In Vitro Analysis of Resistance to Cyclic Load and Preload Distribution of Two Implant/Abutment Screwed Connections

Failures of implant-supported prosthesis represent an eventuality difficult to be managed by clinicians. Fixture or screw fracture, in fact, could give rise to the failure of the prosthodontic rehabilitation; Adell et al¹  reported a fixture fracture rate of 5%, while Schwarz²  reported a failure rate of 12.5% for maxillary implants and 14.3% for mandibular implants.

As regards the screwed systems, screw loosening has been reported to be the most recurrent complication,³  and many researchers state that the associated abutment mobility could compromise the survival of the rehabilitation.^4–6  However, it is generally accepted that the screw loosening occurs most frequently in single crown implant restorations rather than in implant-supported bridges.⁷ 

Preload applied during screw tightening and the fit between implant components are both crucial elements in providing a durable stability to the implant-abutment complex. McGlumphy et al⁸  report that the preload is dependent on the following factors: (1) torque applied, which influences the screw head friction, the thread friction, and the elastic/plastic deformation of the screw; (2) screw head geometry, which influences the screw head friction; and (3) the screw and abutment material, which influence the level of grip. From a biomechanical point of view, the applied torque determines the preload, which in turn is responsible for the clamping forces; such forces arise from both attrition between opposing surfaces (screw and fixture) and elastic/plastic deformation occurring on the mating structures.

From a clinical point of view, cyclic forces applied to the restoration during function could have a separating effect on the implant/abutment joint; in fact, screw loosening is generally due to off axis load.

The aim of the present research is an in vitro evaluation of the preload distribution in screw-retained implant systems under cyclic load.

This study used 35 Xsigñ internal-hexagon-connection implants (Figures 1 and 2) (Dr. Ihde Dental AG, Gommiswald, Switzerland), sized 4.5 × 10 mm (group Xsigñ), and 35 internal-octagon-connection implants SSO (Figures 3 and 4) (Dr. Ihde Dental AG), sized 4.5 × 10 mm (group SSO). Each group of samples was divided into 3 subgroups (2 test and 1 control) as reported in the Table.

The samples belonging to the Xsigñ and SSO test groups were fixed, by means of an acrylic self-curing resin,⁹  to the aluminum sample holder with the major axis perpendicular to the shelf. The insides of each fixture were soaked with artificial saliva, and then a 15° pre-angled titanium abutment was mounted (Figure 5); the torque applied during mounting procedures, according to the manufacturer instructions, was 25 Ncm for Xsigñ implants and 35 Ncm for SSO implants.

Each sample underwent the action of a vertical cyclic load ranged between 20 and 200 N for an average timespan of 280 hours using a universal test machine (Lloyd 30K, Lloyd Instruments Ltd, Segensworth, UK) managed by a dedicated software (Nexigen, Batch Version 4.0 Issue 23, Lloyd Instruments Ltd, Segensworth, UK) (Figure 6). After the application of the cyclic load, the loss of preload was investigated in each sample through the evaluation of the presence or absence of abutment mobility with respect to the fixture. The samples were then removed from the aluminum blocks, included into resin (LR White EM, TAAB Laboratories Equipment Ltd, Aldermaston, UK), sectioned along the mid axis (Micromet, Remet sas, Casalecchio di Reno, Italy), lapped with abrasive paper with decreasing grain (320, 600, 800, 1200) using an LS2 lapping machine (Micromet, Remet sas), metalized with gold (Emitech K 550, Emitech Ltd, Ashford, UK), and observed via scanning electron microscope (EVO 50, Carl Zeiss SMT AG, Germany) in secondary and backscattered electrons mode.

For implants belonging to control groups, the stump was fixed by using a dynamometric wrench adjusted to 25 Ncm for Xsigñ implants, and to 35 Ncm for SSO implants. Each sample was included into resin (LR White EM, TAAB Laboratories Equipment Ltd, Aldermaston, UK), sectioned along the mid axis (Micromet, Remet sas), and lapped with abrasive paper with decreasing grain (320, 600, 800, 1200) using an LS2 lapping machine (Micromet, Remet sas), metalized with gold (Emitech K 550, Emitech Ltd), and observed via scanning electron microscope (EVO 50, Carl Zeiss AG, Oberkochen, Germany) in secondary and backscattered electrons mode.

Five Xsigñ internal-hexagon-connection implants (Dr. Idhe Dental AG), sized 4.5 × 10 mm, and 5 SSO internal-octagon-connection implants (Dr. Ihde Dental AG), sized 4.5 × 10 mm, were used. Each implant was inserted into a block of acrylic self-curing resin (Figure 7) (Jet Kit, Lang Dental Mfg Co, Inc, Wheeling, Ill). Then, a silicon-made mold was constructed and filled with photoelastic liquid (PLM-4R Vishay Micro-Measurements, Raleigh, NC), mixed according to the proportions given by the manufacturer. After polymerization of the photoelastic resin, the samples were removed from the silicon mold and polished using a series of abrasives with decreasing grains (320, 600, 800, 1200) using an LS2 lapping machine (Micromet, Remet sas). A titanium stump was screwed to each photoelastic model, applying a preload of 15 Ncm, and we proceeded to the assessment of stress concentration sites through a circular polarizer and analyzer: each sample underwent photoelastic analysis before and after the application of a load equal to 25 N using a universal test machine (Lloyd 30K, Lloyd Instruments Ltd) managed through a dedicated software (Nexigen, Batch Version 4.0 Issue 23). The images were recorded by means of a digital camera and then archived in electronic format.

After the application of a dynamic load, there was no evidence at a macroscopic level of mechanical failure referred to the interfaces between abutments and fixtures. In all samples, moreover, no abutment-fixture movement was recordable.

At low magnification, it was possible to observe the parts that make up the implant-abutment-screw unit in the 2 different implant systems (Figures 8 and 9). In all samples the absence of gap between the surfaces was observed; Figure 10 is a scanning electron microscope photo of a sample belonging to the Xigñ control group, highlighting the presence of an intimate contact between the base of the abutment and the neck of the fixture. In the same way Figure 11, relative to a sample belonging to the SSO control group, shows the presence of contact between the side of the coils of the fixing screw and the internal thread of the fixture.

After the application of the cyclic load, all samples belonging to the Xsigñ test group show maintenance of preload both in the screw underhead and in correspondence of the internal thread of the fixture (Figures 12 and 13). The structural integrity, in the end, results were maintained also in correspondence of the closing edge of the stump over the fixture (Figure 14).

Here the scanning electron microscope examination carried out on the sections reveals that in 2 samples (13.33%) there was a loss of preload applied to the fixing screw; the scanning electron microscope analysis reveals the presence of a gap equal to 20.424 ± 5.846 μm (mean ± SD) between the screw coils and the internal thread of the fixture (Figures 15 and 16). Examination of the screwhead-stump contact areas also reveals that a plastic deformation occurred in some areas, which is likely responsible for the loss of intimate contact between the parts (Figure 17). Concerning the contact between the base of the stump and the internal surfaces of the fixture's neck, the presence of a slight gap was found in only 1 section (Figure 18). Microscopic examination of the section after the removal of the stump's fixing screw reveals in the end the presence of mechanical damage to the fixture's internal thread (Figure 19).

Analysis of replicas made of photoelastic material gives indications on the distribution of the preload applied to the fixing screw both before and during the application of a load. In nonloaded samples appear areas of stress basically originated by the screw tightening force: in samples belonging to the Xsigñ photometry test group, without a load, a fringe pattern is shown where the base of the stump and the fixing screw connect. However, during a load, the regions in which the stresses are concentrated are represented by the base of the stump, while significant stresses in the contact areas between fixing screw and the fixture's internal thread are not significant (Figure 20).

Concerning the samples belonging to the SSO photometry test group, without a load a fringe pattern consisting of bigger number of bands, concentrated in the base portion of the stump which contacts the internal surfaces of the fixture's neck, is shown; during a load, on top of that, the stress areas concentrate, besides the base of the stump, also along the threaded portion of the fixing screw (Figure 21).

The implant-abutment interface design, besides being a connection system, represents 2 debated topics of scientific literature. Many aspects guide the operator through the choice of one implant system over another. In this work, only the mechanical aspect concerning the maintenance of preload at the interfaces between the fixing screw, the stump, and the internal surfaces of the fixture were taken into consideration. A cyclic load ranging between 20 N and 200 N for 280 hours was used, which is equal to approximately 1 × 10⁶ cycles, because according to Cibirka et al,¹⁰  this corresponds to approximately 1 year of in vivo function. Furthermore, the implants were placed following an axis parallel to the load direction: this choice was determined by the fact that the load was to be applied at the summit of an abutment angled 15°, for a distance from the stump-implant connection of approximately 8 mm. The load applied during dynamic trials resulted so parallel to the major axis of the fixture, but distant from the latter approximately 1.5 mm, in a fashion that generated a torque.

After the application of the cyclic load, the 2 implant systems showed a slightly different behavior: 2 of the 15 samples belonging to the SSO test group turned out to have their fixing screw untightened, while all samples belonging to the Xsigñ test group kept their preload.

According to Ekfeldt et al,³  the loosening of the connection screw represents, overall, the most common complication, and other authors agree that the unscrewing of the abutment represents a problem that compromises the long-term success of the prosthetic rehabilitation.^4–6  In 2004, Cho et al¹¹  report that the loss of preload in standard size diameter (3.75 and 4 mm) reaches 14.5% where a dynamometric wrench is not used for screw tightening. In 1991, Jemt⁷  states that the loosening of the connection screw occurs more frequently in single restorations with respect to bridges.

In implant systems that have a screwed connection, the mechanical continuity between abutment and fixture is basically assured by the preload applied to the screw during tightening procedures and by the grade of fitting, that is the precision, existing between the abutment and the fixture. Relative to the retentive function offered by the connection screw, this basically depends on the preload applied during the tightening, which according to McGlumphy et al⁸  depends on at least 3 factors: (1) the applied torque, which affects directly the attrition under the screwhead, the attrition of the coils, and the grade of elastic/plastic deformation that the system undergoes; (2) the geometry of the screwhead that affects the attrition grade under the screwhead; and (3) the material that makes up the screw and the abutment, and in the end determines the level of grip existing between the 2 structures.

On the basis of what Cho et al¹¹  and McGlumphy et al⁸  point out, the application of an adequate torque level represents an element that affects in a determinant way the connection tightness. In general, the implant systems that are on the market allow the application of torques that range from 25 to 40 Ncm. According to McGlumphy et al⁸  and Haack et al,¹²  the preload applied is included between 60% and 75% of the maximal load beyond which the system undergoes plastic deformation. This value allows the connection's protection from occlusal loads, and at the same time, maximizes the screw resistance to fatigue. Preload values higher than 60%–75% of the elastic deformation limit, but still smaller than the rupture load, introduce plastic deformations that, paradoxically, promote unscrewing. According to Lang et al,¹³  it is possible to increase the preload by reducing the attrition coefficient existing between the system's components.

Also, there is a limit beyond which the maintenance of the preload gets compromised rather than improved. Bickford ¹⁴establishes that the attrition coefficient depends on the angle of the coil and the angle of the helix, according to the following relation: μ = tan θ cos Φ where: μ = attrition coefficient; θ = coil angle; and Φ = helix angle.

This relation establishes the value of the attrition coefficient below which the connection screw, independently from the existence of forces external to the system, loses the preload once it has been tightened. The presence of a high attrition coefficient, on the other hand, does not allow reaching the desired preload, even with a dynamometric wrench. The last consideration assumes an objective meaning in the work of Guda et al¹⁵  who find that in the oral environment the probability to obtain a tightening of the connection screw, which reaches 60%–75% of its elastic limit, is equal to 54.5%; the same parameter decreases to a value of 0.02% when the tightening is performed in a dry environment.

On the basis of these data, the present study chose to use a dynamometric wrench adjusted to values suggested by the manufacturer, in addition to wetting the insides of the implants with artificial saliva before the tightening procedures.

In the present work a different behavior has been observed in the 2 groups of examined samples; in this regard it needs to be considered that the connection screw gives retention to the abutment with respect to the fixture, while the stability of the whole system (ie, the capacity of it to oppose itself to the action of external forces) depends on many factors, likely linked to the design of the connection, which is the base of the abutment, the neck of the fixture and of the screw, and the grade of mechanical precision that the many components present during the interface. The analysis of the samples belonging to the control group highlights the existence of an adequate level of precision: the images relative to the coupling of the base of the stump and the neck of the fixture do not manifest the presence of irregularities in the profile or the presence of spaces between the components.

Regarding the geometric design that the 2 tested systems present, a first difference resides in the fact that the base of the SSO abutment has an octagonal and a conic portion each with an extension of approximately 2 mm, while the Xsigñ has a base that extends for approximately 3.2 mm made up of 2 cylindrical portions separated by a hexagonal one (Figure 22). The scanning electron microscope examination conducted on sections obtained from samples belonging to the control group confirms this aspect and highlights the fact that the Xsigñ stumps lean on the neck of the implant (Figure 9), while the SSO stumps lean directly on the internal aspect of the neck of the fixture (Figure 8). On the basis of these considerations, it is legitimate to conclude that after the application of a load that generates a torque, we have a different way to transfer the load. In the SSO system, the contact between the base of the stump and the neck of the fixture occurs on a surface, which is lesser than the one in the Xsigñ system; this fact accounts for the lower stability level of the SSO system design, and it therefore determines the onset of reciprocal micromovements, responsible for the loss of preload.

During the load application, the distribution of forces from the abutment to the fixture and the connection screw shows diverse results among the two implant systems considered in the present research. Sakaguchi and Borgersen,¹⁶  in a study conducted through the use of finite element analysis, highlight the fact that during the load, the stresses are transferred to the head and the stem of the connection screw, besides the first coil. The photoelastic analysis conducted in the present work partially confirms this finding. On one hand, in both implant systems after load application, a force transfer occurs from the base of the stump to the regions corresponding to the neck of the fixture, and on the other hand a difference related to stress transfer to fixing screws exists. The Xsigñ samples show the presence of chromatic bands right into the most coronal region of the screw, supporting the findings of Sakaguchi and Borgersen¹⁶ ; in the SSO samples, however, the load seems to be distributed to the whole threaded portion of the screw, which means that in the SSO system, given the lower stability offered by the stump, during load application a greater stress transfer occurs to the fixing screw, which loses more easily the applied preload.

Nevertheless, it needs to be considered that the 2 systems studied in this work imply also a different design of the prosthetic crown. In the Xsigñ system, the finishing edge of the crown closes onto the abutment. The SSO system, however, presents a design that contemplates the closing edge of the crown to lean onto the neck of the fixture. In the present work, we decided to apply the load directly to the abutment for reasons concerning standardization of the cyclic tests.

In clinical reality, the functional load is applied to the prosthetic crown, which transfers it to the other parts of the system. It is legitimate, therefore, to conclude that in clinical practice the SSO system could show a different biomechanic behavior, if we do not consider other variables related to the manufacturing of the prosthetic crown, basically represented by: (1) the precision of the closing edge leaning on the neck of the fixture; (2) the passivation of the prosthetic core in relation to the abutment; and (3) the endurance of sealing material used during cementation of the prosthesis. These factors correspond to the distribution of functional loads from the prosthetic crown to the neck of the fixture, minimizing the overload imposed onto the abutment.

The present in vitro study shows that the design of the abutment connection area affects the functional load transfer to the fixture and connection screw; the octagonal/conical morphology connection without the support of the fixture's neck transfers a greater load to the connection screw. The design based on the presence of 2 cylindrical portions with a hexagonal trait and the support of the fixture's neck, however, distributes the load mostly to the fixture than to the connection screw. From a clinical point of view, therefore, the Xsigñ system seems to be more suited to treat lateral-posterior mono-edentulism, while the SSO implant system should be used to treat those cases that require splinting together of 2 or more implants with the aim to minimize the risks of abutment unscrewing.