JDC SCIENTIFIC ARTICLE

Microleakage and Polymerization Shrinkage of

Various Polymer Restorative Materials
David Alain Gerdolle DDS, MS Eric Mortier, DDS, MS
Dominique Droz, DDS, PhD

ABSTRACT
Purpose: The aim of this in vitro study was to evaluate the polymerization shrinkage and
the microleakage of direct resin-based restorative materials commonly used in pediatric
dentistry.
Methods: Standardized Class V cavities overlapping the cementoenamel junction were
prepared on the buccal and the lingual surfaces of 40 extracted human mandibular third
molars (36 specimens, 4 controls). The cavities were restored with 4 different materials: a
packable resin composite (Filtek P60), a compomer (Compoglass F), an ormocer (Admira)
and their associated bonding agents (Scotchbond 1, Excite, and Admira Bond, respec-
tively), and a resin-modified glass ionomer (Fuji II LC). The teeth were then immersed
in methylene blue solution for 48 hours. Dye penetration was evaluated for all materials,
which were analyzed using a multivariate model (α=0.05): influence of microleakage score,
margin location (enamel/cementum), and preparation location (buccal/lingual). Multi-
variate analysis was performed using a polychotomous logistic regression. Polymerization
shrinkage was evaluated by the disk deflective method. The percentage of polymerization
shrinkage (N=3) was evaluated by ANOVA and Tukey test.
Results: Regarding polymerization shrinkage, the P60 demonstrated the lowest value, fol-
lowed by ADM and COF, whereas FLC presented the highest shrinkage-strain (P<.0001).
The preparation location had no significant effect on dye penetration (P=.86). Margin
location (enamel or cementum) had a significant effect on microleakage (odds ratio
[OR]=24.61). Significant differences in the microleakage patterns and scores were also
observed between the 4 restorative materials. Admira exhibited the lowest overall micro-
leakage. In comparing Filtek P60, Compoglass F, and Fuji II LC to Admira, P60 showed
significantly less microleakage (OR=1.30) than Fuji II LC (OR=1.47), whereas Compoglass
F demonstrated the greatest significant overall microleakage (OR=3.15).
Conclusion: Within the experimental conditions of this in vitro study, the microleakage
was significantly lower at the enamel margins than at the cementum margins for the four
restorative materials tested. The ormocer and the packable resin composite exhibited the
best sealing ability, as well as the lowest polymerization shrinkage. It could not be demon-
strated in this study, however, that the higher the polymerization shrinkage was, the lower
the marginal sealing ability was. (J Dent Child 2008;75:125-33)
Received February 3, 2007; Last Revision February 18, 2007; Revision Accepted
March 5, 2007.
KEYWORDS: DENTAL POLYMERS, MICROLEAKAGE, POLYMERIZATION SHRINKAGE

T
oday, resin-based restorative materials (RBRM) are
Dr. Gerdolle is a Private Practionner and General Manager, Swiss
Dental Clinics, Vevey, Switzerland; Dr. Mortier is Fellow Research-
the most frequently used restorative material in both
er, Laboratory of Physics of Materials, UMR CNRS – INPL 7556, Mines pediatric dentistry and general pratice.1,2 Although
School, Nancy, France; Dr. Droz is Senior Lecturer, Department of these materials present numerous benefits to both practitio-
Pedodontics, Faculty of Odontology, University of Nancy 1, Nancy,
France. ners and patients, the conversion of resin monomers into
Correspond with Dr. Gerdolle at [email protected] a polymer network leads to bulk contraction and stresses.3

Journal of Dentistry for Children-75:2, 2008 Microleakage of dental polymer Gerdolle et al 125
The polymerization shrinkage is known to be the main b. micromechanical bonding similar to that of resin
cause of marginal gap formation, subsequent microleakage, composites.20 The retention and the marginal
and pulpal pathosis.4 sealing capacity could be determined by such a
The sealing ability of RBRM to the calcified tooth tissues double adhesion mechanism.
is of great significance in many restorative procedures.5,6 2. The second hybrid material type is represented by
Despite numerous improvements within modern dentin the polyacid-modified resin composites—so-called
adhesive systems, the bond strength and marginal adap- compomers. Compomers have close relationship with
tation of RBRM to dentin remain less predictable than conventional resin composites and are thought to
those to enamel.7 Many restorative cavities, specifically in bond to dentin by micromechanical bonding, as resin
posterior teeth, present margins located in cementum or in composites do.21,22 Moreover, the fluoride content of
dentin.8,9 The difficulty in achieving marginal sealing and these materials might present a cariostatic effect, which
correct hygiene, especially at the cervical margin of deep would last several weeks.23
cavities, increases marginal microleakage and the subsequent In vitro evaluations remain an essential method for
recurrence of caries.10 an initial screening of dental materials and might act as
Although no current product satisfies all the require- a predictable indicator of in vivo leakage.24 The primary
ments of an ideal restorative material, adhesive techniques factors affecting sealing ability and, consequently, clinical
enable some procedures that cannot be provided with durability, are:
amalgam restorations.11 RBRM procedures, however, are 1. chemical variations of the tooth substrate9;
technique sensitive compared to amalgam restorations. 2. adhesive properties of the materials5,9; and
Besides, the restorative material type plays an important 3. differences in the coefficient of thermal expansion of
role in restoration longevity. The type of resin composite restorative materials with the tooth substrate.25
is partly determined by the amount and the size of filler It is known that many different and varied techniques
particles.11 A high amount of filler increases the strength have been used to test the in vitro cavity-sealing ability of
and the elasticity modulus and reduces: polymerization restorations. These have included the use of: dyes, chemical
shrinkage12,13, coefficient of thermal expansion, and wa- tracers, radioactive isotopes, air pressure, bacteria, neutron
tersorption.14,15 Recently, a new generation of composites activation analysis, scanning electron microscopy, artificial
has been proposed with a new filler design, allowing a caries techniques, and electrical conductivity.26
more efficient packing into cavity preparations.16 Packable The purpose of this study was to evaluate, via dye pene-
composites have generally larger filler particles, because tration and in relation to their polymerization shrinkage, the
the resin matrix is chemically modified to allow this slight marginal leakage in Class V cavities of 4 different restorative
increase of filler amount.16 materials associated with their respective adhesive systems:
Another new approach in restorative dentistry has been a packable resin composite, an ormocer, a compomer, and
the introduction of ormocer (organically modified ceramics) a resin-modified glass ionomer.
in 1998. Instead of bisphenol glycidyl methacrylate (Bis-
GMA), urethane dimethacrylate (UDMA) and triethylene
glycol dimethacrylate (TEGDMA), multifunctional ure- METHODS
thane- and thioetheracrylate alkoxysilanes as sol-gel precur-
MICROLEAKAGE EVALUATION
sors have been developed for the synthesis of inorganic-or-
The same operator performed all the procedures according
ganic copolymer ormocer resin composites as RBRM.17 The
to ISO specification 11 405:2003.27 Forty freshly extracted
hydrolysis and the polycondensation reactions of alkoxysilyl
human third molars, stored for less than 3 months, were
groups allow the formation of an inorganic Si-O-Si network,
selected as study specimens. All gingival remnants were
and the acrylate groups are available for photochemically
removed, and the crowns were thoroughly cleaned with
induced organic polymerization. After incorporation of
prophylactic rotary instruments (Screw-in Cups, W&H,
filler particles, the so-called ormocer composites can be
Bürmoos, Austria). Before storage, the teeth were examined
manipulated like hybrid composites.18
under binocular microscope (X10, model S2H; Olympus
About 2 decades ago, 2 hybrid restorative material types
Corp, Tokyo, Japan) to ensure that the specimens were
originated from combining glass ionomer and methacrylate
exempt from any decay, cracks, or previous restorations. In
resin technology19:
order, the teeth were stored in:
1. The first type includes the resin-modified glass iono-
1. 0.1% T chloramine (Prolabo, Paris, France) at 4°C for
mers (RMGI), which resemble the composition of
1 week;
conventional glass ionomers. Resin-modified glass
2. distilled water at 4°C for 3 months at most;
ionomers are said to bond to enamel and dentin with a:
3. distilled water at 23°C±1°C during the last 12 hours
a. chemical bonding mechanism identical to conven-
before use.
tional glass ionomers; and

126 Gerdolle et al Microleakage of dental polymer Journal of Dentistry for Children-75:2, 2008
 Table 1. Materials Used*

Code Material (manufacturer) Type Main components Batch no.

ADM Admira (Voco Organically modified ceramic • Monomers: Bis-GMA, di-UDMA, 03655
D-27457, Cuxhafen, (ormocer) TEGDMA
Germany) • Fillers (78 wt %=56% vol):
Ba-Al-B-silicate glass (90%,
ca 0.7µm), SiO2 (10%)
• 3-dimensionally curing anorganic-
organic copolymers, additive aliphatic
and aromatic dimethacrylates

COF Compoglass F Polyacid-modified resin • Monomers (22.5 wt %): UDMA, B06383

(Vivadent Ets, FL-9494 composite (compomer) poly-EGDMA, CADCADM
Schaan, Liechtenstein) • Fillers (77 wt %): Ba-Al-fluorosilicate
glass (1 µm), ytterbium trifluoride,
spheroid mixed oxide
• Additional contents (0.25 wt %):
Catalysts, stabilizers, pigments

FLC Fuji II LC Resin-modified glass • Powder: Aluminofluorosilicate glass Powder

(GC Corp, Tokyo, Japan) ionomer • Liquid: Polyacrylic acid, HEMA, 2-2-4 030461
trimethyl hexamethylene dicarbonate, Liquid
TEGDMA, water 080771

P60 Filtek P60 Resin composite • Monomers: UDMA, TEGDMA, 9AN

(3M/ESPE dental products, Bis-EMA
St. Paul, Minn) • Fillers (81 wt %=61% vol): ZrSiO4

* Bis-GMA=Bisphenol-A glycidylmethacrylate; Bis-GA=Bisphenol-A glycidylpolyacrylate;

UDMA=urethane dimethacrylate; TEGDMA=triethyleneglycol dimethacrylate;
poly-EGDMA=polyethyleneglycol dimethacrylate; CADCADM=cycloaliphatic dicarboxilic acid dimethacrylate;
Bis-EMA=ethoxylated bisphenol-A glycol dimethacrylate; HEMA=2-hydroxyethyl methacrylate.

In 4 series of 9 teeth, standardized, nonbeveled Class V 3. an ormocer (Admira [ADM], Voco, Cuxhaven, Ger-
(U shaped) cavities were prepared and finished on the buc- many) and the associated bonding agents Scotchbond
cal and lingual surfaces with 90 µm and 20 µm diamond 1 (3M ESPE Dental Products, St Paul. MN, USA),
rotary cutting instruments (no. 802 314 009 and 801 314 Excite (Vivadent, Schaan, Liechtenstein), and Admira
023, Komet, Lemgo, Germany) under constant air-water Bond (Voco, Cuxhaven, Germany) respectively; and
spray. The margins were located: 4. a resin-modified glass ionomer (Fuji II LC [FLC], GC
1. on both sides of the cementoenamel junction (CEJ); Corp. Tokyo, Japan).
2. in enamel for the occlusal aspect of the preparation; and The main components of all the materials are summa-
3. in cementum at the gingival margin. rized in Tables 1 and 2.
The dimensions of each preparation (mesiodistal The specimens were randomly assigned to receive a pair
width=4.6 mm; occlusogingival height=2.6 mm; pulpal of restorative materials according to the possible pairing
depth=2 mm) were verified with a 10-µm accurate electronic combinations (P60/COF, P60/ADM, P60/FLC, COF/
caliper (Digimatic, model no. 500-181U; Mitutoyo Corp, ADM, COF/FLC, ADM/FLC). The random allocations
Tokyo, Japan). Between each step of the experiment, the were carried on by drawing lots. A visible-light polymerizing
teeth were stored in distilled water at 23°C±1°C. Following unit with an irradiating diameter of 9 mm (XL 3000 cur-
the conditioning and priming of the teeth according to the ing light, 3M/ESPE, St Paul, MN) was used to polymerize
manufacturers’ instructions, the cavities were filled with a the bonding agents and the restorative materials. Light
single increment of 4 different materials: activation energy was controlled at regular intervals to as-
1. a packable resin composite (Filtek P60 [P60], 3M sure a minimum value of 600 mW/cm2. The distance of
ESPE Dental Products, St Paul. MN, USA); the light tip from the specimens was maintained between
2. a compomer (Compoglass F [COF], Vivadent, Schaan, 1 and 2 mm.
Liechtenstein);

Journal of Dentistry for Children-75:2, 2008 Microleakage of dental polymer Gerdolle et al 127
 Table 2. Bonding Agents Used*

Code Material (manufacturer) Type Main components Batch no.

ADB Admira Bond ADM Bis-GMA, HEMA, organic acids, complex 03653
(Voco D-27457, 3-dimensionally curing anorganic-organic
Cuxhafen, Germany) copolymers, acetone

EXC Excite (Vivadent Ets SOL2 Bis-GMA, HEMA, dimethacrylate phos- B33276
FL-9494, Schaan, Liechtenstein) phoric acid, acrylate, highly dispersed silica,
ethanol (25 wt % )

SCO Scotchbond 1† P60 Bis-GMA, HEMA, Bis-phenol A glycerolate 0EE

(3M ESPE dental products, dimethacrylate, copolymer of polyacrylic and
St. Paul, Minn) polyitaconic acids, water, ethanol

* B is-GMA=Bisphenol-A glycidylmethacrylate; UDMA=urethane dimethacrylate;

HEMA=2-hydroxyethyl methacrylate; 4-META=4- methacryloxyethyl-trimelittic acid.
† Scotchbond 1=Scotchbond Single Bond in USA

Immediately following polymerization, the specimens c. 2=penetration to the dentin aspect of the preparation
were immersed in distilled water at 23°C±1°C for 24 hours. wall, but not including the pulpal floor; and
Polishing was then conducted with flexible disks (Sof-Lex d. 3=penetration including the pulpal floor of the prepa-
XT Pop On, 3M/ESPE, Seefeld, Germany). The root apexes ration.27
were then sealed with a resin composite (Z100, 3M/ESPE, Statistical analyses were performed using the statistical
Seefeld, Germany) without a bonding agent. The teeth software SAS/STAT, (v. 8.2, SAS Institute Inc, Cary, NC).
were entirely coated with 2 layers of nail varnish (L’Oreal, Using a multivariate model (α=0.05), the following fac-
Paris, France), except for the location of the restorations tors were analyzed: microleakage score, margin location
and 1 mm around the restoration margins. Four teeth were (enamel or cementum), and preparation location (buccal
prepared as controls: or lingual). The model provided regression coefficients of
a. 2 specimens with a filled cavity and entirely coated independent variables (margin location, preparation loca-
with nail varnish served as negative controls; and tion, and restorative materials). Odds ratios (OR) expressed
b. 2 specimens with an empty cavity, and without varnish the effect, when changing an independent variable, on the
coating, served as positive controls. probability of having one unit higher score value, holding
All of the specimens were then soaked in a 1% methylene other variables in the equation model constant (OR are
blue dye (Prolabo, Paris, France) for 48 hours. To evaluate given with the 95% confident limits). Multivariate analysis
dye penetration, after having been rinsed with distilled was performed using an ordinal polychotomous logistic
water, the specimens were sectioned with a 500-µm thick, regression.
slow-speed, diamond-coated disk (Isomet-Plus; Buehler,
Lake Bluff, Ill) under water coolant. A first section was
POLYMERIZATION SHRINKAGE
centered along the mesiodistal axis to separate the buccal
Polymerization shrinkage was measured with the “deflecting
and the lingual surfaces. Three 1-mm thick buccolingual
disk” technique used by Watts and Cash.28 The tested poly-
sections were then made for each half-specimen: (a) 1 in
mer was placed within a 1.64-mm high brass ring—with an
the center of the restoration; (b) 1 in the mesial margin;
internal diameter of 15 mm—and attached with an adhesive
and (c) 1 in the restoration’s distal margin. The sections
to a glass microscope slide, the surface of which had been
were examined on each side under a binocular microscope
sandblasted. A flexible glass cover slip (0.1-mm thick) was
(X10, model S2H, Olympus Corp, Tokyo, Japan). Dye
placed on the ring rim in contact with the tested material. A
penetration was measured on the 6 enamel margins and 6
linear vertical displacement transducer (LVDT GTX 2500,
cementum margins, for a total of 12 measurements per cav-
RDP Electronics, UK; sensitivity>0.1 µm) was gently located
ity. As there were 18 cavities per material, 216 measurements
in contact with the cover slip’s upper surface. The tempera-
were taken for each restorative material. The degree of dye
ture of the glass plate and the specimen platform was set to
penetration was identified according to ISO specification
37°C±0.5°C, and the room temperature was maintained at
11 405:2003:
23°C±1°C with a relative humidity of 50%.
a. 0=no penetration;
The light guide initiated the polymerization from below
b. 1=penetration to the enamel or cementum aspect of
the unset specimen-disk, using a light-curing unit (XL 3000
the preparation wall;

128 Gerdolle et al Microleakage of dental polymer Journal of Dentistry for Children-75:2, 2008
 Curing Light, 3M Dental Products, St Paul, Minn) activated where L0 is the initial specimen height and ΔL is the cover
for 60 seconds, for all the samples (N=3 for each material). slip displacement expressed in µm. Data were statistically
The light activation energy was regularly controlled to as- analyzed by 1-way ANOVA and Tukey tests.
sure a minimum value of 600mW/cm2. For each specimen,
the cover slip was attracted downwards in an axial way, as
shrinkage took place. The displacement of the cover slip was
RESULTS
recorded over time. According to Watts and Cash,28 with MICROLEAKAGE
the displacement of the disk upper surface being uniform, The negative controls showed no evidence of dye penetra-
measurements at the center were representative of the whole. tion, whereas the dye completely penetrated the positive
The cover slip displacement was not only recorded during control cavities. The dye penetration data (Table 3) and the
the light activation time, but also for 100 seconds after light polychotomous stepwise logistic regression results (Table 4)
activation stopped—for a total duration of 160 seconds. are presented in accordance with each element of the study,
The shrinkage-strain, ε(t), was expressed as a percentage, including the: preparation location, various restorative ma-
according to the equation: terials, and margin location.
The buccal or the lingual preparation location on the
ε(t)%=100 x ΔL / L0
teeth had no significant influence on the dye penetration

Table 3. Leakage Scores*

Variables Total Score 0 Score 1 Score 2 Score 3

Materials (code) Total 864 96 328 115 325

Admira 216 27 101 25 63

(ADM)
EM N=108 25 73 7 3
CM N=108 2 28 18 60

Compoglass F 216 17 63 26 110

(COF)
EM N=108 17 62 9 20
CM N=108 0 1 17 90

Fuji II LC 216 20 85 41 70
(FLC)
EM N=108 19 67 17 5
CM N=108 1 18 24 65

Filtek P60 216 32 79 23 82

(P60)
EM N=108 32 61 12 3
CM N=108 0 18 11 79

Margin location Total N=864 96 328 115 325

Enamel N=432 93 263 45 31

Cementum N=432 3 65 70 294

Preparation location Total N=864 96 328 115 325

Buccal N=432 46 167 53 166

Lingual N=432 50 161 62 159

* EM=enamel margins; CM=cementum margins.

Journal of Dentistry for Children-75:2, 2008 Microleakage of dental polymer Gerdolle et al 129
 Table 4. Factors Associated With the Score Variation Table 5. Shrinkage-strain Data*
(N=864)
Mean apparent shrinkage strain (%)
OR (95% CI)* Materials
at 37°C±(SD)
Margin location Enamel 1
t=60 seconds t=160 seconds
Cementum 24.61 (17.94-33.74)
Admira 2.64±0.11a 2.82±0.14a
Materials ADM 1
Compoglass F 2.96±0.02b 3.22±0.01b
COF 3.15 (2.16-4.58)
Fuji II LC 3.25±0.17c
3.46±0.13c
FLC 1.47 (1.03-2.10)
Filtek P60 1.99±0.06d 2.13±0.04d
P60 1.30 (0.91-1.85)

* OR=odds ratio; CI=confidence interval. * Within datasets for each material, superscript letters indicate
homogenous sub-sets (at the 0.05 level).

(P=.86). Conversely, the polychotomous stepwise logistic at the margin locations is presented in Figure 1. In each
regression results established that margin location was the occurrence, the score values were higher at the cementum
strongest and most consistent predictor of increased micro- margin than at the enamel margin. The increase of micro-
leakage (OR=24.61). The statistical model confirmed that leakage at the cementum margins was significantly greater
microleakage may have been dependant upon the material for COF and for P60 than for ADM and FLC (P<.05).
type. Admira significantly exhibited the lowest overall mi-
croleakage. Comparing Filtek P60, Compoglass F, and Fuji
POLYMERIZATION SHRINKAGE
II LC to Admira, P60 showed significantly less microleak-
A representative, time-dependent curve for each material is
age (OR=1.30) than Fuji II LC (OR=1.47). Compoglass
displayed in Figure 2, and the summary data are given in
F, however, demonstrated the greatest significant overall
Table 5. Figure 2 measurements were made at 160 seconds—
microleakage (OR=3.15). The ordinal polychotomous
sufficient duration to obtain a good estimate of the final
stepwise logistic regression revealed statistical interactions
shrinkage-strain. Statistical analysis established the same value
referring to the score variation between the margin location
ranking and an equivalent statistical significance at the differ-
and the luting cements.
ent shrinkage measurement times (60 seconds, 160 seconds).
This study confirmed that the occurrence of microleak-
FLC had the highest polymerization shrinkage (P<.0001),
age is higher at the cementum margin than at the enamel
whereas P60 showed the lowest shrinkage (P<.0001), fol-
one. The effect of the 4 materials on the microleakage score
lowed by ADM (P<.0001) and COF (P<.0001).

DISCUSSION
According to Mannhart et al,7 all the restorations selected
in this study exhibited interfacial microleakage as was ex-
pected.8,16,21,22 A possible explanation relied upon the adhesive
bond, which declined or even broke because of dimensional
changes, occurring during polymerization.29 Therefore, an
interfacial gap would occur if the adhesion of the restorative
material to the tooth structures did not compensate for the
shrinkage stress exerted by the material in the setting’s very
first setting.30
The polymerization of the resin matrix modified the ma-
terial from a flowing viscous-plastic phase into a rigid-elastic
phase—a gel. Resin shrinkage occurred before reaching the
gel point, when the monomer-polymer was still flowable. It
could be partially compensated by a movement of molecules
in the resin composite from the restoration’s free surfaces.29,31
The compensation could not occur after the gelation point,
however, and large stresses consequently developed in the
RBRM. Thus, polymerization contraction stress might be
the primary cause for microleakage when the experimental
Figure 1. Graph of restorative material effect on microleakage restorations were not subjected to thermocycling and/or
score vs margin locations (N=864). mechanical load cycling22 as the present study did. Shrink-

130 Gerdolle et al Microleakage of dental polymer Journal of Dentistry for Children-75:2, 2008
 themselves contributed to increase FLC micro-
leakage scores.
Despite a constant water-storage at 23°C±1°C
between each experimental step, the extracted
teeth tissue did not contain sufficient water
to prevent dehydration.33 Thus, microleakage
might result from a bonding breakdown and
Shrinkage-strain (%)

from dye absorption into the layer of porous

material—a phenomena which is due to the
changes in the physical properties of the mate-
rial.34 Actually, methylene blue penetration was
observed in this study in both the tooth/material
Admira gap and the FLC material. This physicochemical
Compoglass F
behavior might be explained by: hygroscopic ex-
pansion33, reduced setting stress owing to water
Fuji II LC
absorption35; and improved bonding ability or
P60 polymerization during the water storage by the
increased value of flexural modulus.6
These properties, however, may also com-
pensate for the relative weakness of the bond
Time (seconds) strength on dental hard tissues.34 The water
sorption mechanism may affect the mechani-
Figure 2. Time-dependence of shrinkage-strain. cal properties and decrease the sealing ability
of the composite resin materials.36 In fact, the
water solubility of some matrix components
age has to be specially considered when a bulk placement could create some modification in the material
technique in Class V cavity was used, because a large bonding structure and, subsequently, allow tracer penetration.23
area (high C-factor) allowed greater stress development at the This study’s results confirmed that polymerization
adhesive interface.13,30,31 contraction was not the only parameter implicated in the
Nevertheless, if only contraction stresses determined the adhesion mechanisms and durability. Other factors have to
extent of microleakage, FLC used in this study was expected also be considered:
to obtain the highest scores of dye penetration and P60 was 1. the extent of the marginal gap;
expected to show the best sealing ability. ADM demon- 2. varying coefficients of thermal expansion for tooth and
strated the overall lowest dye penetration scores, however, restoration materials at the interfacial level25;
despite a shrinkage-strain value superior to P60. Also, COF 3. degradation of the bonding joint or of the restorative
presented the highest leakage scores, when its shrinkage- materials37; or
strain value was inferior to FLC. The good sealing ability 4. the dissolution of liners or smear layers.25,37
of ADM and P60 could be related to their high filler load, This study agrees with others that demonstrated a stron-
which withstands the polymerization contraction stress.17 ger adhesion to enamel than to dentin or cementum.16,17
In addition, ormocers might undergo less leakage because The crystalline prismatic structure conferred to enamel
of their specific 3-dimensional structure and extremely high an anisotropic behavior. That means that cleavage plan
molecular weight.15 location was perpendicular to the enamel surface and,
Regarding the difference between the marginal leakage thus, a high bond strength.5 Once again, this study clearly
of FLC and COF, a major factor might be the bonding indicates that cementum cannot offer sufficient crystalline
ability to tooth structures. At the same time FLC showed structure to provide a high micromechanical bond with
the highest penetration score and a less overall leakage the RBRM.38 Consequently, the leakage scores for all the
than COF, particularly at the cementum margins. On top studied materials were lower at the enamel margins than at
of an acid-base reaction, COF sets with a free-radical polym- the cementum margins.
erization process. Therefore, the shrinkage stress would be The sealing ability of FLC at the cementum margin was
sufficiently enhanced to produce a marginal gap and more better than that of P60 and COF and might result from a
subsequent leakage. The setting procedure of FLC is essen- complex adhesion mechanism. The short-time condition-
tially achieved by an acid-base reaction.19 A polymerization ing treatment might remove the smear layer and expose
reaction also occurs with the HEMA and urethane dimethac- the collagen fibrils up to a 0.5-µm depth, leaving a large
rylate monomers of the resin matrix that produce additional part of the crystal content linked to the collagen fibrils.19
shrinkage.13 The fact that FLC demonstrated weaker bond A chemical bonding—a consequence of the acid-base
strength to both enamel and dentin could explain the high reaction in the glass ionomer setting reaction—occurred
leakage scores.32 Furthermore, the experimental conditions with the calcium of the hydroxyapatite crystals.19 HEMA

Journal of Dentistry for Children-75:2, 2008 Microleakage of dental polymer Gerdolle et al 131
and urethane dimethacrylate monomers might interdiffuse meet long-term requirements. This appears to be a
in collagen fibrils, providing a micro-mechanical bond in point of major concern, particularly considering the
partly demineralized dentin.19 dental care of young patients.
The RBRMs did not chemically bond to tooth structure, 4. This study, however, could not demonstrate that higher
when glass ionomers do thanks to the acid-base process. The polymerization shrinkage resulted in lower marginal
sealing ability of the RBRMs was said to be mainly based on sealing ability. Thus, polymerization shrinkage was
a micromechanical overlap in the conditioning tooth crys- confirmed not to be the only factor inducing interfacial
talline structures. The diphenyl sulfone derivative in COF microleakage.
might also have this ability.39 COF, however, demonstrated
higher leakage scores than FLC at the cementum margin.
To be able to establish whether a correlation might
ACKNOWLEDGMENTS
The authors wish to thank Dr. C. Loos-Ayav of the Epide-
exist between this study’s results and those previously con-
miology Department, Brabois Hospital, Nancy, France, for
ducted,24 this study conformed to the ISO/TS 11405:2003
her advice and help with the statistical calculations.
standard guidelines.27 The employment of this recognized
approach addressed whether in vitro studies are able to
reproduce the clinical behavior of dental materials. Some REFERENCES
authors claim that in vivo specimens are able to exhibit 1. Hicks MJ, Flaitz CM. Epidemiology of dental caries in
greater microleakage than in vitro specimens, even if the the pediatric and adolescent population: A review of past
latter are thermal cycled or mechanically loaded. 22 Oth- and current trends. J Clin Pediatr Dent 1993;18:43-9.
ers report differences between in vitro and in vivo bond 2. O’Brien WJ. Dental Materials and Their Selection.
strengths, which are related to microleakage.5,24 2nd ed. Carol Stream, Ill: Quintessence; 1997:100-3.
It is accepted that the present study may have underesti- 3. Feilzer AJ, De Gee AJ, Davidson CL. Curing con-
mated the actual leakage by comparing a sectioning method traction of composite and glass ionomer cements. J
with other methods, which screen the entire volume of the Prosthet Dent 1988;59:297-300.
preparations.26 Some authors have indicated that such a 2- 4. Brännström M, Vojinovic O. Response of the dental
dimensional approach may explain the lack of correlation pulp to invasion of bacteria round three filling materi-
between in vitro microleakage measurements and in vivo als. J Dent Child 1976;43:15-21.
performance.40 Moreover, the ISO 11405 specifications27 5. Mason PN, Ferrari M, Cagidiaco MC. Shear bond
may create unfavorable polymerization conditions for the strength of four dentinal adhesives applied in vivo and
RMGIC, especially during uncontrolled dehydration con- in vitro. J Dent 1996;24:217-22.
ditions. Thus, the microleakage may be overestimated for 6. Örtengren U, Elgh U, Spasenoska V, Milleding P,
this material. For these reasons, this experiment has to be Haasum J, Karlsson S. Water sorption and flexural
completed by evaluating the microleakage after a mechanical properties of composite resin cement. Int J Prostho-
loading treatment in an acidic environment. It is hoped that dont 2000;13:141-7.
such parameters might better reflect the clinical situation. 7. Manhart J, Schmidt N, Chen HY, Kunzelmann KH,
Hickel R. Marginal quality of tooth-colored restora-
tions in Class II cavities after artificial aging. Oper
CONCLUSIONS Dent 2001;26:357-66.
Within the experimental results of this in vitro study, the 8. Dietschi D, Scampa U, Campanile G, Holz J. Mar-
following conclusions were drawn: ginal adaptation and seal of direct and indirect Class
1. None of the 4 restorative materials prevented dye pe- II composite resin restorations: An in vitro evaluation.
netration. For all the materials tested, dye penetration Quintessence Int 1995;26:127-38.
was higher at the cementum margin. This observation 9. Fortin D, Swift EJ Jr, Denhy GE, Reinhadt JW. Bond
suggests that the restorative materials tested must be strength and microleakage of current dentin adhesives.
carefully considered for subgingival preparations. Dental Mater 1994;10:253-8.
2. Regarding material types, the ormocer and the pack- 10. Manhart J, Garcia-Godoy F, Hickel R. Direct posterior
able resin composite exhibited the lowest microleakage restorations: Clinical results and new developments.
scores, followed by the resin-modified glass ionomer. Dent Clin North Am 2002;46:303-39.
The compomer, however, showed the highest marginal 11. Wakefield CW, Kofford KR. Advances in restorative
leakage values. materials. Dent Clin North Am 2001;45:7-31.
3. Regarding polymerization shrinkage, the packable resin 12. Davidson CL, Feilzer AJ. Polymerisation shrinkage
composite demonstrated the lowest values, followed and polymerisation shrinkage stress in polymer-based
respectively by the ormocer and the compomer; the restoratives. J Dent 1997;25:435-40.
resin-modified glass ionomer, however, presented the 13. Feilzer AJ, Kakaboura AI, De Gee AJ, Davidson CL.
highest values. In this way, ormocers and packable The influence on the development of setting shrinkage
composites appear to be more suitable than com- stress in traditional and resin-modified glass ionomer
pomers and resin-modified glass ionomer cements to cements. Dent Mater 1995;11:186-90.

132 Gerdolle et al Microleakage of dental polymer Journal of Dentistry for Children-75:2, 2008
14. Feilzer AJ, De Gee AJ, Davidson CL. Relaxation of 27. International Standards Organization. ISO standard
polymerization contraction shear stress by hygroscopic 11405:2003. Dental materials–testing of adhesion to
expansion. J Dent Res 1990;69:36-9. tooth structure. Geneva: ISO/TS, 2003. Available at:
15. Mortier E, Gerdolle DA, Jacquot B, Panighi MM. “http://www.iso.ch/iso/en/prods-services/ISOstore/store.
Importance of watersorption and solubility studies for html”. 2d Edition, 2003-01-20.
the couple bonding agent/resin-based filling material. 28. Watts DC, Cash AJ. Determination of polymerization
Oper Dent 2004;29:669-76. shrinkage kinetics in visible-light-cured materials:
16. Meiers JC, Kazemi R, Meier CD. Microleakage of Methods development. Dent Mater 1991;7:281-7.
packable composite resins. Oper Dent 2001;26:121-6. 29. Chen HY, Manhart J, Hickel R, Kunzelmann KH.
17. Yazici AR, Celik C, Ozgünaltay G. Microleakage of Polymerization contraction stress in light-cured pack-
different resin composite types. Quintessence Int able composite resins. Dent Mater 2001;17:253-9.
2004;35:790-4. 30. Davidson CL, de Gee AJ, Feilzer AJ. The competition
18. Hickel R, Manhart J, Garcia-Godoy F. Clinical results between the composite-dentin bond strength and
and new developments of direct posterior restorations. the polymerisation bond stress. J Dent Res 1984;63:
Am J Dent 2000;13:41D-54D. 1396-9.
19. Gladys S, Van Meerbeck B, Lambrechts P, Vanherle G. 31. Yap AUJ, Wang HB, Siow KS, Gan LM. Polymeriza-
Marginal adaptation and retention of a glass ionomer, tion shrinkage of visible light-cured composites. Oper
resin-modified glass ionomers, and polyacid-modified Dent 2000;25:98-103.
resin composite in cervical Class V lesions. Dent Mater 32. Sidhu SK, Watson TF. Interfacial characteristics of
1998;14:294-306. resin-modified glass ionomer materials: A study on
20. Erickson RL, Glasspoole EA. Bonding to tooth struc- fluid permeability using confocal fluorescence micro-
ture: A comparison of glass ionomer and resin com- scopy. J Dent Res 1998;77:1749-59.
posite systems. In: Hunt P, ed. Glass Ionomers: The 33. Watts DC, Kisumbi BK, Toworfe GK. Dimensional
Next Generation. Proceeding of the 2nd International changes of resin/ionomer restoratives in aqueous or
Symposium on Glass Ionomers: 1994. Philadelphia, neutral media. Dent Mater 2000;16:89-96.
Pa: International Symposia in Dentistry; 1994:101-22. 34. Irie M, Suzuki K. Water storage effect on the marginal
21. Evancuski JW, Meiers JC. Microleakage of Compo- seal of resin-modified glass ionomer restorations. Oper
glass-F and Dyract-AP compomers in Class V prepa- Dent 1999;24:272-8.
rations after salivary contamination. Pediatr Dent 35. Akashi A, Matsuya Y, Unemori M, Akamine A. The
2000;22:39-42. relationship between water absorption characteristics
22. Hakimeh S, Vaidyanathan J, Houpt ML, Vaidyana- and the mechanical strength of resin-modified glass
than TK, Von Hagen S. Microleakage of compomer ionomer cements in long-term water storage. Biomater
Class V restorations: Effect of load cycling, thermal 1999;20:1573-8.
cycling, and cavity shape differences. J Prosthet Dent 36. Momoi Y, McCabe JE. Hygroscopic expansion of resin-
2000;83:194-203. based composites during 6 months of water storage.
23. Geurtsen W. Substances released from dental resin Br Dent J 1994;176:91-6.
composites and glass ionomer cements. Eur J Oral Sci 37. Prati C, Pashley DH, Chersoni S, Mongiorgi R. Mar-
1998;106:687-95. ginal hybrid layer in Class V restorations. Oper Dent
24. Raskin A, D’Hoore W, Gonthier S, Degrange M, 2000;25:228-33.
Déjou J. Reliability of in vitro microleakage tests: A 38. Rasmussen ST, Patchin RE, Scott DB, Heuer AH.
literature review. J Adhesive Dent 2001;3:295-308. Fracture properties of human enamel and dentin. J
25. Yoshida Y, Van Meerbeck B, Nakayama Y, Snauwaert Dent Res 1976;55:154-64.
J, Hellemans L, Lambrechts P, Vanherle G, Wakasa 39. Meyer JC, Cattani-Lorente MA, Dupuis V. Compom-
K. Evidence of chemical bonding at biomaterial-hard ers: Between glass ionomer cements and composites.
tissue interface. J Dent Res 2000;79:709-14. Biomater 1998;19:529-39.
26. Gwinnett JA, Tay FR, Pang KM, Wei SH. Compari- 40. Gale MS, Darvell BW, Cheung KM. Three-dimen-
son of three methods of critical evaluation of micro- sional reconstruction of microleakage pattern using
leakage along restorative interfaces. J Prosthet Dent a sequential grinding technique. J Dent. 1994;22:
1995;74:575-85. 370-5.

Journal of Dentistry for Children-75:2, 2008 Microleakage of dental polymer Gerdolle et al 133