Adverse childhood experiences, child poverty, and adiposity trajectories from childhood to adolescence: evidence from the Millennium Cohort Study

Child and adolescent adiposity has risen in recent decades; in the UK, 9.7% of children entering primary school and 20.2% in the final primary year were obese in 2019. Adiposity beginning in childhood carries short- and long-term health consequences. Multiple biological, psychosocial, and cultural factors contribute to adiposity, complicating etiologic understanding. Adverse childhood experiences (ACEs) are negative experiences in childhood that require substantial adaptation (e.g., abuse, household dysfunction). ACEs are common: nearly half of UK adults report at least one ACE, with similar high prevalence internationally. Research often treats ACEs as a cumulative score (ACE score) and sometimes examines individual ACEs; recent recommendations emphasize moving beyond cumulative scores to identify which adversities drive associations. Prior work shows ACEs relate to worse physical and mental health across the life-course, including consistent associations with adult adiposity (pooled OR ~1.34–1.46) with dose-response relationships. Fewer studies have focused on children/adolescents; a meta-analysis reported pooled OR 1.13 (95% CI 0.92–1.39) for child maltreatment–childhood obesity, and most studies in a recent review found associations, though with inconsistency by type of ACE, sex, age, and adiposity measure. Sex differences are notable; some studies report associations only among girls or only among boys, and mechanisms may differ by sex (e.g., maternal depression directly related to boys’ BMI but indirectly via child depression for girls). Many prior studies used single time-point BMI, missing developmental trajectories and potentially biasing adiposity measurement. Childhood poverty, associated with worse health and higher ACE exposure, may interact with ACEs to influence adiposity, and poverty and ACEs often co-occur, suggesting joint consideration. This study investigates whether early childhood ACEs (cumulative and individual) relate to BMI and FMI trajectories from ages 5/7 to 17 and whether poverty modifies these associations. Hypotheses: higher ACE scores associate with steeper adiposity trajectories; associations differ by individual ACE; and poverty amplifies ACE effects.

The paper reviews evidence that ACEs are prevalent and linked to adverse adult health, including obesity with pooled odds ratios between 1.34 and 1.46 and dose-response relationships. Among children/adolescents, a meta-analysis reported a pooled OR of 1.13 (95% CI 0.92–1.39) for child maltreatment and obesity, and a systematic review found 21 of 24 studies reporting associations, though findings varied by specific ACEs, sex, age group, and adiposity metric. Sex-specific patterns have been reported, with some studies finding associations only in one sex, and potential mechanisms differing (e.g., maternal depression’s direct vs indirect pathways). Prior work has often relied on single time-point BMI, limiting insight into trajectories and adiposity composition. The literature has paid limited attention to poverty’s modifying role, despite its association with both health and ACE exposure and their overlap, motivating examination of ACE–poverty interactions.

Design and data source: Prospective cohort analysis using the UK Millennium Cohort Study (MCS), a nationally representative cohort of children born 2000–2001. Baseline N=18,540; follow-ups at ages 3, 5, 7, 11, 14, and 17 (MCS2–MCS7). Loss to follow-up from baseline: 15.6%, 17.8%, 25.2%, 28.3%, 36.7%, and 42.7% at MCS2–MCS7, with higher non-response in ethnic/disadvantaged areas. Ethical approval obtained and informed consent collected. Exposure: Eight ACEs measured prospectively in early childhood (≤ age 3): (1) Parental separation (absence of natural mother/father at MCS1 or MCS2); (2) Parental depression/anxiety diagnosis (either parent at MCS1 or MCS2); (3) Parental drug use (either parent reporting occasional/regular recreational drug use at MCS2); (4) Parental alcohol misuse (either parent drinking daily or 5–6 times/week at MCS1 or MCS2); (5) Interparental use of force (either parent reporting partner ever used force at MCS1 or MCS2); (6) Parental discord (derived at MCS2 using 6 GRIMS-based items; total score >3 indicating discord); (7) Harsh parenting (5 CTSPC items; sum ≥4 indicating harsh parenting); (8) Physical punishment (mother reports smacking frequency at MCS2: monthly or more = yes). A cumulative ACE score was created: 0, 1, 2, 3+ ACEs. Effect modifier: Poverty measured using McClements equivalized household income below 60% of the national median (MCS1 primarily; supplemented by MCS2 if missing). Outcomes: Adiposity measured by BMI (kg/m²) at ages 5, 7, 11, 14, 17; and Fat Mass Index (FMI = body fat mass/height²) at ages 7, 11, 14, 17. Height, weight, and body fat percentage measured by trained interviewers using Tanita BF-522W scales; prior validation indicates ±5% accuracy and ±1% repeatability. Covariates: Child sex, ethnicity (White; Mixed/Other; Indian; Pakistani/Bangladeshi; Black/Black British), birth weight (kg); maternal prenatal BMI (kg/m²) and maternal age at birth; highest parental occupational social class (NS-SEC 5-category); highest parental education (None; GCSEs; A-Levels; Higher education; Overseas qualifications). Analysis: Linear mixed-effects models with random intercepts and slopes and a quadratic term for time modeled BMI trajectories from age 5 to 17 and FMI from age 7 to 17. Intercepts set at age 5 for BMI and age 7 for FMI. Model selection based on AIC/BIC; maximum likelihood estimation. Complete-case analysis including participants with at least one adiposity measure and complete data on ACEs, poverty, and covariates, yielding N=7,282 for BMI and N=6,912 for FMI. Sex stratification applied due to sex differences; interactions tested via three-way terms (ACE×time×poverty) and time×sex. Analyses conducted in STATA 17.

- Descriptive: Approximately 25%, 33%, 24%, and 18% of boys and 27%, 36%, 23%, and 14% of girls had 0, 1, 2, and 3+ ACEs, respectively. Parental depression was the most common ACE. About 16% lived in poverty. BMI and FMI increased with age; girls had a steeper FMI increase than boys. - ACE score and trajectories: In sex-stratified models adjusted for covariates, higher ACE scores were associated with steeper BMI/FMI trajectories overall, but statistically significant only among boys with 3+ ACEs compared to 0 ACEs: BMI slope β = 0.13 (95% CI 0.02–0.24); FMI slope β = 0.09 (95% CI 0.01–0.19). No baseline differences in BMI (age 5) or FMI (age 7) by ACE score. - Individual ACEs: Parental depression associated with steeper increases in both BMI and FMI for boys and girls (BMI: boys β = 0.15, 95% CI 0.07–0.23; girls β = 0.13, 95% CI 0.05–0.20. FMI: boys β = 0.09, 95% CI 0.03–0.15; girls β = 0.09, 95% CI 0.02–0.16). Boys with parental depression also had higher starting BMI/FMI. Parental separation and physical punishment associated with steeper BMI/FMI increases among girls only (BMI: separation β = 0.25, 95% CI 0.06–0.44; physical punishment β = 0.14, 95% CI 0.03–0.26. FMI: separation β = 0.20, 95% CI 0.03–0.37; physical punishment β = 0.12, 95% CI 0.02–0.22). Interparental use of force associated with higher intercepts (baseline levels) of BMI and FMI in girls and FMI in boys. Parental alcohol misuse associated with flatter increases in BMI and FMI for both sexes (BMI slope: boys β = −0.11, 95% CI −0.20 to −0.02; girls β = −0.18, 95% CI −0.27 to −0.09. FMI slope: boys β = −0.07, 95% CI −0.14 to −0.01; girls β = −0.16, 95% CI −0.23 to −0.08). - Poverty interactions: No significant interactions between ACEs (cumulative or individual) and poverty on adiposity trajectories. However, poverty was associated with steeper BMI/FMI increases in models including individual ACEs.

The study addressed whether early childhood ACEs relate to adiposity trajectories across childhood and adolescence and whether poverty modifies these relationships. Using longitudinal mixed models, results showed limited evidence that cumulative ACEs relate to steeper adiposity trajectories, significant only among boys with 3+ ACEs, and no baseline differences by ACE score. This partially aligns with prior studies reporting weak or sex-specific associations and suggests that effects of early ACEs on adiposity may require time to manifest. Analyses of individual ACEs revealed clearer patterns: parental depression predicted steeper BMI/FMI increases in both sexes, with higher baseline adiposity among boys, while parental separation and physical punishment were linked to steeper increases among girls only, highlighting sex-specific vulnerability and supporting the importance of disaggregating ACEs. The inverse association between parental alcohol misuse and adiposity slope may reflect measurement limitations (frequency captured, not quantity) and socioeconomic patterning of drinking frequency. The absence of ACE–poverty interactions suggests poverty elevates ACE exposure risk but does not amplify ACE effects on adiposity trajectories; nonetheless, poverty itself was linked to steeper adiposity increases in individual ACE models. Mechanistically, sex differences may be influenced by sociocultural factors (e.g., differential help-seeking and social support) and biological pathways (e.g., stress system alterations influencing pubertal timing differently by sex). The findings emphasize the complexity of ACE–adiposity relationships, sex differences, and the value of examining both cumulative and specific ACEs.

This study demonstrates that early childhood ACEs are associated with adiposity development in nuanced, sex-specific ways. Cumulatively, 3+ ACEs predicted steeper BMI and FMI increases among boys only. Individually, parental depression predicted steeper adiposity increases in both sexes; parental separation and physical punishment predicted steeper increases among girls; interparental force related to higher baseline adiposity (girls BMI/FMI; boys FMI); and parental alcohol misuse related to flatter adiposity increases. No ACE–poverty interaction was detected. These results underscore the importance of considering sex and specific ACEs when assessing risk for adverse adiposity trajectories. Future research should elucidate mechanisms (psychosocial and biological), refine ACE measurements (including patterns of co-occurrence), and evaluate interventions that mitigate ACE impacts on obesity risk.

- Measurement of ACEs may entail misclassification: cumulative scores assume equal weighting of adversities and overlook co-occurrence; single-ACE models ignore clustering. - Parental alcohol misuse was defined by frequency, not quantity, likely underestimating misuse and potentially confounding by socioeconomic status. - Limited item availability for parental discord and harsh parenting (abbreviated GRIMS/CTSPC), and parent-reported ACEs may be biased; adversities outside the family (e.g., bullying) not captured. - Complete-case analysis may introduce attrition bias; non-response higher in disadvantaged/ethnic minority groups. - Observational design precludes causal inference.