AUTHOR COPY

Bio-Medical Materials and Engineering 23 (2013) 317–327 317DOI 10.3233/BME-130756IOS Press

A FEM based endosteal implant simulationto determine the effect of peri-implant boneresorption on stress induced implant failure

N. Michailidis a,∗, G. Karabinas b, A. Tsouknidas a, G. Maliaris c, D. Tsipas a and P. Koidis b

a Physical Metallurgy Laboratory, Mechanical Engineering Department, Aristoteles University ofThessaloniki, Thessaloniki, Greeceb Department of Fixed Prosthesis and Implant Prosthodontics, Aristotle University of Thessaloniki,Thessaloniki, Greecec Laboratory for Special Engineering, Department of Electrical and Computer Engineering,Polytechnics School, Democritus University of Thrace, Komotini, Greece

Received 21 December 2011Accepted 4 February 2013

Abstract. Although dental implants exhibit only limited failure rates, their fracture is associated to major modifications of theprosthetic treatment and complex surgery for the removal of the remaining embedded implant part. This investigation aims toassess the developing stress fields in the bone–implant interface during mastication and asses the failure modes of oral implants.

In order to achieve this, a FEM model of an implant was reverse engineered and virtually loaded at the top of the crown fora force spectrum ranging from 75–225 N in a vertical, horizontal and oblique occlusal direction. The calculated stress fieldswere compared with clinically retrieved fractured implants with identical geometrical characteristics and the fracture modes ofboth cases were correlated. The developing stress patterns facilitated the interpretation of the implant failure as the maximumstresses, indicated critical values in both, lingual and buccal sides of the implant–bone interface at a certain critical level ofbone resorption, in which failure occurs.

Keywords: Peri-implant bone resorption, endosteal implant failure, FEM

1. Introduction

Endosteal implants represent the most common type of implant used for dental and facial rehabilita-tion. Gradient stress transition between the implant and the surrounding bone matter is a key successfactor of the restoration process pivotal for long term fixational stability, as mastication induces multi-axial forces and bending moments on clinical crowns, thus resulting in complex loading scenarios. Thestress distribution between prosthesis and peri-implant tissue depends, among others, on the geometrical

*Address for correspondence: Assoc. Prof. Dr.-Eng. Nikolaos Michailidis, Physical Metallurgy Laboratory, MechanicalEngineering Department, School of Engineering, Building D’, 9th floor, Aristotle University of Thessaloniki (BOX: 490),GR 54124 Thessaloniki, Greece. Tel.: +30 2310 995891; Fax: +30 2310 996069; E-mail: nmichail@eng.auth.gr; URL:http://pml.meng.auth.gr/.

0959-2989/13/$27.50 © 2013 – IOS Press and the authors. All rights reserved

HOR COPY

318 N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure

characteristics of the implant [1,2], the type of loading [3], as well as on the surrounding bone qual-ity [4,5]. The resulting stress field in the bone implant interface exerts a crucial effect on the implantfailure rate, associated not only to the fracture of the implant-prosthesis but also to the bone atrophy,as insufficient fixation or excessively increased stress, may hinder osseointegration and thus impendphysiological bone remodeling [6].

The annual number of dental implant placements is constantly increasing and has passed 5.5 mil-lion implants in 2010 compared to 0.8 million in 2004, as published by the American Academy ofImplant Dentistry. These numbers add significance to the precise evaluation of the implant–bone inter-face, rendering Finite Element Method (FEM) techniques as an increasingly useful tool during implantdevelopment, failure assessment and decision making throughout treatment.

Different theories have been proposed for the mechanisms that lead to implant failure and fracture[7,8]. In addition, several FEM based studies have been presented over the last years [9–11] in an effortto reveal the synergistic effects, describing the developing stress fields on both, maxilla or mandibleand embedded prosthesis. A geometrically accurate FEM model however of an endosteal implant andthe prosthetic superstructure, describing the stress transition of the implant–peri-implant structure andassociating these to retrieved fractured parts, has to the best of our knowledge yet to be presented.

2. Methods and materials applied in the study

2.1. FEM model

The FEM model introduced in the present paper is demonstrated in Fig. 1, both in an exploded andmeshed representation. The restored implant geometry (crown, implant screw, abutment and implant)was reverse engineered. The cancellous bone matter was described as a 3 mm thick cylinder surroundingthe implant screw, which was embedded 12.16 mm into the edentulous maxillary bone, the screw thread

Fig. 1. Reverse engineered 3D model and meshed geometry. (Colors are visible in the online version of the article; http://dx.doi.org/10.3233/BME-130756.)

AUTHOR COPY

N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure 319

was shaped trapezoid with a 0.58 mm pitch and 0.28 mm deep. The crowns abutment was fixed in theimplant at a 7.7 mm length and a 2 mm thick abutment distanced the crown from the implant screw.

In order to determine the effect of peri-implant bone resorption on the developing stress fields andasses at which level critical values occur, for both implant and surrounding bone, the FEM model wasredesigned considering a 1 and 2 mm bone retraction at the upper surface of the alveolar bone (as shownin Fig. 1). This resulted in an axial exposure of the implant at its upper extremity and is expected to leadto an increase of local stress due to the bending moment of the multi-axial load applied at the occlusalsurface of the crown.

In order to determine the optimum mesh density in terms of processing time and results accuracyand approximate a mesh independent grid, convergence studies were conducted for every model entityindividually, these considered the element aspect ratio and size allowing the consideration of all modelspecific characteristics. This is of significant importance in order to obtain a realistic and isotropic stresstransition within the bone–implant interface [12,13]. The achieved values are presented in Table 1.

2.2. Materials

In order to consider isotropic material properties for the alveolar bone matter, the model was acceptedto have reached an equilibrium of remodeling. In this way, the structural damage of the bone, causedby the surgical intervention, can be neglected. This assumption is quite accurate, as the osseointegrationphase of the traumatized bone occurs during the first months following the implantation [6]. Neverthe-less, a wide range of material properties can be found in literature concerning cancellous bone, due to thevariety of acquisition protocols that have been presented in individual studies and are further dispersedby disease, gender and age dependent criteria [14]. The material characteristics of the peri-implant boneemployed in the introduced model were considered linear (elastic) as presented in Table 2. The sametable illustrates the material properties of the implant [15].

Table 1

Mesh related figures of the introduced model

No. of elements Max element side length (mm) Min element side length (mm)

Crown 175,287 3.1 × 10−1 1.2 × 10−2

Bone 1,641,511 1.4 × 10−1 8.7 × 10−3

Implant 2,332,838 1.2 × 10−1 8.1 × 10−3

Abutment screw 232,967 9.1 × 10−2 8.3 × 10−3

Abutment 214,217 7.3 × 10−2 7.8 × 10−3

Table 2

Material properties

Young’s modulus (GPa) Poisson’s ratioCrown 100 0.33Bone 13.7 0.3Implant 110 0.33Abutment screw 100 0.33Abutment 100 0.33

AUTHOR COPY

320 N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure

2.3. Boundary conditions and loading scenarios

The assumption of optimal osseointegration also facilitated the consideration of fixed interfaces on allinteracting bone–implant surfaces, which were considered as fully bonded, while the model was fixed atthe cylindrical external face of the bone by constraining all degrees of freedom.

Endosteal implants are continually exposed to multi-axial dynamic and static loads comprising a rathercomplex loading system. Overall, three types of loads were independently applied at the occlusal surfaceof the crown at a vertical, horizontal and oblique occlusal direction to the implant axis. Although thisis not an accurate simulation of the dynamic loads occurring during mastication, as dental implants aresubjected to composite loading scenarios, the Saint Venant principle [16] facilitates the replacementof complex force systems by equivalent static loads, due to the rather small dimensions of the model.For each force direction, three loads were applied 75, 150 and 225 N, covering a broad spectrum ofmastication loads that have been reported over the last years in various investigations [17–19].

3. Results

The developed model was employed to systematically investigate the effect of the prosthesis loadingscenario and the bone resorption on the developing stress fields of the implant. Figure 2 illustrates thecalculated equivalent von Mises stress on the endosteal implant, at a 75 N horizontal load. The effect ofthe bone resorption, at the upper surface of the implant’s axial direction, is exhibited for the buccal andlingual implant side in Fig. 2a and b respectively.

The corresponding figure details indicate that the critical stress develops in all cases just a few mmbeneath the bones’ surface. The maximum stress for the completely embedded implant is within phys-iological values and even in the case of bone, subject up to a 1 mm retraction, well below the implantsyield stress. In alveolar bone however, exhibiting significant resorption (2 mm retraction), the developingstress reaches a critical value even at low force values.

Figure 3 illustrates the development of the calculated stress fields over a cross section of the endostealimplant as a function of the bone resorption progression, the illustrated stress distributions correspondto the worst case scenario of a 225 N horizontal load. The cross-section of the critical area exhibitingthe highest stress, advocates that these intense concentrations are a superficial phenomenon as the stressfield deteriorates to lesser values just a few mm below the implant surface.

The introduced model was further on employed to determine critical regions for stress induced failureof the examined implant. An overview of the maximum calculated equivalent von Mises stress on both,lingual and buccal implant side (healthy bone with no retraction), is presented in Fig. 4 consideringoverall 9 loading scenarios (75, 150 and 225 N in horizontal, vertical and oblique occlusal direction).The linearity of the registered maximum stress versus the applied force can be explained by the elasticproperties used during the simulation.

As indicated by this figure, the application of horizontal forces can be considered as a worst casescenario for the prosthesis. This behavior is also brought forth in all simulations regarding the insuffi-cient alveolar bone matter; Fig. 5 exhibits this biomechanical response in a 3D diagram, correlating thedeveloping max. von Mises stress to applied force and bone resorption.

The presented surfaces lead to conclusions regarding the implant failure, as stress in the magnitudeof 950 MPa should be considered as critical for the implant material. In a further step, the developingstress distributions, as calculated by the introduced FEM model, were compared to surgically retrieved

AU

N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure 321

Fig. 2. FEM calculated stress distribution of the endosteal implant subjected to a 75 N horizontal load and embedded at variousimplant lengths. (Colors are visible in the online version of the article; http://dx.doi.org/10.3233/BME-130756.)

fractured implants, examined with SEM. A characteristic example of an implant subjected to a 225 Nhorizontal load and a 2 mm bone resorption is demonstrated in Fig. 6. The figure illustrates the develop-ing stress distributions on both sides of the implant, indicating that the maximum developing stress onthe buccal side is concentrated in the region between pitch 1 and 2 while the lingual side of the implant,subjected to the same loading scenario, exhibits a critical region between pitch 2 and 3 (see Fig. 6).Based on the calculated results, the analytical model predicts fracture of the implant in this region, asthe maximum von Mises stress is on both sides higher than the materials strength.

The plane defined by the maximum occurring stresses of both, buccal and lingual implant sides ex-hibits good coherence when compared to the fractured surface of an identical implant, surgically re-covered form a patient and examined with SEM, as shown in the bottom right part of Fig. 6. To eluci-date however the etiology of the implant failure, the principal stress development during mastication is

COP

322 N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure

Fig. 3. FEM calculated stress distribution of the endosteal implant function of the bone resorption progression. (Colors arevisible in the online version of the article; http://dx.doi.org/10.3233/BME-130756.)

Fig. 4. Maximum force depended stress occurring on the implant. (Colors are visible in the online version of the article;http://dx.doi.org/10.3233/BME-130756.)

equally critical to the occurring von Mises stress. Figure 7 illustrates the principal stress distribution onthe prosthesis due to a vertical load of 225 N. It is evident that the stress concentration appears in thesame region as indicated by previous figures.

AUTHOR

N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure 323

Fig. 5. Maximum depended stress occurring on the implant as a function of both, applied force and bone resorption. (Colorsare visible in the online version of the article; http://dx.doi.org/10.3233/BME-130756.)

4. Discussion

The model indicated that past the threshold of 2 mm peri-implant bone resorption, both alveolar tissueand prosthesis exhibit critical stress values. This phenomenon is in good coherence with previous studies[20], as bone resorption or atrophy has been associated with implant fracture.

The course of the maximum occurring stress, as illustrated in Fig. 4, sustains that an implant embeddedat an insufficient length will result in intense stress concentrations on the prosthesis that may causeimplant failure. Another phenomenon of importance however is the stress distribution induced on thevicinal bone tissue. As the strength characteristics of cortical and cancellous bone are inferior to theones of the implant material (Ti-alloy), elevated stress due to improper implant fixation will impendphysiological bone remodeling [6] and thus disrupt or hinder the healing process of the “traumatized”alveolar bone.

There exist a consensus in literature that endosteal screw failure originates as a result of the appliedforce, implant properties and structural rigidity of the prosthesis. As the introduced FEM model consid-ered the same implant material and geometry during all loading scenarios, the investigation focused onthe effect of force magnitude and orientation on the developing stress fields. As expected, the simula-tions designated that the application of a horizontal force exerts elevated stress concentrations comparedto a vertical or oblique load of the same magnitude, hitherto several authors have drawn attention to thisfact [21].

AU

324 N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure

Fig. 6. Implant fracture development. (Colors are visible in the online version of the article; http://dx.doi.org/10.3233/BME-130756.)

The analysis visualized the developing von Mises stress fields, indicating critical regions within themodel which are in good coherence to those described in literature [22]. Comparing the analyses of theexamined spectrum of applied force scenarios, reveled a pattern of high stress concentration at the apicalregion of the implant–bone interface. All cases lead to qualitative similar stress distributions, with higherstress concentrations occurring at the buccal side of the implant neck just below its upper fixation point.

The computational results elicit that the biomechanical response of the implant is governed by highlyinterdisciplinary phenomena associated to structure related characteristics, a tendency also observedin previous studies [23]. The developing stress fields indicate areas with stress concentrations, which insome cases exceed the expected physiologic limit, a fact that has been reported to lead to bone resorptionand remodeling [24]. A closer however look at these stress fields, indicated that these intense phenomenaare superficial and rapidly depreciate to normal values just a few µm below the implant surface, thus

AUTHOR

N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure 325

Fig. 7. Principal stresses distribution developing for a 225 N vertical load. (Colors are visible in the online version of the article;http://dx.doi.org/10.3233/BME-130756.)

strengthening the hypothesis that implant failure is significantly accelerated in alveolar bone, exhibitingconsiderable resorption or atrophy. It should however be considered, that the in vivo structural integrity ofthe prosthesis is expected to be further deduced by cyclic fatigue phenomena as well as the aggressivelycorrosive environment in which the implants are embedded.

5. Conclusions

Our study indicated that relatively low stresses develop in the cancellous bone region compared tothe excessive high ones of the vicinal implant material, which can be attributed to the differing materialproperties of bone and implant.

The increased stiffness of the implant material develops high stresses, while the adjacent cancellousbone deforms, partially absorbing the induced load, thus leading to inferior stress distributions. This isevident along the entire fixation length of the implant and all loading scenarios, a phenomenon that hasbeen associated to the cylindrical implant screw geometry [25].

The simulations revealed that the fixation length of the implant (embedded screw length) is of pivotalimportance for long term stability. Poor osseointegrated implants are supported insufficiently and mightlead to fracture even when subjected to low forces.

AUTHOR COPY

326 N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure

Bone resorption, altering the initial implant/crown ratio, has a significant effect on the biomechanicalbehavior of the prosthesis. Implant revelation in the magnitude of 15% of the total embedded length,will lead to increased implant fatigue and possible fracture, even under mild loading scenarios.

The results of our study were verified by comparing the simulated stress fields to fractured implants(clinically retrieved). The correlation affirmed the bio-realistic stress development, as the predicted crit-ical regions calculated by the model cohere to the fracture site of the prosthesis.

References

[1] S. Yokoyama, N. Wakabayashi, M. Shiota and T. Ohyama, The influence of implant location and length on stress distri-bution for three-unit implant-supported posterior cantilever fixed partial dentures, J. Prosthet. Dent. 91 (2004), 234–240.

[2] C. Stellingsma, H.J. Meijer and G.M. Raghoebar, Use of short endosseous implants and an overdenture in the extremelyresorbed mandible: a five-year retrospective study, J. Oral Maxillofac. Surg. 58 (2000), 382–387.

[3] O. Kayabası, E. Yüzbasıoglu and F. Erzincanli, Static, dynamic and fatigue behaviors of dental implant using finiteelement method, Adv. Eng. Softw. 37 (2006), 649–658.

[4] D.C. Holmes and J.T. Loftus, Influence of bone quality on stress distribution for endosseous implants, J. Oral Implantol.23 (1997), 104–111.

[5] S.E. Clift, J. Fisher and C.J. Watson, Finite element stress and strain analysis of the bone surrounding a dental implant:effect of variations in bone modulus, Proc. Inst. Mech. Eng. H 206 (1992), 233–241.

[6] D. Lin, Q. Li, W. Li, N. Duckmanton and M. Swain, Mandibular bone remodeling induced by dental implant, J. Biomech.43 (2010), 287–293.

[7] K. Yokoyama, K. Kaneko, Y. Miyamoto, K. Asaoka, J. Sakai and M. Nagumo, Fracture associated with hydrogen ab-sorption of sustained tensile-loaded titanium in acid and neutral fluoride solutions, J. Biomed. Mater. Res. A 68 (2004),150–158.

[8] M.G. Manda, P.P. Psyllaki, D.N. Tsipas and P.T. Koidis, Observations on an in-vivo failure of a titanium dental im-plant/abutment screw system: A case report, J. Biomed. Mater. Res. B: Appl. Biomater. 89B (2009), 264–273.

[9] B. Simsek, E. Erkmen, D. Yilmaz and A. Eser, Effects of different inter-implant distances on the stress distribution aroundendosseous implants in posterior mandible: A 3D finite element analysis, Med. Eng. Phys. 28 (2006), 199–213.

[10] K. Degerliyurt, B. Simsek, E. Erkmen and A. Eser, Effects of different fixture geometries on the stress distribution inmandibular peri-implant structures: a 3-dimensional finite element analysis, Oral. Surg. Oral. Med. Oral. Pathol. Oral.Radiol. Endod. 110 (2010), 1–11.

[11] G. Limbert, C. van Lierde, O.L. Muraru, X.F. Walboomers, M. Frank, S. Hansson, J. Middleton and S. Jaecques, Tra-becular bone strains around a dental implant and associated micromotions – a micro-CT-based three-dimensional finiteelement study, J. Biomech. 43 (2010), 1251–1261.

[12] A. Tsoukindas, S. Maropoulos, S. Savvakis and N. Michailidis, FEM assisted evaluation of PMMA and Ti6Al4V asmaterials for cranioplasty resulting mechanical behaviour and the neurocranial protection, Biomed. Mater. Eng. 21 (2011),139–147.

[13] A. Tsoukindas, N. Michailidis, S. Savvakis, K. Anagnostidis, K.D. Bouzakis and G. Kapetanos, A FEM modelling tech-nique to determine the mechanical response of a lumbar spine segment under complex loads, J. Appl. Biomech. 28 (2012),448–456.

[14] B. Helgason, E. Perilli, E. Schileo, F. Taddei, S. Brynjólfsson and M. Viceconti, Mathematical relationships between bonedensity and mechanical properties: a literature review, Clin. Biomech. (Bristol, Avon) 23 (2008), 135–146.

[15] T. Kitagawa, Y. Tanimoto, M. Odaki, K. Nemoto and M. Aida, Influence of implant/abutment joint designs on abutmentscrew loosening in a dental implant system, J. Biomed. Mater. Res. B: Appl. Biomater. 75 (2005), 457–463.

[16] D. Iesan and A. Scalia, On Saint Venant’s principle for microstretch elastic bodies, Int. J. Eng. Sci. 35 (1997), 1277–1290.[17] S. Huthmann, C. Staszyk, H.G. Jacob, K. Rohn and H. Gasse, Biomechanical evaluation of the equine masticatory action:

Calculation of the masticatory forces occurring on the cheek tooth battery, J. Biomech. 42 (2009), 67–70.[18] G. Ortug, A new device for measuring mastication force (Gnathodynamometer), Ann. Anat. 184 (2002), 393–396.[19] J. Duyck, Biomechanical characterisation of in vivo load on oral implants, Thesis, Katholieke Universiteit Leuven, Bel-

gium, 2000.[20] R.E. Jung, B.E. Pjetursson, R. Glauser, A. Zembic, M. Zwahlen and N.P. Lang, A systematic review of the 5-year survival

and complication rates of implant-supported single crowns, Clin. Oral Impl. Res. 19 (2008), 119–130.[21] D.C. Holmes, W.R. Grigsby, V.K. Goel and J.C. Keller, Comparison of stress transmission in the IMZ implant system

with polyoxymethylene or titanium intramobile element: a finite element stress analysis, Int. J. Oral Maxillofac. Implants7 (1992), 450–458.

AUTHOR COPY

N. Michailidis et al. / A FEM based endosteal implant simulation on stress induced implant failure 327

[22] J.P. Geng, K.B.C. Tan and G.R. Liu, Application of finite element analysis in implant dentistry: A review of the literature,J. Prosthet. Dent. 85 (2001), 585–598.

[23] M. Cruza, A.F. Lourenc, E.M. Toledo, L.P. da Silva Barra, A.C. de Castro Lemonge and T. Wassall, Finite element stressanalysis of cuneiform and cylindrical threaded implant geometries, Technology and Health Care 13 (2006), 1–18.

[24] N. Inou, Y. Iioka, H. Fujiwara and K. Maki, Functional adaptation of mandibular bone, in: Computational Biomechanics,Springer, Heidelberg, 1996, pp. 23–42.

[25] M.R. Rieger, K. Fareed, W.K. Adams and R. Tanquist, A bone stress distribution for three endosseous implants, J. Prosthet.Dent. 61 (1989), 223–228.