Received: 11 December 2019 Revised: 10 February 2020 Accepted: 19 March 2020

DOI: 10.1111/jerd.12584

RESEARCH ARTICLE

Effect of incisal preparation design on load-to-failure

of ceramic veneers

Sy Yin Chai BDSc, DClinDent1 | Vincent Bennani DDS, PhD1 |

John M. Aarts BHealSc, BEd, MHealSc | Karl Lyons BDS, MDS, PhD1
1
|
Bronwyn Lowe BE, BAppSc, PhD2

1
Faculty of Dentistry, Department of Oral
Rehabilitation, University of Otago, Dunedin, Abstract
New Zealand Objective: This in vitro study aims to evaluate load-to-failure of ceramic veneers with
2
Department of Applied Sciences, University
butt joint (BJ) and feathered edge (FE) incisal preparation designs, and to correlate
of Otago, Dunedin, New Zealand
these results to the failure mode of the restorations.
Correspondence
Materials and methods: One hundred and forty-eight typodont teeth (customized
Dr Vincent Bennani, Faculty of Dentistry,
Department of Oral Rehabilitation, University Nissin A25A-UL19B) were divided into two different preparation configurations BJ
of Otago, PO Box 647, Dunedin 9054,
and FE and two different loading angulations, 0 and 20 . Lithium disilicate ceramic
New Zealand.
Email: [email protected] veneers (IPS e.max CAD, Ivoclar Vivadent) were milled using computer-aided-design-
and-computer-aided-manufacturing (CAD/CAM) techniques. Veneers were bonded
to typodont teeth with resin cement (IPS Variolink Esthetic, Ivoclar Vivadent). Each
group was loaded at the incisal edge using an Instron Universal Testing Machine at a
crosshead speed of 0.01 mm/s till failure.
Results: Pairwise comparison showed veneers from the BJ groups had a significantly
higher load-to-failure value compared to the FE groups. Veneers with a FE prepara-
tion design loaded at 20 angulation had the lowest load-to-failure value.
Conclusions: Within the limitations of the present study, both incisal preparation
designs and loading angulations have significant effects on the load-to-failure values
of ceramic veneers. BJ group exhibits a significantly higher load-to-failure value com-
pared to the FE group.
Clinical significance: BJ incisal preparation is preferred over FE preparation design.

KEYWORDS

butt joint, feathered edge, load-to-failure, preparation design, veneer preparation

1 | I N T RO DU CT I O N laminate veneers have high survival rates.7 Glass-ceramic laminate

veneers show a slightly higher survival rate than feldspathic porcelain
Ceramic veneers have become a popular treatment modality due to veneers, 94% vs 87%, respectively. The complications reported were
their conservative and esthetic nature and clinical performance.1-3 fracture/chipping (4%), debonding (2%), severe marginal discoloration
Compared to resin composite veneers, ceramic veneers are stronger (2%), endodontic problems (2%), and secondary caries (1%).7 A recent
and more durable, as they are less susceptible to wear, marginal frac- retrospective clinical study conducted in a university setting by two
4-6
ture, discoloration, staining, and plaque accumulation. A recent sys- academic staff with a follow-up of 20 years indicated that ceramic
tematic review and meta-analysis of the clinical outcome of ceramic fracture was the main reason for failure of ceramic veneers.1 This
veneers showed that both feldspathic porcelain and glass-ceramic finding is also supported by numerous earlier clinical studies.3,8-16

J Esthet Restor Dent. 2020;1–9. wileyonlinelibrary.com/journal/jerd © 2020 Wiley Periodicals, Inc. 1

2 CHAI ET AL.

The number of fractures increased dramatically as the number of The literature also shows there is lack of correlation between the
years in service increased, from 4% at 5 years to 34% at 10 years.13 in vitro studies and failure mode of these restorations. Most in vitro
The incisal edge was the most common location of the ceramic frac- studies applied only one loading angulation. Wall et al applied two
ture.9,10 Nordbø et al reported failures at the composite-cement inter- loading angulations (130 and 137 ), but the angulation difference was
face and on the buccal surface of the veneer and that failure of insignificant.30
ceramic veneers also occurred due to buccal incisal chipping.10 This in vitro study aimed to compare the load-to-failure values of
Clinically, ceramic veneer fracture has been attributed to numer- ceramic veneers with the FE and the BJ preparation designs under
ous predisposing factors. These factors range from the incisal prepara- standardized conditions and at two loading angulations. Fractographic
tion design, to the type of ceramic, to the luting agent and ratio of failure analysis is performed to investigate the characteristic fracture
porcelain to luting composite thickness, and so on.1,3,11,12,17-19 features and direction of crack propagation of the samples with
Although the incisal preparation design for ceramic veneers has been different incisal preparation designs and under different loading
widely discussed, there is no consensus on whether incisal reduction angulations.
is necessary and how much of the incisal overlap should be pro-
vided.20-24 Laboratory and clinical surveys reported that the two most
common incisal preparation designs provided are the BJ and the 2 | M A T E R I A L S A N D M ET H O D S
FE.25,26 Although, the best level of scientific and clinical evidence is
obtained by randomized controlled trials,27 heterogeneity of clinical The sample teeth (Typodont teeth, A252A-UL 19B and modified
variables has made the direct comparison of the different types of A252A-UL 19B; Nissin Dental Products Inc., Kyoto, Japan) were split
28
incisal preparations impossible. Hence, the next best level of scien- evenly (37 per group) between four different groups, two groups with
tific evidence is obtained via in vitro studies. in vitro studies on FE preparations and two groups with BJ preparations. The customized
mechanical properties of ceramic veneers can be divided into two typodont teeth with standardized buccal veneer preparation with FE
main types of tests load-to-failure29-37 or fatigue.38-40 incisal preparation design (FE group) (A252-UL 19B) were obtained
Of the three in vitro studies which have applied fatigue testing to from the dental manufacturing company (Nissin Dental Products Inc)
their specimens, Bergoli et al and Stappert et al reported no veneer (Figure 1). The incisal edge of a typodont tooth (A252A-UL 19B) was
fractures during initial cyclic loading simulating 4 and 5 years of clini- reduced by 1.5 mm to simulate the BJ incisal preparation and the
cal service, respectively,38,39 whereas the study by Chaiyabutr et al resulting mesial and distal sharp edges were rounded to avoid stress
reported failure of all veneers under 100 000 cycles (equivalent to concentration points on ceramic veneers (Figure 2). The sample was
less than one-half a year of clinical service).40 These inconsistent
results were due to the lack of standardization of the cyclic loading
procedures and the sliding phenomenon, which resulted in fracture or
cracks in abutment teeth rather than ceramic veneers.38-40 A meta-
analysis of in vitro studies on ceramic veneers with different prepara-
tion designs revealed that ceramic veneers with the BJ preparation
has the least effect on the strength of the tooth, and veneers with the
palatal chamfer (PC) preparation are the most susceptible to ceramic
fractures based on static load-to-failure testing.28 In this meta-analysis,
there were no statistically significant differences in the load-to-failure
values between ceramic veneers with FE and the PC preparation
designs, and between the PC and BJ preparation designs. However,
no comparison was made between the FE and BJ preparations in
terms of the load-to-failure value due to the inclusion and exclusion
criteria of this study.28 Other in vitro studies directly comparing
ceramic veneers with the FE and BJ preparation designs either show
no statistically significant difference in load-to-failure values between
the two designs29-31 or statistically higher load-to-failure value with
the BJ preparation design.37 However, it is hard to draw conclusions
based on these studies due to the heterogeneity of study design, lack
of standardization of loading conditions, and inconsistent quality of
abutment teeth.28,31,33,36,38 In addition, numerous studies have failed
to obtain an adequate sample size to allow statistical power analy-
sis.37,41 Various studies have also applied different mounting appara-
F I G U R E 1 Prepared typodont tooth with FE incisal preparation
tus, loading angulations and points, plunger dimension and shape, design by Nissin Dental Products Inc., Kyoto, Japan. FE,
type of load (static vs cyclic), and thermocycling procedure.31,38,42 feathered edge
CHAI ET AL. 3

F I G U R E 2 A, Image of
typodont tooth prior to 1.5 mm
incisal reduction, B, typodont
tooth after modification to
simulate the BJ preparation
design. BJ, butt joint

then sent to Nissin Dental Products Inc. for customization of the BJ spray, dried, and etched with 37% phosphoric acid gel (Total Etch Gel,
group (modified A252A-UL 19B). The sample size had previously been Ivoclar Vivadent) for 30 seconds. After rinsing, Syntac Primer (Ivoclar
determined by statistical power analysis and was based on the results Vivadent) was applied onto the abutment tooth surface using a micro
of the pilot studies. Teeth were assigned to four groups: (a) BJ, 20 brush for 15 seconds. Excess primer was dispersed and dried thor-
loading angulation (BJ 20 ), (b) BJ, 0 loading angulation (BJ 0 ), (c) FE, oughly. Syntac Adhesive (Ivoclar Vivadent) was then applied for
20 loading angulation (FE 20 ), and (d) FE, 0 loading angulation 10 seconds, thoroughly dried but not rinsed off before Heliobond
 
(FE 0 ). The incline of 20 was determined based on our pretesting. A (Ivoclar Vivadent) was applied and dispersed with oil-free air. Variolink
third group at 45 was considered but the 45 loading angulation Esthetic DC (Ivoclar Vivadent) was applied onto the inner surface of
resulted in bending stress and fracture in the neck region of the the ceramic veneer and the veneer was seated on the abutment tooth
typodont teeth. The dimensions of the sample teeth were evaluated following its path of insertion and sustained with finger pressure by a
for consistency using a laboratory putty stent and confirmed using an single operator.45 Excess cement was removed and the veneer was
electronic caliper. tacked in placed by light-curing a small area for 3 seconds (Bluephase,
Ceramic veneers with the FE and BJ incisal preparation designs 650 mW/cm2, LOW mode; Ivoclar Vivadent). The restoration margins
were fabricated with IPS e.max CAD blocks (HT A1/C14; Ivoclar were covered with glycerin gel (Liquid Strip; Ivoclar Vivadent) immedi-
Vivadent, Schaan, Liechtenstein) using the CAD/CAM technique ately after removal of excess cement. The light-curing process was
(Amann Girrbach AG, Koblach, Austria). The buccal thickness of the repeated on the other surfaces (mesiobuccal, mid-buccal, distobuccal,
ceramic veneers of both groups was 0.5 mm. The BJ group had an mesiopalatal, midpalatal, and distopalatal) for 20 seconds.
incisal ceramic thickness of 1.5 mm. The designated cement gap of Wax jigs were created at the root and base area of the typodont
50 μm was incorporated within the design. Prior to the cementation teeth simulating 20 and 0 loading angulations. A silicone mold was
procedure, each veneer was randomly seated on a prepared tooth made of the waxed jig and embedded typodont teeth using regular
with finger pressure to check its marginal fit on middle third area of body polyvinylsiloxane material (Exahiflex, GC Corporation, Sydney,
each surface (mesial, distal, incisal/palatal finish line, and buccal) using Australia) and laboratory silicone putty (Protesil Lab Putty, Oraltec,
an optical light microscope (Alphaphot-2 YS2; Nikon, Tokyo, Japan). Auckland, New Zealand). Self-cure acrylic resin (Vertex Castapress,
None of the mean marginal gap exceeded 100 μm.43 Vertex Dental, Soesterberg, The Netherlands) was mixed according to
The surface preparation (veneer and abutment tooth) and cemen- the manufacturer's instructions, poured into the silicone mold and
tation procedures were confirmed in our pilot studies. Surface prepa- cured in a pressure pot for 30 minutes at 2.5 bar. All specimens were
ration of the ceramic veneers was completed using 5% HF acid (IPS embedded 1 mm below the level of CEJ. Once the acrylic resin had
Ceramic Etching Gel, Ivoclar Vivadent) for 20 seconds, rinsed and fully cured, the angulation of the embedded typodont teeth was
dried before Monobond Plus (Ivoclar Vivadent) was applied for confirmed.
60 seconds prior to air dispersion. The cementation surface of abut- The Instron Universal Testing Machine (Instron Corp, Canton,
ment tooth was lightly sandblasted with 50 μm aluminum oxide parti- Massachusetts) was calibrated before the start of the testing session.
cles at 0.2 MPa (2 bar) pressure for 10 seconds at a distance of 10 cm A customized plunger was attached to the 2500N Instron load cell,
from the sample.44 The sandblasted surface was rinsed with water with the cross-head speed set at 0.01 mm/s. The customized
4 CHAI ET AL.

stainless-steel plunger of 40 mm diameter was designed assuming tested groups are shown in Figure 3. Group BJ 0 yielded the greatest
that anterior wear facets are of 0.5 mm in diameter, with a 100 N load-to-failure value (mean 834 ± SD 105 N), followed by group BJ
loading force and 509 MPa contact pressures. 46
The embedded 20 (681 ± 141 N), FE 0 (663 ± 157 N), and FE 20 (280 ± 90 N).
typodont tooth was mounted and the orientation was standardized The mean load-to-failure value for group FE 20 was significantly
using a protractor. All samples were loaded to failure. A load vs time lower compared to the other three groups. The Kaplan-Meier failure
graph was generated by Exponent texture analysis software (Stable function for each of the groups indicated that the failures from group
Micro Systems, Surrey, UK) during each test. The maximum load-to- FE 20 occurred mostly at loads lower than any of the failures for
failure values were obtained from the first drop in load. 46
A second groups BJ 20 and FE 0 , and entirely before any failures for group BJ
reading of the graphs was conducted by an independent reviewer. 0 . There was no evidence of different patterns of failure for groups
The data were analyzed using Stata Statistical Software (Stata BJ 20 and FE 0 , which had similar median load-to-failure. From a
15.1; Stator LLC, College Station, Texas). Appropriate summary statis- two-way ANOVA looking at the data shown in Figure 3, there was
tics (means and standard deviations for loads, and counts and percent- overall evidence of a difference between the four groups (Wald
ages for type of failure) were provided for load-to-failure results by P < .001). There was evidence that preparation design and loading
preparation-angle combinations. Two-way analysis of variance angulation interacted (P < .001) so that the effect of angle (ie 0 vs
(ANOVA) models were used to assess the overall preparation and 20 ) varied by group and equivalently that the effect of group (ie BJ
angle effects on load-to-failure along with the interaction between vs FE) varied by angle. The load-to-failure value for group FE 20 was
the two factors. Post-hoc tests were investigated where the interac-
tion or main effect was statistically significant. A three-way ANOVA
was used to compare loads between groups by mode of failure with
the same model diagnostics used. Poisson regression (with robust
standard errors) was used to model the risk for the type of failure
(mixed or cohesive) in a similar way. Time until failure was examined
using a Kaplan-Meier curve with a log-rank test used to compare the
groups. All differences were presented accompanied by 95% confi-
dence interval (CI) and two-sided P < .05 was considered statistically
significant. Modes of failure were macroscopically assessed and ana-
lyzed after the tests were completed. Qualitative analysis was initially
carried out by visual interpretation under fiberoptic transillumination
and a stereomicroscope with magnification ranging from ×5 to ×378
(Nikon Stereo Microscope Smz 800, Nikon, New York). The mode of
failure (eg adhesive, cohesive or mixed cohesive, and adhesive) for
veneers of each group was determined and statistically analyzed. One
sample from each group was randomly selected for further analysis F I G U R E 3 Mean load-to-failure values in Newtons for the
under a scanning electron microscope (SEM) (JEOL JSM-6700F; JEOL different groups with 95% confidence intervals
Ltd., Tokyo, Japan). The samples were coated on both sides with
approximately 10 nm of platinum in an Emitech K 575X; Peltier-
cooled high-resolution sputter coater (EM Technologies Ltd., Kent,
England) before they were attached to the SEM specimen holder
using a double-sided carbon tape and viewed in a field emission SEM
(JEOL JSM-6700F; JEOL Ltd) at an accelerating voltage of 5 kV or
10 kV using the lower secondary detector. The analysis of fracture
patterns of the samples was based on the National Institute of
Standards and Technology (NIST) Recommended Practice Guide on
Fractography of Ceramics and Glasses.47

3 | RESULTS

The load-to-failure values of the second measurement by an indepen-

dent reviewer were found to be extremely consistent with the first
measurement (ICC 0.97, 95% CI 0.96, 0.98), with the 95% CI entirely
above the 0.95 level. All analyses in this section use the measure-
ments from the first reviewer only. The load-to-failure means for all FIGURE 4 Failure modes for the different groups
CHAI ET AL. 5

statistically significantly lower compared to the other three groups (all FE 20 exhibited cohesive failure. From a Poisson regression model of
P < .001). All pairwise comparisons were statistically significant failure events, overall, the four groups differed in terms of the distri-
 
(P < .001) except for a comparison between BJ 20 and FE 0 bution of failure mode (Wald P = .006) with evidence of an interaction
(P = .552), where both factors (preparation and angle) are varied simul- (P = .002) showing that the effect of angle varied between the two
taneously. None of the veneers in the four groups de-bonded during preparations. Group differences involving FE 20 were all statistically
testing. significant (all P < .001). No statistically significant differences were
  
The mode of failure for the four groups (BJ 20 , BJ 0 , FE 20 , found for BJ 20 and BJ 0 (P = .650), FE 0 and BJ 20 (P = .179), and

and FE 0 ) was summarized in Figure 4. The mode of failure for all FE 0 and BJ 0 (P = .344). All cohesive failures occurred within the
groups was either mixed or cohesive. The majority of samples in group ceramic veneers. None of the samples had adhesive failure
(debonding/delamination). Of the samples with mixed failure mode,
adhesive failures occurred almost exclusively between the adhesive
resin cement and typodont tooth. From a three-way ANOVA (prepara-
tion-angle-failure mode), within each group the mean load-to-failure
value did not vary between failure modes for group BJ 20 (mixed 704
± 166 N, cohesive 648 ± 105 N, P = 0.185) or BJ 0 (mixed 834
± 101 N, cohesive 814 ± 105 N, P = .607). There was only one mixed
failure for group FE 20 (508 N) but this was well above the mean for
cohesive failures (273 ± 82 N). Similar for group FE 0 , mixed-mode
failures occurred at higher load-to-failure values than cohesive failures
(mixed 715 ± 112 N, cohesive 499 ± 146 N, P < .001).
The SEM images of one sample from each group are shown in
Figure 5 to Figure 8C. The SEM images from the BJ 20 group showed
a mixed failure mode (Figure 5). The SEM analysis showed crack prop-
agation that spread from the ceramic-resin interface layer to the
external surface below the loading zone.
F I G U R E 5 SEM image (palatal view) from the BJ 20 group
The SEM images from the BJ 0 group also showed a mixed fail-
showing the fractured veneer Wake hackles and crack propagation.
BJ, butt joint; dcp, direction of crack propagation; O, origin of ure mode (Figures 6A and 4B,C). Origin of failure was detected at the
fracture; SEM, scanning electron microscope ceramic-resin interface under the loading zone. On right lateral

F I G U R E 6 A, SEM image (incisal view) from the BJ 0 group showing the fractured veneer crack initiation, B, SEM image (palatal view) from
the BJ 0 group showing the fractured veneer cone cracks, C,. SEM image (buccal view) from the BJ 0 group showing the fractured veneer crack
propagation. BJ, butt joint; dcp, direction of crack propagation; O, origin of fracture; SEM, scanning electron microscope

F I G U R E 7 A, SEM image (buccoincisal view) from the FE 20 group showing the fractured veneer crack initiation at cement layer, B, SEM
image (buccal view) from the FE 20 group showing the fractured veneer surface contact stress and crack propagation, C,. SEM image (buccal
view) from the FE 20 group showing the twist hackles and crack propagation. dcp, direction of crack propagation; FE, feathered edge; SEM,
scanning electron microscope
6 CHAI ET AL.

F I G U R E 8 A, SEM image (incisal view) from the FE 0 group showing the fractured veneer loading zone, B, SEM image from the FE 0 group
showing the cone zone and crush zone, C,. SEM image from the FE 0 group showing the fractured veneer cone zone and crack propagation. dcp,
direction of crack propagation; FE, feathered edge; SEM, scanning electron microscope

aspects of the sample, secondary crack features such as crack line reinforcement fibers within a resin matrix.49 The elastic modulus of
(Figure 6A) and cone cracks were noted (Figure 6B). The origin of the the typodont tooth material is similar to the dentine layer. It does
crack and crack propagation is seen in Figure 6C. not replicate the anisotropic structure of a natural tooth. Bonding
The SEM images from the FE 20 group showed the origin of properties of a ceramic veneer to a typodont tooth and a natural
fracture at the ceramic-cement interface under the load contact area tooth would be different, but the comparison is beyond the scope of
(Figure 7A). Mixed failure modes were observed, with crack propaga- the current study. This study also does not take into account the
tion found from the incisal region toward the cervical region of the ageing process of ceramic in the oral environment, thermal, chemical,
veneer. Wake hackles (Figure 7A) and the fractured veneer surface and mechanical stresses, as well as microbiological insults from the
showing evidence of contact stress and crack propagation (Figure 7B). oral biofilm. One of the limitations of this in vitro study is the static
Twist hackles were also observed at the buccal surface of the veneer load-to-failure testing methodology, which does not represent cyclic
(Figure 7C). loading conditions49 or correlate to the clinical failure mecha-
The stereomicroscopic and SEM analyses of the sample from the nism.46,49,50 However, fatigue testing has been shown to result in
FE 0 group showed a mixed failure mode. A crush zone with irregular the sliding of the loading device along the palatal contour of the
fracture planes was observed at the load contact area (Figure 8A). A teeth, resulting in inconsistent failures and early fracture of the abut-
crush zone was identified and cone zone was identified (Figure 8B). ment teeth.31,35 The static load-to-failure testing methodology pro-
The crack was found to propagate from the load contact area toward vides ease of test standardization, overview of fracture behavior of
the internal surface of the veneer (Figure 8C). the tooth-restoration complex, comparisons of material strength,
estimation of failure risk, data collection, and comparison between
studies.18,28,31,35,46,51 Under appropriate conditions, the failure
4 | DISCUSSION mode seen in clinically failed restorations can be reproduced in
in vitro studies.46,52 It has been used as preclinical evaluations for
This study included load-to-failure and fractographic failure analysis assessing and comparing materials, restoration design, and fabrica-
to investigate the fracture features and crack propagation of ceramic tion concepts.53
veneers under in vitro conditions. This in vitro study looked at two Another consideration is the loading angles used in the study.
different veneer designs under two different load angulations to eval- Although 45 loading angulation between incisal edge of veneer and
uate if there is a preferred design. Of the in vitro studies which have plunger orientation best simulates intraoral interincisal angle based on
29-31,37
compared the FE and BJ preparation designs previously, three well-established cephalometric measurements,54,55 on pretesting
studies reported no statistically significant difference in veneer load- attempts to do a 45 loading angulation excessive bending force was
29-31
to-failure values between the two preparation designs, while one placed on the typodont teeth. This excessive bending resulted in
study reported significantly higher load-to-failure values for the BJ breakage of the typodont tooth abutment before any veneer failures.
preparation design. 37
However, to date, simulation protocols intended Hence, a lesser loading angulation of 20 was adopted and the loading
to replicate clinical fractures of ceramic veneers have not been stan- angulations of 0 was applied to simulate an edge-to-edge occlusion.
48
dardized. This study included fractographic failure analysis to inves- This study highlights that the load-to-failure value of ceramic
tigate the characteristic fracture features and crack propagation of veneers is influenced by the incisal preparation design and loading
ceramic veneers under in vitro conditions as reference for future angulation, with the differences between the two preparation groups
studies. increasing when an inclined load is applied. This can be partly
The quality of the tooth samples, tooth preparation, veneer fab- explained by the high flexural strength and fracture toughness of IPS
rication, and loading conditions were standardized in this study. e.max CAD material,56,57 and the thickness of ceramic material at inci-
Although the use of typodont teeth eliminates the disadvantages of sal edge. When an inclined load is applied to the FE 20 group, the
using a natural tooth, it comes with other limitations. The typodont ceramic veneers failed at a significantly lower load-to-failure value
tooth is essentially a homogenous solid structure consisting of compared to the other groups.
CHAI ET AL. 7

The load-to-failure values obtained in this study ranged between 1 Incisal preparation design and loading angulation have a significant
280 and 834 N, and were comparable to previous in vitro load-to- effect on load-to-failure values.
failure studies, which had also investigated lithium disilicate ceramic 2 The BJ group had higher load-to-failure values than the FE group.
veneers.39 The load-to-failure values reported in this study are much 3 FE 20 group exhibited the lowest load-to-failure value compared
greater than the reported maximum bite force in the anterior region to all other groups tested.
of between 100 and 200 N.58 This comparison between the load-to- 4 The failure mode of e.max CAD veneers is dependent on the incisal
failure values and maximum bite force provides reassurance that teeth preparation design and the loading angulation.
employed in the present study are within appropriate parameters. 5 Fractographic surface analysis indicates that minimal thickness of
In this study, all four groups failed by a combination of cohesive incisal ceramic increases the risk of ceramic fracture.
and mixed failure modes. The weakest group, FE 20 , had a higher
percentage of cohesive failures than mixed failures. The failure mode ACKNOWLEDG MENTS
in group FE 20 was unlike the BJ 20 , BJ 0 , and FE 0 groups which The authors thank the New Zealand Dental Association Research
exhibited both mixed and cohesive fracture modes. This difference in Foundation for research grant (#RF8.05 2016), Centre of Materials
fracture pattern could be caused by the reduced thickness of ceramic Science and Technology, University of Otago for use of their research
at the load contact area and angulation of contact stresses inherent to facilities, Ms Liz Girvan, technical staff, Scanning Electron Microscopy,
the FE preparation.47 This is supported by the results of an in vitro University of Otago, for technical support, and Mr Andrew Gray, bio-
study that showed that the thickness of monolithic lithium disilicate statistician, Department of Preventive and Social Medicine, University
ceramic crowns was responsible for such failures, and they rec- of Otago, for his assistance with the statistical analysis.
59
ommended a minimum thickness of 1.0 to 1.2 mm.
The mean load-to-failure values of both BJ 20 and FE 0 groups DISCLOS URE STATEMENT
are statistically significantly higher than the FE 20 group. As well as The authors do not have any financial interest in the companies
the failure mode of FE 20 group being predominantly cohesive, the whose materials are included in this article.

SEM analysis of the FE 20 group showed the origin of failure to be
located at the ceramic-resin interface under the loading zone. These OR CID
results indicate there is a significantly greater risk of ceramic fracture Vincent Bennani https://orcid.org/0000-0002-0878-5207
in veneers with FE incisal preparation design compared to the BJ
design and also that as the loading angulation increases the load-to- RE FE RE NCE S
failure value decreases. Therefore, it would be reasonable to infer that 1. Beier US, Kapferer I, Burtscher D, Dumfahrt H. Clinical performance
veneers with a FE incisal preparation design would have a higher risk of porcelain laminate veneers for up to 20 years. Int J Prosthodont.
2012;25:79-85.
of complications and/or failures compared to a BJ design, particularly
2. Layton DM, Clarke M, Walton TR. A systematic review and meta-
in patients with an incisal angle of 20 or greater. analysis of the survival of feldspathic porcelain veneers over 5 and
The SEM analysis of the samples from the BJ 20 and BJ 0 10 years. Int J Prosthodont. 2012;25:590-603.
groups suggested that fracture origin was at the ceramic-resin inter- 3. Friedman MJ. A 15-year review of porcelain veneer failure‑a
clinician's observations. Compend Contin Educ Dent. 1998;19:
face layer below the loading zone. The crack propagated traveled from
625-632.
the internal toward the external surface of the restoration, toward the 4. Welbury RR. A clinical study of a microfilled composite resin for labial
incisal and buccal directions, and compression curls were noted at the veneers. Int J Paediatr Dent. 1991;1:9-15.
external surface of ceramic veneers (Figures 5 and 6C). These finds 5. Rucker LM, Richter W, MacEntee M, Richardson A. Porcelain and
resin veneers clinically evaluated: 2-year results. J Am Dent Assoc.
were similar to findings reported in other fractographic failure ana-
1990;121:594-596.
lyses.52,60 In contrast, SEM images of the fractured samples from the 6. Meijering AC, Creugers NH, Roeters FJ, Mulder J. Survival of three
FE 0 group showed a crush zone forming under the load contact area types of veneer restorations in a clinical trial: a 2.5-year interim evalu-
(Figure 8A-C). ation. J Dent. 1998;26:563-568.
7. Morimoto S, Albanesi R, Sesma N, Agra C, Braga M. Main clinical out-
Other than load-to-failure values, it is vital to consider the stress
comes of feldspathic porcelain and glass-ceramic laminate veneers: a
distribution of the veneer-tooth system under various loading condi- systematic review and meta-analysis of survival and complication
tions. The BJ group showed consistently more favorable stress distri- rates. Int J Prosthodont. 2016;29:38-49.
bution at all loading angulations. Hence, the BJ incisal preparation 8. Reid JS, Murray MC, Power SM. Porcelain veneers- a sfour-year
follow-up. J Res Dent. 1988;4(60):62-66.
design is preferred to FE preparation design due to more favorable
9. Christensen GJ, Christensen RP. Clinical observations of porcelain
stress distribution and a higher load-to-failure value.
veneers: a three-year report. J Esthet Dent. 1991;3:174-179.
10. Nordbø H, Rygh-Thoresen N, Henaug T. Clinical performance of por-
celain laminate veneers without incisal overlapping: 3-year results.
5 | C O N CL U S I O N S J Dent. 1994;22:342-345.
11. Walls AW. The use of adhesively retained all-porcelain veneers dur-
ing the management of fractured and worn anterior teeth. Part 2:
Within the limitations of this study, the following conclusions can be clinical results after 5 years of follow-up. Br Dent J. 1995;178:
drawn: 337-340.
8 CHAI ET AL.

12. Dumfahrt H, Schaffer H. Porcelain laminate veneers. A retrospective 36. Schmidt KK, Chiayabutr Y, Phillips KM, Kois JC. Influence of prepara-
evaluation after 1 to 10 years of service. Part II: clinical results. Int J tion design and existing condition of tooth structure on load to failure
Prosthodont. 2000;13:9-18. of ceramic laminate veneers. J Prosthet Dent. 2011;105:374-382.
13. Peumans M, De Munck J, Fieuws S, Lambrechts P, Vanherle G, Van 37. Jankar AS, Kale Y, Kangane S, Ambekar A, Sinha M, Chaware S. Com-
Meerbeek B. A prospective ten-year clinical trial of porcelain veneers. parative evaluation of fracture resistance of ceramic veneer with
J Adhes Dent. 2004;6:65-76. three different incisal design preparations‑an in vitro study. J Int Oral
14. Smales RJ, Etemadi S. Long-term survival of porcelain laminate Health. 2014;6:48-54.
veneers using two preparation designs: a retrospective study. Int J 38. Stappert CF, Ozden U, Gerds T, Strub JR. Longevity and failure load
Prosthodont. 2004;17:323-326. of ceramic veneers with different preparation designs after exposure
15. Layton D, Walton T. An up to 16-year prospective study of 304 por- to masticatory simulation. J Prosthet Dent. 2005;94:132-139.
celain veneers. Int J Prosthodont. 2007;20:389-396. 39. Bergoli CD, Meira JB, Valandro LF, Bottino MA. Survival rate, load to
16. Granell-Ruiz M, Fons-Font A, Labaig-Rueda C, Martinez-Gonzalez A, fracture, and finite element analysis of incisors and canines restored
Roman-Rodriguez JL, Sola-Ruiz MF. A clinical longitudinal study with ceramic veneers having varied preparation design. Oper Dent.
323 porcelain laminate veneers. Period of study from 3 to 11 years. 2014;39:530-540.
Med Oral Patol Oral Cir Bucal. 2010;15:e531-e537. 40. Chaiyabutr Y, Phillips KM, Ma PS, Chitswe K. Comparison of load-
17. Peumans M, Van Meerbeek B, Lambrechts P, Vanherle G. Porcelain fatigue testing of ceramic veneers with two different preparation
veneers: a review of the literature. J Dent. 2000;28:163-177. designs. Int J Prosthodont. 2009;22:573-575.
18. Magne P, Kwon KR, Belser UC, Hodges JS, Douglas WH. Crack pro- 41. Zarone F, Apicella D, Sorrentino R, Ferro V, Aversa R, Apicella A. Influ-
pensity of porcelain laminate veneers: a simulated operatory evalua- ence of tooth preparation design on the stress distribution in maxillary
tion. J Prosthet Dent. 1999;81:327-334. central incisors restored by means of alumina porcelain veneers: a 3D-
19. Magne P, Versluis A, Douglas WH. Effect of luting composite shrink- finite element analysis. Dent Mater. 2005;21:1178-1188.
age and thermal loads on the stress distribution in porcelain laminate 42. Hahn P, Gustav M, Hellwig E. An in vitro assessment of the strength
veneers. J Prosthet Dent. 1999;81:335-344. of porcelain veneers dependent on tooth preparation. J Oral Rehabil.
20. Jordan A. Clinical aspects of porcelain laminate veneers: consider- 2000;27:1024-1029.
ations in treatment planning and preparation design. J Calif Dent 43. Nakamura T, Dei N, Kojima T, Wakabayashi K. Marginal and internal
Assoc. 2015;43:199-202. fit of Cerec 3 CAD/CAM all-ceramic crowns. Int J Prosthodont. 2003;
21. Radz GM. Minimum thickness anterior porcelain restorations. Dent 16:244-248.
Clin N Am. 2011;55:353-370. 44. Pallis K, Griggs JA, Woody RD, Guillen GE, Miller AW. Fracture resis-
22. Pini NP, Aguiar FH, Lima DA, Lovadino JR, Terada RS, Pascotto RC. tance of three all-ceramic restorative systems for posterior applica-
Advances in dental veneers: materials, applications, and techniques. tions. J Prosthet Dent. 2004;91:561-569.
Clin Cosmet Investig Dent. 2012;4:9-16. 45. Yucel MT, Yondem I, Aykent F, Eraslan O. Influence of the supporting
23. Walls AW, Steele JG, Wassell RW. Crowns and other extra-coronal die structures on the fracture strength of all-ceramic materials. Clin
restorations: porcelain laminate veneers. Br Dent J. 2002;193:73-82. Oral Investig. 2012;16:1105-1110.
24. Quinn F, McConnell RJ, Byrne D. Porcelain laminates: a review. Br 46. Kelly JR. Clinically relevant approach to failure testing of all-ceramic
Dent J. 1986;161:61-65. restorations. J Prosthet Dent. 1999;81:652-661.
25. Christensen GJ. Have porcelain veneers arrived? J Am Dent Assoc. 47. Quinn GD. Fractography of Ceramics and Glasses. St. Gaithersburg,
1991;122:81. MD: National Institute of Standards and Technology; 2007.
26. Brunton PA, Wilson NH. Preparations for porcelain laminate veneers 48. Kelly JR, Benetti P, Rungruanganunt P, Bona AD. The slippery slope:
in general dental practice. Br Dent J. 1998;184:553-556. critical perspectives on in vitro research methodologies. Dent Mater.
27. Burns PB, Rohrich RJ, Chung KC. The levels of evidence and their 2012;28:41-51.
role in evidence-based medicine. Plast Reconstr Surg. 2011;128: 49. He LH, Purton D, Swain M. A novel polymer infiltrated ceramic for
305-310. dental simulation. J Mater Sci Mater Med. 2011;22:1639-1643.
28. da Costa DC, Coutinho M, de Sousa AS, Ennes JP. A meta-analysis of 50. Anusavice KJ, Kakar K, Ferree N. Which mechanical and physical test-
the most indicated preparation design for porcelain laminate veneers. ing methods are relevant for predicting the clinical performance of
J Adhes Dent. 2013;15:215-220. ceramic-based dental prostheses? Clin Oral Implants Res. 2007;18
29. Hui KK, Williams B, Davis EH, Holt RD. A comparative assessment of (Suppl 3):218-231.
the strengths of porcelain veneers for incisor teeth dependent on 51. Nawafleh N, Hatamleh M, Elshiyab S, Mack F. Lithium disilicate resto-
their design characteristics. Br Dent J. 1991;171:51-55. rations fatigue testing parameters: a systematic review. J Prosthodont.
30. Wall JG, Reisbick MH, Johnston WM. Incisal-edge strength of porce- 2016;25:116-126.
lain laminate veneers restoring mandibular incisors. Int J Prosthodont. 52. Øilo M, Kvam K, Tibballs JE, Gjerdet NR. Clinically relevant fracture
1992;5:441-446. testing of all-ceramic crowns. Dent Mater. 2013;29:815-823.
31. Castelnuovo J, Tjan AH, Phillips K, Nicholls JI, Kois JC. Fracture load 53. Cardelli P, Serafini N, Sinjari B, Murmura G, Ozcan M. Reliability anal-
and mode of failure of ceramic veneers with different preparations. ysis of lithium disilicate crowns: effect of veneering and milling pro-
J Prosthet Dent. 2000;83:171-180. duction workflow. J Prosthodont. 2016;25:623-628.
32. Zarone F, Epifania E, Leone G, Sorrentino R, Ferrari M. Dynamometric 54. Schudy FF. Cant of occlusal plane and axial inclinations of teeth. Angle
assessment of the mechanical resistance of porcelain veneers related Orthod. 1963;33:69-82.
to tooth preparation: a comparison between two techniques. 55. Thulasidas S, Givan DA, Lemons JE, O'Neal SJ, Ramp LC, Liu PR. Influ-
J Prosthet Dent. 2006;95:354-363. ence of implant angulation on the fracture resistance of zirconia abut-
33. Chun YH, Raffelt C, Pfeiffer H, et al. Restoring strength of incisors ments. J Prosthodont. 2015;24:127-135.
with veneers and full ceramic crowns. J Adhes Dent. 2010;12:45-54. 56. Giordano R. Materials for chairside CAD/CAM-produced restora-
34. D'Arcangelo C, De Angelis F, Vadini M, D'Amario M, Caputi S. Frac- tions. J Am Dent Assoc. 2006;137(Suppl):14S-21S.
ture resistance and deflection of pulpless anterior teeth restored with 57. Ritzberger C, Apel E, Höland W, Peschke A, Rheinberger VM. Proper-
composite or porcelain veneers. J Endod. 2010;36:153-156. ties and clinical application of three types of dental glass-ceramics
35. Akog lu B, Gemalmaz D. Fracture resistance of ceramic veneers with and ceramics for CAD-CAM technologies. Materials. 2010;3:3700-
different preparation designs. J Prosthodont. 2011;20:380-384. 3713.
CHAI ET AL. 9

58. Regalo SC, Santos CM, Vitti M, et al. Evaluation of molar and incisor
bite force in indigenous compared with white population in Brazil. How to cite this article: Chai SY, Bennani V, Aarts JM,
Arch Oral Biol. 2008;53:282-286.
Lyons K, Lowe B. Effect of incisal preparation design on load-
59. Yu T, Wang F, Liu Y, Wu T, Deng Z, Chen J. Fracture behaviors of
monolithic lithium disilicate ceramic crowns with different thick-
to-failure of ceramic veneers. J Esthet Restor Dent. 2020;1–9.
nesses. RSC Adv. 2017;7:25542-25548. https://doi.org/10.1111/jerd.12584
60. Zhao K, Wei YR, Pan Y, Zhang XP, Swain MV, Guess PC. Influence of
veneer and cyclic loading on failure behavior of lithium disilicate
glass-ceramic molar crowns. Dent Mater. 2014;30:164-171.