Using exposure-response analyses from a randomized trial, this study found consistent but not statistically significant associations between UDMA-based polyacid-modified (compomer) or standard minifill bisGMA-based dental composite and neuropsychological health in children. Overall, most change scores were slightly poorer among children with greater compomer or composite exposure, which is in contrast to the findings for amalgam exposure levels (Bellingeret al., 2007b
). However, of 4 primary tests of executive function, scores on only one, the test of Letter Fluency, were statistically significantly related to compomer/composite exposure, and the results for other general tests of intelligence, learning, memory, and visual-spatial skills indicate that chance cannot be ruled out as an explanation for these findings.
A strength of this study is that the data were obtained as part of a randomized clinical trial. For this analysis of composite treatment level, the previously reported intent-to-treat and amalgam level analyses are available to evaluate the possibility of confounding by factors related to severity of dental disease, or disease on primary teeth vs. permanent teeth. Of note, the current finding of poorer change scores with higher composite exposure is in contrast to findings from analyses of amalgam exposure-response, where change scores tended to improve (Bellingeret al., 2007b
). Thus, confounding by factors related to severity of dental disease is unlikely. Furthermore, the current findings are consistent with the intent-to-treat results, which showed that children randomized to amalgam fared better on the 3 primary neuropsychological endpoints of the trial (WISC-II full-scale intelligence, General memory index, and Visuomotor global score) (Bellingeret al., 2006
). A possible explanation for these trends is that we are observing actual effects of composites, but the effects on neuropsychological health measures are indeed small, as reflected in both the original trial and these additional analyses, and that they are statistically insignificant due to insufficient power to detect differences of such small magnitude. This possibility would be definitively addressed only by additional randomized controlled trials designed specifically to assess smaller effect sizes with chronic exposure.
In advance of such data, given the lack of statistical significance in our results, associations between composites and worse neuropsychological outcomes may be spurious findings due to chance. A consequence of conducting multiple neuropsychological tests may be a compromise in the statistical significance threshold. The multivariate analysis of variance, which had the benefit of minimizing the likelihood of Type I errors due to multiple tests of executive function, confirmed that the impairment in change scores was not statistically significant.
The rationale for using multiple neuropsychological outcome measures was that various tests are often conducted in a neuropsychological assessment to enhance interpretation of results, particularly for complex constructs such as executive functioning (Golden et al., 2002
). Executive functions include planning, directing and maintaining attention, organization, abstract reasoning and problem-solving, self-regulation, and motor control, and may be confounded by intelligence, memory, and language. Executive functioning was of particular interest given numerous rodent experiments showing effects of bisphenol A on executive function measures, and findings that gestational BPA exposure is associated with poorer executive functioning in girls (Braunet al., 2011
). In the current study, children with greater exposure to compomer or composite fared worse on most tests, but the differences were statistically significant for only one test of executive function, Letter Fluency (sum of 3 trials from the Verbal Fluency test). This test required the child to combine processes related to verbal language within a set of constraints (e.g., to name words beginning with the letter ‘F’) which require the inhibition of inappropriate responses. For another executive function test, the Stroop Color-Word Interference Test, children with greater composites exposure performed worse only on the Color portion (naming the color of a bar). Poor performance on the Color portion alone may be affected by speech motor function, or the individual’s ability to name colors. In the absence of colorblindness, an impaired score on the Color trial may indicate brain dysfunction in the left (dominant) temporal-occipital or the right (non-dominant) posterior area (Goldenet al., 2002
). We found no similar associations with greater amalgam exposure levels in this study.
Previous studies of maternal prenatal urinary BPA measures and child neuropsychological health or behavior have shown inconsistent results. Braun et al. (Braunet al., 2011
, Braun et al., 2009
) found adverse associations with executive functioning and social behavior particularly in girls. Perera et al. (Perera et al., 2012
) found the reverse, whereby higher maternal prenatal BPA concentration was associated with worse parent-reported social behavior scores in boys, but not girls. In a study of 5-week old infants, maternal prenatal BPA levels were not associated with neurobehavioral scores (Yoltonet al., 2011
). Lacking consistency across studies, firm conclusions cannot yet be drawn regarding the role of BPA in human neuropsychological behavior.
In this study, no data were available to evaluate BPA levels, or whether levels of other chemicals that may leach from dental composites, such as bisGMA, UDMA, or TEGDMA (Van Landuytet al., 2011
), increased with greater composite exposure. Prior in vitro and in vivo studies have shown that the minifill composite (Z100) used in NECAT released BPA, bisGMA, bisDMA, and BADGE (Al-Hiyasat et al., 2004
, Martinet al., 2005
, Ortengren et al., 2004
, Pulgar et al., 2000
, Sasaki et al., 2005
, Yap et al., 2004
). For the polyacid-modified composite (Dyract compomer), no detectable BPA or bisGMA were found in eluates from filled tooth samples in one study (Hamid et al., 1998
). For both composites, numerous studies have reported cytotoxic effects (Milhem et al., 2008
, Schweikl et al., 2005
, Sletten and Dahl, 1999
, Wataha et al., 1999
). Resin-composite restorations have comparatively high failure rates (Soncini et al., 2007
). As they undergo degradation, resin components are released, presumably throughout the life of the restoration. Additional studies are needed to measure the long-term release of components of resin-based dental materials and whether the levels absorbed or excreted are associated with adverse health effects.
In conclusion, consistent but not statistically significant associations were observed between methacrylate-based dental composites and small impairments in neuropsychological test score changes over 4- or 5-years of follow-up among children in this randomized clinical trial. Although findings were not statistically significant, they were consistent in the direction suggesting small adverse effects with greater composite exposure. Furthermore, this consistency was not observed for the randomly-assigned amalgam treatment. In light of the previously-reported differences in behavioral outcomes (Bellingeret al., 2008
, Maserejianet al., 2012
), and the non-significant tendency for composites exposure to be associated with poorer neuropsychological test scores both in the present study and in a separate randomized study of children (Bellingeret al., 2006
, DeRouen et al., 2006
), further studies on the safety of methacrylate-based dental composite materials are warranted. Newly-developed dental materials (e.g., ormocer-based, silorane-based) exist, and because they may have distinct properties in chemical composition, biocomptability, and wear (Ilie and Hickel, 2011
, Polydorou et al., 2009
), their potential for toxicity should also be thoroughly tested. In the meantime, well-established factors such as the durability of amalgam, technique-sensitivity needed for composite vs. amalgam, preservation of sound tooth structure, cosmetic preferences of the patient, and differing financial costs should remain among the primary determinants of a decision regarding choice of dental restorative materials.