Eur J Oral Sci 2011; 119: 323330  2011 Eur J Oral Sci

DOI: 10.1111/j.1600-0722.2011.00837.x European Journal of

Printed in Singapore. All rights reserved
Oral Sciences

Amilcar C. Freitas Jr1,2, Estevam A.

Effect of implant connection and Bonfante3, Eduardo P. Rocha2,
Nelson R. F. A. Silva4, Leonard
restoration design (screwed vs. Marotta1, Paulo G. Coelho1
1
Department of Biomaterials and Biomimetics,

cemented) in reliability and failure New York University College of Dentistry, New
York, NY, USA; 2Department of Dental
Materials and Prosthodontics, So Paulo State
modes of anterior crowns University, Araatuba School of Dentistry,
Araatuba, So Paulo; 3Postgraduate Program
in Dentistry, UNIGRANRIO University School
of Health Sciences, Duque de Caxias, Rio de
Janiero, Brazil; 4Department of Prosthodontics,
Freitas AC Jr, Bonfante EA, Rocha EP, Silva NRFA, Marotta L, Coelho PG. Eect of New York University College of Dentistry, New
implant connection and restoration design (screwed vs. cemented) in reliability and York, NY, USA
failure modes of anterior crowns.
Eur J Oral Sci 2011; 119: 323330.  2011 Eur J Oral Sci

The mechanical performance of cemented or screw-retained implant-supported crowns

with an internal or external conguration is yet to be understood. This in vitro study
evaluated the eect of screw-retained and cement-retained prostheses on internal and
external implantabutment connections. Thereby, the reliability and failure modes of
crowns were investigated. Eighty-four implants (Emls; Colosso Evolution system)
were divided into four groups (n=21 each): screw-retained and internal connection
(Si), screw-retained and external connection (Se), cement-retained and internal Estevam A. Bonfante, Postgraduate Program
in Dentistry, UNIGRANRIO University School
connection (Ci), and cement-retained and external connection (Ce). Ti-6Al-4V
of Health Sciences, Duque de Caxias, Rua
abutments were torqued (30 Ncm) to the implants, and maxillary central incisor metal Prof. Jos de Souza Herdy, 1.160 25 de
crowns were torqued (30 Ncm) or cemented (Rely X Unicem; 3M-ESPE) and sub- Agosto, Duque de Caxias, Rio de Janiero
jected to accelerated life-testing in water. Use-level probability Weibull curves and 25071-202, Brazil
reliability for 50,000 cycles at 150 N were calculated. The b values for Si (1.72), Se
(1.50), Ci (1.34), and Ce (1.77) groups indicated that fatigue/damage accumulation Telefax: +551432342566
E-mail: estevamab@gmail.com
accelerated their failure. The Ci group presented the highest reliability, the Se group
presented the lowest reliability, and Si and Ce groups presented intermediate Key words: cement-retained; fractography;
reliability. Screw-retained restorations presented mainly abutment fracture. Cement- implant-supported prostheses; reliability;
retained restorations resulted in failures of the screw in the Ce group, but implant/ screw-retained
screw fracture in the Ci group. Accepted for publication May 2011

Single-tooth replacements in the anterior region using
osseointegrated implants has become increasingly com-
mon in clinical practice, and the choice of the restorative
components and the connection system between the
implants and the restorations must be considered a
paramount factor for long-term clinical success (1). In
order to provide more predictable results regarding
biological, mechanical, and aesthetic aspects, several
connection designs (internal and external connections,
for either screwed or cemented restorations) have been
developed over the years.
However, the literature is still controversial concerning
the choice of connection and/or crown-retaining system.
The advantages of using cement-retained restorations, as
given by several authors, were primarily aesthetics,
passive t of the crown, and the potentially improved
load distribution during function (24). On the other
hand, previous studies (5, 6) also highlighted the
advantages of screw-retained restorations, such as their
retrievability, making the evaluation of oral hygiene and
peri-implant probing easier.
With respect to the implantabutment connection
congurations, dierent possibilities are also available,
including the external and internal connections that are
either screwed or interference-t, or retained by a
combination of both designs (7). Relative to external
connections, the internal implantabutment connections
have been shown to present higher stability and
improved force distribution as a result not only of their
ability to dissipate lateral loads deeply within the
implant, better shielding the abutment screw from stress,
but also because of the longer internal wall engagement
that creates a sti, unied body to resist joint opening
(micromovement) (811).
In vitro studies have shown improved mechanical
behaviour for internal connections after cyclic loading
relative to external connections (1214). While mechan-
ical strength is gained when internal connections are used
vs. external connections, the thinner lateral xture wall
at the connecting part of internal connections may lead
to a higher absolute strain value at the cervical area (13).
Such stress shift has raised concerns regarding an
increased risk of marginal bone resorption or xture
fracture (13).
Although dierent implantabutment connections
and restoration-retaining mechanisms have been used
324 Freitas et al.
extensively in clinical practice, failures of implant
abutment connection present a relatively common
clinical problem (15). The success rate for the implant-
supported restorations was estimated to be 84.9% at
36 months of loading. The success rate was 88.6% for
xed dental prostheses and 86.4% for single crowns. The
most common technical problems included screw loos-
ening, screw fractures, and veneering material failures.
Because mechanical complications comprised mainly
screw loosening and screw fractures, it was suggested
that the connection between the implant and the super-
structure should be improved (16). In a systematic review
that evaluated the 5-yr survival rates of implant-sup-
ported single crowns, it was observed, from the 26
clinical studies included, that the cumulative incidence of
abutment screw or abutment loosening was 12.7% after
5 yr of clinical service in both external and internal
connections (15).
The main challenge in the development of implant
abutment connection designs relies on reducing the
incidence of mechanical failures while improving the
soft/hard tissue and the prosthetic interface (12, 17, 18).
Thus, evaluation of the reliability (the probability of an
item functioning for a given amount of time without
failure) (1921) and the failure modes of currently used
implantabutment connections may provide insight into
the mechanical behaviour of the dierent systems when
used as single-unit anterior crowns. Therefore, the pres-
ent study sought to evaluate the reliability and the failure
modes of anatomically designed maxillary central incisor
crowns by varying the conguration of the retention
system (screw-retained or cement-retained restoration)
and the type of implantabutment connection (internal
or external). The tested hypothesis was that dierent
reliability and failure modes would be found for dierent
implantabutment connection designs used as anterior
single-unit crowns when subjected to step-stress acceler-
ated life testing (SSALT) (22, 23).
Material and methods
Sample preparation
Eighty-four CP Ti (grade II) dental implants (4 mm in
diameter 10 mm in length; internal and external connec-
tions) (Fig. 1A) (Emls; Colosso Evolution System, Itu, SP,
Brazil) were used. The type of retention system, the
conguration of implant connection, and the dimensions are
presented in Table 1 and Fig. 1B. For screw-retained
restorations the abutment and screw were not detachable
from each other. All implants were vertically embedded in
acrylic resin (Orthoresin; Degudent, Mainz, Germany)
poured in a 25-mm-diameter plastic tube, leaving the pros-
thetic platform at the same level as the potting resin surface.
In the screw-retained and internal connection (Si) and
screw-retained and external connection (Se) groups, a
maxillary central incisor crown was waxed in an attempt to
best reproduce its natural morphology and cast in cobalt
chromium (CoCr) metal alloy (cobalt-chrome partial
denture alloy, Wirobond 280; Bego, Bremen, Germany). In
order to standardize the morphology of the crowns, an
impression was taken from the rst waxed pattern and used
A B
C
Fig. 1. (A) Implant connection congurations tested: (left)
internal and (right) external hexagon. (B) Component assem-
bling for the (from left to right) screw-retained and cement-
retained restorations in the respective external and internal
connection groups. (C) Mechanical testing set-up, where the
load was applied at 30 to the long axis of the implant.
by the technician as a guide during waxing of the remaining
crowns. The crowns presented with a thickness of 1.4 mm,
and their nal dimension was 11 mm from cervical to inci-
sal, 7.4 mm proximally, and 6.7 mm buccolingually.
Crowns were then torqued [30 Newton cm (Ncm)] to the
prefabricated abutments. The Ti-6Al-4V abutment screw
was tightened using a torque gauge (Nobel Biocare,
Goteborg, Sweden), according to the manufacturers
instructions (30 Ncm). Abutments (Ti-6Al-4V) and crowns
(in the screw-retained groups) were torqued only once
before mechanical testing.
The crown morphology and dimensions designed for Si
and Se groups were reproduced in the cement-retained and
internal connection (Ci) and cement-retained and external
connection (Ce) groups by customizing the abutments using
a metallic matrix. A metallic matrix that tted all abutments
in both cement-retained groups was fabricated in order to
standardize the shaping of the remaining abutment groups.
The cement-retained crowns were waxed and cast in an
attempt to produce the best internal t to the customized
abutments. Following connection of the corresponding
abutment to implants (30 Ncm torque, as per the manu-
facturers instructions), the cementation surface of the
crowns was blasted with aluminum oxide (particle size
40 lm, using 276 KPa compressed air pressure), cleaned
with ethanol, dried with air free of water and oil, and then
cemented (Rely X Unicem; 3M ESPE, St Paul, MN, USA).
The nal dimensions for both screw-retained and cement-
retained implant-supported restorations were similar for all
groups (Fig. 1B, bottom).
Mechanical testing and reliability analysis
For mechanical testing, the specimens were subjected to 30
o-axis loading (Fig. 1C). Three specimens of each group
underwent single load-to-fracture (SLF) and 18 specimens
 Reliability of implant-supported crown 325

Table 1
Description of the groups tested in the present study

Groups
Si (n=21) Se (n=21) Ci (n=21) Ce (n=21)

Design of implant Screwed/internal Screwed/external Cemented/internal Cemented/external

abutment connection (lot no. 1005153) (lot no. 100593) (lot no. 1005153) (lot no. 100468)
4.0 mm per 4.1 mm per 4.0 mm per 4.1 mm per
height = 4 mm, height = 4 mm, height = 8 mm, height = 8 mm,
thickness = 1 mm thickness = 1 mm thickness = 1 mm thickness = 1 mm
Emls, Colosso PTC Emls, Colosso body Emls, Colosso long Emls, Colosso
body, Ti-6Al-4V of Ti-6Al-4V body of adaptable long body of
(lot no. 081092) transmucous pillar standard pillar, Ti-6Al-4V adaptable standard
(lot no. 100127) (lot no. 1003232) pillar, Ti-6Al-4V
(lot no. 100127)
Screw Colosso Ti-6Al-4V Colosso Ti-6Al-4V Colosso Ti-6Al-4V Colosso Ti-6Al-4V
xation screw xation screw xation screw xation screw
(lot no. 0907122) (lot no. 0907122) (lot no. 1005143) (lot no. 1005143)

Ce, cement-retained and external connection; Ci, cement-retained and internal connection; Se, screw-retained and external
connection; Si, screw-retained and internal connection.

proles, where maximum loads were up to 1250 N) (23, 25).

Bending was determined when plastic deformation occurred
in the abutment screw to an extent determined as failure.
This failure was dened when the compressive displacement
of the indenter surpassed the lower limit of the sample being
Load (N)

tested (standardized in 1 mm), where the machine auto-

matically halted testing.

Statistical analysis
Use-level probability Weibull curves (probability of failure
vs. cycles) with a power law relationship for damage
Cycles
accumulation were calculated (Alta Pro 7; Reliasoft,
Fig. 2. Step-stress load proles for each group: mild, moderate, Tucson, AZ, USA) (26). Reliability (95% two-sided con-
and aggressive. dence bounds) for completion of 50,000 cycles at a 150 N
(27) load was determined for group comparisons.

were subjected to fatigue testing. Single load-to-fracture was

performed at a cross-head speed of 1 mm min)1 in a
universal testing machine (INSTRON 5666; Instron, Can-
ton, MA, USA) with a at tungsten-carbide indenter
applying the load at the incisal edge of the crown. Based
upon the mean load to failure from SLF, three SSALT
proles were determined for the remaining 18 specimens
(Fig. 2) of each group, which were assigned to mild (n=9),
moderate (n=6), and aggressive (n=3) fatigue proles (at a
ratio of 3:2:1, respectively) (2325). These proles are
named based on the stepwise load increase that the specimen
will be fatigued throughout the cycles until a certain level of
load is reached. It means that specimens assigned to a mild
prole will be cycled for longer to reach the same load level
of a specimen assigned to the moderate or aggressive
proles. The prescribed fatigue method was SSALT
under water at 2 Hz using a servo-all-electric system
(TestResources 800L; TestResources, Shakopee, MN, USA)
where the indenter contacted the incisal edge, applied the
prescribed load within the step prole, and lifted-o the
incisal edge. Fatigue testing was performed until failure
(bending or fracture of the abutment or crown screw, and/or
bending, partial fracture or total fracture of the abutment)
or survival (no failure occurred at the end of step-stress
Failure analysis
Images of failed samples were taken using a macro lens
attached to a digital camera (Nikon D-70s; Nikon, Tokyo,
Japan) and utilized for failure mode classication and
comparisons between groups. In order to identify fracto-
graphic markings and to characterize failure origin and
propagation direction, the most representative failed
samples of each group were inspected rst under a polar-
ized-light microscope (MZ-APO stereomicroscope; Carl
Zeiss MicroImaging, Thornwood, NY, USA) and then by
scanning electron microscopy (Model S-3500N; Hitachi,
Osaka, Japan) (28).
Results
SLF and reliability
The mean SD SLF was 430.17 50.22 N for the Si
group, 526.9 120.71 N for the Se group, 486.8
51.78 N for the Ci group, and 468.8 25.15 N for the
Ce group.
326 Freitas et al.
Fig. 3. Graph showing the probability of failure, as a function
of the number of cycles, for the groups tested simulating 50,000
cycles at 150 N. Note the left positioning of the screw-retained
and external connection (Se) group (black dots) relative to the
positioning of the cement-retained and internal connection (Ci)
group (green dots), which indicates the need for more cycles
to reach failure for this group compared with the former.
Cement-retained and external connection (Ce) (blue dots) and
screw-retained and internal connection (Si) (magenta dots)
groups are overlapped in an intermediate position in the chart.
Table 2
Calculated reliability for dierent designs of implant-supported
restorations in the maxillary anterior region given a mission of
50,000 cycles at a load of 150 N
Groups
Output
(50,000 cycles
@ 150 N) Si Se Ci Ce
Upper 0.80 0.19 0.99 0.87
Reliability 0.64** 0.06*** 0.96* 0.76**
Lower 0.40 0.01 0.89 0.57
Ce, cement-retained and external connection; Ci, cement-
retained and internal connection; Se, screw-retained and
external connection; Si, screw-retained and internal connection.
The number of asterisks depicts statistically homogeneous
groups.
The step-stress derived probability Weibull plots and
summary statistics at a 150 N load are presented in
Fig. 3 and Table 2, respectively. The b values and asso-
ciated upper and lower bounds derived from use-level
probability Weibull calculations (probability of failure
vs. number of cycles) of 1.34 (0.802.10) and 1.77 (1.20
2.62) for screw-retained restorations (Si and Se groups,
respectively), and b values of 1.72 (1.142.58) and 1.50
(0.962.35) for cement-retained restorations (Ci and Ce
groups,, respectively) indicated that fatigue was an
accelerating factor for all groups. The b value describes
failure rate changes over time (b < 1: the failure rate
decreases over time, commonly associated with early
failures or failures that occur as a result of egregious
aws; b  1: the failure rate does not vary over time,
associated with failures of a random nature; and b > 1:
the failure rate increases over time, associated with fail-
ures related to damage accumulation) (23, 29, 30).
The step-stress accelerated fatigue permits estimates
of reliability at a given load level (Table 2). The calcu-
lated reliability with 95% CIs for 50,000 cycles at 150 N
showed that the cumulative damage from loads reaching
150 N would lead to implant-supported restoration
survival in 96% of cemented restorations over implants
with an internal connection (group C), whereas only 6%
would survive in group B (screwed restorations over
implants with an external connection) when considering
the given mission (50,000 cycles at 150 N). These values
depict a statistically signicant dierence between Ci
and Se groups. On the other hand, the overlap between
the upper and the lower limits of reliability values in Si
and Ce groups indicates no statistically signicant dif-
ference. For the same mission (50,000 cycles at 150 N),
survival rates of 64% of the specimens in the Si group
and 76% of the specimens in the Ce group were
observed (Table 2).
Failure modes
All specimens failed after SSALT. When component
failures were evaluated together, failures comprised the
combinations of screw bending or fracture, abutment
bending or fracture, and implant fracture. Failure modes
for all groups are presented in Table 3.
For screw-retained restorations (Si and Se groups),
abutment fracture at the rst abutment screw thread
region was the chief failure mode. In these specimens, the
screws and implants were intact after mechanical testing.
For cement-retained restorations, failures were related to
the abutment screw in the Ce group, but involved
implant and screw fracture in the Ci group. In all spec-
imens of the Ce group, abutment screw fracture was
observed at the sixth to seventh abutment-screw thread
region after SSALT, except for one specimen, which
fractured at the rst thread region. In the Ci, group,
abutment screw fracture at the rst thread region was the
chief failure mode. In addition, longitudinal fracture at
the implants neck area was observed whenever the screw
fractured at the rst thread region in the Ci group
(Fig. 4).
Observation of the polarized-light and scanning
electron microscopy micrographs of the fractured surface
of screws and abutments allowed the consistent identi-
cation of fractographic markings, such as compression
curl and beach marks, which allowed identication of the
aw origin and identication of the direction of crack
propagation (Figs 5 and 6).
Discussion
A slightly lower reliability was found for the Si group
(0.64) than for the Ce group (0.76), although the results
were not signicantly dierent. However, the calculated
reliability scores for the Si and Ce groups were signi-
cantly lower than for the Ci group (0.96) and signi-
cantly higher than for the Se group (0.06). It is possible
that the improved stabilization of the implant/abutment/
crown system provided by the increased internal surface
area in the Si conguration was hampered by the pres-
ence of the second screw for crown retention and the
reduced abutment height. When this xation mode was
 Reliability of implant-supported crown 327

Table 3
Failure modes after mechanical testing according to the failure criteria used

Groups
Mechanical test procedure Si Se Ci Ce

SLF (n=3) Screw Screw Screw Screw

3 intact 3 intact 1 bending 3 bending
2 fracture
(1st thread)
Abutment Abutment Abutment Abutment
1 bending 3 fracture 3 intact 3 intact
2 fracture (1st thread)
(1st thread)
Implant Implant Implant Implant
18 intact 18 intact 18 intact 18 intact
SSALT (n=18) Screw Screw Screw Screw
16 intact 17 intact 15 fracture 17 fracture
2 fracture 1 fracture (1st thread) (6th to 7th
(1st thread) (1st thread) 2 fracture thread region)
(head) 1 fracture
1 fracture (1st thread)
(head/1st thread)
Abutment Abutment Abutment Abutment
16 fracture 17 fracture 18 intact 18 intact
(1st thread) (1st thread)
2 intact 1 intact
Implant Implant Implant Implant
18 intact 18 intact 15 fracture/3 intact 18 intact

Ce, cement-retained and external connection; Ci, cement-retained and internal connection; Se, screw-retained and external con-
nection; Si, screw-retained and internal connection; SLF, single load-to-fracture; SSALT, step-stress accelerated life testing.

A B extra screw component that adds to the number of parts

Fig. 4. Macro picture (A) and scanning electron microscopy
micrograph (B) (30 magnication) of a fractured implant from
the cement-retained and internal connection (Ci) group. In all
specimens, the fracture occurred at the buccal side.
evaluated on an external hexagon, where the stabiliza-
tion of the system is mainly held by the abutment and
crown screws (Se) (14), the system resulted in the lowest
reliability. Fixation of the crown with cement in the
external hexagon conguration (Ce) signicantly
increased the reliability relative to Se, and two explana-
tions may be given to account for this nding: (i) by
eliminating the crown screw, only the longer, and
probably more resistant, abutment screw was challenged
by the fatigue-loading scenario; and (ii) the increased
dimension of the abutment on the cemented crowns may
have provided better support for load distribution
(although the nal system dimension and lever arm act-
ing on the whole structure was the same for all groups).
Such an assumption warrants future investigation. The
requiring adequate t/tolerance increases the likelihood
and amplitude of the overall system motion between
parts. In addition, the increased dimension of the abut-
ment on the cemented crowns, and its intimate interac-
tion with the crown as a result of cementation
(minimizing or eliminating mist and motion between
parts), probably decreased the systems susceptibility to
cyclic loading.
For screw-retained restorations (Se and Si), the chief
failure mode was the abutment screw fracture at the rst
abutment-screw thread region. Such failure probably
occurred because bending was accentuated more in the
region of the abutment screw threads than in the crowns
retention screw. However, because the abutment-screw
threads that connect the abutment to the implant are one
piece, it was considered an abutment fracture.
On the other hand, for cement-retained restorations
(Ce and Ci), fracture at the rst thread region of the
abutment screw was the chief failure mode. A potential
reason for the dominating abutment screw failure mode
is the lever created during loading in the connecting
region of the implant/abutment, creating stresses around
the rst abutment-screw thread region. Moreover,
abutment screw fracture was associated with implant
fracture in the Ci group, but not in the Ce group. Such a
fracture was only observed for the Ci group, presumably
because the cement-retained restoration associated with
internal implant connection represented the strongest
interlock between components among groups, as evi-
328 Freitas et al.

A B C

D E F

Fig. 5. Representative failure modes of screw-retained restorations observed in abutments after step-stress accelerated life testing
(SSALT). (A) A fracture occurring at the rst thread region viewed from the abutments long axis. (B) Scanning electron microscopy
micrograph (50) of the fractured surface of the sample shown in panel A. The white dotted arrows show a compression curl that
provides evidence for the fracture origin at the opposing tensile side (large white arrow). Such a feature is representative of exure
failures and results from a travelling crack changing direction as it enters a compression eld. The beach marks (yellow arrows) are
semi-elliptical lines running perpendicular to the overall direction of fatigue crack propagation and marking successive positions of
the advancing crack front, also indicating the direction of crack propagation (dcp) (black arrow). In a reverse-angle view (C) (45) of
the abutment it is possible to note some marks on the lateral surface (lingual side), indicating that the crack propagated from the
lingual direction to the buccal direction. Similar fractographic markings, representative of the same travelling crack history, can be
observed on the fractured surface of abutments from the screw-retained and external connection (Se) group (D,E,F) (50) . Si, screw-
retained and internal connection.

A B C D

E F G H

Fig. 6. Representative failure modes of cement-retained restorations observed in retention screws after step-stress accelerated life
testing (SSALT). (A) A fracture occurring at the rst thread region. (B) A scanning electron microscopy micrograph (50 magni-
cation) of the fractured surface of the sample shown in panel A. The yellow arrows (in panel A) show beach marks and the white
dotted arrows show the compression curl. The large white arrow (fracture origin) at the opposing tensile side indicates the fracture
origin and the direction of crack propagation (dcp) (black arrow) from the lingual direction to the buccal direction. The micrograph
in panel C represents a higher magnication (500) of the boxed area in panel B in a reverse-angle view. The dotted circle shows ne
parallel marks indicating that two parts rubbed together during fatigue testing. The micrograph in panel D is a higher magnication
(500) of the fracture origin shown in panel B. (E) A representative image illustrating a fracture occurring between the sixth and
seventh threads in a sample from the cement-retained and external connection (Ce) group. (F,G) Scanning electron microscopy
micrographs of the fractured surface of the sample shown in panel E, illustrating the fracture origin (the larger white arrow), the
compression curl (the white dotted arrows), and the direction of crack propagation (the black arrow) from the lingual direction to the
buccal direction. The micrograph in panel H is a higher magnication (500) of the boxed area in panel F, showing the fracture
origin. Ci, cement-retained and internal connection.

denced by the signicantly higher reliability. Thus, the
implants from the Ci group were subjected to loads of
higher magnitude and did not support the stress con-
centration at the implants neck area (13, 31). From a
clinical perspective, the resulting failure mode involving
implant fracture presents the only scenario where pros-
thetic replacement is no longer an option under the same
support.
Considering that the reported maximal bite force in
the incisor area may vary from 108 N (32) to 190 N (33,
34), the load used for reliability calculation fell within the
physiologic range.
Our results showed higher fatigue endurance for
cement-retained restorations, regardless of the congu-
ration of implant connection, and are in contrast to those
found by Zarone et al. (2007), who compared screw
restorations and cement-retained restorations using
implants restored with porcelain-fused-to-metal crowns.
It should be pointed out that in their work all the sam-
ples failed by cohesive fractures of the porcelain veneer
and not by implant component failure. In order to
address the question of mechanical performance of
cemented vs. screwed restorations in both external and
internal connections, we utilized a metallic crown, which
eliminated failure of the restorative crown material and
permitted us to analyze the mechanical performance of
the dierent groups.
Concerning the conguration of implant connection
(internal vs. external), higher reliability was observed in
samples with internal connection only when the screw
was retained. It has been shown that deep joints favoured
structural strength of implant systems (12, 14). The
increased stability, and thereby improved stress distri-
bution associated with lower micro-movements, were
identied as the main reasons for the better mechanical
behaviour of internally connected systems (13). It should
be noted, however, that because of engineering design
constraints, such as minimal wall thickness for proper
mechanical performance, dierences in both external and
internal features of the implant, abutment, and screw
designs existed. While from a research standpoint it is
highly desirable that only the connection is changed
(with the connecting screw and the implant remaining
the same), such interplay is unfeasible when one is
attempting to make clinically relevant comparisons
(implants presenting the same diameter, length, and
crown size) between external and internal connections in
most commercially available systems. The reason is that
alterations in the external shape of the implant are
usually performed by manufacturers in order to maintain
tolerances for appropriate t and wall thickness to
ensure internal connection robustness.
Based on fractographic features of failed specimens
observed in polarized-light and scanning electron
microscopy, all fractures were characterized by material
tearing and exhibited gross plastic deformation,
suggesting ductile fractures (35, 36). In a higher magni-
cation of the fractured surface, ne parallel marks
could be observed, indicating the smearing of the tita-
nium alloy surface as a result of two parts rubbing
together during fatigue testing (37). The ductile fractures
Reliability of implant-supported crown 329
caused by stresses exceeding the material yield strength
left marks indicating crack propagation from the lingual
to the buccal direction, where forces naturally occur.
The most common fracture-origin sites observed in
fatigue are stress raisers, including porosity and voids,
thread roots, sharp radii of curvature, through holes,
and surface discontinuities (38). As per our imaging
analysis, the threaded area of crown and abutment
screws appeared to be more prone to fracture as a
result of the presence of thread roots and surface
discontinuities. Moreover, analyses of the samples
fractured surface at higher magnications showed a
concentration of voids in the area of fracture origin.
In an attempt to simulate the oral environment, the
present study comprised fatigue in water, which has
been suggested as an important service-related cause of
failure in metals (38).
As the restoration of single-unit edentulous spaces in
the anterior region with implant-supported crowns is a
challenging scenario in terms of long-term success and
aesthetics (39), it is crucial to acknowledge the functional
and mechanical limitations of the implantabutment
connections (4). Therefore, further evaluations of the
implantabutment stability, combined with fatigue
testing, are warranted. The postulated hypothesis, that
dierent implantabutment connection designs used as
anterior single-unit crowns would result in dierent
reliability and failure modes, was conrmed. Cement-
retained implant-supported restorations with an internal
implant connection showed the highest reliability
between investigated groups after SSALT, while screw-
retained restorations associated with an external con-
nection resulted in the lowest reliability.
Acknowledgements The authors are thankful to CNPq Brazil
(#141870/2008-7), Marotta Dental Studio (Farmingdale, NY,
USA), and Emls Ind & Com Ltda (Itu, SP, Brazil) for their
support. We also would like to thank Dr. George Quinn
(National Institute of Standards and Technology USA) and
Dr. Susanne Scherrer (University of Geneva, School of Dental
Medicine) for their contributions to fractography analysis.
References
1. Zarone F, Sorrentino R, Traini T, Di lorio D, Caputi S.
Fracture resistance of implant-supported screw- versus cement-
retained porcelain fused to metal single crowns: SEM fracto-
graphic analysis. Dent Mater 2007; 23: 296301.
2. Misch CE. Contemporary implant dentistry, 2nd edn. St. Louis:
Mosby, 1998; 549573.
3. Rajan M, Gunaseelan R. Fabrication of a cement- and
screw-retained implant prosthesis. J Prosthet Dent 2004; 92:
578580.
4. Freitas Junior AC, Goiato MC, Pellizzer EP, Rocha EP,
de Almeida EO. Aesthetic approach in single immediate
implant-supported restoration. J Craniofac Surg 2010; 21: 792
796.
5. Michalakis KX, Hirayama H, Garefis PD. Cement-retained
versus screw-retained implant restorations: a critical review. Int
J Oral Maxillofac Implants 2003; 18: 719728.
6. Albrektsson T, Zarb G, Worthington P, Eriksson AR. The
long-term efficacy of currently used dental implants: a review
and proposed criteria of success. Int J Oral Maxillofac Implants
1986; 1: 1125.
330 Freitas et al.
7. Misch C. Principles for screw-retained prostheses. Misch CE
Dental Implant Prosthetics San LuIs: Elsevier Mosby, 2005;
452471.
8. Binon PP. Implants and components: entering the new
millennium. Int J Oral Maxillofac Implants 2000; 15: 7694.
9. Finger IM, Castellon P, Block M, Elian N. The evolution
of external and internal implant/abutment connections. Pract
Proced Aesthet Dent 2003; 15: 625632; quiz 634.
10. Sailer I, Sailer T, Stawarczyk B, Jung RE, Hammerle CH.
In vitro study of the influence of the type of connection on the
fracture load of zirconia abutments with internal and external
implant-abutment connections. Int J Oral Maxillofac Implants
2009; 24: 850858.
11. Da Silva EF, Pellizzer EP, Quinelli Mazaro JV, Garcia
Junior IR. Influence of the connector and implant design on
the implant-tooth-connected prostheses. Clin Implant Dent
Relat Res 2010; 12: 254262.
12. Khraisat A, Stegaroiu R, Nomura S, Miyakawa O. Fatigue
resistance of two implant/abutment joint designs. J Prosthet
Dent 2002; 88: 604610.
13. Maeda Y, Satoh T, Sogo M. In vitro differences of stress
concentrations for internal and external hex implant-abutment
connections: a short communication. J Oral Rehabil 2006; 33:
7578.
14. Steinebrunner L, Wolfart S, Ludwig K, Kern M. Implant-
abutment interface design affects fatigue and fracture strength
of implants. Clin Oral Implants Res 2008; 19: 12761284.
15. Jung RE, Pjetursson BE, Glauser R, Zembic A, Zwahlen
M, Lang NP. A systematic review of the 5-year survival and
complication rates of implant-supported single crowns. Clin
Oral Implants Res 2008; 19: 119130.
16. Muche R, Krausse A, Strub JR. [Success rates of implant
supported prostheses in partially edentulous patients Part II].
Schweiz Monatsschr Zahnmed 2003; 113: 404410.
17. Piermatti J, Yousef H, Luke A, Mahevich R, Weiner S. An
in vitro analysis of implant screw torque loss with external hex
and internal connection implant systems. Implant Dent 2006;
15: 427435.
18. Pjetursson BE, Tan K, Lang NP, Bragger U, Egger M,
Zwahlen M. A systematic review of the survival and compli-
cation rates of fixed partial dentures (FPDs) after an observa-
tion period of at least 5 years. Clin Oral Implants Res 2004; 15:
625642.
19. Guess PC, Zavanelli RA, Silva NR, Bonfante EA, Coelho
PG, Thompson VP. Monolithic CAD/CAM lithium disilicate
versus veneered Y-TZP crowns: comparison of failure modes
and reliability after fatigue. Int J Prosthodont 2010; 23: 434
442.
20. Silva NR, Bonfante EA, Zavanelli RA, Thompson VP,
Ferencz JL, Coelho PG. Reliability of metalloceramic and
zirconia-based ceramic crowns. J Dent Res 2010; 89: 1051
1056.
21. Bonfante EA, Coelho PG, Navarro JM Jr, Pegoraro LF,
Bonfante G, Thompson VP, Silva NR. Reliability and
failure modes of implant-supported Y-TZP and MCR three-
unit bridges. Clin Implant Dent Relat Res 2010; 12: 235
243.
22. Coelho PG, Bonfante EA, Silva NR, Rekow ED, Thompson
VP. Laboratory simulation of Y-TZP all-ceramic crown clinical
failures. J Dent Res 2009; 88: 382386.
23. Coelho PG, Silva NR, Bonfante EA, Guess PC, Rekow ED,
Thompson VP. Fatigue testing of two porcelain-zirconia
all-ceramic crown systems. Dent Mater 2009; 25: 11221127.
24. Nelson W. Accelerated testing: statistical models, test plans and
data analysis. New York: John Wiley & Sons, 1990.
25. Silva NR, Coelho PG, Fernandes CA, Navarro JM, Dias
RA, Thompson VP. Reliability of one-piece ceramic implant.
J Biomed Mater Res B Appl Biomater 2009; 88: 419426.
26. Zhao W, Elsayed E. A general accelerated life model for step-
stress testing. IEEE Trans Reliabil 2005; 37: 10591069.
27. Hellsing G. On the regulation of interincisor bite force in man.
J Oral Rehabil 1980; 7: 403411.
28. Manda MG, Psyllaki PP, Tsipas DN, Koidis PT. Obser-
vations on an in-vivo failure of a titanium dental implant/
abutment screw system: a case report. J Biomed Mater Res B
Appl Biomater 2009; 89: 264273.
29. Silva NR, de Souza GM, Coelho PG, Stappert CF, Clark
EA, Rekow ED, Thompson VP. Effect of water storage time
and composite cement thickness on fatigue of a glass-ceramic
trilayer system. J Biomed Mater Res B Appl Biomater 2008; 84:
117123.
30. ReliaSoft. The Weibull distribution and beta. (http://
www.reliasoft.com/newsletter/1q2001/beta.htm), 2010.
31. van Steenberghe D, Naert I, Jacobs R, Quirynen M.
Influence of inflammatory reactions vs. occlusal loading on peri-
implant marginal bone level. Adv Dent Res 1999; 13: 130135.
32. Helkimo E, Carlsson GE, Helkimo M. Bite force and state of
dentition. Acta Odontol Scand 1977; 35: 297303.
33. Helkimo E, Ingervall B. Bite force and functional state of
the masticatory system in young men. Swed Dent J 1978; 2:
167175.
34. Hagberg C. Assessment of bite force: a review. J Craniomandib
Disord 1987; 1: 162169.
35. Handbook A. Failure analysis and prevention, vol. 11. Materials
Park, OH: ASM International, The Materials Information
Society, 1992.
36. Scott WW Jr. Fractography, Vol. 12, Metals Handbook, 9th
edn. Materials Park: ASM International, 1992.
37. Voort V. Conducting the failure examination (in metallic
components). Metals Engng Q 1975; 15: 3136.
38. Parrington RJ. Fractography of metals and plastics. In:
Scott WW Jr, eds. Practical failure analysis. Vol. 2. Materials
Park: ASM International, 2002; 1646.
39. Sailer I, Zembic A, Jung RE, Hammerle CH, Mattiola
A. Single-tooth implant reconstructions: esthetic factors
influencing the decision between titanium and zirconia
abutments in anterior regions. Eur J Esthet Dent 2007; 2:
296310.