Dental Materials (2005) 21, 445–453

www.intl.elsevierhealth.com/journals/dema

Microleakage of various cementing agents

for full cast crowns
Andree Piwowarczyka,*, Hans-Christoph Lauera, John A. Sorensenb
a
Department of Prosthetic Dentistry, School of Dentistry, Johann Wolfgang Goethe University,
Theodor-Stern-Kai 7, 60590 Frankfurt, Germany
b
Pacific Dental Institute, Portland, OR, USA

Received 5 November 2003; received in revised form 1 June 2004; accepted 13 July 2004

KEYWORDS Summary Objectives. To evaluate microleakage and marginal gaps in full cast
Cementing agent; crown restorations bonded with six different types of cementing agents.
Microleakage; Methods. Sixty non-carious human premolars and molars were prepared in a
Marginal gap; standardized manner for full cast crown restorations. The mesial and distal margins
Full cast crowns were located in dentin, while the vestibular and palatal/lingual margins were located
in enamel. Crowns were made from a high-gold alloy using a standardized technique.
The specimens were randomized to six groups of cementing agents: one zinc-
phosphate cement (Harvard cement), one conventional glass–ionomer cement (Fuji
I), one resin-modified glass–ionomer cement (Fuji Plus), two standard resin cements
(RelyX ARC, Panavia F), and one self-adhesive universal resin cement (RelyX
Unicem). After 4 weeks of storage in distilled water at 37 8C, the specimens were
subjected to 5000 thermocycles ranging from 5 to 55 8C. Then, they were placed in a
silver nitrate solution, embedded in resin blocks, and vertically cut in buccolingual
and mesiodistal direction. Subsequently, the objects were evaluated for micro-
leakage and marginal gap using a high-resolution digital microscope camera.
Results. A number of inter-group differences were statistically significant. RelyX
Unicem showed the smallest degree of microleakage both in enamel and in dentin.
Panavia F und RelyX Unicem were associated with significantly larger marginal gaps
than all other cementing agents. No association was observed between microleakage
and marginal gap other than a weak direct correlation when using Harvard cement on
enamel.
Significance. The cementing agents investigated revealed different sealing
abilities. These differences were not associated with specific types of materials.
Q 2004 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

Introduction

In vitro studies of microleakage are an initial

* Corresponding author. Tel.: C49 69 6301 5640; fax: C49 69
screening method to assess the maximum theoreti-
6301 3711. cal loss of sealing ability in vivo [1]. The occurrence
E-mail address: piwowarczyk@t-online.de (A. Piwowarczyk). of microleakage along the interface has been
0109-5641/$ - see front matter Q 2004 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.dental.2004.07.009
446
related to pulpal problems [2,3], hypersensitivity,
and secondary caries [4], the latter being the most
common reason why restorations are replaced [5].
Few studies have dealt with microleakage in full
cast crowns based on different types of cementing
agents [6–10]. Moreover, the methodologies and
cementing agents used in these studies have been
too diverse to allow direct comparison of the data.
But it is possible to analyze different types of
commercially available cementing agents that can
be used for long-term cementation of fixed restor-
ations. These include zinc-phosphate cements,
conventional glass–ionomer cements, resin-modi-
fied glass–ionomer cements, standard resin
cements, and a recently developed material
described as self-adhesive universal resin cement.
The objective in developing this cement was to
combine the ease of handling offered by glass–
ionomer cements with the favorable mechanical
properties [11], attractive esthetics [12], and good
adhesion of resin cements [13].
A review of the literature by Raskin et al. [14]
showed that 62.5% out of 144 studies on micro-
leakage performed between 1992 and 1998 eval-
uated microleakage in Class V cavities, while only
4.3% evaluated microleakage and marginal gaps as
study parameters involved in crown restorations.
These latter include only a limited number of
studies [9,10] investigating the correlation between
microleakage and marginal gaps in full cast crowns;
no such correlation was found. However, the
significance of the marginal gaps of full cast
restorations and their clinical consequences remain
to be determined.
The objective of the present in vitro study was to
investigate microleakage and marginal gaps associ-
ated with cementing agents in full cast crowns
following artificial aging. Other questions to be
examined were (1) whether various types of
cementing agents produce the same sealing ability
for definite bonding of indirect restorations and (2)
in how far microleakage results differ depending on
whether the margins of the restoration are placed in
dentin or in enamel. A final objective was to
determine whether there was a connection between
microleakage and marginal gaps in full cast crowns.
Materials and Methods
Specimen preparation
A total of 60 freshly extracted non-carious perma-
nent human molars with fully developed roots were
selected for this study. All teeth were stored in
A. Piwowarczyk et al.
distilled water at room temperature immediately
after extraction. Calculus and residual periodontal
tissue were removed using a surgical knife, scaler,
and curette. Subsequently, the teeth were con-
servatively polished using a rotating brush and
pumice.
The preparations for the full cast crowns were
performed using a suitable design and a converging
angle of around 68 to achieve optimal retention and
resistance properties. The occlusal and axial sur-
faces were reduced by approximately 1.2 and
0.8 mm, respectively. The cervical preparation
margins were designed as circular chamfers using
torpedo-shaped diamond burs and water-cooling
(Gebr. Brasseler, Lemgo, Germany; No. 878,314,
size 012). The preparations were finished with the
help of magnifying loupes (magnification !3.5;
Zeiss, Oberkochen, Deutschland; serial #53-20)
using fine and extra-fine diamond burs (Gebr.
Brasseler, Lemgo, Germany; No. 8878.314 and
878EF.314, size 012). Subsequently, a final check
of the preparations was performed. The mesial and
distal margins were located in dentin, while the
vestibular and palatal/lingual margins were located
in enamel.
For the model dies, impressions of the prepared
teeth were taken with Impregum Penta (3M ESPE,
Seefeld, Germany) and poured with Type IV resin-
stabilized extra-hard stone (esthetic-base 300;
dentona AG, Dortmund, Germany; lot #51001742)
following the manufacturer’s mixing instructions.
The stone was allowed to set for 40 min and then
removed from the impression. The resultant stone
dies were trimmed and covered with stone hard-
ener (Classic Hardener Spacer clear; Kerr Lab, Belle
de St Claire, France) followed by two layers of die
spacer (Classic Cement Spacer blue; Kerr Lab, Belle
de St Claire, France) above the preparation margin.
The preparation margin was marked with a non-
graphite pencil under a stereomicroscope at !32
magnification (Zeiss, Oberkochen, Germany; Stemi
2000-C, serial #2004003753). Subsequently, red
cervical wax and blue modeling wax (YETI Dental-
produkte, Engen, Germany) were used to model full
crowns of 0.5 mm thickness. Four round reference
marks approximately 1.5 mm in diameter were
applied in the mesial, distal, vestibular and
lingual/palatal segments centrally around 3 mm
above the preparations located in enamel and
dentin. Five wax copings each were embedded in
a #3 muffle ring with phosphate-bonded investment
(Precibalite plus; Dentona AG, Dortmund,
Germany; lot #0501101). After setting, the muffle
was heated and a centifugal casting performed
using a Type IV high-gold casting alloy (Portadur P4;
Wieland Edelmetalle GmbH, Pforzheim, Germany;
Microleakage of various cementing agents for full cast crowns 447

relative constituents: Au 68.5%, Ag 12.0%, Cu 12.0%,
Pt 6.9%, Zn 0.5%, Ir 0.1%; lots #4492 and 4269).
The castings were divested, trimmed and seated
using well-established procedures routinely used by
dental laboratories for crown restorations. The fit
of the castings was checked with silicone (Fit
Checker; GC, Munich, Germany) and improved if
necessary. Any points visibly pressed into the
silicone were relieved using a small (0.08 mm in
diameter) rotary instrument. The marginal fit of all
castings was verified on the prepared residual teeth
under a stereomicroscope (!32 magnification) and
with an extra-fine probe (EXD5; Hu-Friedy, Chicago,
IL, USA). Subsequently, the interior surface of the
gold frameworks was sandblasted with Al2 O 3
(Hasenfratz, Grafing, Germany; average grain
size, 105 mm; pressure, 0.18 MPa; distance,
10 mm; duration, 10 s).
Our experimental set-up included one zinc-
phosphate cement (Harvard cement), one conven-
tional glass–ionomer cement (Fuji I), one resin-
modified glass–ionomer cement (Fuji Plus), two
standard dual-cure resin cements (RelyX ARC,
Panavia F), and one recently developed dual-
cure self-adhesive resin cement (RelyX Unicem)
(Table 1). All cementing agents were processed
at room temperature (23 8C) strictly following
the manufacturers’ instructions. Two cements
(GC Fuji Plus and RelyX Unicem) were supplied
in pre-measured capsules. Upon activation, these
materials were mechanically triturated with the
rotational mixing machine (CapMix; 3M ESPE,
Seefeld, Germany) for the time recommended by
the manufacturers (10 or 15 s). RelyX ARC was
supplied in pre-measured delivery systems. Pana-
via F was mixed on a mixing block, using a hard
plastic spatula, at a base-to-catalyst paste ratio of
1:1 for approximately 20 s. An oxygen-blocking gel
(Oxyguard II; Kuraray, Osaka, Japan; lot #00373A)
was applied for 3 min when Panavia F was used.
Fuji I was manually mixed at a powder-to-liquid
ratio of 1.8–1.0 for around 20 s. The powder-to-
liquid ratio of Harvard cement was determined by
weighing according to the manufacturer’s instruc-
tions using an analytical balance (G1 mg). Mixing
was performed on a cool slab, over a wide area,
to incorporate small increments of powder into
the liquid for approximately 90 s (Harvard
cement).
Before cementation, all 60 teeth were random-
ized along with their gold frameworks to six test
groups (nZ10). The interior surfaces of the crown
restorations and the prepared teeth were cleaned
with alcohol. All teeth, which were to receive the
resin-modified glass–ionomer cement or a standard
resin cement, were pretreated with dentin

Table 1 Description of cementing agents used in this study.

Materials Type Main compositiona Adhesive system Manufacturer
Harvard cement Zinc-phosphate P: zinc oxide, magnesia; No adhesive system Richter & Hoffmann,
(Batch No. powder cement L: phosphoric acid Berlin, Germany
2112498001, Batch
No. liquid
2111000013)
Fuji I (Batch No. Conventional P: polyacrylic acid, alu- No adhesive system GC Corp., Tokyo,
0001251) glass–ionomer mino-silicate glass; Japan
cement L: polyacrylic acid,
citric acid
Fuji Plus (Batch No. Resin-modified P: alumino-silicate Fuji Plus conditioner GC Corp., Tokyo,
0009214) glass–ionomer glass; L: HEMA, poly- Japan
cement acrylic acid, TEGDMA
RelyX ARC (Batch No. Resin cement Bis-GMA, TEGDMA, silica Single bond adhesive 3M ESPE, Seefeld,
CACA) and zirconium glass (Batch No. 4242) Germany
Panavia F (Batch No. Resin cement BPEDMA, MDP, DMA, ED primer (Batch No. Kuraray, Osaka,
base 00124A, Batch barium, boron and sili- 00108B, Batch No. Japan
No. catalyst 00046A) cium glass, NaF 00115B)
RelyX Unicem (Batch Self-adhesive Phosphoric acid metha- No adhesive system 3M ESPE, Seefeld,
No. 0001) universal resin crylates, dimethacry- Germany
cement lates, inorganic fillers,
fumed silica, initiators
P, powder; L, liquid. Bis-GMA, bisphenol-A diglycidyl ether dimethacrylate; BPEDMA, bisphenol-A polyethoxydimethacrylate; DMA,
aliphatic dimethacrlyate; HEMA, 2-hydroxyethylmethacrylat; MDP, 10-methacryloyloxy decyl dihydrogenphosphate; NaF, sodium
fluoride; TEGDMA, triethylene glycol dimethacrylate.
a
According to the information provided by the manufacturers.
448
adhesives as recommended by the manufacturers
(Table 1). Dentin adhesives were not used for the
zinc-phosphate cement, the conventional glass–
ionomer cement, and the self-adhesive resin
cement. Bonding was performed by loading the
cementing agent into the interior surface of the
restoration and applying finger pressure for 10 s.
Then the frameworks were axially loaded at a
constant weight of 58.8 N for 7 min [15]. Excess
cement was removed with a scaler; marginal fit was
checked both by visual inspection and with a probe.
All tooth-restoration specimens were stored in
distilled water at 37 8C for 4 weeks, then they were
subjected to 5000 thermocycles ranging from 5 to
55 8C (immersion time, 20 s; transfer time, 10 s).
Subsequently, the root surfaces of the restored
teeth were covered with two layers of nail varnish
(ending 2 mm below the crown margin) and sub-
jected to silver nitrate penetration [16]. Then
they were placed into an unimolar silver nitrate
solution (Crystal, Fisher Scientific, Fairfield, NJ,
USA) for 6 h, followed by thorough rinsing, storage
in a photochemical developer (Solutek Corporation,
Boston, MA, USA) for 12 h, and exposure to a 150-W
floodlamp for 6 h.
Next, the tooth-restoration specimens were
embedded in a transparent resin matrix (Buehler
Epoxide; Buehler, Lake Bluff, IL, USA), which was
allowed to harden at room temperature for 24 h.
Each resin block was cut twice in the buccolingual
and the mesiodistal direction along the previously
applied reference marks using a slow-speed diamond
saw (Isomet; Buehler Ltd, Evanston, IL, USA) with
water-cooling. In this way, each specimen featured
eight surfaces (four in enamel and four in dentin) for
analysis of microleakage and marginal gap. The cut
surfaces were once again placed under a 150-W
floodlamp for 5 min, such that all portions of the
silver nitrate penetration acquired a black color.
Evaluation of microleakage and marginal gap
using a high-resolution digital microscope
camera
The microleakage in the area of the tooth–cement
interface was defined as linear penetration of silver
nitrate starting from the restorative crown margins
[17] and was determined with a stereomicroscope
(475052-9901; Zeiss, Oberkochen, Germany) and
AxioCam HR digital microscope camera (Zeiss,
Oberkochen, Germany; software module: Axio
Vision 3.1). The images were taken at a resolution
of 1300!1030 pixels. A micrometer scale (474026;
Zeiss, Oberkochen, Germany) was placed diagonally
across the image for calibration. The maximum
A. Piwowarczyk et al.
potential calibration-related error due to the width
of mapped lines was G0.709 or G0.475% depending
on the magnification factor relative to the absolute
measured values. Depending on the magnification,
2.42 or 3.64 mm/pixel were obtained. Metric assess-
ment of distances was performed using micrometers
(mm) as units. A randomly selected image including
10 measurements was analyzed to determine the
potential mapping error due to the definition of
microleakage length. Given a measured length of
1 mm, a maximum mapping error of G0.01 mm was
obtained irrespective of the magnification factor.
Marginal gaps were measured as defined by
Holmes et al. [18] using the stereomicroscope
(4730129901; Zeiss, Oberkochen, Germany) and
digital microscope camera. The selected magnifi-
cation was based on 1.21 mm equaling one pixel.
Based on the absolute values measured, the
maximum calibration-related error was G0.154%.
A randomly selected image including 10 measure-
ments was analyzed to determine the potential
mapping error due to the definition of marginal gap
length. Given a length of 40 mm, a maximum
mapping error of G0.18 mm was obtained.
Statistical analysis
Descriptive representation of continuous variables
was based on mean values, SD, and minimum–
maximum values. In addition, a number of tests
(skewness test, kurtosis test, and omnibus test)
were used to analyze whether the frequency
distribution of the random sample differed signifi-
cantly from the normal distribution. Since, the bulk
of data was not characterized by a normal
distribution, a non-parametric Kruskal–Wallis test
was used to analyze differences in the various
groups of cementing agents, comparing the
obtained test parameters (Bonferroni-corrected p
values). A non-parametric Wilcoxon’s test was used
to evaluate the differences between enamel and
dentinal substrates. The selected level of statistical
significance was p!0.05. Spearman’s correlation
coefficient (R) was used to assess the correlation
between two continuous variables.
Results
Microleakage with preparation margins
in enamel
Fig. 1 illustrates the mean values and SD for the
microleakage findings obtained with the various
cementing agents. The smallest degree of
Microleakage of various cementing agents for full cast crowns 449

Figure 1 Mean values and SD for microleakage, based Figure 2 Mean values and SD for microleakage, based
on 40 measurements for each material, with preparation on 40 measurements for each material, with preparation
margins located in enamel. margins located in dentin.

microleakage was observed with the dual-cure self- the standard resin cement RelyX ARC and the self-
adhesive resin cement (RelyX Unicem; 0.70G adhesive resin cement RelyX Unicem (p!0.0001),
0.61 mm), followed by the conventional glass– and between the standard resin cement Panavia F
ionomer cement (Fuji I; 0.71G0.32 mm), and the and the self-adhesive resin cement RelyX Unicem
resin-modified glass–ionomer cement (Fuji Plus; (p!0.0001).
0.77G0.48 mm). The greatest degree of microleak-
age was observed with the zinc-phosphate (Har- Microleakage with preparation margins in
vard) cement (1.59G0.64 mm). The Kruskal–Wallis
enamel versus dentin
test revealed statistically significant differences
between the zinc-phosphate (Harvard) cement on
All test groups revealed significant differences in
one hand and all other agents (all p!0.01) with the
microleakage between enamel and dentin. Micro-
exception of standard resin cement (RelyX ARC) on
leakage was invariably more pronounced in dentin
the other. A significant difference was also estab-
than in enamel (Table 2).
lished between one standard resin cement (RelyX
ARC) and the self-adhesive resin cement (RelyX
Unicem) (pZ0.019). Marginal gap

The average marginal gap in the specimens ranged

Microleakage with preparation margins in from 47.70G17.92 mm for the zinc-phosphate
dentin (Harvard) cement up to 74.59G29.15 mm for the
self-adhesive resin cement (RelyX Unicem; Fig. 3).
Fig. 2 illustrates the microleakage findings with The difference between one of the standard resin
preparation margins located in dentin. The smal-
lest degree of microleakage was observed with
the self-adhesive resin cement (RelyX Unicem; Table 2 Comparative microleakage data (mean
values and SD) with preparation margins in enamel or
1.01G0.54 mm), followed by the resin-modified
dentin.
glass–ionomer cement (Fuji Plus; 1.39G0.49 mm),
and the conventional glass–ionomer cement (Fuji Enamel Dentin p value
I; 1.41G0.90 mm). Statistically significant differ- (mm) (mm)
ences were observed between the zinc-phosphate Harvard 1.59 (0.64) 2.01 (0.70) 0.0054
(Harvard) cement and both the conventional glass– cement
ionomer cement (Fuji I, pZ0.006) and the self- Fuji I 0.71 (0.32) 1.41 (0.90) 0.000003
adhesive resin cement (RelyX Unicem, p!0.0001), Fuji Plus 0.77 (0.48) 1.39 (0.49) 0.000008
RelyX ARC 1.11 (0.68) 2.19 (1.04) 0.0000001
between the conventional glass–ionomer cement
Panavia F 0.95 (0.59) 2.11 (1.35) 0.000001
(Fuji I) and the standard resin cement (RelyX ARC, RelyX Unicem 0.70 (0.61) 1.01 (0.54) 0.0098
pZ0.002), between the resin-modified glass–iono-
mer cement Fuji Plus and the standard Inter-group differences were evaluated by Wilcoxon’s test
and were statistically significant in all groups (p!0.05).
resin cement RelyX ARC (pZ0.0281), between
450
Figure 3 Mean values and SD for marginal gap based on
80 measurements for each material.
cements (Panavia F) and the self-adhesive resin
cement (RelyX Unicem) was not statistically signifi-
cant. Both, however, differed significantly from all
other cementing agents (all p!0.01).
Correlation between microleakage
and marginal gap
A weak direct correlation between microleakage
and marginal gap was observed in the enamel of
specimens in which the zinc-phosphate cement
(Harvard) had been used for bonding (correlation
coefficient, RZ0.41). No such correlation was
found in any of the other groups.
Discussion
It is well established that the type of cementing
agent used for bonding has a bearing on microleak-
age [17,19]. It is also known that the composition
and other special characteristics of cementing
agents (e.g. the setting properties and the dentin
adhesive used) determine the degree of leakage.
The smallest degree of microleakage, both in
enamel and in dentin, was obtained with the self-
adhesive resin cement (RelyX Unicem). Apparently,
therefore, this agent is capable of generating an
effective seal at the interfaces between restorative
alloy, cementing agent, and dental tissue without
requiring pretreatment of the prepared tooth
surfaces. The formulation of RelyX Unicem contains
specific multifunctional phosphoric-acid methacry-
lates able to interact with the tooth surface in
multiple ways, such as by forming complex com-
pounds with calcium ions or by different kinds
of physical interaction like hydrogen bonding or
dipole-to-dipole interactions. This variety of inter-
actions seems to enable RelyX Unicem to generate
A. Piwowarczyk et al.
self-adhesion to both enamel and dentin, resulting
in an effective seal of the tooth–cement interface
and hence in the lowest microleakage values in
enamel and dentin of all luting agents examined.
The two standard dual-cure resin cements (RelyX
ARC and Panavia F) were associated with a higher
degree of microleakage—both in enamel and in
dentin—than the conventional glass–ionomer
cement (Fuji I) and the resin-modified glass–
ionomer cement (Fuji Plus). In general, different
microleakage patterns may be caused by different
factors, depending on the type of cementing agent
used and the experimental conditions.
The greater leakage of the resin cements RelyX
ARC and Panavia F compared to the conventional
glass–ionomer cement Fuji I and the resin-modified
glass–ionomer Fuji Plus might be thus attributed to
polymerization shrinkage of the resin cements
[20,21], combined with the coefficients of thermal
expansion of the materials involved (i.e. tooth
substance, cement, metal crown), during aging
[22]. By contrast with the resin cements, conven-
tional glass–ionomer cements like Fuji I are con-
sidered to be dimensionally stable during setting,
and the hydrophilic formulation of resin-modified
glass–ionomer cements like Fuji Plus can compen-
sate for the initial setting contraction by sub-
sequent expansion due to water uptake. This
difference in chemical behavior might explain the
different microleakage results found for the two
standard resin cements, the conventional glass–
ionomer cement, and the resin-modified glass–
ionomer cement.
Microleakage was found to be most pronounced
with the zinc-phosphate cement applied on prep-
aration margins located in enamel. Zinc-phosphate
cement on dentin showed the third-largest micro-
leakage of all cementing agents. A similar trend was
described by a number of authors [6–10] who
compared several types of materials and observed
the greatest degree of microleakage in full cast
crowns bonded with zinc-phosphate cements. Poss-
ible reasons for these unfavorable results for zinc-
phosphate cements include their lack of micro-
mechanical and especially—by contrast with all
other cements investigated in this study—the
absence of chemical retention on dental tissue
[23] and their high solubility [24]. Comparisons
of the clinical performance with regard to micro-
leakage around full cast crowns after 6 months in
function showed that a zinc-phosphate cement
(Flecks; Mizzy, Cherry Hill, NJ, USA) showed
significantly higher values than a resin-modified
glass–ionomer cement (Infinity; Den Mat Corp.,
Santa Maria, CA, USA) [25]. It must be noted,
however, that Kydd et al. [26] have shown that
Microleakage of various cementing agents for full cast crowns
crown restorations bonded with zinc-phosphate
cement remained functional despite microleakage
for over 20 years, with none of the tooth–crown
interfaces revealing carious lesions. Other authors
concluded that the cement dissolving at the
marginal gap does not seem to have any serious
consequences and does not give rise to carious
penetration [27,28].
In the present study, microleakage occurred
exclusively at the cement–enamel/dentin inter-
face. This observation is confirmed by various
studies in vitro and in vivo [9,17,25,26,29,30].
Although no microleakage occurred at the metal–
cement interface, the use of metal primer should
be taken into consideration clinically when metal
restorations are cemented using polymer-based
cements, because there is a chance of a better
marginal seal when using metal primer [31,32].
However, this effect is mainly dependent on the
combination of alloy composition, metal surface
treatment [33], metal primer, and the cement used
[31,32].
Our analysis of microleakage in enamel versus
dentin revealed statistically significant differences
in all study groups, i.e. the degree of microleakage
was invariably higher with preparation margins in
dentin than in enamel. This result is in keeping with
the observations of other authors [8,9]. Although the
retention mechanism on both dentin and enamel is
micromechanical in nature, both substrates are
dissimilar in composition and structure. The bond
between the cementing agent and the dentinal hard
tissue is compromised by the tubular microstruc-
ture, the higher content of organic material, and the
intrinsic humidity of the dentinal substrate [34]. It
must be noted, however, that bonding properties
and tracer penetration may vary from location to
location, depending on the structure and compo-
sition of the enamel or dentinal substrate. Major
determinants include age, location, tissue health,
and dietary factors [35,36].
To be able to perform an immediate relative
comparison between the in vitro results of the study
groups, margins were examined only after autopo-
lymerization for all resin cements. One would have
to take into consideration, however, that light-
polymerization increases the bond between dual-
cure resin cements and dental restorative materials
[37]. Dual-cure resin cements should therefore be
polymerized with a polymerizing light, especially if
the crown margins are located supragingivally and
are accessible.
The mean results in terms of marginal gap on
storing the specimens in water for 4 weeks and
subjecting them to 5000 thermocycles were least
favorable for one standard dual-cure resin
451
cement (Panavia F) and for the self-adhesive
resin cement (RelyX Unicem). On the other hand,
the second-most favorable marginal gaps were
also obtained with resin cement (RelyX ARC).
Consequently, the marginal gap quality of a given
type of cementing agent may be co-determined
by its specific physicochemical properties. Gemal-
mez et al. [38] demonstrated that the marginal
gap of porcelain inlays was larger when a dual-
cure resin cement was used for bonding. White
and Kipnis [39], who investigated the effect of
five different luting agents on the marginal fit of
cast single-crown restorations bonded to
extracted premolars in vitro, observed no signifi-
cant marginal gap differences in the pre-bonding
stage (range 35.1–69.5 mm), but reported that
significant differences did emerge after cementa-
tion. The smallest gaps were observed with a
glass–ionomer cement (Ketac-Cem; 3 M ESPE,
Seefeld, Germany; 82.8G12.6 mm), the largest
gaps with a microfilled resin cement (Thin Film
Cement; Den Mat Corp., Santa Maria, CA, USA)
using oxalate dentin bonding (333.1G45.9 mm).
This finding may conceivably be due to resin
cements rapidly gaining viscosity in the process
of curing. For this reason, White et al. [8]
recommended that resin cements should be
applied swiftly and carefully in clinical practice,
and that indirect restorations should be inserted
with considerable pressure.
The present study also showed that bonding of
the restorations with the zinc-phosphate cement
(Harvard cement), the conventional glass–ionomer
cement (GC Fuji I), the resin-modified glass–
ionomer cement (GC Fuji Plus), and one of the
standard resin cements (RelyX ARC), involved no
significant marginal gap differences. White et al.
[29] demonstrated in vivo that the marginal gaps
obtained with a zinc-phosphate cement (Flecks;
Mizzy, Cherry Hill, NJ, USA) and a resin-modified
glass–ionomer cement (Infinity; Den Mat Corp.,
Santa Maria, CA, USA) with or without dentin
bonding (Tenure; Den Mat Corp.) did not signifi-
cantly differ after 6 months of clinical function. The
clinical implications of marginal gaps—i.e. how
large they have to be to favor penetration of toxins
and bacteria, pulpal damage and secondary caries—
remain to be clarified.
The results obtained did not show any regular
influence of marginal gaps on microleakage. One
would have to assume that the quality of the
marginal gap obtained under the experimental
conditions outlined is not correlated with the
microleakage results obtained. It should be pointed
out that the results of the marginal gap were
smaller than the maximum clinically acceptable
452
marginal gap size of 120 mm (as defined by Mc Lean
and Von Frauenhofer [40]) in all test groups. As in
our own study, White et al. [9] did not find any
correlation between marginal gaps and microleak-
age. Their study examined full cast crowns made of
a non-precious alloy that had been cemented with a
zinc-phosphate cement (Flecks; Mizzy, Cherry Hill,
NJ, USA), a polycarboxylate cement (Durelon; 3M
ESPE, Seefeld, Germany), a glass–ionomer cement
(Ketac-Cem; 3M ESPE, Seefeld, Germany), a micro-
filled resin with dentin bonding (Thin Film Cement
and Tenure; Den Mat Corp., Santa Maria, CA, USA),
and a resin cement (Panavia Ex; Kuraray, Osaka,
Japan). Nor did the study of Lindquist and Connolly
[10] find any correlation between microleakage and
marginal gap when using a zinc-phosphate cement
(Flecks; Mizzy, Cherry Hill, NJ, USA) and a resin-
modified glass–ionomer cement (Vitremer (now
RelyX Luting); 3M ESPE, Seefeld, Germany). How-
ever, more research is needed to determine the
relationships between restoration-associated dis-
eases and the variables of microleakage, marginal
opening, and host factors.
Although we used established protocols to
simulate the oral environment, the real-life
scenario is too complex to be fully reproduced
by experimental set-ups of this type. On balance,
it is reasonable to assume that the data obtained
in the various study groups constituted a viable
basis for comparison. In clinical practice, how-
ever, additional factors such as biocompatibility,
thermal/electric conductivity, ease of use, and,
most important, the specific requirements of each
case (e.g. height of the residual tooth structure
and preparation angle) must enter the equation to
find out which cementing agent is most
appropriate.
Conclusions
1. None of the cementing agents investigated in
this study yielded a perfect seal at the bonding
interface in enamel or dentin.
2. All cementing agents investigated were associ-
ated with higher degrees of microleakage when
the tooth substrate was located in dentin.
3. The self-adhesive universal resin cement (RelyX
Unicem) revealed the smallest degree of micro-
leakage both in enamel and in dentin.
4. After artificial aging differences in the marginal
gap quality were observed.
5. An association between marginal gap and micro-
leakage was not observed.
A. Piwowarczyk et al.
References
[1] Crim GA, Chapman KW. Reducing microleakage in class II
restorations: an in vitro study. Quintessence Int 1994;25:
781–5.
[2] Bergenholtz G, Cox CF, Loesche WJ, Syed SA. Bacterial
leakage around dental restorations: its effect on the dental
pulp. J Oral Pathol 1982;11:439–50.
[3] Cox CF, Keall CL, Keall HJ, Ostro E, Bergenholtz G.
Biocompatibility of surface-sealed dental materials against
exposed pulps. J Prosthet Dent 1987;57:1–8.
[4] Going RE, Sawinski VJ. Microleakage of a new restorative
material. J Am Dent Assoc 1966;73:107–15.
[5] Mjör IA, Moorhead JE, Dahl JE. Reasons for replacement of
restorations in permanent teeth in general dental practice.
Int Dent J 2000;50:362–6.
[6] Tjan AH, Dunn JR, Grant BE. Marginal leakage of cast gold
crowns luted with an adhesive resin cement. J Prosthet
Dent 1992;67:11–15.
[7] Tjan AH, Peach KD, VanDenburgh SL, Zbaraschuk ER.
Microleakage of crowns cemented with glass–ionomer
cement: effects of preparation finish and conditioning
with polyacrylic acid. J Prosthet Dent 1991;66:602–6.
[8] White SN, Yu Z, Kipnis V. The effect of seating force on film
thickness of new adhesive luting agents. J Prosthet Dent
1992;68:476–81.
[9] White SN, Ingles S, Kipnis V. Influence of marginal opening
on microleakage of cemented artificial crowns. J Prosthet
Dent 1994;71:257–64.
[10] Lindquist TJ, Connolly J. In vitro microleakage of cementing
agents and crown foundation material. J Prosthet Dent
2001;85:292–8.
[11] Piwowarczyk A, Lauer HC. Mechanical properties of luting
cements after water storage. Oper Dent 2003;28:535–42.
[12] Li ZC, White SN. Mechanical properties of dental luting
cements. J Prosthet Dent 1999;81:597–609.
[13] Nakabayashi N, Kojima K, Masuhara E. The promotion of
adhesion by the infiltration of monomers into tooth
substrates. J Biomed Mater Res 1982;16:265–73.
[14] Raskin A, D’Hoore W, Gonthier S, Degrange M, Dejou J.
Reliability of in vitro microleakage tests: a literature
review. J Adhes Dent 2001;3:295–308.
[15] Jorgensen KD. Factors affecting the film thickness of zinc-
phosphate cements. Acta Odontol Scand 1960;18:479–90.
[16] Wu W, Cobb EN. A silver staining technique for investigating
wear of restorative dental composites. J Biomed Mater Res
1981;15:343–8.
[17] White SN, Sorensen JA, Kang SK, Caputo AA. Microleakage
of new crown and fixed partial denture luting agents.
J Prosthet Dent 1992;67:156–61.
[18] Holmes JR, Bayne SC, Holland GA, Sulik WD. Considerations
in measurement of marginal fit. J Prosthet Dent 1989;62:
405–8.
[19] Shortall AC, Fayyad MA, Williams JD. Marginal seal of
injection-molded ceramic crowns cemented with three
adhesive systems. J Prosthet Dent 1989;61:24–7.
[20] Feilzer AJ, De Gee AJ, Davidson CL. Curing contraction of
composites and glass–ionomer-cements. J Prosthet Dent
1988;59:297–300.
[21] Inokoshi S, Willems G, Van Meerbeek B, Lambrechts P,
Braem M, Vanherle G. Dual-cure luting composites. Part I:
filler particle distribution. J Oral Rehabil 1993;20:133–46.
[22] Crim GA, Garcia-Godoy F. Microleakage: the effect of
storage and cycling duration. J Prosthet Dent 1987;57:
574–6.
Microleakage of various cementing agents for full cast crowns
[23] Diaz-Arnold AM, Vargas MA, Haselton DR. Current status of
luting agents for fixed prosthodontics. J Prosthet Dent
1999;81:135–41.
[24] Swartz ML, Phillips RW, Pareja C, Moore BK. In vitro
degradation of cements: a comparison of three test
methods. J Prosthet Dent 1989;62:17–23.
[25] White SN, Zhaokun Y, Tom JF, Sangsurasak S. In vivo
microleakage of luting cements for cast crowns. J Prosthet
Dent 1994;71:333–8.
[26] Kydd WL, Nicholls JI, Harrington G, Freeman M. Marginal
leakage of cast gold crowns luted with zinc-phosphate
cement: an in vivo study. J Prosthet Dent 1996;75:9–13.
[27] Mesu FP, Reedijk T. Degradation of cementing agents
measured in vitro and in vivo. J Dent Res 1983;62:1236–40.
[28] Pluim LJ, Arends J. The relation between salivary proper-
ties and in vivo solubility of dental cements. Dent Mater
1987;3:13–18.
[29] White SN, Yu Z, Tom JF, Sangsurasak S. In vivo marginal
adaption of cast crowns luted with different cements.
J Prosthet Dent 1995;74:25–32.
[30] Shiflett K, White SN. Microleakage of cements for stainless
steel crowns. Pediatr Dent 1997;19:262–6.
[31] Yoshida K, Kamada K, Sawase T, Atsuta M. Effect of three
adhesive primers for a noble metal on the shear bond
strengths of three resin cements. J Oral Rehabil 2001;28:
14–19.
453
[32] Antoniadou M, Kern M, Strub JR. Effect of a new metal
primer on the bond strength between a resin cement and
two high-noble alloys. J Prosthet Dent 2000;84:554–60.
[33] Cobb DS, Vargas MA, Fridrich TA, Bouschlicher MR. Metal
surface treatment: characterization and effect on compo-
site-to-metal bond strength. Oper Dent 2000;25:427–33.
[34] Swift Jr EJ, Perdigao J, Heyman HO. Bonding to enamel and
dentin: a brief history and state of the art. Quintessence Int
1995;26:95–110.
[35] Pashley DH. Clinical correlations of dentin structure and
function. J Prosthet Dent 1991;66:771–81.
[36] Zuidgeest TG, Herkstroter FM, Arends J. Mineral density and
mineral loss after demineralization at various locations in
human root dentine. A longitudinal microradiographic
study. Caries Res 1990;24:159–63.
[37] Peters AD, Meiers JC. Effect of polymerization mode of a
dual-cured resin cement on time-dependent shear bond
strength to porcelain. Am J Dent 1996;9:264–8.
[38] Gemalmaz D, Ozcan M, Yoruc AB, Alkumru HN. Marginal
adaptation of a sintered ceramic inlay system before and
after cementation. J Oral Rehabil 1997;24:646–51.
[39] White SN, Kipnis V. Effect of adhesive luting agents on the
marginal seating of cast restorations. J Prosthet Dent 1993;
69:28–31.
[40] Mc Lean JW, Von Frauenhofer VA. The estimation cement
film thickness by an in vivo technique. Br Dent J 1971;131:
107–11.