d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) e218–e228

Available online at www.sciencedirect.com

jo ur n al homep age : w ww.int l .e lsev ierhea l th .com/ journa ls /dema

Biomechanical evaluation of internal and external hexagonplatform switched implant-abutment connections:An in vitro laboratory and three-dimensional finiteelement analysis

Amilcar C. Freitas-Júniora,b,∗, Eduardo P. Rochac, Estevam A. Bonfanted,Erika O. Almeidaa,c, Rodolfo B. Anchietac, Ana P. Martini c, Wirley G. Assuncãoc,Nelson R.F.A. Silvaa, Paulo G. Coelhoa

a Department of Biomaterials and Biomimetics, New York University College of Dentistry, NY, USAb Postgraduate Program in Dentistry, Potiguar University - School of Health Sciences, Natal, RN, Brazilc Department of Dental Materials and Prosthodontics, São Paulo State University, Aracatuba School of Dentistry, Aracatuba, SP, Brazild Postgraduate Program in Dentistry, Unigranrio University - School of Health Sciences, Duque de Caxias, RJ, Brazil

a r t i c l e i n f o

Article history:

Received 6 September 2011

Received in revised form

11 May 2012

Accepted 17 May 2012

Keywords:

Dental implants

Platform switching

Biomechanics

Reliability

Finite element analysis

a b s t r a c t

Objectives. The aim of this study was to assess the effect of abutment’s diameter shifting on

reliability and stress distribution within the implant-abutment connection for internal and

external hexagon implants. The postulated hypothesis was that platform-switched implants

would result in increased stress concentration within the implant-abutment connection,

leading to the systems’ lower reliability.

Methods. Eighty-four implants were divided in four groups (n = 21): REG-EH and SWT-EH (reg-

ular and switched-platform implants with external connection, respectively); REG-IH and

SWT-IH (regular and switched-platform implants with internal connection, respectively).

The corresponding abutments were screwed to the implants and standardized maxillary

central incisor metal crowns were cemented and subjected to step-stress accelerated life

testing. Use-level probability Weibull curves and reliability were calculated. Four finite ele-

ment models reproducing the characteristics of specimens used in laboratory testing were

created. The models were full constrained on the bottom and lateral surface of the cylin-

der of acrylic resin and one 30◦ off-axis load (300 N) was applied on the lingual side of the

crown (close to the incisal edge) in order to evaluate the stress distribution (svM) within the

implant-abutment complex.

Results. The Beta values for groups SWT-EH (1.31), REG-EH (1.55), SWT-IH (1.83) and REG-IH

(1.82) indicated that fatigue accelerated the failure of all groups. The higher levels of �vM

within the implant-abutment connection observed for platform-switched implants (groups

SWT-EH and SWT-IH) were in agreement with the lower reliability observed for the exter-

nal hex implants, but not for the internal hex implants. The reliability 90% confidence

intervals (50,000 cycles at 300 N) were 0.53(0.33–0.70), 0.93(0.80–0.97), 0.99(0.93–0.99) and

0.99(0.99–1.00), for the SWT-EH, REG-EH, SWT-IH, and REH-IH, respectively.

∗ Corresponding author at: Department of Biomaterials & Biomimetics, New York University College of Dentistry,345E 24th Street, Room 812, New York, NY 10010, USA. Tel.: +1 55 84 88402345.

E-mail address: ac.freitas.jr@gmail.com (A.C. Freitas-Júnior).0109-5641/$ – see front matter © 2012 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.dental.2012.05.004

d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) e218–e228 e219

The postulated hypothesis was partially accepted. The higher levels of stress observed

within implant-abutment connection when reducing abutment diameter (cross-sectional

area) resulted in lower reliability for external hex implants, but not for internal hex implants.

emy

1

Iihbtamoli(bftmp

[bili

cipssc[ssoeiasiaf[

aif[poiis

the increasingly step-wise rapidness in which a specimen is

© 2012 Acad

. Introduction

n the first year after implant insertion and loading, early peri-mplant bone loss commonly leads to a reduction in boneeight, shown to vary as a function of quality and quantity ofone, implant and abutment designs, implant’s surface struc-ure, insertion depths, arch region, and other factors [1–3]. Anttempt to hinder this process has resulted in the develop-ent of the platform switching concept, which consists in use

f an abutment of smaller diameter connected to a implant ofarger diameter. This connection shifts the perimeter of themplant-abutment junction inward towards the central axisthe middle of the implant), potentially improving the distri-ution of forces and placing the implant-abutment gap awayrom the peri-implant bone [4,5]. It has been suggested thathe inward shift of the implant-abutment gap may physically

inimize the impact of the inflammatory cell infiltrate in theeri implant tissues, potentially reducing bone loss [2,6–11].

From a biomechanical perspective, previous in vitro studies12–17] have shown reduced levels of stress on peri-implantone in platform-switched implants relative to matched

mplant-abutment diameters. Such potential for crestal boneevel preservation has been shown in animal [18–20] and clin-cal studies [11,21,22].

On the other hand, complications with implant-abutmentonnections is still a common clinical problem, especiallyn single-tooth replacements [18,23,24]. When consideringlatform-switched implants, previous studies [14,17] havehown an increased stress on the abutment and fixationcrew, which may compromise the system biomechani-al performance. Controversially, several published studies1,12,13,15,16,19,25–27] related to the mechanics of platform-witched implants have been restricted to analyzing thetress distribution on peri-implant bone and not on theverall system biomechanical behavior. To date, studiesvaluating the mechanical behavior of platform-switchedmplants considering the stress distribution in implant-butment complex are scarce and restricted to computerimulations [14,17,28,29], which do not consider several clin-cal variables (influence of fatigue damage accumulationnd wet environment) previously reported as importantactors to reproduce clinically observed failure modes30].

Since the main challenges in the development of implant-butment connection designs comprises reducing thencidence of mechanical failures while improving the inter-ace between soft tissue and implant-abutment junction31,32], the evaluation of reliability and failure modes sup-orted by evaluation of stress distribution in each componentf platform-switched connections may provide insight

nto the mechanical behavior of different configurations ofmplant-abutment connection. Therefore, the present studyought to assess the effect of abutment’s diameter shifting

of Dental Materials. Published by Elsevier Ltd. All rights reserved.

(regular and switched-platform) on reliability and failuremodes of anatomically correct maxillary central incisorcrowns varying the geometry of implant connection (internaland external hexagon). In order to evaluate the stress distribu-tion within implant-abutment complex (implant, abutmentand fixation screw), a three-dimensional finite element anal-ysis was performed considering the variables. The postulatedhypothesis was that platform-switched implants wouldresult in increased stress concentration within the implant-abutment connection, leading to the systems’ lower reliabilitywhen subjected to step-stress accelerated life testing(SSALT).

2. Materials and methods

2.1. In vitro laboratory study: single load-to-fracture(SLF) and step-stress accelerated-life testing (SSALT)

Eighty-four commercially pure titanium grade 2 dentalimplants (SIN implants, São Paulo, SP, Brazil) were dis-tributed in four groups (n = 21 each) varying the abutmentdiameter (switched or regular platform) and the type ofimplant connection (internal or external hexagon) (Fig. 1and Table 1): (1) SWT-EH (switching platform and externalhexagon implant); (2) REG-EH (regular platform and externalhexagon implant); (3) SWT-IH (switching platform and inter-nal hexagon implant); and (4) REG-IH (regular platform andinternal hexagon implant).

All implants were vertically embedded in acrylic resin(Orthoresin, Degudent, Mainz, Germany), poured in a 25-mm-diameter plastic tube, leaving the top platform in thesame level of the potting surface (Fig. 2). All groups wererestored with standardized central incisor metallic crowns(CoCr metal alloy, Wirobond® 280, BEGO, Bremen, Germany)cemented (Rely X Unicem, 3M ESPE, St. Paul, MN, USA) onthe abutments, which presented identical height but differentdiameters (Table 1).

For mechanical testing, the specimens were subjected to30◦ off-axis loading (Fig. 2C). Three specimens of each groupunderwent single-load-to-fracture (SLF) testing at a cross-head speed of 1 mm/min in a universal testing machine(INSTRON 5666, Canton, MA, USA) with a flat tungsten car-bide indenter applying the load on the lingual side of thecrown, close to the incisal edge. Based upon the mean loadto failure from SLF, three step-stress accelerated life-testingprofiles were determined for the remaining 18 specimens ofeach group which were assigned to a mild (n = 9), moderate(n = 6), and aggressive (n = 3) fatigue profiles (ratio 3:2:1, respec-tively) [30,33]. Mild, moderate and aggressive profiles refer to

fatigued to reach a certain level of load, meaning that speci-mens assigned to a mild profile will be cycled longer to reachthe same load of a specimen assigned to either moderate or

e220 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) e218–e228

Fig. 1 – Three-dimensional models of implant-abutment connections to be tested in the present study. (A) and (B) SWT-EHand REG-EH (switching and regular platform connected to an external hex implant, respectively). (C) and (D) SWT-IH and -

rnal

REG-IH (switching and regular platform connected to an inte

aggressive profiles. In the present study, the profiles startedat a load that was approximately 30% of the mean value ofSLF and ended at a load that was approximately 60% of thesame value. The rationale for utilizing at least three profiles forthis type of testing was based on the need to distribute failureacross different step loads and allows better prediction statis-tics, narrowing confidence bounds. The prescribed fatiguemethod was step-stress accelerated life-testing (SSALT) underwater at 9 Hz with a servo-all-electric system (TestResources800L, Shakopee, MN, USA) where the indenter contacted the

crown surface, applied the prescribed load within the stepprofile and lifted-off the crown surface. Thus, during SSALTeach specimen was submitted to constant stress during a

Fig. 2 – (A) Component assembling for the switching and regularexternal and internal connection groups: (1 and 2) SWT-EH and Rexternal hex implant, respectively); (3 and 4) SWT-IH and REG-IHhex implant, respectively). (B) Implant connection configurationsright) internal hexagon; poured in a 25-mm-diameter plastic tubwas applied at 30◦ to the long axis of the implant.

hex implant, respectively).

predetermined length of time. The stress on this specimenis thus increased step by step until failure (bending or fractureof the fixation screw and/or abutment) or survival (no failureoccurred at the end of step-stress profiles, where maximumloads were up to 600 N) [30,33]. Based upon the step-stress dis-tribution of the failures, the fatigue data were analyzed usinga power law relationship for damage accumulation and theuse level probability Weibull curves (probability of failure vs.cycles) at a use stress load were determined for life expectancycalculations by using the software Alta Pro 7 (Reliasoft,

Tucson, AZ) [34]. The master Weibull curves obtained fromthe SSALT fatigue data were used to determine the reliabil-ity (the probability of an item functioning for a given amount

platform (from left to right) restorations in the respectiveEG-EH (switching and regular platform connected to an

(switching and regular platform connected to an internal embedded in acrylic resin: (top, left) external and (top,e (bottom). (C) Mechanical testing set-up, where the load

d e n t a l m a t e r i a l s 2 8 ( 2

Tabl

e

1

–

Ch

arac

teri

stic

s

of

the

com

pon

ents

use

d

in

the

pre

sen

t

stu

dy.

Com

pon

ents

SWT-

EH

REG

-EH

SWT-

IH

REG

-IH

Imp

lan

tEx

tern

al

hex

(SU

R

5011

)

Exte

rnal

hex

(SU

R

5011

)

Inte

rnal

hex

(SIH

S

5511

)

Inte

rnal

hex

(SIH

S

5511

)5.

0

mm

dia

met

er

by

11.5

mm

len

gth

5.0

mm

dia

met

er

by

11.5

mm

len

gth

5.0

mm

dia

met

er

by

11.5

mm

len

gth

5.0

mm

dia

met

er

by

11.5

mm

len

gth

Ø

pro

sth

etic

pla

tfor

m

=

5.0

mm

Ø

pro

sth

etic

pla

tfor

m

=

5.0

mm

Ø

pro

sth

etic

pla

tfor

m

=

5.5

mm

Ø

pro

sth

etic

pla

tfor

m

=

5.5

mm

Abu

tmen

tC

emen

ted

(Al 4

151)

Ø

pla

tfor

m

=

4.1

mm

Cem

ente

d

(Al 5

051)

Ø

pla

tfor

m

=

5.0

mm

Cem

ente

d

(Al 4

501)

Ø

pla

tfor

m

=

4.5

mm

Cem

ente

d

(Al 5

501)

Ø

pla

tfor

m

=

5.5

mm

Ø

nec

k’s

regi

on

=

2.9

mm

Ø

nec

k’s

regi

on

=

2.9

mm

Scre

wfi

xati

on

scre

w

(PT

Q20

08)

fixa

tion

scre

w

(PT

Q20

08)

fixa

tion

scre

w

(PT

QH

16)

fixa

tion

scre

w

(PT

QH

16)

0 1 2 ) e218–e228 e221

of time without failure, 90% two-sided confidence bounds) oftested specimens for completion of a mission of 50,000 cyclesat 210 N and 300 N load [35] for group comparisons. For themission reliability and parameters calculated in the presentstudy, the 90% confidence interval range were calculated asfollows:

IC = E(G) ± Z˛sqrt(Var(G)) (1)

where CB is the confidence bound, E(G) is the mean estimatedreliability for the mission calculated from Weibull statistics,Z˛ is the z value concerning the given CB level of significance,and Var(G) is the value calculated by the Fisher Informationmatrix [33,36].

Macro images of failed samples were taken with a digi-tal camera (Nikon D-70s, Nikon, Tokyo, Japan) and utilized forfailure mode classification and comparisons between groups.In order to identify fractographic markings and characterizefailure origin and direction of crack propagation, the mostrepresentative failed samples of each group were inspectedfirst under a polarized-light microscope (MZ-APO stereomi-croscope, Carl Zeiss MicroImaging, Thornwood, NY, USA) andthen by scanning electron microscopy (SEM) (Model S-3500N,Hitachi, Osaka, Japan) [37,38].

2.2. Three-dimensional finite element analysis(3D-FEA)

Four virtual 3D models were created using computer-aideddesign (CAD) software (SolidWorks 2010, Dassault SystèmesSolidWorks Corp., Concord, MA, USA) following design anddimensions observed in groups SWT-EH, REG-EH, SWT-IH andREG-IH. Each 3D CAD model represented all characteristics ofthe implant-abutment connection in order to reproduce theexperimental conditions prevailing as a result of the mechan-ical tests (Fig. 1). The components of the models consisted ofa maxillary central incisor crown (Co–Cr alloy), a 50 �m-thick[39] resin cement layer (Rely X Unicem), an abutment (titaniumalloy), a fixation screw (titanium alloy), an implant (titaniumalloy), and a cylinder created in the CAD software with thesame dimensions of the plastic tubes used in the in vitro lab-oratory study (Fig. 3A). The anatomically correct crown wasgenerated from microcomputed tomography images in .dicomformat (�CT40, Scanco Medical AG, Bruttisellen, Switzerland)and its cementation surface was designed to fit the abutmentsin all groups. The implant insertion hole in the cylinder (acrylicresin) was obtained by a Boolean subtraction (Fig. 3B).

The components were assembled, imported into FEA soft-ware (Ansys Workbench 12.0, Swanson Analysis Inc., Houston,PA, USA), meshed (Fig. 3C) (number of parabolic tetrahedralelements [40] between 254,513 and 288,543; and number ofnodes between 433,816 and 492,803) and tested for conver-gence prior to mechanical simulation. It was considered thatthe convergence criterion between meshes refinement was achange of less than 6% in the maximum simulated von Misesequivalent stress (�vM) of the implant/abutment/screw com-

ponents [41].

The FEA model assumptions were that: (1) all solids werehomogeneous, isotropic and linearly elastic; (2) there wereno slip conditions (perfect bonding) among components (set

e222 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) e218–e228

Fig. 3 – (A) 3D CAD models of the implant-abutment complex including fixation screw, abutment and implant. (B) CompleteCAD model with a cement-retained crown over an implant which was embedded into the acrylic resin cylinder. The redarrow represents a 30◦ off-axis load (300 N) applied on the crown surface, and the blue arrows around the cylinder representthe fixation (full constraint) on the bottom and lateral surface of the cylinder of acrylic resin. (C) Finite element mesh of the

e m

model. On the right there is a higher magnification (2×) of th

implant-abutment-screw, elastic modulus (E) = 110 GPa andPoisson’s ratio (v) = 0.35) [42]; (3) there was a uniform cementlayer (E = 8 GPa, v = 0.33) [43]; (4) there was a crown (E = 220 GPa,v = 0.30) [44] with similar dimensions (13 mm height witha mesiodistal width of 8.8 mm and buccal-lingual width of7.1 mm) in all FEA models; (5) there were no flaws in anycomponents; (6) the boundary conditions of the model weredefined on the bottom and lateral surface of the cylinderof acrylic resin (E = 1.37 GPa, v = 0.30)1 to represent the con-strained of x, y and z directions (displacement = 0) (Fig. 3B).As in the mechanical tests, one 30◦ off-axis load (300 N) wasapplied on the lingual side of the crown, close to the incisaledge (Fig. 3B). Regions of higher von Mises equivalent stress(�vM) were determined within implant-abutment connectionfor all models.

3. Results

3.1. In vitro laboratory study (SLF and SSALT)

The SLF mean ± standard deviation values for groupSWT-EH was 1090.01 N ± 140.49 N, 1204.95 N ± 49.78 N forgroup REG-EH, 960.69 N ± 113.85 N for group SWT-IH and818.8 N ± 105.85 N for group REG-IH.

The step-stress accelerated fatigue allows estimation ofreliability at a given load level (Table 2). The calculated reli-ability with 90% confidence intervals for a mission of 50,000cycles at 300 N showed that the cumulative damage fromloads reaching 300 N would lead to restoration survival in 53%of specimens in group SWT-EH, whereas 93% would survive

in group REG-EH. These values depict a statistically signifi-cant difference between groups SWT-EH and REG-EH. On theother hand, the overlap between the upper and lower limits

1 Manufacturer’s information.

esh showed in the boxed area.

of reliability values in groups SWT-IH and REG-IH indicatesno statistically significant difference in reliability of implant-supported restorations with internal connections, regardlessof abutment diameter (switching or regular platform). For thegiven mission, a survival of 99% of the specimens would beobserved in both groups (SWT-IH and REG-IH). As shown inTable 2, from 99% to 100% of the specimens would survivegiven a mission of 50,000 cycles at 210 N, indicating no statis-tically significant difference in reliability among all groups.

The step-stress derived probability Weibull plots at a 300 Nload are presented in Fig. 4. The Beta (ˇ) values and associatedupper and lower bounds derived from use level probabil-ity Weibull calculation (probability of failure vs. number ofcycles) of 1.31 (0.75–2.28) and 1.83 (1.01–3.32) for platform-switched implants (groups SWT-EH and SWT-IH, respectively),and values of 1.55 (0.78–3.06) and 1.82 (1.02–3.25) for regularplatform implants (groups REG-EH and REG-IH, respectively)indicated that fatigue was an accelerating factor for all groups.The Beta value describes failure rate changes over time ( < 1:Failure rate is decreasing over time, commonly associated with“early failures” or failures that occur due to egregious flaws;

∼ 1: failure rate that does not vary over time, associated withfailures of a random nature; > 1: Failure rate is increasingover time, associated with failures related to damage accu-mulation) [30,45,46].

3.2. Failure modes

All specimens failed after SLF and SSALT. Failure modes for allgroups are presented in Table 3. For restorations over externalhex implants (groups SWT-EH and REG-EH) screw fracture atthe third thread region was the chief failure mode (Fig. 5C). In

these specimens, abutments and implants were intact aftermechanical tests. For restorations over internal hex implants(groups SWT-IH and REG-IH), screw and abutment fracture atthe narrowest region (cervical collar of the abutment, diameter

d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) e218–e228 e223

Table 2 – Calculated reliability (upper and lower limits) for tested groups given two different missions: 50,000 cycles at300 N load and 50,000 cycles at 210 N load.

Output SWT-EH REG-EH SWT-IH REG-IH

50,000 cycles @ 300 N 0.53 (0.33–0.70)a 0.93 (0.80–0.97)b 0.99 (0.93–0.99)c 0.99 (0.99–1.00)c

50,000 cycles @ 210 N 0.99 (0.94–0.99)c 0.99 (0.98–0.99)c 1.00 (0.99–1.00)c 1.00 (0.99–1.00)c

The superscript letters (a, b and c) depicts statistically homogeneous groups.

Fig. 4 – This graph shows the probability of failure as a function of number of cycles (time) for tested groups simulating amission of 50,000 cycles at 300 N. Note the left position of the SWT-EH group (green) relative to REG-EH group (blue), andSWT-IH group (pink) relative to REG-IH group (black), which indicates the need for more cycles to failure in regular-platformg

omg

tcfc

roups compared to the switched-platform groups.

f 2.9 mm) (Fig. 6B and C) were observed in all specimens afterechanical tests. No implant fracture was observed in any

roup.Observation of the polarized-light and SEM micrographs of

he screw’s fractured surface allowed the consistent identifi-

ation of fractographic markings, such as compression curl,atigue striations and dimples, which allowed the identifi-ation of flaw origin and the direction of crack propagation

Table 3 – Failure modes after mechanical testing (single-load-to(SSALT)) according to the used failure criteria.

Groups SWT-EH REG-EH

SLF(n = 3)

Screw: 3 fracture Screw: 3 fracture

Abutment: 3 intact Abutment: 3 intact

Implant: 3 intact Implant: 3 intact

SSALT(n = 18)

Screw: 18 fracture Screw: 18 fracture

Abutment: 18 intact Abutment: 18 intact

Implant: 18 intact Implant: 18 intact

(Fig. 7). As per our imaging analysis of the specimen’s frac-tured surface, all fractures were characterized by materialtearing and exhibited gross plastic deformation, suggestingductile fractures (Figs. 6C and 7A and B). The resulting duc-tile fractures occurred as stresses exceeded the material yield

strength leaving telltale fractographic marks that indicatedcrack propagation from lingual to buccal (Fig. 7C), whereocclusal forces naturally occur in the anterior region. Although

-fracture (SLF) and step-stress accelerated life-testing

SWT-IH REG-IH

Screw: 3 fracture Screw: 3 fractureAbutment: 3 fracture Abutment: 3 fractureImplant: 3 intact Implant: 3 intact

Screw: 18 fracture Screw: 18 fractureAbutment: 18 fracture Abutment: 18 fractureImplant: 18 intact Implant: 18 intact

e224 d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) e218–e228

Fig. 5 – Images illustrating the peak of stress for the fixation screw in all groups. (A) 3D CAD model with abutment intransparency showing the contact area (black arrow) at the third thread region of the screw. (B) Peak of von Mises equivalentstress (�vM) at the third thread region of the screw. (C) Macro picture of the screw fractured at the third thread region.

Fig. 6 – Images illustrating the peak of stress for the abutments. (A) Peak of von Mises equivalent stress (�vM) located at theexternal region (lingual side) of the cervical collar of the abutment. (B) Macro picture of an abutment fractured at thenarrower region of the abutment (cervical collar region). In all specimens, the fracture occurred in this region. (C) SEMmicrograph of the region of fracture.

d e n t a l m a t e r i a l s 2 8 ( 2 0 1 2 ) e218–e228 e225

Fig. 7 – Representative fractured screw after SSALT depicting: (A and B) Macro image and SEM micrograph, respectively,showing a fracture occurring at the third thread region viewed from the screw’s long axis. (C) is a SEM micrograph (60×) ofthe fractured surface of sample shown in (B). The white dotted circle shows a compression curl which evidences fractureorigin at the opposing tensile side (white box), indicating the direction of crack propagation (dcp) (white arrow). (D) is ahigher magnification (250×) of the boxed area in (C) showing the fracture origin. (E and F) are higher magnifications (2000×and 1500×, respectively) of the fractured surface showing typical fractographic features of metallic materials: (E) fatigues

aipfatmoa[awa

3

T(Tmdcto(iI

ditions. Previous studies [8–11] have demonstrated thatplatform-switched abutments may not only reduce the earlyperi-implant bone loss and increase the biomechanical

Table 4 – von Mises equivalent stress (�vM) in MPawithin the implant-abutment connection.

Component Implant Abutment Screw

triations and (F) dimpled surface appearance.

part may fail in a brittle manner, ductile fracture morphologys frequently observed away from the origin. For example, com-ression curl is a fractographic feature representative of flexureailures and results from a traveling crack changing directions it enters a compression field [47]. Usually it evidences frac-ure origin at the opposing tensile side (Fig. 7D). At higher

agnifications (from 500× to 2500×), fatigue striations werebserved (Fig. 7E). They emanated outward from the originnd marked successive positions of the advancing crack front37]. Also in a higher magnification (1500×) a dimpled surfaceppearance created in some areas on the fractured surfaceas observed, exemplifying a typical ductile fracture in metal

lloys, commonly created by microvoid coalescence [37].

.3. 3D-FEA

he values for �vM within implant-abutment compleximplant, abutment and fixation screw) are presented inable 4, and showed that the stress distribution on abut-ent and screw was strongly influenced by the abutment

iameter (regular and switched-platform) and type of implantonnection (external and internal hexagon). When reducinghe abutment diameter, an increase in the �vM of 41.08% was

bserved in the abutment connected to external hex implant

SWT-EH), while an increase in the �vM of 53.27% was observedn the abutment connected to internal hex implant (SWT-IH).n the fixation screw, increases of 19.67% and 11.57% were

observed in the �vM for SWT-EH and SWT-IH, respectively. Norelevant differences in the levels of �vM were observed in theimplant body when considering the variables of this study.

The highest level of stress was observed in the fixationscrew for all models. In the fixation screw, the peak of �vM wasconcentrated at the third thread region in all groups (Fig. 5),whereas in the abutment the peak of �vM was located on thelingual region at the cervical collar (Fig. 6A).

4. Discussion

The concept of platform switching is increasingly soughtbecause it can be advantageous in several clinical con-

SWT-EH 228 182 365REG-EH 225 129 305SWT-IH 216 166 270REG-IH 216 108.3 242

s 2 8

e226 d e n t a l m a t e r i a l

support available to the implant, but also may improve esthet-ics. Baumgarten and coworkers [6] suggested that the platformswitching technique is useful when shorter implants are used,when implants are placed in esthetic zone, and when a largerimplant is desirable but prosthetic space is limited. How-ever, as per our laboratory and mechanical simulation results,such attempt to minimize the bone remodeling and resorptionand simultaneously improve esthetics by reducing abutmentdiameter may result in higher stress concentration in theconnection components. Our simulation findings are in agree-ment with other recent FEA studies [14,17].

Considering the relevance of a fatigue resistant implant-abutment connection for the long-term clinical success, thepresent study evaluated the effect of platform switching con-cept for external and internal hex implants in the reliabilityof maxillary central incisor crowns using SSALT. This methodconsists on a mechanical test for shortening the life of materi-als or hastening the degradation of their performance. Unlikeother methods, the aim of such testing is to quickly obtaindata which, properly modeled and analyzed, yield desiredinformation on component life or performance under nor-mal use. In addition, the SSALT method allows the predictionwith confidence intervals (based on calculation of a masterWeibull distribution) of the life expectancy of a given materialunder specified loading. We have used a life-stress relation-ship model allowing the extrapolation of a use level probabilitydensity function from life data obtained at increased stresslevels. These models describe the path of a particular lifecharacteristic of the distribution from one stress level toanother. For the Weibull distribution, the scale parameter (eta)is considered to be stress-dependent. Therefore, the life-stressmodel for data that fits the Weibull distribution is assignedto eta. Our results showed that fatigue damage accumulationaccelerated the failures of all tested designs in the presentstudy, as evidenced by the resulting > 1 (also called theWeibull shape factor). Furthermore, a statistically significantlower reliability (given a mission of 50,000 cycles at 300 N load)was found for platform-switched implants with external hex(SWT-EH), but not for platform-switched implants with inter-nal hex (SWT-IH).

These findings may be explained based in the associa-tion among stress distribution and system’s reliability aroundthe weakest component of the implant-abutment connection:The fixation screw. The higher levels of stress (�vM) in theabutment screw observed for the external hexagon connec-tion was associated with a lower reliability after mechanicaltesting for both regular and switched-platform systems (300 Nload simulation). However, it can be assumed that the slightincrease (11.57%) in stress levels (�vM) observed in the fixa-tion screw when reducing abutment diameter over an internalhex implant (SWT-IH) was not significant to result in lowermechanical reliability. The lower values for reliability observedin groups SWT-EH and REG-EH were due to lower loads initi-ating prosthetic component failure when compared to groupswith internal hex implants (SWT-IH and REG-IH).

Worth noting is that all previous considerations were per-

formed under mission of 50,000 cycles at 300 N load. If amission of 50,000 cycles at 210 N load is considered (meanvalue for incisal bite force) [35], the cumulative damage fromloads reaching 210 N would lead to restoration survival of

( 2 0 1 2 ) e218–e228

99–100% of the specimens after 50,000 cycles. Thus, under nor-mal occlusion conditions, almost all tested specimens wouldpresent satisfactory fatigue endurance in the wet environmentused in the present study. In an attempt to simulate the oralenvironment, fatigue in water was performed, which has beensuggested as an important service-related cause of failure inmetals [37].

Screw fracture at the third thread region was the chief fail-ure mode in all groups. Those resulted as the upper part ofthe implant’s connection in contact with the fixation screwunder off-axis loading causes the presence of a lever aroundthe third thread region of the screw (Fig. 5). Moreover, fractureof the abutments were also observed in groups with internalhex implants. Those fractures were always located on the lin-gual side at the narrowest region below the cervical collar.The narrowest part of a component is usually its weakest partbecause it is the region where the maximum stresses occur,because of the smallest cross-sectional area. In the presentstudy the peak of �vM was located at the external region of thecervical collar (Fig. 6) because a perfect bonding was consid-ered between abutment and implant. In our FEA simulationthere was no separation of these components when submit-ted to tensile forces and higher tensile stresses were generatedat this region (external area of the cervical collar). Those hightensile stresses are not real, given that in the physical testing(SSALT) the abutments moves away from the implant plat-form (at palatal region), but does not pull the implant. Futuresimulations with more complex models capable to addresssuch limitation are warranted. Moreover, it has been previ-ously reported that the failure location is related to the designcharacteristics of the implant-abutment combination, whichis commonly located in the threaded region or areas that rep-resent a critical point for prosthetic component’s endurancedue to the shift in geometry along its length and subtle alter-ation in cross-sectional area [23].

Despite the stress distribution observed in the 3D-FEAbeing obtained from single static loading, such as in SLFtests, which does not represent the cyclic loading observedin oral environment and in fatigue tests (SSALT), our resultssuggest improved stress distribution within the implant-abutment connection of regular-platform models regardlessof the methodology (in vitro study or finite element analy-sis). Thus, the improved stress distribution may presumablybe the reason for better mechanical behavior of internallyconnected systems compared to the externally connectedcounterparts. Concerning the geometry of implant connec-tion (internal vs. external), higher reliability was observed inspecimens with internal connection regardless of the abut-ment diameter. These findings are in agreement with otherstudies that pointed that deep joints show increased stabil-ity favoring structural strength of implant systems [24,32,48].It should be noted, however, that due to engineering designconstraints such as minimum wall thickness for propermechanical performance of each of the different connectionsystems, differences in both external and internal features ofthe implant, abutment, and screw designs will exist. While

from a research standpoint it is highly desirable that only theconnection is changed with the connecting screw and implantremaining the same, such interplay is unfeasible for whenone is attempting to make clinically relevant comparisons

8 ( 2

(smetf

tcmamwiai

5

TwiraatfFr

A

T1M(

r

d e n t a l m a t e r i a l s 2

implants presenting the same diameter, length, and crownize) between external and internal connections in most com-ercially available systems, as alterations in the implant

xternal shape is usually performed by manufacturers in ordero maintain tolerances for appropriate fit and wall thicknessor the internal connection robustness.

According to the literature [7], there are potential limita-ions for using platform-switched implants, e.g. the need foromponents that have similar designs (the screw access holeust be uniform) and the need for enough space to develop

proper emergence profile. Considering that the replace-ent of single-unit edentulous spaces in the anterior regionith implant-supported restorations is a challenging scenario

n terms of long-term success and esthetics, it is crucial tocknowledge the functional and mechanical limitations of themplant-abutment connections.

. Conclusions

he postulated hypothesis that platform-switched implantsould result in increased stress concentration within the

mplant-abutment connection, leading to the systems’ lowereliability on laboratory mechanical testing was partiallyccepted. The higher levels of stress observed within implant-butment connection when reducing abutment diameter, andherefore its cross-sectional area, resulted in lower reliabilityor external hex implants, but not for internal hex implants.ailure modes were similar when comparing switching andegular platforms.

cknowledgements

his investigation was supported in part by Research Grant41870/2008–7 from CNPq – Brazil. The authors are thankful toarotta Dental Studio (Farmingdale, NY, USA) and SIN implants

São Paulo, SP, Brazil) for their support.

e f e r e n c e s

[1] Bozkaya D, Muftu S, Muftu A. Evaluation of load transfercharacteristics of five different implants in compact bone atdifferent load levels by finite elements analysis. Journal ofProsthetic Dentistry 2004;92:523–30.

[2] Hermann F, Lerner H, Palti A. Factors influencing thepreservation of the periimplant marginal bone. ImplantDentistry 2007;16:165–75.

[3] Manz MC. Factors associated with radiographic vertical boneloss around implants placed in a clinical study. Annals ofPeriodontology 2000;5:137–51.

[4] Lopez-Mari L, Calvo-Guirado JL, Martin-Castellote B,Gomez-Moreno G, Lopez-Mari M. Implant platformswitching concept: an updated review. Medicina Oral,Patologia Oral Y Cirugia Bucal 2009;14:e450–4.

[5] Lazzara RJ, Porter SS. Platform switching: a new concept inimplant dentistry for controlling postrestorative crestalbone levels. International Journal of Periodontics and

Restorative Dentistry 2006;26:9–17.

[6] Baumgarten H, Cocchetto R, Testori T, Meltzer A, Porter S. Anew implant design for crestal bone preservation: initial

0 1 2 ) e218–e228 e227

observations and case report. Practical Procedures &Aesthetic Dentistry:PPAD 2005;17:735–40.

[7] Gardner DM. Platform switching as a means to achievingimplant esthetics. New York State Dental Journal2005;71:34–7.

[8] Cappiello M, Luongo R, Di Iorio D, Bugea C, Cocchetto R,Celletti R. Evaluation of peri-implant bone loss aroundplatform-switched implants. International Journal ofPeriodontics and Restorative Dentistry 2008;28:347–55.

[9] Luongo R, Traini T, Guidone PC, Bianco G, Cocchetto R,Celletti R. Hard and soft tissue responses to theplatform-switching technique. International Journal ofPeriodontics and Restorative Dentistry 2008;28:551–7.

[10] Duarte AR, Rossetti PH, Rossetti LM, Torres SA, BonachelaWC. In vitro sealing ability of two materials at five differentimplant-abutment surfaces. Journal of Periodontology2006;77:1828–32.

[11] Canullo L, Fedele GR, Iannello G, Jepsen S. Platformswitching and marginal bone-level alterations: the results ofa randomized-controlled trial. Clinical Oral ImplantsResearch 2010;21:115–21.

[12] Schrotenboer J, Tsao YP, Kinariwala V, Wang HL. Effect ofmicrothreads and platform switching on crestal bone stresslevels: a finite element analysis. Journal of Periodontology2008;79:2166–72.

[13] Baggi L, Cappelloni I, Di Girolamo M, Maceri F, Vairo G. Theinfluence of implant diameter and length on stressdistribution of osseointegrated implants related to crestalbone geometry: a three-dimensional finite elementanalysis. Journal of Prosthetic Dentistry 2008;100:422–31.

[14] Tabata LF, Assuncao WG, Adelino Ricardo Barao V, de SousaEA, Gomes EA, Delben JA. Implant platform switching:biomechanical approach using two-dimensional finiteelement analysis. Journal of Craniofacial Surgery2010;21:182–7.

[15] Vargas LC, Almeida EO, Rocha E, Freitas Jr AC, Anchieta RB,Kina S, et al. Regular and platform switching. Bone stressanalysis with varying implant diameters. Journal of OralImplantology 2011.

[16] Rodriguez-Ciurana X, Vela-Nebot X, Segala-Torres M,Rodado-Alonso C, Mendez-Blanco V, Mata-Bugueroles M.Biomechanical repercussions of bone resorption related tobiologic width: a finite element analysis of threeimplant-abutment configurations. International Journal ofPeriodontics and Restorative Dentistry 2009;29:479–87.

[17] Maeda Y, Miura J, Taki I, Sogo M. Biomechanical analysis onplatform switching: is there any biomechanical rationale?Clinical Oral Implants Research 2007;18:581–4.

[18] Jung RE, Pjetursson BE, Glauser R, Zembic A, Zwahlen M,Lang NP. A systematic review of the 5-year survival andcomplication rates of implant-supported single crowns.Clinical Oral Implants Research 2008;19:119–30.

[19] Becker J, Ferrari D, Herten M, Kirsch A, Schaer A, Schwarz F.Influence of platform switching on crestal bone changes atnon-submerged titanium implants: a histomorphometricalstudy in dogs. Journal of Clinical Periodontology2007;34:1089–96.

[20] Cochran DL, Bosshardt DD, Grize L, Higginbottom FL, JonesAA, Jung RE, et al. Bone response to loaded implants withnon-matching implant-abutment diameters in thecanine mandible. Journal of Periodontology 2009;80:609–17.

[21] Hurzeler M, Fickl S, Zuhr O, Wachtel HC. Peri-implant bone

level around implants with platform-switched abutments:preliminary data from a prospective study. Journal of Oraland Maxillofacial Surgery 2007;65:33–9.

s 2 8

e228 d e n t a l m a t e r i a l

[22] Degidi M, Iezzi G, Scarano A, Piattelli A. Immediately loadedtitanium implant with a tissue-stabilizing/maintainingdesign (’beyond platform switch’) retrieved from man after 4weeks: a histological and histomorphometrical evaluation. Acase report. Clinical Oral Implants Research 2008;19:276–82.

[23] Quek HC, Tan KB, Nicholls JI. Load fatigue performance offour implant-abutment interface designs: effect of torquelevel and implant system. International Journal of Oral andMaxillofacial Implants 2008;23:253–62.

[24] Steinebrunner L, Wolfart S, Ludwig K, Kern M.Implant-abutment interface design affects fatigue andfracture strength of implants. Clinical Oral ImplantsResearch 2008;19:1276–84.

[25] Hsu JT, Fuh LJ, Lin DJ, Shen YW, Huang HL. Bone strain andinterfacial sliding analyses of platform switching andimplant diameter on an immediately loaded implant:experimental and three-dimensional finite elementanalyses. Journal of Periodontology 2009;80:1125–32.

[26] Chang CL, Chen CS, Hsu ML. Biomechanical effect ofplatform switching in implant dentistry: athree-dimensional finite element analysis. InternationalJournal of Oral and Maxillofacial Implants 2010;25:295–304.

[27] Crespi R, Cappare P, Gherlone E. Radiographic evaluation ofmarginal bone levels around platform-switched andnon-platform-switched implants used in an immediateloading protocol. International Journal of Oral andMaxillofacial Implants 2009;24:920–6.

[28] Pessoa RS, Vaz LG, Marcantonio Jr E, Vander Sloten J, Duyck J,Jaecques SV. Biomechanical evaluation of platform switchingin different implant protocols: computed tomography-basedthree-dimensional finite element analysis. InternationalJournal of Oral and Maxillofacial Implants 2010;25:911–9.

[29] Canay S, Akca K. Biomechanical aspects of bone-leveldiameter shifting at implant-abutment interface. ImplantDentistry 2009;18:239–48.

[30] Coelho PG, Silva NR, Bonfante EA, Guess PC, Rekow ED,Thompson VP. Fatigue testing of two porcelain-zirconiaall-ceramic crown systems. Dental Materials 2009;25:1122–7.

[31] Pjetursson BE, Tan K, Lang NP, Bragger U, Egger M, ZwahlenM. A systematic review of the survival and complicationrates of fixed partial dentures (FPDs) after an observationperiod of at least 5 years. Clinical Oral Implants Research2004;15:625–42.

[32] Khraisat A, Stegaroiu R, Nomura S, Miyakawa O. Fatigueresistance of two implant/abutment joint designs. Journal ofProsthetic Dentistry 2002;88:604–10.

[33] Nelson W. Accelerated testing: statistical models, test plansand data analysis. New York: John Wiley & Sons; 2004.

[34] Zhao WEE. A general accelerated life model for step-stresstesting. IEEE Transactions On Reliability 2005;37:1059–69.

( 2 0 1 2 ) e218–e228

[35] Paphangkorakit J, Osborn JW. The effect of pressure on amaximum incisal bite force in man. Archives of Oral Biology1997;42:11–7.

[36] Abernethy R. The New Weibull Handbook. 5th ed. NorthPalm Beach: Dr. Robert B. Abernethy; 2006, 350 p.

[37] Parrington RJ. Fractography of metals and plastics. Practicalfailure analysis. Materials Park: ASM International; 2002,16-46.

[38] Manda MG, Psyllaki PP, Tsipas DN, Koidis PT. Observationson an in vivo failure of a titanium dental implant/abutmentscrew system: a case report. Journal of Biomedical MaterialsResearch Part B Applied Biomaterials 2009;89:264–73.

[39] Li LL, Wang ZY, Bai ZC, Mao Y, Gao B, Xin HT, et al.Three-dimensional finite element analysis of weakenedroots restored with different cements in combination withtitanium alloy posts. Chinese Medical Journal2006;119:305–11.

[40] de Almeida EO, Rocha EP, Freitas Jr AC, Freitas Jr MM. Finiteelement stress analysis of edentulous mandibles withdifferent bone types supporting multiple-implantsuperstructures. International Journal of Oral andMaxillofacial Implants 2010;25:1108–14.

[41] Huang HL, Hsu JT, Fuh LJ, Tu MG, Ko CC, Shen YW. Bonestress and interfacial sliding analysis of implant designs onan immediately loaded maxillary implant: a non-linearfinite element study. Journal of Dentistry 2008;36:409–17.

[42] Huang HL, Fuh LJ, Ko CC, Hsu JT, Chen CC. Biomechanicaleffects of a maxillary implant in the augmented sinus: athree-dimensional finite element analysis. InternationalJournal of Oral and Maxillofacial Implants 2009;24:455–62.

[43] Coelho PG, Bonfante EA, Silva NR, Rekow ED, Thompson VP.Laboratory simulation of Y-TZP all-ceramic crown clinicalfailures. Journal of Dental Research 2009;88:382–6.

[44] Erkmen E, Meric G, Kurt A, Tunc Y, Eser A. Biomechanicalcomparison of implant retained fixed partial dentures withfiber reinforced composite versus conventional metalframeworks: a 3D FEA study. Journal of the MechanicalBehavior of Biomedical Materials 2011;4:107–16.

[45] ReliaSoft. The Weibull Distribution and Beta. 2010.[46] Silva NR, de Souza GM, Coelho PG, Stappert CF, Clark EA,

Rekow ED, et al. Effect of water storage time and compositecement thickness on fatigue of a glass-ceramic trilayersystem. Journal of Biomedical Materials Research Part B,Applied Biomaterials 2008;84:117–23.

[47] Quinn GD. Fractography of ceramics and glasses.Washington: U.S. Government Printing Office; 2007.

[48] Maeda Y, Satoh T, Sogo M. In vitro differences of stressconcentrations for internal and external heximplant-abutment connections: a short communication.Journal of Oral Rehabilitation 2006;33:75–8.