Biocompatibility of Dental Materials

Enamel

Mature human enamel is highly mineralized (96% by weight), with 1% of its weight being composed of organic molecules and 3% being water. The organic matrix of enamel consists of at least two types of glycoproteins: amelogenins and enamelins. After synthesis by ameloblasts, the calcified organic matrix of enamel does not appear to be maintained in any way by cellular synthetic mechanisms, contrary to other calcified tissues such as dentin, bone, and cementum. Enamel rods have a specific orientation with each other, and this orientation provides maximal strength. Because of its high mineral (hydroxylapatite) content, enamel is much more brittle than dentin and is solubilized to a greater extent by acid solutions. This property is used to advantage with bonding agents, when acids are used to etch the enamel to provide micromechanical retention of resin composites. The differential etching that occurs is a consequence of the different orientation of enamel rods on the enamel surface. Permeability of enamel to most oral molecules is quite low, and in this sense enamel “seals” the tooth to outside agents. However, recent evidence indicates that enamel is not totally impermeable. Peroxides in bleaching agents have been shown to penetrate intact enamel in just a few seconds.

Dentin and Pulp

Because of their intimate anatomical relationship, most researchers consider dentin and pulp to be a single tissue. The dentinal matrix (both calcified and uncalcified) forms the greatest bulk of the tooth. Calcified dentin is about 20% organic, 70% inorganic, and 10% aqueous by weight. Collagen constitutes approximately 85% of the organic portion of dentin, and hydroxyapatite is the main inorganic compound. The dentinal matrix also contains many proteins, including collagenous proteins (mainly type I collagen with smaller amounts of type V and type I trimer collagens), noncollagenous dentinspecific proteins (phosphophoryns, dentin sialoprotein, and dentin matrix protein-1), and several nonspecific proteins associated with mineralized tissues (e.g., osteocalcin and osteopontin).

The dentinal matrix surrounds dentinal tubules that are filled with odontoblastic processes. These processes stem from cells that reside in the pulp of the tooth. The tubules traverse the region between the dentoenamel junction (DEJ) and the pulp. The numbers of tubules per cross-sectional area range from about 20,000/mm² near the DEJ to 50,000/mm² near the pulp. Tubule diameter varies from about 0.5 μm at the DEJ to about 2.5 μm near the pulp (Fig. 5-2). Some odontoblastic processes extend through the dentin tubules to the DEJ, but the percentage of processes that reach the DEJ is a matter of some conjecture.

A serumlike fluid fills the dentinal tubules. This fluid has continuity with the extracellular fluid of the pulp tissue. Pulpal circulation maintains an intercellular hydraulic pressure of about 24 mm Hg (32.5 cm H₂O), which causes fluid flow in the tubules to be directed from the pulp outward toward the DEJ when enamel is removed. External hydrostatic and osmotic pressures can also cause fluid movement toward or away from the pulp. The positive or negative displacement of this fluid through exposed dentinal tubules is capable of affecting either odontoblasts or pulpal nerve endings. These effects are the basis of the hydrodynamic theory of hyperalgesia (pulpal hypersensitivity).

During cavity preparation by a dentist, a “smear layer” is formed by the action of the bur or hand instruments on the calcified dentin matrix (Fig. 5-3). This mat of organic and inorganic particles occludes the dentinal tubules to some extent. The smear layer is quite effective in reducing hydrostatic pressures, but less effective in reducing diffusion, especially if the layer is interrupted or defective. The smear layer can be removed by acid etching, which also demineralizes the openings of the tubules (Fig. 5-4). The dentinal tubules establish continuity with the pulpal fluid to facilitate the diffusion of molecules, both natural or from materials, into and out of the pulp. The smear layer, dentinal tubules, and dentinal matrix are all important in the application of dentinal bonding agents and the ability of components of the bonding agents to reach and affect pulpal tissues.

Moderately deep cavity preparation will damage odontoblasts by severing the odontoblastic processes. Deep cavity preparation may destroy most of the dentin and kill the primary odontoblasts. A number of investigators think that the extracellular matrix (ECM) of dentin and pulp is largely responsible for the differentiation of secondary odontoblasts that form reparative dentin. The source of secondary odontoblasts is not known, but much of the proliferative activity of granulation tissue following pulpal insult is found in perivascular areas proximal to the core of the pulp. In monkeys, the minimal amount of time between pulp injury and replacement odontoblast differentiation is about 5 days. Nerves and blood vessels, which arborize from the core of the pulp as they approach the odontoblastic layer, may influence the extent of the inflammatory response and the amount of new dentinal matrix formed during dentin repair.

In the absence of a smear layer, ions released from materials and/or bacterial products diffuse toward the pulp against the pressure gradient (diffusional permeability, see later discussion). Bacteria can sometimes be seen within tubules below a carious lesion or at the base of a prepared cavity, with or without a restoration (Fig. 5-5). When toxic bacterial or chemical products traverse the dentin, odontoblasts and the pulpal connective tissue usually respond first by focal necrosis (0 to 12 hours), which may be followed by an acute, but more widespread, pulpitis (12 hours to several days). This response may resolve naturally if the injurious agent is removed or the tubules are blocked. If the pulpitis does not resolve, it may spread to more completely involve the pulp in liquefaction necrosis (especially if the pulpitis results from bacterial products) or in chronic inflammation. Finally, both acute complete pulpitis and acute exacerbation of chronic pulpitis may lead to sequelae such as dental periapical lesions and osteomyelitis. These conditions may be reviewed in oral pathology textbooks.

Dentin Permeability

Much has been learned about dentin permeability in the last three decades. In practical terms, two types of dentin permeability occur. The first is fluid convection; that is, movement of fluid through the dentinal tubules. Fluid convection toward the pulp will occur under positive hydraulic pressure when a crown or inlay is being seated. If the dentinal tubules are open, this produces a sharp, localized pain in the pulp from stimulation of A-fibers. Fluid convection away from the pulp will occur with negative osmotic pressures when concentrated solutions, such as sucrose or saturated calcium chloride, are exposed to open dentinal tubules. Clinically, this situation occurs with cervical abrasion or carious lesions. Convection of fluids across dentin varies with the fourth power of the radius of the dentinal tubule (r ⁴), and thus is very sensitive to the diameter of the tubules. In general, coronal dentin exhibits greater convective permeability than root dentin. Axial wall dentin is more permeable than dentin in the floor of cavities, and dentin near pulp horns (where tubule diameter is greatest) is more permeable than dentin at a distance from the pulp horns. The presence of a smear layer or of cavity liners, sealers, resin bonding agents, crystals such as calcium oxalate, and even debris and bacteria in the dentinal tubules can dramatically reduce fluid convection.

The second type of dentin permeability is diffusion. Patent dentinal tubules, no matter how small the diameter, establish a diffusion gradient through which ions and molecules can move, even against positive hydraulic pressure. Diffusion is proportional to the length of the dentinal tubules and thus, roughly, to the thickness of the dentin between cavity preparation and the pulp. Smear layers created in cavity preparation are better than cavity liners and sealers at limiting diffusional permeability. However, if the smear layer is incomplete, interrupted, or removed, or if there is a disruption in a cavity liner, sealer, or base, then diffusion of molecules toward the pulp will occur.

Diffusion of natural and synthetic molecules through dentin has been studied. In general, diffusion through a given thickness of dentin is proportional to the molecular size of the molecule. Consequently, molecules the size of urea (molecular weight [mw] 60), phenol (mw 94), and glucose (mw 180) diffuse more easily than molecules the size of dextran (mw 20,000) and albumin (bovine serum albumin, mw 68,000). Through diffusion, small or globular molecules such as albumin, gamma globulin, and bis-glycidyldimethacrylate (Bis-GMA, an oligomer of resin composites) are diluted 2000 to 10,000 times on the pulpal side of the dentin by 0.3 to 0.4 mm of dentin. Large, fibrous molecules such as fibrinogen are diluted 25,000 to 125,000 times by the same dentin thickness. Most molecules are probably adsorbed to some extent by dentin. Some molecules and atoms or ions, such as tetracycline, zinc, H₂O₂, and fluorescein, are adsorbed to a greater extent than the biological molecules and resin monomers mentioned earlier. Finally, the capillary beds and vascular dynamics in most healthy pulp are probably capable of removing relatively large amounts of cytotoxic chemicals and bacterial products once they diffuse through the dentin. However, if the pulp is already damaged (inflamed because of caries or trauma), edema and sluggish circulation probably compromise the removal of these materials. Much still needs to be learned about the dynamics and significance of diffusion and adsorption of both host and foreign molecules through dentin.

Cytotoxicity Tests

Cytotoxicity tests assess cell death caused by a material by measuring cell number or growth before and after exposure to that material. Cells are plated in a well of a cell-culture dish, where they attach. The material is then placed in the test system. If the material is not cytotoxic, cells will remain attached to the well and will proliferate over time. If the material is cytotoxic, cells may stop growing, exhibit cytopathic features (Figs. 5-6 and 5-7), or detach from the well. If the material is a solid, then the density (number of cells per unit area) of cells may be assessed at different distances from the material, and a “zone” of inhibited cell growth may be described (Fig. 5-8). Cell density can be assessed qualitatively, semiquantitatively, or quantitatively. Substances such as Teflon or cell-culture treated polystyrene can be used as negative (noncytotoxic) control materials, whereas materials such as plasticized polyvinyl chloride can be used as positive (cytotoxic) control materials. Control materials should be well defined and commercially available to facilitate comparisons among other testing laboratories.

Another group of tests is used to measure cytotoxicity by a change in membrane permeability (Fig. 5-9). Membrane permeability is the ease with which a dye can pass through a cell membrane. This test is used on the basis that a loss in membrane permeability is equivalent to or very nearly equivalent to cell death. The advantage of the membrane permeability test is that it identifies cells that are alive (or dead) under the microscope. This feature is important because it is possible for cells to be physically present, but dead (when materials fix the cells). There are two basic types of dyes used. Vital dyes are actively transported into viable cells, where they are retained unless cytotoxic effects increase the permeability of the membrane. It is important to establish that the dye itself does not exhibit cytotoxicity during the time frame of the test. Nonvital dyes are not actively transported, and are only taken up if membrane permeability has been compromised by cytotoxicity. Many types of vital dyes have been used, including neutral red and Na₂⁵¹CrO₄. The use of neutral red and Na₂[⁵¹Cr]O₄ are particularly advantageous because they are neither synthesized nor metabolized by the cell. Examples of nonvital dyes include trypan blue and propidium iodide.

Tests for Cell Metabolism or Cell Function

Some in vitro tests for biocompatibility use the biosynthetic or enzymatic activity of cells to assess cytotoxic response. Tests that measure deoxyribonucleic acid (DNA) synthesis or protein synthesis are common examples of this type of test. The synthesis of DNA or protein by cells is usually analyzed by adding radioisotope-labeled precursors to the medium and quantifying the radioisotope (e.g., ³H-thymidine or ³H-leucine) incorporated into DNA or protein. A commonly used enzymatic test for cytotoxicity is the MTT test. This test measures the activity of cellular dehydrogenases, which convert a chemical called MTT, via several cellular reducing agents, to a blue, insoluble formazan compound (Fig. 5-10). If the dehydrogenases are not active because of cytotoxic effects, the formazan will not form. Production of formazan is quantified by dissolving it and measuring the optical density of the resulting solution. Alternatively, the formazan can be localized around the test sample by light or electron microscopy. Other formazan-generating chemicals have been used, including NBT, XTT, and WST.

Recently developed indicator dyes such as alamarBlue have been formulated to quantitatively measure cell proliferation using an oxidationreduction (redox) indicator that yields a colorimetric change and a fluorescent signal in a response to a metabolic activity. These new generation dyes permit continuous monitoring of cells over time. Additional in vitro tests to measure gene activation, gene expression, cellular oxidative stress, and other specific cell functions have been proposed. However, these tests are not yet routinely used to assess the biocompatibility of materials.

Tests That Use Barriers (Indirect Tests)

Most of the cytotoxicity tests presented thus far are performed with the material in direct contact with the cell culture. Researchers have long recognized that in vivo, direct contact often does not exist between cells and the materials. Separation of cells and materials may occur from keratinized epithelium, dentin, or extracellular matrix. Thus, several in vitro barrier tests have been developed to mimic in vivo conditions. One such test is the agar overlay method (Fig. 5-11) in which a monolayer of cultured cells is established before adding 1% agar or agarose (low melting temperature) plus a vital stain, such as neutral red, to fresh culture media. The agar forms a barrier between the cells and the material, which is placed on top of the agar. Nutrients, gas, and soluble toxic substances can diffuse through the agar. Solid test samples or liquid samples adsorbed onto filter paper can be tested with this assay for up to 24 hours. This assay correlates positively with the direct-contact assays described previously and the intramuscular implantation test in rabbits. However, the agar may not adequately represent barriers that occur in vivo. Furthermore, because of variability of the agar’s diffusion properties, it is difficult to correlate the intensity of color or width of the zone around a material with the concentration of leachable toxic products.

A second barrier assay is the Millipore filter assay. This technique establishes a monolayer of cells on filters made of cellulose esters. The culture medium is then replaced with medium containing about 1% agar, and this mixture is allowed to gel over the cells. Finally, the filter-monolayer-gel is detached and turned over so that the filter is on top for placement of solid or soluble test samples for 2 or more hours. After exposure to the test samples, the filter is removed and an assay is used to determine the effect of the sample on a cellular metabolic activity. The succinyl dehydrogenase assay described previously can be used with this test. Like the agar overlay test and the cell contact tests, toxicity in the Millipore filter test is assessed by the width of the cytotoxic zone around each test sample. This test also has the drawback of arbitrarily influencing the diffusion of leachable products from the test material. The agar diffusion and Millipore filter tests can provide, at best, a qualitative cytotoxic ranking among materials.

Dentin barrier tests have shown improved correlation with the cytotoxicity of dental materials in usage tests in teeth, and are gradually being developed for screening purposes (Fig. 5-12). A number of studies have shown that dentin forms a barrier through which toxic materials must diffuse to reach pulp tissue. Thus pulpal reaction to zinc oxide–eugenol is relatively mild as compared with the more severe reactions to the same material in direct contact with cells in in vitro assays and tissue in implantation tests. The thickness of the dentin correlates directly with the protection offered to the pulp. Thus, assays have been developed that incorporate dentin disks between the test sample and the cell assay system. The use of dentin disks offers the added advantage of directional diffusion between the restorative material and the culture medium.

Other Assays for Cell Function

In vitro assays to measure immune function or other tissue reactions have also been used. The in vivo significance of these assays is yet to be ascertained, but many show promise for being able to reduce the number of animal tests required to assess the biocompatibility of a material. These assays measure cytokine production by lymphocytes and macrophages, lymphocyte proliferation, chemotaxis, or T-cell rosetting to sheep red blood cells. Other tests measure the ability of a material to alter the cell cycle or activate complement. The activation of complement is of particular concern to researchers working on artificial or “engineered” blood vessels and other tissues in direct contact with blood. Materials that activate complement may generate inflammation or thrombi, and may propagate a chronic inflammatory response. Whereas concerns about these types of response to dental materials are fewer, it is possible that activation of complement by resins or metals or their corrosion products may prolong inflammation in the gingiva or pulp.

Mutagenesis Assays

Mutagenesis assays assess the effect of a biomaterial on a cell’s genetic material. There are a wide range of mechanisms by which a material can affect a cell’s genes. Genotoxic mutagens directly alter cell DNA through various types of mutations. Each chemical may be associated with a specific type of DNA mutation. Genotoxic chemicals may be mutagens in their native states, or may require activation or biotransformation to be mutagens, in which case they are called promutagens. Epigenetic mutagens do not alter the DNA themselves, but support tumor growth by altering the cell’s biochemistry, altering the immune system, acting as hormones, or other mechanisms. Carcinogenesis is the ability to cause cancer in vivo. Mutagens may or may not be carcinogens, and carcinogens may or may not be mutagens. Thus, quantitation and relevance of tests that measure mutagenesis and carcinogenesis are extremely complex. A number of government-sponsored programs evaluate the ability of in vitro mutagenesis assays to predict carcinogenicity.

The Ames test is the most widely used short-term mutagenesis test and the only short-term test that is considered thoroughly validated. It uses mutant stocks of Salmonella typhimurium that require exogenous histidine. Native stocks of bacteria do not require exogenous histidine. Exclusion of histidine from the culture medium allows a chemical to be tested for its ability to convert the mutant strain to a native strain. Chemicals that significantly increase the frequency of reversion back to the native state have a reportedly high probability of being carcinogenic in mammals because they significantly alter genetic material. Performance of this test requires experience in the field and special strains of Salmonella to produce meaningful results. Several strains of Salmonella are used, each to detect a different type of mutation transformation. Furthermore, chemicals can be “metabolized” in vitro using homogenates of liver enzymes to simulate the body’s action on chemicals before testing for mutagenicity.

A second test for mutagenesis is the Styles’ cell transformation test. This test on mammalian cells was developed to offer an alternative to bacterial tests (Ames test), which may not be relevant to mammalian systems. This assay quantifies the ability of potential carcinogens to transform standardized cell lines so they will grow in soft agar. Untransformed fibroblasts normally will not grow within an agar gel, whereas genetically transformed cells will grow below the gel surface. This characteristic of transformed fibroblasts is the only characteristic that correlates with the ability of cells to produce tumors in vivo. At least four different continuous cell lines (Chang, BHK, HeLa, WI-38) have been used. In 1978, Styles claimed 94% “accuracy in determining carcinogenic or noncarcinogenic activity” when testing 120 compounds in two cell lines. However, there has been some difficulty in reproducing these results.

In a recent report, four short-term tests (STTs) for gene toxicity were compared (Table 5-2). The Ames test was the most specific (86% of noncarcinogens yielding a negative result). The Ames test also had the highest positive predictability (83% of positives were actually carcinogens) and displayed negative predictability equal to that of other STTs (i.e., 51% of all Ames test negatives were noncarcinogenic). However, the results were in agreement (concordance) with rodent carcinogenicity tests for only 62% of the chemicals. Also, the Ames test was sensitive to only 45% of the carcinogens; that is, it missed over half of the known carcinogens. The other three STTs were assays for chromosomal aberration, sister chromatid exchange in CHO cells, and the mouse lymphoma L5178Y cell mutagenesis assay. The sister chromatid exchange method, the mouse lymphoma mutagenesis assay, and the Ames test had 73%, 70%, and 45% sensitivity, respectively. However, because the Ames test is widely used, extensively described in the literature, and technically easier to conduct in a testing laboratory than the other tests, it is most often conducted in a screening program. These studies suggest that not all carcinogens are genotoxic (mutagenic) and not all mutagens are carcinogenic. Thus, although STTs for mutagenesis are helpful for predicting some carcinogens, STTs cannot predict all of them.

Dental Pulp Irritation Tests

Generally, materials to be tested on the dental pulp are placed in class 5 cavity preparations in intact, noncarious teeth of monkeys or other suitable animals. Care is taken to prepare uniformly sized cavities. After anesthesia and a thorough prophylaxis of teeth, cavities are prepared under sterile conditions with an efficient water-spray coolant to ensure minimal trauma to the pulp. The compounds are placed in an equal number of anterior and posterior teeth of the maxilla and mandible to ensure uniform distribution in all types of teeth. The materials are left in place from 1 to 8 weeks. Zinc oxide–eugenol (ZOE) and silicate cement have been used as negative and positive control materials, respectively.

At the conclusion of the study, the teeth are removed and sectioned for microscopic examination. The tissue sections are evaluated by the investigators without knowledge of the identity of the materials, and necrotic and inflammatory reactions are classified according to the intensity of the response. The thicknesses of the remaining dentin and reparative dentin for each histological specimen is measured with a photomicrometer and recorded. The response of the pulp is evaluated based on its appearance after treatment. The severity of the lesions is based on disruption of the structure of the tissue and the number of inflammatory cells (usually both acute and chronic) present. Pulpal response is classified as either slight (mild hyperemia, few inflammatory cells, slight hemorrhage in odontoblastic zone), moderate (definite increase in number of inflammatory cells, hyperemia, and slight disruption of odontoblastic zone), or severe (decided inflammatory infiltrate, hyperemia, total disruption of odontoblastic layer in the zone of cavity preparation, reduction or absence of predentin, and perhaps even localized abscesses). As with dental caries, the mononuclear cells are usually most prominent in the inflammatory response. If neutrophils are present, the presence of bacteria or bacterial products must be suspected. Some investigators now use ZOE cements to “surface-seal” the restorations to eliminate the effects of microleakage on the pulp.

Until recently, most dental-pulp irritation tests have involved intact, noncarious teeth, without inflamed pulps. There has been increased concern that inflamed dental pulp tissue may respond differently than normal pulps to liners, cements, and restorative agents. Efforts have been made to develop techniques that identify bacterial insults to the pulp. Usage tests that study teeth with induced pulpitis allow evaluation of types and amount of reparative dentin formed and will probably continue to be developed.

Dental Implants in Bone

At present, the best estimations of the success and failure of implants are gained from three tests: (1) penetration of a periodontal probe along the side of the implant, (2) mobility of the implant, and (3) radiographs indicating either osseous integration or radiolucency around the implant. Currently, an implant is considered successful if it exhibits no mobility, no radiographic evidence of peri-implant radiolucency, minimal vertical bone loss, and absence of persistent peri-implant soft-tissue complications. Previously, investigators argued that formation of a fibrous connective tissue capsule around a subperiosteal implant or root cylinder was the natural reaction of the body to a material. They argued that this was actually an attachment similar to the periodontal ligament and should be considered a sign of an acceptable material. However, in most cases it resembled the wall of a cyst, which is the body’s attempt to isolate the implanted material as the material slowly degrades and leaches its components into tissue. Currently, for implants in bone, implants should be completely encased in bone, the most differentiated state of that tissue. Fibrous capsule formation is a sign of irritation and chronic inflammation.

Mucosa and Gingival Usage Tests

Because various dental materials contact gingival and mucosal tissues, the tissue response to these materials must be measured. Materials are placed in cavity preparations with subgingival extensions. The materials’ effects on gingival tissues are observed at 7 days and again after 30 days. Responses are categorized as slight, moderate, or severe. A few mononuclear inflammatory cells (mainly lymphocytes) in the epithelium and adjacent connective tissue constitute a slight response. A moderate response is indicated by numerous mononuclear cells in the connective tissue and a few neutrophils in the epithelium. A severe reaction evokes a significant mononuclear and neutrophilic infiltrate and thinned or absent epithelium.

A difficulty with this type of study is the frequent presence of some degree of preexisting inflammation in gingival tissue. Bacterial plaque is the most important factor in causing this inflammation. Secondary factors are surface roughness of the restorative material, open or overhanging margins, and overcontouring or undercontouring of the restoration. One way to reduce the interference of inflammation caused by plaque is to perform dental prophylaxis before preparing the cavity and placing the material. However, prophylaxis and cavity preparation will themselves cause some inflammation of soft tissues. Thus, if margins are placed subgingivally, time for healing (typically 8 to 14 days) must be allowed before assessing the effects of the restorative agents.

ANSI/ADA Specification 41

Three categories of tests are described in the 1982 ANSI/ADA Specification: initial, secondary, and usage tests. This document uses the testing scheme shown in Fig. 5-13, B. The initial tests include in vitro assays for cytotoxicity, red blood cell membrane lysis (hemolysis), mutagenesis and carcinogenesis at the cellular level, and in vivo acute physiological distress and death at the level of the organism. Based on the results of these initial tests, promising materials are tested by one or more secondary tests in small animals (in vivo) for inflammatory or immunogenic potential (e.g., dermal irritation, subcutaneous and bony implantation, and hypersensitivity tests). Finally, materials that pass secondary tests and still hold potential are subjected to one or more in vivo usage test (placement of the materials in their intended contexts, first in larger animals, often primates, and finally, with Food and Drug Administration approval, in humans). The ANSI/ADA Specification No. 41, 1982 Addendum has two assays for mutagenesis—the Ames test and the Styles cell transformation test.

ISO 10993

In the 1980s, international efforts were initiated by several organizations to develop international standards for biomedical materials and devices. Several multinational working groups, including scientists from ANSI and the International Standards Organization (ISO) were formed to develop these standards. The final document (ISO 10993) was published in 1992. Further revision of the dental components of this document resulted in the publication of ISO 7405:1997 (Preclinical evaluation of biocompatibility of medical devices used in dentistry—Test methods for dental materials). This is the most recent standard available for biological testing methods for dental materials. The original ISO 10993 contained 12 parts, each dealing with a different aspect of biological testing for all types of medical and dental devices. For example, part 2 addressed animal welfare requirements; part 3 addressed tests for genotoxicity, carcinogenicity, and reproductive toxicity; and part 4 dealt with tests for interactions with blood. The standard divided tests into “initial” and “supplementary” tests to assess the biological reaction to materials. Initial tests are tests for cytotoxicity, sensitization, and systemic toxicity. Some of these tests are done in vitro, others in animals in nonusage situations. Supplementary tests assessed things such as chronic toxicity, carcinogenicity, and biodegradation. Most of the supplementary tests are to be done in animal systems, many in usage situations. The selection of tests for a specific material is left up to the manufacturer, who must present and defend the testing results. Guidelines for the selection of tests are given in part 1 of the standard and are based on how long the material will be present; whether it will contact body surface only, blood, or bone; and whether the device communicates externally from the body.

The current version of the ISO standard is available from the International Organization for Standardization (www.iso.ch), Case Postale 56, CH-1211 Geneva 20, Switzerland, with reference to document ISO 7405:1997—Dentistry (Preclinical evaluation of biocompatibility of medical devices used in dentistry—Test methods for dental materials). ANSI/ADA Specification No. 41 was also recently revised to conform to the ISO standard, and was reaffirmed in 2001 as ANSI/ADA Specification 41-1979 (Reaffirmed 2001) (Recommended Standard Practices for Biological Evaluation of Dental Materials and 41a, Addendum). The ANSI/ADA specification, which governs biocompatibility testing in the United States, is available from the Council on Scientific Affairs, American Dental Association (www.ada.org), 211 E. Chicago Avenue, Chicago, IL 60611, or the American National Standards Institute (www.ansi.org), 1819 L Street NW, Washington, DC 20036.

Microleakage

There is evidence that restorative materials may not bond to enamel or dentin with sufficient strength to resist the forces of contraction during polymerization, wear, or thermal cycling. If a bond does not form, or debonding occurs, bacteria, food debris, or saliva may be drawn into the gap between the restoration and the tooth by capillary action. This effect has been termed microleakage. The importance of microleakage in pulpal irritation has been extensively studied. Several early studies reported that various dental restorative materials irritated pulp tissue in animal tests. However, several other studies hypothesized that the products of microleakage, not the restorative materials, caused the irritation. Subsequently, numerous studies showed that bacteria present under restorations and in dentinal tubules might be responsible for pulpal irritation. Other studies showed that bacteria or bacterial products such as lipopolysaccharides could cause pulp irritation within hours of being applied to dentin.

Finally, a classic animal study shed light on the roles of restorative materials and microleakage on pulpal irritation. Amalgam, composite, zincphosphate cement, and silicate cement were used as restorative materials in class 5 cavity preparations in monkey teeth. The materials were placed directly on pulp tissues. Half of the restorations were surface-sealed with ZOE cement. Although some irritation was evident in all restorations at 7 days, after 21 days, the sealed restorations showed less pulpal irritation than those not sealed, presumably because microleakage had been eliminated. Only zinc phosphate cement elicited a long-term inflammatory response. Furthermore, the sealed teeth exhibited a much higher rate of dentin bridging under the material. Only amalgam seemed to prevent bridging. This study suggests that microleakage plays a significant role in pulpal irritation, but that the materials can also alter normal pulpal and dentinal repair.

Recently, the concept of nanoleakage has been put forward. Like microleakage, nanoleakage refers to the leakage of saliva, bacteria, or material components through the interface between a material and tooth structure. However, nanoleakage refers specifically to dentin bonding, and may occur between mineralized dentin and a bonded material in the very small spaces of demineralized collagen matrix into which the bonded material did not penetrate. Thus nanoleakage can occur even when the bond between the material and dentin is intact. It is not known how significant a role nanoleakage plays in the biological response to materials, but it is thought to play at least some role and it is suspected of contributing to the hydrolytic degradation of the dentin-material bond, leading ultimately much more serious microleakage.

The full biological effects of restorative materials on the pulp are still not clear. Restorative materials may directly affect pulpal tissues, or may play an auxiliary role by causing sublethal changes in pulpal cells that make them more susceptible to bacteria or neutrophils. It is clear, however, that the design of tests measuring pulpal irritation to materials must include provisions for eliminating bacteria, bacterial products, and other microleakage. Furthermore, the role of dentin in mitigating the effects of microleakage remains to be fully understood. Recent research has focused on the effects that resin components have on the ability of odontoblasts to form secondary dentin. Other research has established the rates at which these components traverse the dentin (see the next section on dentin bonding).

Dentin Bonding

Traditionally, the strength of bonds to enamel have been higher than those to dentin. Bonding to dentin has proved more difficult because of its composition (being both organic and inorganic), wetness, and lower mineral content. The wettability of demineralized dentin collagen matrix has also been problematic. Because the dentinal tubules and their resident odontoblasts are extensions of the pulp, bonding to dentin also involves biocompatibility issues.

After being cut, such as in a cavity preparation, the dentin surface that remains is covered by a 1- to 2-μm layer of organic and inorganic debris. This layer has been named the smear layer (see Fig. 5-3). In addition to covering the surface of the dentin, the smear layer debris is also deposited into the tubules to form dentinal plugs. The smear layer and dentinal plugs, which appear impermeable when viewed by electron microscopy, reduce the flow of fluid (convective transport) significantly. However, research has shown that diffusion of molecules as large as albumin (66 kDa) will occur through the smear layer. The presence of this smear layer is important to the strength of bonds to restorative materials and to the biocompatibility of those bonded materials.

Numerous studies have shown that removing the smear layer improves the strength of the bond between dentin and restorative materials with contemporary dentin bonding agents, although earlier research with older bonding agents showed the opposite effect. A variety of agents have been used to remove the smear layer, including acids, chelating agents such as ethylenediaminetetraacetic acid (EDTA), sodium hypochlorite, and proteolytic enzymes. Removing the smear layer increases the wetness of the dentin and requires that the bonding agent be able to wet dentin and displace dentinal fluid. The precise mechanism by which bonding occurs remains unclear. However, it appears that the most successful bonding agents are able to penetrate into the layer of collagen that remains after acid etching. There, they create a “hybrid layer” of resin and collagen in intimate contact with dentin and dentinal tubules. The strength of the collagen itself has also been shown to be important to bond strengths.

From the standpoint of biocompatibility, the removal of the smear layer may pose a threat to the pulp for three reasons. First, its removal juxtaposes resin materials and dentin without a barrier, and therefore increases the risk that these materials can diffuse and cause pulpal irritation. Second, the removal of the smear layer makes any microleakage more significant because a significant barrier to the diffusion of bacteria or bacterial products toward the pulp is removed. Third, the acids used to remove the smear layer are a potential source of irritation themselves. Nevertheless, removal of the smear layer is now a routine procedure because superior bond strengths are achieved. Some recent techniques that etch and “bond” direct pulp exposures make biocompatibility of the bonding agents even more critical, because the dentinal barrier between the materials and the pulp is totally absent.

Biocompatibility of acids used to remove the smear layer has been extensively studied. Numerous acids, including phosphoric, hydrochloric, citric, and lactic acids, have been used to remove the smear layer. The effect of these acids on pulp tissues depends on a number of factors, including thickness of dentin between the restoration and the pulp, strength of the acid, and degree of etching. Most studies have shown that dentin is a very efficient buffer of protons, and that most of the acid never reaches the pulp if sufficient dentin remains. A dentin thickness of 0.5 mm has proved adequate in this regard. Citric or lactic acids are not as well buffered, probably because these weak acids do not dissociate as efficiently. Usage tests that have studied the effects of acids have shown that phosphoric, pyruvic, and citric acids produce moderate pulpal inflammatory responses, but this resolves after 8 weeks. Recent research has shown that in most cases the penetration depth of acid into the dentin is less than 100 μm. However, the possibility of adverse effects of these acids cannot be ruled out, because odontoblastic processes in the tubules may be affected even though the acids do not reach the pulp itself.

Bonding Agents

Numerous bonding agents have been developed and are applied to cut dentin during restoration of the tooth. There have been a number of studies of biocompatibility of these bonding systems. Many of these reagents are cytotoxic to cells in vitro if tested alone. However, when placed on dentin and rinsed with water between applications of subsequent reagents as prescribed by the manufacturer, cytotoxicity is reduced. Longer-term in vitro studies suggest, however, that components of the bonding agents may penetrate up to 0.5 mm of dentin and cause significant suppression of cellular metabolism for up to 4 weeks after application. This suggests that residual unbound constituents may cause adverse reactions.

Hydroxyethyl methacrylate (HEMA), a hydrophilic resin contained in several bonding systems, is at least 100 times less cytotoxic in tissue culture than Bis-GMA. Studies using long-term in vitro systems have shown, however, that adverse effects of resins occur at much lower concentrations (by a factor of 100 or more) when exposure times are increased to 4 to 6 weeks. Many cytotoxic effects of resin components are reduced significantly by the presence of a dentin barrier. However, if the dentin in the floor of the cavity preparation is thin (<0.1 mm), there is some evidence that HEMA is cytotoxic in vivo.

Other studies have established the in vitro cytotoxicity of most of the common resins in bonding agents, such as Bis-GMA, triethylene glycol dimethacrylate, urethane dimethacrylate (UDMA), and others. Combinations of HEMA and other resins found in dentin bonding agents may act synergistically to cause cytotoxic effects in vitro. There have been very few clinical studies of diffusion of hydrophilic and hydrophobic resin components through dentin. These studies indicated that some diffusion of these components occurs in vivo as well. Interestingly, there has been one report that some resin components enhance the growth of oral bacteria. If substantiated, this would cause concern about the ability of resin-based materials to increase plaque formation.

Resin-Based Materials

For tooth restorations, resin-based materials have been used as cements and restorative materials. Because they are a combination of organic and inorganic phases, these materials are called resin composites. In vitro, freshly set chemically cured and light-cured resins often cause moderate cytotoxic reactions in cultured cells over 24 to 72 hours of exposure, although several newer systems seem to have minimal toxicity. The cytotoxicity is significantly reduced 24 to 48 hours after setting and by the presence of a dentin barrier. Several studies have shown that some materials are persistently cytotoxic in vitro even up to 4 weeks, whereas others gradually improve, and a few newer systems show little toxicity even initially. In all cases, cytotoxicity is thought to be mediated by resin components released from the materials. Evidence indicates that the light-cured resins are less cytotoxic than chemically cured systems, but this effect is highly dependent on the curing efficiency of the light and the type of resin system. In vivo, usage tests have been used to assess the biological response to resin composites. The pulpal inflammatory response to hemically cured and light-cured resin composites is low to moderate after 3 days when they were placed in cavities with approximately 0.5 mm of remaining dentin. Any reaction diminished as the postoperative periods increased to 5 to 8 weeks and was accompanied by an increase in reparative dentin (Fig. 5-15). With a protective liner or a bonding agent, the reaction of the pulp to resin composite materials is minimal. The longer-term effects of resins placed directly on pulpal tissue are not known, but are suspected to be less favorable.

Amalgam and Casting Alloys

Dental amalgam has been used extensively for dental restorations. Biocompatibility of amalgam is thought to be determined largely by the corrosion products released while in service. Corrosion, in turn, depends on the type of amalgam, whether it contains the γ₂ phase, and its composition. In cell culture screening tests, free or nonreacted mercury from amalgam is toxic. With the addition of copper, amalgams become toxic to cells in culture, but low-copper amalgam that has set for 24 hours does not even inhibit cell growth. Implantation tests show that low-copper amalgams are well tolerated, but high-copper amalgams cause severe reactions when in direct contact with tissue. In usage tests, the response of the pulp to amalgam in shallow cavities or in deeper but lined cavities is minimal, and amalgam rarely causes irreversible damage to the pulp. However, pain results from using amalgams in deep, unlined cavity preparations (0.5 mm or less remaining dentin), with an inflammatory response occurring after 3 days.

In cavities with 0.5 to 1.0 mm of dentin remaining in the floor, the cavity preparation should be lined for two other reasons. First, thermal conductivity with amalgam is significant and can be a problem clinically. Second, margins of newly placed amalgam restorations show significant microleakage. Marginal leakage of corrosion and microbial products is probably enhanced by the natural daily thermal cycle in the oral cavity, although long-term sealing of the margins occurs through the buildup of these corrosion products.

A number of high-copper amalgams are in clinical use. Usage tests reported that after 3 days, the pulpal response to these materials appears similar to that elicited by low-copper amalgams in deep, unlined cavities. At 5 weeks they provoke only slight pulpal response. At 8 weeks the inflammatory response was reduced. Bacterial tests on the high-copper amalgam pellets have revealed little inhibitory effect on serotypes of Streptococcus mutans, thus suggesting that the elements were not released in amounts necessary to kill these microorganisms. Although the high-copper amalgams seem biologically acceptable in usage tests, liners are suggested for all deep cavities.

Although not a commonly used material, amalgams based on gallium rather than mercury have been developed to provide direct restorative materials that are mercury-free. In cell culture, these alloys appear to be no more cytotoxic than traditional high-copper amalgams. These restorations release significant amounts of gallium in vitro, but the biological significance of this release is not known. In implantation tests, gallium alloys caused significant foreign body reaction. Clinically, these materials show much higher corrosion rates than standard amalgams. This leads to surface roughness and discoloration. There are few reports of pulpal responses to these materials.

Cast alloys have been used for single restorations, fixed partial dentures, ceramic-metal crowns, and removable partial dentures. The gold content in these alloys ranges from 0 wt% to 85 wt%. These alloys contain several other noble and non-noble metals that may have an adverse effect on cells if they are released from the alloys. However, metal ions released from these materials are most likely to contact gingival and mucosal tissues, while the pulp is more likely to be affected by the cement retaining the restoration.

Glass Ionomers

Glass ionomer has been used as both a cement (luting agent) and as a restorative material. Light-cured ionomer systems have been introduced; these systems use Bis-GMA or other oligomers as pendant chains on the polyacrylate main chain. In screening tests, freshly prepared ionomer is mildly cytotoxic, but this effect is reduced over time. The fluoride release from these materials, which is probably of some therapeutic value, causes cytotoxicity in vitro. Some researchers have reported that certain systems are more cytotoxic than others, but the reasons for this are not clear. The overall pulpal biocompatibility of glass ionomer materials has been attributed to the weak nature of the polyacrylic acid. It is unable to diffuse through dentin due to its high molecular weight. In usage tests the pulp reaction to these cements is mild, and histological studies show that any inflammatory infiltrate from ionomer is minimal or absent after 1 month. There have been several reports of pulpal hyperalgesia for short periods (days) after placing glass ionomers in cervical cavities. This effect is probably the result of increased dentin permeability after acid etching.

Liners, Varnishes, and Nonresin Cements

Calcium hydroxide cavity liners come in many forms, ranging from saline suspensions with a very alkaline pH (above 12) to modified forms containing zinc oxide, titanium dioxide, and resins. Resin-containing preparations can be polymerized chemically, but light-activated systems have are also available. The high pH of calcium hydroxide in suspension leads to extreme cytotoxicity in screening tests. Calcium hydroxide cements containing resins cause mild-to-moderate cytotoxic effects in tissue culture in both the freshly set and long-term set conditions. Inhibition of cell metabolism is reversible in tissue culture by high levels of serum proteins, suggesting that protein binding or buffering in inflamed pulp tissue may play an important role in detoxifying these materials in vivo. The initial response after exposing pulp tissue to these highly alkaline aqueous pulp-capping agents is necrosis to a depth of 1 mm or more. The alkaline pH also helps to coagulate any hemorrhagic exudate of the superficial pulp. Shortly after necrosis occurs, neutrophils infiltrate into the subnecrotic zone. After 5 to 8 weeks, only a slight inflammatory response remains. Within weeks to months, however, the necrotic zone undergoes dystrophic calcification, which appears to be a stimulus for dentin bridge formation. When resins are incorporated into the compound, these calcium hydroxide compounds become less irritating and are able to stimulate reparative dentin bridge formation more quickly than the Ca(OH)₂ suspension alone. Significantly, this occurs with no zone of necrosis, and reparative dentin is laid down adjacent to the liner (Fig. 5-16). This indicates that replacement odontoblasts form the dentin bridge in contact with the liner. However, some of these materials evidently break down with time and create a gap between the restoration and the cavity wall. Resin-containing calcium hydroxide pulp capping agents are the most effective liners now available for treating pulp exposures, and after treatment, the uninfected pulp undergoes a relatively uncomplicated wound-healing process.

Numerous investigators have analyzed the effects of applying thin liners such as copal varnishes and polystyrenes under restorations. These materials are not generally used under resin-based materials, because resin components dissolve the thin film of varnish. Because liners are used in such thin layers, they do not provide thermal insulation, but do initially isolate the dentinal tubule contents from the cavity preparation. They may also reduce penetration of bacteria or chemical substances for a time. However, because of the thinness of the film and formation of pinpoint holes, the integrity of these materials is not as reliable as that of other cavity liners.

Zinc phosphate has been widely used as a cement for castings and orthodontic bands, and as a cavity preparation base. Because the thermal conductivity of this cement is approximately equal to that of enamel (considerably less than that of metals), it has also been used to build up remaining tooth structure. In vitro screening tests indicate that zinc phosphate cement elicits strong-to-moderate cytotoxic reactions that decrease with time. Leaching of zinc ions and a low pH may explain these effects. The dilution of leached cement products by dentin filtration has been shown to protect the pulp from most of these cytotoxic effects. Focal necrosis, observed in implantation tests with zinc phosphate cements injected into rat pulp, confirm the cytotoxic effects of this cement when it contacts pulp tissue. In usage tests in deep cavity preparations, moderate-to-severe localized pulpal damage is produced within 3 days, probably because of the initial low pH (4.2 at 3 minutes). However, the pH of the set cement approaches neutrality after 48 hours. By 5 to 8 weeks, only mild chronic inflammation is present, and reparative dentin has usually formed. Because of the initially painful and damaging effects on the pulp of this cement when placed in deep cavity preparations, the placement of a protective layer of a dentin bonding agent, ZOE, varnish, or calcium hydroxide is recommended under the cement.

Zinc polyacrylate cements (polycarboxylate cements) were developed to combine the strength of zinc phosphate cements with the adhesiveness and biocompatibility of ZOE cements. In short-term tissue culture tests, cytotoxicity of freshly set and completely set cements has correlated with both the release of zinc and fluoride ions into the culture medium and with a reduced pH. Some researchers suggest that this cytotoxicity is an artifact of tissue culture because the phosphate buffers in the culture medium encourage zinc ions to leach from the cement. Supporting this theory, cell growth inhibition can be reversed if EDTA (which chelates zinc) is added to the culture medium. Furthermore, inhibition of cells decreases as the cement sets. In addition, concentrations of polyacrylic acid above 1% appear to be cytotoxic in tissue culture tests. On the other hand, subcutaneous and bone implant tests over a 1-year period have not indicated long-term cytotoxicity of these cements. Thus, other mechanisms such as buffering and protein-binding of these materials may neutralize these effects in vivo over time. Polyacrylate cements evoke a pulpal response similar to that caused by ZOE, with a slight-to-moderate response after 3 days and only mild, chronic inflammation after 5 weeks. Reparative dentin formation is minimal with these cements, and thus they are recommended only in cavities with intact dentin in the floors of the cavity preparations.

ZOE cements have been used in dentistry for many years. In vitro, eugenol from ZOE fixes cells, depresses cell respiration, and reduces nerve transmission with direct contact. Surprisingly, it is relatively innocuous in usage tests with class 5 cavity preparations. This is not contradictory for a number of reasons. The effects of eugenol are dose dependent and diffusion through dentin dilutes eugenol by several orders of magnitude. Thus, although the concentration of eugenol in the cavity preparations just below the ZOE has been reported to be 10⁻² M (bactericidal), the concentration on the pulpal side of the dentin may be 10⁻⁴ M or less. This lower concentration reportedly suppresses nerve transmission and inhibits synthesis of prostaglandins and leukotrienes (anti-inflammatory). In addition and as described before, ZOE may form a temporary seal against bacterial invasion. In cavity preparations in primate teeth (usage tests), ZOE caused only a slight-tomoderate inflammatory reaction within the first week. This was reduced to a mild, chronic inflammatory reaction, with some reparative dentin formation (within 5 to 8 weeks), when cavities were deep. For this reason, it has been used as a negative control substance for comparison with restorative procedures in usage tests.

Bleaching Agents

Bleaching agents have been used on nonvital and vital teeth for many years, but their use on vital teeth has increased astronomically in recent years. These agents usually contain some form of peroxide (generally carbamide or hydrogen peroxide) in a gel that can be applied to the teeth either by a dentist or at home by a patient. The agents may be in contact with teeth for several minutes to several hours depending on the formulation of the material. Home bleaching agents may be applied for weeks to even months in some cases. In vitro studies have shown that peroxides can rapidly (within minutes) traverse the dentin in sufficient concentrations to be cytotoxic. The cytotoxicity depends to a large extent on the concentration of the peroxide in the bleaching agent. Other studies have even shown that peroxides can rapidly penetrate intact enamel and reach the pulp in a few minutes. In vivo studies have demonstrated adverse pulpal effects from bleaching, and most reports agree that a legitimate concern exists about the long-term use of these products on vital teeth. In clinical studies, the occurrence of tooth sensitivity is very common with the use of these agents, although the cause of these reactions is not known. Bleaching agents will also chemically burn the gingiva if the agent is not sequestered adequately in the bleaching tray. This is not a problem with a properly constructed tray, and long-term, low-dose effects of peroxides on the gingival and periodontal tissues are not known.

Reactions to Ceramic Implant Materials

Most ceramic implant materials have very low toxic effects on tissues, because they are in an oxidized state and are corrosion resistant. As a group, not only are they minimally toxic, but they are nonimmunogenic and noncarcinogenic. However, they are brittle and lack impact and shear strength.

Reactions to Implant Metals and Alloys

Pure metals and alloys are the oldest type of oral implant materials. All metal implants share the quality of strength. Initially, metallic materials were selected on the basis of ease of fabrication. Over time, however, biocompatibility with bone and soft tissue and the longevity of the implant have become more important. Although a variety of implant materials have previously been used (including stainless steel and chromium-cobalt-molybdenum), the only metallic dental implant materials in common use today are titanium and titanium alloys.

Titanium is a pure metal, at least when first cast. In less than a second the surface forms a thin film of various titanium oxides. This oxide layer is corrosion resistant and allows bone to osseointegrate. The major disadvantage of this metal is that it is difficult to cast. It has been wrought into various forms, but this process introduces metallic impurities into the surface that may adversely affect bony response unless extreme care is taken during manufacturing. Titanium implants have been used with success as root forms that are left unloaded under the mucosa for several months before they are used to support a prosthesis. With frequent recall and good oral hygiene, implants have been maintained in healthy tissue for longer than two decades. Titanium-aluminum-vanadium alloys (Ti-6Al-4V) have been used successfully in this regard as well, but questions remain about the liability of released aluminum and vanadium.

Although titanium and titanium alloy implants have corrosion rates that are markedly less than other metallic implants, they do release titanium into the body. Currently there is no evidence that these released elements are a problem locally or systemically. The issue remains unresolved.

In soft tissue, the bond epithelium forms with titanium is morphologically similar to that formed with the tooth, but this interface has not been fully characterized. Connective tissue apparently does not bond to the titanium, but does form a tight seal that seems to limit ingress of bacteria and bacterial products. Techniques are being developed to limit down-growth of the epithelium and loss of bone height around the implant, as this will ultimately cause implant failure. Peri-implantitis is now a documented disease around implants and involves many of the same bacteria as periodontitis. The role of the implant material or its released components in the progression of peri-implantitis is not known.