Dental Sealants and Flowable Composite Restorations and Psychosocial, Neuropsychological, and Physical Development in Children

Abstract

Purpose

Dental sealant materials may release their components, including bisphenol-A (BPA), intra-orally, but long-term health effects are uncertain. The New England Children’s Amalgam Trial (NECAT) found that composite restorations were associated with psychosocial, but not neuropsychological or physical, outcomes. The previous analysis did not consider sealants and preventive resin restorations (PRRs), which were routinely placed during NECAT. This analysis examines sealant/PRR exposure in association with psychosocial and other health outcomes.

Methods

NECAT recruited 534 children aged 6–10 years and provided dental care during 5-year follow-up. Annually, examiners conducted psychosocial and neuropsychological tests, and measured body mass index (BMI) and fat percentage (BF%). Associations between surface-years (SY) of sealants/PRRs and outcomes were tested using multivariable models.

Results

Cumulative exposure level to sealants and/or PRRs was not associated with psychosocial assessments (e.g. total problems: Child Behavior Checklist, 10-SY β=−0.2, SE=0.3, P=0.6) or neuropsychological tests (e.g. full-scale IQ, 10-SY β=0.1, SE=0.2, P=0.6). There were no associations for changes in BMI-for-age z-score (P=0.4), BF% (girls 10-SY β= −0.2 SE=0.3; boys 10-SY β= −0.1 SE=0.3), or menarche (10-SY hazard ratio=0.91, 95%CI 0.83–1.01, P=0.08).

Conclusions

This analysis showed no significant associations between exposure level of dental sealants or PRRs and behavioral, neuropsychological, or physical development in children during 5-year follow-up.

INTRODUCTION

Dental sealants on primary and permanent molar teeth are an important component of children’s oral health and caries prevention.^1–3 Despite their proven benefits, recent concerns regarding chemical exposures from dental materials have resulted from laboratory findings indicating that the resins used in sealants are possible exposure sources to bisphenol-A (BPA) and other compounds that have potential adverse effects.^4–7 BPA is a well-known endocrine disruptor and causes a variety of adverse health effects at low dose exposure levels in laboratory animal experiments.⁸ The main monomer used in many manufactured sealants today, bisphenol-A-glycidyl-dimethacrylate (bisGMA), has also been shown to cause local cytotoxicity, DNA damage, and estrogenic activity.^9–13 For example, BisGMA exposure significantly reduced fertility, with reductions in sperm counts, weights of testis and preputial glands, and overall body weight in adult male mice,¹⁴ and significantly increased embryo resorption rates and ovary weights in adult female mice.¹⁵ Observational human cohort studies have identified associations between urinary BPA levels and various adverse health measures, ranging from behavior problems in children to coronary artery disease in adults.^16–22 However, additional longitudinal studies are needed to help examine causal associations.

Clinical studies measuring salivary and/or urinary bisGMA, BPA and other compounds before and after placement of sealants or composite restorations have shown increased biomarker levels up to 30 hours after their placement in adults.^23–27 Children have higher levels of urinary BPA than all other age groups in the U.S. and may be most sensitive to exposure effects.²⁸ Previously, we examined the association of composite restoration placement with adverse health outcomes over five years, using data from the New England Children’s Amalgam Trial (NECAT). NECAT randomized 534 children to receive either amalgam or composite for posterior tooth restorations, collecting detailed data to allow an examination of treatment levels over time. While results showed no adverse effects for amalgam,²⁹ children randomized to composites, especially those with greater exposure to bisGMA-based composites, had worse psychosocial health measures during follow-up.^{30, 31} Further examination of neuropsychological test scores and physical development measures showed no consistent associations between composites and those outcomes, indicating specificity to behavioral health effects.^{32, 33}

In addition to restorative fillings, all children participating in NECAT received pit and fissure sealants during the 5-year trial, and, as needed, flowable composite for preventive resin restorations (PRRs) for shallow caries. Compared to composites used for restorations, sealants are designed to be highly flowable, to facilitate their infiltration onto the surfaces of dental pits and fissures. The high viscosity of sealants and flowable composites is attained by using substantially greater amounts of resin monomers that contain bisGMA.

No prior studies have examined exposure to dental sealants in association with long-term adverse health effects in children. The objective of this manuscript was to examine whether, consistent with previous findings for packable composite restorations in NECAT,^{30, 31} flowable composite (sealants and PRRs) treatment levels were associated with the behavioral outcome measures. We also examined associations between sealants/PRRs and neuropsychological and physical development measures during the 5-year follow-up.

METHODS

Study Population

NECAT was a randomized safety clinical trial of amalgam vs. composite for posterior restorations conducted among 534 children at six community dental clinics in urban Boston and rural Maine from 1997–2005. Eligible children were aged 6–10 years at baseline, had no existing amalgam restorations; ≥2 posterior teeth with caries requiring restoration on occlusal surfaces; fluent in English, and, by parent-report, no physician-diagnosed psychological, behavioral, neurological, immunosuppressive or renal disease. Written informed consent was obtained from the parent/guardian, and signed assent was obtained for children aged eight years or older. The study was approved by the institutional review boards of New England Research Institutes (Watertown, MA, USA) and all participating sites. Additional details on the study procedures have been published.^{29, 34, 35}

Dental Materials and Interventions

Participants received comprehensive dental care semi-annually during their 5-year participation. Standard dental care included exams, cleaning, fluoride application, sealant placement, and restorative treatment. Dental procedures and materials were standardized across study sites, following manufacturer’s indications for use. The flowable composite material used for preventive sealing of sound pits and fissures of posterior teeth, referred to as a sealant, was Ultraseal XT by Ultradent (South Jordan, Utah, USA). The flowable composite used for PRRs, which treated shallow, incipient caries that did not extend into the dentin, was Revolution by Kerr (Orange, Calif., USA). Composite restoration materials used were Z100 by 3M ESPE (St. Paul, Minn., USA) for permanent teeth, and Dyract AP compomer by Dentsply Caulk (Milford, Del., USA) for primary teeth. Sealants and PRRs were placed as needed regardless of assigned treatment group for restorations.

Health Outcome Measures

One supervising psychologist (D.C.B.) trained and certified examiners to conduct psychosocial and neuropsychological tests and continuously monitored them for quality control throughout the trial. The study used two validated instruments for psychosocial assessments at baseline and follow-up: (1) Child Behavior Checklist (CBCL) parent-report,³⁶ and (2) Behavior Assessment for Children Self-Report (BASC-SR).³⁷ Both are widely used in screening children and adolescents for psychosocial problems,³⁸ yielding global T-scores (mean 50, SD=10) and core syndrome scores. Based on our previous findings, the primary outcomes of interest were BASC-SR scores on the four global scales (Total emotional symptoms index, Clinical maladjustment, School maladjustment, Personal adjustment) measured at the end of 5-year follow-up. We did not analyze change since baseline for BASC-SR because most (63 percent) of the subjects were <8 years of age at baseline, and BASC-SR was not developed or validated for children under eight years of age. Secondary behavioral outcomes were BASC-SR subscale scores, CBCL change scores, and the percentage of children with at-risk or clinically-significant scores at the end of follow-up.

For neuropsychological assessments, the study used a battery of validated tests of executive functioning, intelligence, memory, visual-spatial skills, verbal fluency, and problem-solving, which together help enhance interpretation of this complex outcome.³⁹ Neuropsychological outcomes of interest were WISC-III Full-Scale IQ and factors of verbal comprehension, perceptual organization, freedom from distractibility, processing speed, the Wechsler Individual Achievement Test (WIAT) reading and math scores, and the following tests of executive function: Trail-Making Test, Letter Fluency subtest of Verbal Fluency, Stroop Color-Word Interference Test, and Wisconsin Card Sorting Test.

At baseline and annual follow-up visits, NECAT data collectors measured height, weight, and body fat percentage, using a calibrated bioimpedance scale (model TBF-551, Tanita Corp of America, Inc.).⁴⁰ We calculated body mass index (kg/m²) and BMI-for-age z-score using the Centers for Disease Control (CDC) data files, defining overweight or obese as a BMI-for-age z-score ≥ 85^th percentile.^{41, 42} Female participants reported menarche status annually as no/yes, and if yes, month and year of menarche.

Statistical Analysis and Power

We excluded children who were missing data on placement of sealants (n=44), PRR (n=7), or both (n=33), resulting in 450 children in the total analytic cohort. We initially examined sealants and PRRs separately, and then combined them to create a measure of total flowable composite exposure, because (i) both contained bisGMA as a primary source monomer according to the material safety data sheets and (ii) PRRs were often placed using the same material as that used for preventive sealants. We calculated the cumulative treatment levels per subject using a surface-years (SY) measure, which takes each treated surface and weights it by the length of time that it was present in the mouth. We categorized surface-years into four categories: none, and among the exposed, tertile of exposure level.

For each health outcome, we conducted statistical analyses separately and excluded children with missing outcome data (see results tables for sample sizes in each analysis). Numerous sensitivity analyses were conducted, such as: (i) excluding children (N=39) who had pre-existing sealants at baseline, (ii) excluding children (also N=39, coincidentally) who had no sealants or PRRs during the study, because this deviation from standard clinical procedures may indicate an unmeasured source of confounding from children in this minority, and (iii) for analyses of psychosocial outcomes, excluding 302 children who ever received composite restorations during the study, given our previously reported findings of an association between composite restorations and psychosocial measures. Results of these sensitivity analyses confirmed those of the primary analyses and therefore are not presented.

Outcomes were analyzed as changes in score between baseline and follow-up for all measures except the BASC-SR (as explained above) and in the exploratory analysis of menarche. Physical development outcomes were analyzed separately for boys and girls. Repeated measurements of body fat percentage and body mass index based on annual study visits were used in linear mixed effects models. Models included subject-specific intercepts to account for natural heterogeneity in the population, age-specific slopes to account for differences in growth trajectory by baseline age and, if statistically significant, piece-wise linear splines to account for changes in growth at puberty.⁴³ Age at menarche was analyzed among girls from the Maine clinical site only (N=113), because menarche data collection was incomplete at the Boston site.

We examined associations using multivariable models adjusting for relevant sociodemographic and dental variables. The following variables were included in the multivariable models, depending on the outcome of interest: age, sex, race/ethnicity (white, black, Hispanic, or other), household income, socioeconomic status, primary care giver’s marital status, geographic study region (urban Boston or rural Maine), blood lead level, fruit and vegetable intake, maternal alcohol/tobacco/drug exposure during pregnancy, and composite restorations (surface-years). Variables were retained in models if they were associated with the outcome at P<0.2 level (see footnotes of Tables 3–5 for variables included for each specific outcome). Using cross-product terms in the multivariable models, we tested interactions between sealant/PRR treatment levels and age, study site (Boston or Maine), or sex (for psychosocial and neuropsychological outcomes). Results are expressed as beta coefficients indicating the difference in test score or growth metric associated with a 10-SY increase in sealant/PRR exposure, or adjusted mean value for each SY category, with standard errors (SE). For menarche, results are expressed as hazard ratio for time to menarche. The complete case analysis for psychosocial and neuropsychological measures had 80% power at alpha = 0.05 to detect a correlation of at least 0.14 between composites exposure level and changes in test scores. For physical development outcomes, the analysis had 80% power to detect a minimum correlation of 0.20 among boys and 0.16 among girls. For analyses of menarche, the subsample of 113 girls achieved 80% power to detect a regression coefficient equal to 0.0149 (hazard ratio of 1.015 for positive associations or 0.985 for inverse associations). All analyses were conducted using SAS v.9.3 (Cary, NC) at alpha=0.05.

Table 3.

Association between Flowable Sealants and Preventive Resin Restorations (PRRs) and Psychosocial Function Global Scores^*

	Sealants and PRRs, total surface-years				10 surface-years βeta (SE), P-value
		Tertile 1	Tertile 2	Tertile 3	Total Sealant/PRR		Sealants		PRRs
	0	0.1–33.9	34–48.9	≥49
BASC Year 5 T-Scores, adjusted mean (SE)
No. children	18	134	127	129
Surface-years, median	0	20.5	40.5	59.5	β (SE)	P	β (SE)	P	β (SE)	P
Emotional symptoms index†	47.8 (2.5)	49.0 (1.6)	47.8 (1.6)	47.3 (1.5)	−0.4 (0.2)	0.06	−0.3 (0.2)	0.21	−0.5 (0.4)	0.24
Clinical maladjustment‡	47.1 (2.7)	48.6 (1.7)	47.1 (1.7)	46.9 (1.7)	−0.4 (0.2)	0.09	−0.2 (0.2)	0.38	−0.7 (0.4)	0.11
School maladjustment§	51.7 (3.0)	54.2 (1.9)	53.4 (1.9)	51.9 (1.8)	−0.3 (0.3)	0.29	−0.2 (0.3)	0.58	−0.5 (0.5)	0.35
Personal adjustment¶	49.3 (2.5)	48.1 (1.6)	48.5 (1.6)	49.5 (1.5)	0.3 (0.2)	0.16	0.2 (0.2)	0.37	0.3 (0.4)	0.40
Child Behavior Checklist 5-Year Change Scores, adjusted mean (SE)
No. children	11	114	111	106
Surface-years, median	0	20.5	40.5	60
Competence#	−3.4 (3.7)	−0.8 (2.4)	−2.0 (2.4)	0.1 (2.4)	0.2 (0.3)	0.56	0.1 (0.3)	0.71	0.2 (0.5)	0.65
Total Problem Behaviors**	3.8 (3.5)	−2.9 (2.3)	−1.9 (2.3)	−2.9 (2.3)	−0.2 (0.3)	0.57	−0.2 (0.3)	0.50	−0.1 (0.5)	0.84
Core Syndromes
Withdrawn	0.9 (1.9)	−2.2 (1.3)	−1.9 (1.3)	−2.0 (1.2)	−0.1 (0.2)	0.49	−0.1 (0.2)	0.40	0.1 (0.3)	0.83
Anxious/Depressed	−0.1 (2.0)	−1.8 (1.3)	−1.5 (1.3)	−1.9 (1.3)	−0.1 (0.2)	0.70	−0.2 (0.2)	0.26	0.3 (0.3)	0.23
Delinquent Behaviors	0.6 (2.2)	−1.7 (1.4)	−1.9 (1.4)	−1.7 (1.4)	−0.1 (0.2)	0.60	−0.1 (0.2)	0.45	−0.1 (0.3)	0.76

^*Multivariable generalized linear models adjusted for age, sex, race/ethnicity (white, black, Hispanic, or other), socioeconomic status (continuous score), geographic study region (rural or urban), baseline blood lead level, composite restorations (continuous surface-years), primary care giver’s marital status (married/living with partner, separated/divorced/widowed, or single), and maternal alcohol/tobacco/drug exposure during pregnancy (yes/no).

^†Emotional symptoms index is a global indicator composed of 2 scales from Clinical Maladjustment (Anxiety and Social Stress), 2 from Personal Maladjustment (Interpersonal Relations and Self-Esteem), and 2 from no other domain (Depression and Sense of Inadequacy). Higher scores indicate more problems.

^‡Clinical Maladjustment summary score is composed of Anxiety, Social Stress, Atypicality, Locus of Control, and for ages >=12y, Somatization. Higher scores indicate more problems.

^§School Maladjustment summary score is composed of Attitude to School, Attitude to Teachers, and for ages >=12 y, Sensation Seeking. Higher scores indicate more problems.

^¶Personal Adjustment summary score is composed of Interpersonal Relations, Relations with Parents, Self-esteem, and Self-reliance. Lower scores indicate more problems.

^#Competence summary score is composed of Activities, Social Adaptation, and School competence subscales. Lower scores indicate more problems (i.e., decreased competence). These scores do not contribute to the Total Problem Behaviors score from the CBCL.

^**Total Problem Behaviors score is composed of Internalizing and Externalizing Problems scales, as well as 4 core syndromes: Social Problems, Thought Problems, Attention Problems, and Sex Problems. Higher scores indicate more problems on these global and core syndrome scores.

Abbreviations: BASC=Behavior Assessment for Children; PRR= Preventive resin restoration; SE=standard error.

Table 5.

Adjusted Mean (SE) Changes in Body Mass Index Z-Score and Body Fat Percentage during 5-Year Follow-Up, by Total Cumulative Exposure to Flowable Sealants and Preventive Resin Restorations^*

Males	Sealants and PRRs, total surface-years category
		Tertile 1	Tertile 2	Tertile 3
	0	0.1–29.9	30.0–45.9	≥46.0
No. males	9	69	61	75
Surface-years, median	0	19.5	39.5	55.0
BMI z-score	−0.02 (0.23)	0.08 (0.1)	−0.17 (0.1)	−0.04 (0.09)
Body fat percentage	6.21 (2.84)	3.82 (1.21)	3.70 (1.23)	3.94 (1.12)
Females	0	0.1–23.9	24–35.9	≥36
No. females	13	77	83	78
Surface-years, median	0	15.0	39.0	60.5
BMI z-score	0.17 (0.18)	0.25 (0.08)	0.30 (0.07)	0.24 (0.08)
Body fat percentage	6.43 (2.23)	8.40 (0.96)	7.69 (0.88)	7.50 (0.96)

^*No statistically significant associations (P>0.05) between sealant/PRR exposure and body mass or body fat changes. Table presents multivariable-adjusted mean changes over follow-up. All models were adjusted for age and either baseline BMI z-score (for BMI models) or height and baseline body fat percentage (for body fat models). For boys, models additionally were adjusted for household income (<$20k, $20–40k, or >$40k) and race/ethnicity. For girls, models were additionally adjusted for daily servings of fruits and vegetables.

RESULTS

The mean (SD) age of study participants was 7.4 (1.3) years at baseline, and most had mixed primary and permanent dentition. Table 1 presents baseline sociodemographic characteristics by category of the total cumulative received sealant/PRR by the end of the study.

Table 1.

Baseline Characteristics of Participants, Overall and by Total Cumulative Exposure (Surface-years) to Flowable Composites (Sealants or Preventive Resin Restorations) during the 5-Year Trial

	Total	By Sealant/PRR Exposure Level (Surface-years category)
			Tertile among exposed
			Tertile 1	Tertile 2	Tertile 3
		0	0.1–33.9	34–48.9	≥49
N	450	39	136	138	137
Age, mean (SD)	7.4 (1.3)	7.4 (1.2)	7.2 (1.4)	7.2 (1.2)	7.9 (1.3)
Sex, n (%)
Female	240 (53.3%)	22 (56.4%)	72 (52.9%)	77 (55.8%)	69 (50.4%)
Male	210 (46.7%)	17 (43.6%)	64 (47.1%)	61 (44.2%)	68 (49.6%)
No. of carious teeth, mean (SD)	5.3 (2.9)	6.1 (2.9)	5.9 (3.2)	5.1 (2.6)	4.9 (2.6)
No. of carious surfaces, mean (SD)	9.4 (6.7)	11.9 (7.2)	10.9 (7.8)	9.0 (6.2)	7.5 (5.0)
No. of sealed surfaces, mean (SD)*	4.0 (3.9)	0	2.2 (2.4)	4.5 (3.5)	6.5 (4.3)
No. children with composite fillings	15	2	3	6	4
No. of composite fillings, mean (SD)	0.1 (0.5)	0.1 (0.5)	0.1 (0.4)	0.1 (0.7)	0.1 (0.5)
Race/ethnicity, n (%)†
Non-Hispanic White	291 (64.7%)	20 (51.3%)	100 (73.5%)	93 (67.4%)	78 (56.9%)
Non-Hispanic Black	87 (19.3%)	8 (20.5%)	22 (16.2%)	25 (18.1%)	32 (23.4%)
Hispanic (non-mixed)	28 (6.2%)	3 (7.7%)	3 (2.2%)	9 (6.5%)	13 (9.5%)
Other	44 (9.8%)	8 (20.5%)	11 (8.1%)	11 (8.0%)	14 (10.2%)
Socioeconomic status, n (%)‡
Low	124 (27.6%)	15 (38.5%)	33 (24.3%)	35 (25.4%)	41 (29.9%)
Medium	209 (46.4%)	16 (41.0%)	70 (51.5%)	58 (42.0%)	65 (47.5%)
High	117 (26.0%)	8 (20.5%)	33 (24.3%)	45 (32.6%)	31 (22.6%)
Geographic Location, n (%)
Urban (Boston, Mass.)	230 (51.1%)	21 (53.9%)	53 (39.0%)	77 (55.8%)	79 (57.7%)
Rural (Farmington, Maine)	220 (48.9%)	18 (46.2%)	83 (61.0%)	61 (44.2%)	58 (42.3%)
Drinking water source, n (%)
Bottled	119 (30.3%)	4 (30.8%)	33 (26.4%)	35 (26.9%)	47 (37.6%)
Tap	153 (38.9%)	5 (38.5%)	61 (48.8%)	50 (38.5%)	37 (29.6%)
Mixed	113 (28.8%)	4 (30.8%)	31 (24.8%)	42 (32.3%)	36 (28.8%)
Don’t know	8 (2.0%)	0 (0.0%)	0 (0.0%)	3 (2.3%)	5 (4.0%)
Fruits and vegetables servings/day, mean (SD) ¶	1.3 (0.6)	1.3 (0.5)	1.3 (0.7)	1.2 (0.5)	1.3 (0.6)
Blood lead level, mean (SD) mg/dL¶	2.3 (1.8)	2.6 (2.6)	2.2 (1.4)	2.4 (1.7)	2.4 (2.0)
Birth weight, mean (SD) g¶	3340 (543)	3152 (484)	3415 (569)	3327 (517)	3329 (549)
Body fat %, mean (SD)	22.7 (10.4)	21.5 (9.0)	20.6 (9.2)	22.8 (10.4)	24.8 (11.6)
Body mass index, mean kg/m2 (SD)	18.0 (3.9)	17.3 (2.9)	17.3 (3.1)	17.9 (4.2)	18.9 (4.2)
Body mass index-by-age Z-score (SD)	0.6 (1.1)	0.1 (1.2)	0.4 (1.1)	0.6 (1.1)	0.7 (1.1)

^*Sealed with either preventive sealant or preventive resin restoration (PRR).

^†Race/ethnicity was self-reported by the parent of the child. The “other” category included individuals who identified themselves as Asian, Native American, multiracial (specified) or other (specified).

^‡Socioeconomic status index was calculated using household income and education level of the primary caregiver and standardized to the U.S. population.⁶²

^¶Three children were missing baseline data on marital status of primary caregiver, fruit/vegetable intake, and gum chewing frequency. Seven children were missing data on baseline blood lead level. Forty-one children were missing data on birth weight, which was self-reported by the parent.

During the 5-year follow-up, most children received sealants (87 percent) or PRRs (59 percent), with 61 percent receiving both. Table 2 describes the sealant/PRR treatment levels received during the study among all subjects. At study entry, 39 (7.3 percent) children had sealants or PRRs in place (mean 4.3 surfaces); 37 of these 39 children had additional sealants or PRRs placed during the study. There were no significant differences in levels of sealant/PRR treatment by assigned restoration treatment group (composite or amalgam). As expected, there was an inverse correlation between sealant placement and restoration treatment; children with more sealants placed had fewer restorations during follow-up (Pearson correlation coefficient = −0.18, P<0.001). In contrast, PRR treatment levels were not correlated with restorative fillings (P=0.8).

Table 2.

Placement of Flowable Composites (Sealants and Preventive Resin Restorations) at the Initial Study Visits and during 5-Year Follow-up

	Sealant	PRR	Total (Sealant and/or PRR)
No. sealed surfaces, mean (SD)
Initial Study Visit	3.3 (3.6)	0.9 (1.7)	4.2 (3.8)
Final Year 5 Visit (present or placed at final visit)	8.0 (6.0)	2.0 (2.6)	10.3 (6.6)
All Visits (ever treated)	10.9 (8.2)	2.7 (3.3)	13.6 (8.9)
Surface-years of exposure by end of trial	29.5 (20.5)	7.1 (10.0)	37.7 (22.0)
No. (%) children ever having the material placed during the trial	397 (88.2%)	290 (64.4%)	411 (91.3%)

There were no associations between sealant and/or PRR surface-years exposure and the primary outcomes of global behavior scores at the end of follow-up (Table 3). Results were similar in additional analyses (data not shown) examining primary and permanent teeth separately, posterior-occlusal surfaces, excluding children who received composite restorations, or excluding children with pre-existing sealants/PRRs. Analyses of scores for specific core syndromes (e.g., anxiety, depression, sensation seeking, interpersonal relations) showed similar results of no consistent associations with sealant or PRR exposure. We furthermore examined cut-points of clinically meaningful scores (normal, at-risk, or clinically significant) and found no consistent or significant associations with sealants/PRR. Lastly, there were no interactions between sealant/PRR exposure and age, study site, or gender.

Neuropsychological test scores over follow-up were not associated with tertile of sealant or PRR treatment levels (Table 4). Compared to children who ever received any sealants or PRRs, the minority of children receiving no sealants or PRRs during the study (<6 percent of children) had distinctly worse 5-year changes in WISC-III IQ (e.g., mean 1.7 point decrease vs. 2.2–3.5 point increase) and WIAT (e.g., mean 4.7 point decrease vs. 1.3–1.9 point decrease), yet a smaller increase in errors on the executive function Wisconsin Card Sorting Test (e.g., mean total errors 3.8 vs. 17.2–17.8 increase). Thus, linear tests for trend including both non-exposed and exposed children were not appropriate. No statistically significant associations or trends were observed for test score changes among those with any sealant/PRR treatment. The sensitivity analyses confirmed these findings.

Table 4.

Adjusted Mean (SE) Changes in Neuropsychosocial Test Scores between Baseline and 4- or 5-Year Follow-Up, by Total Cumulative Exposure to Flowable Sealants and Preventive Resin Restorations^*

	Sealants and PRRs, total surface-years category
5-Year Change Score Tests	Year 5
		Tertile 1	Tertile 2	Tertile 3
	0	0.1–33.9	34.0–48.9	≥49.0
No. children	23	136	138	137
Surface-years, median	0	21.3	40.5	59.5
WISC-III Full-Scale IQ	−1.7 (2.1)	3.5 (0.8)	2.2 (0.8)	2.6 (0.8)
WISC-III Factors
Verbal Comprehension	−1.0 (0.8)	0.2 (0.3)	0.0 (0.3)	−0.1 (0.3)
Perceptual Organization	−3.4 (2.7)	3.7 (1.0)	3.0 (0.9)	3.2 (1.0)
Freedom from Distractibility	1.4 (3.0)	3.9 (1.1)	1.2 (1.0)	4.5 (1.1)
Processing Speed	0.9 (3.4)	5.7 (1.2)	6.2 (1.2)	6.6 (1.2)
WIAT
Reading	−4.7 (2.7)	−1.9 (1.0)	−1.9 (1.0)	−1.3 (1.0)
Mathematics	−5.3 (3.0)	−0.7 (1.1)	−4.0 (1.0)	−3.2 (1.1)
4-Year Change Scores Tests	Year 4
		Tertile 1	Tertile 2	Tertile 3
	0	0.1–23.9	24.0–35.9	≥36.0
No. children	26	127	148	136
Surface-years, median	0	14.5	30.0	44.5
Trail-Making Test
Part B - Part A	−24.7 (13.2)	−33.4 (5.0)	−30.7 (4.4)	−36.8 (4.9)
Verbal Fluency
Letter fluency	13.5 (2.0)	13.7 (0.8)	12.3 (0.7)	12.9 (0.8)
Stroop Color-Word Interference
Color	15.7 (3.0)	18.2 (1.1)	19.3 (0.9)	19.3 (0.9)
Color-Word	13.9 (2.4)	13.8 (0.9)	12.5 (0.7)	13.6 (0.8)
Wisconsin Card Sorting Test
No. of categories achieved	0.8 (0.4)	1.1 (0.2)	1.1 (0.1)	1.1 (0.1)
Total perseverative errors	7.9 (4.0)	19.0 (1.6)	20.3 (1.4)	20.4 (1.6)
Total errors	3.8 (4.1)	17.2 (1.6)	17.8 (1.4)	17.6 (1.6)

^*No statistically significant associations (P>0.05) between sealant/PRR exposure and neuropsychological test score changes among children in tertiles 1–3. Means are adjusted for age, sex, socioeconomic status (continuous score), geographic study region (urban or rural), baseline blood lead level, and composite restorations (surface-years).

Abbreviations: WISC-III=Wechsler Intelligence Scale for Children-Third Edition. WIAT=Wechsler Individual Achievement Test

During the 5-year study, the prevalence of children whose BMI-for-age category was overweight or obese increased for boys from 35 percent at baseline to 43 percent, and for girls from 32 percent at baseline to 47 percent. There were no significant associations between sealant/PRR exposure levels and changes in BMI or body fat percentage for girls or boys (Table 5). Among girls, menarche occurred slightly later with greater exposure to sealants/PRRs (age-adjusted hazard ratio=0.91, 95% CI 0.83, 1.01; P-value=0.08).

DISCUSSION

This prospective cohort analysis of children’s exposure to flowable composites found no associations between dental sealant and PRR exposure levels and behavioral, neuropsychological, or physical development over 5-years follow-up. To our knowledge, this is the first analysis of these health outcomes in relation to dental sealant materials in children. Over eight million children in the U.S. have dental sealants, with sealants prevalence increasing across age, race, sex, and poverty level subgroups.⁴⁴ Approximately 10–25 percent of sealants have been shown to completely fail or have partial retention within 2–3 years of placement, though sealant longevity depends on patient characteristics, sealant material type and placement techniques.^{2, 45–48} Laboratory studies have shown that monomers used in sealants are released during placement,^49–53 and it is plausible that continued low amounts leach over time with chemical and mechanical interactions in the oral environment.^7,54 These compounds have been shown to cause various cytotoxic, embryotoxic, genotoxic and endocrine effects over a range of concentrations in vitro and animal experiments.^{9–13, 55–57} Thus, a thorough evaluation and assessment of the long-term safety of these materials in standard use in preventive dentistry is important for dental and general public health.

We had previously shown associations of bisGMA-based dental composite with psychosocial health outcomes in the NECAT participants. In the current analysis of bisGMA-based sealants, most estimates, except for menarche,³² were in the opposite direction of estimates for composite.^{31, 33} Although this may be an artifact of the inverse correlation between composites and sealants, the correlation between sealants and composite was relatively small. It is possible that non-dental factors, such as residual confounding by unmeasured socioeconomic factors or oral health attitudes related to having sealant placement, contributed. However, without relevant biomarker data, it is difficult to explain the lack of similar findings between composite and sealant materials. That is, it is possible that the different results for the composite and the sealant were due to differences in material manufacturing, including the precise monomers and other components used, and/or the resulting integrity (i.e., degradation) of the materials once in the mouth and exposed to the oral environment. Flowable composites have higher resin content and in general, inferior mechanical properties, compared to standard hybrid composites.⁵⁸ However, standard non-flowable composites that are placed in bulk and undergo bulk polymerization could be subject to less complete curing,⁵⁹ hence more exposure to resin monomers. Previous studies of leaching from dental material components have shown increased saliva levels of BPA and related compounds within a few hours after treatment, but the extent of leaching over time as materials degrade is unknown.^23,27 If negligible leaching occurred after the initial exposure during placement, this might be a reason for the lack of association with the outcomes measured in this study. Thus, a limitation of NECAT for these secondary data analysis was that urinary concentrations of BPA, bisGMA, TEGDMA, HEMA, or other compounds that may leach from sealants were not measured. Biomarker data would help elucidate whether differences in results were due to differences in the internal dose of chemical exposures, or to artifacts of unknown or unmeasured factors associated with dental treatment. Additional studies that include biomarker measurement along with prospective longitudinal data on dental treatment and health outcomes are needed for a comprehensive examination of the long-term safety of dental materials.

For this secondary data analysis, we combined data on two different materials used for sealants and PRRs, to evaluate total flowable composites exposure. Although both materials contain bisGMA, differences in mechanical properties could lead to differences in leaching of components. Considering that this study was not designed to investigate sealant materials, we conducted numerous sensitivity analyses to assess the robustness of the findings, e.g., excluding subgroups of children (those with pre-existing sealants, those not receiving any sealants) and examining strata of age, sex, or race; results consistently showed no associations with these psychosocial, neuropsychological, or growth outcomes. The minority of children who did not receive sealants had, on average, worse change compared to children who followed routine study procedures to receive any sealants. This difference may be due to the small sample size of children without sealants, or to unmeasured factors related to not receiving sealants. For that reason, we focused our interpretation of results on categories of sealant exposure, particularly among those ever treated (i.e., tertiles of exposure). Overall, these results showed no effects from sealants, and furthermore confirmed the caries prevention benefits of sealant placement.

While sealant level was not statistically significantly associated with age of menarche, the direction indicating later menarche was consistent with our previous findings for randomization to composite restoration treatment.³² In animal experiments, BPA alters reproductive development, often accelerating onset of puberty, but also causes altered reproductive parameters such as number of estrous cycles and days of estrus ⁶⁰ and impaired follicle formation.⁶¹ Our analysis raises the question of possible effect on later age of menarche, but it is not sufficient to draw any conclusions. The menarche analyses were exploratory due to limitations in the sample (i.e., a subset of girls from one study site) and follow-up age window, in that not all girls had reached menarche within the five year study period. Furthermore, although children randomized to composite restorations had later menarche compared to those randomized to amalgam, there was no association between composite restoration treatment levels and age of menarche.³²

Strengths of the study include that the data were obtained as part of a randomized clinical trial. The dental procedures, including materials used and placement procedures, and the treatment levels were strictly monitored for the trial. In addition, health assessments for the outcomes analyzed here were conducted using validated and routinely calibrated measures. This analysis was well-powered to detect associations between sealants and the health outcome measures. Given the eligibility criteria of two or more caries, the trial did not include a U.S. population-based representative sample of children; however, its focus on children in need of dental treatment resulted in a sample of children who represent the specific targets of sealant programs, i.e., children at high risk of caries.

In summary, this analysis does not provide evidence to alter current practice regarding dental sealants. It remains uncertain whether the association between composite restorations and adverse psychosocial outcomes was caused by higher long-term release of compounds from restorative composite compared to preventive sealants, by the particular combination of compounds in the composite, or by some unknown confounding factor or chance difference between composite and amalgam treatment groups in the clinical trial. Future research should include biomarker measures along with treatment and longitudinal health outcomes data.

Conclusions

Based on the results of the secondary data analysis reported here, the following conclusions can be made from this study:

1. Cumulative treatment levels of bisGMA-containing dental sealants and PRRs were not associated with psychosocial health or behavior problems in children or adolescents after 5 year follow-up.

2. Cumulative treatment levels of bisGMA-containing dental sealants and PRRs were not associated with changes in neuropsychological test scores during the 5 year study.

3. Cumulative treatment levels of bisGMA-containing dental sealants and PRRs were not associated with changes in body composition or body mass index during the 5 year study.

4. No evidence was provided to suggest alterations to standard procedures and materials used for preventive dental sealants or PRRs in children.

Abbreviations

BADGE

Bisphenol A diglycidyl ether

bisGMA

Bisphenol A-Glycidyl Methacrylate

BPA

Bisphenol A

MANOVA

Multivariate analysis of variance

NECAT

New England Children’s Amalgam Trial

PRR

Preventive resin restoration

TEGDMA

Triethylene glycol dimethacrylate

WISC-III

Wechsler Intelligence Scale for Children-Third Edition

WIAT

Wechsler Individual Achievement Test