I. INTRODUCTION

Dental amalgam has been used in the routine dental care of hundreds of millions of Americans, both children and adults, for the past 150 years. Amalgam is the most widely used dental restorative material because it can be applied in a broad range of clinical situations and is durable, easy to use, relatively insensitive to variations in handling technique and oral conditions, and inexpensive compared to alternative materials. More than 200 million restorative procedures were performed by U.S. dentists in 1990, of which amalgam restorations accounted for approximately 96 million (Nash,1991).

Dental amalgam has a much longer service record than most drugs and biomaterials in use today and, except for gold, all other dental restorative materials. There is more information about dental amalgam than about any other dental restorative material presently used. Yet, concerns are raised periodically about the safety of dental amalgam relative to one of its ingredients—elemental mercury.

These concerns have stimulated a comprehensive scientific assessment by the U.S. Public Health Service (PHS) of the benefits and risks of amalgam. To conduct this assessment, the Assistant Secretary for Health (ASH) charged the PHS Committee to Coordinate Environmental Health and Related Programs (CCEHRP) to examine the potential health risks of dental amalgam. This task was assumed by the standing Risk Assessment Subcommittee of CCEHRP. In order to facilitate a complete review of amalgam, an ad hoc subcommittee of CCEHRP was established in March 1991 to examine the benefits of amalgam. The reports of these two committees will be reviewed by the Risk Management Subcommittee of CCEHRP, which will develop an overall PHS statement on the risks and benefits of dental amalgam.

The present report was prepared by the Ad Hoc Subcommittee on the Benefits of Dental Amalgam. This subcommittee assessed the benefits of dental amalgam to oral and general health. In so doing, the subcommittee also compared dental amalgam to dental restorative materials that are or potentially may be available for achieving similar health and functional benefits.

Mercury Exposure

Dental amalgam is a mixture of approximately equal parts of elemental liquid mercury (43 to 54 percent) and an alloy powder containing a mixture of other metals, predominantly silver, but also tin and copper, with smaller amounts of zinc, palladium, or indium sometimes present. The relative proportion of these ingredients may vary, but the ingredients themselves have remained essentially the same through the years.

Mercury is distributed widely in the environment; it is found in food, air, and prescribed drugs and medicines. Mercury from dental amalgam restorations has generally been reported to contribute a relatively small percentage of an individual's total daily mercury exposure (Vostal, 1972; Shibko et al, 1976; WIIIiams, 1981; Olsson and Bergman, 1987; Mackert, 1987; Snapp, et al., 1989). Some studies, however, suggest that the relative contribution is higher (Clarkson et al., 1988; World Health Organization, 1991). The health effects from exposure to different levels of elemental mercury have been documented for decades, with much of the information derived from case reviews of exposure among industrial workers. The mercury in dental amalgam was considered to be inert until the development of highly sensitive devices for measuring mercury vapor, which permitted the discovery of previously undetectable levels of mercury vapor from dental restorations (Gay et al., 1979).

Whether such levels pose a risk to health is difficult to determine. Nevertheless, key questions must be raised. What are the risks of dental amalgam? What are the benefits of dental amalgam? Are these benefits and risks comparable to those of other dental restorative materials? Should we continue to use dental amalgam? If not, what are the implications of replacing dental amalgam with other restorative materials?

Federal Reviews

The Federal Government has sought over the years to explore concerns and to review existing data on dental amalgams. The present effort is the first formal attempt by PHS to specifically focus on the benefits of dental amalgam use.

During the 1980s, the National Institute of Dental Research (NIDR) sponsored a number of workshops, conferences and meetings addressing the safety of dental restorative materials and indications for their use. Recommendations from individual meetings often have led to subsequent consultations and deliberations. For example, as a result of an NIDR consultant meeting in 1983 that focused on mercury toxicity, the NIDR supported additional meetings on the biocompatibility of metals (1984); biocompatibility, toxicity, and hypersensitivity in dentistry and dental amalgam

and mercury toxicity (1985); the criteria for placement and replacement of dental restorations (1987); and the possible systemic responses from dental amalgam (1991). All of these meetings involved experts and recognized authorities from around the world.

Continuing to address the safety of all dental restorative materials, the NIDR, in collaboration with the Office of Medical Applications of Research, National Institutes of Health (NIH), convened in August 1991 a workshop on the effects and side effects of dental restorative materials. Other recent meetings have been held by the National Institute of Environmental Health Sciences and the Food and Drug Administration (FDA).

In addition, the American Dental Association (ADA) has conducted numerous professional symposia and has published scientific articles relating to dental amalgam and other restorative materials. Collectively, these deliberations have resulted in continued support for the routine use of dental amalgam as a restorative material, as evidenced by recent statements from a number of professional and voluntary associations (Consumer Union, 1990; ADA, 1991; National Multiple Sclerosis Society, 1991).

Since the inception of the U.S. Medical Device Amendments of 1976, the FDA has regulated the components of dental amalgam. An FDA amalgam task force was formed in 1984 to monitor the scientific literature and provider- and patient-supplied information related to amalgam use.

A recent meeting of the FDA Dental Products Panel on March 15, 1991, brought together world experts in mercury toxicity, medical and dental experts, and others in order to address issues related to amalgam safety and toxicity. While reaffirming the safety of amalgam for current use as a restorative material, the FDA panel called for additional research to answer some of the important questions that have been raised in animal studies. The FDA also participated in the present CCEHRP subcommittees to assess the risks and benefits of dental amalgam and will take the lead in determining appropriate future regulatory actions.

Changing Patterns of Oral Health and Dental Practice

The oral health of the American public and the practice of dentistry have undergone dramatic changes in the past 200 years. During the 18th and early 19th centuries, the public was resigned to the ravages of dental caries and, ultimately, the loss of many—and often all—teeth. Toothaches were treated commonly by extraction. By the middle of the 19th century, the development of restorative dentistry enabled individuals to retain teeth that became carious. The subsequent discovery of nitrous oxide and local anesthetics, along with improved methods of cutting tooth structure, further enhanced dental treatment. Extractions were still common, but replacement of portions of teeth, or even entire teeth, through more modern flexed and removable partial dentures became possible.

In the 20th century, restorative and prosthetic dentistry became more sophisticated, facilitated by the continuing development of clinical equipment, techniques and materials, including the high-speed drill. Over the past several decades, attention has been focused on preventing dental caries. The use of topical and systemic fluorides, improved oral hygiene products and practices, dietary modifications and dental sealants has contributed to dramatic declines in dental caries among school-aged children. In fact, smooth-surface caries (in contrast to pit and fissure caries) have been reduced to negligible levels in most children (NIDR, 1989). There also is some evidence of a decline in dental caries among young adults (Brown and Swango,1991) and clear evidence of a decline in tooth loss for Americans of all ages (Ismail et al., 1987; Meskin and Brown, 1988; Brown and Swango, 1991).

Further, among those who continue to experience caries, there are consistent clinical reports that, compared to prior years, lesions are smaller, easier to treat and require less destruction of healthy tooth structure in order to restore form and function. These trends suggest that the number, size and frequency of replacement of restorations and, thus, overall exposure to dental restorative materials will decrease.

Preliminary evidence for this trend comes from a recent survey indicating that a significant decline in the use of dental amalgam has taken place since 1979 (Nash, 1991). This comes at a time when newer materials, such as posterior composites, pit and fissure sealants, preventive resin restorations and glass ionomer cements, are being integrated into dental practice. Some of these materials have improved adhesive characteristics, so that removal of tooth structure can be minimized. Others, such as high copper amalgam alloys, which have been in widespread use for some years, demonstrate improved physical properties. Some of these new alloys also contain less mercury than the dental amalgam used several decades ago. If caries rates continue to decline and new biocompatible materials are proven to be effective, then fewer restorations will be needed and materials other than dental amalgam will be used relatively more frequently.

Nevertheless, it must be recognized that serious dental caries problems remain in the population. Analysis of epidemiological data suggests that the dental caries that occur among children today are concentrated in certain segments of the population. For those populations at high risk (PHS, 1990), the rate and severity of caries reflect patterns documented among the general population a generation ago.

As the population ages and adults retain more teeth, root caries are likely to become an increasing concern, along with coronal caries in adults whose general health is compromised or who suffer the side effects of medication or therapy, such as radiation treatment for head and neck cancer.

Finally, one must remember that there are so many restorations already in the mouths of patients that decades of replacement work lie ahead. Few restorations can be expected to last the lifetime of an individual. As the lifespan of individual Americans continues to increase, so will the need for replacement restorations. Even today, up to two-thirds of the restorations currently provided are estimated to be replacements (Maryniuk and Kaplan, 1986), contributing significantly to the more than $30 billion that is spent for dental care each year in the United States (Nash, 1991).

Approach to the Study

In developing this report, the Ad Hoc Subcommittee on the Benefits of Dental Amalgam acknowledged the changing environment of dental practice and oral health. This environment makes any study of the benefits of dental amalgam and other restorative materials complex. Inherent in the assessment of benefits is an assessment of the risks incurred. In addition, a relative comparison of dental amalgam to other available materials is warranted, which necessitates a discussion of the potential risks of other materials that might be used.

The subcommittee sought to address the following key questions:

The subcommittee addressed potential benefits to the patient, the public and the provider.

Several approaches for accessing information on the benefits of dental amalgam were used. A literature search was performed using the Medline system to identify articles published from 1980 through July 1991. Experts in the field of dental materials were asked to identify text materials that might be relevant. Several review papers were commissioned and prepared by expert consultants. These papers included useful bibliographies and provided validation for the overall scope and direction of the report. The papers and report of the NIH/NIDR technology assessment conference on the effects and side effects of dental restorative materials, held in August 1991, also were reviewed to assure that the scientific assessment of benefits and of the biocompatibility of dental restorative materials was as current as possible.

The scientific material reviewed for this report includes well-quantified, prospective studies using objective assessment methods; cross-sectional studies reporting data for a given point, or points, in time; retrospective studies reporting the longevity of restorations; laboratory reports; and articles published in rigorously reviewed scientific journals. The subcommittee's conclusions and recommendations reflect an overall assessment of the relevant science on the use and benefits of dental amalgam and other dental restorative materials.

The chapters that follow focus on comparisons between the characteristics of available and emerging restorative materials used in the restoration of posterior teeth and those of dental amalgam; the biocompatibility of dental restorative materials which are potential alternatives to amalgam; the relative costs and benefits of dental amalgam and other restorative materials; and policy and research implications regarding dental amalgam and alternative dental restorative materials. An extensive list of references also is provided.

II. MATERIALS, METHODS, AND INDICATIONS FOR THE RESTORATION OF POSTERIOR TEETH

Dentists today have numerous materials from which to select when restoring teeth, including amalgam, composite, glass ionomer cement, gold foil, cast metals, ceramics, and metalceramics. Specific clinical situations, however, dictate a much narrower range of appropriate restoration options.

The clinical decision as to which restorative material to place is complex, involving factors relating to the tooth, the patient, the clinician, and the properties of the restorative materials. Individual restorative materials ideally are applied in a defined set of clinical circumstances, and it is not possible to freely substitute one material for another and expect long-term success.

Because it is anticipated that this report will be read by individuals with dental knowledge ranging from limited to expert, this section begins with a background discussion of some of the factors that must be considered when selecting the appropriate restorative material for a specific clinical situation. These factors include the diagnosis of dental canes, treatment and material options, the properties of dental restorative materials, longevity of materials, and clinical decision-making in determining when a restoration has failed.

Finally, this section provides a brief description of all the currently available posterior restorative materials, including their individual advantages and disadvantages, as well as indications and contraindications for their use.

Diagnosis of Dental Caries (Tooth Decay)

Caries-producing bacteria are continuous residents of the oral cavity for people who have teeth and, thus, the opportunity for caries to manifest itself is always present for these individuals. Individuals may seek care from a dentist when they become aware of caries in their mouth. Pain, discoloration, a bad taste or odor, a sharp tooth or restoration edge felt by the tongue, or a dark spot on a tooth can trigger such a visit. The majority of caries, however, is likely to be asymptomatic (painless) and is found by the dentist during a periodic examination—the traditional "checkup." This process includes a direct visual and tactile examination, often supplemented with radiographic information, which permits assessment of the teeth even in areas that cannot be visualized directly.

When there is questionable, or very early evidence of caries, seen as a "white spot," or slight sticking of the explorer in a pit or fissure of a tooth, application of conservative preventive procedures, such as demineralizing fluoride applications, dental sealants, or preventive resin restorations (to be discussed in detail later) might be employed. These procedures, combined with followup observation of the suspected areas at later appointments, are a realistic alternative to preparing and restoring the tooth. At this point oral bacterial screening for strep mutans and lactobacillus may also be appropriate for assessing the caries disease risk for the individual.

Treatment and Material Options

For much of this century it was believed that dental caries could be treated away with restorations (Anusavice, 1989). Clearly, this is not the case. The long-term consequences of the insertion of the first restoration in any tooth always must be a consideration in the treatment decision (Lutz et al., 1987). Dental restorations have a limited clinical durability. As restorations need replacement, increasing amounts of tooth structure are lost and the patient may enter into a repetitive restorative cycle with larger restorations, weaker teeth, and more complex therapy (Elderton and Davies, 1984). indeed, it has been estimated that as many as two-thirds of restorations placed each year are replacements for existing restorations (Maryniuk and Kaplan, 1986). As the cavity size expands, the range of restorative materials to effectively employ becomes limited, and the option of appropriately placing a more economical direct restorative material that conserves tooth structure is lost.

Where active dental caries is evident (some longstanding caries may be arrested or nonactive and not require treatment), the dentist must decide whether or not to restore the tooth and, if restoration is required, which restorative material to employ for the anticipated situation.

Many factors must be considered relative to the placement of a restoration.

These many factors, several of which legitimately could be viewed differently by different patients and dentists, make it desirable to have a variety of options available for consideration. It is neither feasible nor desirable to use a single approach.

The range of acceptable treatment options for the patient who has overt caries includes: 1) The tooth can be restored; 2) the tooth can be extracted; or 3) no treatment can be rendered. A decision to have the tooth extracted or to forego treatment has both short- and long-term consequences, which are usually negative. Traditionally, this decision is one in which patients have participated actively through informed consent.

For teeth that are to be restored, the second decision concerning which procedure and material to use is traditionally one in which patients have been involved less fully. Dentists generally offer patients a "case presentation" outlining overall treatment options. For individual restorations, however, the specific choice of procedures and materials routinely has been made by the dentist.

Although caries is the predominant reason for restoration of teeth, several other clinical conditions, such as tooth fracture, restoration failure, and trauma, also may require restoration. The most common clinical conditions, treatment options, and restorative material options are summarized Table 1.

The Search for the Ideal Restorative Material

Despite modern dental materials and techniques, the oral cavity presents a demanding environment for restorative materials. Restorative materials break down for a variety of reasons including: dietary factors, masticatory stresses, acid-base shifts, temperature changes, failure of the tooth structure itself, the adhesive nature of plaque, the complex and different structures of cementum, dentin, and enamel, and interaction with other materials. The consequences of breakdown include recurrent caries, surface wear, leakage at the tooth-restoration interface (often referred to as microleakage), marginal fracture, bulk fracture, discoloration, corrosion, lack of biocompatibility, and sensitivity of the pulp to bacteria, chemicals, temperature, and pressure. Indeed, no test system is available that can duplicate readily the combined stresses of the oral cavity over a lifetime. Yet, even though the ideal restorative material does not exist, ideal characteristics can be outlined, as suggested below.

Although this list is extensive, undoubtedly there are additional desirable characteristics for a dental restorative material. Given the number and range of characteristics, it is not surprising that no restorative material available today meets all, or even most, of the requirements for each category of ideal properties.

Direct and Indirect Dental Restorative Materials

Dental restorations may be classified as direct or indirect. Direct restorative materials are inserted into cavity preparations in a soft, pliable state and then set hard. For direct restorations, the tooth is prepared and the filling material is placed during the same appointment. Direct restorations usually require less destruction of intact tooth tissues than indirect restorations. Direct fillings are appropriate only when sufficient tooth structure remains to maintain the integrity of the restorative material. The greater the loss of tooth structure, the more likely that an indirect restoration is indicated. Amalgam, resin based composite materials, glass ionomer cements, and compacted gold foil are examples of direct restorative materials.

Indirect restorations, such as inlays, onlays, and crowns, are fabricated in a dental laboratory on models made from impressions of the tooth prepared by the dentist. These restorations generally require multiple visits and placement of temporary restorations in the prepared teeth between appointments. In contrast to direct restorative materials, all indirect restorations are cemented as one-piece restorations and so require the removal of all undercuts, undermined tooth structure, and, often, significant amounts of healthy tooth tissues in order to produce parallel walls of the cavity preparation to allow insertion of the restoration and to provide adequate bulk of restorative material for strength. The two-step procedure and laboratory costs make indirect restorations significantly more expensive for the patient.

Recently, techniques have been developed where a composite inlay is prepared in the mouth, hardened outside the mouth, and cemented into the tooth during the same visit. A relatively new and not widely available technique, CADCAM (Computer-Aided Design and Computer-Aided Manufacture), uses a computer to record the prepared tooth optically and to direct the grinding of a ceramic (porcelain) block to produce an inlay, onlay, or crown for cementation at the same visit (Mörmann et al., 1990). These techniques eliminate the need to make an impression or temporary restoration, but are also significantly more expensive than a direct restorative material and are not generally available in dental offices.

Longevity and the Diagnosis of Failure in Restorative Materials

The longevity of a restoration depends upon many factors varying according to tooth type, location, condition, type of restoration, age of the patient, materials used, clinician capability, and the proper diagnosis of restoration failure. One of the major reasons that clear-cut longitudinal longevity data are deficient is the lack of objective measures for determining when a restoration has failed.

The dentist's decision to replace or repair a restoration involves numerous factors, including breakdown in marginal integrity, presence of recurrent or new caries, unacceptable esthetics, excessive wear, and pain symptoms. Criteria for quantifying the clinical failure of restorations have not been well defined, and diagnostic techniques used to determine the quality or functional status of restorations are grossly inadequate. Such criteria are necessary for definitively determining such factors as the clinical significance of leakage at the restoration tooth interface, bacterial colonization, presence of caries under restorations, and breakdown of marginal integrity at interproximal and subgingival margins (Anusavice, 1989). Currently, the decision to classify any of the above clinical conditions as a failure requiring replacement draws on the individual dentist's clinical judgment, which has been shown to be highly variable and not defined clearly (Maryniuk, 1984). Merrett and Elderton (1984) found that great variation exists among dentists in their decision to replace a restoration and that a third of these decisions would not be agreed upon by a randomly selected second dentist.

Surveys by Mjör (1979, 1980) of 85 dentists in private practice detailed the reasons for replacing amalgam and composite restorations. The primary reason for replacing amalgam restorations was recurrent caries (58 percent). Marginal degradation (9 percent), isthmus (bulk) fracture (13 percent), and tooth fracture (12 percent) also were commonly cited. For composites, poor marginal adaptation and anatomic form (40 percent), recurrent caries (20 percent), and discoloration (19 percent) were the most commonly cited reasons for failure. In these studies, the resin-based restorations were predominantly of a Class III type, with a few Class V restorations (see glossary for definitions).

Clear-cut data on longevity also are lacking because of the difficulty in designing studies that include the many pertinent variables, such as quality of the restoration, patient hygiene and dietary habits, materials used, operator proficiency, and conditions under which the restorations were placed. Maryniuk (1984) reviewed longevity data from 21 published studies of various restorative materials and, on the basis of study design, validity of data, and failure criteria, concluded that because of methodological flaws in these studies and discrepancies in the determination of failure, no generalizable information is available to describe and predict the lifespan of restorations.

It must be emphasized that great variations exist and limited data are available from general practice, especially in regard to posterior composite restorations. It is also important to consider that the quality of restorative materials has improved considerably in the past 15 to 20 years, especially for the composite resins. Many of the studies assessing longevity utilized materials, both composite and amalgam, that are no longer in clinical use, having been replaced by superior materials. The anticipated longevity of improved composite restorations placed in general practice has not yet been established.

Factors Influencing the Success of a Restorative Material

The long-term clinical success of a restoration is attributable to diverse factors that can be grouped into three general categories—patient, clinician, and restorative material (Figure 1).

It is not possible to rank these major categories in order of significance because the principal cause of restoration failure will vary considerably among patients, dentists, and materials.

There may even be batch-to-batch variation within the same material. Success or failure may well be due to various combinations of factors, and the relative contribution of each factor has not been clarified. For example, if one factor was improved 20 or 30 percent, would there be a corresponding increase in restoration longevity? The following discussion reviews the relative importance of factors within each of the three categories.

Patient characteristics

These factors play an important role in the long-term clinical success of a restoration. Cooperation by the patient during a procedure allows moisture control and visual access, and aids in proper tooth preparation and placement of the restoration. The size of the restoration, dietary factors, personal prevention practices, and damaging oral habits, such as bruxing or ice-chewing, are also important.

Several studies have found a statistically significant correlation between recurrent caries and poor patient oral hygiene and have concluded that the oral hygiene status of the patient should be a major determining factor in clinical decision making (Goldberg et al., 1981; Eriksen et al., 1986).

Dental clinician factors

Numerous studies have demonstrated that the dentist's skill affects the longevity of restorations (Abramowitz, 1966; Elderton, 1976; Lavelle, 1976; Smales and Gerke, 1986). These studies have concluded, for example, that faulty preparation, contouring, and overhangs account for a significant number of restoration failures. It has been demonstrated that previous experience with a given technique and procedure is important for clinical success. There is generally a learning curve when using new materials. It has become increasingly difficult for dentists to remain familiar with the full range of available materials because of the rapid pace of new materials development. This factor likely contributes to inappropriate use of some materials, improper placement of restorations, and, most assuredly, limited data upon which clinicians can make decisions about the use of materials.

Technique Sensitivity

Small changes in manipulation can produce large differences in the quality and performance of a restoration. This effect is known as technique sensitivity. In general, materials that are technique-sensitive demonstrate variation in physical properties, mechanical properties, handling characteristics, and/or clinical performance based on relatively small procedural changes. Some materials are more technique-sensitive than others. Also, materials may perform well under ideal laboratory or study conditions, while under "average" dental practice condition performance may vary significantly.

Some materials are so technique-sensitive that widely variable results can occur even within a single practice (Smales and Gerke, 1986). Technique sensitivity also is an issue in the dental laboratory where, for example, a number of problems may occur in the processing of metal and porcelain crowns, which may ultimately result in delayed failure of the porcelain metal prosthesis after it has been cemented on the tooth.

One of the major reasons amalgam has long been the most widely used restorative material is its relatively low technique sensitivity compared to other dental restorative materials (Jordan, 1985), although studies have shown large differences in the strength of amalgam based on mixing time and speed (Brackett et al., 1987). Yet, variations in mixing, placement, and contamination are not generally as critical as with most other restorative materials. For example, even a slight amount of moisture may result in the immediate failure of a gold foil or greatly reduced adhesion and physical properties of a composite. A moisture-contaminated amalgam may have reduced physical properties and a shorter lifespan, but still provide reasonable service. Although amalgam is easy to manipulate and place, the best results and longer life of the restoration are obtained when placed under ideal clinical conditions.

Letzel and Vrijhoef (1984) concluded that the amalgam alloy, patient, and operator each had a significant influence on marginal quality of amalgam restorations over a 5-year period. The patient and operator effects decreased with time, whereas the type of alloy exhibited a stronger effect with time. Mjör (1986) suggested that handling effects are a most important factor in producing long-lasting amalgam restorations.

Figure 2 presents a hypothetical plot of the percentage of restoration failures from materials that are highly technique-sensitive, moderately technique-sensitive, or technique-insensitive. This figure also could represent the failure curves of three dentists with little experience, moderate experience, or extensive experience using the same technique-sensitive product.

Rates of wear

Gradual wear of the teeth is a natural process. The rate of wear depends on individual factors such as the abrasiveness of the diet, oral habits (e.g., bruxism or grinding of teeth for extended periods of time), toothbrushing, and other factors. The challenge for the dental materials scientist and the clinical practitioner is to match the rate of wear of the restorative material with that of tooth enamel. If the restorative material wears faster than the enamel, there is a chance for supereruption or shifting of the opposing tooth and greater stress transfer to the supporting tooth structure, which may result in tooth fracture. If the restorative material is harder than the enamel, such as porcelain and base metals, rapid loss of enamel may occur in the opposing teeth.

Various measurement techniques have been developed to determine the wear rates of restorative materials (Cvar and Ryge, 1971; Goldberg et al., 1980; Leinfelder et al., 1983). However, it should be recognized that wear resistance is one of the most difficult properties to evaluate in materials science. The mechanism of clinical wear has proven difficult to duplicate in the laboratory and may vary with time, depending on tooth location, chewing patterns, restoration size, material handling, and other factors. A major problem in drawing meaningful conclusions from data on clinical wear is the discrepancy between the types of studies conducted and the data obtained by different research groups studying the same materials (Jones, 1990).

Various studies have demonstrated that factors such as the width and complexity of posterior restorations, the finishing and polishing techniques, and occlusal stress are significant in the wear of materials (Berry et al., 1981; Mjor, 1981; Osborne and Gale, 1981; Mahler and Nelson, 1984; Qvist et al., 1986; Reel and Mitchell, 1987). The longevity and durability of posterior amalgam, composite, or glass ionomer restorations are related to their size, configuration, and location. Small restorations and those placed in nonstress-bearing situations are more durable. With time, however, larger restorations and remaining tooth cusps are more likely to fail because of the larger functional area of the restoration. In general, there also is more stress the farther a restoration is placed posteriorly in the mouth. Greater stress leads to a more rapid breakdown and need for replacement (Reel and Mitchell, 1987).

Amalgam and gold wear at a similar rate as tooth enamel. Adequately glazed or polished porcelain and glass ceramic also wear favorably compared with tooth enamel. However, if the glaze is lost or the porcelain is not repolished after adjustments have been made, then these materials have been demonstrated to increase wear of the opposing teeth. The low impact and fracture resistance and the poor wear resistance of glass ionomer limit its use in posterior teeth to Class V restorations and smaller cavities in primary (baby) teeth. Composite still is considered to have problems with excessive wear under stress, which limits its use, in posterior situations, to minimal-stress-bearing situations. Eriksen et al. (1986) also reported that composites were associated with a greater risk of caries than amalgam or cast gold restorations. Although earlier formulations of posterior composites exhibited high wear rates, more recent products have wear sates similar to that of amalgam (Qvist et al., 1990). Long-term clinical trials will be needed, however, before drawing final conclusions, since material properties determined under in vitro conditions are not always identical to those demonstrated under clinical conditions.

Leakage Along the Tooth-Restoration Interface

Despite the controversy over the significance of marginal gaps, leakage at the tooth-restoration interface has not been perceived as a significant problem with amalgam restorations. Corrosion products from amalgam from along the restoration-tooth interface, suppressing the penetration of fluids, debris, and microorganisms, thereby, making the restoration "self-sealing" (Phillips, 1984).

Despite significant improvements over earlier formulations, the greatest problem with existing composites is polymerization shrinkage (tendency of the material to contract as it sets), which breaks the seal formed between the material and the tooth structure and allows gaps to form at the tooth-restoration interface, especially adjacent to margins that extend into dentin. Polymerization shrinkage results in stresses in the tooth (Jensen and Chan, 1985), the resin itself (Bowen et al., 1983), and the interracial region between the tooth and the restoration. Thermal stress also has been shown to increase marginal leakage around composite restorations (Momoi et al., 1990), as has the use of composites with higher viscosity and lower water-sorption values (Crim, 1989).

Repairability

People are living longer and tooth loss across all ages is decreasing (Ismail et al., 1987; NIDR, 1989; Brown and Swango, 1991). Given that each time a restoration is replaced, more tooth structure is lost, it is highly desirable to increase the serviceable lifetime of a restoration. An important decision by the dentist, which greatly affects the longevity of a given restoration, is whether to remove an entire defective restoration or to repair only the defective portion.

Traditionally, dentists have regarded "repair" as "patchwork dentistry" and have frowned on the practice. Repair, generally, has not been considered acceptable in the dental school curriculum and only recently has been suggested in dental textbooks (Baum et al., 1981 and 1985; Sturdevant et al., 1985) and the dental literature (Cowan, 1983; Boyd, 1989; Ettinger, 1990). Thus, restorations defective in only one area, but otherwise acceptable, have been completely removed, resulting in more loss of tooth structure.

Lack of standards to determine restoration failure, and the lack of sensitive diagnostic tests to detect recent caries cause dentists to err on the side of caution when faced with an uncertain diagnosis. Matynink and Kaplan (1986) and Boyd (1989) found that dentists more frequently replace restorations placed by other dentists than those placed by themselves. Additionally, Elderton (1977, 1984) found that the replacement amalgam can be as deficient as the original, even when done by an experienced clinician.

Guidelines or criteria for repair of restorations are not well established. Subjective judgment cannot be standardized easily. For example, a bad restoration margin judged by dentist A may be judged acceptable by dentist B. Likewise, a color mismatch may be acceptable to patient A and dentist A, but not to patient B or dentist B.

All available, direct restorative materials possess certain properties that, in the oral environment, result in eventual breakdown. It is a major advantage if a material can be easily, effectively, and economically repaired in order to extend the serviceable life of the restoration. Most of the direct restorative materials, including amalgam and composite, are repaired easily. For patients at low risk for decay having good diet, proper oral hygiene, and an acceptable saliva flow rate, repair can be a more conservative and preferable option than replacement.

Dental Materials for Restoring Posterior Teeth

The restorative materials available for posterior restorations are described briefly below and summarized according to their relative advantages, disadvantages, clinical indications, and contraindications. Table 2 provides a quick summary of the most frequently used materials for restoring posterior teeth.

Amalgam

Dentists have more than a century of experience using amalgam as a direct filling material. Amalgam is strong and durable enough to withstand the pressures of chewing; it is relatively inexpensive and easy to place; and it has properties that may help prevent recurrent caries (Phillips, 1984; 0rstavik, 1985). Dental amalgam is widely considered to be unesthetic, however, and questions regarding its safety have been raised virtually from the time of its first use.

Although amalgam has a range of defined optimal uses, its low cost to patients, ease of manipulation, and durability allow it to be used in areas where a stronger or more esthetic material ideally would be placed. For example, large amalgam fillings are often used even when a casting would be stronger. Lost cusps are replaced with amalgam when a cast onlay would be more durable and long-lasting. Incipient caries are restored with amalgam when a preventive resin and sealant would conserve tooth structure and function.

Composite

Composites have excellent esthetic properties and are applied most frequently in anterior tooth cavities. In the 1980's, the mechanical and physical properties of composite resins, fillers, coupling agents, and bonding agents were improved, and a number of brands have been approved by the American Dental Association for posterior restorations in nonstress-bearing situations. When used in large restorations, including virtually all posterior situations, an incremental filling technique must be utilized to ensure complete polymerization and to minimize the effects of shrinkage of the resin on the final size of the restoration. Compared with amalgam restorations, the longer time necessary to properly complete this procedure has implications relative to moisture

contamination and financial cost to the patient.

Exacting techniques are necessary for the successful placement of a composite resin. Composite restorations rely upon mechanical and chemical adhesion of the material to the tooth surface to seal margin areas and, thus, are sensitive to moisture contamination during placement. The difficulties presented in controlling saliva and the moisture normally present on tissues of the tooth create an unfavorable surface for adhesion. This is a major consideration in clinical decision making, because moisture control is difficult in many patients and in the most posterior areas of the dentition. Marginal leakage and the formation of recurrent caries are likely consequences of moisture contamination.

Problems with excessive wear under stress and high technique sensitivity still limit composite use in posterior situations; however, they are popular with individuals who strongly value esthetics. Additionally, composites have been advocated as an alternative for persons concerned about the mercury content of amalgam. This situation may result in the inappropriate use of composite in stressbearing situations.

Increasingly, individuals desire attractive, as well as functionally satisfactory, teeth. Composite resin currently has limited, but important, applications as a posterior restorative material. Its use in treating incipient lesions in conjunction with sealants is an important step in the long-term conservation of tooth structure. Unfortunately, as a recent worldwide survey has shown, the teaching of placement techniques for posterior composites is limited. Professional dental education rarely includes significant opportunities for students to gain clinical experience in the use of composite resins as posterior restorative materials (Wilson and Setcos, 1989). There are indications, however, that these opportunities are increasing and the use of composite for specific posterior restorative situations, such as preventive resins for lesions in minimal- stress-bearing areas, likely will become a more integral part of the dental curriculum as further research data become available.

Pit and Fissure Sealants and Preventive Resin Restorations

A contemporary report on dental restorative materials must include a discussion of sealants and preventive resins. Although technically a preventive measure, sealants increasingly play an important role in a conservative restorative treatment strategy, in which the goal is to preserve healthy tooth structure.

Some of the pits and fissures of teeth largely are fused during tooth development, while others may remain microscopically open and impossible to clean. The latter fissures are potential sites for the colonization of cariesforming bacteria, despite the best oral hygiene efforts. Sealants are resin materials that flow easily and, when applied to the acid-etched surfaces of pits and fissures of posterior teeth, bond to the enamel and seal the pits and/or fissures from bacterial invasion and debris.

The decline in caries rates experienced over the past 30 yeas in the United States has resulted largely from the addition of fluoride the drinkng water and to dentrifices (PHS, DHHS, 1991). Fluoride, however, has its greatest effect on the smooth surfaces of the teeth and lesser benefit protecting pits and fissures. Graves and Burt (1975) found that more than 91 percent of the callous surfaces in permanent fist moles of children up to -grade 6 were in pits and fissures. The National Children's Oral Health Survey of 1979-80 reported that 84 percent of the cases experience of 5- to 17-year-old children occurred in pit and fissure surfaces (NIDR, 1981).

Many studies have demonstrated the efficacy of pit and fissure sealants in reducing caries. Horowitz et al. (1977) reported a 37-percent reduction in occlusal caries after 5 years. Meurman et al. (1978) reported a 59.8-percent reduction in cases after 5 years and Simonsen (1987) reported a 47-percent reduction in cases after 10 years.

Despite numerous published studies on the safety and effectiveness of sealants, the dental profession has been slow to adopt their use. The 1985-86 National Children's Oral Health Survey found that less than 7-percent of children 5 to 17 years old had received sealants (NIDR, 1989). Dentists have cited a number of reasons for their reluctance to place sealants, including concern about sealing in cases. Several studies, however, have demonstrated that sealants can be applied over incipient active cases, resulting in a rapid drop in viable bacteria count and elimination of the nutrition source, rendering the bacteria nonviable and stopping further progression of the disease (Jeronimus et al., 1975; Handelman et al., 1976; Going et al., 1978; Mertz-Fairhurst et al., 1987). Studies also have cited a perceived lack of cost-effectiveness (Stiles et al., 1976; Messer and Nustad, 1979; Lennon et al., 1980; Simonsen, 1982), and lack of third-party coverage as reasons why sealants have not been accepted more widely (ADA,1981). Other cost-effectiveness studies, however, indicate decreased long-term expense for sealed teeth, as compared to unsealed teeth (Stiles et al., 1976; Simonsen, 1989). The more cases-prone the population, the more effective is this treatment modality.

The 1983 National Institutes of Health Consensus Development Conference on Dental Sealants in the Prevention of Tooth Decay concluded that pit and fissure sealants were a safe and effective means for preventing pit and fissure cases. "Expanding the use of sealants would substantially reduce the occurrence of dental cases ... and improve the health of the public and reduce expenditures for the treatment of dental disease" (NIH, 1984).

Investigators are examining other uses for sealants, such as sealing over the surface of amalgam restorations to reduce or eliminate the release of mercury vapor from the surface. Promising results also have been reported in improving wear rate and marginal integrity and in reducing bacterial leakage for both posterior composites and amalgams by applying sealants over the surface (Mertz-Faithurst and Ergle, 1991; Dickinson et al., 1990).

Preventive Resin Restorations (PRR) utilize a combination of composite and sealant to treat early caries in pits and fissures. Despite the name of preventive resin, this technique is employed after caries has formed and the caries is judged to be deeper into dentin than appropriate for management by fissure sealant alone (Anusavice, 1989). In the interest of conserving tooth structure, PRR involves only removing the affected tooth structure, acid-etching the enamel, placing composite in the prepared cavity, and using sealant in the remaining pits and fissures. These conservative restorations are minimal in size and are used in nonstress bearing situations. The PRR can be considered an alternative and, in most situations, preferable to the placement of conservative Class I amalgams (Anusavice, 1989; Simonsen, 1990; Mjor, 1991). A treatment pattern starting with early identification of caries, fissure sealants, and preventive resins will conserve tooth structure and help to forestall, or significantly defer, the need for major restorative care later.

Few data are available on the long term clinical evaluation of preventive resins. In one study, Houpt et al. (1986) demonstrated a 72-percent survival rate for PRR after 5 years. Simonsen and Landy (1987) have also reported favorable results. These studies, however, are small, and comparisons of preventive resins with restorations for which they are generally substituted, (i.e., Class I amalgam) are needed.

Glass Ionomer

Glass ionomers were introduced commercially about 10 years after dental composites and enamel-bonding materials came to the market. Composites proved to have a competitive edge over glass ionomers as restorative materials because of their higher strength.

The original glass ionomers had a number of clinical drawbacks that limited their acceptance. Clinical failings were related to manipulation, setting sequence, early moisture sensitivity, esthetics, and surface texture. Consequently, glass ionomes, as restorative materials, did not gain the acceptance of dentists to the same extent as composites.

For a few important reasons, glass ionomers recently have gained wider acceptance as a restorative material for defined situations. They bond chemically to tooth structure and release fluoride. Patient response to glass ionomers is usually excellent because the placement technique can be extremely conservative and requires little, if any, drilling (Hunt 1990); the procedure is usually quick and painless and often does not require local anesthesia; and the resulting restoration is fairly esthetic.

Developments in the formulation of glass ionomers have made them useful as a cavity-lining material and for cementation and preventive applications, as well as for their original intended use as a direct filling material. As a filling material, glass ionomers are perhaps best used in restoring deciduous teeth and in Class V restorations involving gingival erosion and abrasion defects in adults. The use of glass ionomer may play an increasingly important role in the growing geriatric population which is retaining their teeth longer, but facing a concomitant increased risk of root caries.

While glass ionomer appear to be satisfactory in many anterior applications and primary teeth, their use continues to be limited in permanent posterior teeth, particularly with stress-bearing restorations. Limitations include low tensile strength, low impact and fracture resistance (brittleness), and degradation.

Glass ionomers are not recommended for restorations where toughness and resistance to wear are major considerations (Sulong and Aziz, 1990). It has been recognized, generally, that the wear resistance of glass ionomer is inadequate in areas of occlusal contact. Clinical studies have shown that a gradual loss of contour can be expected because of chemical degradation and surface wear (McLean, 1980). One study of a glass ionomer product, using a commercial composite resin as a control, reported that the glass ionomer abraded about three times more rapidly, by volume, than the composite (Smales and Joyce, 1978).

In the early to mid-1980s, it was found that the introduction of metal fibers or powder in the glass ionomer system (glass-cermet cements) significantly improved abrasion resistance (McLean, 1984). The addition of silver alloy powder to glass ionomer, in particular, resulted in a number of improvements in its physical properties (Simmons, 1990). The silver cermet material has a light gray color, which is no more unesthetic than silver amalgam, but it has a major disadvantage in that it has a low fracture toughness, making it of limited value in regions subjected to the stresses of mastication (Croll, 1990; McLean and Gasser, 1985).

Glass ionomes, including cermets, are technique sensitive (Knibb and Plant, 1989; Mount, 1990b; Smales et al., 1990; Smales and Gerke, 1990; Watson, 1990). The setting reaction and maturation of glass ionomer restorations are relatively slow. Even with the most skillful placement technique, however, the success of a glass ionomer restoration may hinge on the composition of commercial glass ionomer materials, which may vary widely from manufacturer to manufacturer (Smith, 1990).

Although glass ionomer exhibit significantly less polymerization shrinkage than composites, some curing contraction generally occurs, leading to the formation of marginal gaps (Feilzer et al., 1988; Saunders et al., 1990). Marginal leakage associated with glass ionomer can be reduced still further if the restoration is covered with a thin layer of posterior composite resin (Guelmann et al., 1989).

Gold Foil

For centuries, gold foil has been applied to various surfaces for ornamentation or utility. Early use of foil also included adaptation to teeth where defects existed. With time, as new instruments became available and better skills were developed, more and more uses were found for this material in dentistry. Newer forms of the gold appeared and made easier the meticulous task of condensation, first with powdered gold (Baum, 1965), then with other forms of electrolytic-formed gold (mat gold).

Properly placed, direct-filling gold restorations are excellent replacements and can be expected to last for 20 years or more. Their clinical indications, however, are limited. Most frequently, they are placed into small cavities in nonstress-bearing situations, or to repair defective margins of cast gold inlays, onlays, and crowns. Large restorations of foil are difficult to place. In addition, pure gold is too soft and ductile to withstand the forces that are exerted on most posterior restorations. Furthermore, larger restorations in the anterior of the mouth are not esthetic.

The major difficulties with direct gold restorations are the technique sensitivity of placement, the skill and meticulous attention required of the dentist, the potential damage to the pulp and/or periodontal tissues because of trauma during placement, and the overall cost to the patient in time and money.

Although many dentists still believe that this material should continue to be placed and that the technique should be taught, the use of gold foil is limited and diminishing. Its use is declining primarily because of the high cost associated with this technique, the limited number of applications for its use, and the availability of acceptable alternative materials, primarily composite, glass ionomer, or amalgam.

Cast Metal and Metal-Ceramic Restorations

Cast metal restorations such as inlays, onlays, and crowns are indirect restorations generally requiring two or more appointments. The successful fabrication and placement of these restorations depend on close attention by the dentist and laboratory technician to minute details in a multiprocedural, step-by-step process. Each restoration is designed carefully to restore anatomy, function, appearance, and comfort.

The decision to restore with inlays, onlays, crowns, and/or bridges depends on many factors, including the degree of tooth destruction, esthetic needs, missing teeth, oral hygiene, and the financial capability and desires of the patient. There is over a sixfold increase in price for cast restorations in teeth that could be restored with amalgam.

Cast metal posterior inlays only cover a portion of the occlusal surface. It is believed that these inlays weaken the tooth and may lead to cuspal fracture (Norman, 1991). Therefore, onlays or crowns that cover and protect the cusps are the recommended restoration for highstress-bearing situations where there is inadequate natural tooth remaining to support a direct restorative material and where one or more cusps need replacement.

Since tooth preparations for full crowns are easier for the dentist to prepare and are less likely to involve the pulp than tooth preparations for an onlay, they are becoming the cast restoration of choice when cuspal coverage is indicated.

The selection of casting alloys depends on the location of the tooth in the mouth, the presence and type of adjacent restorations and opposing teeth, the need for esthetics, and the patient's financial capability.

Casting alloys for metal-cesmic restorations are divided into three categories: high noble, with at least 60-percent noble metal content and at least 40-percent gold; noble metals, with at least 25-percent noble metal; and predominantly base metal, which has less than 25-percent noble metal. The noble metals in casting alloys are primarily gold, platinum, and palladium (ADA, 1984).

Base metal alloys, which can include nickel, beryllium, cobalt, and chromium have gained widespread use, especially in the United States, because of their low cost and superior physical properties. These properties include: high mechanical strength, resistance to sag when fired with porcelain at high temperatures, porcelain bond strength, thermal compatibility between porcelain and metal, and resistance to corrosion. A survey of dentists in Minnesota by Olin et al. (1989) revealed that 62 percent of dentist prescriptions written in that year were for base metal alloys.

Fabricating fixed prosthetics like crowns, inlays, and onlays is extremely technique-sensitive, and the skill and attention to detail by both the dentist and technician play a major role in the longevity of these devices.

Metal-ceramic restorations (porcelain fused to metal, PFM) combine the strength of cast metal with the esthetics of porcelain. In these restorations, porcelain is baked onto a thin coping (cast metal substructure) prepared from an impression of the tooth. Metal-ceramic restorations have been successfully employed for single crowns and multiunit bridges for the past 30 years. These restorations are used for more than 60 percent of the crown and bridge restorations performed (Anusavice, 1991).

One of the main disadvantages of metal-ceramic crowns is the high abrasive potential of ceramics relative to opposing natural teeth or other dental materials. Mahalick et al. (1971) reported a high wear rate of enamel-porcelain surface interactions, as compared to gold alloy against enamel. DeLong et al. (1986) reported a high coefficient of friction between enamel and dental porcelain and concluded that the wear of porcelain appears to be one order of magnitude (10X) greater then that of dental amalgam. When the surface of the porcelain is roughened through occlusal adjustment, care must be taken to restore a highly polished surface or severe wear of the opposing tooth structure may result.

The longevity of noble metal inlays compared with amalgam was reported in two studies. Jahn et al. (1989) found no significant difference between gold inlays and amalgam after 2 years. Mjor et al. (1990) reviewed a number of clinical trials and reported longevities of cast metal restorations that ranged from slightly less than to 90 percent greater than that of amalgam restorations. Schwartz d al. (1970) reported a mean lifetime of 10.3 years for full metal crowns. Recurrent caries was the primary cause of failure for 58 percent of the crowns. Kerchbaum and Voss (1977) estimated that only 3 percent of PFM restorations failed over a 10-year period. When properly fabricated, however, it is likely that a cast metal or metal-ceramic restoration will be in service for many years longer than large, direct restorations.

The failure rates reported for PFM restorations appear to be relatively low (Kerchbaum and Voss, 1977; Coomaert et al., 1984; Glantz et al., 1984; Leempoel etal.,l985;Christiansen,1986). The reasons for failures of PFM crowns and bridges fall into five major categories: (1) clinical deficiencies, (2) laboratory deficiencies, (3) inadequate dentist-technician communication, (4) technique sensitivity of materials, and (5) patient factors. The principal cause of failure varies considerably among dentists and among laboratory technicians.

Although 70 percent of the dentists indicated that PFM crowns with porcelain occlusion on maxillary first molars were highly successful, only 26 percent indicated that they would have used PFM crowns with porcelain occlusal surfaces for their own personal treatment. Most of these dentists preferred, for their own maxillary first molars, a three-quarter gold crown (53 percent), compared with a PFM crown with metal occlusion (7 percent), a seven-eighths gold crown (11 percent), or a full-gold crown (1 percent) Christensen, 1986). This preference is likely because of the potential for increased wear if the porcelain surface loses its glaze or polish.

The success of any cemented restoration will depend on the strength and lack of solubility of the luting agent (cement), as well as the ability to achieve an extremely close fit between the tooth and restoration. A tight junction must be established between the restoration and the finish line of the preparation on the tooth. A space of only 50 microns between the restoration and tooth will result in a visible cement line. This cement line eventually will result in a defective seal that will permit progressive dissolution of the cement from beneath the restoration. When the cement dissolves, food particles, oral fluids, and bacteria can enter the defect and may cause caries in the supporting tooth (Zander, 1957).

There are limits to the use of PFM and cast metal restorations. For the most part, they are used only on permanent teeth in adults because the necessary removal of tooth structure for proper fabrication would threaten pulp vitality in children and even many young adults. Also, the restorations are costly, amounting to more than eight times the cost of amalgam.

Ceramic Restorations

Approximately 30 years ago, the term glass ceramic was given to certain formulations of porcelain which, by the controlled nucleation and growth of crystals at elevated temperature, fanned a polycrystalline material. Compared with feldspathic porcelain, the resultant material exhibited greater strength and toughness, a variable coefficient of thermal expansion, greater ease of fabrication, lower processing shrinkage, better translucency control, good thermal shock resistance, and excellent chemical durability. Its use in dentistry has expanded rapidly.

Dental porcelain and newer glass ceramics have a multitude of applications. Different types are employed in the construction of artificial denture teeth, full crowns, inlays, onlays, and laminate veneers, and as the esthetic veneer over a metal substructure for crowns and bridges. Although some porcelains and glass-ceramics have been considered for bridges, the failure rates to date have been unacceptably high.

Strength tests on the newer glass ceramics encouraged manufacturers to develop all-ceramic crowns for posterior teeth. However, the strength of all ceramic crowns is significantly less than that of porcelain-fused-to-metal (PFM) crowns. Thus, all-ceramic crowns should be restricted to lower-stress situations, such as the anterior teeth and in patients with smaller biting force and no history or evidence of bruxing. For posterior teeth, all-ceramic crowns should be considered only for low-stress conditions in which PFM and metal crowns are unacceptable.

Dental ceramics generally are used to restore extensively damaged, diseased, or fractured teeth because of their excellent esthetics, wear resistance, chemical inertness, and low thermal conductivity. In addition, they match the characteristics of tooth structure fairly well. Ceramic restorations represent one of the few esthetic choices for treatment of small-to-large defects in posterior teeth. Compared with glass ionomers, dental ceramics are more durable, less technique sensitive, and more predictable from an esthetic viewpoint, but they are more costly by a factor of more than six. Compared with office-produced composites (direct) and lab-processed composites (indirect), ceramics are more color-stable, higher in flexural strength, more resistant to abrasion, potentially more abrasive to opposing enamel, and again, over six times more costly. Compared with all-metal or ceramic-metal crowns, ceramic restorations are more esthetic, but generally have a shorter life expectancy.

Ceramic inlays, onlays, full crowns, and veneers have become popular alternatives over the past 5 years because of improvements in physical properties, cementation techniques, and an increased public demand for esthetic materials.

One of the problems encountered with the use of ceramic materials in dentistry has been their inherent brittleness and low tensile strength. A small flaw can enhance tensile stress that can initiate a crack and cause fracture of the restoration.

Although newer formulations are significantly stronger than earlier types of porcelain, data on clinical success are not available.

CAD/CAM—One of the potential uses of glass ceramics is in the production of machined inlays, onlays, and crowns by means of CAD/CAM systems, as described earlier. The precision of defining and machining the marginal area of CAD/CAM prostheses has been reported to be in the range of 0 to 250 m (Mörmann et al., 1987). Recent improvements in hardware and software have considerably improved the overall precision of this method; however, the technique is very demanding (Roulet and Herder, 1990). Rekow et al. (1991) stated that crowns produced using a CAD/CAM system can fit at least as well as those produced with ideal casting conditions.

Herder (1988) reported that postoperative pain occurred in 31 percent of the cases after inlays were cemented in place, even though marginal openings could be detected in only 0.3 percent of the cases overall. This pain disappeared in all cases within a period of 4 to 12 weeks. Of greater concern is the observation by Herder (1988) float, after 6 months, submargination (loss of material at the restoration-tooth junction) occurred in 19 percent of the interproximal sites and 50 percent of the occlusal sites, which was explained by the excessive wear of the resin cementation material (Roulet, 1987). In summary, inadequate long-term clinical data are available from controlled studies, and no data are available to indicate the performance of these inlays under routine private-practice conditions. CAD/CAM is not widely available and CAD/CAM restorations will likely be similar in price or slightly higher than metal-ceramic restorations.

Because the failure rates of all-ceramic restorations are relatively high, the esthetic demands for posterior restorations are not sufficient to recommend their general use in preference to metallic restorations, especially for molar sites. Metal ceramic restorations, which are indicated for moderate-to-high stress conditions, can be recommended when esthetics are of concern.

III. BIOCOMPATIBILlTY OF DENTAL RESTORATIVE MATERIALS

Ideally, a dental material that is to be used in the oral cavity should be harmless to all oral tissues—gingiva, mucosa, pulp, and bone. Furthermore, it should contain no toxic, leachable, or diffusible substance that can be absorbed into the circulatory system, causing systemic toxic responses, including teratogenic or carcinogenic effects. The material also should be free of agents that could elicit sensitization or an allergic response in a sensitized patient.

Rarely, unintended side effects may be caused by dental restorative materials as a result of toxic, irritative, or allergic reactions. They may be local and/or systemic. Local reactions involve the gingiva, mucosal tissues, pulp, and hard tooth tissues, including excessive wear on opposing teeth from restorative materials. Systemic reactions are expressed generally as allergic skin reactions. Side effects may be classified as acute or chronic.

In this chapter, the Ad Hoc Subcommittee on the Benefits of Dental Amalgam addresses these biocompatibility issues in relation to all dental posterior restorative materials used to replace missing tooth structure. Only local reactions with regard to dental amalgam are considered. Potential systemic side effects from dental amalgam use are addressed in the report of the Risk Assessment Subcommittee.

Standards and Testing

The oral environment is especially hostile for dental restorative materials. Saliva has corrosive properties, and bacteria are ever present. This environment demands appropriate biological tests and standards for evaluating any material that is developed and intended to be used in the mouth. Such tests and standards, which have been developed in the past 10 to 15 years, serve as the basis for recommending any dental restorative material (Stanley, 1985; Mjör, 1991).

Until a few years ago, almost all national and international dental standards and testing programs focused entirely on physical and chemical properties. The physical and chemical requirements set forth in the specifications for dental materials have been based on published clinical studies and clinical use of the materials; that is, the specifications lag behind materials development. Today, however, dental materials standards require biological testing as well. The science of dental materials now encompasses a knowledge and appreciation of certain biological considerations associated with the selection and use of materials designed for use in the oral cavity (Phillips, 1991).

In accordance with existing standards, all dental materials should pass primary tests (screening to indicate cellular response), secondary tests (evaluating tissue responses), and usage tests in animals before being evaluated clinically in humans.

Evaluation of dental products for safety and efficacy has historically been the purview of both the ADA and the

FDA. The U.S. Medical Device Amendments of 1976 were the first regulations that emphasized the need for biological standardization and testing of dental, as well as medical, materials. In accordance with these regulations, all dental materials are reviewed for safety and effectiveness and classified by the FDA as Class I, II, or III, according to risk.

Class I materials are those considered to be of low risk in causing adverse reactions and, thus, require only "general controls," such as good manufacturing practices and record-keeping by the producer. Materials in Class II must satisfy the requirements outlined in the current ANSI/ADA specifications. The most extensive testing is required for Class III materials, which includes full safety and efficacy assessments prior to marketing. The FDA regulates the components of dental amalgam and, with the advice of a panel of experts, classifies the alloy component into class II (special controls) and the USP grade mercury into Class I (general controls). The amalgamated product is not classified.

restorative materials. This program is intended to record systematically any side effects from medical and dental devices and to establish a database from which their potential adverse effects can be evaluated. These data can then be used to determine the types of regulatory actions that should be taken in the future. Use of this system by the dental profession has been low.

In the past few years there has been an increasing demand for safety evaluation and control of dental restorative materials. However, the task is difficult.. In general, qualitative and quantitative information about substances released intraorally from dental alloys and other dental materials is meager (Hensten-Pettersen, 1986; Klotzer and Reuling, 1990). Verified diagnoses of side effects are not often established because the mild nature of the reactions are not viewed as justifying more extensive testing involving several medical specialties. Published studies of side effects among patients therefore are mostly inconclusive, especially because much information is based solely on questionnaire surveys among patients and dentists.

Questionnaires do not provide objective information on side effects that may be attributable to dental treatments because of varying respondent ability to observe, evaluate, and clearly describe symptoms and because the symptoms could be caused by factors other than dental treatment. Few large-scale studies have been conducted to evaluate systematically the frequency and severity of side effects of restorative materials, and most of the existing clinical citations of side effects are case reports. Although these are important in providing the basis for larger epidemiological studies, only systematic, cross-sectional, and, possibly, longitudinal studies truly can establish the magnitude and nature of side effects associated with restorative materials.

A balanced discussion of the biocompatibility of dental amalgam requires consideration of the relative biocompatibility of other restorative materials that potentially could serve as alternatives to amalgam This chapter includes a review of the biocompatibility of other dental restorative materials as well, focusing on those used for posterior restorations. These include resin-based composites, glass ionomer materials, gold foil and dental casting alloys, ceramics, and other materials. Current standards and testing of dental materials, potential side effects, and biocompatibility are also presented.

Side Effects

Side effects to dental materials are believed to be rare and, generally, those that have been reported are mild (Kallus and Mjör 1990; Hensten-Pettersen and Jacobsen, 1991). Yet, given the millions of treatments provided, many individuals potentially may be affected. Consideration must be given to the relative biocompatibility of all dental restorative materials.

The incidence and severity of side effects of restorative materials have been included as part of a few general studies on dental materials. Two basically different research approaches have been followed, one focusing on the general population and one on defined risk groups.

One approach has been to evaluate side effects in dental patients and retrospective dentist reports of clinical experience (Kallus and Mjor,l991). In these studies, no systemic toxic reactions to dental restorative materials have been reported. Local reactions that have been reported are not severe, the most common being lichenoid reactions in the oral mucosa and skin reactions such as rashes, dermatitis, and eczematous lesions. These reactions depend on the chemical composition of the materials used and their degradation products, absorption, accumulation, and other factors associated with leachable substances from the restoration.

The other approach has been to study personnel (e.g., dental personnel) who handle restorative materials as part of their daily work (Ahlbom et al., 1986; Nylander et al., 1986 and 1989; Ericson and Kallen, 1990; Hensten-Pettesen and Jacobsen, 1990; Munksgard et al., 1990). Studies of dental personnel are presented in the Risk Assessment Subcommittee Report.

Local Reactions

Lichenoid/white or erosive red lesions in the oral mucosa have been reported in direct topographical relation to dental amalgam, composite, and other restorative materials (Banoczy et al., 1979; Bolewska et al., 1990; Lundström, 1984; Lind et al., 1986; Holmstrup, 1991). Hietanen et al. (1987), on the other hand, found no evidence of hypersensitivity to dental restorative materials in patients with oral lichen planus.

In part, these local reactions may be allergic in origin (Lied et al., 1986; Lind, 1988; Kaaber, 1991), occurring at the site of exposure or distant from the site of exposure, or they may be toxic in nature (Hensten-Pettesen and Jacobsen, 1991), having a direct, irritating effect at the site of exposure. In either case, the cause often is difficult to ascertain. It must be recognized that toxic reactions are dose dependent, while allergic reactions are virtually dose-independent.

Based on published case reports and surveys of adverse reactions, most verified adverse effects of dental materials are allergic reactions (Kaaber, 1990). Dental materials contain components that are common allergens, such as chromium, cobalt, mercury, eugenol, components of resin-based materials, colophonium,and formaldehyde. Direct toxic effects also may occur, for example, from formaldehyde-containing materials (Brodin et al., 1982) and as enhanced tissue responses to methyl methacrylate in formaldehyde-sensitized individuals (Kallus, 1984). However, it is important to keep in mind that the presence of an allergen or a toxic component in a material is not a verification of the reason for a reaction per se. Even in patients with a known hypersensitivity to specific substances, other contributing factors may elicit a reaction. On the other hand, the more potent the allergen or toxic component, the more likely will be the association with adverse reactions.

Concentration and length of exposure are two important considerations. As pointed out by Paracelsus more than 400 years ago, dose is a critical factor in toxicology. Many materials, including table salt, water, and mercury, can be toxic if given in sufficiently high concentrations. Even potentially toxic amounts of materials appear to be well managed by normal physiological clearing mechanisms if the amount of exposure per unit time is low. In the oral cavity, concentration versus time is mitigated by the filtering effects of dentin, the smear layer of cutting debris, and/or the base material between the source of the toxin and the pulp (Stanley, 1990).

Besides concentration and filtering effects, one must consider a number of procedural influences involved with providing a new restoration. For instance, a composite restoration typically includes mechanical cutting, pressures from placement or curing, drying effects, bacterial exposure, acid-etching procedures, enamel and/or dentin bonding, and light-curing steps. Any of these procedures may produce pulpal reactions that are not associated with the filling material itself.

Biocompatibility Factors: Local Reactions

In Designing Restorative Materials, dental scientists give particular attention to several key factors relating to a material's biocompatibility with the human organism. These include potential tissue responses, leakage of bacteria at the tooth-filling interface, shrinkage of materials, and stress created in the tooth structure from restoration procedures.

Tissue Responses. Restorative materials may elicit responses from the pulp, gingiva, and oral mucosa. The pulp may be irritated in a number of ways: by cutting, mechanical procedures involved in preparing the tooth cavity, and the restorative material itself, potential leaching of a material's components, improper placement of the restoration, leakage of bacteria at the margins of the restoration caused by an inadequately placed or incompletely cured material, and agents (e.g., acid) used to prepare the tooth cavity and secure bonding of the restorative material. Severe and prolonged irritation may be irreversible and lead to permanently damaged pulp tissue.

Since the landmark studies of Langeland in the 1950s, knowledge of the biology of the human dental pulp and its capacity to recover from injury has increased tremendously (Langeland, 1957; Stanley, 1984, 1989). Although some restorative materials have been known to cause pulp lesions when placed as far as 1.5 mm from the pulp, most only produce significant and often irreversible lesions when placed less than 1.0 mm, and usually less than 0.5 mm, from human pulp tissue (Stanley, 1991).

Appropriate lining agents are useful for preventing severe lesions. Some agents (e.g., calcium hydroxide) act to stimulate formation of secondary dentin, while others (e.g., glass ionomer cement) protect the prepared dentin and enamel from leakage around the restoration and invasion of bacteria into pulpal tissues.

Acid etching of dentin during treatment may elicit pulpal effects by increasing the permeability of dental tissues to restorative materials and microbial products. These effects depend largely on the particular acid used and the skill of the practitioner (Stanley, 1988 and 1989).

Similar considerations apply to gingival and mucosal tissue. Effects may be temporary in response to the procedure or longer lasting in response to the amount of material placed and agents used.

Leakage of Bacteria. The presence of bacteria at the tooth-filling interface and the consequences of the penetration of microorganisms into the dentin and pulp because of leakage around the margins of a restoration have received considerable attention. Some authors believe that infection due to the penetration of microorganisms around the restoration, and especially beneath it, is the greatest threat to the pulp, rather than the toxicity

of any restorative material (Brannstrom and Nyborg, 1971; Brannstrom and Vojinovic, 1976; Bergenholtz et al., 1982). Pulpal lesions that become more severe 1 week or more after a dental restoration has been placed may be due to marginal leakage.

Severe pulp lesions, in the short term, may be related to the toxicity of the restorative material used. All restorative materials may leach to some extent, and the amount, toxicity, and allergenicity of components that do leach vary considerably (Mjör,l991). When the pulp becomes devitalized after a restorative procedure, consideration must be given to the combined effects of the mechanical and thermal injury induced during cutting of the tooth substance, the toxicity of the restorative materials, and bacterial action (Qvist and Stoltze, 1982).

Shrinkage of Materials. This problem, particularly relevant to composites, may occur when a restorative material is bonded to the surface of a tooth, creating stress as the polymer sets and pulls on the tooth. The larger the cavity and the larger the mass of the restoration, the more extensive the shrinkage can be. Shrinkage may be the cause of postrestoration sensitivity. The degree of bonding and the shrinkage of material will affect the extent of marginal opening that allows bacteria to leak under the material, especially at the critical gingival marginal area, thus potentiating pulpal irritation and even recurrent decay. Eventually, it may be necessary to replace the restoration.

Stress from Restoration Procedure. As a restorative material is condensed into a cavity or a restoration is cemented to a tooth, material may be forced into open dentinal tubules and pressure gradients may arise that place force on live tissue. Individuals may demonstrate initial tissue reactions to these procedures, but these generally subside within hours or days after a procedure is completed.

Restorative Materials

The biocompatibility of specific dental restorative materials is summarized below.

Dental Amalgam

Because of its extensive use, there is more information available about the biocompatibility of dental amalgam than about any other dental restorative material. Local soft tissue reactions to dental amalgam fillings are addressed in this report. Potential, systemic biological effects are addressed in the report of the Risk Assessment Subcommittee.

Over the years, amalgam has provided excellent clinical service with few documented adverse effects. Mercury from dental amalgam does not seem to contribute to any pulpal responses (Stanley, 1991). Leakage also has not been perceived as a significant problem with amalgam restorations. In fact, corrosion products from amalgam form along the restoration-tooth interface, suppressing the penetration of fluids, debris, and microorganisms (Phillips, 1984) and, over time, improving the adaptation of dental amalgam to the tooth structure.

Amalgam restorations, in general, have been considered to be either inert or only mildly irritating to the pulp or body tissues in dogs, rats, and humans (Manley, 1942; Schroff, 1946-47; James and Schour, 1955; Silberkweit et al., 1955; Massler, 1956; Welder et al., 1956). Any pulpal response to amalgam seems to be related mainly to the physical insertion of the amalgam, that is, the pressure of condensation (Stanley, 1991), and is usually of short duration. Skogedal and Mjor (1979) indicate that alloys containing the highest percentages of copper cause slightly more pulpal responses after 1 to 2 months in monkeys than conventional amalgam.

In 1962, Swerdlow and Stanley reported extreme degrees of leukocytic accumulation in the pulps of human teeth restored initially with amalgam, which resolved as early as 15 days after the restoration, suggesting that the physical insertion of the amalgam was a contributing factor, rather than the properties of the amalgam itself. They also demonstrated that the pressure of grinding procedures and dehydration can contribute to intensified pulpal responses (Stanley and Swerdlow, 1960).

In 1968, Soremark et al., using radioactive mercury (Hg¹⁹⁷), showed that mercury reached the pulp in humans by 6 days, if no liner was used, and that the rate of mercury diffusion into enamel and dentin was related inversely to the degree of mineralization of the tooth, which is higher, generally, in older patients. However, Kurosaki and Fusayama (1973) showed that mercury from amalgam in humans and dogs did not reach the pulp; they thus postulated that the mercury does not dissolve, but, rather, penetrates back into the amalgam and reacts further with previously unreacted alloy cores. Stephen and Ingram (1969) reported similar findings, as did van der Linden and van Aken (1973). Only zinc and tin occurred in high concentrations in the dentin beneath the amalgam restorations.

Resin-Based Composites

Composite materials that are certified or accepted by the ADA are required to pass a variety of tests. However, like amalgam, longitudinal, in viva research on the biocompatibility of composite resins is scant, particularly on those developed for posterior restorations (Bayne,l991). Composite material, however, has been shown to elicit a chronic inflammatory response in viva (Nasjleti et al., 1983), to be cytotoxic in cell culture (Hensten-Pettersen and Helgeland, 1977, 1981; Mjor, 1977; Wennberg and HenstenPettersen, 1981; Kasten et al., 1982), to be potentially allergenic (Nathanson and Lockart, 1979; Kallus et al., 1983; School, 1991), and to inhibit RNA synthesis (Caughman et al., 1990).

Composite materials are associated with many organic compounds whose long-term allergenic and toxicity potentials have not been established (Anusavice, 1989). The organic matrix contains, in addition to a variety of different dimethacrylates, a number of reactive chemicals to make the materials optimal as dental restorative materials. These components include initiators, such as benzoyl peroxide or camphorquinones; accelerators, such as toluidines, anilines, aminobenzoic acid, and others, depending on whether the polymerization is chemically or light induced; inhibitors, such as hydroquinonmonomethylether or 2,6 ditertiary butyl-p-cresol; plasticizers, such as dibutylphytlate; and pigments which are metal salts (Munksgaard, 1989). Many of these components are found in household glues and, thus, sensitization and allergic reactions to these components may occur on the basis of dental or other exposures. Chemicals from both the resin (Inoue and Hayashi, 1982) and filler (Soderholm, 1983) components of composite have been shown to leach out from the set material. Degradation and wear of resin-based composites release their components, including the fillers, silanized layer, and polymer matrix. Minute amounts of these materials may be swallowed, exposing components and fragments of restorative material to stomach acids and enzymes. Subsequent dissolution and absorption of ionic species under this condition have just begun to be explored by Freund (1990) and others, and the significance is unknown. Also, minute amounts of formaldehyde may form as a degradation product of resin-based composite materials (Øysaed et al., 1988).

Incomplete polymerization is an inherent problem with resin-based composites, and it predisposes the material to degradation and leaching into adjacent tissue. Incomplete polymerization occurs when a number of reactive groups do not participate in the polymerization (Ruyter and Svendsen, 1978). In addition, any surface layer exposed to oxygen/air will be polymerized incompletely (Ruyter and Svendsen, 1978; Ferracane and Greener, 1984) and such layers will release an amount of monomer or degradation product from the composite corresponding to the thickness of the unpolymerized layer (Øysæd et al., 1988). It is important to obtain as complete polymerization as possible through the entire restoration in order to minimize pulpal responses (Stanley, 1984). The level of pulpal response to composite resins is intensified especially in deep cavity preparations when an incomplete curing of the resin permits an even higher concentration of residual unpolymerized monomer to leach into the pulp (Swartz et al., 1983; Visible Light Bonding, 1985).

Great strides have been made in the curing and polymerization of resin-based composites, but an ideal system has not yet been obtained.

During the past 20 years, pulp and dentin reactions to composite materials have been related more to bacterial leakage than to the toxicity of the material (Brännström and Nyborg, 1971; Bergenholtz et al., 1982; Brännström, 1985; Bergenholtz, 1989; Stanley, 1989). Leakage, adverse pulp reactions, and the development of recurrent caries are associated with polymerization shrinkage of composites and imperfect adhesive bonding of the material to the tooth cavity (Bower, 1991). Thermal stress also increases marginal leakage around composite restorations (Momoi et al., 1990), as does the use of composites with higher viscosity and lower water-sorption values (Cnm, 1989). Although there is less leakage with heat-and-light-treated composite inlays (Wends, 1991; Shortall et al., 1989; Biedem~an, 1989), the problems associated with marginal gaps have not been solved completely (Cheung, 1990).

Pulp studies of individual components show slight, but varied responses (Stanley et al., 1979). Early developed composite materials produced severe pulp reactions (Langeland et al., 1966), but most studies of pulp reactions to modem materials show no, or moderate, reactions (Mjör and Wennberg, 1985; Qvist and Thylstrup, 1989), although severe pulp reactions also have been reported (Qvist et al., 1989). A number of factors will affect the result (Qvist and Stoltze, 1982), and it is recommended, generally, that a base, or liner, be used in conjunction with composite restorations. A report on the use of plastic materials as retrograde root fillings revealed slight tissue reactions (Andreasen et al., 1989).

Pain/toothache has been reported following the insertion of composite restorations (Boksman and Jordan, 1986;

Wilson et al., 1986; Leinfelder, 1991), especially large ones (Qvist and Thylstrup, 1989). Again, a number of factors may affect the pain, such as polymerization shrinkage (which can cause severe strain to develop within the tooth), the effect of acid on the dentin, leakage, and reactions to the materials per se.

Gingival reactions following contact with composite materials have not been described. However, inflammatory reactions adjacent to unfilled, cold-cured acrylic resin have been noted, while heatcured resins are well accepted (Podshodly, 1968). The permeability of the gingival epithelium (Squier, 1973) allows penetration of leachable components and, thus, there is potential for toxic and allergic reactions with composite materials. Lichenoid reactions in the oral mucosa in contact with resin-based composite materials have been described (Lied, 1988). Such lesions usually heal spontaneously when the restoration is replaced with a different type of restorative material.

In addition to toxic components, such as remaining monomer in cold-cured plastics, plaque adhesion to resin-based materials may play a significant role in gingival reactions. It has been demonstrated in vitro and in viva that more plaque attaches to plastic restorative materials than to other materials or enamel (Skjorland, 1973; Sonju and Skjorland, 1976). Plaque at the restoration/tooth junction also contains elevated levels of cariogenic bacteria (Svanberg et al., 1990).

The allergic reactions associated with resin-based materials can affect dental personnel working with the materials, as well as patients (Malmgren and Medin, 1981; Hensten-Pettesen and Lyberg, 1986; Munksgaard, 1989; Hensten-Pettesen, 1989;Kaaber, 1990). Documentation and conclusive diagnosis of individual patient reactions are difficult and sometimes confused by confounding factors or multiple allergies (Hensten-Pettesen and Mjor, 1989).

A classic problem for usage studies is to isolate the effects of the material of interest from the effects of other materials that are part of the overall procedure. Cavity lines, enamel acid-etch material, and bonding agents are used routinely with composite. In addition, long-term effects of bacterial leakage confound measurements of potential chemical effects of the filling material and may be the primary cause for pulpal responses to composite filling materials (Skogedal and Eriksen, 1976). Excessive acid etching before placing a composite also may cause irritating effects by permitting the ingress of bacteria (Brännström, 1981).

The pulpal effects of composite materials and procedures currently are a relatively minor concern for most clinicians, but postoperative sensitivity and loss of vitality associated with posterior composite restorations have been reported (Bowen, 1991). These reports have resulted in a renewed emphasis on careful cavity preparation and careful use of restorative materials and lines (Council on Dental Materials, Instruments and Equipment, 1986; Bales, 1987; O'Hara et al., 1988; Swift, 1989).

Glass Ionomer Materials

Almost all of the clinical and toxicological information on glass ionomer has been developed on lines, bases, and cements, which were the first widespread clinical applications of this material. Clinical trials have focused mainly on restoration retention and integrity.

Smith and Ruse (1986) attempted to identify the mechanisms of potential sensitivity related to glass ionomer use. They measured the pH of cements following mixing and concluded that the initially low pH may produce chemically irritating conditions for the dental pulp. The actual pH depends importantly on manipulation procedures, such as the mixing ratio of components (Mount, 1986). Woolford (1989) also observed that the pH of glass ionomer cements remained very low during the fist hour after setting, noting differences between a variety of commercial products. Brännström et al. (1991) commented that the low pH could occur for a long time and probably complicated the evaluation of other biological properties of glass ionomers.

When glass ionomer cements first were introduced, pulpal responses were classified as bland, moderate, and less irritating than with other cements or composite resins. Clinical studies show that such cements may cause early inflammatory reactions on newly prepared dentin, which resolve within a few days. Screening tests in cell cultures indicate that glass ionomers can be cytotoxic and therefore, protective calcium hydroxide liners are recommended when working near the pulp and when the thickness of remaining dentin is not certain (Kawahara et al., 1979; ~1son and Prosser, 1982; Mount, 1988; Draheim, 1988; Muller et al., 1990; Caughman et al., 1990). Liners are recommended particularly when using glass ionomer cements as luting agents for indirect restorative materials since glass ionomer, when used as a luting agent, requires the material to be more viscous and, thus, more irritating. Still, it is thought that the high molecular weight of the polymer liquid, as well as other aspects of its composition (e.g., the use of weaker acids and less toxic monomers), help guard against permeation of the material through the dentinal tubules to the pulp (Klötzer, 1975; Dahl and Tronstad, 1976; Wilson, 1977; Tobias et al., 1978; Beagrie, 1979; Beagrie and Bránnström, 1979; Kawahara et al., 1979; Nordenvall et al., 1979; Wilson and Prosser, 1982; Mount, 1984; Van de Voorde et al., 1988). With a new, visible light-cured composition Kanaoka et al. (1991) did not find adverse responses in cell cultures.

A more severe pulp response has been reported with the powder-liquid ratios used for the luting cement (Hensten-Pettersen and Helgeland, 1977; Meryon et al., 1983). Both the proximity of the pulp and treatment of the bacterial layer covering the tooth will affect this response. Numerous in vitro cytotoxicity studies have shown that fresh-mined glass ionomer cements cause more damage than set cements; the longer the set before placing them in contact with cell cultures, the less the effect on cell cultures. Also, the more powder that is incorporated into the mix, the less toxic the mix will be to the cell cultures (Dahl and Tronstad, 1976; Hensten-Pettersen and Helgeland, 1977; Mjör et al., 1977; Tobias et al., 1978; Kawahara et al., 1979; Cooper, 1980; Meryon et al., 1983; Hume and Mount, 1988).

As with other materials, hydraulic pressure and etching during placement of the restoration may cause irritation of the pulp. Undue reactions in gingival tissue related to the use of glass ionomer cements, however, have not been reported from clinical practice. It is thought that the relatively good adhesion of this material accounts for its high biocompatibility. Leakage appears to be largely prevented and, thus, invasion of bacteria at the tooth-filling interface is minimized.

Leaching of component materials may be advantageous for glass ionomers. When glass ionomers are used as a luting agent or a restorative material, fluoride is released slowly, thereby inhibiting caries formation at the margins of and beneath restorations. The mechanism of action is not clear. DeSchepper et al. (1989a, 1989b) concluded that the effect might come as much from the hydrogen ion concentration as from the fluoride ion release. Levels of hydrogen and fluoride ion release are not constant. Hydrogen ion release is related primarily to the setting reaction of traditional formulations. Fluoride ion release is related to the degree of solubilization and diffusion of the glass particle components. The level of release decreases with time (Cooley and McCourt, 1991).

Early human clinical trials by Plant and Jones (1976) in Class I sites in premolars resulted in no sensitivity, but there was irritation in 5 percent of the pulps. In that study, clearly there was adequate remaining dentin thickness to provide a substantial barrier to any potential chemical insults. Nordenvall et al. (1979) compared glass ionomer to composite in contralateral tooth pairs (same tooth type at opposite sides of the mouth) Goldfoil and reported that, in cases in which pulpal inflammation was present, bacteria also were present in the restored site. Browne et al. (1983) reported a high correlation of pulpal inflammation with bacterial microleakage. They concluded that any potential chemical irritation was of only minor importance. Plant et al. (1988) evaluated a range of cementing media for inflammation and sensitivity. They detected at least some cases of bacterial microleakage for all

materials. Even though there was pulpal inflammation detected in 15 of 37 teeth after extraction, there was no sensitivity reported by any of the patients at any time. This is further evidence that sensitivity should not be used as a measure of biological activity. Osborne and Berry (1986, 1990) have been monitoring glass ionomer filling materials as Class III and V restorations for 3 years. There have been no reports of any sensitivity at any recall time. In a study designed to evaluate the effects of immediate finishing of glass ionomer restorations (Matis et al., 1988), there were no reports of sensitivity problems. Powell et al. (1990) examined 108

Class V abrasion/erosion lesions restored with glass ionomer filling materials. Sensitivity was examined in detail,

distinguishing hot and cold sensitivity as well as evaluating the effects of patient age and tooth site. Posterior teeth were more sensitive to cold. Younger patients showed more preoperative and postoperative sensitivity. Most teeth, but not all, became less sensitive by being restored Sensitivity appeared to be worse at cervical margins. All of the conclusions of this study correlate well with the hypothesis that the mechanism of fluid flow in dentinal tubules is the main cause of sensitivity.

Gold foil

Gold foil is a stable and relatively insoluble restorative material. In extremely rare circumstances (estimated at 1:1 million), patients sensitized to gold may react to gold restorations. These reactions include burning sensations of the oral mucous membrane in contact with the gold alloy, lichenoid lesions, and general systemic reactions (Pregert et al., 1979; Holland-Moritz et al., 1980; Castelain and Castelain, 1987).

The insertion of gold foil may result in pulpal reactions, but these are generally thought to be caused by the forces of condensation (Swerdlow and Stanley, 1962; Thomas et al., 1969; Stanley, 1984), thermal conductivity, cavity preparation, dehydration of the cavity, and micro leakage. Dowden and Langeland (1983) reported, however, that pulpal inflammation, destruction of odontoblasts, and hemorrhage were attributable to the toxicity of gold.

Casting alloys

Gold alloys and other alloys used in cast dental restorations and solders contain a number of elements, either intentionally added or as impurities. Allergic reactions have been described for many of these metals, including palladium (Phlelepeit and Legrum, 1986), nickel (Council on Dental Materials, 1982; Henstein-Pettersen et al., 1984; Femandez et al., 1986), chromium (Hildebrand, 1985), and cobalt (de Melo et al., 1983; Hildebrand, 1985).

Approximately 10 percent of women and 1 percent of men are sensitive to nickel (Merck Index, 1983). The extensive use of base metal casting alloys containing nickel for fixed restorations has been of major concem to the dental profession, but relatively few case reports substantiate this concern (Kalkwarf, 1984; Hensten-Pettersen, 1984; Femandez et al., 1986, Lamster et al., 1987). Allergy to gold-based restorations is reported more commonly than allergic reactions to nickel-containing dental alloys (Tomell, 1962; Elgart and Higdon, 1971; Schof et al., 1971; Young, 1974; Klaschka, 1975; Fenton and Jeffry, 1978; Fregert et al., 1979; Holland-Moritz et al., 1980; Izumi, 1982).

Palladium-based alloys have been reported as causative agents in cases of stomatitis (van Loon et al., 1984), oral lichenoid reactions (Downey, 1989), and disseminated urticaria (van Joost and Roesyanto-Mahadi, 15 90). Palladium allergy seems to occur in patients who are sensitive also to nickel (van Ketel and Niebber, 1981; Nakayama, 1982; van Loon et al., 1984, 1986; Stenman and Bergman, 1989; Augthun et al., 1990), but not consistently (Castelain and Castelain, 1987). Studies of T-lymphocyte levels in patients exposed to amalgam and nickel-containing alloys (Eggleston, 1984) and of the effect of fixed prosthodontic restorations made of silverpalladium alloys on serum immunoglobulins IgA, IgG, and IgM (Vitsentzos et al., 1988) are inconclusive. All casting alloys, except unalloyed titanium, seem to have a potential for eliciting adverse reactions in individual hypersensitive patients.

Chromium/cobalt alloys have an excellent history of biocompatibility, although there are some reports of tissue sensitivity in a very limited population (Merck Index, 1983). More extensive studies have been performed in patients before and after replacement of amalgam restorations with gold-based inlays. These studies found no significant effects on blood cells, erythrocyte components, electrolyte balance, liver function, inflammatory activity, immune stimulation, tissue damage, and kidney function (Molin, 1990). No evidence of toxicity or tissue reaction has been shown to alloys with a low gold content. Only limited data have been generated on the biological response to high-copper casting alloys.

Removable partial dentures made of base metal alloys have the potential of eliciting adverse reactions in patients allergic to cobalt, chromium, or nickel, but the incidence is uncertain. Patients with denture stomatitis related to the metal part of the prosthesis, and who have been patch-tested for contact allergy to nickel, cobalt, and chromium, often react to two, or all three, of the metals (Re, 1960; Brencllinger and Tarsitano, 1970; Levantine, 1974; Wood, 1974; Kaaber et al., 1979). Elevated cobalt and chromium levels have been observed in the saliva and tongue scrapings of patients with cobalt-chromium removable partial dentures (Stenman, 1982; de Melo et al., 1983), but the significance is unknown.

Toxic metals, such as beryllium and cadmium, also may be present in dental alloys, but no adverse effects have been reported in patients. Indium, the most common substitute for zinc, does not appear to have adverse biological effects (Merck Index, 1983). Likewise, there appear to be no adverse effects from alloys containing iron, molybdenum, manganese, or gallium. Titanium, the metal of choice for metallic implants, and alloys of titanium are biocompatible (Norman, 1991).

For the most part, metal ions, when placed on culture media, present an inhibited zone with various organisms (e.g., they show cytotoxicity or cell damage). These metals include chromium, cobalt, copper, mercury, nickel, tin, and zinc, all of which are used in dentistry. However, these metals are not found in dental restorations as metal ions, but as "eliminated structures." In addition, the alloying of these metals reduces their potential for ion production.

Ceramics

The relative incidence of biological side effects of dental ceramics compared with other restorative materials is considered to be low. In general, conventional dental ceramics are considered to be the most inert of all materials used for dental restorations. Ceramic restorative materials are not known to cause biological reactions, except for wear on the opposing dentition and/or restorations. No long-term data on the biocompatibility of these restorations are available (Roulet and Herder, 1991).

Additional Materials Used in the Restoration of Teeth

The fabrication of indirect cast restorations of alloys, fused and CAD/CAM prepared ceramic restorations, and in

direct composites involves many separate procedures that bring the oral tissues into contact temporarily with a wide variety of materials. These materials include impression materials, tissue retraction cord and astringents, and plastic or metal temporary restoration materials.

Other materials, such as luting agents (cements), last as long as the restoration itself. For all cemented restorations, pulp, dentin, and, to some degree, gingival reactions may be more dependent on the luting cement than on the material used to make the restoration. The biological response varies with the type of luting agent used and the methods of handling. Pulpal response to luting agents also may be related to hydraulic pressures produced during cementation. Most of these materials are subject to the same biocompatibility standards as the posterior restorative materials discussed above; however, the scope of the discussion in this report is limited to materials used in the long-term replacement of missing tooth structure.

Summary and Conclusion

Many of the biocompatibility considerations pertaining to dental restorative materials are sized in Table 1. All materials in current use are considered acceptable, in terms of their biocompatibility with local tissues, when properly handled and placed. Adverse systemic reactions are believed to be rare and self-limiting and tend to be of an allergenic nature. Local reactions have been documented in a small percentage of individuals, and systemic toxic reactions have been reported in the scientific literature.

* For example, class 1, 111, or V

** For example, mesial-occlusal-distal surfaces in a molar

Estimates of the costs of various restorative treatment scenarios are open to criticism, not only because they are based on "soft" data, but also because the types of treatment options used in the calculations are selected subjectively and will vary in clinical decision making. Several different models are used below to describe the long-term cost implications of providing restorative care over a 60 year period to an initially carious tooth Comparative fees derived from the ADA Survey of Dental Practice, 1990 (unpublished, preliminary data), are used. The relative costs of different scenarios for restoring posterior teeth involve comparative one-time costs (i.e., costs at the time of inserting a restoration) and cumulative, relative costs of selecting a particular clinical course over a 60-year period.

Figure 1 illustrates the costs over a 60 year period by extrapolating the values for amalgam, composite, and gold castings based on the longevity data given above.

Based on this extrapolation of costs, an approximate 2.5-fold increase in cumulative costs would be incurred by selecting composite instead of amalgam as a posterior restorative material. The cumulative costs of cast gold restorations would be about 50-percent greater than composite after 60 years.

Figure 2 is a revised model of anticipated costs that incorporates greater average longevity for dental amalgams (15 years) and composites (10 years) and, thus, fewer required replacements over a lifetime. Although adequate data do not yet exist for substantiating these expected restoration longevities for the future, early reports suggest that such longevities are achievable with improved dental restorative materials, application of more objective replacement criteria, and regular and consistent dental care. The impact of these changes on the model are substantial and suggest that the cumulative cost differential between amalgam and composite would be reduced to about 2:1.

If the relative fees shown in Figure 2 are used in a somewhat more clinically relevant model (Figure 3), the apparent price difference between initial amalgam and composite restorations would be increased beyond that depicted in Figure 2, because a comparatively expensive endodontic procedure will become necessary after three composite restoration replacements. Endodontics is included because of the potential clinical need to resolve expected pulp reactions and/or provide retention for restorations. With amalgam, an endodontic procedure would not be needed theoretically until age 75, the end point for the model. If the endodontic procedure and cast restoration were provided at this point, the relative cumulative costs between the composite and amalgam scenarios would be slightly less than 2:1. However, the endodontic procedure may not be accepted by the patient in this situation.

Figure 3 reflects the "countdown" theory suggested by Lutz et al. (1987) and reinforced by Simonsen (l991). This perspective proposes that teeth that are restored initially will undergo additional restoration over their lifetimes, involving more and more loss of tooth structure— and, eventually, possible root canal therapy and restoration by a crown.

Undoubtedly, much tooth loss will be caused by repeated restoration failures, which may result, in some cases, from a patient's inability to cover the cost of extensive and repeated treatment. In fact, cross-sectional data indicate that regular dental care users aged 35 to 44 lose an average of 4.5 teeth and irregular dental care users lose an average of 8.7 teeth during these 10 years (Kroeze, 1989).

Conclusions

Over the past century, the use of dental amalgam has provided substantial oral health benefits to the U.S. population. Indeed, the availability of amalgam during this period is perhaps the primary factor in the restoration to health and subsequent long-term retention of hundreds of millions of decayed teeth. Alternative materials are available and are being used increasingly in many situations where amalgam typically has been the material of choice. The combined effects of declining dental caries and use of alternative dental restorative materials have resulted in a dramatic 38-percent decline in the annual placement of dental amalgams by U.S. dentists between 1979 and 1990. There is reason to believe that this overall trend will continue.

All of the alternative materials to dental amalgam are more expensive than amalgam on a one-time basis as well as over the lifetime of an individual, and the general use of these materials instead of amalgam will result in markedly higher treatment costs.

The use of alternative restorative materials rather than dental amalgam to restore teeth in those seeking care would increase the annual national expenditures for dental services by more than $12 billion. The one-time direct costs for replacing all existing dental amalgams in the U.S. population would be enormous and impractical. Additional indirect costs would be substantial.

V. CONCLUSIONS AND RECOMMENDATIONS

The changing environment in which dentistry is being practiced will continue to have a dramatic impact on how dental amalgam is used and how its benefits are assessed. Declining dental caries rates in children and young adults indicate a need to reassess assumptions about the optimal approach to managing dental caries in the population. Historically, high rates of dental caries have led to a common view that caries attack was unavoidable and that, once a lesion was initiated, it would continue to increase in size if left untreated. The best long-term treatment was believed to be complete excision of carious tooth structure and adjacent sound tooth structure that might become carious in the future. The perception that there are "self-cleansing" areas of the teeth that do not predispose to carious attack was the rationale for extending cavity preparations beyond the extent of the carious lesion. This rationale was intuitive, however, and has proven to not be justified scientifically.

Effective preventive methods and the emergence of improved restorative materials permit a more conservative restorative approach and, generally, a wider spectrum of appropriate clinical choices than in the past. Although there is no single, ideal dental restorative material, certain materials offer advantages when used in specific clinical situations. For example, when minimal carious lesions occur in nonstress-bearing areas of posterior teeth, composite resins may be used as an alternative to dental amalgam, and they provide the advantage of preserving the maximum amount of sound tooth structure.

It is apparent that dentists will be treating patients with markedly varying oral health needs in the coming years. Some patients will present with rather low levels of dental caries that are not extensive in size. These patients will benefit from concerted prevention efforts and the use of smaller, nontraditional cavity preparations in posterior teeth, quite often employing newer dental restorative materials such as composite and glass ionomer cements.

Other patients will demonstrate higher levels and more extensive types of dental caries and/or many dental restorations that require replacement. Patients with extensive caries or in need of replacement restorations still will require aggressive preventive interventions but, generally, they will not be able to be managed as conservatively as patients with few, nonextensive caries. Restorations requiring replacement largely reflect the more destructive era of dental caries and the more extensive restorative approaches of the past. Once a large restoration has been placed, it cannot be replaced with a smaller one.

The qualitative value of a sound tooth versus a minimally restored tooth, a minimally restored tooth versus a moderately restored tooth, or a moderately restored tooth versus a totally rehabilitated tooth should not be overlooked. When dental caries are found in early stages or simply suspected, "wait and watch" is a rational alternative to definitive restoration, especially if patients can adopt more healthful practices and dentists can offer preventive interventions that may arrest early lesions.

The shift away from amalgam as the material of choice in many clinical situations has begun already and can be justified scientifically based on declining caries rates and the emergence of new and improved materials and methods. There continue to be, however, substantial oral health benefits that accrue to individuals and the population from the use of dental amalgam.

Based on a review of scientific evidence presented in this report, several broad recommendations can be made about the prevention and management of dental caries in the contemporary environment.

Recommendations for Research

The U.S. Public Health Service, as well as other Federal departments (the Department of Defense, the Department of Veteran Affairs), continues to sponsor and conduct research on dental amalgam and other restorative materials, and the National Institute of Dental Research Long-Range Research Plan for the Nineties points to areas of additional research interest in restorative materials. The following broad array of research recommendations was identified by the Ad Hoc Subcommittee on the Benefits of Dental Amalgam as important areas to pursue, based on a review of the relevant scientific literature conducted during development of this report.

REFERENCES