While the amount of literature in this area is large, the extent of the uncertainty is demonstrated by just a few key references. In 1979, the National Institutes of Health held a consensus conference on the removal of third molars.¹⁰ At that conference, the literature and evidence up to that time were reviewed, and expert testimony was received. On that basis, the committee members agreed that third molars with nonrestorable carious lesions, recurrent infection, cysts and tumors, as well as those contributing to the resorption of adjacent teeth and destruction of bone, are candidates for removal. Removal of third molars to prevent crowding of anterior teeth was not supported. There was, however, no consensus on removal of impacted teeth when there is no evidence of pathology. It was agreed that even though early removal might reduce morbidity, further studies were needed and that the economic implications also needed more study.

In 1996, Daley²⁰ published an extensive article on third-molar extraction based on a comprehensive review of the literature. Daley stated that "... third molars without associated pathological or developmental conditions are sacrificed, usually in adolescents and young adults, like no other human tissue, in the name of preventive dental care." He went on to conclude that "in more than 98 percent of cases, there is no apparent benefit to prophylactic third-molar extraction in adolescents. The concept that all third molars (functional or nonfunctional) should be extracted prophylactically should be abandoned."

The ADA’s 1997 Dental Practice Parameters⁷ includes a section on impacted and unerupted teeth. It provides no specific criteria, nor does it directly address the third-molar controversy. It provides only the rather ambiguous statement that "after consideration of the individual circumstances, the dentist should decide whether the impacted/unerupted tooth should be monitored or treated."

A genuine uncertainty exists as to the most appropriate course of action for pathology-free impacted third molars.

Guidelines published by the American Academy of Pediatric Dentistry are similarly nonspecific.⁸ In the guidelines’ 1998–1999 revision, it is recommended that "third molars which are determined to be potential or active problems should be considered for treatment by the appropriately trained dentist. Diagnostic criteria for extraction should be those currently accepted by the dental profession."

A 1998 article on impacted teeth published by the American Academy of Oral and Maxillofacial Surgeons makes it clear that in the Academy’s view, "All impacted teeth are pathologic processes ... ," and the article claims that "... scientific evidence [shows that] impacted teeth are a pathological entity and surgical management is the treatment of choice."³

A markedly different conclusion, also claiming to be based on the current state of the science, was arrived at in 1999 by the National Institute for Clinical Excellence, or NICE, in the United Kingdom.^30,31 NICE conducted a review to provide guidance to the United Kingdom’s National Health Service, or NHS, on the removal of third molars. The stated aim of the review was "to provide a summary of existing evidence on prophylactic removal of impacted wisdom teeth, in terms of the incidence of surgical complications associated with prophylactic removal, and the morbidity associated with retention." This report concluded that "there is no reliable research evidence to support the prophylactic removal of pathology-free impacted third molars." Based on this report, the NHS has advised dentists in England and Wales that "the routine practice of prophylactic removal of pathology-free impacted third molars should be discontinued in the NHS." The report also states that "surgical removal of impacted third molars should be limited to patients with evidence of pathology. Such pathology includes unrestorable caries, nontreatable pulpal and/or periapical pathology, cellulitis, abcess and osteomyelitis, internal/external resorption of the tooth or adjacent teeth, fracture of the tooth, disease of follicle including cyst/tumor, tooth/teeth impeding surgery or reconstructive jaw surgery, and when a tooth is involved in or within the field of tumor resection." The report further states that plaque is not an indication for surgery, and that "the evidence suggests that a first episode of pericoronitis, unless particularly severe, should not be considered an indication for surgery."

These examples highlight the fact that genuine uncertainty exists as to the most appropriate course of action for pathology-free impacted third molars. We conducted this study to see whether this uncertainty in the dental literature is reflected in clinical practice. While acknowledging that this debate continues, we do not attempt to suggest what the best approach is, but rather attempt to describe current practice in more detail than has been possible previously.

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION LIMITATIONS CONCLUSIONS REFERENCES

 
The source of treatment data for this article is insurance claims filed with Delta Dental Plan of Michigan, or DDPM, and its subsidiaries for treatment provided from July 1991 through December 1999. During this period, DDPM processed claims for approximately 100 million procedures provided to about 7.4 million patients. Claims were received from all 50 states. More than 3.8 million of these patients were Michigan residents, nearly 1 million were from Ohio, more than 500,000 were from Indiana, and more than 2 million were from states other than Michigan, Ohio or Indiana. The covered people were employees or retirees and their dependents, and included a wide range of employees in public and private employment.

For the data included in this analysis, the treating dentists received fee-for-service payment, and payment levels most commonly were 100 percent for diagnostic and preventive services and 80 to 90 percent for most restorative and oral surgical services. There were no special limitations pertaining to third molars. Throughout the period of this study, coverage did not change in any major way that would be expected to affect patterns of third-molar removal. We conducted individual analyses using further subsets of these claims data; for example, continuously insured people in the longitudinal analyses and subgroups of various ages in many of the analyses. We detail these subsets and their corresponding results in the next section.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION LIMITATIONS CONCLUSIONS REFERENCES

 
Figure 1 shows why the issue of third molars is of interest. In 1999, third molars were the most commonly extracted permanent teeth in DDPM-insured patients who resided both in and outside Michigan. In fact, among the 1,846,083 continuously insured people who resided in Michigan during 1999, third-molar extractions accounted for 89,846 (45.3 percent) of the 198,215 permanent teeth extracted. Outside Michigan, among 1,551,000 continuously insured people, third molars accounted for 56,449 (31.6 percent) of the 178,799 permanent teeth extracted. The difference in these percentages primarily is the result of differences in the age mix of patients between the two groups; Michigan had a somewhat higher proportion of adolescent beneficiaries. Figure 2 shows the number of third-molar extractions per patient; demonstrates that in both Michigan and the rest of the country, the overwhelming majority of these third molars were extracted from patients 15 to 25 years of age, peaking at age 18 years; and shows that the patterns were virtually the same between Michigan and the rest of the country when age was controlled for.

View larger version (58K): [in this window] [in a new window]   Figure 1. Extraction of permanent teeth by tooth number in 1999.

View larger version (57K): [in this window] [in a new window]   Figure 2. Average number of third-molar extractions per patient (range 0–4) by age in Michigan and non-Michigan residents, 1999.

Regarding the pattern of these extractions, Friday was the preferred day of the week for extractions, and Figure 3 demonstrates that the favored month was August, followed by July, December and June. In December 1999, the peak days for third-molar extraction were the Monday before Christmas and the Monday and Tuesday between Christmas and New Year’s Day; these days were the three days in 1999 with the highest number of third-molar extractions.

View larger version (58K): [in this window] [in a new window]   Figure 3. Third-molar extractions by month in patients 16 to 21 years of age, 1999.

We limited these longitudinal analyses to patients from Michigan, because it was the only state for which we had large numbers of people who were continuously insured since at least 1991. Also, only in Michigan were the numbers sufficient to allow for inferences about referral patterns for individual dental practitioners.

Claims data from 13,343 of these 1978 birth cohort patients, who were continuously insured during the entire 8.5-year period, show that 5,929 (44.4 percent) had one or more third molars removed. The table shows further that during this 8.5-year period, 80.4 percent of those who had third-molar extractions had all four teeth removed, and of these 94.4 percent had all four teeth removed in a single visit.

Because our focus for this part of the analysis was dental practices, we further limited the analysis to patients who received all of their routine care from only one general or pediatric dentist. In this way, we were able to characterize the patterns of referral for specialty care (in this case, third-molar extraction), according to the referring general or pediatric dentist; approximately 90 percent of third-molar extractions were performed by oral surgeons. As 33,144 of these patients each received their routine care from one dentist, we could appropriately associate each of their patterns of care with a single dentist. We further limited the analysis to dentists who treated at least 10 of these patients, to ensure that the patterns of care attributed to them were based on more than a few patients. The result was that we were able to analyze the patterns of third-molar removal among 22,356 adolescents in 1,048 dental practices, and the practice patterns were based on an average of 21.3 patients per practice.

Figure 4 shows the distribution of these 1,048 dental practices by percentage of adolescent patients who had third molars removed during the 8.5-year period. For example, Figure 4 shows that in 75 dental practices, 10 percent or fewer of the insured adolescent patients had their third molars extracted during the 8.5-year period. In fact, in 22 of these 75 practices, none of the insured adolescents had their third molars removed. At the other end of the spectrum, in 57 practices more than 70 percent of the adolescent patients had their third molars removed. Most practices were somewhere between these extremes.

View larger version (48K): [in this window] [in a new window]   Figure 4. Distribution of 1,048 dental practices in Michigan by percentage of adolescent patients who had third-molar extractions from July 1991 through December 1999.

View larger version (48K): [in this window] [in a new window]   Figure 5. Distribution of 1,026 dental practices in Michigan by average age of adolescent patients when they had their third molars extracted.

View larger version (49K): [in this window] [in a new window]   Figure 6. Percentage of third-molar extractions that are partial or full bony impactions (procedure codes 7230, 7240 and 7241), by age of patient.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION LIMITATIONS CONCLUSIONS REFERENCES

Another important pattern that was revealed by our analysis is that there is substantial variation among dental practices in whether their patients have their third molars removed and in the timing of removal. It appears that adolescent patients in some dental practices rarely are referred for third-molar removal, while in others most or all adolescent patients have their third molars removed. In addition, while the average age of third-molar removal is 18 years, in some practices the average is several years younger, while in still others it is several years older. This difference in the average age at which third molars are removed has implications for the cost of the procedure and presumably for the difficulty of the extractions. Almost all of the third molars extracted from patients 16 years of age or younger are submitted with the procedure codes 7230, 7240 or 7241, which are for extractions that are full or partial bony impactions. In contrast, in patients just a few years older, the mix shifts to less expensive, and presumably less difficult, procedures.

While almost 90 percent of third molars are removed by oral surgeons, it clearly is the general practitioners and pediatric dentists who are responsible for making the referrals to these specialists.

These results also support many earlier suggestions in the literature of revisiting the question about removal of pathology-free third molars. While almost 90 percent of third molars are removed by oral surgeons, it clearly is the general practitioners and pediatric dentists who are responsible for making the referrals to these specialists. These insurance claims demonstrate that there is wide divergence in and a lack of consensus as to whether or not dentists recommend that their patients have their third molars removed and the age at which they recommend that these teeth be removed.

This variation is important because it is reasonable to infer several fundamentally different referral patterns from these data. Some providers seem to refer all, or nearly all, of their adolescent patients for third-molar removal. It seems unlikely that all of these patients, especially those younger than 17 years of age, have active pathology associated with all four third molars. In other practices, referral for third-molar extractions appears to be much less common, and in others it is rare. Because the vast majority of these extractions are referred to specialists, the referring dentist does not have a direct economic stake in the decision. We reasonably assume that the decision to refer is based largely on each dentist’s perception of the trade-off between the risks, discomfort and costs associated with early removal and the potential risks, discomfort and costs associated with retention of these teeth. If the argument is to be made that either approach is defensible, then an obvious question arises as to why one would not choose the alternative that has the lowest cost and subjects the fewest people to surgical procedures.

	LIMITATIONS

TOP ABSTRACT METHODS RESULTS DISCUSSION LIMITATIONS CONCLUSIONS REFERENCES

 
It is important to acknowledge that this study is limited to an insured population. We believe that it would be interesting to compare these findings with the current patterns of treatment for noninsured people. Although dental benefit coverage should not determine treatment decisions, the peak age of 18 years for removal of third molars also may be associated with the potential loss of dental coverage for those who lose dependent status because of leaving school.

	CONCLUSIONS

TOP ABSTRACT METHODS RESULTS DISCUSSION LIMITATIONS CONCLUSIONS REFERENCES

 
The controversy over routine removal of pathology-free third molars continues. In this article, it was not our intent to suggest which is the best approach, but rather to show practice as it exists for a large group of insured people. The range of opinion is wide, from removal of only pathologically involved third molars, according to narrow definitions of what constitutes pathology, to removing all four teeth at an early age, even before eruption would normally occur.

These diverse opinions reflect the range of practice of general practitioners, as they are the ones who control the system through their referral patterns. The apparent lack of consensus should be resolved so the profession can maintain the public’s confidence that the recommended care is based on sound evidence.