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 information 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).
(1) Introduction
Over the past 30 years, restorative dentistry has seen a revolution in techniques, materials available, prevalence of disease, and patient priorities. Today’s dentist is able to prevent 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 changing 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 composite materials integrate with the dentin-pulp complex. It is the purpose of this review to present recent and relevant information 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).
(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 restorations worldwide, is testimony to their apparent biocompatibility. 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 materials may have systemic estrogenic effects (Schafer et al., 1999), may elicit allergic reactions (Katsuno et al., 1996), or may possibly even act as carcinogens (Schweikl and Schmalz, 1999). Therefore, it is imperative that we reach a more precise definition 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 crystals), 30 vol% organic components (mostly type I collagen), and 20 vol% fluid (Mjör et al., 2001). The collagen fibrils are arranged in a network, forming a matrix for the hydroxyapatite crystals. Spaces of 20–50 nm separate fibrils about 20–100 nm in diameter (Eick et al., 1997). This network is uniformly mineralized and forms the bulk of the dentin (intertubular dentin). In dentin immediately adjacent to the tubules (peritubular dentin), the mineralized component predominates, and the collagen network is sparse. The dentinal tubules run from the dentin-enamel junction and converge on one another toward the pulp of the tooth (Outhwaite et al., 1976). In cross-section, the density of the tubules is about 15,000/mm2 near the DEJ to over 65,000/mm2 near the pulp (Garberoglio and Brännström, 1976). Furthermore, the tubular diameter increases from 0.5 μm near the DEJ to over 2.5 μm near the pulp. The convergence of the tubules and their increased diameter toward the pulp are responsible for an increase in dentin permeability near the pulp (Fig. 1). The permeability of the dentin allows for both outward pulpal fluid flow and inward diffusion of chemical and bacterial products. Pashley (1990) has used the Hagen-Poiseuille equation to show that fluid filtration varies with the fourth power of the radius of the dentin tubule, and that the driving force is the fluid pressure gradient. The inward diffusion of bacterial or products of material degradation may cause deleterious reactions in the dental pulp. However, the dentin acts as a diluter of diffusing substances. According to the Fick equation, the rate of diffusion is dependent on the applied concentration but is inversely proportional to the dentin thickness. The surface area available for diffusion, the temperature, and the chemical characteristics of the diffusing molecules all affect diffusion (Pashley, 1985). Dentinal diffusion of bacterial or material products is a critical factor in the assessment of material biological or microbiological risks. It has been clearly established from in vitro and in vivo studies that outward flow of dentinal fluid cannot completely compensate for the inward diffusion of chemicals or bacteria (Fig. 2), and that the thickness of dentin remaining between prepared surfaces and the pulp is the critical variable in determining whether dentin can protect the pulp (Holz and Baume, 1973; Pashley and Matthews, 1993).
The dental pulp consists of a loose connective tissue that occupies the central part of the tooth. At the periphery, the odontoblasts line the dentin, and their processes extend into the dentinal tubules for at least several hundred microns (Fig. 3). Odontoblasts are connected to each other by gap junctions, desmosomes, and tight junctions and are highly specialized in the synthesis and the secretion of organic molecules and the mineralization of the dentin. Turner et al.(1989) have shown that a functional barrier exists between the odontoblasts that prevents the passage of macromolecules from the pulp into the predentin and dentin. They also demonstrated that this functional barrier 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 hydrodynamic theory, dentinal pain is induced by rapid fluid shifts across dentinal tubules (Brännström and Aström, 1972). These shifts are caused by temperature, pressure, or mechanical perturbation and result in a mechanical deformation of the odontoblasts and nearby nerves. The mechanical deformation of A-delta nerves is responsible 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 pressure in the pulp (called pulpal pressure) is the result of vascular pressure, and recent research indicates that normal pulpal 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 pressure 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 normal 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 secondary 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 production 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 important 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 patients, in whom pulpal insults are longstanding and reparative 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 triethylene glycol dimethacrylate (TEGDMA), among others. Monomers may be slightly soluble in water, but are commonly 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 composites 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 sufficiently thin to receive the full power density of the curing light. 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 methacrylate monomer double-bonds actually remains unreacted in the polymer (Asmussen, 1982; Imazato et al., 2001). Any unpolymerized monomer in the composite is a potential biological liability 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 contemporary 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 composite, 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 products into the restoration.
The risk of biological harm from degraded or unpolymerized 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 exposure 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 identification. 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 components 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 estimated dose exposure: If the dose to cause an adverse response is greater than the estimated exposure by a comfortable safety margin, the likelihood of an adverse event occurring in an exposed population is small, and the material may be deemed to have a low risk of biological problems.
Although a few in vivo studies have attempted to document the biological risks of resin-based materials, most information 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 concentrations of 11 components of dental resins on mouse fibroblasts. 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 mitochondrial 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 mediators 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 significant 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 mitochondrial activity and total protein content per cell, even at concentrations of 1–10% of those used in short-term experiments. These results confirmed that risk assessment of dentin adhesives 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 activated 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 polymerization 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 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 complex procedures have been required for bonding to dentin because of the completely different nature of the dentin substrate. Bonding to dentin is further complicated by the formation 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 dissolved 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 systems (Fig. 6).
Some manufacturers have developed “one-bottle” adhesives that contain mixtures of organic solvents and resins (HEMA, BisGMA, TEGDMA, UDMA) to impregnate the collagen-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 with the HEMA-water primers (Fig. 7). Therefore, these mixtures 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 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 systems, they are simpler to use and may provide more consistent 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 adhesives 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, priming, and bonding into a single step. These ‘all-in-one’ adhesives are in their infancy and will likely undergo significant evolution in coming years.
There are different methods for measuring 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 polymerization. Their results clearly showed that resin-based materials 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 directly 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 material 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 materials, 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 adhesives 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% compared with flat dentin, where no polymerization stress is present (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 promote 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 quality of resin-dentin bonds. However, there is increasing evidence 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 complex 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 clinical 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 bacteria, fluids, molecules, or ions between a cavity wall and the restorative material. Microleakage gaps are many micrometers 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 degradation. However, the most likely cause of microleakage is from the volumetric shrinkage that occurs concurrently with polymerization 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-factor 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 influenced by the kinetics of the polymerization reaction. A resin-based material will flow plastically to accommodate shrinkage 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 curing 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 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 collapse and will not be effectively infiltrated by the primer or adhesive resin. To a lesser degree, primers that use organic solvents 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 incomplete 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 compromised. 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 infiltrate 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 nanoleakage 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 applications 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 protection 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 fundamental 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 significantly during cavity preparation. Treatment of caries lesions involves removal of infected tissues but requires the preservation 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 biological risks of each adhesive system under relevant clinical conditions. Shallow cavities located in superficial or sclerotic dentin do not pose a major biological risk, because the permeability of the dentin is low and the thickness of the remaining dentin is adequate to prevent any adverse effects from diffusing 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-etching 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 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 generate 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 extremely 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, clinically, pulpal cells that have chronically suffered from exposure to toxic components of resins will respond differently to bacterial 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 overstressed pulp leads inevitably to pain and significant restorative preparation time and costs.
(4.2) Clinical perspectives on resin-tooth bonding (minimizing microbiological risks)
Clinically, the biggest problem with bonding resin-based materials 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 restorations 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 balanced 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 bacteria 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., 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 proteolytic properties, hydrogen peroxide is oxidative, and both diffuse 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 routine 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 preparation 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 composite restorations on top of the caries lesions without the removal of the caries lesion. The results of this study seriously 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 bonding, 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 sclerotic and caries-affected dentin (Inoue et al., 2001). Therefore, total-etching adhesives are the material of choice for most clinical 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 confusion. 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 systems 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-rinsing step, self-etching adhesives are less sensitive to moisture 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 during composite shrinkage (Choi et al., 2000). Despite differences among materials, research supports the concept of thick adhesive 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 adhesives. Poor management of the shrinkage stresses that develop during the curing of the restorative material can cause the failure 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 inserted into the cavity.
Early in the development of resin-based materials, the concept of incremental addition of material to the cavity, combined with the use of the so-called “directed-cure”, was proposed 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 (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 polymerization 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 curing, 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 polymerization stresses (Bouschlicher and Rueggeberg, 2000). In an effort to cure larger increments of composite faster, investigators 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 during curing are probably higher, because the composite reaches 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-curing 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 practitioners in combination with more superficial layers of light-cured resins.
A variety of restorative techniques has been used clinically to control polymerization shrinkage stresses. These techniques 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 possible 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 polymerization 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 contraction 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 successful 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 protect 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 laboratory-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 cements for indirect restorations.
In direct restorations, several techniques have been used to limit polymerization stresses, thereby reducing microbiological 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 filling material is incrementally polymerized onto it. The adhesive 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 clinicians 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 configuration 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 bonding 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 adhesive materials are complex and interactive and not fully understood. 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 evaluation 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 materials 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 developed. Adhesives and adhesive strategies will likely be adapted or even customized to deal with a variety of bonding substrates, 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 susceptible 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 mixture of oxiranes and polyol that can polymerize by light activation (Eick et al., 2002). With these new chemical structures, 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 materials has provided the clinician with many techniques and materials with which to restore tooth structure, esthetics, and function. 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 biological 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 methods, clinical techniques, and materials that may help us better estimate these risks and improve the reliability of resin-based restorations.
Schematic of convergence of tubules toward the pulp. (A) Periphery of the dentin. Most surface area is occupied by intertubular dentin (☆), with a few tubules surrounded by hypermineralized peritubular dentin (
In vivo diffusion through dentin of a solution of silver nitrate. The dentin-pulp complex. The dentin and pulp exist together as an integrated unit. The functional barrier that develops between the odontoblasts prevents the passage of macromolecules from the pulp into the predentin and dentin (HE staining × 40). Tertiary dentin. Tertiary dentin formation (arrows) is regarded as an important defense mechanism of the pulp-dentin complex in response to either pathological or physiological insults. The presence of tertiary dentin reduces dentin permeability. 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. × 2400). Modified SEM illustration of bonding to dentin with conventional (three-step) total-etching adhesives. Modified SEM image of bonding to dentin with one-bottle (two-step) total-etching adhesives. Modified SEM image of bonding to dentin with (two-step) self-etching adhesives. 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. × 10,000).
). (B) Near the pulp, the increase in tubule diameter has occurred largely at the expense of the peritubular dentin. This substrate has a high protein content. As the remaining dentin is made thinner (from A to B), the permeability of the dentin increases, because both diameter and density of dentinal tubules are increased. Reprinted with permission from Elsevier.
) inside the cavity (HE staining × 40). 
) into the pulp area and into the capillary system active at clearing the pulp (HE staining × 40).
