pyrig

No Co

ht
t fo
rP

by N
ub

Q ui
Marginal Adaptation and Fracture Resistance of lica
tio
n
te ot

n
Root-canal Treated Mandibular Molars with Intracoronal
ss en c e fo r

Restorations: Effect of Thermocycling and Mechanical

Loading
Thomas Hitza/Mutlu Özcanb/Till N. Göhringc

Purpose: This study evaluated the marginal adaptation, fracture modes, and loads to failure of different mesio-oc-
clusal-distal (MOD) restorations in root canal treated molars in vitro.
Materials and Methods: Forty mandibular first molars were randomly assigned to five groups (n = 8): UTR= un-
treated, RCT-AM= root-canal treated (RCT)+amalgam filling; RCT-COM= RCT+direct composite resin restoration; RCT-
FRC= RCT+composite resin restoration with two layers of multidirectional woven glass fibers; RCT-CER= RCT+ceramic
inlay. All teeth were subjected to thermocycling and mechanical loading (TCML) in a computer-controlled masticator
(1,200,000 loads, 49 N, 1.7 Hz, 3000 temperature cycles of 5°C to 50°C). Marginal adaptation was evaluated before
and after TCML by scanning electron microscopy at 200X at the tooth/restoration interfaces. After TCML, all specimens
were loaded to failure in a universal testing machine at 0.5 mm/min. Failure types were analyzed and scored (reparable:
modes 1 and 2; catastrophic: modes 3 to 5). Data were analyzed with ANOVA and Bonferroni correction (α= 0.05).
Results: Marginal adaptation decreased significantly after TCML in all groups. The highest decrease was observed in
RCT-FRC (before: 89.9 ± 2.2%, after TCML: 49.8 ± 11.9%) and the lowest in RCT-CER (before TCML: 90.3 ± 2.6% after
TCML 80.4 ± 9.0%). Loads to failure (in N) were in descending order as follows: UTR: 3048 ± 905; RCT-CER: 1853 ±
477; RCT-AM: 1447 ± 363; RCT-FRC: 1066 ± 306; RCT-COM 960 ± 228. While UTR resulted in statistically higher
loads to failure than those of all other groups (p < 0.05), RCT-CER showed significantly higher results than those of
RCT-COM (p < 0.05) and RCT-FRC (p < 0.05). The UTR group showed exclusively reparable failures in the form of either
superficial (mode 1) or complete cusp chipping (mode 2), but all restored teeth exhibited catastrophic failures involv-
ing fractures along or through the restoration (modes 3 to 5).
Conclusion: None of the restored teeth in any group were able to bear the same load level as the natural teeth. With
regard to marginal adaptation and fracture resistance, luted ceramic inlays were advantageous. The integration of
FRC in the restoration had no positive effect on the fracture resistance. FRC application showed a negative effect on
marginal adaptation of composite resin restorations.
Keywords: ceramic, fiber reinforced composite, fracture resistance, inlay, intracoronal restoration, marginal adapta-
tion, root canal treated teeth.

J Adhes Dent 2010; 12: 279-286. Submitted for publication: 19.05.08; accepted for publication: 15.05.09.
doi: 10.3290/j.jad.a17712

a Associate Dentist, Clinic for Fixed and Removable Prosthodontics and Dental
Material Science, Center for Dental and Oral Medicine, University of Zürich,
Switzerland.
b Professor, Clinic for Fixed and Removable Prosthodontics and Dental Material
Science, Center for Dental and Oral Medicine, University of Zurich, Switzer-
land.
c Senior Lecturer, Clinic of Preventive Dentistry, Periodontology and Cariology,
Center for Dental and Oral Medicine, University of Zürich, Switzerland.
Correspondence: PD Dr.med.dent. Till N. Göhring, Clinic of Preventive Den-
tistry, Periodontology and Cariology, Center for Dental and Oral Medicine, Uni-
versity of Zürich, Plattenstrasse 11, CH-8032 Zürich, Switzerland. Tel:
+41-44-634-34-70, Fax:+41-43-211-33-22. e-mail: till.goehring@zzmk.uzh.ch
A fter root canal treatment, the dentist is often faced
with a deep cavity and remaining fragile dental hard
tissues. After access cavity preparation, root-canal treat-
ment and post-space preparation, a cumulative loss of
stability and fracture resistance was observed in vitro.23
The loss of the pulp chamber opening during endodontic
access preparation diminishes flexural cusp strength.28
Furthermore, pulpless teeth may be loaded more heavily
than their vital counterparts before a pain response is trig-
gered,29 which may predispose fracture. According to the
cohort studies on survival of restorative materials, it ap-
pears that endodontically treated teeth are more prone to
loss of restoration and/or fracture than vital teeth.11,20,38
However, discrepant results have also been published.36

Vol 12, No 4, 2010 279


Hitz et al
Table 1 Experimental groups and the materials used in this study
Group n Adhesive Restorative material
UTR 8 – – –
RTC-AM 8 Syntac1 Amalgam (Dispersalloy2)
RTC-COM 8 Syntac1 Composite (Tetric Ceram1)
RTC-FRC 8 Syntac1 Composite (Tetric Ceram1)
RTC-CER 8 Syntac1 Ceramic (ProCad1)
1Ivoclar
Vivadent, Schaan, Liechtenstein; 2Dentsply, Konstanz, Germany.
UTR= untreated; RCT-AM= root-canal treated (RCT)+amalgam filling; RCT- COM= RCT+direct composite resin restoration; RCT-FRC=
RCT+composite resin restoration with two layers of multidirectional woven glass fibers; RCT-CER= RCT+ceramic inlay
There is still no consensus on the optimal restoration ap-
proach following root-canal treatment. A retrospective clini-
cal study indicated a six times higher long-term survival rate
for crowned root-canal treated teeth than for non-crowned
ones.1 Unfortunately, root-canal treatment followed by crown
restorations are time-consuming, expensive treatment op-
tions, and cannot be easily afforded by every patient. Hence,
immediately after root canal treatment, the tooth can be re-
stored with an intra-coronal filling instead of a crown, pro-
viding the patient is informed and consents to the tooth be-
ing crowned in the long term. Intracoronal amalgam restora-
tions, retained by macromechanically retentive prepara-
tions, can keep teeth in service for years or decades. How-
ever, their poor marginal sealing properties are described as
reasons for leakage.2 On the other hand, the clinical rele-
vance of marginal opening value and caries incidence is a
difficult issue to investigate clinically. Since the oral hygiene
strategies and habits of patients vary between individuals,
the consequences of large marginal gaps may not neces-
sarily be correlated with the caries incidence. This has long
been known from amalgam restorations.26 Although there is
currently no clinical evidence to support the concept that
marginal gaps lead to secondary caries, the presence of
marginal openings may still carry some clinical importance
where oral hygiene is less than ideal.8
With the advances in adhesive techniques, such restora-
tions can also be accomplished using tooth-colored materi-
als. However, especially in deep cavities found after en-
dodontic treatment, a large cavity configuration factor (C-fac-
tor) might amplify the problems with composite resin restora-
tions due to their shrinkage during polymerization.10 Incre-
mental restorative application techniques may reduce
stresses on cavity walls during polymerization, but their ap-
plication becomes more complicated with increasing cavity
depth.24 When the restored tooth is overloaded, the bulk ma-
terial might lead to a propagation of the fracture into the un-
derlying dental hard tissues, yielding catastrophic fracture.
The progression of an onset fracture may be stopped by the
integration of glass fibers that are able to spread energy and
change the direction of a fracture line. Hence, such fibers
could prevent a restoration from dramatic failure.17 The in-
tegration of polyethylene fibers was described to have a sig-
280
pyrig
No Co
ht
t fo
rP
by N
ub
Q ui
lica
Reinforcement Luting material tio
te ot n
n
ss e n c e
fo r
–
– –
– –
Glass-Fiber (Vectris Frame1) –
– Composite (Tetric1)
nificant positive effect on fracture resistance of direct com-
posite restorations in vitro.5 Ceramic restorations, on the
other hand, provide a higher modulus of elasticity that re-
duces plastic deformation of the tooth compared to com-
posite resin materials.28 In adhesively luted ceramic restora-
tions, polymerized composite resin is reduced significantly
to an approximately 50- to 100-μm wide luting space. This
reduces stress on cavity walls during polymerization.24 One
other restoration option for root-canal treated teeth is the
chairside CAD/CAM technique that allows manufacturing
intra-coronal restorations in one appointment. This ap-
proach may eliminate the risk of possible tooth fractures and
bacterial contamination of the root canal filling during the
provisional restoration phase.6
The most common reasons for failure of endodontically
treated teeth are secondary caries and fractures of the
restoration or dental hard tissues.13 It has also been re-
ported that marginal integrity of restorations in vitro is one
of many important factors which helps to predict their long-
term clinical outcome.12 Therefore, one objective of this
study was to evaluate the marginal adaptation of different
intracoronal restorations in human first mandibular molars
before and after thermomechanical loading. Additionally,
load to failure and modes of failure were also studied and
compared to intact human molars. The following hypotheses
were tested: a) marginal adaptation of all kinds of adhesive
restorations are comparable to one another but superior to
amalgam restorations, b) fracture resistance increases with
adhesive restoration procedures and is further improved by
incorporating glass-fiber layers.
MATERIALS AND METHODS
For this study, 40 extracted mandibular first molars of
comparable dimensions were selected by visual inspec-
tion, digital caliper measurement (CAPA 150, Tesa; Re-
nens, Switzerland) and radiographs (Digora, Soredex;
Helsinki, Finland). Thirty-two molars were randomly di-
vided into 4 experimental groups of 8 teeth each. Experi-
mental groups and the materials used in this study are
listed in Table 1. The remaining 8 caries-free mandibular
The Journal of Adhesive Dentistry

first molars were used as controls, ie, they were not pre-
pared at all (UTR). In all groups, the teeth had radiographi-
cally visible root canals, no cervical or root caries, and
similar dimensions measured at the cementoenamel junc-
tion. Teeth with extremely curved roots and wide or atypi-
cally shaped root canals were excluded. All teeth were
stored in 0.1 M thymol solution from extraction until treat-
ment. The patients had been informed before extraction
that their teeth would be used for research purposes. The
extraction had no influence on the individual treatment
plans of the patients. All teeth were cleaned with scalers,
nylon bristle brushes, and pumice. The roots of all teeth
were covered with an air-thinned 0.3-mm layer of polyvinyl-
siloxane (President light, surface activated, Coltène Whale-
dent; Altstätten, Switzerland) to simulate a periodontal
ligament. They were then centrally mounted on scanning
electron microscopy (SEM) specimen carriers (Balzers
Union; Balzers, Liechtenstein) with autopolymerized resin
(Paladur, Hereaus Kulzer; Hanau, Germany) with a center-
ing device (PPK, Zurich, Switzerland). The distance be-
tween the cementoenamel junction and the resin was 3
mm to simulate the biological width.
Root Canal Treatment
While the teeth in the UTR group were not endodontically
treated, in the remaining four groups, all teeth had under-
gone endodontic treatment. After access cavity prepara-
tion with a high-speed contra-angled handpiece (Sirius;
Micro-Mega; Besancon, France) and a diamond bur (FG
8514, Intensiv; Grancia, Switzerland), a step-down proce-
dure was performed using Gates Glidden burs (sizes 3 to
1; Maillefer; Ballaigues, Switzerland) in a low-speed con-
tra-angled handpiece (Micro-Mega) for the first 3 mm.
Nickel-titanium files (#20; NitiFlex, Maillefer) were in-
serted, and the working length was assessed with digital
radiographs (Digora). Root canal preparation was per-
formed with machine-driven rotary files (Profile .04,
Dentsply; Konstanz, Germany) and EDTA glide solution (RC
Prep Endodontic Lubricant, Stone Pharmaceuticals; Phila-
delphia, PA, USA). The master apical rotary instrument was
#35 in mesial and #45 in distal canals. After each file, the
canal was rinsed with sodium hypochlorite (1% wt). Follow-
ing root canal preparation, the canals were rinsed with
17% EDTA (Pulpdent; Watertown, MA, USA), dried with
paper points (Dr. Wild & Co; Basel, Switzerland) and obtu-
rated using cold lateral condensation with gutta-percha
points (#35) in mesial and in distal canals (#45) (Roeko;
Langenau, Germany), accessory point size A (Roeko), and
a sealer (AH-Plus, Dentsply; Konstanz, Germany). The ac-
cess cavities were covered with a temporary restorative
(Cavit, 3M ESPE; Seefeld, Germany) and the teeth were
stored in tap water at 36°C for at least 24 h.
Cavity Preparation
In all endondontically treated teeth, standardized non-
beveled mesio-occluso-distal (MOD) cavities were initially
prepared with water-cooled coarse diamond burs (100
μm; FG8614, Intensiv). One proximal cervical finishing line
was located 1 mm in root dentin and another in the
enamel 1.5 mm above the cementoenamel junction. All
Vol 12, No 4, 2010
pyri
No CoHitz et gal
ht
t fo
rP
by N
ub
Q ui
lica
tio
te otn
n
6 7 ss e n c e
fo r
5*
3 4
2*
1
RCT-COM & RCT-FRC
Fig 1 After interproximal walls of the Class II cavities were re-
stored with three increments, the resulting deep central Class I
cavities of RCT-COM were filled with five separately polymerized
increments (1,3,4,6,7). In RCT-FRC, multidirectional glass fiber
weavers (Vectris Frame: 2* and 5*) were adapted to the compos-
ite resin layer of 1 and 4 before polymerization.
remnants of the pulp chamber roof were removed. No lin-
ers or bases were used to establish adhesion to all inner
cavity surfaces. Cavities were finished with finishing burs
(25 μm; FG3614, Intensiv) and the proximal boxes were
finished with ultrasonic tips (PCS-Set and Master Piezon
400, EMS, Nyon, Switzerland). After finishing, all finishing
lines were located in the dental hard tissues and all cavi-
ties to be filled had similar dimensions (Fig 1).
Restorative Procedures
In order to mimic a realistic operative setting, the prepared
teeth were placed in a custom-made typodont model
(PPK), with adjustable adjacent teeth, during the restora-
tive procedures.
In groups RCT-AM, RCT-COM and RCT-FRC, enamel was
etched for 30 s with 35% phosphoric acid (Ultraetch, Ultra-
dent Products; South Jordan, UT, USA), then rinsed for 40 s
with water, and dried with oil-free air. The adhesive system
(Syntac Primer, Lot F51870, Syntac Adhesive, Lot F57527,
Heliobond, Lot F58115, Ivoclar Vivadent; Schaan, Liechten-
stein) was applied according to the manufacturer’s instruc-
tions and photopolymerized for 60 s (Optilux 500, standard
light tip). In RCT-AM, metallic matrix bands (Palodent Sec-
tional Matrix System, Dentsply Caulk; Milford, DE, USA) were
placed and fixed with wooden wedges. Amalgam (Disper-
salloy, Lot 010425, Dentsply; Konstanz, Germany) was trit-
urated and applied. Twenty-four hours after placement,
restorations were polished with rubber tips. In RCT-COM and
RCT-FRC, transparent plastic matrices (Hawe Neos; Bioggio,
Switzerland) and light-reflecting wedges (Luciwedge Adapt,
Hawe Neos) were placed for approximal restoration con-
touring. Proximal boxes were filled with three separately and
transdentally photopolymerized composite resin increments
281

Hitz et al
Table 2 Marginal adaptation before and after thermomechanical loading (TCML) in percentagefor
Group n Continuous margin (pooled)
Before TCML (Mean ± SD) Significance*
RCT-AM 8 55.0 ± 9.3
RCT-COM 8 92.1 ± 3.6
RCT-FRC 8 89.9 ± 2.2
RTC-CER 8 90.3 ± 2.6
*No statistically significant differences were found in groups with the same letters (p < 0.05). For group descriptions, see Table 1.
(Tetric Ceram A2, Lot F09945, Ivoclar Vivadent), resulting in
deep central Class I cavities with approximately 1-mm-thick
proximal walls. These central cavities were filled with one 1-
mm composite increment (Fig 1, layer 1), which covered the
cavity floor and was polymerized from the occlusal direction.
Subsequently, two separately polymerized oblique incre-
ments were placed (Fig 1, layer 3 and 4). Each increment
was polymerized for 60 s (Optilux 500, Turbo light tip, >
1000 MW/cm2, distance < 1mm). The result was a Class I
cavity with a composite resin floor at approximately the
same height as the cavity floor of a vital tooth. Two oblique
increments with transdental polymerization were used to
complete the restoration (Fig 1, layers 6 and 7). In RCT-FRC
before polymerization of the first (Fig 1, layer 1) and the third
central increment (Fig 1, layer 4), weavers (weavers consist
of four layers of loosely woven glass fiber mats in a liquid
polymer matrix) of multidirectional glass fibers (Vectris
Frame, Lot E94005, Ivoclar Vivadent) were cut to fit and
adapted to the unpolymerized composite before polymer-
ization (Fig 1, layers 2* and 5*). Hence, the first weaver was
located at the bottom, and the second was located at the for-
mer roof of the pulp chamber. The restoration was completed
with two oblique occlusal increments (Fig 1, layers 6 and 7).
In RCT-CER, cavities were coated with Cerec Liquid (Vita
Zahnfabrik; Bad Säckingen, Germany) and covered with a
thin layer of titanium dioxide (ProCad Contrast, Ivoclar Vi-
vadent). Optical impressions were made and inlays were
constructed (correlation mode, Cerec 3D, Sirona; Bensheim,
Germany). Inlays were milled from leucite reinforced glass
ceramic (ProCad 300 I14B3/B4, Lot 24540 Ivoclar Vi-
vadent). After try-in, subsurfaces were etched with 5% hy-
drofluoric acid (Vita Ceramic Etch, Vita Zahnfabrik) for 60 s.
Silane (Monobond S, Lot E34242, Ivoclar Vivadent) was ap-
plied, and after 60 s, solvent was evaporated with oil-free air.
Subsurfaces were then covered with a thin film of an adhe-
sive (Heliobond). Inlays were protected from light until in-
sertion. The cavities were cleaned carefully with water spray.
Cavity finishing lines were etched and the adhesive system
was applied as in all other groups. Care was taken to avoid
any pooling of the adhesive (Heliobond). After polymeriza-
tion for 60 s, composite resin (Tetric A2, Lot D00163, Ivoclar
Vivadent) was carefully adapted to cavity floors and walls
and the inlays were 7/8 inserted with the help of ultrasound
(SP-Tip, Piezon Master 400, EMS; Nyon, Switzerland). Sur-
282
pyrig
No Co
ht
t
by N
Pu
bli
Q ui
cat
ion
After TCML (Mean ± SD) Significance*
te ot
n
ss e n c e fo r
C 31.2 ± 5.3 D
A 69.1 ± 9.9 B
A 49.8 ± 11.9 C
A 80.4 ± 9.0 B
plus was removed with a probe and the inlays were inserted
to the end position with ultrasound. Small amounts of sur-
plus were not removed and composite was polymerized
transdentally from mesio-buccal, buccal, distobuccal, disto-
lingual, lingual, mesiolingual mesio-occlusal, occlusal and
disto-occlusal for 60 s each (Optilux 500, Turbo light tip).
Contouring, finishing, and polishing procedures were per-
formed under a stereomicroscope (Stemi 1000, Carl Zeiss
AG, Oberkochen, Germany) at 12X magnification. Finishing
diamond burs (15 μm and 8 μm), flexible abrasive disks
(Sof-Lex, 3M ESPE), and abrasive polishing brushes (Occlu-
brush, Hawe Neos) were used.
Impressions were made (President Light Body surface-ac-
tivated impression material; Coltène Whaledent; Altsätten,
Switzerland) and were filled with epoxy resin (Stycast 1266;
Emerson and Cuming; Westerlo, Belgium). These replicas
were compared with those made after thermomechanical
loading.
Forty palatal cusps from maxillary second molars were
used as antagonists. They were placed into the specimen
carriers, as described above, and randomly divided into 5
groups consisting of 8 specimens each. After restoration, all
specimens and antagonists were stored in tap water at 37°C
for two weeks.
Loading
The specimens were loaded mechanically in the center of
the occlusal surface in a computer-controlled masticator
(CoCoM 2, PPK) with 1.2 million vertical loads of 49 N at
1.7 Hz and 6000 simultaneous cycles of thermal stress at
temperatures from 5°C to 50°C. Each thermal cycle took
120 s. Immediately after thermomechanical loading, repli-
cas were made, sputtered with gold for 1 min (Sputter SCD
030, Balzers Union), and the entire restoration margin was
examined by scanning electron microscopy (SEM) at 15 kV
from a working distance of 17 mm (Amray 1810T, Amray;
Bedford, MA, USA). The researcher was calibrated in the
established procedures by an experienced operator, with
whom a re-calibration was performed for every group. For
the evaluation of marginal adaptation, the researcher was
blinded to the group membership of each specimen. For
this purpose, all specimens were numbered by a third per-
son who was not involved in this study.
The Journal of Adhesive Dentistry
 pyri
No CoHitz et gal

ht
t fo
Table 3 Loads to failure and failure characteristics rP

by N
ub

Q ui
lica
Group n Loads to failure (N) Failure characteristics tio
te ot n

n
First crack Total failure
Mean ± SD Signifi- Mean ± SD Signifi Minimum Maximum s s
fo r
Reparable Catastrophicen c e
cance cance*

UTR 8 1362 ± 820 no 3048 ± 905 A 1859 4298 8 –

RCT-AM 8 1000 ± 329 no 1447 ± 363 BC 848 2138 – 8
RCT-COM 8 855 ± 208 no 960 ± 228 C 620 1220 – 8
RCT-FRC 8 781 ± 196 no 1066 ± 304 C 658 1559 – 8
RTC-CER 8 1362 ± 428 no 1853 ± 477 B 1150 2758 – 8

*No statistically significant differences were found in groups with the same letters (p < 0.05).

Marginal adaptation was assessed for the following char-
acters and expressed as a percentage of the total margin
length examined: continuous margin (no gap, no interruption
to continuity), non-continuous “imperfect” margins (gap due
to adhesive or cohesive failure; fracture of the restorative
material or fracture of the dental hard tissue related to
restoration margins) at 200X magnification. These observa-
tions were made in occlusal, proximal, and cervical enamel
and cervical dentin-restoration areas to identify the most vul-
nerable areas of the restoration. After analysis of marginal
adaptation, specimens were placed in a custom-made car-
rier and loaded axially in a universal testing machine
(Schenk Trebel; Baden, Switzerland) with a 10-mm steel
sphere. The sphere had three-point occlusal contact. A 0.5-
mm-thick piece of tin foil between the steel sphere and
crown allowed a more equal load distribution and avoided
loading peaks on small surface areas (Fino tin layer 0.50, Fi-
no; Bad Bocklet, Germany). The crosshead speed was 0.5
mm/min. Teeth were transilluminated by cold light (Intralux
4000-1, Volpi; Schlieren, Switzerland) during loading to de-
tect visible fractures.
Additionally, sounds and visible load drops were record-
ed (SE 130, Ernst Roser; Bern, Switzerland). The load was
recorded in Newtons and mean values were calculated per
group. After load tests, the specimens were analyzed for the
failure mode: “reparable” tooth fracture that might clinical-
ly allow a new direct restoration, overlay or crown place-
ment, and “catastrophic” tooth/root fracture that might ne-
cessitate tooth extraction. Classification was based on a
two-examiner agreement. Failures were identified mainly in
5 modes: mode 1: superficial cusp chipping; mode 2: com-
plete cusp chipping; mode 3: fracture along the tooth
restoration interface; mode 4: fracture through the restora-
tion; mode 5: mixed fracture partly through the restoration
and partly along the tooth/restoration interface.
Statistical Analysis
After testing for normal distribution, the results of SEM
analysis were tested for statistical significance with factor-
ial (between groups) and repeated measures (before and
after thermomechanical loading) of analysis of variance
(ANOVA). Loads to failure were compared using one-way
ANOVA. Post-hoc comparisons were performed with t-tests.
The Bonferroni correction was applied for multiple testing.
For all statistical analyses, the level of significance was set
at 95%.
RESULTS
All teeth and all restorations survived thermomechanical
loading in the computer-controlled masticator without loss
of retention or visible fractures, and were used for analysis
of marginal adaptation and static load test.
Data of the separately assessed interface areas were
pooled due to the minimal and nonsignificant differences
within each group. Before thermomechanical loading, amal-
gam restorations had significantly lower percentages of con-
tinuous margins (55 ± 9.3%) than the adhesive restorations
(p < 0.05), which did not differ significantly among each oth-
er (89.9 ± 2.2 to 92.1 ± 3.6%) (p ≥ 0.05). After thermome-
chanical loading, marginal adaptation decreased signifi-
cantly in all groups ranging between 10% (RCT-CER) and
40% (RCT-FRC) (p < 0.05). Differences were significant be-
tween all groups (p < 0.05) except between RTC-COM and
RTC-CER (p < 0.05) (Table 2).
During the load-to-failure tests, first sounds were record-
ed between 781 ± 196 N (RTC-FRC) and 1362 ± 821 N
(UTR), which were not significantly different from each other
(p < 0.05). Major load drops and visible (by cold light illumi-
nation) damage occurred between 960 ± 227 N (RTC-COM)
and 3048 ± 905 N (UTR). The restored teeth fractured sig-
nificantly earlier than the untreated controls (Table 3). The
results for teeth restored with amalgam, composite, and
glass-fiber reinforced composite neither showed significant
differences nor was there a difference between ceramic and
amalgam (p > 0.05). The teeth restored with intracoronal
composite fillings with (p < 0.05) or without (p < 0.05) glass
fibers fractured significantly earlier than the teeth restored
with ceramic inlays.
The UTR group showed exclusively reparable failures in
the form of either superficial (mode 1) or complete cusp

Vol 12, No 4, 2010 283

 pyrig
Hitz et al
No Co

ht
t fo
Table 4 Types and frequency of failure modes of intracoronal restorations rP

by N
ub

Q ui
lica
UTR RCT-AM RCT-COM RCT-FRC RCT-CER tio
Fracture Modes Mode n Mode n Mode n Mode n Mode
te n
ot n

n
Repairable ss e n c e
fo r
Mode 1
Superficial cusp-
4
chipping
l* b*
Mode 2
Complete cusp-chipping 4

Catastrophic
Mode 3
Fracture along the tooth- 2 3 2 4
restoration interface

Mode 4
Fracture through the 0 0 2 2
restoration

Mode 5
Mixed fracture partly
through the restoration,
partly along the tooth-
restoration interface *m

6 5 4 2
*d

*b: buccal, *d: distal, *l: lingual, *m: mesial

chipping (mode 2). The restored teeth in all other groups with
different approaches and materials exhibited catastrophic
failures involving fractures along or through the restoration,
leading to root fractures (modes 3 to 5) (Table 4).
DISCUSSION
In general, the two main reasons for failure of teeth re-
stored after root canal treatment are coronal leakage and
fractures of the remaining hard tissues. In this study, both
events were analyzed under dynamic and static loads. In
an effort to approximate the clinical situation as much as
possible, natural teeth of similar dimensions were se-
lected and the periodontal ligament was simulated. Speci-
mens were loaded with a steel sphere that contacted the
functional and nonfunctional cusps in a position close to
that found clinically. In addition, a tin foil was placed be-
tween the load cell and the occlusal surface to avoid punc-
tual load peaks. However, all laboratory studies present
limitations, and the results should thus be interpreted
carefully. One limitation of the study is the direction of the
load application. The dynamic and static loads were ap-
plied only vertically. Therefore, lateral forces that may
occur clinically during clenching were not simulated. Fur-
thermore, the restorations were placed in a typodont
model with adjacent teeth.16 Environmental working condi-
tions might lead to more difficulties during restoration
placement in vivo. Although clinical long-term data are
missing, the high degree of standardization allows some
conclusions to be drawn from this study.
In the first part of this study, marginal adaptation of in-
tracoronal fillings in endodontically treated mandibular mo-
lars was analyzed before and after cyclic thermomechanical
loading. Even with the use of an adhesive bonding resin,
which was recommended to reduce microleakage signifi-
cantly at the tooth/amalgam interface,34 marginal adapta-
tion of amalgam fillings was poor. This finding led to the ac-
ceptance of the first hypothesis. Good and stress-resistant
marginal adaptation was achieved with ceramic inlays and
to a slightly lesser degree with incrementally placed direct
composite restorations. These findings are in accordance
with an earlier study.24 During SEM analysis, the operator
was blinded to the group membership of the specimens. A
complete blinding was not possible because the luting ce-
ment layer of the ceramic inlay group might be indicative,
which could be considered a limitation of such studies.
The incorporation of glass fibers resulted in significantly
poorer marginal adaptation after thermomechanical load-
ing. One reason for this might be that with the use of glass
fibers, two different materials having different physical prop-
erties were combined. Another reason might be a weaken-
ing effect on the restoration due to the difficult handling
properties of the glass fiber weavers used. These weavers
consist of four layers of loosely woven glass fiber mats in a
liquid polymer matrix. The weavers were nearly impossible
to adapt to polymerized composite resin. Hence, it was de-
cided to adapt the fibers to the more sticky unpolymerized

284 The Journal of Adhesive Dentistry


composite resin. This approach worked better but it was still
difficult to manipulate. During the fiber placement proce-
dure, it was not possible to avoid air and flaws in the glass
fiber and composite layer. This layer therefore might be
weaker and more elastic than the bulk composite. During oc-
clusal loading, this weak base might bend and result in high
stress at the interface between the restoration margin and
cavity finishing line. Degradation of the monomer matrix
during thermomechanical cycling could also affect the fiber
adhesion to the composite. Consequently, a significant de-
crease of marginal adaptation was observed. Although there
are some studies on marginal adaptation of glass fiber re-
inforced restorations, none of these studies evaluated glass
fiber reinforced inlays or direct fillings. Only full crowns,4
fixed dental prostheses, and inlay-3 or slot-inlay-retained
fixed dental prostheses15 were analyzed. Two studies with in-
tracoronal fiber-reinforced composite restorations did not
analyze marginal quality.5,7 The focus of these studies was
fracture resistance only. On the other hand, recent studies
that evaluated fracture resistance of endodontically treated
molars used posts33 or full coverage crowns.31 Therefore, a
direct comparison with this study could not be made. Nev-
ertheless, due to the less favorable marginal adaptation
findings in the group RCT-FRC but the nonsignificant findings
between RCT-COM and RCT-CER after thermomechanical cy-
cling, the hypothesis regarding the marginal quality of ad-
hesive restorations could only be partially accepted.
In the second part of our study, nonrestored natural mo-
lars and all intracoronally restored teeth were statically
loaded to failure. None of the restored teeth reached the
load to failure values of nonrestored molars. Controversy ex-
ists in literature as to how much load a restoration or a re-
stored tooth must be able to bear during oral function and
parafunction. Biting forces are reported to occur between
150 and 665 N.9 Forces vary from 216 to 847 N in the pos-
terior region, with the maximum in the first molar re-
gion.21,37,38 It has been reported that maximum biting force
can range between 244 and 1243 N.14 In view of these val-
ues, the tested restorations might withstand mastication
forces. However, the achieved values were about only one-
half (RCT-AM, RCT-CER) or one-third (RCT-COM, RCT-FRC) of
the values obtained with natural, untreated molars in our
test setup. Bindl et al6 calculated a critical limit of 1330 N
fracture load for a first molar. This value is in accordance
with first audible and visible cracks in the control group of
this study. Considering this value, only machined adhesive
ceramic inlays (RCT-CER) fulfilled the strength requirements
at least with respect to initial cracks. Interestingly, adhesive
techniques were not able to strengthen the endodontically
treated molars more than the amalgam restorations. Neither
mean load to failure nor failure patterns were positively in-
fluenced by adhesive restorations. This finding is in contrast
to some previous studies,22,30 but supports the results of
other studies.18,27,35 The variations in results could be at-
tributed to the differences in the testing parameters such as
crosshead speed, or the size and position of the load cell.25
In this study, a natural tooth substance acted as the antag-
onist during dynamic loading, whereas in all other studies,
a metal jig was used which may affect the crack propagation.
Vol 12, No 4, 2010
pyri
No CoHitz et gal
ht
t fo
The incorporation of glass fibers did not improve
r Pfracture
by N
resistance in our study. This finding opposes that of auprevi-
bli
Q ui
ous study, in which the use of polyethylene fibers signifi- cat
i
cantly improved fracture resistance of composite te restora- on
ot
n
ssof ce
fo r
tions.5 However, the values measured for total failure en fiber
reinforced composites (1066 ± 304 N) were comparable to
that study (944 ± 121 N). The different results for the unre-
inforced composite resin restorations may be attributed to
the different application protocols used. While in this study,
the incremental application technique might have compen-
sated for the high C-factor of the deep cavities. Belli et al5
used a bulk technique, which might have led to high stress-
es within the cavity. In another study, the positive influence
of the incremental composite resin layering led to no differ-
ence in fracture resistance irrespective of the use of fiber re-
inforcement.7 However, others have found little difference in
stress development between incremental and bulk place-
ment protocols for composite resins.19 Although the fiber
types, their adhesion to the composite, and architecture
may all affect the stress distribution, based on the marginal
adaptation as well as load-to-failure findings, no additional
benefit of fiber application could be detected on the out-
come, leading to rejection of the second hypothesis.
In terms of marginal adaptation and load-to-fracture find-
ings of this study, adhesively luted ceramic restorations
would seem to be the most recommended treatment option
among the others. However, the load bearing capacity of in-
tracoronally restored root-canal-treated molars was signifi-
cantly lower than that of intact molars, indicating that none
of the restoration approaches could restore the fracture re-
sistance of endodontically treated teeth to the level of nat-
ural teeth. These in vitro findings may support clinical stud-
ies in which significantly lower survival rates were reported
for teeth that were restored but not crowned after endodon-
tic treatment.1,32
Load-to-fracture findings may not be verified in a clinical
setup, but the failure modes found in this study need to be
confirmed by clinical observations in endodontically treated
and restored teeth. Especially the observed catastrophic
failures (modes 3 to 5) may not necessarily clinically need
extraction. However, these kinds of failures in this in vitro set-
up reached the cavity floor and propagated to the root, indi-
cating a failure type that may lead to extraction of the tooth.
In the opinion of the authors, the dramatic loss of sound
dental hard tissues during crown preparation for full cover-
age restorations should be avoided where possible. When in-
tracoronal restorations do not perform as well as expected,
the next step would be to cover all cusps with an overlay
restoration. This approach will be studied in a future inves-
tigation.
CONCLUSIONS
Within the limitations of this in vitro study on intracoronal
restorations in endondontically treated molars with mesio-
occlusal-distal Class II cavities, with respect to marginal
adaptation and loads-to-failure, adhesively luted ceramic
restorations showed the best performance. However, the
mean loads-to-failure for all restored teeth were signifi-
285

Hitz et al
cantly lower than the values for intact teeth. The incorpora-
tion of glass fibers in direct composite resin restorations
did not have a positive effect on fracture resistance. In-
stead, a negative effect on marginal adaptation was ob-
served.
REFERENCES
1. Aquilino SA, Caplan DJ. Relationship between crown placement and the
survival of endodontically treated teeth. J Prosthet Dent 2002;87:256-263.
2. Bauer JG, Henson JL. Microleakage: a measure of the performance of di-
rect filling materials. Oper Dent 1984;9:2-9.
3. Behr M, Rosentritt M, Leibrock A, Schneider-Feyrer S, Handel G. In-vitro
study of fracture strength an marjinal adaptation of fibre-reinforced adhe-
sive fixed partial dentures. J Dent 1999;27:163-168.
4. Behr M, Rosentritt M, Sikora MI, Karl P, Handel G. Marginal adaptation and
fracture resistance of adhesively luted glass fibre-composite reinforced
molar crowns with different inner crown surfaces. J Dent 2003;31:503-
508.
5. Belli S, Erdemir A, Ozcopur M, Eskitascioglu G. The effect of fibre insertion
on fracture resistance of root filled molar teeth with MOD preparations re-
stored with composite. Int Endod J 2005;38:73-80.
6. Bindl A, Richter B, Mörmann WH. Survival of ceramic computer-aided de-
sign/manufacturing crowns bonded to preparations with reduced macrore-
tention geometry. Int J Prosthodont 2005;18:219-224.
7. Cobankara FK, Unlu N, Cetin AR, Ozkan HB. The effect of different restora-
tion techniques on the fracture resistance of endodontically-treated mo-
lars. Oper Dent 2008;33:526-533.
8. Cömleko l u ME, Dündar M, Özcan M, Güngör MA, Gökçe B, Artunç C. Influ-
ence of cervical finish line type on marginal adaptation of zirconia ceramic
crowns. Oper Dent 2009;5:1-5.
9. Craig RG. Restorative Dental Materials, ed 10. St. Louis: Mosby, 1997.
10. Feilzer AJ, De Gee AJ, Davidson CL. Setting stress in composite resin in re-
lation to configuration of the restoration. J Dent Res 1987;66:1636-1639.
11. Fennis WM, Kuijs RH, Kreulen CM, Roeters FJ, Creugers NH, Burgersdijk
RC. A survey of cusp fractures in a population of general dental practices.
Int J Prosthodont 2002;15:559-563.
12. Frankenberger R, Krämer N, Lohbauer U, Nikolaenko SA, Reich SM. Mar-
ginal integrity: is the clinical performance of bonded restorations pre-
dictable in vitro? J Adhes Dent 2007;9:107-116. Erratum in: J Adhes Dent
2007;9:546.
11. Gher ME Jr, Dunlap RM, Anderson MH, Kuhl LV. Clinical survey of fractured
teeth. J Am Dent Assoc 1987;114:174-177.
12. Gibbs CH, Anusavice KJ, Young HM, Jones JS, Esquivel-Upshaw JF. Maxi-
mum clenching force of patients with moderate loss of posterior tooth sup-
port: a pilot study. J Prosthet Dent 2002;88:498-502.
13. Göhring TN, Peters OA, Lutz F. Marginal adaptation of inlay-retained adhe-
sive fixed partial dentures after mechanical and thermal stress: an in vitro
study. J Prosthet Dent 2001;86:81-92.
14. Göhring TN, Schönenberger KA, Lutz F. Potential of restorative systems
with simplified adhesives: quantitative analysis of wear and marginal
adaptation in vitro. Am J Dent 2003;16:275-282.
15. Göhring TN, Zappini G, Mayer J, Zehnder M. Glass-fiber frameworks for
fixed partial dentures: laser-interferometrical in vitro analysis. Quintes-
sence Int 2004;35:668-675.
16. Joynt RB, Wieczkowski G Jr, Klockowski R, Davis EL. Effects of composite
restorations on resistance to cuspal fracture in posterior teeth. J Prosthet
Dent 1987;57:431-435.
17. Kuijs RH, Fennis WM, Kreulen CM, Barink M, Verdonschot N. Does layering
minimize shrinkage stresses in composite restorations? J Dent Res
2003;82:967-971.
18. Hansen EK, Asmussen E. In vivo fractures of endodontically treated poste-
rior teeth restored with enamel-bonded resin. Endod Dent Traumatol
1990;6:218-225.
286
pyrig
No Co
ht
19.
t foof dentition. Acta
Helkimo E, Carlsson GE, Helkimo M. Bite force and state
rP
by N
Odontol Scand 1977;35:297-303.
ub resis-
Q ui
20. Hürmüzlü F, Kiremitçi A, Serper A, Altundaflar E, Siso SH. Fracture lica
tance of endodontically treated premolars restored with ormocer and pack-tio
able composite. J Endod 2003;29:838-840. te ot n
n
21. Lang H, Korkmaz Y, Schneider K, Raab WH. Impact of endodontic
fo r e
ss e n ctreat-
ments on the rigidity of the root. J Dent Res 2006;85:364-368.
22. Lutz F, Krejci I, Barbakow F. Quality and durability of marginal adaptation in
bonded composite restorations. Dent Mater 1991;7:107-113.
23. Naumann M, Metzdorf G, Fokkinga W, Watzke R, Sterzenbach G, Bayne S,
Rosentritt M. Influence of test parameters on in vitro fracture resistance of
post-endodontic restorations: a structured review. J Oral Rehabil
2009;36:299-312.
24. Özer L, Thylstrup A. What is known about caries in relation to restorations
as a reason for replacement? A review. Advances in Dent Res 1995;9:394-
402.
25. Qualtrough AJ, Cawte SG, Wilson NH. Influence of different transitional
restorations on the fracture resistance of premolar teeth. Oper Dent
2001;26:267-272.
26. Panitvisai P, Messer HH. Cuspal deflection in molars in relation to en-
dodontic and restorative procedures. J Endod 1995;21:57-61.
27. Randow K, Glantz PO. On cantilever loading of vital and non-vital teeth. An
experimental clinical study. Acta Odontol Scand 1986;44:271-277.
28. Reeh ES, Douglas WH, Messer HH. Stiffness of endodontically-treated
teeth related to restoration technique. J Dent Res 1989;68:1540-1544.
29. Salameh Z, Ounsi HF, Aboushelib MN, Sadig W, Ferrari M. Fracture resis-
tance and failure patterns of endodontically treated mandibular molars
with and without glass fiber post in combination with a zirconia-ceramic
crown. J Dent 2008;36:513-519.
30. Sorensen JA, Martinoff JT. Intracoronal reinforcement and coronal cover-
age: a study of endodontically treated teeth. J Prosthet Dent 1984;51:780-
784.
31. Sorrentino R, Monticelli F, Goracci C, Zarone F, Tay FR, García-Godoy F, Fer-
rari M. Effect of post-retained composite restorations and amount of coro-
nal residual structure on the fracture resistance of endodontically-treated
teeth. Am J Dent 2007;20:269-274.
32. Staninec M, Holt M. Bonding of amalgam to tooth structure: tensile adhe-
sion and microleakage tests. J Prosthet Dent 1988;59:397-402.
33. Steele A, Johnson BR. In vitro fracture strength of endodontically treated
premolars. J Endod 1999;25:6-8.
34. Valderhaug J, Jokstad A, Ambjørnsen E, Norheim PW. Assessment of the
periapical and clinical status of crowned teeth over 25 years. J Dent
1997;25:97-105.
25. Waltimo A, Kemppainen P, Könönen M. Maximal contraction force and en-
durance of human jaw-closing muscles in isometric clenching. Scand J
Dent Res 1993;101:416-421.
26. Waltimo A, Könönen M. A novel bite force recorder and maximal isometric
bite force values for healthy young adults. Scand J Dent Res 1993;
101:171-175.
Clinical relevance: For the restoration of mesio-oc-
clusal-distal Class II cavities in endondontically treated
molars, adhesively luted ceramic inlays could be rec-
ommended over amalgam fillings, direct composites, or
composite resin restorations with two layers of multidi-
rectional woven glass fibers, with respect to marginal
adaptation and loads-to-failure. Intact teeth, however,
showed significantly higher mean load-to-failure values
than those of all restored teeth.
The Journal of Adhesive Dentistry