Does Fluoride in Compomers Prevent Future Caries in Children?

Abstract

Compomer restorations release fluoride to help prevent future caries. We tested the hypothesis that compomer is associated with fewer future caries compared with amalgam. The five-year New England Children’s Amalgam Trial recruited 534 children aged 6-10 yrs with ≥ 2 carious posterior teeth. Children were randomized to receive compomer or amalgam restorations in primary posterior teeth, placed with a fluoride-releasing bonding agent. The association between restorative material and future caries was assessed by survival analysis. Average follow-up of restorations (N = 1085 compomer, 954 amalgams) was 2.8 ± 1.4 yrs in 441 children. No significant difference between materials was found in the rate of new caries on different surfaces of the same tooth. Incident caries on other teeth appeared slightly more quickly after placement of compomer restorations (p = 0.007), but the difference was negligible after 5 yrs. Under the conditions of this trial, we found no preventive benefit to fluoride-releasing compomer compared with amalgam.

Introduction

Compomer dental restorations were developed to combine the mechanical and esthetic properties of composites with the fluoride-releasing advantage of glass-ionomer cements. The fluoride released into the mouth by compomer is intended to help protect against future caries (Eichmiller and Marjenhoff, 1998; Burke et al., 2006). It has been shown that compomer does in fact release fluoride into the mouth (Chung et al., 1998; Grobler et al., 1998; Ylp and Smales, 1999; Karantakis et al., 2000; Asmussen and Peutzfeldt, 2002; Marczuk-Kolada et al., 2006), which has a preventive effect against future caries compared with composite materials without fluoride (Chung et al., 1998; Dionysopoulos et al., 1998; Donly and Grandgenett, 1998; Hicks et al., 2000; Torii et al., 2001; Attar and Onen, 2002; Gonzalez Ede et al., 2004; Yaman et al., 2004). Two survey articles (Burke et al., 2006; Wiegand et al., 2006) concluded that although fluoride-releasing materials, including compomer, inhibit the formation of caries in vitro, such effects have not yet been determined in vivo. Additionally, there is very limited literature comparing compomer with amalgam in this regard, despite the continued widespread use of amalgam for dental restorations.

Although amalgam contains no fluoride to protect against future caries, it has very different properties compared with compomer/composite, which may protect against future caries by other means. One prior study (Marks et al., 1999) has compared recurrent caries by compomer and amalgam materials in primary molars using a split-mouth design. After 3 yrs of follow-up, out of 17 restorations with each material, one restoration of each type required replacement due to recurrent caries. However, the sample size of this study was too small for a small to moderate effect to be detected. Therefore, the relative potential for prevention of future caries for compomer and amalgam has not yet been well-established.

The purpose of this analysis is to compare the incidence of new caries after children are randomized to receive compomer or amalgam restorations, with data collected prospectively as part of the New England Children’s Amalgam Trial (NECAT). Data from this trial have already shown a greater need for replacement of primary posterior restorations due to recurrent caries on the same tooth surface after placement of compomer compared with amalgam restorations (3.0% vs. 0.5%, p = 0.002) (Soncini et al., 2007). This finding may have resulted from properties of compomer restorations, such as microleakage, outweighing any potential beneficial effects of released fluoride. However, it remains possible that fluoride from compomers succeeds in protecting other teeth in the mouth. It is important to examine this possibility because, if there is a benefit to compomer over amalgam in preventing future caries on other surfaces, such a benefit may outweigh the risk of more frequent replacement of compomers. Indeed, an assessment of dental material choice must consider the overall picture of future dental procedures, including new restoration placement as well as replacement of existing restorations.

Materials & Methods

Study Design and Participants

A detailed discussion of the design of NECAT has been previously published (The Children’s Amalgam Trial, 2003; Bellinger et al., 2006). The study was approved by the institutional review boards of the New England Research Institutes, The Forsyth Institute, and all dental clinics. English-speaking children in Boston, MA, and Farmington, ME, aged 6-10 yrs at baseline, were eligible if they had no known prior or existing amalgam restorations, ≥ 2 posterior teeth with dental caries requiring restorations on occlusal surfaces, and no clinical evidence of existing psychological, behavioral, neurological, immunosuppressive, or renal disorders. Parental consent and child assent were obtained for 534 children.

Dental Treatment and Clinical Procedures

A complete dental examination was scheduled every 6 mos during the five-year trial, between September 1997 and March 2005. Participating children were provided preventive and restorative dental care. Sealants (Ultradent XT plus; Ultradent, South Jordan, UT, USA ) were placed on all sound permanent and primary molars with deep pits and fissures and were replaced as needed in all children. Restorations were continually placed over the course of the trial as needed, according to assigned treatment. Children were randomized to receive either amalgam (n = 267) or composite/compomer (n = 267) restorations at baseline and during the course of the trial. For children assigned to the composite/compomer group, compomer was placed in primary dentition, while composite was placed in permanent dentition. For children in the amalgam group, composite/compomer was used in the anterior dentition if required by standard clinical practice guidelines. The amalgam material used was Dispersalloy (Dentsply Caulk, Milford, DE, USA), and the compomer material was Dyract (Dentsply Caulk).

One dentist (J.A.S.) treated 97% of the Boston-area participants, with two additional dentists treating the remaining seven children at one Boston site. Three dentists treated rural Maine participants during the course of the trial. Clinical variability was minimized by centralized training of all dental personnel and the use of standard pediatric dental procedures as specified in the NECAT protocol and procedures manual. The same technique was used in the placement of all restorations, with rubber dam used the majority of the time. Following the complete removal of decay, the tooth was acid-etched with 30% phosphoric acid for 20 sec and washed thoroughly. A bonding agent (Optibond; Kerr, Orange, CA, USA) that has a fluoride-release characteristic was applied and light-cured for 30 sec. The restoration was then placed according to the manufacturer’s indications. Although fluoride-containing bonding agents are less commonly used in the placement of amalgam compared with compomer restorations, the use of bonding agents had to be standardized in this clinical trial (designed to compare the safety of the two types of materials themselves), to eliminate possible confounding (e.g., by bonding agent).

Statistical Analysis

Only exposure to posterior primary restorations was considered for this analysis, since NECAT used compomer only in primary teeth and amalgam only in posterior teeth. However, determination of future caries, after placement of the restoration, included all teeth (including permanent and anterior teeth).

Each restoration contributed follow-up from the date of initial placement to the date of exfoliation, extraction, or the child’s last dental visit (whether at year 5 or before withdrawal from the trial), whichever occurred first. Because restorations were placed at the baseline dental visit, as well as during follow-up visits over the five-year trial, the start of follow-up time varied by restoration. We estimated the date of exfoliation by averaging the dates of the last dental visit with the primary tooth and the first dental visit with the corresponding permanent tooth; since dental exams were performed every 6 mos and documented the status of each tooth, the date of exfoliation is accurate to within 3 mos. Restorations placed with no subsequent follow-up (i.e., at the last dental visit before withdrawal, tooth exfoliation, or at the end of the trial; N = 80) were excluded from all analyses and descriptive statistics.

The time from placement of each restoration to the development of each new carious surface recorded while the initial restoration was still in the mouth was calculated. There were two outcomes for this analysis: the times until (1) new caries on a different surface of the same tooth, and (2) new caries on any different teeth. To test whether these outcomes varied by type of restoration material (compomer vs. amalgam, as treated), we used a random-effects accelerated failure time model with proportional hazards (Cox and Oakes, 1984). The models adjusted for the following covariates when they were significant (p < 0.05) or changed the effect of restoration material by > 10%: age, gender, socio-economic status (a weighted average of household income and education), and the number of restorations (of either type) in the mouth at the time the restoration was placed. The random effect in this survival model was the child, to account for the potential correlation between restorations in the same mouth. Survival curves were plotted to show the percents of compomer and amalgam restorations that had future caries develop either on the same tooth or on different teeth after placement through the follow-up time period.

Results

At baseline, the average age of participants was 7.9 ± 1.3 yrs (Table). The mean number of total caries at baseline was 9.5 decayed tooth surfaces, of which 7.8 were in primary teeth and 1.7 were in permanent teeth. The sample was gender-balanced and racially diverse.

Table.

Baseline Characteristics of New England Children’s Amalgam Trial Participants (N = 534), by Assigned Treatment^a

Baseline Characteristic	Amalgam Group (n = 267)	Composite Group (n = 267)
Study site, n (%)
Boston	144 (53.9)	147 (55.1)
Maine	123 (46.1)	120 (44.9)
Carious surfaces,
mean (SD) range	9.8 (6.9) 2-39	9.3 (6.2) 2-36
Age, mean (SD), yrs	7.9 (1.3)	7.9 (1.4)
Gender, n (%)
Female	131 (49.1)	156 (58.4)
Male	136 (50.9)	111 (41.6)
Race, n (%)b
Non-Hispanic white	165 (64.0)	158 (60.3)
Non-Hispanic black	49 (19.0)	49 (18.7)
Hispanic	15 (5.8)	23 (8.8)
Other	29 (11.2)	32 (12.2)
Household income, n (%)
≤ $20,000	74 (29.2)	86 (33.1)
$20,001 - $40,000	113 (44.7)	109 (41.9)
> $40,000	66 (26.1)	65 (25.0)
Education of primary caregiver, n (%)
< High school	34 (13.2)	38 (14.6)
High school graduate	197 (76.4)	194 (74.3)
College graduate	18 (7.9)	17 (6.5)
Post-college degree	9 (3.5)	12 (4.6)

^aFor race, data were available for 520 participants (N = 520); for income, N = 513; for education, N = 519.

^bRace was self-reported by the parents of the children.

Over the five-year trial, 1085 compomer and 954 amalgam restorations with follow-up were placed on posterior primary surfaces in 441 children (243 receiving compomer, 218 receiving amalgam, with 20 receiving both). The average length of restoration follow-up was 2.8 ± 1.4 yrs, with a range of 0.03-6.3 yrs. Although this was a five-year study with dental visits every 6 mos, children sometimes scheduled visits at longer intervals, resulting in greater than 5 yrs of follow-up.

Survival curves for time until new caries after placement of either compomer or amalgam restorations were plotted by treatment group (Fig.). Random-effects survival analysis showed no significant effect of dental material on new caries on the same tooth (p = 0.98, Fig., a). However, there was a very small, but statistically significant (p = 0.007), effect of dental material on new caries formed on different teeth, with longer time until the formation of new caries following placement of amalgam restorations (Fig., b).

Figure.

Rates of new caries after restoration placement, by treatment group, in the New England Children’s Amalgam Trial (N = 2039 restorations). (a) Rate of new caries on a different surface of the same tooth. P = 0.98, calculated from a random-effects accelerated-failure time model with proportional hazards, adjusted for gender, socio-economic status, and number of decayed and filled surfaces in the mouth. (b) Rate of new caries on a different tooth. P = 0.007, calculated from a random-effects accelerated-failure time model with proportional hazards, adjusted for age.

In comparison of new caries following those primary restorations with five-year follow-up (N = 211: 109 composite ± 102 amalgam), amalgam was associated with a slightly higher percentage of new caries on different surfaces of the same tooth after 5 yrs of follow-up (17.7% vs. 14.7% of restorations). The mean number of new caries on different teeth 5 yrs after restoration placement was also slightly higher for amalgam (4.2 vs. 3.5 new carious surfaces). In analyses that examined the percentage of sound teeth in the mouth at baseline that developed caries during the five-year follow-up, there was no difference by restoration material; 8.5% of sound teeth in either arm developed caries. This relatively low percent may be explained by the exfoliation of sound primary teeth over the five-year follow-up. Almost all children presented with caries during follow-up, though often on teeth that were either already decayed or not present in the mouth at baseline.

Discussion

This paper presents, for the first time in a large sample, analyses to compare the formation of future caries after placement of compomer vs. amalgam restorations. The NECAT data allowed us to investigate this association properly, in vivo, by recruiting a large cohort of children with a high rate of initial restorations, randomly assigning them to amalgam or compomer, thus removing selection biases, and providing regular dental care during the course of the five-year trial.

This dataset previously showed greater need for replacement of primary posterior restorations due to recurrent caries on the same tooth surface after placement of compomer compared with amalgam restorations (3.0% vs. 0.5%, p = 0.002) (Soncini et al., 2007). Adding to this, the current analysis demonstrated that, under the restoration placement conditions of the trial, there was no protective effect of compomer relative to amalgam in its ability to prevent future caries on different surfaces/teeth. Rather, incident caries on other teeth (i.e., other than the tooth where the restoration was placed) appeared more quickly after the placement of compomer restorations, but the difference, though statistically significant, was very small, and the survival curves showed little difference between compomer and amalgam at the end of follow-up. In contrast, among restorations with a full 5 yrs of follow-up, the mean number of new caries on other teeth or on other surfaces in the same tooth after placement of compomer was actually slightly lower, compared with amalgam. Thus, overall, there was no clear or consistent effect of these dental materials on the formation of new caries on different surfaces/teeth in our study. We conclude that there does not appear to be an advantage to the fluoride-releasing compomer compared with amalgam restorations placed with the Optibond bonding agent for the prevention of new caries in children.

A limitation in the generalizability of these findings is that a bonding agent (Optibond) with a fluoride-release characteristic was used for all restorations, such that all restored surfaces received some fluoride exposure. Additionally, the sealant used on children in both treatment groups also released fluoride into the mouth. Therefore, this was a comparison of compomer and amalgam in an oral environment that had an unmeasured, but relatively constant, base level of fluoride. However, the Optibond bonding agent releases fluoride into the tooth, as opposed to the compomer restoration, which releases fluoride into the mouth, resulting in a different route of fluoride exposure.It is possible that either the small amount of fluoride releasedinto the tooth from the Optibond liner protects against future caries in a magnitude similar to the fluoride-releasing properties of compomer, or, despite the received benefit of additional fluoride from compomer, compomer is still no more successful at preventing future caries due to other properties. For example, the problem of microleakage, common to compomer and composites, may be a factor contributing to the formation of recurrent caries (Burgess et al., 2002; Estafan and Agosta, 2003; Soncini et al., 2007). It has been noted that “despite the cariostatic effects possibly achieved from fluoride-releasing materials, secondary caries is still one of the main reasons for clinical failure of restorations” (Wiegand et al., 2006).

Additional study comparing compomer with amalgam in other oral environments would be worthwhile. Still, because children with multiple dental treatment needs, as in our study, are most in need of protection against future caries, the use of fluoride agents at various levels of treatment may represent an increasingly common standard of care (American Academy of Pediatric Dentistry, 2005). In light of our findings, the evidence to date suggests no advantage of compomer over amalgam in protecting against the formation of future caries. Therefore, although there is reason to prefer the use of compomer rather than composite in primary teeth, because of the cariostatic property of fluoride (Chung et al., 1998; Dionysopoulos et al., 1998; Donly and Grandgenett, 1998; Hicks et al., 2000; Torii et al., 2001; Attar and Onen, 2002; Gonzalez Ede et al., 2004; Yaman et al., 2004), compomer does not perform better than amalgam in this respect when amalgam is placed with a fluoride-releasing bonding agent. Considering the results presented here, together with our previously published results on new restoration placement and replacement of existing restorations (Soncini et al., 2007), our study suggests that placement of compomer restorations in a high-risk population results in more future dental needs compared with amalgam.