In a 1988 retrospective study of endodontically treated teeth, Hansen⁷ found that 181 premolars with mesial-occlusodistal, or MOD, amalgam restorations had a significantly higher incidence of cusp fracture than did the 40 premolars with MOD resin-based composite restorations. However, he stated that the results of the study "should be cautiously interpreted, especially since the number of resin-restored teeth was rather small" and also because restorations for endodontically treated posterior teeth ideally should have cuspal coverage, unlike the intracoronal restorations in his study. In two subsequent and much larger studies of endodontically treated posterior teeth, Hansen and colleagues⁸ and Hansen and Asmussen⁹ found that the percentage of amalgam restorations without cuspal fracture after 20 years (66 percent) was about the same as the percentage of composite restorations without cusp fracture after only 10 years (72 percent). Long-term studies not limited to endodontically treated teeth have shown relatively low rates of cusp fracture in amalgam-restored teeth. In one such study of about 600 teeth restored with amalgam, only 1.5 percent had a cusp fracture after five years.¹⁰ Another study, larger and longer-term, showed that only 1.8 percent of 1,415 teeth with Class II amalgam restorations had cusp fractures after 10 years.¹¹ After 15 years, in 1,213 teeth available to be examined, 3.8 percent of the restorations had failed because of enamel fracture alone, and 5.0 percent had failed because of enamel fracture only or because of a combination of enamel fracture, isthmus fracture and/or caries.¹²

In every general dental practice, dentists observe complete cusp fractures in teeth restored with both resin-based composite and amalgam. They also observe teeth restored with either of these materials that do not have complete cusp fractures. We attempted to determine the actual prevalence of cusp fractures in amalgam-restored teeth versus that in resin-based composite–restored teeth in a general dental practice.

	SUBJECTS AND METHODS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
From September 2001 until February 2002, one or the other of two general dentists (M.J.W. and M.M.S.) examined 1,902 consecutively seen adult patients (18 years of age and older). They examined posterior teeth only with either amalgam or resin-based composite restorations and with at least one virgin cusp present and/or one cusp missing and unrestored (cusp fracture). For these teeth, they recorded the patient’s name, age, sex, type of restoration (amalgam or resin-based composite), number of restored surfaces and the presence or absence of a cusp fracture. They defined "cusp" as a virgin cusp and "cusp fracture" as a completely absent and unrestored cusp. If there was a cusp fracture, they recorded whether caries was associated with it (since the caries itself could have caused the cusp to fracture). The dentists included no teeth that had both types of restorations and no teeth that had other types of restorations, even if they also had amalgam or resin-based composite restorations. The dentists obtained data for 10,869 restorations, of which 10,082 were amalgam and 787 were resin-based composite.

	RESULTS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Table 1 shows the number of subjects and the number of teeth with amalgam and composite restorations along with the percentage of teeth with cusp fractures (including both teeth with and without caries) for each of five age groups. To evaluate statistical significance, we used the ² test with the Yates correction. Data for male and female subjects did not differ significantly and have been pooled in the table. For statistical analyses, we pooled the data for subjects aged 18 through 54 years, as well as data for subjects aged 55 through 96 years.

Table 2 shows the frequency of cusp fractures in younger (younger than 55 years) and older (55 years or older) subjects with several types of restorations. Comparison of the percentages for teeth without caries (the first two rows of data) suggests a lower percentage of cusp fractures in teeth with a single restored surface than in teeth with multisurface restorations; we found this difference to be significant only in younger subjects with amalgam-restored teeth (P < .001).

Finally, the data suggest that for older subjects only, fracture rates may have been higher in composite-restored teeth than in amalgam-restored teeth. This difference was significant only when we pooled the data for teeth with one and with two or more restored surfaces (P < .05).

When we examined the data for teeth both with and without caries (the third and fourth rows of data in Table 2) and compared them with the data for the teeth without caries, we found that the same effects were evident and the same contrasts were statistically significant.

	DISCUSSION

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Our results indicate that cusp fracture rates are influenced by age and by the number of surfaces of a tooth that have been restored. The findings provide no evidence that fracture rates are higher in amalgam-restored teeth than in composite-restored teeth.

The number of restorations per subject was essentially uniform in those whose age ranged from 25 through 69 years and then decreased in older subjects (Table 1). Elderly subjects are more likely to have had extractions or crowns, and this probably explains the decreased number of restorations per elderly subject.

The percentage of restored teeth with cusp fractures increased after the age of 55 years (Table 1). Apparently before that age, the incidence of new fractures roughly matched the rate at which fractured teeth were repaired or lost. Older subjects had larger restorations (Table 2) and probably more brittle teeth, both of which may have contributed to their higher cusp fracture rates.

Our study had several weaknesses. One of these was the fact that we did not know the age of the restorations and, therefore, could not determine the likelihood that teeth would suffer a cusp fracture during a defined period of restoration. Instead, we studied the prevalence of cusp fractures, in effect recording the number of cusp fractures that had not yet been repaired in teeth that had not yet been lost. Clearly, such data provide a less-than-perfect basis for estimating the relative importance of the several factors that cause cusp fractures.

The percentage of restored teeth with cusp fractures increased after the age of 55 years.

In addition, we did not record types (molars or premolars, maxillary or mandibular) or vitality of teeth. Endodontically treated teeth are more likely to undergo cusp fracture than are vital teeth. Although we recorded the restoration class, which may be somewhat related to restoration size (with Class II restorations generally being larger than Class I restorations), we did not measure restorations’ width or depth. As with most retrospective studies, there was no way to determine some preparation features (such as rounded versus sharp line angles). Smaller and shallower restorations are more resistant to cusp fracture than are wider and deeper restorations.^13,14 Similarly, preparations with sharp line and point angles are more likely to be associated with cusp fracture than are those with rounded line and point angles. All of these factors could have influenced the results of the study.

Both resin-based composite and amalgam materials and technology have undergone dramatic changes within the last decade. Dentin bonding now is used routinely for both resin-based composite and amalgam restorations, and it allows for smaller preparation designs, which make teeth more resistant to fracture.^15–19 If some of the restorations in our study were placed more than a decade ago, it is probable that the composite restorations were not dentin-bonded, and it is probable that the amalgam restorations were not bonded at all. This could have affected the likelihood of cusp fracture. Tooth preparation techniques for both resin-based composite and amalgam²⁰ have changed from sharp to rounded line angles, which also may help prevent cusp fractures.

Another weakness of our study was the relatively small number of composite-restored teeth (787) that we examined, as compared with amalgam-restored teeth (10,082). Our sample probably is typical in this respect. Although more resin-based composite restorations are being placed now than in the past, there still are far more amalgam restorations than posterior resin-based composite restorations in use, and there probably are far more existing amalgam restorations in patients than there are posterior resin-based composite restorations.

In both amalgam- and composite-restored teeth, the incidence of cusp fracture was greater in teeth with two or more restored surfaces than on teeth with a single restored surface (Table 2). This is as expected, because restorations on several surfaces usually will weaken a tooth more than will a restoration on a single surface.

Older subjects had higher fracture rates in teeth restored with composite than in those restored with amalgam (Table 2). However, for several reasons, this finding needs to be replicated before it is accepted. First, the total number of cusp fractures in composite-restored teeth in older subjects was only eight, which is too few to provide a reliable basis for any conclusion even if the result is mathematically statistically significant. Second, our statistical method of testing selected contrasts without an overall test for significance can yield "false positives." Third, older subjects had a smaller proportion of resin-based composite restorations than did younger subjects, which indicates they may have had a different history of dental restoration decisions. For these reasons, we recommend that little significance be attached to the elevated cusp fracture rates observed in composite-restored teeth in our older subjects.

Our study specifically compared resin-based composite restorations with amalgam restorations, and the results indicate that amalgam restorations do not lead to elevated rates of cusp fracture. Additionally, there are several aforementioned long-term prospective studies showing that amalgam-restored teeth do not exhibit high rates of cusp fracture.

Why, then, do so many dentists believe that teeth restored with amalgam are subject to high rates of cusp fracture but that teeth restored with resin-based composites are not? Consider first this statement by DiTolla⁴: "I thought back over my first six years in practice and realized that 99 percent of the teeth that required crowns all had silver fillings." Next, recognize that our experience in the present study is similar; we saw 190 cusp fractures in amalgam-restored teeth but only 18 fractures in composite-restored teeth. But in the same period, the number of teeth without cusp fracture was 9,892 among those restored with amalgam versus only 769 among those restored with resin-based composite. So among the teeth that had no cusp fracture, there were more than 12 times the number of amalgam-restored teeth as composite-restored teeth.

Among the teeth that had no cusp fracture, there were more than 12 times the number of amalgam-restored teeth as composite-restored teeth.

As a rule, dentists probably pay less attention to restorations that are functioning well, concentrating on restorations that are failing and the extensive procedures (such as crown placement) that such failure necessitates. Practicing dentists spend only about 10 percent of their time on diagnostic services, but about 90 percent of their time on other services.²¹ As a result, the perception that a huge number of amalgam-restored teeth break may be different from the reality that only a small percentage of amalgam-restored teeth break. Among the teeth with cusp fracture in our subjects, there were 10 times as many with amalgam restorations as there were with resin-based composite restorations; however, among the restorations that were functioning well, there were 12 times as many amalgam restorations as resin-based composite restorations.

Psychologists call this type of error "underuse of base rates," and it has been known for some time.²² The general idea is that people often draw conclusions on the basis of their experience (for example, gray cars are involved in more accidents than are yellow cars) without appropriately factoring in the important role of unequal base rates (in this example, the fact that there are more gray than yellow cars on the road). Similarly, any practicing dentist would have the experience of seeing many more cusp fractures in amalgam-restored teeth than in composite-restored teeth, which might lead to an incorrect conclusion unless a careful and numeric appraisal of the base rates of resin-based composite and amalgam restorations was carried out. Obviously, although the "underuse of base rates" process may be involved, our study does not directly test this hypothesis. All it does is provide data about the actual relative rates of cusp fractures in teeth restored with amalgam and with resin-based composite.

The use of amalgam restorations remains somewhat controversial.^23,24 Some dentists have stated that the death of amalgam is "imminent"^25,26 and that posterior resin-based composite restorations are now placed more frequently than amalgam restorations,²⁷ but neither of these assertions is true. The number of posterior resin-based composite restorations has increased dramatically in recent years, but does not yet approach the number of amalgam restorations placed in the United States. In 1999, there were 71 million amalgam restorations placed by U.S. dentists in private practice, but only 46 million posterior resin-based composite restorations.²⁸ Sales of amalgam were higher (although only slightly) in 2001 than in 2000.²⁹

	CONCLUSION

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Among dentists, the belief is widespread that amalgam restorations result in more cusp fractures than do resin-based composite restorations. In our study of 10,082 teeth restored with amalgam and 787 teeth restored with resin-based composite, the prevalence of cusp fracture was 1.88 and 2.29 percent, respectively, showing essentially no difference in cusp fracture rate. Because the majority of restorations presently in use in posterior teeth are amalgam, dentists see more cusp fractures in teeth restored with amalgam than in composite-restored teeth; this may account for the persisting belief in the destructive effects of amalgam restorations. Our results suggest that dentists can be confident that neither amalgam nor resin-based composite restorations are associated with a high rate of cusp fracture.