Abstract
Introduction:
Childhood physical, sexual, and emotional abuse are linked to adult obesity, and little is known about what protective factors might mitigate this association.
Methods:
Data from female (n=4,247) and male (n=1,982) participants in the longitudinal Growing Up Today cohort study from 1996 to 2013 were used to examine whether factors found to promote mental health resilience after abuse also operate as buffers (modifiers) of the abuse–weight status association. At age 20–25 years, participants were asked about their history of child abuse before age 18 years. Potential resilience factors (modifiers) included childhood family SES, neighborhood safety, supportive relationships with adult non-family members, quality of maternal relationship, family structure, religious service attendance, and prayer/meditation. Associations between child abuse and BMI at age 25–32 years were modeled using linear regression, adjusted for sociodemographic variables and baseline BMI. Potential modifiers were tested with interaction terms. Analyses were run in 2019–2020.
Results:
Relative to no abuse, severe abuse was associated with 0.9 kg/m2 (95% CI=0.5, 1.2) higher adult BMI, corresponding to a 46% increased risk of obesity (95% CI=1.28, 1.67). Less severe abuse was not significantly associated with BMI (β=0.1, 95% CI= −0.2, 0.4). There were no significant interactions between modifiers and abuse.
Conclusions:
Factors previously found to promote resilience to mental health sequelae after abuse did not modify the association of severe child abuse with higher weight status.
INTRODUCTION
Physical, sexual, and emotional abuse in childhood are associated with significant adult mental and physical health problems,1–5 including obesity, a leading cause of morbidity and mortality.4,5 Yet, there is also growing evidence in the mental health literature of individuals’ potential for resilience after abuse, defined as “good outcomes in spite of serious threats to adaptation or development.”6 However, an examination of the literature indicated that no published studies have examined resilience to obesity after child abuse.
Protective factors have been identified that appear to support mental health despite abuse.7–9 Poor mental health may be an important part of the pathway from abuse to weight gain. Abuse may disrupt energy homeostasis through a variety of mental health–related behavioral and physiological effects on weight.10 For example, abuse may interrupt development of affect regulation—that is, the ability to cope with distress—leading to obesogenic maladaptive coping behaviors11–14 such as overeating of highly palatable (high-fat/high-sugar) foods.15–17 Thus, factors that support positive mental health despite exposure to abuse may also protect against excessive weight gain.11
Resources found to support positive mental health after abuse include higher SES,18 positive interpersonal supports,18,19 neighborhood quality,20 and practices such as prayer or meditation.21,22 In one study of resilience to overweight and obesity in children exposed to adversity, positive contextual factors, such as neighborhood safety, attenuated the adversity–weight association.23 However, this study did not include childhood abuse in its adversity measure. Thus, the aim of the current analysis is to identify factors that modify and buffer the association of childhood or adolescent physical, sexual, and emotional abuse with adult weight status. Given findings in the mental health literature, the following potential protective factors are examined: higher family social status, neighborhood safety, presence in child’s life of a supportive non-family adult, child’s reported quality of relationship with their mother, 2 biological parents in the home, living with extended family, religious service attendance, and prayer or meditation. It is hypothesized that there will be an interaction between each modifier and abuse such that the presence of the modifier will be associated with a reduced association between child abuse and adult weight status.
METHODS
Study Sample
The Growing Up Today Study (GUTS) cohort comprises children of participants in the longitudinal Nurses’ Health Study II. In 1996, Nurses’ Health Study II participants with children aged 9–14 years were asked to consent to their children’s participation in GUTS; 18,526 (54%) consented, and their children (N=26,765) were invited to participate. Of these, 16,882 (n=9,039 who identified as female and n=7,843 who identified as male) returned the baseline GUTS questionnaire. GUTS participants were followed through online and mailed questionnaires annually or biennially from 1996 through 2007, in 2010, and in 2013. Child abuse was assessed retrospectively in the 2007 survey (age 20–25 years) and body weight was assessed in 2013 (age 25–32 years). Potential modifiers were primarily from surveys between 1996 and 1999, when participants were aged <18 years. However, 2 important interpersonal factors—participants’ relationship with their mothers and retrospective reports of a supportive non-family adult during childhood—were available only in 2005 and 2007, respectively. Details on variables are provided in the Measures section.
Analyses included participants who completed the questionnaires on which the exposure and outcome were assessed (N=6,804; n=4,247 female and n=1,982 male participants). Participants missing abuse history (n=260 female and n=213 male participants, 7%) or body weight (n=55 female and n=47 male participants, 1%) were excluded, leaving an analytic sample of 6,229. In addition, modifier analyses included participants with non-missing data on the relevant modifier variable (n=5,222–6,229) (Appendix Table 1). Data analyses occurred in 2019–2020.
This study was approved by the IRB of Brigham and Women’s Hospital and was exempted from review by the University of Minnesota IRB.
Measures
The 2007 GUTS survey assessed physical, sexual, and emotional abuse during childhood (age 0–10 years) and adolescence (age 11–17 years), which were combined to examine abuse prior to age 18 years. Coding of abuse variables followed that of previous studies of GUTS data.24
Physical abuse was assessed with 2 items from the Conflict Tactics Scales,25 asking participants whether, in childhood or adolescence, an adult in their family: (1) pushed, grabbed, or shoved them or (2) kicked, punched, or hit them with something in a way that hurt their body, or physically attacked them in some other way. Physical abuse was categorized as none, less severe (pushed/shoved but not kicked/punched/physically attacked in either childhood or adolescence), or severe (kicked/punched/physically attacked in childhood or adolescence).
Sexual abuse was ascertained with questions by Finkelhor et al.26 asking, during childhood or adolescence: (1) whether they were touched in a sexual way by an adult or older child or were you forced to touch an adult or older child in a sexual way when you did not want to? and (2) did an adult or older child force you or attempt to force you into any sexual activity by threatening you, holding you down or hurting you in some way when you did not want to? Sexual abuse was categorized as none, sexual touching but no forced sexual activity in childhood or adolescence, and forced sexual activity in childhood or adolescence.
Emotional abuse was measured with 3 items from the Childhood Trauma Questionnaire,27 asking participants how often, in childhood or adolescence, (5 points from never to very often) an adult in their family yelled and screamed at them, said hurtful or insulting things to them, or punished them in a way that seemed cruel. Responses were summed and categorized into 3 levels that mirrored the prevalence of less severe and more severe abuse observed in other cohorts28,29: no abuse (<75th percentile), less severe abuse (75th–90th percentile), and severe abusive (>90th percentile).
Timing of abuse (any versus none) was examined by creating a series of indicator variables for: no type of abuse in either childhood or adolescence; physical, sexual, or abuse in childhood only; physical, sexual, or emotional abuse in adolescence only; and physical, sexual, or emotional abuse in both time periods.
Following a previous analysis of female participants in this cohort,24 a combined abuse variable was coded as: none (no physical, sexual, or emotional abuse in childhood or adolescence), less severe only (less severe physical abuse, sexual touching only, or less severe emotional abuse in childhood or adolescence, with no severe abuse), and severe (severe physical abuse, forced sex, or severe emotional abuse in childhood or adolescence).
Participant BMI in kilograms divided by meters squared (kg/m2) was based on self-reported weight and height in 2013 (age 25–32 years). Self-reported weight has been shown to have good validity in adolescents and young adults.30,31
In 1999 (age 12–17 years) and 2001 (age 14–19 years), participants were asked where their family’s social status ranked relative to others in the U.S. (range=0–10, with lower numbers corresponding to higher status, described as at the top of the ladder are the people who are best off—they have the most money, the highest amount of schooling, and the jobs that bring the most respect). The average of responses at these 2 timepoints was taken. The continuous measure was used for tests of interaction, and dichotomized at the sex-specific median for stratified analyses.
Participants were asked to respond to the statement: It’s not safe enough in my neighborhood to get out and get some exercise, at 3 time points: 1997 (age 10–15 years), 1998 (age 11–16 years), and 1999 (age 12–17 years). Response options were always safe, usually safe, usually unsafe, or really unsafe. The average of responses over these 3 reports was taken and treated as a continuous variable in interaction analyses. For stratified analyses, responses were dichotomized into always safe versus all other responses.
At age 20–25 years (2007), participants were asked: In childhood (before age 11)/in adolescence (age 11–17) how often did an adult who was NOT in your family make you feel that you were important or special? Responses options were never, rarely, sometimes, often, or very often.26 The original ordinal variable was used to test interactions, and was dichotomized at often or very often for stratified analyses.
At age 18–23 years (2005), participants were asked to rate (strongly disagree to strongly agree) how satisfied they were with their mother’s: (1) love and affection, (2) emotional support, (3) sharing things in common with participant, (4) time spent with participant, (5) ability to resolve conflicts, (6) respect for participant, (7) time for fun with participant, (8) communication, and (9) general relationship. Scores were summed and the total was treated as continuous for interaction analyses and dichotomized at the sex-specific median for stratified analyses.
The 1999 questionnaire (age 12–17 years) asked participants to report the frequency (never, less than once a month, 1–3 times a month, once a week, more than once a week) with which they attended religious services. Responses were dichotomized into regular attenders (responses of weekly or more) or irregular/non-attenders.
The 1999 questionnaire (age 12–17 years) asked participants to report the frequency (never, less than once a week, 1–6 times a week, once a day or more) with which they practiced prayer or meditation. Responses were dichotomized into regular prayer/meditation (responses of weekly or more) or irregular/no prayer/meditation.
In 1996 (age 9–14 years), participants were asked: Which adults do you live with most of the time? Response options (of which respondent could chose multiple) were: mother, father, stepmother, stepfather, grandmother, grandfather, other relative, and other adults. Family structure was categorized as “both parents” (participant indicated living with mother and father most of the time) versus other family arrangements. A separate variable was created to represent extended family or adults other than parents in the household. Respondents living with a grandmother, grandfather, other relative, or other adults were designated as living with extended family.
Age in years at 1996 baseline and non-White race were included in sociodemographic adjusted models as potential common correlates of abuse and BMI. Perceived social status was included as a covariate in analyses where it was not being examined as a predictor or modifier. Because children with higher weight status may be more vulnerable to abuse, and are also at greater risk for heavier adult weight,32 BMI percentile at baseline (participant’s baseline BMI standardized to Centers for Disease Control and Prevention growth data) was included as an adjustment.33 However, because some abuse-related weight gain may have occurred prior to study baseline (age 9–14 years), parameters controlling for baseline BMI may be over-adjusted; thus, sociodemographic-adjusted and baseline BMI–adjusted results are presented separately.
Statistical Analysis
The following approach was taken to identify what modifiers might buffer abuse-related weight gain. First, the overall associations of abuse with BMI were estimated. Second, the association of each hypothesized protective factor with BMI was estimated. Finally, interactions between abuse and modifiers found to predict BMI were tested; continuous or ordinal versions of the modifiers were used where possible. Models stratified by each dichotomized modifier were also run for assessment of patterns across strata of modifiers.
For the first step, linear models regressed BMI at age 25–32 years on: (1) each type of abuse separately, categorized by severity; (2) abuse by timing (childhood, adolescence, or both); and (3) combined abuse (no abuse, less severe abuse only, more severe abuse). Crude models were run first, then sociodemographics (age, race, and perceived social status) were added, and then baseline BMI was added. Adjusted models were then run adding all potential protective factors; models examining different types of abuse also mutually adjusted for the other abuse types.
Robust SEs were used to adjust for correlation between siblings (approximately 15% of those in the analytic sample also had a sibling in the sample), which were retained in the sample because their presence did not change results. Two supplemental analyses were run: Modified Poisson models were used to estimate the association of abuse with obesity, and sex X abuse interactions were tested and found non-significant. Thus, results are presented for male and female participants combined.
RESULTS
Of 6,229 individuals included in the analytic sample, 1,554 (24.9%) experienced less severe abuse and 1,237 (19.8%) experienced severe abuse prior to age 18 years. The sample had an average BMI of 25.3 at age 25–32 years (Table 1), and 14.6% had obesity.
Table 1.
Distribution of Analysis Variables Across Abuse Categories in the Growing Up Today Study (N=6,229)
| Variable | Child or adolescent abuse | ||
|---|---|---|---|
| None (n=3,438) | Less severe onlya (n=1,554) | More severeb (n=1,237) | |
| Mean (SD) or % | Mean (SD) or % | Mean (SD) or % | |
| Female | 66.2% | 58.6% | 63.8% |
| Nonwhite | 2.8% | 2.5% | 4.4% |
| Age in 1996 (baseline), years | 11.5 (1.6) | 11.5 (1.6) | 11.5 (1.6) |
| Age in 2013, years | 27.8 (1.7) | 27.8 (1.7) | 27.8 (1.7) |
| BMI percentile in 1996 (baseline) | 55.0 (29.2) | 54.8 (29.6) | 57.2 (30.3) |
| BMI in 2013 (kg/m2) | 25.0 (5.1) | 25.3 (5.3) | 26.1 (6.0) |
| Perceived social status (range 1 to 10, with 1 being highest status) | 3.7 (1.1) | 3.9(1.2) | 4.0(1.2) |
| Quality of relationship with mom (range 1 to 5, higher=better relationship) | 3.4 (0.7) | 3.2 (0.8) | 2.8 (1.0) |
| Important non-family adult | 72.0% | 65.8% | 55.4% |
| Neighborhood always safe | 91.3% | 88.4% | 85.3% |
| Participant lives with both parents most of the time | 91.4% | 88.7% | 82.5% |
| Participant lives with extended family or other adults | 3.0% | 3.3% | 4.9% |
| Attends religious services weekly | 59.3% | 56.6% | 52.8% |
| Prays or meditates weekly | 66.9% | 63.7% | 60.1% |
1+ type of less severe abuse (physical: pushed/grabbed/shoved; sexual: non-consensual touching; emotional: 75th percentile—<90th percentile score); no severe abuse.
1+ type of severe abuse (physical: kicked/punched/hit; sexual: forced sex; emotional: ≥90th percentile score).
After adjustment for sociodemographic variables and baseline BMI, those with less severe abuse had a similar BMI to those without abuse (β=0.1 kg/m2, 95% CI= −0.2, 0.4). Those with severe abuse had a 0.9 kg/m2 (95% CI=0.5, 1.2) higher BMI relative to those with no abuse (Table 2); this BMI difference translated into a risk ratio for obesity of 1.46 (95% CI=1.28, 1.67). Emotional abuse had the strongest association with BMI (less severe: β=0.8, 95% CI=0.5, 1.2; more severe: β=1.1, 95% CI=0.7, 1.6), and the only association that was maintained after mutual adjustment for all abuse types. Neither sexual nor physical abuse was related to BMI after adjustment for emotional abuse (Table 2). Abuse examined by timing (childhood versus adolescence versus both) showed that only those who experienced abuse in both childhood and adolescence (n=1,871) had an elevated BMI in adulthood (β=0.8, 95% CI=0.5, 1.1) ( Table 2).
Table 2.
Linear Regression Results of BMI at Age 25–32 Years as a Function of Type and Timing of Abuse Exposure in Childhood (Age 0–10 Years) and/or Adolescence (Age 11–17 Years) in the Growing Up Today Study (N=6,229)
| Abuse categorization | N | Mean BMI | Sociodemographic-adjusted b | Additionally adjusted for baseline BMIb | Mutually adjusted for protective factors and abuse typesc |
|---|---|---|---|---|---|
| βa (95% CI) | βa (95% CI) | βa (95% CI) | |||
| Combined | |||||
| None | 3,438 | 25 | 0 (ref) | 0 (ref) | 0 (ref) |
| Less severe onlyd | 1,554 | 25.2 | 0.2 (−0.2, 0.5) | 0.1 (−0.2, 0.4) | 0.1 (−0.3, 0.4) |
| More severee | 1,237 | 26.1 | 1 (0.6, 1.4) | 0.9 (0.5, 1.2) | 0.7 (0.3, 1.1) |
| Physical | |||||
| None | 4,157 | 25.1 | 0 (ref) | 0 (ref) | 0 (ref) |
| Less severe (pushed, grabbed, shoved) | 1,375 | 25.3 | 0.2 (−0.2, 0.5) | 0.3 (0.0, 0.5) | −0.1 (−0.4, 0.3) |
| More severe (kicked, punched, hit) | 730 | 26.3 | 1.2 (0.7, 1.7) | 1.1 (0.6, 1.5) | 0.2 (−0.4, 0.7) |
| Sexual | |||||
| None | 5,651 | 25.3 | 0 (ref) | 0 (ref) | 0 (ref) |
| Less severe (non-consensual touching) | 429 | 25.3 | 0.1 (−0.5, 0.6) | 0.0 (−0.5, 0.5) | −0.3 (−0.8, 0.3) |
| More severe (forced sex) | 304 | 25.6 | 0.3 (−0.4, 1.0) | 0.3 (−0.3, 0.9) | 0.0 (−0.6, 0.6) |
| Emotional | |||||
| None | 4,811 | 25 | 0 (ref) | 0 (ref) | 0 (ref) |
| Less severe (75th−<90th percentile) | 843 | 25.9 | 1 (0.6, 1.4) | 0.8 (0.5, 1.2) | 0.7 (0.3, 1.2) |
| More severe (≥90th percentile) | 671 | 26.5 | 1.6 (1.0, 2.1) | 1.1 (0.7, 1.6) | 1.0 (0.4, 1.6) |
| Any vs none in childhood and/or adolescence | |||||
| No abuse in childhood or adolescence | 2,914 | 24.9 | 0 (ref) | 0 (ref) | 0 (ref) |
| Abuse in childhood only | 986 | 25.2 | 0.1 (−0.3, 0.5) | 0 (−0.3, 0.4) | 0.1 (−0.3, 0.4) |
| Abuse in adolescence only | 443 | 25 | 0 (−0.6, 0.6) | 0.1 (−0.4, 0.5) | 0.0 (−0.6, 0.5) |
| Abuse in childhood and adolescence | 1,871 | 25.9 | 0.9 (0.6, 1.2) | 0.8 (0.5, 1.1) | 0.7 (0.3, 1.0) |
Difference between BMI at age 25–32 years compared to those with no abuse.
Adjusted for baseline age, non-White race, and perceived SES in adolescence.
Adjusted for baseline age, non-White race, and perceived SES in adolescence, baseline BMI, quality of maternal relationship, presence of supportive non-family adult, neighborhood safety, living with both parents, living with extended family, religious service attendance, prayer/meditation; models of individual abuse types (physical, sexual, emotional) are mutually adjusted for other types.
1+ type of less severe abuse (physical: pushed/grabbed/shoved; sexual: non-consensual touching; emotional: 75th percentile—<90th percentile score); no severe abuse.
1+ type of severe abuse (physical: kicked/punched/hit; sexual: forced sex; emotional: ≥90th percentile score).
Table 3 shows the association of each hypothesized modifier with BMI. After adjustment, individuals with higher social status and presence of a supportive non-family adult had an approximately 0.5 kg/m2 lower BMI than those without (higher social status: 95% CI= −0.8, −0.3; supportive non-family adult: 95% CI= −0.9, −0.3). Extended family in the home was associated with an approximately 0.8 kg/m2 higher BMI. Those who reported a more positive relationship with their mothers had a slightly lower BMI (−0.3, 95% CI= −0.5, −0.1) than those who reported a less positive relationship. Neighborhood safety, religious service attendance, prayer/meditation, and family structure showed null associations with BMI and were not pursued further as modifiers.
Table 3.
Linear Regression Results of BMI at Age 25–32 Years as a Function of Hypothesized Protective Factors in Childhood (Ages 0–18 Years) in the Growing Up Today Study (N=6,229)
| Hypothesized protective factor | N | Mean BMI | Crude | Model Sociodemographic-adjustedb | Additionally adjusted for baseline BMIb |
|---|---|---|---|---|---|
| βa (95% CI) | βa (95% CI) | βa (95% CI) | |||
| Social status | |||||
| Lower | 3,035 | 25.7 | 0 (ref) | 0 (ref) | 0 (ref) |
| Higher | 3,194 | 24.9 | −0.8 (−1.0, −0.5) |
−0.8 (−1.0, −0.5) |
−0.5 (−0.8, −0.3) |
| Neighborhood safety | |||||
| Not always safe | 2,070 | 25.5 | 0 (ref) | 0 (ref) | 0 (ref) |
| Always safe | 3,914 | 25.2 | −0.3 (−0.6, 0.0) |
−0.3 (−0.7, 0.0) |
−0.2 (−0.5, 0.1) |
| Adult outside the family made participant feel important or special | |||||
| Never, rarely, sometimes | 1,924 | 25.7 | 0 (ref) | 0 (ref) | 0 (ref) |
| Often or very often | 4,305 | 25.0 | −0.7 (−1.0, −0.4) |
−0.6 (−0.9, −0.3) |
−0.6 (−0.9, −0.3) |
| Relationship with mother | |||||
| Less positive | 3,037 | 25.4 | 0 (ref) | 0 (ref) | 0 (ref) |
| More positive | 2,799 | 25.2 | −0.2 (−0.5, 0.1) |
−0.2 (−0.5, 0.1) |
−0.3 (−0.5, −0.01) |
| Religious service attendance | |||||
| Less than weekly | 2,225 | 25.2 | 0 (ref) | 0 (ref) | 0 (ref) |
| Weekly or more | 3,004 | 25.3 | 0.1 (−0.2, 0.4) |
0.1 (−0.2, 0.4) |
0.2 (0.0, 0.5) |
| Prayer/meditation | |||||
| Less than weekly | 1,817 | 25.4 | 0 (ref) | 0 (ref) | 0 (ref) |
| Weekly or more | 3,424 | 25.2 | −0.2 (−0.6, 0.0) |
−0.2 (−0.5, 0.1) |
0.0 (−0.2, 0.3) |
| Family structure | |||||
| As child, lived with single parent or step parents | 653 | 25.7 | 0 (ref) | 0 (ref) | 0 (ref) |
| As child, lived with both parents | 5,576 | 25.2 | −0.5 (−1.0, −0.1) |
−0.4 (−0.8, 0.1) |
−0.1 (−0.5, 0.3) |
| Extended family | |||||
| No extended family | 6,024 | 25.2 | 0 (ref) | 0 (ref) | 0 (ref) |
| Extended family | 205 | 26.4 | 1.2 (0.4, 2.0) |
1.1 (0.3, 1.9) |
0.8 (0.1, 1.5) |
Difference between BMI at age 25–32 years compared to those with no abuse.
Adjusted for baseline age, non-White race, and perceived social status in adolescence (perceived social status not adjusted in models examining social status as a predictor).
There were no statistically significant interactions of modifiers with abuse in predicting BMI (Table 4). Consistent with interaction findings, stratified analyses showed mostly similar abuse–BMI estimates in those with and without each modifier, with some minor exceptions. For example, among those with no supportive non-family adult, there was a 1.0 kg/m2 (95% CI=0.4, 1.6) greater BMI associated with abuse, versus 0.6 kg/m2 (95% CI=0.2, 1.0) among those with a supportive relationship.
Table 4.
BMI at Age 25–32 Years as a Function of Abuse Exposure (None, Less Severea, More Severeb) in Childhood or Adolescence (Age <18 Years) Across Levels of Hypothesized Modifiers (N=6,229c)
| Abuse exposure | N | Mean BMI | P interaction | Crude | Model Sociodemographic-adjustede | Additionally adjusted for baseline BMIe |
|---|---|---|---|---|---|---|
| βd (95% CI) |
βd (95% CI) |
βd (95% CI) |
||||
| Unstratified / overall model | – | |||||
| None | 3,438 | 25.0 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 1,554 | 25.2 | 0.2 (−0.1, 0.5) |
0.2 (−0.2, 0.5) |
0.1 (−0.2, 0.4) |
|
| More severe | 1,237 | 26.1 | 1.0 (0.7, 1.4) |
1.0 (0.6, 1.4) |
0.9 (0.5, 1.2) |
|
| Social status | 0.60 | |||||
| Higher | ||||||
| None | 1,889 | 24.7 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 745 | 25.0 | 0.2 (−0.2, 0.7) |
0.2 (−0.2, 0.6) |
0.2 (−0.2, 0.6) |
|
| More severe | 560 | 25.4 | 0.7 (0.2, 1.2) |
0.6 (0.1, 1.1) |
0.6 (0.2, 1.0) |
|
| Lower | ||||||
| None | 1,549 | 25.4 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 809 | 25.5 | 0.1 (−0.4, 0.6) |
0.2 (−0.3, 0.6) |
0.1 (−0.3, 0.5) |
|
| More severe | 677 | 26.6 | 1.3 (0.7, 1.8) |
1.3 (0.7, 1.8) |
1.0 (0.6, 1.5) |
|
| Adult outside the family made participant feel important/special | 0.10 | |||||
| Often or very often | ||||||
| None | 2,536 | 24.9 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 1,051 | 25.1 | 0.2 (−0.1, 0.6) |
0.1 (−0.2, 0.5) |
0.2 (−0.2, 0.5) |
|
| More severe | 718 | 25.6 | 0.7 (0.2, 1.1) |
0.7 (0.2, 1.2) |
0.6 (0.2, 1.0) |
|
| Never, rarely, sometimes | ||||||
| None | 902 | 25.4 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 503 | 25.5 | 0.1 (−0.5, 0.7) |
0.1 (−0.6, 0.7) |
−0.1 (−0.6, 0.5) |
|
| More severe | 519 | 26.8 | 1.4 (0.7, 2.0) |
1.3 (0.6, 2.0) |
1.0 (0.4, 1.6) |
|
| Relationship with mother | 0.66 | |||||
| More positive | ||||||
| None | 1,836 | 24.9 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 593 | 25.4 | 0.5 (0.0, 1.0) |
0.4 (−0.1, 0.9) |
0.3 (−0.1, 0.7) |
|
| More severe | 370 | 26.3 | 1.4 (0.7, 2.0) |
1.2 (0.6, 1.9) |
0.9 (0.4, 1.5) |
|
| Less positive | ||||||
| None | 1,407 | 25.2 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 858 | 25.2 | −0.1 (−0.5, 0.4) |
−0.1 (−0.6, 0.3) |
−0.0 (−0.5, 0.3) |
|
| More severe | 772 | 26.0 | 0.7 (0.2, 1.3) |
0.7 (0.2, 1.3) |
0.7 (0.2, 1.1) |
|
| Extended family | 0.26 | |||||
| Extended family | ||||||
| None | 97 | 26.9 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 48 | 26.4 | −0.6 (−2.4, 1.3) |
−0.2 (−2.1, 1.8) |
0.2 (−1.6, 1.9) |
|
| More severe | 60 | 26.5 | −0.4 (−2.5, 1.7) |
−0.2 (−2.4, 1.9) |
0.7 (−1.2, 2.6) |
|
| No extended family | ||||||
| None | 3,341 | 25.0 | 0 (ref) | 0 (ref) | 0 (ref) | |
| Less severe | 1,506 | 25.2 | 0.2 (−0.1, 0.6) |
0.2 (−0.2, 0.5) |
0.1 (−0.2, 0.4) |
|
| More severe | 1,177 | 26.1 | 1.1 (0.7, 1.5) |
1.0 (0.6, 1.4) |
0.9 (0.5, 1.2) |
1+ type of less severe abuse (physical: pushed/grabbed/shoved; sexual: non-consensual touching; emotional: 75th percentile—<90th percentile score); no severe abuse.
1+ type of severe abuse (physical: kicked/punched/hit; sexual: forced sex; emotional: ≥90th percentile score).
Sample sizes for stratified models may differ due to missingness of individual modifier data.
Difference between BMI at age 25–32 years compared to those with no abuse.
Adjusted for baseline age, non-race, and perceived social status in adolescence (perceived social status not adjusted in models examining social status as a modifier).
DISCUSSION
In this cohort, exposure to severe childhood/adolescent abuse was associated with greater adult body weight and obesity risk. Notably, emotional abuse appeared to have the strongest association with BMI, whereas sexual abuse was not found to have an association. Once adjusted for emotional abuse, physical abuse associations were also close to the null. Prior results in this cohort have found sexual abuse to be modestly associated with BMI among female participants.24 Thus, the null finding is likely due to the inclusion of male participants. Prior findings have also found emotional abuse to be the type of abuse most strongly related to BMI. Several protective factors were found to be associated with reduced BMI in adulthood. None of these factors significantly attenuated the average weight gain associated with abuse.
There is limited literature with which to compare these findings. One study23 examined resilience to overweight associated with adverse family experiences, finding a modest attenuation of the adversity–weight status association when certain protective factors (e.g., neighborhood safety) were present. However, interactions were not significant. Findings of the present study are similar, with qualitative comparisons across stratified models suggesting a slight attenuation with the presence of certain protective factors, but no significant interaction detected. Numerous researchers have found a range of factors that appear to protect mental health after abuse.9,18–22,34–36 Results here indicate that similar factors may not improve weight outcomes in survivors of child abuse.
Strengths of this study include a large community-based cohort with substantial numbers of both male and female participants. With >15 years of follow-up starting in childhood, numerous potential resilience factors were measured and available for investigation. Finally, the longitudinal design allowed for clarity about the temporal order of childhood abuse and BMI, with adjustment for BMI in childhood to account for reverse causation.
Limitations
Several limitations of this study warrant attention. First, most of the potential modifiers were assessed at a single point in time via self-report, which may have prevented detection of a modifying effect. Nevertheless, these variables were almost all associated with BMI in the expected direction, suggesting they captured the domains of interest. Second, abuse during childhood and adolescence was retrospectively self-reported, with the potential for recall and social desirability biases. Although retrospective reports are potentially problematic,37 given the challenges of prospective assessment38 and known biases in official reports,39,40 retrospective self-reports are currently the standard in large epidemiologic studies41–43 and have been found to be similarly associated with obesity as prospectively assessed or objectively confirmed abuse.4 Third, BMI was also self-reported. Correlations between self-reported and measured BMI are very high,31 thus errors in self-report are unlikely to substantially impact results. Finally, the GUTS sample is largely White and consists of children of nurses, limiting generalizability of the results.
CONCLUSIONS
Abuse in childhood, particularly emotional abuse, appears to predict later life weight gain. Factors found to foster resilience to abuse-related mental health detriments were not found to have a buffering impact on weight gain. Continued investigations of ways to mitigate the negative health effects of abuse are critically needed, including investigation of potentially modifiable mediators such as depressive symptoms and disordered eating. Further, prevention of child abuse itself must remain a central public health priority.
Supplementary Material
ACKNOWLEDGMENTS
Dr. Mason’s time on this research was supported by the Building Interdisciplinary Research Careers in Women’s Health Grant (#K12HD055887) from the Eunice Kennedy Shriver National Institutes of Child Health and Human Development (NICHD), Office of Research on Women’s Health, and National Institute on Aging, NIH, administered by the University of Minnesota Deborah E. Powell Center for Women’s Health. The content is solely the responsibility of the authors and does not necessarily represent the office views of NICHD or NIH. Data collection was also supported by NIH grants R01 HD066963 and R01 HD049889.
Footnotes
The authors have no conflicts of interest to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Hughes K, Bellis MA, Hardcastle KA, et al. The effect of multiple adverse childhood experiences on health: a systematic review and meta-analysis. Lancet Public Health. 2017;2(8):e356–e366. 10.1016/s2468-2667(17)30118-4. [DOI] [PubMed] [Google Scholar]
- 2.Chen LP, Murad MH, Paras ML, et al. Sexual abuse and lifetime diagnosis of psychiatric disorders: systematic review and meta-analysis. Mayo Clin Proc. 2010;85(7):618–629. 10.4065/mcp.2009.0583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Maniglio R Child sexual abuse in the etiology of depression: a systematic review of reviews. Depress Anxiety. 2010;27:631–642. 10.1002/da.20687. [DOI] [PubMed] [Google Scholar]
- 4.Danese A, Tan M. Childhood maltreatment and obesity: systematic review and meta-analysis. Mol Psychiatry. 2013;19(5):544–554. 10.1038/mp.2013.54. [DOI] [PubMed] [Google Scholar]
- 5.Hemingsson E, Johansson K, Reynisdottir S. Effects of childhood abuse on adult obesity: a systematic review and meta-analysis. Obes Rev. 2014;15(11):882–893. 10.1111/obr.12216. [DOI] [PubMed] [Google Scholar]
- 6.Masten AS. Ordinary magic: resilience processes in development. Am Psychol. 2001;56(3):227–238. 10.1037//0003-066x.56.3.227. [DOI] [PubMed] [Google Scholar]
- 7.Gartland D, Riggs E, Muyeen S, et al. What factors are associated with resilient outcomes in children exposed to social adversity? A systematic review. BMJ Open. 2019;9(4):e024870. 10.1136/bmjopen-2018-024870. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Meng X, Fleury M-J, Xiang Y-T, Li M, D’Arcy C. Resilience and protective factors among people with a history of child maltreatment: a systematic review. Soc Psychiatry Psychiatr Epidemiol. 2018;53(5):453–475. 10.1007/s00127-018-1485-2. [DOI] [PubMed] [Google Scholar]
- 9.Afifi TO, MacMillan HL. Resilience following child maltreatment: a review of protective factors. Can J Psychiatry. 2011;56(5):266–272. 10.1177/070674371105600505. [DOI] [PubMed] [Google Scholar]
- 10.Hemmingsson E A new model of the role of psychological and emotional distress in promoting obesity: conceptual review with implications for treatment and prevention. Obes Rev. 2014;15(9):769–779. 10.1111/obr.12197. [DOI] [PubMed] [Google Scholar]
- 11.Michopoulos V, Powers A, Moore C, Villarreal S, Ressler KJ, Bradley B. The mediating role of emotion dysregulation and depression on the relationship between childhood trauma exposure and emotional eating. Appetite. 2015;91:129–136. 10.1016/j.appet.2015.03.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Palmisano GL, Innamorati M, Vanderlinden J. Life adverse experiences in relation with obesity and binge eating disorder: a systematic review. J Behav Addict. 2016;5(1):11–31. 10.1556/2006.5.2016.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tasca GA. Attachment and eating disorders: a research update. Curr Opin Psychol. 2019;25:59–64. 10.1016/j.copsyc.2018.03.003. [DOI] [PubMed] [Google Scholar]
- 14.Mason SM, Flint AJ, Field AE, Austin SB, Rich-Edwards JW. Abuse victimization in childhood or adolescence and risk of food addiction in adult women. Obesity (Silver Spring). 2013;21(12): E775–E781. 10.1002/oby.20500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Torres SJ, Nowson CA. Relationship between stress, eating behavior, and obesity. Nutrition. 2007;23(11–12):887–894. 10.1016/j.nut.2007.08.008. [DOI] [PubMed] [Google Scholar]
- 16.Dallman MF, Pecoraro NC, La Fleur SE. Chronic stress and comfort foods: self-medication and abdominal obesity. Brain Behav Immun. 2005;19(4):275–280. 10.1016/j.bbi.2004.11.004. [DOI] [PubMed] [Google Scholar]
- 17.Adam TC, Epel ES. Stress, eating and the reward system. Physiol Behav. 2007;91(4):449–458. 10.1016/j.physbeh.2007.04.011. [DOI] [PubMed] [Google Scholar]
- 18.Afifi TO, MacMillan HL, Taillieu T, et al. Individual- and relationship-level factors related to better mental health outcomes following child abuse: results from a nationally representative Canadian sample. Can J Psychiatry. 2016;61(12):776–788. 10.1177/0706743716651832. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Collishaw S, Pickles A, Messer J, Rutter M, Shearer C, Maughan B. Resilience to adult psychopathology following childhood maltreatment: evidence from a community sample. Child Abus Negl. 2007;31(3):211–229. 10.1016/j.chiabu.2007.02.004. [DOI] [PubMed] [Google Scholar]
- 20.Jaffee SR, Caspi A, Moffitt TE, Polo-Tomás M, Taylor A. Individual, family, and neighborhood factors distinguish resilient from non-resilient maltreated children: a cumulative stressors model. Child Abuse Negl. 2007;31(3):231–253. 10.1016/j.chiabu.2006.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Howell KH, Miller-Graff LE. Protective factors associated with resilient functioning in young adulthood after childhood exposure to violence. Child Abus Negl. 2014;38(12):1985–1994. 10.1016/j.chiabu.2014.10.010. [DOI] [PubMed] [Google Scholar]
- 22.Chandy JM, Blum RW, Resnick MD. Female adolescents with a history of sexual abuse: risk outcome and protective factors. J Interpers Violence. 1996;11(4):503–518. 10.1177/088626096011004004. [DOI] [Google Scholar]
- 23.Lynch BA, Finney Rutten LJ, Wilson PM, et al. The impact of positive contextual factors on the association between adverse family experiences and obesity in a national survey of children. Prev Med. 2018;116:81–86. 10.1016/j.ypmed.2018.09.002. [DOI] [PubMed] [Google Scholar]
- 24.Mason SM, MacLehose RF, Katz-Wise SL, et al. Childhood abuse victimization, stress-related eating, and weight status in young women. Ann Epidemiol. 2015;25(10):760–766.e2. 10.1016/j.annepidem.2015.06.081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Strauss MA, Gelles RJ. Physical Violence in American Families: Risk Factors and Adaptations to Violence in 8,145 Families. New Brunswick: Transaction Publishers; 1990. [Google Scholar]
- 26.Finkelhor D, Moore D, Hamby SL, Straus MA. Sexually abused children in a national survey of parents: methodological issues. Child Abuse Negl. 1997;21(1):1–9. 10.1016/s0145-2134(96)00127-5. [DOI] [PubMed] [Google Scholar]
- 27.Bernstein DP, Fink L, Handelsman L, et al. Initial reliability and validity of a new retrospective measure of child abuse and neglect. Am J Psychiatry. 1994;151(8):1132–1136. 10.1176/ajp.151.8.1132. [DOI] [PubMed] [Google Scholar]
- 28.Felitti VJ, Anda RF, Nordenberg D, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: the Adverse Childhood Experiences (ACE) Study. Am J Prev Med. 1998;14(4):245–258. 10.1016/s0749-3797(98)00017-8. [DOI] [PubMed] [Google Scholar]
- 29.Rich-Edwards JW, Mason S, Rexrode K, et al. Physical and sexual abuse in childhood as predictors of early-onset cardiovascular events in women. Circulation. 2012;126(8):920–927. 10.1161/circulationaha.111.076877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Strauss RS. Comparison of measured and self-reported weight and height in a cross-sectional sample of young adolescents. Int J Obes Relat Metab Disord. 1999;23(8):904–908. 10.1038/sj.ijo.0800971. [DOI] [PubMed] [Google Scholar]
- 31.Goodman E, Hinden BR, Khandelwal S. Accuracy of teen and parental reports of obesity and body mass index. Pediatrics. 2000;106(1):52–58. 10.1542/peds.106.1.52. [DOI] [PubMed] [Google Scholar]
- 32.Brisbois TD, Farmer AP, McCargar LJ. Early markers of adult obesity: a review. Obes Rev. 2012;13(4):347–367. 10.1111/j.1467-789x.2011.00965.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC Growth Charts for the United States: methods and development. Vital Health Stat 11. 2002;(246):1–190. [PubMed] [Google Scholar]
- 34.Williams J, Nelson-Gardell D. Predicting resilience in sexually abused adolescents. Child Abus Negl. 2012;36(1):53–63. 10.1016/j.chiabu.2011.07.004. [DOI] [PubMed] [Google Scholar]
- 35.Bellis MA, Hardcastle K, Ford K, et al. Does continuous trusted adult support in childhood impart life-course resilience against adverse childhood experiences - a retrospective study on adult health-harming behaviours and mental well-being. BMC Psychiatry. 2017;17(1):110. 10.1186/s12888-017-1305-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Banyard VL, Williams LM. Women’s voices on recovery: a multi-method study of the complexity of recovery from child sexual abuse. Child Abus Negl. 2007;31(3):275–290. 10.1016/j.chiabu.2006.02.016. [DOI] [PubMed] [Google Scholar]
- 37.Reuben A, Moffitt TE, Caspi A, et al. Lest we forget: comparing retrospective and prospective assessments of adverse childhood experiences in the prediction of adult health. J Child Psychol Psychiatry. 2016;57(10):1103–1112. 10.1111/jcpp.12621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Shenk CE, Noll JG, Peugh JL, Griffin AM, Bensman HE. Contamination in the prospective study of child maltreatment and female adolescent health. J Pediatr Psychol. 2016;41(1):37–45. 10.1093/jpepsy/jsv017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Gilbert R, Widom CS, Browne K, Fergusson D, Webb E, Janson S. Burden and consequences of child maltreatment in high-income countries. Lancet. 2009;373(9657):68–81. 10.1016/s0140-6736(08)61706-7. [DOI] [PubMed] [Google Scholar]
- 40.MacMillan HL, Jamieson E, Walsh CA. Reported contact with child protection services among those reporting child physical and sexual abuse: results from a community survey. Child Abuse Negl. 2003;27(12):1397–1408. 10.1016/j.chiabu.2003.06.003. [DOI] [PubMed] [Google Scholar]
- 41.Rich-Edwards JW, Mason S, Rexrode K, et al. Physical and sexual abuse in childhood as predictors of early-onset cardiovascular events in women. Circulation. 2012;126(8):920–927. 10.1161/circulationaha.111.076877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Shin SH, Miller DP. A longitudinal examination of childhood maltreatment and adolescent obesity: results from the National Longitudinal Study of Adolescent Health (AddHealth) Study. Child Abuse Negl. 2012;36(2):84–94. 10.1016/j.chiabu.2011.08.007. [DOI] [PubMed] [Google Scholar]
- 43.Cohen AK, Chaffee BW, Rehkopf DH, Coyle JR, Abrams B. Excessive gestational weight gain over multiple pregnancies and the prevalence of obesity at age 40. Int J Obes (Lond). 2014;38(5):714–718. 10.1038/ijo.2013.156. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.