The use of dental amalgam has declined, but in most of the world, amalgam is the most widely used and widely taught direct restorative material for load-bearing posterior restorations. There are few national regulations on the use of amalgam; however, there are several nations where few amalgam restorations are placed. Long-term studies have shown that under optimum conditions, posterior restorations of amalgam and resin composite last longer than reported previously and that amalgam restorations outlast composite restorations. In general practice settings, posterior amalgam and composite restorations both have lower longevities.
This article is a review of the literature on posterior amalgam restorations published during the period between 1996 and 2006. During this period, research interest on amalgam significantly declined. A Medline search of articles with “amalgam” in the title, “dental” anywhere, and the subject “dentistry” yielded 1054 citations (1.4% of all dental citations) between 1986 and 1995 but only 553 citations (0.81% of all dental citations) between 1996 and 2005. During the latter period, there were only two comprehensive reviews of the literature on dental amalgam, and both appeared early in the period . Several articles referred to amalgam in the context of reviewing the advantages and disadvantages of alternative restorative materials, however .
Because there have been many recent reviews of amalgam biocompatibility and the effects amalgam waste on the environment , this article focuses solely on amalgam restorations. Similarly, because recent reviews have focused on dental amalgam in primary teeth , the focus of this article is on amalgam in permanent teeth. Because of space limitations, an update on the metallurgical, physical, and mechanical properties of dental amalgam must await another venue.
Current usage
In 2004, Burke reviewed trends in amalgam and composite usage around the world. The following discussion summarizes and updates Burke’s excellent review.
North American dentists
Several reports suggested that the overall use of dental amalgam in the United States has declined significantly during the last decade . In one state, the number of resin composite restorations exceeded the number of amalgam restorations in 1999 . Amalgam continues to be the most widely used direct restorative material for posterior load-bearing restorations, however. In 1999, US dentists placed 71 million amalgam restorations compared with 46 million posterior resin composite restorations . The number of posterior composites was up from 13 million in 1990; the number of amalgam restorations was down from 99 million placed in that year . From 1990 to 1999, amalgam restorations declined from 88.4% to 60.6% of the sum of amalgam and posterior composite restorations.
North American dental schools
The best judgment of dental educators may be of interest. In a 1997 survey, 53 of 54 North American dental schools responding reported that they taught the use of resin composite to restore posterior teeth . Thirty-seven percent of the schools devoted less than 5% of operative dentistry curriculum time to teaching class I and II composite restorations; 85% of the schools reported that they spent less than 20% of available curriculum time on these restorations. Only 30% of the surveyed schools taught three-surface class II posterior composites in molars. This study did not explicitly ask about the percentage of the operative curriculum devoted to teaching amalgam restoration. It is plausible that increased curriculum time for posterior composite restorations is an indicator of increased probability that composite will be selected over amalgam.
The trend was—and continues to be—toward greater emphasis on resin composite for posterior restorations. For example, in a 2005 survey, 68% of 47 US dental schools reported that they used resin composite for three-surface class II restorations . This study also found that in 80% of US schools amalgam was taught first and that amalgam was used in 60% of the posterior restorations placed by students. A recent survey suggested that Canadian dental schools have a similar philosophy for direct posterior restorations . Amalgam continues to be favored among Canadian educators: in all schools responding, amalgam and resin composite posterior restorations were taught, with either equal or greater emphasis being placed on amalgam .
European dentists
The use of amalgam in the United Kingdom is similar to that in the United States. In a 2001 survey of 654 British general dentists, 35% reported that they “sometimes” used resin composites in extensive load-bearing restorations in molar teeth . Fifteen percent responded “often,” and 1% responded “always” to this question. Presumably amalgam was used when resin composite was not. In a smaller survey, 30 UK dentists reported that 87% of class II and 67% of class I restorations were amalgam .
Amalgam is used less frequently in some Scandinavian countries. In 2002, Ylinen and Lofroth reported that only 28% of Finnish dentists and 40% of Swedish dentists used amalgam. In the two other Scandinavian countries, however, amalgam was used by most dentists (88% of Danish dentists and 92% of Norwegian dentists). Use of amalgam is particularly low in Finland, where a 2000 survey returned by 548 dentists reported all the restorations they placed in a single working day. Amalgam accounted for only 8% of the class I restorations (resin composite: 80%) and 9% of the class II restorations (composite: 80%). When asked what material they would use to restore an occlusal lesion in the lower second molar in a 20-year-old patient, amalgam was the choice of 52.4% of 173 Danish dentists, 19.9% of 759 Norwegian dentists, and 2.9% of 923 Swedish dentists . A 2005 report commissioned by the Swedish government found that amalgam fillings were no longer used in children and young people and that by weight amalgam made up only 6% of all Swedish fillings .
European dental schools
Responding to a 1997 survey, 100 of 104 (96%) European dental schools reported that they taught resin composites for class I restorations . Seventy-nine percent of European schools taught three-surface class II posterior composites restorations; however, 56% of these schools devoted no more than 20% of the curriculum time for direct restorations to posterior composites. Only 38% of the surveyed schools taught three-surface class II posterior composites in molars. Overall, the European schools were similar to the North American schools in that amalgam was still taught for class I and II restorations, and at most schools, most of the curriculum time was spent on amalgam. A 2006 survey of dental schools in the United Kingdom suggested that the teaching of resin composite for posterior restoration continues to increase . In this study, 9 of 15 schools (60%) reported that they taught three-surface class II resin composites.
The general trend is that amalgam continues to be taught in European dental schools. One dental school in the Netherlands has gradually reduced the amount of curriculum time devoted to dental amalgam as a restorative material . In 2001, it stopped teaching amalgam altogether.
Dentists and dental schools in the rest of the world
Cross-sectional surveys of Australian dentists revealed that between 1984 and 1999, the use of amalgam gradually declined from 57.8% to 23.3% of all restorative services rendered . In a 2002 survey of 560 randomly selected Australian dentists, 32% reported that they “sometimes” used resin composites in extensive load-bearing restorations in molar teeth ; 29% responded “often” and 12% responded “always” to this question. The former two categories revealed greater use of resin composite in Australia (41% “often” or “always”) than in a similarly designed study conducted in the United Kingdom (16% “often” or “always”) . These data suggest a greater move away from amalgam in Australia than has been seen in Europe or the United States.
An even larger move away from amalgam has taken place in Japan. Unfortunately, there are only two reports of this in the English language literature, neither of which provides data . Both articles report that amalgam is little used in general practice, which may be because of fear of mercury that gripped the Japanese public in the aftermath of the poisoning of inhabitants of Minamata and Niigata in the mid-1950s . Victims had consumed methyl mercury–contaminated fish. Given the abandonment of amalgam, it is interesting that most Japanese dental schools do not view resin composite a suitable material in extensive class II restorations. In a 1997 survey, 25 of 27 Japanese dental schools taught resin composite for class I restorations, but less than 19% of the schools considered resin composite a suitable restorative material for three-surface class II restorations .
Data from the rest of the world are spotty. In some countries, dental amalgam may still be the major restorative material for load-bearing restorations. For example, a 1997 survey of 241 Jordanian dentists showed that dental amalgam was used for 88.8% of all class I and class II restorations . In other countries, the trend is more like that seen in North American and Europe. For example, a 1999 survey revealed that 97% of 65 Brazilian schools surveyed considered resin composite suitable for class I restorations. Like faculty representing northern hemisphere dental schools, only 33% of Brazilian respondents considered resin composite suitable for three-surface class II restorations .
Regulation of amalgam use by governments
During the 1990s, anti-amalgam newsletters and Web sites reported that dental amalgam had been banned in Europe, especially in Germany and Sweden. Wahl discussed and refuted these rumors. Similarly, after surveying regulatory agencies in ten countries, Burke concluded that there “were few restrictions worldwide to the placement of dental amalgam.” The tightest current restrictions seem to be in Denmark, where amalgam use is limited to molar teeth . Sweden, Norway, Austria, and Germany recommend that amalgam restorations not be placed in pregnant women . Germany also recommends that amalgam not be placed in patients with renal impairment . Most of the other nations surveyed, including the United States, United Kingdom, Australia, Finland, and Ireland, have issued no recommendations for restrictions on amalgam use.
Although Sweden does not currently regulate amalgam, its national health system has not reimbursed dentists for amalgam restorations since 1999 . This decision has greatly reduced use of amalgam. Sweden also has announced that its overall goal is to phase out use of mercury, including dental amalgam . A 2004 report commissioned by the Swedish government confirmed this goal for mercury in general but recommended that dental amalgam be exempted from the general ban until December 31, 2008 . A 2005 report commissioned by the Swedish government concluded that a phase out of dental amalgam restorations will not have a significant effect because amalgam is already used infrequently .
Regulation of amalgam use by governments
During the 1990s, anti-amalgam newsletters and Web sites reported that dental amalgam had been banned in Europe, especially in Germany and Sweden. Wahl discussed and refuted these rumors. Similarly, after surveying regulatory agencies in ten countries, Burke concluded that there “were few restrictions worldwide to the placement of dental amalgam.” The tightest current restrictions seem to be in Denmark, where amalgam use is limited to molar teeth . Sweden, Norway, Austria, and Germany recommend that amalgam restorations not be placed in pregnant women . Germany also recommends that amalgam not be placed in patients with renal impairment . Most of the other nations surveyed, including the United States, United Kingdom, Australia, Finland, and Ireland, have issued no recommendations for restrictions on amalgam use.
Although Sweden does not currently regulate amalgam, its national health system has not reimbursed dentists for amalgam restorations since 1999 . This decision has greatly reduced use of amalgam. Sweden also has announced that its overall goal is to phase out use of mercury, including dental amalgam . A 2004 report commissioned by the Swedish government confirmed this goal for mercury in general but recommended that dental amalgam be exempted from the general ban until December 31, 2008 . A 2005 report commissioned by the Swedish government concluded that a phase out of dental amalgam restorations will not have a significant effect because amalgam is already used infrequently .
Longevity of amalgam restorations
When Mjor and colleagues reviewed the longevity of posterior restorations in 1990, it was evident that median survival times of amalgam restorations in posterior teeth varied greatly among studies. Sixteen years later, restoration longevity data can appear just as chaotic. For example, in 2004, Manhard and colleagues reviewed clinical studies of various restorative materials placed in posterior teeth, including 41 studies of amalgams and 50 studies of resin composites (see also their earlier reviews ). They found that the ranges of annual failure rates were wide: 0 to 7.4% for amalgams and 0 to 9.0% for composite. From these studies they calculated mean failure annual rates of 3% (standard deviation, 1.9%) for amalgams and 2.2% (standard deviation, 2%) for posterior composites . This does not mean that composites fared better than amalgams; the two failure rates are not statistically different. One might erroneously conclude, however, that posterior composite restorations would be at least as successful in posterior restorations as amalgam.
Manhard and colleagues concede that it is “problematic to directly compare different studies from different authors,” but they are not explicit about some of the pitfalls of combining data from different studies . For example, as Mackert and Wahl noted, many of the cited studies are relatively short-term (≤5 years). Such studies are biased because they exclude failure modes that occur more frequently later in a restoration’s life (marginal degradation, secondary caries, bulk fracture, and tooth fracture) . Manhard and Hickel’s mean annual failure rates combine data from two different types of studies: (1) controlled longitudinal clinical trials, in which restorations are placed and maintained under conditions that are favorable to longevity and (2) uncontrolled studies in general practice, in which restorations have been placed and maintained under conditions less favorable to longevity. The former shows whether a restorative material has the potential to be used successfully and the latter shows whether that potential is actually being achieved . To meaningfully compare the longevity of posterior amalgam and composite restorations, one must be sure that the restorations to be compared have been studied under similar conditions.
Longitudinal studies
The best way to estimate the longevity of restorations is to conduct longitudinal trials . Unfortunately, longitudinal studies are expensive, require long-term commitment of personnel and other resources, and may be plagued by loss of patients . As a result, few studies of dental restorative materials have continued long enough to obtain long-term data. Short-term studies may underreport types of failure (eg, secondary caries and fatigue fracture) that are likely to become more important after many years in vivo. When new failure mechanisms become operative late in a restoration’s life, short-term studies overestimate restoration longevity .
In the following sections and in the accompanying tables, longitudinal studies in which restorations have been followed for at least 8 years are emphasized. To help the reader compare results, failure rates have been extrapolated to median survival times. When median survival times have been determined from life tables, it is noted in the tables. It should be cautioned, however, that extrapolated data, even from long-term studies, assume that past performance will predict future behavior. The future is not certain: the failure rate may speed up as new failure mechanisms become operative as time progresses, or conversely, the failure rate may slow as early failures eliminate the restorations most at risk of failure from the study population.
Longitudinal studies in optimum setting
Studies conducted in these settings, typically dental schools, tend to show a material’s durability under optimum conditions . Patients are often dentally aware. They are often dental students, dental school staff, or conscientious patients who are judged especially likely to return for recall appointments. Typically, operator variability is reduced by using only a few (usually less than six) dentists. These dentists are often teaching staff who are well calibrated and likely to adhere closely to study protocols. Importantly, these dentists seldom function under tight time constraints like those in private practice.
Several 5- to 8-year longitudinal studies of posterior amalgam restorations appeared during the 1980s. The results of these studies suggested that in optimum settings dental amalgam restorations might last much longer than previously thought. For example, amalgam restorations in a set of studies reviewed by Letzel and colleagues had median survival times of 11.4 to 87.5 years for low-copper amalgams and between 19.2 and more than 150 years for high-copper amalgam restorations ( Table 1 ) . During the last 15 years, longitudinal studies of even longer duration have appeared. In longitudinal prospective trials, class I and II amalgam restorations were found to have median survival times of 57.5 years , 65.8 years , and 69 years .
| Authors | Year | Study type a | Study setting b | Study duration (y) | No. of dentists | No. of restorations | Median survival time (y) | Survival estimate method c |
|---|---|---|---|---|---|---|---|---|
| Studies of class I and II amalgam restorations | ||||||||
| Osborne & Norman | 1990 | P | + | 13 | 1 | 181 | 57.5 | A |
| Letzel et al | 1997 | P | + | 5–15 | 7 | 3244 | 65.8 | E |
| Collins et al | 1998 | P | ++ | 8 | 1 | 53 | 69.0 | A |
| Dawson & Smales d | 1992 | R | ++++ | 0–17 | many | 1345 | 14.4 | B |
| Lucarotti et al d | 2005 | R | ++++ | 0–12 | many | 76,418 | 11.9 | E |
| Bjertness & Sonyu d | 1990 | R | ++ | 0–17 | 4 | 782 | 44.7 | F |
| Hawthorne & Smales d | 1997 | R | +++ | mean 25 | 20 | 1728 | 22.5 | B |
| Smales d | 1991 | P | ++ | 8–10 | many | 1476 | 62.5 | F |
| Class II restorations only | ||||||||
| Gruythuysen et al | 1996 | P | + | 15 | 3 | 1213 | 44.1 | A |
| Jokstad & Mjor | 1991 | P | ++ | 9.5 | 7 | 469 | 25.0 | E |
| Smales | 1991 | P | ++ | 15 | many | 664 | 27.2 | F |
| Lucarotti et al —distal-occlusal & mesial-occlusal restorations | 2005 | R | ++++ | 0–12 | many | 16,680 | 9.8 | E |
| Lucarotti et al —mesial-occlusal-distal restorations | 2005 | R | ++++ | 0–12 | many | 147,087 | 8.8 | E |
a P, prospective; R, retrospective.
b +, controlled; ++, closer to controlled; +++, closer to general practice; ++++, general practice.
c A, Survival time extrapolated from percentage of restorations surviving at end of study. B, Survival time is taken directly from a life table. C, Survival time is extrapolated from a life table. D, Survival time is taken directly from survival plots calculated by the Kaplan-Meier method. E, Survival time is extrapolated from survival plots calculated by the Kaplan-Meier method. F, Survival time is extrapolated from actuarial life tables.
d Some classes III and V but predominantly classes I and II.
Longitudinal studies in general practice settings
Two relatively recent longitudinal studies have shown that amalgam restorations do not survive as long in general practice settings as in clinical trials. The first study was a retrospective longitudinal analysis of all types of amalgam restorations placed in Australian Air Force clinic patients between 1972 and 1988 . They found a median survival time of just 14.4 years. The sites in which the restorations were placed were not reported. It is presumed that most of the restorations were class I or II. Class III and V restorations included in this study would most likely have increased survival time.
The second report of the survival of amalgam restorations placed in general practices appeared recently. In a retrospective longitudinal study, Lucarotti and colleagues used insurance payment records to follow a large number of restorations placed by the General Dental Service of the United Kingdom between 1990 and 2001. The median survival time of single surface amalgam restorations (presumably mostly class I restorations) was 11.9 years.
Longer survival times are sometimes reported for studies conducted in what seems to be general practice populations. When case selection is scrutinized, however, one concludes that the data are not typical of general practices. For example, Bjertness and Sonju conducted a retrospective longitudinal study of records from the general practices of six Norwegian dentists. This 17-year study yielded a 44.7-year median survival time for amalgams of unknown composition. The survival time may have been increased by the use of a study population that was limited to patients who returned annually for examination. Such conscientiousness suggests that the selected patients have a high dental awareness. Patient oral hygiene may have been better than is typical in general practice populations. That four of the dentists worked part time at a dental school also may have increased the durability.
As was the case in Bjertness and Sonju’s study, in their retrospective study of restoration longevity in three Australian practices, Hawthorne and Smales selected patients who had “a continuous attendance history.” They found a median survival time of 22.5 years. The selection of highly conscientious patients may have increased survival time. On the other hand, more than 64% of the restorations were class II amalgams. The predominance of class II amalgam in the sample may explain why the median survival time was less than that found by Bjertness and Sonju. Note, however, that Bjertness and Sonju did not report the distribution of restorations by class, so one does not know for sure that one study is more class II–rich than the other.
Smales reported on the 10-year durability of a set of amalgam restorations placed in an Australian dental school clinic by dental students and staff. The study setting was neither a general practice nor a well-controlled clinical trial. The median survival time of the amalgams restorations in this setting was 62.5 years. This long durability suggests that the dental school setting may be closer to the optimum setting of a controlled clinical trial than it is to a general practice setting.
Longitudinal studies of class II restorations
Under optimum conditions, class I and II amalgam restorations are found to have median survival times between 57 and 70 years. As might be expected, similar trials of just class II restorations yield shorter survival times. In one such study, the median survival time was 44.1 years ( Table 1 ) . In another study, median survival time was 25 years . The survival time of the latter may have been reduced by two of the six operators who placed their restorations in general practice settings. A study under slightly less than optimum conditions gave similar results. In a study conducted in an Australian dental hospital, where restorations were placed by a large number of student and staff dentists, Smales found a medium survival time of 27.6 years. These survival times are longer than found general practice settings, however. In a large retrospective longitudinal study by Lucarotti and colleagues , class II amalgams placed in general practices in the United Kingdom were found to have median survival times of 9.8 years for distal-occlusal and mesial-occlusal restorations and 8.8 years for mesial-distal-occlusal restorations.
Longitudinal studies of extensive posterior restorations
Restorations in which one or more cusp has been restored with amalgam exhibit even shorter survival times. Table 2 provides some details of longitudinal studies of such restorations. Three prospective longitudinal studies were conducted in optimum settings; two are in good agreement. In one study, the median survival time was found to be 14.9 years . In a second study, the median survival time was found to be 12.5 years for molars with all cusps covered and 14.5 years for molars with only partial cusp coverage . In a third study, however, a longer median survival time was found: 27.4 years for amalgams with a least one cusp covered with amalgam . In this last study, the investigator also reported that the survival time for complex amalgams was not significantly different than class II amalgam restorations in the same patient pool . This observation suggested that the extensive amalgams may have included fewer cusps than other studies.
