most clinical methodologies are subjective and have many variables, they are also dependent on operator skills. More accurate techniques have been reported in the literature; however, they are generally not feasible or practical for clinical use. Considering this difficulty in assessing the fit of implant fixed prostheses clinically, any fabrication method should first be tested and verified by research. Thus, in vitro studies provide indispensable data regarding a specific method of fabrication and its accuracy. This information will indicate the anticipated fit of each technique, what degree of distortion is acceptable, and what could be the implications of such a degree of distortion to the dental technician and the clinician.^9,42  The measuring procedures described in this article can be used for reproducible testing of the passive fit in newly developed products, both in vitro and in vivo.

Finally, it can be affirmed that each screw procedure causes distortion of the abutments, and there is further distortion by primary connection. Further studies are necessary to achieve an ideal passive fit of a screw-retained superstructure on four implants placed in the interforaminal region of the mandible.

The measuring technique using strain gauges for comparative checking and localization of mechanical stress in bar restorations can record reproducible measurements of the axial distortion of implant-supported superstructures. There was no significant difference between the two sides when comparing the accuracy and fit of a screw-retained bar. Distortions were present for both cast-on abutment and lost-wax casting techniques.