Implant-Abutment Connection Designs for Anterior Crowns...

Implant-Abutment Connection Designsfor Anterior Crowns: Reliability and

Failure ModesLucas S. Machado, DDS, MS,* Estevam A. Bonfante, DDS, MS, PhD,† Rodolfo B. Anchieta, DDS, MS,*

Satoshi Yamaguchi, PhD,‡ and Paulo G. Coelho, DDS, BS, MS, MSMtE, PhD§

The success of dental implant ther-apy should be evaluated fromboth esthetic and mechanical per-

spectives.1 Considering that both es-thetics and biomechanics are affectedby the achievement and maintenanceof osseointegration,2 engineering param-eters of the implant, such as macroge-ometry and type of implant-abutmentconnection, are under constant develop-ment.1,3–10

The external hexagon (EH) systemwas initially designed to provide a rota-tional torque transferring mechanismsuited for surgical placement of theimplant into the osteotomy. Althoughproviding ways for prosthesis indexingand antirotational capabilities, the exten-sion of use of EH from full-mouthsplinted reconstructions to single-unitprosthesis leads to concerns regardingthe mechanical challenges imposed byloads mainly borne by the abutment

screw.11 Internal connections such asinternal hexagon (IH) and Morse taper(MT) were further developed so thatloads could be dissipated along theimplant walls in contact with the abut-ment surface, also providing a shield forthe abutment screw.12 MT includes aninterference fit connection type in a coni-cal geometry of the abutment and the cor-responding surface of the implant well.13

The choice of implant-abutmentconnection design (external or internal)and prostheses fixation mode (ce-mented vs screwed) may affect eitherbiological or technical complicationrates, especially in single crowns,

which are not splinted and subjectedto multidirectional loading that chal-lenges the connection components andrestoration structural integrity.1,14–16 Ithas been reported that the single-unitimplant-supported restoration repre-sents the scenario where the incidenceof screw loosening or fracture is thehighest, especially in external connec-tions.4,17,18 In addition, the design of theimplant-abutment connection may alsoinfluence the biological response ofhard and soft tissues.18,19

In an attempt to reduce prostheticand biological complications, differentimplant-abutment connection designs

*PhD Candidate, Department of Biomaterials and Biomimetics,College of Dentistry, New York University, New York, NY.†Assistant Professor, Department of Prosthodontics, Universityof Sao Paulo - Bauru College of Dentistry, Bauru, SP, Brazil.‡Postdoctoral Researcher, Department of Biomaterials andBiomimetics, College of Dentistry, New York University, NewYork, NY; Assistant Professor, Department of BiomaterialsScience, Osaka University Graduate School of Dentistry, Osaka,Japan.§Assistant Professor, Department of Biomaterials andBiomimetics, College of Dentistry, New York University, NewYork, NY; Director for Research, Department of Periodontologyand Implant Dentistry, New York University College of Dentistry,New York, NY.

Reprint requests and correspondence to: Estevam A.Bonfante, DDS, MS, PhD, Department ofProsthodontics, University of Sao Paulo - Bauru Collegeof Dentistry. Al. Octávio Pinheiro Brisola, 9–75, Bauru,SP 17012–901, Brazil, Phone: 55(14) 8153-0860, Fax:55(14) 3234-2566, E-mail: [email protected]

ISSN 1056-6163/13/00000-001Implant DentistryVolume 0 � Number 0Copyright © 2013 by Lippincott Williams & Wilkins

DOI: 10.1097/ID.0b013e31829f1f2d

Objectives: To investigate theeffect of implant-abutment connec-tion types on reliability and failuremodes of anterior single-unitcrowns.

Methods: Fifty-four implantswere divided in 3 groups (n ¼ 18each): external hexagon (EH), inter-nal hexagon (IH), and Morse taper(MT) connection. Abutments werescrewed to the implants, and maxil-lary central incisor metal crownswere cemented and subjected tostep-stress accelerated life testing.

Results: The beta values derivedfrom use-level probability Weibullcalculations for groups IH (2.52),EH (1.67), and MT (0.88) indicatedthat fatigue influenced the failureonly of IH and EH groups. The

reliability for a mission of 100,000cycles at 175 N was 0.99 (0.98–1.00), 0.84 (0.62–0.94) and 0.97(0.87–0.99) for the EH, IH, andMT, respectively. The characteristicstrength was not significantly differ-ent between EH (290 N) and IH (251N) but significantly higher for MT(357 N). For IH and EH groups, fail-ure involved screw fracture, and theMT implants primary failure modewas abutment fracture.

Conclusions: Reliability washigher for the EH and MT relative toIH groups, whereas the characteristicstrength was significantly higherfor implants with MT connection.(Implant Dent 2013;0:1–6)Key Words: dental implants, reli-ability, Weibull, fractography

IMPLANT DENTISTRY / VOLUME 0, NUMBER 0 2013 1

Copyright � Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

17

pressure), cleaned with ethanol, driedwith air free of water and oil, and thencemented (Rely X Unicem; 3M ESPE,St Paul, MN). The final dimensions forEH, IH, and IC were the same.

Mechanical Testing andReliability Analysis

To test the implant-abutment con-nection in a challenging scenario,mechanical testing was undertaken withall specimens placed at a 30-degree axialinclination, as per ISO 14801:2007,where the indenter contacted the crownsurface and applied the prescribed loadwithin the step profile. The intent was toprovide a bending component duringloading, which may occur during inter-cuspal position but more commonlypresent in protrusive or laterotrusive

mandibularmovements, namelyanteriorguidance.29 Based on the mean load tofailure of previous studies, SSALT pro-files were determined.1,18,27,28 Thisfatigue testing approach consists of test-ing the samples at stress levels higherthan used stress to facilitate failures ina timely manner. The results of thesetests are then analyzed so that a profileof the failure behavior of the specimensat used stresses can be determined basedon the behavior of the samples at theaccelerated stresses.26

The profiles were designated mild,moderate, and aggressive, with thenumber of specimens assigned to eachgroup in the ratio 3:2:1, respectively.Therefore, of the 18 samples per group,9 were allocated in the mild, 6 in themoderate, and 3 in the aggressive

profiles. Mild, moderate, or aggressiveprofile refer to the increasingly step-wise rapidness in which a specimen isfatigued to reach a certain level of load,meaning that specimens assigned toa mild profile will be cycled longer toreach the same load of a specimenassigned to either moderate or aggres-sive profile.26 The rationale for using atleast 3 profiles for this typeof testingwasbased on the need to distribute failuresacross different step loads, allowing bet-ter prediction statistics, narrowing confi-dence bounds. The prescribed fatiguemethod was SSALT under water at 9Hzwith a servo-all-electric system (Tes-tResources 800L, Shakopee, MN),where the indenter contacted the crownsurface and applied the prescribed loadwithin the step profile.

The specimens were evaluated atthe completion of each fatigue step-stress cycle or final failure (determinedby setting the machine to stop whena compression lower limit of approxi-mately 0.5 mm was reached). Criteriaused for failure were bending or fractureof the fixation screw, partial fracture ortotal fracture of the abutment, andfracture of an implant.

Based on the step-stress distributionof the failures, use-level probabilityWeibull curves (probability of failurevs cycles) with use stress of 175 N and90% 2-sided confidence intervals werecalculated and plotted (Alta Pro 7;ReliaSoft, Tucson, AZ), using a powerlaw relationship for damage accumula-tion. Reliability for amission of 100,000cycles at 175N (90%2-sided confidenceinterval) was calculated for comparisonbetween the groups. The Weibull mod-ulus 2-sided 90% confidence intervalswere calculated using the Fisher matrixmethod.26 For the parameters calculatedin this study, the 90% confidence inter-val range were calculated as follows:

IC 5 E ðGÞ 6 Za sqrtðVar ðGÞÞ ð1Þ

where IC is the confidence bound, E(G) isthe mean estimated reliability for the mis-sion calculated fromWeibull statistics, za isthe z value concerning the given IC level ofsignificance, andVar (G) is the value calcu-lated by the Fisher information matrix.30,31

Fig. 2. Contour plot (Weibull modulus vs characteristic strength) for group comparisons. Notethe overlap between IH and EH.

Table 3. Failure Modes After Mechanical Testing (SSALT) According to the UsedFailure Criteria

EH (n ¼ 18) IH (n ¼ 18) MT (n ¼ 18)

Implant 10 ¼ fracture (hexagon) 18 ¼ intact 18 ¼ intact6 ¼ fracture (body)

2 ¼ intactAbutment 18 ¼ intact 18 ¼ intact 7 ¼ fracture

7 ¼ bending4 ¼ intact

Abutment screw 12 ¼ fracture 18 ¼ fracture 4 ¼ fracture6 ¼ intact 14 ¼ intact



have been developed and are availablefor use. However, the literature con-cerning themechanical behavior of EH,IH, and MT implant-abutment connec-tions is still sparse and may be contra-dictory.20 Most studies concerning themechanical behavior of implant con-nections have been limited to staticnumerical simulations,14,21,22 and onlya fewhave considered the role of fatiguein the mechanisms of failure.14,21,22

Finite element analysis, for instance,has shown improved stability for theabutment and the lowest stress con-centration in the abutment screw forMT connections23 relative to EH andIH. Conversely, a recent study com-paring the EH, IH, and MT implantsshowed that EH presented signifi-cantly higher fatigue resistance thanIH and MT.20

The main challenge in the devel-opment of implant-abutment connec-tion designs relies on reducing/eliminating the incidence of mechani-cal failures in the implant-prostheticdevices and improving the response ofbone and soft tissues.18,24,25 Thus, theevaluation of reliability and failuremodes could provide insight intothe mechanical behavior of differentconfigurations of implant-abutmentconnections.1,18,26–28

This study aims to evaluate thereliability and failure modes of maxil-lary central incisor metal crowns onEH, IH, and MT implant-abutmentconnection configurations when sub-jected to step-stress accelerated lifetesting (SSALT). The null hypothesiswas that there would be no differencein reliability or failure modes betweenthe different implant-abutment con-nections of implants from the samemanufacturer when subjected toSSALT.

MATERIALS AND METHODS

Experimental DesignFifty-four implants of 4.0mmdiam-

eter and 10 mm length with the sameexternal macrogeometry were used forthis study (Implacil de Bortoli, Ltda.,Sao Paulo, Brazil). They were dividedaccording to the implant-abutment con-nection design as follows: EH, IH, andMT (Table 1).

Sample PreparationImplantswithEH, IHandMTdesign

were secured to their respective prefabri-cated abutments (Implacil de Bortoli,Ltda) and torqued according to the man-ufacturer’s instructions (25N$cm). Then,they were vertically embedded in poly-methyl-methacrylate resin (OrthodonticResin; Dentsply Caulk, Philadelphia,PA) leaving 1 mm of the implant-abut-ment finishing line exposed above thepotting surface.

All groups were restored with stan-dardized central incisormetallic crownscast in Co-Cr alloy (Wirobond 280;BEGO, Bremen, Germany). To repro-duce the anatomy of the crowns (n¼ 54total), an impressionwas taken from thefirst waxed pattern and used by the tech-nician as a guide during waxing of theremaining crowns. Before cementation,crowns were sandblasted with alumi-num oxide (particle size # 40 mm,using 276 KPa compressed air

Table 1. Characteristics of the Groups Used in This Study

Groups EH (n ¼ 18) IH (n ¼ 18) MT (n ¼ 18)

Design of implant-abutmentconnection (Implacilde Bortoli, Ltda)

Conical implantwith externalhexagonconnection

Conical implantwith internalhexagonconnection

Conical implantwith internalconicalconnection

Fig. 1. Use-level probability Weibull for tested groups showing the probability of failure asa function of number of cycles given a mission of 100,000 cycles at 175 N.

Table 2. Calculated Reliability for a Mission of 100,000 Cycles at 175 N Load

EH IH MT

Mission of 100,000 cycles at 400 NUpper 1.00 0.94 0.99Mean 0.99 0.84 0.97Lower 0.98 0.62 0.87

BetaUpper 2.56 3.91 1.68Mean 1.67a 2.52a 0.88b

Lower 1.08 1.62 0.46

Equal letters (a and b) represent statistically homogeneous groups.

2 IMPLANT-ABUTMENT CONNECTION DESIGNS FOR ANTERIOR CROWNS � MACHADO ET AL


18

pressure), cleaned with ethanol, driedwith air free of water and oil, and thencemented (Rely X Unicem; 3M ESPE,St Paul, MN). The final dimensions forEH, IH, and IC were the same.

Mechanical Testing andReliability Analysis

To test the implant-abutment con-nection in a challenging scenario,mechanical testing was undertaken withall specimens placed at a 30-degree axialinclination, as per ISO 14801:2007,where the indenter contacted the crownsurface and applied the prescribed loadwithin the step profile. The intent was toprovide a bending component duringloading, which may occur during inter-cuspal position but more commonlypresent in protrusive or laterotrusive

mandibularmovements, namelyanteriorguidance.29 Based on the mean load tofailure of previous studies, SSALT pro-files were determined.1,18,27,28 Thisfatigue testing approach consists of test-ing the samples at stress levels higherthan used stress to facilitate failures ina timely manner. The results of thesetests are then analyzed so that a profileof the failure behavior of the specimensat used stresses can be determined basedon the behavior of the samples at theaccelerated stresses.26

The profiles were designated mild,moderate, and aggressive, with thenumber of specimens assigned to eachgroup in the ratio 3:2:1, respectively.Therefore, of the 18 samples per group,9 were allocated in the mild, 6 in themoderate, and 3 in the aggressive

profiles. Mild, moderate, or aggressiveprofile refer to the increasingly step-wise rapidness in which a specimen isfatigued to reach a certain level of load,meaning that specimens assigned toa mild profile will be cycled longer toreach the same load of a specimenassigned to either moderate or aggres-sive profile.26 The rationale for using atleast 3 profiles for this typeof testingwasbased on the need to distribute failuresacross different step loads, allowing bet-ter prediction statistics, narrowing confi-dence bounds. The prescribed fatiguemethod was SSALT under water at 9Hzwith a servo-all-electric system (Tes-tResources 800L, Shakopee, MN),where the indenter contacted the crownsurface and applied the prescribed loadwithin the step profile.

The specimens were evaluated atthe completion of each fatigue step-stress cycle or final failure (determinedby setting the machine to stop whena compression lower limit of approxi-mately 0.5 mm was reached). Criteriaused for failure were bending or fractureof the fixation screw, partial fracture ortotal fracture of the abutment, andfracture of an implant.

Based on the step-stress distributionof the failures, use-level probabilityWeibull curves (probability of failurevs cycles) with use stress of 175 N and90% 2-sided confidence intervals werecalculated and plotted (Alta Pro 7;ReliaSoft, Tucson, AZ), using a powerlaw relationship for damage accumula-tion. Reliability for amission of 100,000cycles at 175N (90%2-sided confidenceinterval) was calculated for comparisonbetween the groups. The Weibull mod-ulus 2-sided 90% confidence intervalswere calculated using the Fisher matrixmethod.26 For the parameters calculatedin this study, the 90% confidence inter-val range were calculated as follows:

IC 5 E ðGÞ 6 Za sqrtðVar ðGÞÞ ð1Þ

where IC is the confidence bound, E(G) isthe mean estimated reliability for the mis-sion calculated fromWeibull statistics, za isthe z value concerning the given IC level ofsignificance, andVar (G) is the value calcu-lated by the Fisher information matrix.30,31

Fig. 2. Contour plot (Weibull modulus vs characteristic strength) for group comparisons. Notethe overlap between IH and EH.

Table 3. Failure Modes After Mechanical Testing (SSALT) According to the UsedFailure Criteria

EH (n ¼ 18) IH (n ¼ 18) MT (n ¼ 18)

Implant 10 ¼ fracture (hexagon) 18 ¼ intact 18 ¼ intact6 ¼ fracture (body)

2 ¼ intactAbutment 18 ¼ intact 18 ¼ intact 7 ¼ fracture

7 ¼ bending4 ¼ intact

Abutment screw 12 ¼ fracture 18 ¼ fracture 4 ¼ fracture6 ¼ intact 14 ¼ intact



19

distribution. An instructive graphicalmethod to determine whether these datasets are from different populations(based on nonoverlap of confidencebounds) is the utilization of a Weibullparameter contour plot (Weibull mod-ulus [m] vs characteristic strength [h]).As presented in Figure 2, the MT grouppresented a significantly higher charac-teristic strength compared with EH andIH groups, whereas no difference wasobserved between EH and IH groups.The Weibull modulus (m) was m ¼16.58 for EH, m ¼ 13.72 for IH, and,m ¼ 6.80 for MT. The characteristicstrength (h) was h ¼ 290.8 N for EH,h¼ 251N for IH, andh¼ 357.3 forMT(Fig. 2).

Failure ModesAll specimens failed after SSALT.

When component failures were evalu-ated together, failures comprised thecombination of abutment screw bendingor fracture, abutment fracture, andimplant fracture. Observed failuremodes are described in Table 3. ForEH and IH groups, failure predomi-nantly involved abutment screw frac-ture. The abutments remained intactafter mechanical testing for EH and IHgroups. TheMT group presented a high-er number of abutment fractures in com-parisonwith theEHand IHgroups alongwith few failures in abutment screw.Thehighest number of implant failuresoccurred in theEHgroup at the hexagon.Observation of the polarized-light andSEM micrographs of the fractured sur-face of the abutment screws allowed theconsistent identification of fractographicmarkings, such as compression curlwithits opposite tensile side indicating thefracture origin, and the direction of crackpropagation (Figs. 3–5).

DISCUSSION

It has been proposed that differentimplant-abutment connectionswill pres-ent advantages and disadvantages inclinical and laboratory practice.9,15,16

Ultimately, their reliability is the param-eter that will determine their successfrom an overall biomechanical perfor-mance as the highest number of failuresof dental implant prosthetics arise fromconnection failures.33 This study

evaluated the reliability and failuremodes of different implant connectionsfrom the same manufacturer. For thispurpose, a fatigue testing method thatresults in failures that are remarkablysimilar to clinical failures was used.

The life-stress relationship modelallows the extrapolation of a use-levelprobability density function from lifedata obtained at increased stress levels.These models describe the path of a par-ticular life characteristic of the distribu-tion from 1 stress level to another.14 FortheWeibull distribution, the scale param-eter (h) is considered to be stress depen-dent.14 Therefore, the life-stress modelfor data that fits the Weibull distributionis assigned to h.14

The current findings may be ex-plained based on the association amongstress distribution and systems’ reliabil-ity around the weakest component of theimplant-abutment connection: the abut-ment screw.9,14–16 It can be speculatedthat higher levels of stress in the abut-ment screw lead to failures in implantswith EH and IH configuration. Althoughno significant difference in reliabilitywas found between the different implantconnection designs, 2 important factorsshould be considered based on the failuremodes: (1) from a clinical perspective,both the internally connected systems(IH andMT) resulted in the best scenariowhen the restorative overall system isconsidered; that is, implants never frac-tured and repair would be limited to pros-thetic components; and (2) apart from thesame reliability, the MT’s failure wasmainly dictated by damage accumulationrather than fatigue per se. In addition,significantly higher fatigue loads wereneeded for failure of MT compared withEH and IH groups. Because all previousconsiderations were performed consider-ing a mean value of incisal bite force,14,34

the cumulative damage from loads reach-ing 175 N would lead to acceptable per-centage of restoration survival for the allspecimens as per our simulation.

This means that, under functionalocclusion conditions,35 almost all testedspecimens would present satisfactoryfatigue endurance in the wet environ-ment used in this study.

The results of the accelerated lifetesting suggest that the geometric inherentdifferences of each design did play a role

in the use-level probability calculations,which showed that failure rates increasedover time and were related to damageaccumulation in theEHand IHsystembutnot in the MT system (where the systemstrength was more likely to play a role onfailure distribution as per the low betavalue observed). Therefore, a fatigue-associated failure behavior was observedfor the different tested systemswithEHorIH connections, as evidenced by theresulting b . 1 (also called the Weibullshape factor).28 The results of this studyare in contrast with those recently pre-sented, where fatigue resistance of EH,IH, and MT groups presented the sameoutcome,20 and intrinsic differences in theimplants and components geometry mayhave accounted for the different results.

The failure modes were differentfor all groups as each geometric con-figuration and manufacturing toleran-ces do result in different regions andlevel of stress concentration. Fracturesof the abutment screw were more oftenobserved in the IH group. Conversely,EH implant fractures were associatedwith complex fracture scenarios, suchas fractures of the screw that at timeswere associated with failures in thebody or implant hexagon. Differentfrom the other groups, the MT pre-sented more susceptibility to failure ofthe abutment, although occurring atsignificantly higher characteristicstrength values than EH and IH, likelybecause of the larger contact areabetween the abutment and the implantinternal walls relative to IH. In sucha scenario, despite the larger area forstress dissipation during loading, asloading increases, the abutment integrityis challenged by the higher resistanceprovidedby the implantwall thickness atthe cervical region.5 From a clinical per-spective, the complex and multicompo-nent failures observed for the EH groupmay result in catastrophic implant loss.When a technical complication occurs,and theydo seem to increase over time,36

it is desirable that they remain limited toprosthetic restorative components ratherthan in the implant.

CONCLUSIONS

The postulated null hypothesis thatdifferent reliability and failure modes



Failure AnalysisThe failed samples were inspected

in polarized light (MZ-APO stereomi-croscope; Carl Zeiss MicroImaging,Thornwood, NY) and classified accord-ing to the proposed failure criteria forcomparisons between groups. To iden-tify fractography markings and to char-acterize failure origin and propagationdirection, the most representative failedsamples of each group were inspectedunder a scanning electron microscope(SEM) (S-3500N; Hitachi, Osaka,Japan).1,32

RESULTS

The step-stress use-level probabilityWeibull plot and summary statistics ata 175-N load are presented in Figure 1and Table 2, respectively. The step-stress accelerated fatigue permits esti-mates of reliability at a given load level.The calculated reliability with 90% con-fidence intervals for a mission of100,000 cycles at 175 N showed thatthe cumulative damage from loadsreaching 175 N would lead to 99%implant-supported restoration survivalin the EH, 84% in the IH, and 97% inthe MT groups.

The mean beta (b) values (confi-dence interval range) and associatedupper and lower bounds derived fromuse-level probability Weibull calcula-tion (probability of failure vs numberof cycles) were 1.67 (1.56–1.08), 2.52(3.9–1.62), and 0.88 (0.46–1.68) forgroups EH, IH, and MT, respectively.These values indicated that fatigue wasnot an accelerating factor only for fail-ures of group MT, whereas EH and IHpresented failure distributions that wereinfluenced by damage accumulation.The b (called the Weibull shape factor)describes failure rate changes over timewhere b , 1: failure rate is decreasingover time, commonly associated with“early failures” or failures that occurbecause of egregious flaws; b; 1: fail-ure rate that does not vary over time,associated with failures of a randomnature; andb. 1: failure rate is increas-ing over time, associated with failuresrelated to damage accumulation.18

The load-at-failure data during stepstress for each sample were then usedto calculate a probability Weibull

Fig. 3. Representative fractured screw after SSALT of group (IH): (A) SEM micrograph (503)showing a fracture occurring in the abutment screw viewed from its long axis. The whitedotted area shows a compression curl, which evidences fracture origin at the opposing tensileside (white box) and indicates the direction of crack propagation (dcp) (white arrow). B and C,Higher magnification (4503) of the boxed area presented in (A). D and E, Higher magnifi-cations (1503 and 4503, respectively) of the fractured surface showing the dimpled surfaceappearance because of microvoid coalescence (pointer).

Fig. 4. Representative fatigue fractured implant from the external hexagon group: (A) SEMmicrograph (253) showing a fracture occurring in implant viewed from its long axis. The whitedotted area shows a compression curl, which evidences fracture origin at the opposing tensileside (white box) and indicates the direction of crack propagation (dcp) (white arrow). B and C,Higher magnification (8003) of the boxed area and compression curl shown in (A) wherefractographic marks such as fatigue striations are evident. D and E, Higher magnifications(18003) of the fractured surface showing the direction of crack propagation (dcp) and fatiguestriations (pointer).

Fig. 5. Representative SEM micrograph (453) of a fractured MT abutment.



20

distribution. An instructive graphicalmethod to determine whether these datasets are from different populations(based on nonoverlap of confidencebounds) is the utilization of a Weibullparameter contour plot (Weibull mod-ulus [m] vs characteristic strength [h]).As presented in Figure 2, the MT grouppresented a significantly higher charac-teristic strength compared with EH andIH groups, whereas no difference wasobserved between EH and IH groups.The Weibull modulus (m) was m ¼16.58 for EH, m ¼ 13.72 for IH, and,m ¼ 6.80 for MT. The characteristicstrength (h) was h ¼ 290.8 N for EH,h¼ 251N for IH, andh¼ 357.3 forMT(Fig. 2).

Failure ModesAll specimens failed after SSALT.

When component failures were evalu-ated together, failures comprised thecombination of abutment screw bendingor fracture, abutment fracture, andimplant fracture. Observed failuremodes are described in Table 3. ForEH and IH groups, failure predomi-nantly involved abutment screw frac-ture. The abutments remained intactafter mechanical testing for EH and IHgroups. TheMT group presented a high-er number of abutment fractures in com-parisonwith theEHand IHgroups alongwith few failures in abutment screw.Thehighest number of implant failuresoccurred in theEHgroup at the hexagon.Observation of the polarized-light andSEM micrographs of the fractured sur-face of the abutment screws allowed theconsistent identification of fractographicmarkings, such as compression curlwithits opposite tensile side indicating thefracture origin, and the direction of crackpropagation (Figs. 3–5).

DISCUSSION

It has been proposed that differentimplant-abutment connectionswill pres-ent advantages and disadvantages inclinical and laboratory practice.9,15,16

Ultimately, their reliability is the param-eter that will determine their successfrom an overall biomechanical perfor-mance as the highest number of failuresof dental implant prosthetics arise fromconnection failures.33 This study

evaluated the reliability and failuremodes of different implant connectionsfrom the same manufacturer. For thispurpose, a fatigue testing method thatresults in failures that are remarkablysimilar to clinical failures was used.

The life-stress relationship modelallows the extrapolation of a use-levelprobability density function from lifedata obtained at increased stress levels.These models describe the path of a par-ticular life characteristic of the distribu-tion from 1 stress level to another.14 FortheWeibull distribution, the scale param-eter (h) is considered to be stress depen-dent.14 Therefore, the life-stress modelfor data that fits the Weibull distributionis assigned to h.14

The current findings may be ex-plained based on the association amongstress distribution and systems’ reliabil-ity around the weakest component of theimplant-abutment connection: the abut-ment screw.9,14–16 It can be speculatedthat higher levels of stress in the abut-ment screw lead to failures in implantswith EH and IH configuration. Althoughno significant difference in reliabilitywas found between the different implantconnection designs, 2 important factorsshould be considered based on the failuremodes: (1) from a clinical perspective,both the internally connected systems(IH andMT) resulted in the best scenariowhen the restorative overall system isconsidered; that is, implants never frac-tured and repair would be limited to pros-thetic components; and (2) apart from thesame reliability, the MT’s failure wasmainly dictated by damage accumulationrather than fatigue per se. In addition,significantly higher fatigue loads wereneeded for failure of MT compared withEH and IH groups. Because all previousconsiderations were performed consider-ing a mean value of incisal bite force,14,34

the cumulative damage from loads reach-ing 175 N would lead to acceptable per-centage of restoration survival for the allspecimens as per our simulation.

This means that, under functionalocclusion conditions,35 almost all testedspecimens would present satisfactoryfatigue endurance in the wet environ-ment used in this study.

The results of the accelerated lifetesting suggest that the geometric inherentdifferences of each design did play a role

in the use-level probability calculations,which showed that failure rates increasedover time and were related to damageaccumulation in theEHand IHsystembutnot in the MT system (where the systemstrength was more likely to play a role onfailure distribution as per the low betavalue observed). Therefore, a fatigue-associated failure behavior was observedfor the different tested systemswithEHorIH connections, as evidenced by theresulting b . 1 (also called the Weibullshape factor).28 The results of this studyare in contrast with those recently pre-sented, where fatigue resistance of EH,IH, and MT groups presented the sameoutcome,20 and intrinsic differences in theimplants and components geometry mayhave accounted for the different results.

The failure modes were differentfor all groups as each geometric con-figuration and manufacturing toleran-ces do result in different regions andlevel of stress concentration. Fracturesof the abutment screw were more oftenobserved in the IH group. Conversely,EH implant fractures were associatedwith complex fracture scenarios, suchas fractures of the screw that at timeswere associated with failures in thebody or implant hexagon. Differentfrom the other groups, the MT pre-sented more susceptibility to failure ofthe abutment, although occurring atsignificantly higher characteristicstrength values than EH and IH, likelybecause of the larger contact areabetween the abutment and the implantinternal walls relative to IH. In sucha scenario, despite the larger area forstress dissipation during loading, asloading increases, the abutment integrityis challenged by the higher resistanceprovidedby the implantwall thickness atthe cervical region.5 From a clinical per-spective, the complex and multicompo-nent failures observed for the EH groupmay result in catastrophic implant loss.When a technical complication occurs,and theydo seem to increase over time,36

it is desirable that they remain limited toprosthetic restorative components ratherthan in the implant.

CONCLUSIONS

The postulated null hypothesis thatdifferent reliability and failure modes



21

would be found for different implant-abutment connection designs when sub-jected to SSALT was partially accepted.Although the reliability was not differ-ent between EH, IH, and IC groups, thefailure modes were different.

DISCLOSURE

The authors claim to have nofinancial interest, either directly orindirectly, in the products or informa-tion listed in the article.

ACKNOWLEDGMENTS

The implants used in this studywere donated by Implacil de Bortoli,Sao Paulo, Brazil.

REFERENCES

1. Almeida EO, Freitas Júnior AC,Bonfante EA, et al. Effect of microthreadpresence and restoration design (screwversus cemented) in dental implant reliabil-ity and failure modes. Clin Oral ImplantsRes. 2013;24:191–196.

2. Abuhussein H, Pagni G, Rebaudi A,et al. The effect of thread pattern uponimplant osseointegration. Clin Oral Im-plants Res. 2010;21:129–136.

3. Shin YK, Han CH, Heo SJ, et al.Radiographic evaluation of marginal bonelevel around implants with different neckdesigns after 1 year. Int J Oral MaxillofacImplants. 2006;21:789–794.

4. Piao CM, Lee JE, Koak JY, et al. Mar-ginal bone loss around three different implantsystems: Radiographic evaluation after 1year. J Oral Rehabil. 2009;36:748–754.

5. Freitas AC Jr, Bonfante EA, RochaEP, et al. Effect of implant connection andrestoration design (screwed vs. cemented)in reliability and failure modes of anteriorcrowns. Eur J Oral Sci. 2011;119:323–330.

6. Freitas Júnior AC, Bonfante EA,Silva NR, et al. Effect of implant-abutmentconnection design on reliability of crowns:regular vs. horizontal mismatched plat-form. Clin Oral Implants Res. 2012;23:1123–1126.

7. Freitas-Júnior AC, Almeida EO,Bonfante EA, et al. Reliability and failuremodes of internal conical dental implantconnections. Clin Oral Implants Res.2013;24:197–202.

8. Freitas-Júnior AC, Bonfante EA,Martins LM, et al. Effect of implant diame-ter on reliability and failure modes of molarcrowns. Int J Prosthodont. 2011;24:557–561.

9. Almeida EO, Freitas AC Jr, BonfanteEA, et al. Mechanical testing of implant-

supported anterior crowns with differentimplant/abutment connections. Int J OralMaxillofac Implants. 2013;28:103–108.

10. Almeida EO, Júnior AC, BonfanteEA, et al. Reliability evaluation of alumina-blasted/acid-etched versus laser-sintereddental implants. Lasers Med Sci. 2013;28:851–858.

11. Maeda Y, Satoh T, Sogo M. In vitrodifferences of stress concentrations forinternal and external hex implant-abutmentconnections: A short communication.J Oral Rehabil. 2006;33:75–78.

12. Becker W, Becker BE, NewmanMG, et al. Clinical and microbiological find-ings, that can cause failure of dental im-plants [in German]. Quintessenz. 1991;42:9–21.

13. Bozkaya D, Müftü S. Mechanics ofthe tapered interference fit in dental im-plants. J Biomech. 2003;36:1649–1658.

14. Freitas-Júnior AC, Rocha EP,Bonfante EA, et al. Biomechanical evaluationof internal and external hexagon platformswitched implant-abutment connections:An in vitro laboratory and three-dimensionalfinite element analysis. Dent Mater. 2012;28:e218–e228.

15. Gracis S, Michalakis K, Vigolo P,et al. Internal vs. external connections forabutments/reconstructions: A systematicreview. Clin Oral Implants Res. 2012;23(suppl 6):202–216.

16. Sailer I, Muhlemann S, Zwahlen M,et al. Cemented and screw-retainedimplant reconstructions: A systematicreview of the survival and complicationrates. Clin Oral Implants Res. 2012;23(suppl 6):163–201.

17. Goodacre CJ, Kan JY,Rungcharassaeng K. Clinical complica-tions of osseointegrated implants.J Prosthet Dent. 1999;81:537–552.

18. Freitas AC Jr, Bonfante EA,Martins LM, et al. Reliability and failuremodes of anterior single-unit implant-supported restorations. Clin Oral ImplantsRes. 2012;23:1005–1011.

19. Jones AA, Cochran DL. Conse-quences of implant design. Dent Clin NorthAm. 2006;50:339–360.

20. Ribeiro CG, Maia ML, Scherrer SS,et al. Resistance of three implant-abutmentinterfaces to fatigue testing. J Appl OralSci. 2011;19:413–420.

21. Tabata LF, Assunção WG, BarãoVAR, et al. Implant platform switching: Bio-mechanical approach using two-dimensionalfinite element analysis. J Craniofac Surg.2010;21:182–187.

22. Maeda Y, Miura J, Taki I, et al. Bio-mechanical analysis on platform switching:Is there any biomechanical rationale? ClinOral Implants Res. 2007;18:581–584.

23. Pessoa RS, Muraru L, Júnior EM,et al. Influence of implant connection type

on the biomechanical environment ofimmediately placed implantsdCT-basednonlinear, three-dimensional finite elementanalysis. Clin Implant Dent Relat Res.2010;12:219–234.

24. Khraisat A, Stegaroiu R, NomuraS, et al. Fatigue resistance of two implant/abutment joint designs. J Prosthet Dent.2002;88:604–610.

25. Piermatti J, Yousef H, Luke A, et al.An in vitro analysis of implant screw torqueloss with external hex and internal connec-tion implant systems. Implant Dent. 2006;15:427–435.

26. Bonfante EA, Coelho PG, NavarroJM Jr, et al. Reliability and failure modes ofimplant-supported Y-TZP and MCR three-unit bridges. Clin Implant Dent Relat Res.2010;12:235–243.

27. Coelho PG, Silva NR, Bonfante EA,et al. Fatigue testing of two porcelain-zirconia all-ceramic crown systems. DentMater. 2009;25:1122–1127.

28. Martins LM, Bonfante EA, ZavanelliRA, et al. Fatigue reliability of 3 single-unitimplant-abutment designs. Implant Dent.2012;21:67–71.

29. Okeson JP. Chapter? Mechanicsof mandibular movement. In: Managementof Temporomandibular Disorders andOcclusion, 6th ed. St. Louis, MO: MosbyElsevier: 2006:81–94.

30. Abernethy R. The New WeibullHandbook. North Palm Beach, FL: Dr.Robert B. Abernethy; 2006.

31. Nelson W. Accelerated Testing:Statistical Models, Test Plans and DataAnalysis. New York, NY: John Wiley &Sons; 2004.

32. Manda MG, Psyllaki PP, TsipasDN, et al. Observations on an in-vivo failureof a titanium dental implant/abutmentscrew system: A case report. J BiomedMater Res B Appl Biomater. 2009;89:264–273.

33. Jung RE, Pjetursson BE, GlauserR, et al. A systematic review of the 5-yearsurvival and complication rates of implant-supported single crowns. Clin Oral Im-plants Res. 2008;19:119–130.

34. Paphangkorakit J, Osborn JW.The effect of pressure on a maximum inci-sal bite force in man. Arch Oral Biol. 1997;42:11–17.

35. Klineberg IJ, Trulsson M, MurrayGM. Occlusion on implantsdIs therea problem? J Oral Rehabil. 2012;39:522–537.

36. Adell R, Lekholm U, Rockler B,et al. A 15-year study of osseointegratedimplants in the treatment of the edentulousjaw. Int J Oral Surg. 1981;10:387–416.



22

Implant-Abutment Connection Designs for Anterior Crowns...

Documents

Transcript of Implant-Abutment Connection Designs for Anterior Crowns...