OBJECTIVES:

To explore the relative contributions of genetic and environmental influences on dental caries risk and to investigate fetal and developmental risk factors for dental caries.

METHODS:

We recruited children from 250 twin pregnancies midgestation and collected demographic, health, and phenotypic data at recruitment, 24 and 36 weeks’ gestational age, birth and 18 months, and 6 years of age. 25-hydroxyvitamin D was quantified in mothers at 28 weeks’ gestation and in infants at birth. Dental caries and enamel defects were measured at six years of age. We compared concordance for the presence of any caries and advanced caries in monozygotic and dizygotic twin pairs. To investigate environmental risk factors for caries, we fitted multiple logistic regression models using generalized estimating equations to adjust for twin correlation.

RESULTS:

A total of 345 twins underwent dental assessment, with 111 (32.2%) showing signs of any caries and 83 (24.1%) having advanced caries. There was no evidence of higher concordance in monozygotic twins compared with dizygotic twins, with a difference of 0.05 (95% confidence interval −0.14 to 0.25; P = .30) and 0.00 (95% confidence interval −0.26 to 0.26; P = .50) for any caries and advanced caries, respectively, suggesting that environmental factors, rather than genetics, are the predominant determinant of caries risk. After adjusting for potential confounders, lack of community water fluoridation, hypomineralized second primary molars, dichorionic placenta, and maternal obesity were associated with caries.

CONCLUSIONS:

Environmental rather than genetic factors drive dental caries risk and arise as early as prenatal life.

What’s Known on This Subject:

Understanding of early-life risk factors for dental caries is limited by lack of prospective studies that adequately control for confounding. Caries is believed to be highly heritable, but genetic studies to date have failed to identify strong associations.

What This Study Adds:

Robust exposure data and comprehensive statistical methods were used to identify potentially modifiable environmental risk factors from the prenatal period onward. Environment exposures may be more important than genetics in determining dental caries risk in children.

Oral health is an integral part of general health. In addition to impacting the ability to eat, speak, smile, grow, and learn, oral diseases have been linked with noncommunicable diseases (NCDs) such as cardiovascular disease and diabetes.1 Despite advances in prevention and management, 60% to 90% of schoolchildren worldwide experience dental caries, potentially resulting in pain, infection, and hospitalization.2 Toothache may result in school absence, poor nutrition, compromised growth and development, and impaired quality of life for the child and their caregivers.1 Furthermore, childhood dental caries is the strongest indicator of future poor oral health, and therefore caries prevention in children is important for oral health outcomes in adulthood.3 

Dental caries is a dynamic process that occurs when demineralization of dental hard tissues, triggered by a sugar-driven dysbiosis of the dental plaque microbiome, overwhelms remineralization by protective factors in the mouth.4 Initial demineralization of tooth enamel is often subclinical but can lead to the development of carious lesions, which range from incipient areas of increased enamel opacity and porosity to frank cavitation.

Genetic factors are linked to dental caries,5 but the implications of these associations for caries risk, both at individual and population levels, have not been widely considered. Studying familial (including twin) aggregation of complex conditions, such as dental caries, is a valuable tool for further optimizing prevention when both genetic and environmental factors are likely to be important.6 Monozygotic twins share all genetic variation, whereas dizygotic twins, like siblings, share 50% on average. Comparison of concordance in monozygotic and dizygotic pairs can help determine the influence of shared, nonshared, and genetic factors to the variation in risk of dental caries.7 

Consistent with the developmental origins of health and disease paradigm, early-life exposures may increase caries risk through biological early-life programming.8 Reframing prevention of dental caries in terms of developmental origins of health and disease may provide novel targets for prevention before disease onset but are hampered by a lack of prospective studies.

In this prospective longitudinal twin cohort, we aimed to investigate fetal and developmental risk factors for dental caries and the relative contributions of environmental and genetic influences on dental caries risk.

The Peri/postnatal Epigenetic Twins Study (PETS) is a longitudinal birth cohort of 250 mothers and their twin children established in 2007.9 Women, pregnant with twins, were recruited midgestation. Questionnaires regarding maternal prepregnancy weight, illness (including infection), medication use, stress, alcohol intake, and smoking were collected at 3 time points during pregnancy (Supplemental Fig 2). We determined the socioeconomic status (SES) of participants at birth by linking to the Index of Relative Socio-Economic Disadvantage, one of the Socio-Economic Indexes For Areas (SEIFA) developed by the Australian Bureau of Statistics on the basis of census data, via postal codes.10 The children were reviewed (and hospital records accessed) in the immediate postnatal period to obtain obstetric, birth anthropometric, and neonatal data. Chorionicity (ie, whether twins had separate [dichorionic] or shared [monochorionic] placentas) was determined from ultrasound scans and placental examination at delivery. All different-sex twins were assumed to be dizygotic. We determined zygosity for all same-sex twins using 12-marker microsatellite polymerase chain reaction with DNA from umbilical cord and/or buccal samples when available.11 

A total of 244 twin pairs were reviewed at age 18 months, and data on breastfeeding duration, illnesses (including infection, infantile eczema, asthma, and food allergy), hospitalization, and medication use were obtained from parents. At age 6 years, dental examinations were performed, and data were collected on dietary sugar intake (Dietary Sugar Intake section of the Supplemental Information) and oral hygiene. Access to community water fluoridation was determined by using residential postal codes.

We determined 25-hydroxyvitamin D levels from serum collected from mothers at 28 weeks’ gestation and from twin offspring at birth from serum or plasma from cord blood. For all samples, 25-hydroxyvitamin D levels were determined by using the LIAISON 25 OH Vitamin D TOTAL kit (DiaSorin; Saluggia, Vercelli, Italy). A subset of newborn serum samples were analyzed in 2011, and the remaining available serum and plasma samples were analyzed in 2017, with appropriate adjustment for batching effects between (1) different samples type and (2) measurements at different time points (Vitamin D Levels at Birth section of the Supplemental Information).

Ethics approval was obtained from the Royal Children’s Hospital Human Research Ethics Committee (33174 A), and informed consent was obtained from a parent or guardian.

Dental examinations were performed on-site at the research facility or, for participants unable to travel, at home by 2 trained and calibrated oral health professionals (M.J.S. and P.L.) (Dental Examinations section of the Supplemental Information, Supplemental Tables 6 and 7). We recorded dental caries using the International Caries Detection and Assessment System (ICDAS), which allows for quantification of carious lesions, from early to large cavitated lesions with significant destruction of tooth structure. The common developmental defect of enamel, hypomineralized second primary molars (HSPMs), was recorded by using standardized criteria.12 

We collected and managed the study data using Research Electronic Data Capture tools13 and analyzed data using Stata 15 (Stata Statistical Software: Release 15; Stata Corp, College Station, TX). Two binary outcome variables, the presence or absence of (1) any caries (including noncavitated lesions and/or past treatment) and (2) advanced caries, (established carious lesions with ICDAS codes 4–6 and/or past treatment) were derived from the ICDAS index (Methods section of the Supplemental Information). For each outcome variable, we classified twin pairs as concordant (both children affected) or discordant (1 twin affected). To explore the role of genetic and unmeasured environmental factors, we estimated and compared casewise concordances with 95% confidence intervals (CIs) for monozygotic and dizygotic twins. To estimate similarities for monozygotic and dizygotic twins after adjusting for known risk factors, we fitted a multiple logistic regression model using generalized estimating equations (GEEs) (Data Analysis section of the Supplemental Information). To investigate associations between environmental risk factors and presence of any or advanced caries, we fitted logistic regression models using GEEs to adjust for twin correlation. The associations were reported as odds ratios (ORs) with 95% CI.

We selected biologically plausible exposure variables from data collected during pregnancy, at birth, and at 18 months of age on the basis of previous evidence of association with dental caries. Exposure variables with P < .1 in simple regression models were included in the final multiple regression models to adjust for confounding. As an exploratory study, we adopted an inclusive approach to model building, aiming to identify potential factors rather than exclude factors.

Within-pair analyses (to explore the role of categorical nonshared risk factors, such as early-life hospitalization and antibiotic use, in caries for discordant twin pairs) could not be performed because there were too few twin pairs discordant for both outcome and exposure. To determine if observed associations with nonshared continuous variables were likely to be causal or due to unmeasured shared or unshared factors, we fitted within- and between-pair models using GEEs.14 

A total of 345 twin children (69% of the original cohort) participated in the dental examinations (Fig 1). One child was uncooperative with the caries assessment, resulting in 172 complete twin pairs (101 dizygotic and 71 monozygotic pairs). Most children (n = 277; 80.3%) were aged 6 or 7 years (range: 6–9 years), and 185 (53.6%) were girls. The mean SEIFA for the cohort was 1014.3 (SD 57.9), indicating that the sample had a higher SES and less variation than the Australian average.15 

FIGURE 1

Study retention from recruitment to dental examinations at 6 years of age.

FIGURE 1

Study retention from recruitment to dental examinations at 6 years of age.

Close modal

A total of 111 (32.2%) children had any caries, with a mean of 3.0 teeth (median = 2) affected. Thirty-nine twin pairs were concordant and 33 pairs were discordant for the presence of any caries. A total of 83 (24.1%) children had advanced caries, with those affected having a mean of 2.8 teeth (median = 2) with advanced caries. Twenty-six twin pairs were concordant and 31 twin pairs were discordant for advanced caries, with the twin who was unaffected having either no caries or only early caries (ICDAS caries codes 1–3).

The overall concordance for any caries was 0.70 (95% CI 0.61 to 0.80). There was no evidence of higher unadjusted concordance in monozygotic twins (0.74; 95% CI 0.58 to 0.89) compared with dizygotic twins (0.69; 95% CI 0.56 to 0.81), with a difference of 0.05 (95% CI −0.14 to 0.25; P = .30). The overall concordance for advanced caries was 0.63 (95% CI 0.51 to 0.75; Supplemental Fig 3). There was no evidence of higher unadjusted concordance in monozygotic twins (0.63; 95% CI 0.43 to 0.82) compared with dizygotic twins (0.63; 95% CI 0.47 to 0.78), with a difference of 0.00 (95% CI −0.26 to 0.26; P = .50). A logistic regression GEE model adjusted for known risk factors revealed no differences in concordance between monozygotic and dizygotic twins (Results section of the Supplemental Information).

Maternal stress during pregnancy, cord attachment, smoking beyond the first trimester of pregnancy, age at examination, home visit for examination, nonfluoridated town water, and the presence of HSPMs were all associated (P < .1) with any caries in unadjusted regression models (Table 1). The evidence for an association between covariates HSPMs (OR 2.15; 95% CI 1.04 to 4.47; P = .04), nonfluoridated town water (OR 5.98; 95% CI 1.59 to 22.55; P = .01), and monochorionicity (OR 0.37; 95% CI 0.17 to 0.78; P = .01) and the outcome (any caries) did not attenuate after adjusting for confounding (Table 2, Supplemental Table 11).

TABLE 1

Simple (Unadjusted) Logistic Regression for Risk Factors for Any and Advanced Caries

FactorAny CariesAdvanced Caries
OR (95% CI)POR (95% CI)P
Age     
 Q1 (n = 70) 6.00–6.28 Reference .01 Reference .035 
 Q2 (n = 69) 6.28–6.43 0.58 (0.22 to 1.50) — 0.46 (0.17 to 1.25) — 
 Q3 (n = 68) 6.43–6.66 1.75 (0.72 to 4.24) — 1.39 (0.56 to 3.49) — 
 Q4 (n = 70) 6.68–7.00 0.74 (0.30 to 1.81) — 0.58 (0.21 to 1.62) — 
 Q5 (n = 68) 7.02–9.05 2.50 (0.99 to 6.30) — 1.81 (0.72 to 4.55) — 
Dizygotic twin (n = 203) 1.53 (0.85 to 2.77) .16 1.14 (0.62 to 2.14) .66 
Female sex (n = 185) 0.73 (0.47 to 1.15) .17 0.6 (0.37 to 0.97) .04 
SEIFA (per 100 U) (n = 343) 1.19 (0.69 to 2.03) .53 1.05 (0.58 to 1.91) .87 
Maternal obesity (n = 44) 1.56 (0.70 to 3.46) .27 2.02 (0.90 to 4.58) .09 
Maternal stress score     
 Q1 (n = 72) 3–16 Reference .04 Reference .18 
 Q2 (n = 66) 17–21 0.65 (0.26 to 1.65) — 0.70 (0.27 to 1.80) — 
 Q3 (n = 62) 22–24 0.41 (0.15 to 1.10) — 0.39 (0.13 to 1.14) — 
 Q4 (n = 80) 25–30 0.76 (0.33 to 1.77) — 0.65 (0.27 to 1.59) — 
 Q5 (n = 49) 31–45 1.92 (0.78 to 4.73) — 1.49 (0.56 to 3.92) — 
Maternal infection during pregnancy (n = 205) 1.00 (0.56 to 1.78) .996 1.11 (0.60 to 2.07) .73 
Maternal antibiotic use during pregnancy (n = 66) 1.74 (0.87 to 3.52) .12 1.49 (0.71 to 3.14) .29 
Maternal vitamin D at 28 wk (20 nmol/L) n = 329 1.07 (0.79 to 1.44) .68 1.36 (0.97 to 1.91) .08 
Maternal smoking in second or third trimester (n = 46) 2.17 (1.00 to 4.69) .05 2.07 (1.00 to 4.31) .05 
Maternal alcohol intake during pregnancy (n = 199) 0.80 (0.45 to 1.42) .45 1.15 (0.62 to 2.12) .67 
Gestational age (n = 345), wk 0.93 (0.82 to 1.05) .23 0.97 (0.86 to 1.09) .61 
Vaginal delivery (n = 119) 0.95 (0.52 to 1.72) .86 0.78 (0.41 to 1.47) .44 
Chorionicity     
 Monochorionic (n = 96) 0.36 (0.17 to 0.74) .01 0.47 (0.21 to 1.03) .06 
 Dichorionic (n = 217) Reference — Reference — 
Cord attachment     
 Central cord (n = 141) Reference .05 Reference .19 
 Peripheral cord (n = 122) 1.49 (0.90 to 2.47) — 1.40 (0.81 to 2.40) — 
 Velamentous cord (n = 37) 0.68 (0.35 to 1.30) — 0.64 (0.25 to 1.66) — 
BW (standardized, per unit) (n = 345) 1.10 (0.89 to 1.37) .38 1.08 (0.83 to 1.40) .57 
NICU or SCN (n = 146) 1.48 (0.87 to 2.51) .14 1.21 (0.67 to 2.19) .52 
Birth vitamin D (20 nmol/L) (n = 241) 1.19 (0.87 to 1.63) .28 1.50 (1.04 to 2.15) .03 
Breastfeeding (any) (n = 301) 0.54 (0.21 to 1.35) .19 0.89 (0.31 to 2.59) .83 
Formula feeding (any) (n = 302) 1.07 (0.47 to 2.48) .86 1.39 (0.40 to 4.85) .60 
Hospitalization in first 18 mo (n = 67) 1.10 (0.60 to 2.03) .76 1.14 (0.56 to 2.30) .72 
Infection in first 18 mo (n = 171) 1.19 (0.70 to 2.00) .52 0.90 (0.50 to 1.61) .71 
Antibiotics in first 18 mo (n = 75)a 0.92 (0.54 to 1.57) .75 0.80 (0.41 to 1.57) .52 
Sugar intake     
 Low (n = 83) Reference — Reference — 
 Medium (n = 214) 1.36 (0.71 to 2.57) .35 1.77 (0.83 to 3.76) .14 
 High (n = 38) 1.36 (0.57 to 3.29) .49 2.00 (0.74 to 5.37) .17 
Brushing frequency     
 Twice daily (n = 204) Reference — Reference — 
 Once a day (n = 106) 1.24 (0.64 to 2.41) .52 1.09 (0.56 to 2.13) .79 
 Once every 2–4 d (n = 21) 0.50 (0.13 to 1.89) .31 0.32 (0.08 to 1.33) .12 
No water fluoridation (n = 26) 5.56 (1.83 to 16.93) .003 7.27 (2.31 to 22.85) .001 
Home visit (n = 62) 1.99 (0.96 to 4.14) .07 1.83 (0.86 to 3.90) .12 
HSPM (n = 68) 2.67 (1.47 to 4.86) .001 2.85 (1.43 to 5.67) .003 
FactorAny CariesAdvanced Caries
OR (95% CI)POR (95% CI)P
Age     
 Q1 (n = 70) 6.00–6.28 Reference .01 Reference .035 
 Q2 (n = 69) 6.28–6.43 0.58 (0.22 to 1.50) — 0.46 (0.17 to 1.25) — 
 Q3 (n = 68) 6.43–6.66 1.75 (0.72 to 4.24) — 1.39 (0.56 to 3.49) — 
 Q4 (n = 70) 6.68–7.00 0.74 (0.30 to 1.81) — 0.58 (0.21 to 1.62) — 
 Q5 (n = 68) 7.02–9.05 2.50 (0.99 to 6.30) — 1.81 (0.72 to 4.55) — 
Dizygotic twin (n = 203) 1.53 (0.85 to 2.77) .16 1.14 (0.62 to 2.14) .66 
Female sex (n = 185) 0.73 (0.47 to 1.15) .17 0.6 (0.37 to 0.97) .04 
SEIFA (per 100 U) (n = 343) 1.19 (0.69 to 2.03) .53 1.05 (0.58 to 1.91) .87 
Maternal obesity (n = 44) 1.56 (0.70 to 3.46) .27 2.02 (0.90 to 4.58) .09 
Maternal stress score     
 Q1 (n = 72) 3–16 Reference .04 Reference .18 
 Q2 (n = 66) 17–21 0.65 (0.26 to 1.65) — 0.70 (0.27 to 1.80) — 
 Q3 (n = 62) 22–24 0.41 (0.15 to 1.10) — 0.39 (0.13 to 1.14) — 
 Q4 (n = 80) 25–30 0.76 (0.33 to 1.77) — 0.65 (0.27 to 1.59) — 
 Q5 (n = 49) 31–45 1.92 (0.78 to 4.73) — 1.49 (0.56 to 3.92) — 
Maternal infection during pregnancy (n = 205) 1.00 (0.56 to 1.78) .996 1.11 (0.60 to 2.07) .73 
Maternal antibiotic use during pregnancy (n = 66) 1.74 (0.87 to 3.52) .12 1.49 (0.71 to 3.14) .29 
Maternal vitamin D at 28 wk (20 nmol/L) n = 329 1.07 (0.79 to 1.44) .68 1.36 (0.97 to 1.91) .08 
Maternal smoking in second or third trimester (n = 46) 2.17 (1.00 to 4.69) .05 2.07 (1.00 to 4.31) .05 
Maternal alcohol intake during pregnancy (n = 199) 0.80 (0.45 to 1.42) .45 1.15 (0.62 to 2.12) .67 
Gestational age (n = 345), wk 0.93 (0.82 to 1.05) .23 0.97 (0.86 to 1.09) .61 
Vaginal delivery (n = 119) 0.95 (0.52 to 1.72) .86 0.78 (0.41 to 1.47) .44 
Chorionicity     
 Monochorionic (n = 96) 0.36 (0.17 to 0.74) .01 0.47 (0.21 to 1.03) .06 
 Dichorionic (n = 217) Reference — Reference — 
Cord attachment     
 Central cord (n = 141) Reference .05 Reference .19 
 Peripheral cord (n = 122) 1.49 (0.90 to 2.47) — 1.40 (0.81 to 2.40) — 
 Velamentous cord (n = 37) 0.68 (0.35 to 1.30) — 0.64 (0.25 to 1.66) — 
BW (standardized, per unit) (n = 345) 1.10 (0.89 to 1.37) .38 1.08 (0.83 to 1.40) .57 
NICU or SCN (n = 146) 1.48 (0.87 to 2.51) .14 1.21 (0.67 to 2.19) .52 
Birth vitamin D (20 nmol/L) (n = 241) 1.19 (0.87 to 1.63) .28 1.50 (1.04 to 2.15) .03 
Breastfeeding (any) (n = 301) 0.54 (0.21 to 1.35) .19 0.89 (0.31 to 2.59) .83 
Formula feeding (any) (n = 302) 1.07 (0.47 to 2.48) .86 1.39 (0.40 to 4.85) .60 
Hospitalization in first 18 mo (n = 67) 1.10 (0.60 to 2.03) .76 1.14 (0.56 to 2.30) .72 
Infection in first 18 mo (n = 171) 1.19 (0.70 to 2.00) .52 0.90 (0.50 to 1.61) .71 
Antibiotics in first 18 mo (n = 75)a 0.92 (0.54 to 1.57) .75 0.80 (0.41 to 1.57) .52 
Sugar intake     
 Low (n = 83) Reference — Reference — 
 Medium (n = 214) 1.36 (0.71 to 2.57) .35 1.77 (0.83 to 3.76) .14 
 High (n = 38) 1.36 (0.57 to 3.29) .49 2.00 (0.74 to 5.37) .17 
Brushing frequency     
 Twice daily (n = 204) Reference — Reference — 
 Once a day (n = 106) 1.24 (0.64 to 2.41) .52 1.09 (0.56 to 2.13) .79 
 Once every 2–4 d (n = 21) 0.50 (0.13 to 1.89) .31 0.32 (0.08 to 1.33) .12 
No water fluoridation (n = 26) 5.56 (1.83 to 16.93) .003 7.27 (2.31 to 22.85) .001 
Home visit (n = 62) 1.99 (0.96 to 4.14) .07 1.83 (0.86 to 3.90) .12 
HSPM (n = 68) 2.67 (1.47 to 4.86) .001 2.85 (1.43 to 5.67) .003 

Q, quintile; SCN, special care nursery; —, not applicable.

a

Medications consumed in hospital were not included because these data were not collected.

TABLE 2

The ORs, 95% CIs, and P Values for Factors Found to Be Associated With Any Caries and Advanced Caries After Adjusting for Confounding by Covariates Identified in the Unadjusted Logistic Regression

FactorAny Caries Experience (n = 268)Advanced Caries (n = 265)
Adjusted OR (95% CI)PAdjusted OR (95% CI)P
HSPM 2.16 (1.04 to 4.47) .04 2.43 (1.11 to 5.36) .03 
Nonfluoridated water 5.98 (1.59 to 22.55) .01 6.26 (1.74 to 22.54) .01 
Placenta (monochorionic) 0.37 (0.17 to 0.78) .01 — — 
Maternal obesity — — 2.68 (1.19 to 6.08) .02 
FactorAny Caries Experience (n = 268)Advanced Caries (n = 265)
Adjusted OR (95% CI)PAdjusted OR (95% CI)P
HSPM 2.16 (1.04 to 4.47) .04 2.43 (1.11 to 5.36) .03 
Nonfluoridated water 5.98 (1.59 to 22.55) .01 6.26 (1.74 to 22.54) .01 
Placenta (monochorionic) 0.37 (0.17 to 0.78) .01 — — 
Maternal obesity — — 2.68 (1.19 to 6.08) .02 

—, not applicable.

Maternal and newborn vitamin D levels, chorionicity, maternal obesity, smoking beyond the first trimester of pregnancy, age at examination, sex, nonfluoridated town water, and the presence of HSPMs were all associated (P < .1) with advanced caries in the unadjusted regression models (Table 1). Because maternal and newborn vitamin D levels were highly correlated, only maternal vitamin D was included in the multiple regression model. Nonfluoridated town water (OR 6.26; 95% CI 1.74 to 22.54; P = .01), maternal obesity (OR 2.68; 95% CI 1.19 to 6.08; P = .02), and HSPMs (OR 2.43; 95% CI 1.11 to 5.36; P = .03) were strongly associated with advanced lesions after adjusting for confounding (Table 2, Supplemental Table 12).

A within-pair analysis of birth weight (BW), a continuous variable, and both caries outcome variables failed to reveal any association, with a mean within-pair difference between twins who were affected and unaffected of −58.9 g (95% CI −163.6 to 45.8; P = .62) and −36.1 g (95% CI −152.5 to 80.3; P = .53) in the 33 and 31 pairs discordant for any and advanced caries, respectively.

From the subset of children with vitamin D levels measured at birth, the mean within-pair difference between twins who were affected and unaffected was 0.52 nmol/L (95% CI −4.97 to 6.01; P = .84) in the 21 pairs discordant for any caries and 1.62 nmol/L (95% CI −5.46 to 8.70; P = .63) in the 16 pairs discordant for advanced caries. Within- and between-pair analyses were only applied for advanced caries given the weak association identified in the unadjusted regression model (Table 1). Although there was no evidence of an association between advanced caries and within-pair differences in birth vitamin D levels (OR 0.95; 95% CI 0.38 to 2.36; P = .91), there was strong evidence for an association for between-pair differences (OR 1.91; 95% CI 1.18 to 3.08; P = .01) after adjusting for the effect of HSPMs, maternal obesity, nonfluoridated town water, and chorionicity. Results were robust to the exclusion of outlying and influential observations (Data Analysis section of the Supplemental Information).

This study revealed no difference in concordance between monozygotic and dizygotic twins, indicating that shared and nonshared environmental factors predominate over genetic factors in determining variation in caries risk in children. Dental caries is likely to be a genetically complex phenotype, with small contributions from many loci. These genetic factors may include variants in loci for enamel formation, saliva, immunity, and taste.16,17 Our findings reveal that despite the biological plausibility, genetic factors are relatively less important determinants of caries risk than shared environmental factors. This finding has important clinical implications because the perceived genetic nature of dental caries may lead to a sense of determinism that impedes rather than motivates behavioral change.18 If replicated, findings from our study will help clinicians motivate such change by revealing that caries risk is modifiable.

Previous twin studies of dental caries have been retrospective and have not fully capitalized on the advantages of analyzing twin data, being focused only on the genetic contribution or heritability. Bretz et al19 reported a heritability of 70% for the prevalence rate of surface-based caries in 388 pairs aged 1 to 8 years in a low SES community with nonfluoridated water. Authors of a follow-up study also reported that lesion progression was modestly heritable, with estimates of 30% to 51%, and that heritability of sweet taste and caries was unrelated to early childhood caries.20,21 However, using heritability estimates as low as 30% and as high as 70% to support genetic etiology is potentially misleading. Rather, the great potential of the classic twin model is determining the overall genetic influence on caries risk, compared with environmental factors. Therefore, in our study, associations with genetic factors, although plausible in dental caries, are less relevant caries risk at an individual level compared with environmental factors and, indeed, may distract from addressing modifiable environmental factors.6 Nevertheless, genetic studies may be used to additionally inform mechanistic understanding because environmental exposures operate in a genetic context. In addition, with our study, we can only comment on the relative influence of genetic and environmental factors for the conditions of this study, in particular, for the age range of participants, which was 6 years. Genetic and environmental influences are likely to vary with age.

With our study, we emphasize the parallels between dental caries, one of the most ubiquitous chronic diseases of childhood, and other NCDs, in particular, the role of early-life environmental factors on disease risk. Our findings reveal that for dental caries, an evolutionary mismatch between human development and environmental change may be relevant, as suggested for other NCDs (such as allergy and psychiatric disorders).22 An analysis of historical skeletons, from before farming (Mesolithic) to medieval periods, reveals that dental caries is one of the first signs of this mismatch, with the change from hunting and gathering to farming and, later, the industrial revolution leading to a shift to a disease-associated microbiome with reduced diversity.23 Given the significant morbidity and mortality from NCDs,24 including dental caries, a cohesive global strategy to address environmental risk factors is pertinent.

We used statistical models fitted to data from twins to identify a number of modifiable environmental risk factors, including those in the prenatal period. These modifiable factors should be considered when determining the caries risk of individuals as well as when designing public health initiatives. Community water fluoridation is widely recognized for its socially equitable reduction in caries experience, and the strong associations between lack of community water fluoridation and both caries outcomes clearly support its effectiveness as a population health measure.25 

We identified maternal obesity in pregnancy as a modifiable risk factor for childhood caries, in keeping with previous cohort studies.26 The relationship between maternal and child obesity and dental caries is complex because it is difficult to delineate whether the increased caries risk is due to biological influences on the child or developing fetus, transfer of dietary and/or lifestyle habits, or confounding by social and other unknown factors. Aspects of the intrauterine environment, such as maternal obesity, may lead to epigenetic changes that result in fetal programming, which, in turn, may increase future susceptibility to dental caries.27 

Authors of several studies have reported that HSPMs and their related condition in the permanent teeth, molar incisor hypomineralization (MIH), are risk factors for caries.28 HSPMs are clinically detectable immediately after tooth eruption at 2 to 3 years, so early dental examinations are important to identify children at risk. Developmental defects of enamel, such as HSPMs, are due to early (prenatal) exposures during tooth enamel formation.29 As such, the association between caries and HSPMs strengthens the case for early programming of caries risk. The concordance for caries is low, suggesting that the nonshared environment is relatively important for caries risk. Nonshared factors account for phenotypic differences between twins and may arise as early as the prenatal period, for example, because of differential cord attachment affecting nutritional supply to the embryo.30 Although HSPMs are the only nonshared risk factor identified here, further studies exploring these nonshared early-life factors are warranted.

Our study has some limitations. Although a high retention rate was maintained, the sample size limited power and precision of some findings. Although the exposure data were obtained prospectively, the outcome variables (any and advanced caries) are based on measurements at a single time point and do not capture lifetime caries experience. Community water fluoridation does not necessarily correspond to consumption of fluoridated water, which is influenced by amount of water consumption, source of drinking water, and level of fluoride in drinking water. Dental examinations did not include radiographs, and therefore some carious lesions and restorations, particularly on approximal surfaces, may have not been detected. A longitudinal measurement of caries development from tooth eruption onward would allow for analysis of the period of maximal influence of risk factors in early life. Despite efforts to minimize batch effects, we cannot discount imprecision in the vitamin D measurements. Only 1 area-level measure of SES was used, and including household and personal SES in future studies may be more informative regarding possible social gradients in caries risk. Finally, replication of our findings in singleton studies and other populations would be informative.

In this twin study, we report that shared and, to a lesser degree, nonshared environmental factors appear to be the most important determinants of caries risk, with a likely modest contribution from genetic factors. Water fluoridation, maternal obesity, and HSPMs may be important and modifiable risk factors for dental caries in young children. Interventions in which these early-life factors are targeted may help address the persistently high caries rates worldwide. These findings can help pediatricians and other health professionals involved in the care of children instigate preventive modalities early in life, before the onset of clinical disease and damage to dental tissues.

Dr Silva conceptualized and designed the study, collected data, conducted the analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Drs Kilpatrick and Craig conceptualized and designed the study, collected and interpreted the data, and critically reviewed the manuscript; Dr Manton conceptualized and designed the study, interpreted the data, and critically reviewed the manuscript; Dr Leong conceptualized and designed the study, designed the data collection instruments, collected data, and reviewed and revised the manuscript; Dr Burgner conceptualized and designed the study, interpreted the data, contributed to data analysis, and critically reviewed and revised the manuscript; Dr Scurrah conceptualized and designed the study, conducted data analysis and interpretation, and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: Supported by the National Institute of Dental and Craniofacial Research of the National Institutes of Health (award R01DE019665). The Peri/postnatal Epigenetic Twins Study was supported by grants from the Australian National Health and Medical Research Council (grants 437015 and 607358), the Bonnie Babes Foundation (grant BBF20704), the Financial Markets Foundation for Children (grant 032-2007), the Victorian government’s Operational Infrastructure Support Program, the Australian and New Zealand Society for Paediatric Dentistry (Victorian branch), and the University of Melbourne Paediatric Dentistry Fund. Dr Silva is supported by a National Health and Medical Research Council Postgraduate Health Research Scholarship. Dr Scurrah is supported by a Centre of Research Excellence grant in twin research and a National Health and Medical Research Council project grant (1084197). Funded by the National Institutes of Health (NIH).

We thank all twins and their families and Richard Saffery, Tina Vaiano, Jane Loke, Anna Czajko, Chrissie Robinson, Hillary Ho, and Supriya Raj for their expertise and assistance.

     
  • BW

    birth weight

  •  
  • CI

    confidence interval

  •  
  • GEE

    generalized estimating equation

  •  
  • HSPM

    hypomineralized second primary molar

  •  
  • ICDAS

    International Caries Detection and Assessment System

  •  
  • MIH

    molar incisor hypomineralization

  •  
  • NCD

    noncommunicable disease

  •  
  • OR

    odds ratio

  •  
  • PETS

    Peri/postnatal Epigenetic Twins Study

  •  
  • SEIFA

    Socio-Economic Indexes For Areas

  •  
  • SES

    socioeconomic status

1
Finucane
D
.
Rationale for restoration of carious primary teeth: a review.
Eur Arch Paediatr Dent
.
2012
;
13
(
6
):
281
292
[PubMed]
2
Petersen
PE
.
The World Oral Health Report 2003: continuous improvement of oral health in the 21st century–the approach of the WHO Global Oral Health Programme.
Community Dent Oral Epidemiol
.
2003
;
31
(
suppl 1
):
3
23
[PubMed]
3
Poulton
R
,
Caspi
A
,
Milne
BJ
, et al
.
Association between children’s experience of socioeconomic disadvantage and adult health: a life-course study.
Lancet
.
2002
;
360
(
9346
):
1640
1645
[PubMed]
4
Selwitz
RH
,
Ismail
AI
,
Pitts
NB
.
Dental caries.
Lancet
.
2007
;
369
(
9555
):
51
59
[PubMed]
5
Vieira
AR
,
Modesto
A
,
Marazita
ML
.
Caries: review of human genetics research.
Caries Res
.
2014
;
48
(
5
):
491
506
[PubMed]
6
Hopper
JL
,
Dite
GS
,
Byrnes
GB
. Risks to relatives.
Encyclopedia of Life Sciences
.
Chichester, UK
:
John Wiley & Sons
;
2008
:
1
6
7
Neale
MC
,
Eaves
LJ
,
Kendler
KS
.
The power of the classical twin study to resolve variation in threshold traits.
Behav Genet
.
1994
;
24
(
3
):
239
258
[PubMed]
8
Gluckman
PD
,
Hanson
MA
,
Buklijas
T
.
A conceptual framework for the developmental origins of health and disease.
J Dev Orig Health Dis
.
2010
;
1
(
1
):
6
18
[PubMed]
9
Saffery
R
,
Morley
R
,
Carlin
JB
, et al
.
Cohort profile: the peri/post-natal epigenetic twins study.
Int J Epidemiol
.
2012
;
41
(
1
):
55
61
[PubMed]
10
Australian Bureau of Statistics
. Socio-Economic Indexes for Areas (SEIFA), data cube only, 2006: Table 3. Postal area (POA) index of relative socio-economic disadvantage, 2006. Available at: www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/2033.0.55.0012006?OpenDocument. Accessed August 2, 2016
11
Becker
A
,
Busjahn
A
,
Faulhaber
HD
, et al
.
Twin zygosity. Automated determination with microsatellites.
J Reprod Med
.
1997
;
42
(
5
):
260
266
[PubMed]
12
Ghanim
A
,
Elfrink
M
,
Weerheijm
K
,
Mariño
R
,
Manton
D
.
A practical method for use in epidemiological studies on enamel hypomineralisation.
Eur Arch Paediatr Dent
.
2015
;
16
(
3
):
235
246
[PubMed]
13
Harris
PA
,
Taylor
R
,
Thielke
R
,
Payne
J
,
Gonzalez
N
,
Conde
JG
.
Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support.
J Biomed Inform
.
2009
;
42
(
2
):
377
381
[PubMed]
14
Carlin
JB
,
Gurrin
LC
,
Sterne
JA
,
Morley
R
,
Dwyer
T
.
Regression models for twin studies: a critical review.
Int J Epidemiol
.
2005
;
34
(
5
):
1089
1099
[PubMed]
15
Australian Bureau of Statistics
. 2039.0 - Information Paper: An Introduction to Socio-Economic Indexes for Areas (SEIFA). Available at: http://www.abs.gov.au/ausstats/abs@.nsf/mf/2039.0. Accessed March 25, 2019
16
Shaffer
JR
,
Wang
X
,
Feingold
E
, et al
.
Genome-wide association scan for childhood caries implicates novel genes.
J Dent Res
.
2011
;
90
(
12
):
1457
1462
[PubMed]
17
Werneck
RI
,
Mira
MT
,
Trevilatto
PC
.
A critical review: an overview of genetic influence on dental caries.
Oral Dis
.
2010
;
16
(
7
):
613
623
[PubMed]
18
Marteau
TM
,
Lerman
C
.
Genetic risk and behavioural change.
BMJ
.
2001
;
322
(
7293
):
1056
1059
[PubMed]
19
Bretz
WA
,
Corby
PM
,
Hart
TC
, et al
.
Dental caries and microbial acid production in twins.
Caries Res
.
2005
;
39
(
3
):
168
172
[PubMed]
20
Bretz
WA
,
Corby
PM
,
Melo
MR
, et al
.
Heritability estimates for dental caries and sucrose sweetness preference.
Arch Oral Biol
.
2006
;
51
(
12
):
1156
1160
[PubMed]
21
Bretz
WA
,
Corby
PM
,
Schork
NJ
, et al
.
Longitudinal analysis of heritability for dental caries traits.
J Dent Res
.
2005
;
84
(
11
):
1047
1051
[PubMed]
22
Barouki
R
,
Gluckman
PD
,
Grandjean
P
,
Hanson
M
,
Heindel
JJ
.
Developmental origins of non-communicable disease: implications for research and public health.
Environ Health
.
2012
;
11
:
42
[PubMed]
23
Adler
CJ
,
Dobney
K
,
Weyrich
LS
, et al
.
Sequencing ancient calcified dental plaque shows changes in oral microbiota with dietary shifts of the Neolithic and Industrial revolutions.
Nat Genet
.
2013
;
45
(
4
):
450
455, 455e1
[PubMed]
24
World Health Organization
.
Global Action Plan for the Prevention and Control of Noncommunicable Diseases 2013-2020
.
Geneva, Switzerland
:
World Health Organization
;
2013
25
Burt
BA
.
Fluoridation and social equity.
J Public Health Dent
.
2002
;
62
(
4
):
195
200
[PubMed]
26
Wigen
TI
,
Wang
NJ
.
Maternal health and lifestyle, and caries experience in preschool children. A longitudinal study from pregnancy to age 5 yr.
Eur J Oral Sci
.
2011
;
119
(
6
):
463
468
[PubMed]
27
Wu
Q
,
Suzuki
M
.
Parental obesity and overweight affect the body-fat accumulation in the offspring: the possible effect of a high-fat diet through epigenetic inheritance.
Obes Rev
.
2006
;
7
(
2
):
201
208
[PubMed]
28
Americano
GC
,
Jacobsen
PE
,
Soviero
VM
,
Haubek
D
.
A systematic review on the association between molar incisor hypomineralization and dental caries.
Int J Paediatr Dent
.
2017
;
27
(
1
):
11
21
[PubMed]
29
Silva
MJ
,
Scurrah
KJ
,
Craig
JM
,
Manton
DJ
,
Kilpatrick
N
.
Etiology of molar incisor hypomineralization - a systematic review.
Community Dent Oral Epidemiol
.
2016
;
44
(
4
):
342
353
[PubMed]
30
Plomin
R
,
Daniels
D
.
Why are children in the same family so different from one another?
Behav Brain Sci
.
1987
;
10
(
1
):
1
16

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data