Salivary Epigenetic Measures of Body Mass Index and Social Determinants of Health Across Childhood and Adolescence

Abstract

Importance: Children who are socioeconomically disadvantaged are at increased risk for high body mass index (BMI) and multiple diseases in adulthood. The developmental origins of health and disease hypothesis proposes that early life conditions affect later-life health in a manner that is only partially modifiable by later-life experiences.

Objective: To examine whether epigenetic measures of BMI developed in adults are valid biomarkers of childhood BMI and if they are sensitive to early life social determinants of health.

Design, setting, and participants: This population-based study of over 3200 children and adolescents aged 8 to 18 years included data from 2 demographically diverse US pediatric cohort studies that combine longitudinal and twin study designs. Analyses were conducted from 2021 to 2022.

Exposures: Socioeconomic status, marginalized groups.

Main outcome and measure: Salivary epigenetic BMI, BMI. Analyses were conducted to validate the use of saliva epigenetic BMI as a potential biomarker of child BMI and to examine associations between epigenetic BMI and social determinants of health.

Results: Salivary epigenetic BMI was calculated from 2 cohorts: (1) 1183 individuals aged 8 to 18 years (609 female [51%]; mean age, 13.4 years) from the Texas Twin Project and (2) 2020 children (1011 female [50%]) measured at 9 years of age and 15 years of age from the Future of Families and Child Well-Being Study. Salivary epigenetic BMI was associated with children's BMI (r = 0.36; 95% CI, 0.31-0.40 to r = 0.50; 95% CI, 0.42-0.59). Longitudinal analysis found that epigenetic BMI was highly stable across adolescence but remained both a leading and lagging indicator of BMI change. Twin analyses showed that epigenetic BMI captured differences in BMI between monozygotic twins. Moreover, children from more disadvantaged socioeconomic status (b = -0.13 to -0.15 across samples) and marginalized racial and ethnic groups (b = 0.08-0.34 across samples) had higher epigenetic BMI, even when controlling for concurrent BMI, pubertal development, and tobacco exposure. Socioeconomic status at birth relative to concurrent socioeconomic status best predicted epigenetic BMI in childhood and adolescence (b = -0.15; 95% CI, -0.20 to -0.09).

Conclusion and relevance: This study demonstrated that epigenetic measures of BMI calculated from pediatric saliva samples were valid biomarkers of childhood BMI and may be associated with early-life social inequalities. The findings are in line with the hypothesis that early-life conditions are especially important factors in epigenetic regulation of later-life health. Research showing that health later in life is linked to early-life conditions has important implications for the development of early-life interventions that could significantly extend healthy life span.

Figure 1.. Associations Between Epigenetic Body Mass Index (BMI) and Measured BMI

Cross-sectional associations between scaled epigenetic BMI and measured BMI. Results are presented for 3 samples: 8- to 18-year-old children from the Texas Twin Project (TTP) and 9-year-old children and 15-year-old children from Future of Families and Child Well-Being (FFCW) study (FFCW 9 y; FFCW 15 y). Epigenetic BMI and BMI z scores were scaled in the full sample of each study and time point (A). Within monozygotic twin-pair associations between scaled epigenetic BMI and measured BMI, results were based on 183 monozygotic twin pairs from TTP (B). Association of within-person longitudinal changes in scaled epigenetic BMI and within-person change in BMI from age 9 years to age 15 years. Results based on 1904 longitudinal observations from FFCW (C).

Figure 2.. Cross-Lagged Model of Epigenetic Body Mass Index (BMI) and BMI in the Future of Families and Child Well-Being Study

BMI is the US Centers for Disease Control and Prevention z score for BMI for age. Epigenetic BMI (Epi-BMI) was residualized for cell distributions and technical artifacts and normalized with a mean = 0, SD = 1. Coefficients not shown are controls that influence age 9 years’ BMI and age 9 years’ epigenetic BMI, including self-reported race, sex assigned at birth, socioeconomic status at birth, and if mom smoked while pregnant (see eTable 6 in Supplement 1). SDs of the residual error variances of BMI and epigenetic BMI are in parenthesis after their variance estimate. Cov indicates covariance; var, variance. ^aP < .05

Figure 3.. Socioeconomic Inequalities in Children’s Epigenetic Body Mass Index (BMI) Profiles

Cross-sectional associations between scaled family-level socioeconomic status and salivary epigenetic BMI. Results are presented for 3 samples: A, data from 8- to 18-year-old children from the Texas Twin Project (TTP) and B, data from 9-year-old children and 15-year-old children from Future of Families and Child Well-Being (FFCW) study (FFCW 9 y; FFCW 15 y). Socioeconomic status z scores and epigenetic BMI were scaled in the full sample of each study and time point.

Figure 4.. Cross-Lagged Model of Socioeconomic Status (SES) and Epigenetic Body Mass Index (BMI) Profiles in the Future of Families and Child Well-Being Study

Epigenetic BMI (Epi-BMI) was residualized for cell distributions and technical artifacts. SES and epigenetic BMI are normalized with a mean = 0, SD = 1. Coefficients not shown are controls that influence age 9 years SES and age 9 years epigenetic BMI, including self-reported race, sex assigned at birth, and if mom smoked while pregnant (see eTable 10 in Supplement 1). SDs of the residual error variances of SES and epigenetic BMI are in parenthesis after their variance estimate. Cov indicates covariance; var, variance. ^aP < .05 ^bP < .01 ^cP < .001