Jean Perriard W. Anselm Wiskott Aissa Mellal Susanne S. Scherrer John Botsis Urs C.

Belser

Fatigue resistance of ITI implantabutment connectors a comparison of the standard cone with a novel internally keyed design

Authors afliations: Jean Perriard, H. W. Anselm Wiskott, Susanne S. Scherrer, Urs C. Belser, Division of Fixed Prosthodontics, School of Dental Medicine, University of Geneva, Switzerland Aissa Mellal, John Botsis, Laboratory for Applied Mechanics and Reliability Analysis, Swiss Federal Institute of Technology, Lausanne, Switzerland Correspondence to: Urs C. Belser Department of Prosthodontics University of Geneva School of Dental Medicine lemy-Menn 19, rue Barthe 1205 Geneva Switzerland Tel: 41 22 382 91 28 e-mail: Urs. Belser/medecine.unige.ch

Key words:dental implants; nite element analysis; mechanical stress Abstract:The Straumann Company has recently supplemented its standard morse-taper conguration with an octagonal internal key. During the restorative phase of implant treatment, this additional feature was designed to ensure positional duplicability between the laboratory and the clinical environments. It was, however, unclear whether this keying mechanism would decrease the mechanical strength of the connection between the implant and the abutment. This applies to keyed male and female parts but also to combinations of the new and the standard designs. Specially constructed specimens analogs representing all three combinations were tted with a T-shaped bar, preangled to 15 and subjected to vertical force applications provided by a servohydrolic fatigue tester. The loading frequency was 2 Hz and the maximum cycle number was 106. The data were evaluated using the staircase technique. The specimens were also modeled and analyzed numerically using nite element procedures. The samples failure locations were recorded and the displacement vs. cycle number plots were patterned in four groups. The fatigue tests and staircase analysis showed no difference in mechanical resistance between the standard and the internally keyed connectors. The nite element models revealed a stress concentration located at the apical edges of the octagonal connector. However, it appeared that this phenomenon was based on computational rather than mechanical grounds. The locations of the failure sites were distributed randomly across the structures, thereby indicating the absence of a locus of minor resistance. The patterns of the displacement vs. cycle number could not be attributed to specic combinations between the standard and the internally keyed designs. It was concluded that both connectors are equal in their mechanical resistance to bending and torquing forces.

Date:

Accepted 12 November 2001

To cite this article:

Perriard J, Wiskott WA, Mellal A, Scherrer SS, Botsis J, Belser UC. Fatigue resistance of ITI implantabutment connectors a comparison of the standard cone with a novel internally keyed design Clin. Oral Impl. Res, 13, 2002; 542549 Copyright C Blackwell Munksgaard 2002 ISSN 0905-7161

Prosthetic components are subjected to a complex pattern of horizontal and vertical force combinations (Graf & Geering 1977). Yet all force components do not have the same impact with respect to material resistance and incidence of failure. Force vectors that are directed along the main axis of the implant are compressive in nature and remain well below the materials resistance in compression (Glantz et al. 1993). By contrast, the bucco-lingual force components will result in bending of the

material and it is the tension and shear stresses thus developed that may cause failure of the structure. Further, in contrast to axial loading (Richter 1995), the bending effect is dependent on the height of the restoration and augments linearly as the length of the lever increases (Richter 1998). It thus follows that implant connector designs should be designed to ensure optimal load transfer of bending forces. One of the essential features of the ITIA implant system (Straumann, Waldenburg,

542

Perriard . Standard vs. internally keyed implant connectors

oratory experiments (Norton 1997) and in clinical environments (Felton et al. 1999; Levine et al. 1999). By design, a morse taper is rotation symmetric and thus lacks an antirotational keying mechanism. This precludes accurate laboratory transfers whenever the path of draw determined by the implant needs to be altered to better accommodate an abutment. To provide the ITI implant with such a keying device, the Straumann Company added an internal octagon mid-level of the cone of the implant body. Both the standard and the new cone designs are shown in Fig. 1(a, b). It was unclear, however, whether this additional feature would decrease the load transferring capacity of the joint. Therefore, the present study was initiated to test the hypothesis that no difference existed between the standard cone and the new internally keyed design.
Fig.1. Principle of internal keying. a) Standard biconal design. b) New, internally keyed, Octa design.

106 cycles of the three groups tested; stated differently, for each connector design in determining the load level at which 50% of specimens failed and 50% survived 106 load cycles. Conclusions were drawn after statistical comparison of the three means. The test specimens were divided into three groups: O implant (S) and abutment (S), O implant (O) and abutment (S), O implant (O) and abutment (O). Nine specimens were used in preruns to adjust the machine settings, 20 specimens were used for the OO combination and 10 for both the SS and OS pairs. In addition, the three experimental conditions were modeled using nite element procedures, the locations of the fracture sites were recorded and the displacement vs. number of cycles plots were analyzed.
Mechanical testing

Material and methods

Switzerland) is the conical interlock that connects the endosseous implant to the various types of prosthodontic attachments and which the company refers to as morse taper (Sutter et al. 1993). Morse taper is a term that stems from the tooling industry and which designates a keying mechanism in which a cone is tted within a cone (Schlosser 2001). The gripping action is due to the intimate contact and friction that develops in both elements when the male cone is gently tapped into the female element. This type of attachment is widely used to securely fasten drill bits or chucks to the rotating arbors of lathes or drill presses. The taper of the cone is indicated in degrees or in percent (d radius vs. d unit length). Percentages of 47% are typical. Also, such morse taper designs are characterized by their long shank, resulting in length to diameter ratios of 5 : 1. The ITIs cone total convergence angle is 16, its height is 2.3 mm and its diameter is 2.25 mm. Thus it is not a true morse taper as utilized in industrial applications but rather a biconal type of keying mechanism whose effectiveness is signicantly increased by the preload generated on the fraying surfaces of the cones by torque controlled bolting of the abutment into the endosseous implant. Such a conguration has proven highly suited to the load transfer of bucco-lingual bending forces both in labPrinciple

Specimen setup

The mechanical principle of the experiment was to cyclically load combinations of the standard (S) and the new, octagon keyed (O) designs of implant and abutment analogs and determine their resistance to fatigue failure. Rejecting or accepting the null hypothesis (i.e. no difference between both connector designs) consisted in comparing the mean resistance to failure (i.e. breakage) at

The setup of the specimens is diagrammatically shown in Fig. 2. The specimens (i.e. implant and abutment analogs) were inclined by 15 off the vertical (Merz et al. 2000) and the abutments were tted with 20-mm horizontal bars yielding a T-shaped arrangement. Loading was applied in a vertical direction at 5 mm off-center onto one end of the horizontal bar. This arrangement thus generated both a bending and a torquing moment on the conical joints. The T-bar arrangement was deemed

Fig.2. Principle of testing setup. The specimen was embedded into a resin-lled cylinder which was angled 15 off the vertical. The T-bar allowed a force application at 5 mm off-center. Both features combined allowed a torquing moment to be applied to the specimen.

543 |

Clin. Oral Impl. Res. 13, 2002 / 542549

Perriard . Standard vs. internally keyed implant connectors

necessary to determine how the specimens would react when loaded counterclockwise (untightening). Therefore, in the present setup the worst case situation (i.e. unscrewing) was used. For testing, the specimens were embedded in a cylindrical container lled with polymethyl methacrylate (PMMA) resin (Technovit 4071, Heraeus Kulzer, Wehrheim, Germany). In analogy with clinical implantbone relationships, the implant body was positioned so that the resin was level with the border between the polished and the rough portion of the implant. The implant carried a standard 6 taper 7-mm height abutment which was preloaded to 40 Ncm. The abutment was tted with a gold coping whose purpose was to interface the softer titanium abutment with the harder steel Tbar. The T-bar was machined with a halfround top. Since the denitive inclination of the specimen was to be experimentally determined rst, the half-round conguration provided a resting surface normal to the force vector applied but independent of specimen angulation. The steel bar was used for all test runs. A new gold coping was used after two runs were completed. New implant and abutment analogs were utilized for each test.
Testing machine

nus loadings at 2.05 Hz (2 Hz was chosen initially but at that frequency the feedback loop that controls the oil circuitry was unstable, hence the 2.05 Hz). The maximum load force was set according to the staircase procedure as described below.
Staircase procedure

The testing machine employed was a servohydraulic fatigue tester (Hydropuls, Schenk, Darmstadt, Germany). This machine was designed to generate various types of loading modes (sinus, square, ramp, preprogrammed) via an actuator bar that was applied to the specimens. It could be operated in load or displacement control. The machines settings as well as the parameters pertaining to the experiment (time, number of cycles, loads and displacements) were preset using an ancillary PC. To ensure proper function, the machine required that the specimens always be preloaded to at least 20 N, that is, that they would not be fully unloaded during cyclic force application. For the present experiments, the machine was set to generate si-

The procedure consists in determining the load level (Lm) at which 50% of the samples survive 106 stress cycles and 50% fail. 106 cycles is an arbitrarily set number whose theoretical and practical basis has been explained (Wiskott et al. 1994). The staircase procedure is a straightforward technique that applies to quantal (i.e. fail or not-fail) data. The method requires the samples to be tested consecutively in that the outcome (fail or not-fail) of a given samples test run determines the load level applied to the next sample tested. If the previous sample survived 106 cycles, the next sample is run at the previous load augmented by a predetermined amount. If the previous sample failed, the next sample is run at the previous level minus the predetermined amount. This generates an upand-down pattern of fail and not-fail loads, hence the name staircase. After suitable arrangement of the data, the mean (50% failures and 50% run-outs) and the standard deviation are calculated. The computational aspects of the technique have been described elsewhere (Wiskott et al. 1994; Dieter 1961; Draughn 1979). At the onset, both an entry force level and an increment/decrement must be determined before the test sequence is

started. The entry force level was determined using the equation F M/d sina where F is force applied, M is torque applied during abutment tightening, d is lever length and a is the inclination of the specimen. For M 40 Ncm, d 5 mm and a 15, the estimated force would be 308.8 N. Using this value, it was decided to set the entry force level to 2/3 of the maximum, that is, 205.9 N. The increment/decrement was set to 20 N.
Data analysis

The results of the staircase analyses in terms of Lm and standard deviation were computed. Means were compared by tting with 95% condence intervals. Means whose intervals did not overlap were considered statistically equal.
Numerical analysis

To identify zones of stress concentration, the SS, OS and OO combinations of connectors were modeled using a nite element software (I-DEAS Master MS8, SDRC, Cincinnati, OH, USA) (Curnier 1994) that was installed on a HP 735 workstation in a Unix environment. For modeling, the components were assumed homogeneous, linearly elastic and isotropic. Material characteristics are listed in Table 1. The models were a tridimensional mesh of tetrahedra incorporating approximately 16000 elements and 3600 nodes (combination including the Octa design increased the number of cells required). A layer of friction elements was inserted be-

Table 1. Material parameters used in the numerical analyses E (GPa)

Titanium Steel PMMA 110 206.8 2.38

Poissons ratio
0.3 0.29 0.41

Fig.3. Fatigue resistance of octagon octagon (OO), smooth smooth (SS) and octagon smooth (OS) combinations. Both the OO and the SS combination had overlapping condence intervals. The SO combination was superior to both OO and SS.

544 |

Clin. Oral Impl. Res. 13, 2002 / 542549

Perriard . Standard vs. internally keyed implant connectors

tween the implant and the abutment. Such elements model the transfer of pressures with respect to the relative motion of both

surfaces in contact. The coefcient of friction was set to 0.5 (Abkowitz et al. 1960). The components were rst subdivided into

substructures which were subsequently meshed automatically by the software. There was no need to specically mesh the

Fig.4. FE analysis of SS combinations. Peak stress was 499 Mpa.

Fig.5. FE analysis of SO combinations. Peak stress was 562 Mpa.

Fig.6. FE analysis of OO combinations. Peak stress was 26?900 Mpa.

545 |

Clin. Oral Impl. Res. 13, 2002 / 542549

Perriard . Standard vs. internally keyed implant connectors

T-bar since the torquing force generated by the lever system of the bar was computed and integrated into the numerical simulation. Calculations were performed using an applied force of 300 N.

Location of fracture sites

Discussion
Stress concentrations

Location of fracture sites

The location of the fracture sites was recorded. This information was of signicance to determine whether there was a denite locus of minor resistance inside the connector or whether fracture occurred at random within the structure.

The location of the failure sites is summarized in Fig. 7. Both ssures (dotted lines) and overt fractures (solid lines) were observed. They occurred in the implant at the level of the tip of the screw; in the screw threads, close to the junction with the cone; at three levels of the SynOcta male part and in the solid cone. For all combinations there were at least three sites where failure occurred. No preferential location was detected.

Displacement vs. cycle number plots Displacement recording vs. number of cycles

For each load level applied during staircase analysis, every 50 cycles, the system recorded the cycle number and the displacement of the machines actuator bar. This was deemed necessary to identify possible deciencies in the specimen setup. It was also meant to determine whether a systematic difference existed between specimens that fractured and those that did not. It was hypothesized that specimens that eventually failed might present a growing ssure which would translate into an increasing displacement of the actuator bar.

Plots depicting actuator bar displacement vs. number of cycles were generated for all specimens. There was no systematic behavior that characterized either failed vs. not-failed specimens or which differentiated the three combinations. Nevertheless, four patterns were observed (Fig. 8): stability over the whole run; slow increase of actuator bar displacement, no failure; rapid increase of actuator bar displacement followed by fracture; stability followed by fracture.

The data presented above indicate no denite trend with respect to the null hypothesis (no difference between SS and O O) in any of the tests performed, with one exception, the inordinately high stress concentration on the mating surface of the Octa connectors apical angles. When considering the numerical values obtained (2.69104 MPa peak stress) any OO combination should fail under a load in excess of 6 N (600 g, that is) given a tensile resistance of titanium of 500 MPa (Ashby & Jones 1986). While the magnitude of the values computed can largely be attributed to deciencies in meshing of the model, the phenomenon observed should not be rejected off-hand since it is known that angles generate stress concentrations (Broek 1988). (Incidentally, these do also appear on the cones mating surfaces in the OS combination and may be at the origin of the ssures shown in Fig. 7). So much so that the machining process developed by the manufacturer includes substantial rounding of the angle of the SynOcta male part as shown in Fig. 9. Due to the ge-

Results
Fatigue resistance

The fatigue resistance of the three combinations is presented in Fig. 3. The condence intervals at 95% are also shown. While the internally keyed and the standard designs clearly had overlapping condence intervals, the SO combination presented a superior resistance to force application.

Numerical analysis (Figs46)

Numerical analysis depicts the stress variations (MPa) inside the structures under an applied load of 300 N. The stress intensity bar on the right indicates the highest and lowest stresses appearing in the structure while the 2- and 3D mappings specify the locations of the stresses. While the SS and the SO congurations are essentially similar in terms of the stresses induced, the software identied extreme stress concentrations in the OO combination. These appear in the apical portion of the line angles of the Octa structure (Fig. 6).

Fig.7. Locations of failure sites. Encircled numbers denote the number of occurrences. Dotted lines represent ssures. Solid lines represent fractures.

546 |

Clin. Oral Impl. Res. 13, 2002 / 542549

Perriard . Standard vs. internally keyed implant connectors

Fig.9. SynOcta element from the bottom. Note rounding of the external edges of the screw head.

not mechanically driven. This consideration holds especially in light of the fact that no component fractured in that zone during our tests.

turer is located in a section of the cone that bears the smallest magnitudes of stresses and thus that the risk of failure is greatly reduced.
Mechanics of bolted surfaces

Failure modes and fracture sites

Fig.8. Displacement vs. number of cycle curves. Four patterns were observed. Pattern a: stability over the entire run. Pattern b: slow increase of maximum actuator bar displacement. Pattern c: rapid increase of actuator bar displacement. Pattern d: slight increase followed by fracture.

ometry of the Octa connector, the FE model constructed for the present study could not be simplied to 2D or axisymmetric meshings but included the intricacies of the keying mechanism. Nevertheless, the FE analysis conducted here was static in that it ignored the plastic deformation of the metal. Yet this applies particularly to commercially pure (i.e. unalloyed) titanium where rapid deformation in zones of pressure contact is expected. Hence it is highly likely that the stress phenomenon in the FE software is mathematically and

The four patterns observed in the displacement vs. number of cycle plots (Fig. 8) presumably represent different types of alterations inside the structures. Patterns b and c most likely indicate the progression of a ssure at a fairly slow rate (b) and at a faster rate (c). It is probable that specimen b would have failed at a higher cycle number. The specimens following pattern a were stable and no ssuring occurred. Pattern d, in which displacement actually decreased before fracture, can be explained on the basis of work-hardening, which tends to stiffen the metal (and also to increase its brittleness). The heterogeneity observed in the fracture sites basically eliminates the possibility of a locus of minor resistance. This nding positively valuates the design chosen.

Stresses in conical joints

In a previous study, we had modeled the behavior of cemented conical joints under lateral loading (Wiskott et al. 1999). It was shown that crowns rotated around an axis located mid-level of the cone and that the stress levels in that zone decreased to a minimum. It follows that the internal keying mechanism designed by the manufac-

The various keying mechanisms proposed in dental implantology require some understanding of the basic mechanics of bolted joints. Consider two at plates that are bolted together and onto which a tensile force is applied parallel to the long axis of the bolt, in effect pulling the at plates apart. Initially, when the bolt is tightened, the fraying surfaces (i.e. the portions of the surfaces that come in contact) are drawn together and develop a compressive force onto the mating parts. The joint is now in equilibrium with the compressive force across the fraying surfaces equal to the tensile stress inside the bolt. This internal stress is called pretension and the force on the mating surface is referred to as preload. If an external tensile force is applied onto the plates, the assembly responds in a somewhat unexpected way in that the applied force is not concentrated onto the bolt but distributed along the entire fraying surface. To some extent, the net effect of the preload is to alter the behavior of the two plates as if they were a single component. This effectively shields the bolt from large variations in tension and therefore substantially enhances its resistance to fatigue failure. The magnitude of shielding depends on the joints material and geometry but it is not unusual that as much as 90% of the applied load is dissipated via

547 |

Clin. Oral Impl. Res. 13, 2002 / 542549

Perriard . Standard vs. internally keyed implant connectors

the mating surfaces, leaving only 10% to be borne by the screw. From the above, it follows that, in order to ensure an optimal mechanical continuum, the pretension in the screw must be as high as possible, often in the order of 6070% of the ultimate tensile strength of the screw (Haack et al. 1995). This effect has been put to use in the cone in cone design of the ITI connector (Sutter et al. 1993). However, due to the machining tolerances (Binon 1995) and their near-parallel design, the vertical sides of the Octa connector do not actually carry load and thus the role of the internal keying mechanism is to ensure positional duplicability between laboratory and clinical phases of treatment but not to provide any noteworthy contribution as an antirotational device during function.

canique. Les localisations des sites avec de calcul que me chec e taient distribue es au hasard le long des structures e sisindiquant ainsi labsence dun endroit de moindre re `les de de placement vs cycle ne pouvaient tance. Les mode ` des associations spe s a ciques entre les tre attribue pas e ` clef interne. Les deux connexions `les standards et a mode sistance me canique aux sont donc semblables dans leur re forces de pliage et de torsion.

bra con llave sino a las combinaciones de los disen os nuendar. vos y esta logos especialmente Se ajustaron unos especimenes ana construidos representando las tres combinaciones posibles a una barra con forma de T, preangulada a 15 grados y sometida a una fuerza vertical suministrada por un probador de fatiga servohidraulico. La frecuencia de carga fue ximo de ciclos fue de 106. Los de 2Hz y el numero ma cnico e la escalera. Los datos se evaluaron usando la te ricamente especimenes se modelaron y analizaron nume usando procedimientos de elementos nitos. Se recogieron las localizaciones de los fracasos de las muestras y el desplazamiento frente al nu mero de ciclos se agruparon en cuatro patrones. lisis de la escalera no evidenLas pruebas de fatiga y el ana nica entre los cociaron diferencias en la resistencia meca ndar y los de llave interna. Los modelos elenectores esta n de estre s mentos nitos evidenciaron una concentracio localizada en los bordes apicales del conector octogonal. meno se baso en De todos modos, parece que este feno nicos. Las datos computacionales mas que en datos meca localizaciones de los lugares de fracaso se distribuyeron aleatoriamente a lo largo de las estructuras indicando por ello la ausencia de un lugar de menor resistencia. Los patrones de desplazamiento frente a los ciclos no se pucas entre los disedieron atribuir a combinaciones espec ndar y los de llave interna. n os esta que ambos conectores son iguales en su reSe concluyo nica a las fuerzas de doblaje y de torque. sistencia meca

Zusammenfassung
Die Firma Straumann hat krzlich die Standard-Konusform mit einer achteckigen internen Kantenbahn erweitert. Diese zustzliche Eigenschaft wurde entwickelt, damit whrend der restaurativen Phase der Implantattherapie die Uebertragung der Implantatposition zwischen Labor und der klinischen Umgebung und umgekehrt gesichert werden kann. Es ist jedoch unklar, ob dieser Verschlsselungsmechanismus die mechanische Strke der Verbindung zwischen dem Implantat und dem Sekundrteil beeinusst. Dies betrifft sowohl die mnnlichen und weiblichen Teile mit Verschlsselungsmechanismus als auch die Kombinationen von neuen und Standardteilen. Speziell konstruierte Testanaloge, welche alle drei Kombinationsmglichkeiten wiederspiegelten, wurden an einem T-frmigen Balken befestigt oder um 15 abgewinkelt und vertikalen Krften, welche von einem servohydraulischen Belastungstester generiert wurden, ausgesetzt. Die Belastungsfrequenz betrug 2 Hz und die maximale Anzahl Belastungszyklen betrug 106. Die Daten wurden mittels Treppenstufen-Technik ausgewertet. Die Testkrper wurden ebenfalls als Computermodell numerisch mittels der Finite-Element-Verfahren analysiert. Die Stellen der Misserfolge bei den Testkrpern wurden aufgezeichnet und die Graphiken der Verschiebung gegenber der Anzahl Belastungszyklen wurden gemss Muster in 4 Gruppen aufgeteilt. Der Ermdungstest und die Treppenstufenanalyse zeigten keinen Unterschied in der mechanischen Widerstandsfhigkeit zwischen den Standdardteilen und den intern gesicherten Verbindungen. Die Finite-Element-Modelle ergaben eine Stresskonzentration, welche im Bereich der apikalen Kanten des achteckigen Verbinders lokalisiert war. Es schien jedoch, dass dieses Phnomen mehr auf computertechnischen als auf mechanischen Grnden beruhte. Die Regionen der Misserfolge waren entlang der Strukturen zufllig verteilt. Dies zeigt, dass keine schchste Stelle existiert. Die Muster der Displazierung gegenber den Belastungszyklen konnten nicht spezischen Kombinationen zwischen den Standardteilen und den intern gesicherten Teilen zugeordnet werden. Es wird die Schlussfolgerung gezogen, dass beide Verbindungen in Bezug auf mechanischen Widerstang gegen Biege- und Drehrfte gleichwertig sind.

Conclusion
The data gathered in the present study do not provide a basis for rejecting the null hypothesis of no difference between the standard cone and the newly designed internally keyed (Octa) design. Therefore both connectors are considered equal in their mechanical resistance to bending and torquing forces.

sume Re
cemment remplace sa La compagnie Straumann a re conguration standard par une clef interne octogonale. te effecDurant la phase de restauration, cette addition a e e pour assurer la possibilite de reproduire la position tue tait pas certain que ce entre laboratoire et clinique. Il ne canisme de clef pouvait diminuer la force me canique me de la connexion entre limplant et le pilier. Ceci sappliles et femelles de la clef mais aussi que aux parties ma `les nouveaux et stanaux combinaisons entre les mode cimens construits spe cialement repre sendards. Les spe te fabrique s avec une tant les trois combinaisons ont e ` 15 degre ` angule e a s et sujette a barre en forme de T pre es par un syste `me de fatigue des forces verticales apporte quence de charge e tait de 2Hz et servohydraulique. La fre es ont le nombre de cycles maximum de 106. Les donne te e value es en utilisant la technique par e chelons. Les e chantillons e taient aussi modele s et analyse s nume rie le ments nis. Les quement en utilisant des processus de checs ont e te note es et le de placement localisations des e te analyse par plots en quatre groupes. Les vs cycle a e chelons nont mis en tests de fatigue et lanalyse par e vidence aucune diffe rence entre les connecteurs stane sistance me canidards et les internes au niveau de la re `les de le ments nis ont mis en e vidence que. Les mode e dans les parties apiune concentration du stress localise t cales de la connexion octogonale. Cependant, il appara nome `ne e tait bien plus base sur des proble `mes que ce phe

Resumen
a Straumann ha suplementado recientemente La compan ndar de morse-taper con una llave n esta su conguracio octogonal interna. Durante la fase restaurativa del trata esta caracter stica adiciomiento de implantes, se disen o nal para asegurar la duplicidad posicional entre las condinicas. De todos modos no estaciones de laboratorio y cl a la ba claro si este mecanismo de llave disminuir nica de la conexio n entre el implante y resistencia meca el pilar. Esto se aplica no solo a las partes macho y hem-

548 |

Clin. Oral Impl. Res. 13, 2002 / 542549

Perriard . Standard vs. internally keyed implant connectors

References
Abkowitz, S., Burke, J.J. & Hiltz, R.H. (1960) Titanium in industry. New York: Van Nostrand. Ashby, M.F. & Jones, D.R.H. (1986) Engineering materials. 2 An introduction to microstructures, processing and design. Oxford: Pergamon Press. Binon, P.P. (1995) Evaluation of machining accuracy and consistency of selected implants, standard abutments, and laboratory analogs. International Journal of Prosthodontics 8: 162178. Broek, D. (1988) The practical use of fracture mechanics. Dordrecht: Kluwer Academic Publishers. Curnier, A. (1994) Computational methods in solid mechanics. Dordrecht: Kluwer Academic Publishers. Dieter, G. (1961) Mechanical metallurgy. New York: McGraw-Hill. Draughn, R.A. (1979) Compressive fatigue limits of composite restorative materials. Journal of Dental Research 58: 10931096. Felton, D.A., Chee, W., Johnson, P.F. & Sullivan, D.Y. (1999) Cemented versus screw-retained implant prostheses: which is better? International Journal of Oral Maxillofacial Implants 14: 137141. Glantz, P.O., Rangert, B., Svensson, A. et al. (1993) On clinical loading of osseointegrated implants. A methodological and clinical study. Clinical Oral Implants Research 4: 99105. Graf, H. & Geering, A.H. (1977) Rationale for clinical application of different occlusal philosophies. Oral Science Review 10: 110. Haack, J.E., Sakaguchi, R.L., Sun, T. & Coffey, J.P. (1995) Elongation and preload stress in dental implant abutment screws. International Journal of Oral Maxillofacial Implants 10: 529536. Levine, R.A., Clem, D.S., Wilson, T.G., Higginbottom, F. & Solnit, G. (1999) Multicenter retrospective analysis of the ITI implant system used for single-tooth replacements: results of loading for 2 or more years. International Journal of Oral Maxillofacial Implants 14: 516520. Merz, B.R., Hunenbart, S. & Belser, U.C. (2000) Mechanics of the implant-abutment connection: an 8-degree taper compared to a butt joint connection. International Journal of Oral Maxillofacial Implants 15: 519 526. Norton, M.R. (1997) An in vitro evaluation of the strength of an internal conical interface compared to a butt joint interface in implant design. Clinical Oral Implants Research 8: 290298. Richter, E.J. (1995) In vivo vertical forces on implants. International Journal of Oral Maxillofacial Implants 10: 99108. Richter, E.J. (1998) In vivo horizontal bending moments on implants. International Journal of Oral Maxillofacial Implants 13: 232244. Schlosser (2001) Morse taper, Vol. 2001. http://showcase.netins.net/web/schlosser/mttaper.htm. Sutter, F., Weber, H., Sorenson, J. & Belser, U. (1993) The new restorative concept of the ITI dental implant system: Design and engineering. International Journal of Periodontic and Restorative Dentistry 13: 409431. Wiskott, H.W., Krebs, C., Scherrer, S.S., Botsis, J. & Belser, U.C. (1999) Compressive and tensile zones in the cement interface of full crowns: a technical note on the concept of resistance. Journal of Prosthodontics 8: 8091. Wiskott, H.W., Nicholls, J.I. & Belser, U.C. (1994) Fatigue resistance of soldered joints: a methodological study. Dental Materials 10: 215220.

549 |

Clin. Oral Impl. Res. 13, 2002 / 542549