Obesity and adverse childhood experiences in relation to stress during the COVID-19 pandemic: an analysis of the Canadian Longitudinal Study on Aging

The study investigates how obesity and adverse childhood experiences (ACEs) relate to stress experienced during the COVID-19 pandemic among older adults, and whether ACEs modify the relationship between obesity and stress. Stress and obesity have a bidirectional, life-course relationship influenced by individual factors such as genetics, lifestyle, medications, and mental health. Obesity can increase chronic stress through comorbidities and weight stigma. ACEs—encompassing abuse, neglect, and family dysfunction—are linked to both obesity and stress and can exert effects into older adulthood. The purpose is to quantify associations between obesity, ACEs, and pandemic-related stress, and to test effect modification of obesity-stress associations by ACEs.

Prior work highlights a well-established association between stress and obesity through behavioral pathways (e.g., diet, physical inactivity), neuroendocrine mechanisms (e.g., glucocorticoid activation), and mental health comorbidity. Obesity-related stigma may further drive chronic stress. ACEs have been consistently associated with adverse health, including obesity and heightened stress responsivity later in life, with dose-response patterns especially among those with multiple ACEs. Conceptual models (e.g., van der Valk et al.) frame obesity-stress interrelations across the life course, where both proximal (e.g., body weight) and distal (e.g., early adversity) exposures accumulate to influence later outcomes. Evidence from disaster and pandemic research indicates elevated stress and negative perceptions during COVID-19, with potential heterogeneity by prior adversity. However, whether ACEs modify the obesity-stress link during population-level stressors remained unclear prior to this study.

Design and cohort: Longitudinal analysis within the Canadian Longitudinal Study on Aging (CLSA) COVID-19 Questionnaire Study. Participants aged 50–96 years who completed the CLSA COVID-19 Exit Survey (Sept–Dec 2020) and had pre-pandemic measures were included (n = 23,972). Exposure: Obesity was assessed at CLSA Follow-up 1 (2015–2018). Body mass index (BMI) was measured for most participants; for self-reported BMI, Statistics Canada correction equations (sex-specific, accounting for sociodemographic variables) were applied to mitigate self-report bias. BMI categories followed WHO cut-offs: normal weight (<24.9 kg/m²), overweight (25.0–29.9), obesity class I (30.0–34.9), class II (35.0–39.9), class III (≥40.0). Underweight was grouped with normal due to small numbers. Modifier/exposure: ACEs were assessed at CLSA Follow-up 1 via 11 yes/no items covering abuse (physical, sexual, emotional), neglect, exposure to intimate partner violence, parental death, parental divorce/separation, and living with a family member with mental health problems. A cumulative ACE score (0–11) was created and analyzed in categories (e.g., 0, 1, 2, 3, 4–8), with most analyses focusing on dose-response; a dichotomy (0 vs. ≥1 ACE) was used for interaction testing. Outcomes: Pandemic-related stress at the CLSA COVID-19 Exit Survey (Sept–Dec 2020) comprised: (1) stressors across domains—health (range 0–3), resources (0–4), caregiving (0–2)—and a cumulative stressor score (0–12) formed by summing selected experienced items (e.g., job loss, difficulty accessing food/healthcare, inability to care for self/others, family relationship breakdown); and (2) perceived consequences of the pandemic, assessed with a single item and dichotomized as negative/very negative versus neutral/positive/very positive. Covariates: A priori confounders included age group (50–64, 65–74, 75–96), sex, racial background (white/non-white), physical activity (WHO-guideline-based dichotomy from PASE), total household income (<$50k, $50–<100k, $100–<150k, ≥$150k), alcohol consumption (did not drink last 12 months/occasional/regular ≥ monthly), and depression risk (CES-D short scale, cut-point per instrument). Statistical analysis: Logistic, Poisson, and negative binomial regression models estimated RRs and 95% CIs for associations of obesity and ACEs with stress outcomes. ACEs and obesity were modeled jointly and separately; models adjusted for confounders. Effect modification of obesity-stress associations by ACEs was evaluated on additive (RERI) and multiplicative (ratio of RRs, RRR) scales, with 95% CIs for RERI via delta method. Sensitivity analyses examined maltreatment versus family dysfunction ACE subsets. Analyses were conducted in SAS 9.4.

- Sample and stress prevalence: Of 23,972 participants, over three quarters (76%) reported at least one pandemic-related stressor and 63% perceived the pandemic’s consequences as negative/very negative. - Obesity and stress: A dose-response relationship was observed; higher obesity classes were associated with greater risk of reporting additional stressors overall and within health and resources domains. Example: class III obesity versus healthy weight was associated with increased overall stressors (adjusted RR = 1.67; 95% CI: 1.12–2.39). Obesity was modestly associated with perceiving the pandemic as negative/very negative. - ACEs and stress: 61% reported ≥1 ACE. There was a strong dose-response between ACE count and stress across all outcomes. Those with 4–8 ACEs versus none were more likely to perceive the pandemic as negative/very negative (adjusted RR = 1.32; 95% CI: 1.19–1.47) and had higher risks of additional stressors in total and across domains. - Modification by ACEs: No consistent evidence that ACEs modified the association between obesity and stress on additive or multiplicative scales. One exception was a significant multiplicative interaction for the health domain indicating lower likelihood of an additional health stressor among those with obesity and ACEs compared with those with obesity and no ACEs (RRR ~0.74; 95% CI: 0.58–0.96). - Sex differences (supplemental): Sex modified associations of obesity with stress; females with class III obesity were less likely than males with class III obesity to report stressors across domains and to perceive the pandemic as negative (e.g., health domain RRR = 0.79; 95% CI: 0.63–1.00; total stressors RRR = 0.80; 95% CI: 0.70–0.91).

The study addressed whether obesity and ACEs independently and jointly relate to pandemic-related stress among older adults. Findings support that both higher obesity and greater ACE exposure are associated with increased stressor burden during the COVID-19 pandemic, aligning with life-course frameworks linking stress and obesity and with literature on the enduring impacts of early adversity. However, ACEs did not consistently modify obesity-stress associations, suggesting largely independent contributions during the first pandemic year. The observed sex modification indicates heterogeneity by sex in stress experiences among those with severe obesity. These results underscore the importance of considering both proximal (current obesity) and distal (early-life adversity) exposures when identifying individuals at heightened risk during widespread stressors like pandemics. Public health and clinical strategies could leverage these insights to target screening and support for those at increased risk of adverse stress responses.

This study demonstrates dose-response associations between higher obesity classes and greater pandemic-related stressors, and between higher ACE counts and both increased stressors and more negative pandemic perceptions. ACEs did not consistently modify the obesity-stress association, though sex modified some obesity-stress relationships. The findings highlight subgroups at elevated risk during large-scale stressors and point to the need for interventions that address stress coping and stigma-related pathways among people with obesity and support for those with histories of adversity. Future research should examine longer-term outcomes of pandemic-related stress, potential mechanisms (including weight stigma and neuroendocrine pathways), temporal dynamics beyond the first pandemic year, and targeted interventions to disrupt the stress-obesity cycle.

- Generalizability: The sample was predominantly of white racial background, potentially limiting representativeness. - Timing and recall: Stressors and perceived consequences were recalled for the early pandemic period (Sept–Dec 2020); experiences may have evolved over time. ACEs were self-reported retrospectively, introducing potential recall/information bias. - Measurement validity: Stress measures, although adapted from prior disaster research, were not specifically validated in the CLSA sample. - BMI measurement: BMI used from CLSA Follow-up 1 (2015–2018) may not reflect BMI at the time of the 2020 Exit Survey. Self-reported BMI (for a subset) required correction using equations from 2005, which may not fully remove bias. - Potential selection bias: Older adults with limitations may have been less likely to participate. - Residual confounding: Despite adjustment, unmeasured or imperfectly measured confounders may remain. - Interaction power/timing: Lack of observed ACE modification may reflect timing (first pandemic year) or insufficient power for some strata/outcomes.