BIOLOGICAL RISKS OF RESIN-BASED MATERIALS

TO THE DENTIN-PULP COMPLEX

Serge Bouillaguet
Department of Cariology and Endodontics, Departement of Dental Materials, School of Dental Medicine, University of Geneva, 19 Rue Barthélemy-Menn, CH-1205 Geneva, Switzerland;
serge.bouillaguet@medecine.unige.ch

ABSTRACT: Over the past 30 years, restorative dentistry has seen a revolution in materials, restorative techniques, and patient
priorities. This revolution has been made possible with the development of new resin-based materials which can be bonded to
the tooth structure. Not all of these changes have been without controversy or concern, and some have raised questions about
the biological safety of these new materials and techniques. It is the purpose of this review to present recent and relevant infor-
mation about the biological risks and consequences of resin-tooth bonding and how these risks are affected by the material, its
clinical properties, and its manipulation by the practitioner. These biological risks are complex and interactive, and are still
incompletely defined. In broad terms, these risks can be divided into those stemming from the toxicological properties of the
materials themselves (direct biological risks) and those stemming from microbiological leakage (indirect biological risks).

Key words. Biocompatibility, composites, bonding, acid-etching, microleakage, nanoleakage.

(1) Introduction
O ver the past 30 years, restorative dentistry has seen a rev-
olution in techniques, materials available, prevalence of
disease, and patient priorities. Today's dentist is able to pre-
vent damage from caries by using materials and techniques
that were unknown in 1970. Furthermore, the dentist's
approach to cavity preparation in the management of caries is
radically different from what it was in the past. Whereas
'extension for prevention' was the main philosophy then,
today the ultraconservative preservation of tooth structure is
the primary goal (Staehle, 1999). A second major force chang-
ing dentistry has been the attitude of patients. Patients no
longer seek dental treatment exclusively for pain. Rather, they
are interested in better esthetics, whiter teeth, and remodeled
"smiles" (Lutz and Krejci, 2001). This restorative revolution has
been made possible with the development of new resin-based
materials that can be bonded to tooth structure (Roulet and
Degrange, 2001).
Not all of these changes in the restorative revolution have
been without controversy or concern. The use of new materials
with new chemistries, the etching of dentin, and the need to
ensure complete polymerization and sealing of the restoration
to the tooth have raised questions about the biological safety of
new materials and techniques. Research over the past 10 years
has partially defined the mechanisms by which resin compos-
ite materials integrate with the dentin-pulp complex. It is the
purpose of this review to present recent and relevant informa-
tion about the biological risks and consequences of resin-tooth
bonding and how these risks are affected by the material, its
clinical properties, and its manipulation by the practitioner.
These biological risks are complex and interactive, and are still
incompletely defined. In broad terms, these risks can be divid-
ed into those stemming from the toxicological properties of the
materials themselves (direct biological risks) and those stem-
ming from microbiological leakage (indirect biological risks).
(2) The Direct Biological Risks
of Resin-based Materials
The low number of reported biological problems with resin-
based materials, despite the placement of millions of restora-
tions worldwide, is testimony to their apparent biocompatibil-
ity. However, there are also reports of post-placement tooth
sensitivity (Unemori et al., 2001), local immunological effects
(Jontell et al., 1995), apoptotic reactions (Goldberg et al., 1994),
and long-term pulpal inflammation (Hebling et al., 1999). There
are other reports, less well-documented, that resin-based mate-
rials may have systemic estrogenic effects (Schafer et al., 1999),
may elicit allergic reactions (Katsuno et al., 1996), or may possi-
bly even act as carcinogens (Schweikl and Schmalz, 1999).
Therefore, it is imperative that we reach a more precise defini-
tion of the direct biological risks associated with the use of
resin-based materials.
(2.1) THE DENTIN-PULP COMPLEX
The primary focus for the definition of the direct biological
risks of resin-based materials is the dentin-pulp complex
(Pashley, 1996). Despite the prevailing and accepted thought
that this complex acts anatomically and functionally as a unit,
it is instructive for us to consider the unique properties of each
component of the dentin-pulp complex, to understand how
resin-based materials interact with it.
Dentin is a mineralized tissue that surrounds the dental
pulp and the processes of the odontoblasts. On average, dentin
contains approximately 50 vol% mineral (hydroxyapatite crys-
tals), 30 vol% organic components (mostly type I collagen),
and 20 vol% fluid (Mjör et al., 2001). The collagen fibrils are

15(1):47-60 (2004) Crit Rev Oral Biol Med 47

 (Garberoglio and
Brännström, 1976).
Furthermore, the
tubular diameter
increases from 0.5 ␮m
near the DEJ to over
2.5 ␮m near the pulp.
The convergence of
the tubules and their
increased diameter
toward the pulp are
responsible for an
increase in dentin per-
meability near the
pulp (Fig. 1). The per-
meability of the
dentin allows for both
outward pulpal fluid
flow and inward dif-
fusion of chemical
and bacterial prod-
ucts. Pashley (1990)
Figure 1. Schematic of convergence of tubules toward the pulp. (A) Periphery of the dentin. Most surface area is occu- has used the Hagen-
pied by intertubular dentin (✰), with a few tubules surrounded by hypermineralized peritubular dentin (✪). (B) Near Poiseuille equation to
the pulp, the increase in tubule diameter has occurred largely at the expense of the peritubular dentin. This substrate show that fluid filtra-
has a high protein content. As the remaining dentin is made thinner (from A to B), the permeability of the dentin tion varies with the
increases, because both diameter and density of dentinal tubules are increased. Reprinted with permission from fourth power of the
Elsevier. radius of the dentin
tubule, and that the
driving force is the
arranged in a network, forming a matrix for the hydroxy- fluid pressure gradient. The inward diffusion of bacterial or
apatite crystals. Spaces of 20-50 nm separate fibrils about 20- products of material degradation may cause deleterious reac-
100 nm in diameter (Eick et al., 1997). This network is uni- tions in the dental pulp. However, the dentin acts as a diluter
formly mineralized and forms the bulk of the dentin (inter- of diffusing substances. According to the Fick equation, the
tubular dentin). In dentin immediately adjacent to the tubules rate of diffusion is dependent on the applied concentration but
(peritubular dentin), the mineralized component predomi- is inversely proportional to the dentin thickness. The surface
nates, and the collagen network is sparse. The dentinal tubules area available for diffusion, the temperature, and the chemical
run from the dentin-enamel junction and converge on one characteristics of the diffusing molecules all affect diffusion
another toward the pulp of the tooth (Outhwaite et al., 1976). (Pashley, 1985). Dentinal diffusion of bacterial or material
In cross-section, the density of the tubules is about products is a critical factor in the assessment of material bio-
15,000/mm2 near the DEJ to over 65,000/mm2 near the pulp logical or microbiological risks. It has been clearly established
from in vitro and in vivo studies
that outward flow of dentinal
fluid cannot completely com-
pensate for the inward diffu-
sion of chemicals or bacteria
(Fig. 2), and that the thickness
of dentin remaining between
prepared surfaces and the pulp
is the critical variable in deter-
mining whether dentin can
protect the pulp (Holz and
Baume, 1973; Pashley and
Figure 2. In vivo diffusion through dentin of a solution Matthews, 1993).
of silver nitrate. (A) Diffusion of silver particles The dental pulp consists of
(arrows) across dentinal tubules 35 min after the a loose connective tissue that
application of a silver nitrate solution (✺) inside the occupies the central part of the
cavity (HE staining x 40). (B) Penetration of silver par-
ticles (✺) into the pulp area and into the capillary sys-
tooth. At the periphery, the
tem active at clearing the pulp (HE staining x 40). odontoblasts line the dentin,
and their processes extend into
the dentinal tubules for at least
several hundred microns (Fig.

48 Crit Rev Oral Biol Med 15(1):47-60 (2004)

3). Odontoblasts are connected to each
other by gap junctions, desmosomes,
and tight junctions and are highly spe-
cialized in the synthesis and the secre-
tion of organic molecules and the min-
eralization of the dentin. Turner et al.
(1989) have shown that a functional
barrier exists between the odonto-
blasts that prevents the passage of
macromolecules from the pulp into the
predentin and dentin. They also
demonstrated that this functional bar-
rier may become permeable during
cavity preparation. This barrier may
also be important for the function of
the odontoblasts as a perceptual
organ. According to the hydrodynam-
ic theory, dentinal pain is induced by
rapid fluid shifts across dentinal
tubules (Brännström and Aström,
1972). These shifts are caused by tem- Figure 3. The dentin-pulp complex. The dentin
perature, pressure, or mechanical per- and pulp exist together as an integrated unit.
turbation and result in a mechanical The functional barrier that develops between
deformation of the odontoblasts and the odontoblasts prevents the passage of
nearby nerves. The mechanical defor- macromolecules from the pulp into the pre-
mation of A-delta nerves is responsible dentin and dentin (HE staining x 40).
for brief, sharp pain that characterizes
dentin sensitivity (Närhi, 1990). The
central part of the dental pulp contains
cells, fibers, vessels, ground substance, and interstitial fluid
quite similar to that of other connective tissues. Tissue pres-
sure in the pulp (called pulpal pressure) is the result of vas-
cular pressure, and recent research indicates that normal pul-
pal pressure is about 15 cm H2O (Ciucchi et al., 1995). In the
presence of inflammation, the pulpal vascular beds become
more permeable, leading to localized increases in tissue pres-
sure and increased pain (Heyeraas and Berggreen, 1999). The
increased permeability of pulpal blood vessels also promotes
the release of blood plasma proteins into the pulp, and
inflamed pulpal fluid is therefore more protein-rich than nor-
mal pulpal fluid. These plasma proteins (mostly albumin and
globulin) may bind or agglutinate inside the dentinal tubules
(Hanks et al., 1994).
After tooth formation is complete, the odontoblasts
maintain the dentin and continuously and slowly deposit
and mineralize new secondary dentin. The secretion of sec-
ondary dentin occurs rhythmically, with a daily rate of
approximately 5 microns. If the odontoblasts are irritated by
trauma, bacterial infection, or material degradation products,
the odontoblasts form tertiary dentin over 4-6 weeks. The
exact cellular source and induction mechanisms of the pro-
duction of tertiary dentin are not fully understood, but if the
insult that caused the damage is removed before pulpal
necrosis occurs, then the formation of tertiary dentin re-
establishes a barrier between the insult and the pulp (Baume,
1980). Thus, tertiary dentin formation, which is much faster
than secondary dentin formation, is regarded as an impor-
tant defense mechanism of the pulp-dentin complex in
response to either pathological or physiological insults (Fig.
4). The long-term evolution and treatment of the dentin-pulp
complex are central considerations of most dental restorative
procedures, but are becoming especially important in older
Figure 4. Tertiary dentin. Tertiary dentin for-
mation (arrows) is regarded as an important
defense mechanism of the pulp-dentin com-
plex in response to either pathological or
physiological insults. The presence of tertiary
dentin reduces dentin permeability.
patients, in whom pulpal insults are longstanding and repar-
ative processes are much less effective (Burke and
Samarawickrama, 1995).
(2.2) RESIN-BASED MATERIALS
AND POLYMERIZATION
The BisGMA molecule is the basis for most current resin-
based materials. Several reviews of the composition and
properties of current composite restorative materials have
been recently published (Ferracane, 1995). Composites are a
mixture of a polymerized resin network reinforced by a glassy
filler. The polymer is formed by polymerization of monomers
like BisGMA, urethane dimethacrylate (UDMA), and triethyl-
ene glycol dimethacrylate (TEGDMA), among others.
Monomers may be slightly soluble in water, but are common-
ly quite hydrophobic. Dentinal bonding agents used to bond
composite resins to tooth substrates often contain monomers
similar to those in the composite, but nearly all contain or use
hydroxyethyl methacrylate (HEMA). HEMA is amphoteric
and displaces water in the dentin but is also miscible with
most of the monomers of the composite. The biocompatibility
of dentin-bonding agents is imperative, since they are placed
on etched dentin near the pulp, where tubular density and
diameter are greatest. Bonding agents are also at greatest risk
for incomplete cure, since they are thin and oxygen inhibition
of polymerization is a significant factor (Rueggeberg and
Margeson, 1990).
Light-activated polymerization for contemporary compos-
ites and adhesives is accomplished with the use of blue light
between 450 and 500 nm in wavelength. Typically, 500-800
mW/cm2 of light for 30-40 sec (15-24 Jcm-2) is necessary to
polymerize an increment of composite, which must be suffi-
ciently thin to receive the full power density of the curing light.

15(1):47-60 (2004) Crit Rev Oral Biol Med 49

Although increments of 1 to 3 mm thick are generally used, it
is important to note that a complete polymerization is never
achieved. Theoretically, a 100% conversion of monomer to
polymer is possible, but as much as 25 to 50% of the methacry-
late monomer double-bonds actually remains unreacted in the
polymer (Asmussen, 1982; Imazato et al., 2001). Any unpoly-
merized monomer in the composite is a potential biological lia-
bility if it leaches from the composite toward the pulp of the
tooth (Hume and Gerzina, 1996). More recent evidence also
suggests that extracellular or salivary enzymes may degrade
polymerized networks over time, making the hydrolyzed
products available to tissues (Santerre et al., 2001).
The shrinkage that accompanies polymerization of con-
temporary composites is a significant problem to the overall
biocompatibility of these materials. Nearly all composites
shrink linearly from 0.6-1.4%, depending on the type of com-
posite, the rate of cure, and the amount and nature of the filler
(De Gee et al., 1993; Davidson and De Gee, 2000). Although
shrinkage has been substantially reduced in modern composite
formulations (Labella et al., 1999), shrinkage places stress on
any bonds that have been formed between the restoration and
the tooth (Davidson et al., 1984). If these bonds are broken, then
a gap will form that will allow for percolation of bacterial prod-
ucts into the restoration.
The risk of biological harm from degraded or unpoly-
merized monomers is dependent on several key factors.
First, the component must be free of the polymer to diffuse
into the pulpal tissues. Second, the component must have
properties, such as solubility, that encourage its diffusion
into the pulp. Third, the time and dose of the pulpal expo-
sure must be sufficient to cause a biological reaction, and
finally, the component must have biological properties in
cells that cause problems.
(2.3) BIOCOMPATIBILITY CONCEPTS
Williams has defined biocompatibility as the ability of a
material to perform with an appropriate host response in a
specific application (Williams, 1990). This definition assumes
a risk-benefit balance that needs to be evaluated. The first
step in the assessment of risk is to determine the hazard
posed by components of resin-based restorative materials.
Dose-response assessment is a key step in hazard identifica-
tion. This assessment is achieved with in vitro cytotoxicity
tests, tests for inflammation, tests for immune response,
genotoxic (mutagenicity), and, finally, gene expression in
odontoblast-like cell lines (Hanks et al., 1996). The
ADA/ANSI Doc. 41 (1982) and ISO 10993 (1993) describe
these different tests. The second step in risk assessment is to
determine the doses of the chemicals that will be released by
the material. For adhesive resins, a "dentin-barrier test" has
been developed to determine the concentrations of compo-
nents of dental materials that might reach pulpal tissues
(Hanks et al., 1988). The second tier of tests also includes
intracutaneous reactivity, skin sensitization, and dental
usage tests. Characterizing the risk constitutes the final step
of the process. The dose response is compared with the esti-
mated dose exposure: If the dose to cause an adverse
response is greater than the estimated exposure by a com-
fortable safety margin, the likelihood of an adverse event
occurring in an exposed population is small, and the materi-
al may be deemed to have a low risk of biological problems.
50 Crit Rev Oral Biol Med
Although a few in vivo studies have attempted to docu-
ment the biological risks of resin-based materials, most infor-
mation on the hazards posed by the components of resin-
based restorative materials has been gained from in vitro
studies. As early as 1991, Hanks et al. reported the toxic con-
centrations of 11 components of dental resins on mouse fibro-
blasts. Later, Ratanasathien et al. (1995) evaluated the effects
of simultaneous exposures of cells to several resins. They
demonstrated the additive cytotoxic effects produced by
HEMA when used as a solvent for BisGMA. The synergism
between these 2 molecules has been shown to affect the
apparent toxicity of each individual resin component for the
cultured cells. These unique experiments established that
resins or combinations of resins alter fibroblast mitochondri-
al activity. Rakich et al. (1999) demonstrated that resin
monomers are also a hazard to inflammatory cells that are
common in the pulpal tissue, and Noda et al. (2003) have
shown that resins alter the secretion of inflammatory media-
tors from human macrophages. Other studies have shown
that HEMA is able to diffuse rapidly through dentin in vitro
in sufficient concentrations to cause cytotoxicity (Bouillaguet
et al., 1996), and that bonding agents, as used clinically, elute
sufficient amounts of monomer through dentin to cause sig-
nificant cellular toxicity after 1 wk (Bouillaguet et al., 1998).
The persistent cytotoxicity observed after 1 wk reinforced the
need for evaluation of the long-term effects of the resin
monomers on cellular systems. Indeed, long-term studies
that used sublethal concentrations of HEMA (Bouillaguet et
al., 2000a), TEGDMA, or BisGMA (Lefebvre et al., 1999) for 5-
6 wks showed that resins clearly altered cellular mitochond-
rial activity and total protein content per cell, even at concen-
trations of 1-10% of those used in short-term experiments.
These results confirmed that risk assessment of dentin adhe-
sives must also be considered with a long-term view.
(3) The Microbiological Risks
of Resin-Tooth Restorations
Post-operative sensitivity, pulpitis, and secondary caries are the
three major post-operative problems known to occur after the
placement of resin-based restorations (Mjör et al., 2000). Post-
operative sensitivity is presumably caused by minute fluid
movements through open or unsealed tubules which are acti-
vated by temperature, osmotic changes, or by occlusal loads
(Pashley et al., 1996; Paphangkorakit and Osborn, 2000).
Pulpitis and pulpal necrosis can occur because of the chemical
risks of the materials but are more likely to occur when micro-
organisms penetrate the gap formed as a result of resin poly-
merization shrinkage (Bergenholtz, 2000). Secondary caries
results when bacterial colonization of marginal gaps allows for
the dissolution of tooth structure. All of these clinical problems
are eliminated or greatly reduced when the dentin or enamel is
impregnated with resins, thereby eliminating marginal gaps
and leakage and effectively sealing the tooth (Nakabayashi and
Pashley, 1998).
(3.1) BONDING AND SEALING
Buonocore's acid-etch technique, introduced in 1955 and
refined later by Silverstone (1975), has been shown to provide
a good seal between resin-based materials and etched enamel.
This seal is a result of the penetration of an adhesive into
microporosities created by differential etching of enamel
15(1):47-60 (2004)

Figure 5. Bonding resin-based materials to enamel. Acid-etching of
enamel prior to adhesive application allows for a good wetting of
the surface by the hydrophobic resin and a good penetration into the
microporosities created by the acid (orig. mag. x 2400).
prisms. The goals of enamel etching are to clean the enamel
and to remove the enamel smear layer. Etching results in a
high-energy surface that allows for good wetting by the
hydrophobic bonding resin and good penetration of the resin
into the microporosities (Fig. 5).
Whereas bonding of resin-based materials to acid-etched
enamel has become routine and reliable, different, more com-
plex procedures have been required for bonding to dentin
because of the completely different nature of the dentin sub-
strate. Bonding to dentin is further complicated by the forma-
tion of a smear layer during cavity preparation (Pashley, 1989).
However, good dentin bonding and sealing are possible with
the use of adhesives with complex chemistries. Use of these
adhesives requires multi-step and demanding attention to
clinical details (Van Meerbeek et al., 1998a). Two categories of
adhesive systems are currently available: total-etching and
self-etching adhesives.
Total-etching adhesives
These require relatively high concentrations of acids (32-37%
phosphoric acid) applied to dentin in a separate etching step.
After 15 sec of etching, a water rinse removes the acid and dis-
solved mineral and leaves the acid-insoluble collagen fibers
suspended in the water. This collagen network is highly
hydrophilic and particularly sensitive to dehydration and
shrinkage (Pashley et al., 1993). The next step in the bonding
process is to embed these fibrils with resins. One approach is
to use an aqueous solution of hydrophilic monomers such as
HEMA in an intermediate step called priming (Nakabayashi
and Takarada, 1992). When gently dried with air, the HEMA-
water-collagen mixture will slowly dehydrate but will remain
fully expanded to allow for the subsequent incorporation of
the adhesive resin (Pashley et al., 2000). This bonding strategy
is used by the so-called three-step total-etching adhesive sys-
tems (Fig. 6).
Some manufacturers have developed "one-bottle" adhe-
sives that contain mixtures of organic solvents and resins
(HEMA, BisGMA, TEGDMA, UDMA) to impregnate the col-
lagen-water network. These organic solvents (acetone or
alcohol) quickly displace water in the collagen network,
because the driving force for water removal is greater than
15(1):47-60 (2004) Crit Rev Oral Biol Med
Figure 6. Modified SEM illustration
of bonding to dentin with conven-
tional (three-step) total-etching adhe-
sives. (A) Acid etching of the dentin.
The smear layer has been removed,
and both peritubular and intertubu-
lar dentin is demineralized. The
exposed collagen fibers are highly
hydrophilic (blue) and particularly
sensitive to dehydration. (B) Priming
of the dentin. The water has been
replaced by hydrophilic primers
(orange) which have impregnated
the collagen fibers. Priming with water-based primers is a slow dif-
fusing process. The evaporation of the water solvent will leave the
collagen fibers coated and stiffened by the resins. The substrate has
changed from hydrophilic to hydrophobic. (C) Application of the
adhesive resin. The hydrophobic resin (red) diffuses into the denti-
nal tubules and impregnates the entire depth of the demineralized
dentin before being polymerized.
with the HEMA-water primers (Fig. 7). Therefore, these mix-
tures achieve a dynamic dehydration, because the stiffening
of the collagen fibers and the incorporation of the bonding
resin occur simultaneously (Maciel et al., 1996). However,
recent research indicates that one-bottle adhesives increase
the shrinkage of wet-decalcified dentin, thereby reducing
infiltration of resin monomers (Nakajima et al., 2002). The
Figure 7. Modified SEM image of bonding to dentin with one-bottle
(two-step) total-etching adhesives. (A) Acid-etched dentin (blue). (B)
Application of the primer/adhesive mixture. The organic solvents
used in these mixtures (green) quickly displace water in the collagen
network because of the diffusion gradient they create between the
water of the collagen and solvent in the bonding agent. Therefore, the
incorporation of the primer and the bonding resin inside the collagen
network can occur simultaneously.
51

Figure 8. Modified SEM image of bonding to dentin with (two-step) self-etching adhesives. (A) The smear-layer
covering the dentin surface and the smear-plugs occluding the dentinal tubules are impregnated by the self-etch-
ing primer (light blue). (B) Application of the acidic primer. The acidic resin (blue) has dissolved and impregnat-
ed the smear layer. The resin has also penetrated the dentinal tubules. (C) Dentinal substrate after application of resin-based materials
the adhesive resin. Theoretically, the adhesive resin (purple) infiltrates to the same extent as the acidic primer
exposed the collagen.
advantage of these systems is the elimination of priming as a
separate step, simplification of the procedure, and savings in
clinical time.
Whether a separate priming step is used or not, when
adhesive resins penetrate the intertubular demineralized
dentin and polymerize around the collagen fibrils, they form
the so-called 'hybrid layer' (Nakabayashi et al., 1982). Dentin
hybridization also occurs at the periphery of dentin tubules,
where the peritubular dentin was dissolved and resin plugs are
formed. This process is referred to as the hybridization of the
resin tag (Nakabayashi and Pashley, 1998). The intimate
hybridization of both the intertubular and peritubular dentin
contributes to the sealing and bonding of resin-based materials
to dentin.
Self-etching adhesives
These are an alternative clinical approach to total-etching
systems. Self-etching adhesives contain acidic monomers
combined with hydrophilic monomers that simultaneously
etch and prime the dentin. For most systems of this nature,
the etch-prime step is followed by the application of the
adhesive resin. Theoretically, the adhesive resin infiltrates to
the same depth as the acidic primer exposed the collagen in
dentin (Fig. 8). This hypothesis has been recently confirmed
by laser Raman microscopy (Miyazaki et al., 2002). Because
self-etching adhesives eliminate the rinsing of the etchant
and the drying of the water necessary in the total-etching sys-
tems, they are simpler to use and may provide more consis-
tent clinical results. Hybrid layers formed by self-etching
adhesives on sound dentin are generally thinner than those
produced by total-etching systems. Further, the resin tags are
shallower, and the sealing and bonding may rely mostly on
intertubular hybridization in normal dentin (Inoue et al.,
2000). Self-etching adhesives do not bond as well to enamel
as do total-etching systems, and recent research indicates that
the quality of the resin-dentin bonds formed by such adhe-
sives is directly related to the aggressiveness of the system
(Tay and Pashley, 2001). Manufacturers are currently trying
to perfect new adhesive systems that condense etching, prim-
ing, and bonding into a single step. These 'all-in-one' adhe-
sives are in their infancy and will likely undergo significant
evolution in coming years.
52 Crit Rev Oral Biol Med
There are differ-
ent methods for mea-
suring bonding and
sealing of resins to
dentin. Ciucchi et al.
(1997a) evaluated the
size and volume of the
gap formed in vitro
between resin-based
materials and dentin
during polymeriza-
tion. Their results
clearly showed that
that bond to dentin
had the smallest gaps.
However, none of the
materials was without
some gaps, indicating
that polymerization
shrinkage forces exceeded the dentin bond strengths in at
least some areas of the restoration. Other studies have direct-
ly measured the ability of dentin adhesives to limit fluid flow
through dentin and therefore seal the dentin tubules
(Bouillaguet et al., 2000b). The results showed that no mater-
ial completely sealed the dentin, but that most contemporary
adhesive systems significantly reduced fluid movement (by
> 95% in many cases). Collectively, these studies indicate that
dentin-resin bonds are critical to maintain a seal and to resist
polymerization shrinkage stresses, thereby limiting the
microbiological risks. Another method for measuring the
ability of adhesives to resist polymerization shrinkage forces
assesses their microtensile bond strengths to dentin (Sano et
al., 1994a). A microtensile bond test is used because it is the
most accurate measure of composite-dentin bond strength
(Pashley et al., 1999). Generally, for the comparison of mate-
rials, dentin adhesives are applied on flat dentin to avoid the
influence of cavity geometry on bonding. Recent research has
indicated that the conventional three-step total-etching adhe-
sives were best able to bond composite to dentin under these
circumstances. The two-step total-etching system and the
self-etching system gave comparable bond strengths, but the
one-step self-etching system was not as reliable as the other
systems (Bouillaguet et al., 2001a; Inoue et al., 2001). Almost
all in vitro bonding studies are done on flat dentin surfaces,
yet, clinically, such surfaces are seldom encountered. Other
studies have evaluated the influence of cavity geometry on
bonding, in Class I or Class II MOD cavities, and showed that
bond strengths to cavity walls were reduced by 20% com-
pared with flat dentin, where no polymerization stress is pre-
sent (Yoshikawa et al., 1999; Bouillaguet et al., 2001b). The
authors cautioned in interpreting bond strengths obtained on
flat surfaces, because these studies probably overestimate
dentin bond strengths in most cases. This point is important,
since the flat system is often used by manufacturers to pro-
mote their products. The ability of the materials to bond to
dentin is further compromised by their often complex and
technique-sensitive nature. The ability of an operator to
negotiate these complexities is therefore an important factor
in the successful management of these materials. Few studies
have investigated the influence of the operator on the quali-
ty of resin-dentin bonds. However, there is increasing evi-
15(1):47-60 (2004)
dence that the influence of the operator is of paramount
importance in the performance of dentin-bonding agents
(Ciucchi et al., 1997b; Finger and Balkenhol, 1999; Bouillaguet
et al., 2002).
Studies show that contemporary dentin adhesives have the
potential to provide a good, but not complete, seal of the
dentin. The type of product is important, as is the configuration
of the cavity preparation with respect to the ultimate bond
strength and seal obtained. The clinical environment is com-
plex and often compromises the conditions necessary to obtain
the best dentin seals and the lowest microbiological risks. Thus,
the risk of microbiological contamination remains in the clini-
cal situation.
(3.2) MICROLEAKAGE AND NANOLEAKAGE
Failure of dentin adhesives to seal the dentin and the enamel
results in microleakage or nanoleakage. Leakage has been
shown to occur at the margins of the restoration, but may also
be limited to internal aspects of the restoration. Thus, both the
marginal (peripheral) seal and the internal dentinal seal are
important to the longevity of resin-based restorations.
Although a few in vivo studies have attempted to document
the presence of leakage (Ryge, 1981), most information on
microleakage has been gained from in vitro studies. As
defined by Kidd (1996), microleakage is the passage of bacte-
ria, fluids, molecules, or ions between a cavity wall and the
restorative material. Microleakage gaps are many microme-
ters wide and result from either a lack of primary bonding or
the secondary loss of bonding. Primary bonding may be lost
with time because of occlusal forces or hydrolytic degrada-
tion. However, the most likely cause of microleakage is from
the volumetric shrinkage that occurs concurrently with poly-
merization of the resin. If the resin-tooth bond is too weak,
polymerization forces will debond the resin from the tooth,
and microleakage will result. The ability of the resin-tooth
bond to resist polymerization shrinkage forces depends on
many complex and interacting factors. The nature of the resin
shrinkage first depends on the shape of the cavity preparation
and the ratio of bonded to unbonded (or free) surfaces (Feilzer
et al., 1987; Davidson and Feilzer, 1997). This so-called C-fac-
tor is a clinically relevant predictor of the risk of microleakage
development. Restorations with high C-factors (> 3.0) are at
greatest risk for debonding and microleakage (Yoshikawa et
al., 1999). The stress at the tooth-resin interface is also influ-
enced by the kinetics of the polymerization reaction. A resin-
based material will flow plastically to accommodate shrink-
age until it reaches the so-called gel-point, after which flow
cannot occur and the stress of polymerization contraction will
be directly transmitted to the tooth-resin interface. If the cur-
ing is done rapidly, as with high-intensity curing units, then
the gel-point is reached earlier and more shrinkage stress is
transmitted to the interface. Therefore, high polymerization
rates are more likely to cause debonding and microleakage
(Yoshikawa et al., 2001).
Unlike microleakage, nanoleakage may result even when
the bond between the tooth and resin is intact. If the adhesive
resin does not completely infiltrate the demineralized dentin,
some of the collagen network will contain small nanospaces
between the hybrid layer and the mineralized dentin. These
spaces have been verified by experiments with silver nitrate
and scanning electron microscopy (Sano et al., 1994b). The
15(1):47-60 (2004) Crit Rev Oral Biol Med
spaces appear to be contiguous because the silver nitrate can
diffuse well into interface, even when no interfacial gap
(microleakage) is present. Sano and co-workers coined the
term "nanoleakage" to distinguish this type of leakage from
microleakage (Sano et al., 1995). All adhesive systems exhibit
some degree of nanoleakage, although some systems have
more nanoleakage than others.
In total-etching systems, the water used to rinse the acid
must be removed with air before priming occurs. If too little
water is removed, then bonding is compromised, because the
primer and adhesive resin cannot penetrate the hydrophilic
environment (Tay et al., 1996) and cannot polymerize. If too
much water is removed, then the collagen network will col-
lapse and will not be effectively infiltrated by the primer or
adhesive resin. To a lesser degree, primers that use organic sol-
vents such as acetone also cause a shrinkage of the network
(Nakajima et al., 2002). Any factor that limits infiltration of the
collagen-water network by resin results in at least some
nanoleakage. Nanoleakage may also result from the incom-
plete diffusion of high-molecular-weight resin monomers into
the primed collagen network, simply because of inadequate
time for the diffusion to occur. If the resin adhesive contains
fillers, then its ability to penetrate the network is further com-
promised. Regions of demineralized dentin that have not been
successfully embedded with resin have been implicated as
weak links in dentin-resin bonding. Furthermore, the exposed
collagen network may make the resin-collagen hybrid layer
more susceptible to hydrolytic degradation over the long term
(De Munck et al., 2003).
With self-etching systems, the risk of nanoleakage is
lower, because the acidic monomer that etches the dentin is
also the primer. Thus, the adhesive resin is more likely to infil-
trate to the complete depth of the etching. However, traces of
acid or solvent may remain impregnated within the adhesive
and subsequently inhibit the polymerization of the
monomers. Further, recent research indicates that the newer
self-etching adhesives are semi-permeable membranes
because of the high hydrophilicity of these resins (Tay et al.,
2002, 2003). The existence of both microleakage and nanoleak-
age has been well-documented. It is clear that microleakage
has deleterious consequences for resin-based restorations by
greatly increasing the microbiological risks (Bergenholtz,
2000). However, the clinical consequences of nanoleakage are
less clearly understood.
(4) Clinical Perspectives on Resins
and Resin-bonding/sealing
Adhesive resins can be used safely for numerous clinical appli-
cations if care is taken to control substrates, chemistry, and
polymerization. Further, adhesive systems have the potential
to seal restorations and consequently to offer an effective pro-
tection to the dentin-pulp complex against microbiological
risks. This potential, however, is not realized because of the
complexity of the bonding procedures, which are often poorly
understood by the average clinician.
(4.1) CLINICAL PERSPECTIVES ON RESIN TOXICITY
(MINIMIZING DIRECT BIOLOGICAL RISKS)
Although factors (such as remaining dentin thickness, dentin
permeability, and dentin location) that alter the diffusion and
influence the toxicity of resins have been identified, one funda-
53
mental clinical problem is that a dentist has only a subjective
idea of these factors when preparing a cavity for a resin-based
restoration. Fortunately, some clinical recommendations can be
made to minimize direct biological risks.
Cavity preparation
Cavity preparations in vital teeth are usually performed under
local anesthesia. Local anesthetics contain vasoconstrictors that
may compromise pulpal blood clearance that is normally very
efficient (Kim and Dörscher-Kim, 1990). The dentinal fluid
pressure, which is normally outward from the pulp and tends
to reduce ingress of substances, is therefore reduced. Thus,
direct biological risks of resin-based materials may increase sig-
nificantly during cavity preparation. Treatment of caries lesions
involves removal of infected tissues but requires the preserva-
tion of the hypermineralized dentin (transparent layer) located
at the front of the lesion. This layer is much less permeable than
tubular dentin and therefore offers much more resistance to the
diffusion of materials toward the pulp.
Selecting an adhesive
To minimize direct biological risks associated with the use of
resin-based materials, one should carefully evaluate the biolog-
ical risks of each adhesive system under relevant clinical con-
ditions. Shallow cavities located in superficial or sclerotic
dentin do not pose a major biological risk, because the perme-
ability of the dentin is low and the thickness of the remaining
dentin is adequate to prevent any adverse effects from diffus-
ing materials (Mjör and Ferrari, 2002). Therefore, total-etching
adhesives that provide reliable bonding to enamel and dentin
are recommended. On the other hand, deep cavities closer to
the pulp are more challenging for the clinician because of the
intrinsic permeability and wetness of the dentinal substrate.
Gwinnett and Tay (1998) observed a persistent inflammation
and granulomatous reaction in human pulp in response to the
application of a total-etching adhesive to deep dentin. They
also reported the presence of resin globules displaced into the
dentin tubules and penetrating the pulp. In deep dentin, the
etching as a preliminary step of the bonding process will make
the substrate even more permeable and hydrophilic. Increased
hydrophilicity limits the wetting of the tubule wall by the
resins, may allow the dentin surface to be contaminated by
dentinal fluid, or may interfere with the polymerization
process. Therefore, the use of self-etching adhesives systems is
indicated for young, deep, permeable dentin, because self-etch-
ing adhesives often leave some residual smear plug material in
the tubules which limits the diffusion of uncured monomers
toward the pulp (Tay et al., 2000).
Conversion of monomers
It is generally accepted that the better the polymerization, the
lower the biological risks (Kaga et al., 2001). Clinically, the
polymerization of resin-based materials is achieved with light
energy, and there is a great deal of interest in developing high-
power curing units. With these units, the dentist can cure
faster and the material may have better biological properties,
because increased conversion rates of monomer to polymer
are expected. However, the use of high-output energy lights is
controversial, because it is not clear if the energy emitted by
the unit is totally absorbed by the photo-initiators (CQ or
PPD) to initiate polymerization. Wavelengths outside those
54 Crit Rev Oral Biol Med
necessary to activate these photo-initiators do not improve
the cure of the resin, but do increase the overall risk to the
pulp from secondary generation of heat (Hannig and Bott,
1999). Most recent developments in light-curing units are
focusing on blue-emitting diodes (LEDs), which do not gen-
erate heat (Nomura et al., 2002). However, temperature rise
may also occur because of the exothermic polymerization of
the composite material.
Long-term degradation
The long-term clinical degradation of resin-based materials
and dental adhesives is not known in detail but has been
reported for some materials (Hashimoto et al., 2000). Clinically,
the long-term toxic effects of resin-based materials are extreme-
ly difficult to distinguish from the effects of microleakage and
bacterial contamination. It is likely that both factors contribute
to pulpal stress and disease. Furthermore, it is likely that, clin-
ically, pulpal cells that have chronically suffered from exposure
to toxic components of resins will respond differently to bacte-
rial challenge compared with healthy cells. In vitro evidence
suggests that these interactions between resin components and
bacterial products may increase or decrease the body's ability
to respond appropriately (Rakich et al., 1999). Such interactions
are critical to the biocompatibility of any material, and such
data are not clinically available. Ideally, the dentist would like
to know the inflammatory status of the pulp and the history of
exposure to components of material or bacteria. These factors
are significant to the patient, because the death of an over-
stressed pulp leads inevitably to pain and significant restora-
tive preparation time and costs.
(4.2) CLINICAL PERSPECTIVES ON RESIN-TOOTH
BONDING (MINIMIZING MICROBIOLOGICAL RISKS)
Clinically, the biggest problem with bonding resin-based mate-
rials to teeth is that the clinician will have no indication of how
successful the bond is until many years later. There is no way
for the clinician to measure the strength of the bond, the seal of
the dentin, or the presence of bacteria beneath the restoration.
Yet each of these factors is critical to the longevity and overall
success of the restoration. This section will focus on the clinical
strategies used to minimize the microbiological risks with
resin-based restorations.
Complete removal of micro-organisms
A preliminary step in the placement of resin-based restora-
tions is the complete removal of micro-organisms inside the
cavity. This requirement is based on the concept that bacterial
infection or re-infection from residual micro-organisms
beneath the restoration induces pulpal inflammation and
necrosis (Bergenholtz, 2000). Bacterial removal has to be bal-
anced with the conservation of the inner part of the caries
lesion (transparent layer). Caries-disclosing solutions are used
for this purpose. However, these dyes cannot detect bacteria
within the dentin tubules. Thus, caries-disclosing solutions are
useful but cannot guarantee the complete elimination of bac-
teria from a cavity preparation. Current clinical practice also
advocates the use of dental rubber dam to avoid the bacterial
contamination of the cavity that may occur from outside the
cavity preparation (e.g., saliva). Aside from a better control of
the operating field, the use of a rubber dam has a positive
effect on the quality of some adhesive systems (Hitmi et al.,
15(1):47-60 (2004)
1999). However, the prevention of external contamination of
the cavity preparation is often not straightforward, because
bacterial contamination may also come from water lines of
dental units (Tonetti-Eberle et al., 2001).
Using disinfectants
Previous studies have shown that rinsing cavity surfaces with
sodium hypochlorite solutions (3-10%) or hydrogen peroxide
(3%) reduces bacterial load. Sodium hypochlorite has proteo-
lytic properties, hydrogen peroxide is oxidative, and both dif-
fuse through dentin (Hanks et al., 1994). Thus, these agents may
kill bacteria within dentin tubules, but may also carry a certain
biological risk of their own (Costa et al., 2001). Because there is
some evidence that cavity disinfectants such as hypochlorite
interfere with bonding and polymerization of resins, the rou-
tine use of these agents is not recommended (Lai et al., 2001;
Osorio et al., 2002). Furthermore, acid-etching and self-etching
resins are probably bactericidal to some degree, because most
bacteria cannot survive in extremely low pH conditions
(Murray et al., 2002). Therefore, cavity disinfection with
hypochlorite or peroxide may be superfluous.
Embedding bacteria with resins
For many years, controversy has raged about the ability of
residual bacteria to survive or multiply in a cavity prepara-
tion sealed with resins. However, the work of Mertz-Fairhurst
and co-workers (1995) clearly demonstrated that Class I caries
can be arrested by the placement of sealed posterior compos-
ite restorations on top of the caries lesions without the
removal of the caries lesion. The results of this study serious-
ly challenged the need for cavity disinfection if a sealed
restoration can be obtained. However, a complete seal of the
cavity may be compromised by infected dentin, poor bond-
ing, and polymerization shrinkage. Therefore, relying on the
integrity of the seal to limit bacterial growth may not always
be wise in practical terms.
Selecting an adhesive system
Bacterial leakage and sealing of dentin are interdependent, and
good sealing always results in a lower microbiological risk.
Many reports have confirmed the superiority of total-etching
adhesives over self-etching adhesives in terms of bond strength
to dentin (Van Meerbeek et al., 2001). This superiority has also
been confirmed for enamel bonding and resin bonding to scle-
rotic and caries-affected dentin (Inoue et al., 2001). Therefore,
total-etching adhesives are the material of choice for most clin-
ical applications. However, inadequate bonding with total-
etching systems can be observed when resin penetration is
incomplete. Clinically, the biggest drawback of total-etching
systems is the control of moisture. Achieving the appropriate
amount of dentin wetness causes much of the clinical confu-
sion. Overdrying or overwetting the tooth will significantly
compromise the quality of the resin bond to dentin (Van
Meerbeek et al., 1998b). These decreased bond strengths are
caused primarily by a decreased intertubular permeability of
dentin to adhesives (Fig. 9). Adhesives have various abilities to
accommodate overwettness or overdryness. Water-ethanol sys-
tems are favorable in this regard compared with acetone-based
systems. Water-ethanol systems are therefore considered more
user-friendly (Perdigão and Frankenberger, 2001).
Because they eliminate the need for a separate etching-
15(1):47-60 (2004) Crit Rev Oral Biol Med
Figure 9. Defective bonding. SEM micrograph of a specimen that
was overdried after the etching gel was rinsed off. In such cases, the
adhesive resin cannot penetrate the demineralized dentin because of
the collapse of the collagen network (arrows) (orig. mag. x 10,000).
rinsing step, self-etching adhesives are less sensitive to mois-
ture conditions than are total-etch systems. This fact has been
the primary driving force for the development and clinical use
of the self-etching systems. However, most current research
also agrees that the quality of self-etching bonds to enamel,
sclerotic dentin, and caries-affected dentin is inferior to that
obtained by a total-etching system (Yoshiyama et al., 2002).
Although the cause of poorer bonding is not completely
known, it is likely that the relatively weak acidity of the acidic
primer plays a role, because this weaker acid would not etch
these substrates as well as would phosphoric acid (Tay et al.,
2000).
In addition to the wetness of the dentin, the thickness of
the adhesive layer contributes to the strength and durability of
the bonds (Abdalla and Davidson, 1993). The adhesive resin
should be spread uniformly onto surfaces, with an optimal
thickness to provide sealing and to act as a stress absorber dur-
ing composite shrinkage (Choi et al., 2000). Despite differences
among materials, research supports the concept of thick adhe-
sive layers acting as stress absorbers (Zheng et al., 2001).
Control for polymerization shrinkage stresses
The management of shrinkage stresses during polymerization
is a critical factor in the clinical performance of dental adhe-
sives. Poor management of the shrinkage stresses that develop
during the curing of the restorative material can cause the fail-
ure of a restoration that is otherwise well-managed and well-
placed. Among current adhesive materials, the shrinkage of the
resin is unavoidable to some degree. However, proper clinical
management can minimize the impact of polymerization
shrinkage on the clinical performance of the restoration. Two
factors in the management of polymerization shrinkage are the
method of curing and the manner in which composite is insert-
ed into the cavity.
Early in the development of resin-based materials, the
concept of incremental addition of material to the cavity, com-
bined with the use of the so-called "directed-cure", was pro-
posed as a clinical solution to volumetric shrinkage (Lutz et
al., 1992). In recent years, several new light-curing concepts
have been introduced with the goal of improving composite
properties and reducing stress from polymerization shrinkage
55
(Versluis, 2000). Pulse-delay curing relies on the concept that
an initial, low-energy pulse of curing light will start, but not
complete, the polymerization reaction. This slows polymer-
ization and allows shrinkage stresses to be dissipated by flow
of the material, before the gel point of the polymer has been
reached. After some time, a higher-energy light pulse is
applied to complete the polymerization (Sahafi et al., 2001).
Exponential curing is conceptually similar to pulse-delay cur-
ing, except that the application of light is continuous, with
intensities that are exponentially modulated from low to high.
There is some evidence that these strategies do reduce poly-
merization stresses (Bouschlicher and Rueggeberg, 2000). In
an effort to cure larger increments of composite faster, inves-
tigators have developed newer lights with high outputs.
Plasma-arc-curing (PAC) lights may emit 2000-2500 mW/cm2
and therefore are purported to cure composites in much
shorter times (3-10 sec). However, the shrinkage stresses dur-
ing curing are probably higher, because the composite reach-
es its gel-point early in the polymerization process, and all
subsequent shrinkage stress is then transferred to the resin-
tooth interface. Finally, one must remember that the self-cur-
ing composites (also called chemical curing) offer the clinical
advantages of a relatively slow cure rate (therefore limiting
shrinkage stress) and a complete cure independent of cavity
dimension, depth, or accessibility to light (Feilzer et al., 1993).
Therefore, self-cure resins have been used by some practition-
ers in combination with more superficial layers of light-cured
resins.
A variety of restorative techniques has been used clinical-
ly to control polymerization shrinkage stresses. These tech-
niques can be divided into direct and indirect strategies. The
direct method cures the composite in situ, whereas indirect
methods fabricate and cure most of the bulk of the restoration
in a model or die of some type. The motivation for the indirect
technique is that the composite can be cured with times,
intensities of light, and temperatures that would not be possi-
ble clinically. In this manner, it has been proposed that most
of the shrinkage occurs before the restoration is cemented.
The problem with this technique is that a tremendous poly-
merization stress occurs within the luting resin cement,
because an extremely high C-factor may occur if the cavity
design is not appropriate. Further, previous research indicates
that, in a rigid situation (e.g., inlay cementation), the contrac-
tion stresses that develop during cementation are strongly
related to the resin layer thickness and the compliance of the
substrate (Alster et al., 1995, 1997). Thus, ironically, indirect
strategies may be worse than direct strategies in terms of
polymerization stresses. The indirect method may be success-
ful if the cavity can be designed to maximize free surfaces or
the inlay can be fabricated to allow for some free cementation
space. The dual-bonding technique has been used to cement
indirect restorations. In this technique, the clinician must pro-
tect the pulp of the tooth during the time the inlay is being
fabricated. The adhesive layer is applied to the cavity surfaces
before an impression is taken for the fabrication of a laborato-
ry-made restoration. Thus, the dentin is sealed and the pulp is
protected against bacterial leakage, thereby reducing
microbiological risks (Paul and Schärer, 1997). The restoration
is then luted with adhesive resins during the second visit. The
use of slow-curing cements (dual-curing cements) will help to
reduce the polymerization stresses during cementation,
although some clinicians recommend only light-cured
56 Crit Rev Oral Biol Med
cements for indirect restorations.
In direct restorations, several techniques have been used
to limit polymerization stresses, thereby reducing microbio-
logical risks. Generally, a total bonding strategy is used in
these types of restorations. For the total bonding concept, the
entire cavity surface is covered with the adhesive, and the fill-
ing material is incrementally polymerized onto it. The adhe-
sive layer is thick enough to absorb polymerization shrinkage
stresses. Choi and co-workers (2000) have reported that stress
was significantly absorbed and relieved by the application of
an increasing thickness of low-stiffness adhesive. Some clini-
cians have also advocated the use of flowable materials at the
base of the restoration to absorb these stresses. In general, by
carefully curing the different increments of composites inside
a low configuration factor cavity, the clinician can maintain
stresses at a low level using this technique. When the configu-
ration factor is higher (e.g., in Class 1 or 5 cavities), shrinkage
stresses increase the risk that polymerization stresses will put
the integrity of the restoration at risk. Under such conditions,
the use of the selective bonding concept may be indicated
(Krejci and Stavridakis, 2000). The concept of selective bond-
ing is to pre-determine the location of the failure in case of
excessive stress. The goal is for the dentin to remain sealed
even if polymerization stresses become acute in an area of the
internal part of the restoration.
(5) Perspective on Bonding Resins
to the Dentin-Pulp Complex
The various issues that influence the biocompatibility of adhe-
sive materials are complex and interactive and not fully under-
stood. However, there are several developments in evaluation
methods, clinical techniques, and materials that may help us
better estimate these risks and improve the reliability of resin-
based restorations. One significant problem with today's eval-
uation of biological risks is the inability of current in vivo or
animal tests to adequately predict the long-term response of
the human pulp. Improvements in in vitro tests, including tests
for leakage, could give them a greater potential to evaluate the
biological risks of new materials. The application of new mate-
rials and techniques will be optimized only if the dentist can
properly and rapidly diagnose the problems. The future will
probably include much-improved diagnostic tools, like new
techniques to detect caries. Another likely diagnostic tool will
probably focus on the measurement of the remaining dentin
thickness (RDT). The RDT is paramount to the selection and
clinical success of an appropriate treatment (About et al., 2001;
de Souza et al., 2003). In addition to new diagnostic methods,
new materials and dental adhesives are likely to be devel-
oped. Adhesives and adhesive strategies will likely be adapt-
ed or even customized to deal with a variety of bonding sub-
strates, including enamel, dentin, sclerotic dentin, and caries-
affected dentin. Newer adhesives will probably have new
chemistries, focus on both chemical and mechanical bonding,
be more water-resistant, easier to manipulate, and less sus-
ceptible to operator error. In addition to new adhesives, the
future will likely bring new resin restorative materials with
reduced polymerization shrinkage and shrinkage stress.
Some recent developments in dental composite research have
focused on the use of resin-based materials containing a mix-
ture of oxiranes and polyol that can polymerize by light acti-
vation (Eick et al., 2002). With these new chemical structures,
15(1):47-60 (2004)
suitable formulations can be designed for the development of
dental composites with acceptable mechanical and biological
properties.
Summary
Over the last decades, the development of resin-based materi-
als has provided the clinician with many techniques and mate-
rials with which to restore tooth structure, esthetics, and func-
tion. The clinical success of these new restorative techniques
has been attributed to the ability of resin-based materials to seal
the resin-tooth interface in the absence of any adverse biologi-
cal effect. Although recent literature indicates that the risks of
acute pulpal toxicity to resins are unlikely, it is clear that today's
tests are not adequate to predict long-term clinical biological
risks. The formation of a perfect seal around resin-based
restorations is further required to offer an effective protection
to the dentin-pulp complex against microbiological risks.
Although most adhesive systems have the potential to seal
restorations, research has shown that sealing of cavities with
resin-based materials is not always predictable. Fortunately,
there are several anticipated developments in evaluation meth-
ods, clinical techniques, and materials that may help us better
estimate these risks and improve the reliability of resin-based
restorations.
REFERENCES
Abdalla AL, Davidson CL (1993). Shear bond strength and
microleakage of new dentin bonding systems. Am J Dent 6:295-
298.
About I, Murray PE, Franquin JC, Remusat M, Smith AJ (2001). The
effect of cavity restoration variables on odontoblast cell num-
bers and dental repair. J Dent 29:109-117.
ADA/ANSI (1982). Council on Dental Materials, Instruments and
Equipment. Addendum to American National Standards/
American Dental Association Document No. 41 for recom-
mended standard practices for biological evaluation of dental
materials. Chicago, IL: ADA, pp. 1-63.
Alster D, Feilzer AJ, De Gee AJ, Davidson CL (1995). Tensile
strength of thin resin composite layers as a function of layer
thickness. J Dent Res 74:1745-1748.
Alster D, Venhoven BA, Feilzer AJ, Davidson CL (1997). Influence
of compliance of the substrate materials on polymerization con-
traction stress in thin resin composite layers. Biomaterials 18:337-
341.
Asmussen E (1982). Factors affecting the quantity of remaining
double bonds in restorative resin polymers. Scand J Dent Res
90:490-496.
Baume LJ (1980). The biology of pulp and dentine. A historic, ter-
minologic-taxonomic, histologic-biochemical, embryonic and
clinical survey. Monogr Oral Sci 8:1-220.
Bergenholtz G (2000). Evidence for bacterial causation of adverse
pulpal responses in resin-based dental restorations. Crit Rev
Oral Biol Med 11:467-480.
Bouillaguet S, Wataha JC, Hanks CT, Ciucchi B, Holz J (1996). In
vitro cytotoxicity and dentin permeability of HEMA. J Endod
22:244-248.
Bouillaguet S, Virgillito M, Wataha J, Ciucchi B, Holz J (1998). The
influence of dentine permeability on cytotoxicity of four den-
tine bonding systems, in vitro. J Oral Rehabil 25:45-51.
Bouillaguet S, Wataha JC, Virgillito M, Gonzalez L, Rakich DR,
Meyer JM (2000a). Effect of sub-lethal concentrations of HEMA
(2-hydroxyethyl methacrylate) on THP-1 human monocyte-
15(1):47-60 (2004) Crit Rev Oral Biol Med
macrophages, in vitro. Dent Mater 16:213-217.
Bouillaguet S, Duroux B, Ciucchi B, Sano H (2000b). Ability of
adhesive systems to seal dentin surfaces: an in vitro study. J
Adhes Dent 2:201-208.
Bouillaguet S, Gysi P, Wataha JC, Ciucchi B, Cattani M, Godin C, et
al. (2001a). Bond strength of composite to dentin using self-
etching, conventional and one step adhesive systems. J Dent
29:55-61.
Bouillaguet S, Ciucchi B, Jacoby T, Wataha JC, Pashley DH (2001b).
Bonding characteristics to dentin walls of class II cavities, in
vitro. Dent Mater 17:316-21.
Bouillaguet S, Degrange M, Cattani M, Godin C, Meyer JM (2002).
Bonding to dentin achieved by general practitioners. Schweiz
Monatsschr Zahnmed 112:1006-1011.
Bouschlicher MR, Rueggeberg FA (2000). Effect of ramped light
intensity on polymerization force and conversion in a photoac-
tivated composite. J Esthet Dent 12:328-339.
Brännström M, Aström A (1972). The hydrodynamics of the den-
tine; its possible relationship to dentinal pain. Int Dent J 22:219-
227.
Burke FM, Samarawickrama DY (1995). Progressive changes in the
pulpo-dentinal complex and their clinical consequences.
Gerodontology 12:57-66.
Choi KK, Condon JR, Ferracane JL (2000). The effects of adhesive
thickness on polymerization contraction stress of composite. J
Dent Res 79:812-817.
Ciucchi B, Bouillaguet S, Holz J, Pashley DH (1995). Dentinal fluid
dynamics in human teeth, in vivo. J Endod 21:191-194.
Ciucchi B, Bouillaguet S, Delaloye M, Holz J (1997a). Volume of the
internal gap formed under composite restorations in vitro. J
Dent 25:305-312.
Ciucchi B, Bouillaguet S, Holz J, Roh S (1997b). The battle of the
bonds 1995. Rev Mens Suisse Odontostomatol 107:32-36.
Costa CA, Edwards CA, Hanks CT (2001). Cytotoxic effects of
cleansing solutions recommended for chemical lavage of pulp
exposures. Am J Dent 14:25-30.
Davidson CL, Feilzer AJ (1997). Polymerization shrinkage and
polymerization shrinkage stress in polymer-based restoratives.
J Dent 25:435-440.
Davidson CL, de Gee AJ (2000). Light-curing units, polymerization,
and clinical implications. J Adhes Dent 2:167-173.
Davidson CL, de Gee AJ, Feilzer A (1984). The competition between
the composite-dentin bond strength and the polymerization
contraction stress. J Dent Res 63:1396-1399.
De Gee AF, Feilzer AJ, Davidson CL (1993). True linear polymer-
ization shrinkage of unfilled resins and composites determined
with a linometer. Dent Mater 9:11-14.
De Munck J, Van Meerbeek B, Yoshida Y, Inoue S, Vargas M, Suzuki
K, et al. (2003). Four-year water degradation of total-etch adhe-
sives bonded to dentin. J Dent Res 82:136-140.
de Souza Costa CA, Lopes do Nascimento AB, Teixeira HM (2003).
Response of human pulps following acid conditioning and
application of a bonding agent in deep cavities. Dent Mater
18:543-551.
Eick JD, Gwinnett AJ, Pashley DH, Robinson SJ (1997). Current
concepts on adhesion to dentin. Crit Rev Oral Biol Med 8:306-
335.
Eick JD, Kostoryz EL, Rozzi SM, Jacobs DW, Oxman JD,
Chappelow CC, et al. (2002). In vitro biocompatibility of oxi-
rane/polyol dental composites with promising physical prop-
erties. Dent Mater 18:413-421.
Feilzer AJ, De Gee AJ, Davidson CL (1987). Setting stress in com-
posite resin in relation to configuration of the restoration. J Dent
Res 66:1636-1639.
Feilzer A, De Gee AJ, Davidson CL (1993). Setting stresses in com-
57
 posite for two different curing modes. Dent Mater 9:2-5.
Ferracane JL (1995). Current trends in dental composites. Crit Rev
Oral Biol Med 6:302-318.
Finger WJ, Balkenhol M (1999). Practitioner variability effects on
dentin bonding with an acetone-based one-bottle adhesive. J
Adhesiv Dent 1:311-314.
Garberoglio R, Brännström M (1976). Scanning electron microscop-
ic investigation of human dentinal tubules. Arch Oral Biol
21:355-362.
Goldberg M, Lasfargues JJ, Legrand JM (1994). Clinical testing of
dental materials—histological considerations. J Dent 22(Suppl
2):S25-S28.
Gwinnett AJ, Tay F (1998). Early and intermediate time response of
the dental pulp to an acid etch technique in vivo. Am J Dent
11(Spec No):S35-S44.
Hanks CT, Craig RG, Diehl ML, Pashley DH (1988). Cytotoxicity of
dental composites and other materials in a new in vitro device.
J Oral Pathol 17:396-403.
Hanks CT, Strawn SE, Wataha JC, Craig RG (1991). Cytotoxic
effects of resin components on cultured mammalian fibroblasts.
J Dent Res 70:1450-1455.
Hanks CT, Wataha JC, Parsell RR, Strawn SE, Fat JC (1994).
Permeability of biological and synthetic molecules through
dentine. J Oral Rehabil 21:475-487.
Hanks CT, Wataha JC, Sun Z (1996). In vitro models of biocompat-
ibility: a review. Dent Mater 12:186-193.
Hannig M, Bott B (1999). In-vitro pulp chamber temperature rise
during composite resin polymerization with various light-cur-
ing sources. Dent Mater 15:275-281.
Hashimoto M, Ohno H, Kaga M, Endo K, Sano H, Oguchi H (2000).
In vivo degradation of resin-dentin bonds in humans over 1 to 3
years. J Dent Res 79:1385-1391.
Hebling J, Giro EM, Costa CA (1999). Human pulp response after
an adhesive system application in deep cavities. J Dent 27:557-
564.
Heyeraas KJ, Berggreen E (1999). Interstitial fluid pressure in nor-
mal and inflamed pulp. Crit Rev Oral Biol Med 10:328-336.
Hitmi L, Attal JP, Degrange M (1999). Influence of the time-point of
salivary contamination on dentin shear bond strength of 3
dentin adhesive systems. J Adhes Dent 1:219-232.
Holz J, Baume LJ (1973). Essais biologiques relatifs à la compat-
ibilité des produits d'obturation intermédiaire avec l'organe
pulpo-dentinaire. Rev Mens Suisse Odontostomatol 83:1369-
1458.
Hume WR, Gerzina TM (1996). Bioavailability of components of
resin-based materials which are applied to teeth. Crit Rev Oral
Biol Med 7:172-179.
Imazato S, McCabe JF, Tarumi H, Ehara A, Ebisu S (2001). Degree
of conversion of composites measured by DTA and FTIR. Dent
Mater 17:178-183.
Inoue S, Van Meerbeek B, Vargas M, Yoshida Y. Lambrechts P,
Vanherle G (2000). Adhesion mechanisms of self-etching adhe-
sives. In: Advanced adhesive dentistry. 3rd International
Kuraray symposium. Tagami J, Toledano M, Prati C, editors.
Cirimido, Italy: Kuraray Co. Ltd., pp. 131-148.
Inoue S, Vargas MA, Abe Y, Yoshida Y, Lambrechts P, Vanherle G,
et al. (2001). Microtensile bond strength of eleven contemporary
adhesives to dentin. J Adhes Dent 3:237-245.
ISO 10993-5 (1993). Biological evaluation of medical devices. Part 5:
tests for cytotoxicity: in vitro models. Geneva: International
Organization for Standardization.
Jontell M, Hanks CT, Bratel J, Bergenholtz G (1995). Effects of
unpolymerized resin components on the function of accessory
cells derived from the rat incisor pulp. J Dent Res 74:1162-1167.
Kaga M, Noda M, Ferracane JL, Nakamura W, Oguchi H, Sano H
58 Crit Rev Oral Biol Med
(2001). The in vitro cytotoxicity of eluates from dentin bonding
resins and their effect on tyrosine phosphorylation of L929 cells.
Dent Mater 17:333-339.
Katsuno K, Manabe A, Itoh K, Nakamura Y, Wakumoto S,
Hisamitsu H, et al. (1996). Contact dermatitis caused by 2-
HEMA and GM dentin primer solutions applied to guinea pigs
and humans. Dent Mater J 15:22-30.
Kidd EAM (1996). Microleakage: a review. J Dent 24:199-205.
Kim S, Dörscher-Kim J (1990). Haemodynamic regulation of the
dental pulp. In: Dynamic aspects of the dental pulp. Molecular
biology, pharmacology and pathophysiology. Inoki R, Kudo T,
Olgart L, editors. London: Chapman and Hall Publishing, pp.
167-188.
Krejci I, Stavridakis M (2000). New perspectives on dentin adhe-
sion—the different ways of bonding. Pract Perio Aesthet Dent
12:727-732.
Labella R, Lambrechts P, Van Meerbeek B, Vanherle G (1999).
Polymerization shrinkage and elasticity of flowable composites
and filled adhesives. Dent Mater 15:128-137.
Lai SC, Mak YF, Cheung GS, Osorio R, Toledano M, Carvalho RM,
et al. (2001). Reversal of compromised bonding to oxidized
etched dentin. J Dent Res 80:1919-1924.
Lefebvre CA, Wataha JC, Bouillaguet S, Lockwood PE (1999).
Effects of long-term sublethal concentrations of dental
monomers on THP-1 human monocytes. J Biomater Sci (Polymer
ed) 10:1265-1274.
Lutz F, Krejci I (2001). Quality guidelines of operative dentistry: the
Swiss approach. In: Advances in operative dentistry: challenges
for the future. Vol. 2. Roulet JF, Wilson NHF, Fuzzi M, editors.
Chicago: Quintessence Publishing Co., pp. 305-326.
Lutz F, Krejci I, Barbakow F (1992). Restoration quality in relation
to wedge-mediated light channeling. Quintessence Int 23:763-
767.
Maciel KT, Carvalho RM, Ringle RD, Preston CD, Russell CM,
Pashley DH (1996). The effects of acetone, ethanol, HEMA, and
air on the stiffness of human decalcified dentin matrix. J Dent
Res 75:1851-1858.
Mertz-Fairhurst EJ, Adair SM, Sams DR, Curtis JW Jr, Ergle JW,
Hawkins KI, et al. (1995). Cariostatic and ultraconservative seal-
er restorations: nine-year results among children and adults.
ASDC J Dent Child 62:97-107.
Miyazaki M, Onose H, Moore BK (2002). Analysis of the dentin-
resin interface by use of laser Raman spectroscopy. Dent Mater
18:576-580.
Mjör IA, Ferrari M (2002). Pulp-dentin biology in restorative den-
tistry. Part 6: Reactions to restorative materials, tooth-restora-
tion interfaces, and adhesive techniques. Quintessence Int 33:35-
63.
Mjör IA, Moorhead JE, Dahl JE (2000). Reasons for replacement of
restorations in permanent teeth in general dental practice. Int
Dent J 50:361-366.
Mjör IA, Sveen OB, Heyeraas KJ (2001). Pulp-dentin biology in
restorative dentistry. Part 1: normal structure and physiology.
Quintessence Int 32:427-446.
Murray PE, Smyth TW, About I, Remusat R, Franquin JC, Smith AJ
(2002). The effect of etching on bacterial microleakage of an
adhesive composite restoration. J Dent 30:29-36.
Nakabayashi N, Takarada K (1992). Effect of HEMA on bonding to
dentin. Dent Mater 8:125-130.
Nakabayashi N, Pashley DH (1998). Hybridization of dental hard
tissues. Chapter 1. Tokyo: Quintessence Books, pp. 1-20.
Nakabayashi N, Kojima K, Mashuara E (1982). The promotion of
adhesion by the infiltration of monomers into tooth substrate. J
Biomed Mater Res 16:265-273.
Nakajima M, Okuda M, Pereira PN, Tagami J, Pashley DH (2002).
15(1):47-60 (2004)
 Dimensional changes and ultimate tensile strengths of wet
decalcified dentin applied with one-bottle adhesives. Dent
Mater 18:603-608.
Närhi MVO (1990). Intradental sensory units. In: Dynamic aspects
of the dental pulp. Molecular biology, pharmacology and
pathophysiology. Inoki R, Kudo T, Olgart L, editors. London:
Chapman and Hall Publishing, pp. 137-149.
Noda M, Wataha JC, Lockwood PE, Volkmann KR, Kaga M, Sano
H (2003). Sublethal, 2-week exposures of dental material com-
ponents alter TNF-alpha secretion of THP-1 monocytes. Dent
Mater 19:101-105.
Nomura Y, Teshima W, Tanaka N, Yoshida Y, Nahara Y, Okazaki M
(2002). Thermal analysis of dental resins cured with blue light-
emitting diodes (LEDs). J Biomed Mater Res (Appl Biomater)
63:209-213.
Osorio R, Ceballos L, Tay F, Cabrerizo-Vilchez MA, Toledano M
(2002). Effect of sodium hypochlorite on dentin bonding with a
polyalkenoic acid-containing adhesive system. J Biomed Mater
Res 60:316-324.
Outhwaite WC, Livingston MJ, Pashley DH (1976). Effects of
changes in surface area, thickness, temperature and post-
extraction time on human dentine permeability. Arch Oral Biol
21:599-603.
Paphangkorakit J, Osborn JW (2000). The effect of normal occlusal
forces on fluid movement through human dentine in vitro. Arch
Oral Biol 45:1033-1041.
Pashley DH (1985). The dentin-predentin complex and its perme-
ability: physiologic overview.. J Dent Res 64(Spec Iss):613-620.
Pashley DH (1989). Dentin: a dynamic substrate—a review.
Scanning Microsc 3:161-174.
Pashley DH (1990). Dentin permeability: theory and practice. In:
Experimental endodontics. Spangberg L, editor. Boca Raton,
FL: CRC Press, pp. 19-49.
Pashley DH (1996). Dynamics of the pulpo-dentin complex. Crit
Rev Oral Biol Med 7:104-133.
Pashley DH, Matthews WG (1993). The effects of outward forced
convective flow on inward diffusion in human dentine in vitro.
Arch Oral Biol 38:577-582.
Pashley DH, Ciucchi B, Sano H, Horner JA (1993). Permeability of
dentin to adhesive agents. Quintessence Int 24:618-631.
Pashley DH, Matthews WG, Zhang Y, Johnson M (1996). Fluid
shifts across human dentine in vitro in response to hydrody-
namic stimuli. Arch Oral Biol 41:1065-1072.
Pashley DH, Carvalho RM, Sano H, Nakajima M, Yoshiyama M,
Shono Y, et al. (1999). The microtensile bond test: a review. J
Adhes Dent 1:299-309.
Pashley DH, Zhang Y, Agee KA, Rouse CJ, Carvalho RM, Russell
CM (2000). Permeability of demineralized dentin to HEMA.
Dent Mater 16:7-14.
Paul SJ, Schärer P (1997). The dual bonding technique: a modified
method to improve adhesive luting procedures. Int J Periodont
Rest Dent 17:536-545.
Perdigão J, Frankenberger R (2001). Effect of solvent and rewetting
time on dentin adhesion. Quintessence Int 32:385-390.
Rakich DR, Wataha JC, Lefebvre CA, Weller RN (1999). Effect of
dentin bonding agents on the secretion of inflammatory medi-
ators from macrophages. J Endod 25:114-117.
Ratanasathien S, Wataha JC, Hanks CT, Dennison JB (1995).
Cytotoxic interactive effects of dentin bonding components on
mouse fibroblasts. J Dent Res 74:1602-1606.
Roulet JF, Degrange M (2001). Adhesion—the silent revolution.
Chicago: Quintessence Publishing Co., p. 11.
Rueggeberg FA, Margeson DH (1990). The effect of oxygen inhibi-
tion on an unfilled/filled composite system. J Dent Res 69:1652-
1658.
15(1):47-60 (2004) Crit Rev Oral Biol Med
Ryge G (1981). Clinical criteria. Int Dent J 30:347-358.
Sahafi A, Peutzfeldt A, Asmussen E (2001). Effect of pulse-delay
curing on in vitro wall-to-wall contraction of composite in
dentin cavity preparations. Am J Dent 14:295-296.
Sano H, Shono T, Sonoda H, Takatsu T, Ciucchi B, Carvalho R, et al.
(1994a). Relationship between surface area for adhesion and
tensile bond strength—evaluation of a micro-tensile bond test.
Dent Mater 10:236-340.
Sano H, Shono T, Takatsu T, Hosoda H (1994b). Microporous
dentin zone beneath resin-impregnated layer. Oper Dent 19:59-
64.
Sano H, Takatsu T, Ciucchi B, Horner JA, Matthews GW, Pashley
DH (1995). Nanoleakage: leakage within the hybrid layer. Oper
Dent 20:18-25.
Santerre JP, Shajii L, Leung BW (2001). Relation of dental compos-
ite formulations to their degradation and the release of
hydrolyzed polymeric-resin-derived products. Crit Rev Oral
Biol Med 12:136-151.
Schafer TE, Lapp CA, Hanes CM, Lewis JB, Wataha JC, Schuster GS
(1999). Estrogenicity of bisphenol A and bisphenol A
dimethacrylate in vitro. J Biomed Mater Res 45:192-197.
Schweikl H, Schmalz G (1999). Triethylene glycol dimethacrylate
induces large deletions in the hprt gene of V79 cells. Mutat Res
438:71-78.
Silverstone LM (1975). The acid etch technique: in vitro studies
with special reference to the enamel surface and enamel-resin
interface. In: Proceedings of an international symposium on the
acid etch technique. Silverstone LM, Dogon IL, editors. St Paul,
MN: North Central Publ. Co., pp. 13-39.
Staehle HJ (1999). Minimally invasive restorative treatment. J Adhes
Dent 1:267-284.
Tay FR, Pashley DH (2001). Aggressiveness of contemporary self-
etching systems. I: Depth of penetration beyond dentin smear
layers. Dent Mater 17:296-308.
Tay FR, Gwinnett AJ, Wei SH (1996). The overwet phenomenon: an
optical, micromorphological study of surface moisture in the
acid-conditioned, resin-dentin interface. Am J Dent 9:43-48.
Tay FR, Carvalho R, Sano H, Pashley DH (2000). Effect of smear
layers on the bonding of a self-etching primer to dentin. J Adhes
Dent 2:99-116.
Tay FR, King NM, Suh BI, Carvalho RM, Itthagarun A (2002).
Single-step adhesives are permeable membranes. J Dent 30:371-
382.
Tay FR, Pashley DH, Yiu C, Cheung C, Hashimoto M, Itou K, et al.
(2003). Types of nanoleakage and potential implications: evi-
dence from unfilled and filled adhesives with the same resin
composition. Am J Dent (in press).
Tonetti-Eberle B, Pauli-Uhlmann A, Mombelli A (2001). Quality of
water of dental units. A survey in the region of Berne,
Switzerland. Schweiz Monatsschr Zahnmed 111:1160-1164.
Turner DF, Marfurt CF, Sattelberg C (1989). Demonstration of
physiological barrier between pulpal odontoblasts and its
perturbation following procedures: a horseradish peroxidase
tracing study in the rat routine restorative. J Dent Res. 68:1262-
1268.
Unemori M, Matsuya Y, Akashi A, Goto Y, Akamine A (2001).
Composite resin restoration and postoperative sensitivity:
clinical follow-up in an undergraduate program. J Dent
29:7-13.
Van Meerbeek B, Perdigão G, Lambrechts P, Vanherle G (1998a).
The clinical performance of adhesives. J Dent 26:1-20.
Van Meerbeek B, Yoshida Y, Lambrechts P, Vanherle G, Duke ES,
Eick JD, et al. (1998b). A TEM study of two water-based
adhesive systems bonded to dry and wet dentin. J Dent Res
77:50-59.
59
Van Meerbeek B, Vargas M, Inoue S, Yoshida Y, Peumans M,
Lambrechts P, et al. (2001). Adhesives and cements to promote
preservation dentistry. Oper Dent 26(Suppl 6):119-144.
Versluis A (2000). Evaluation of polymerization vectors and FEA
models. In: Advanced adhesive dentistry. 3rd International
Kuraray symposium. Tagami J, Toledano M, Prati C, editors.
Cirimido, Italy: Kuraray Co. Ltd., pp. 31-37.
Williams DF, editor (1990). Biocompatibility: an overview. In:
Concise encyclopaedia of medical and dental materials. Oxford:
Pergamon, pp. 51-59.
Yoshikawa T, Sano H, Burrow MF, Tagami J, Pashley DH (1999).
60 Crit Rev Oral Biol Med
Effects of dentin depth and cavity configuration on bond
strength. J Dent Res 78:898-905.
Yoshikawa T, Burrow MF, Tagami J (2001). A light curing method
for improving marginal sealing and cavity wall adaptation of
resin composite restorations. Dent Mater 17:359-366.
Yoshiyama M, Tay FR, Doi J, Nishitani Y, Yamada T, Itou K, et al.
(2002). Bonding of self-etch and total-etch adhesives to carious
dentin. J Dent Res 81:556-560.
Zheng L, Pereira PN, Nakajima M, Sano H, Tagami J (2001).
Relationship between adhesive thickness and microtensile
bond strength. Oper Dent 26:97-104.
15(1):47-60 (2004)