BMI Trajectories Across Adolescence: Investigating the Role of Maltreatment Type and Cortisol Reactivity

• In Event: 3-069 - Multimodal Investigation of Childhood Adversity, Stress System Regulation, and Adolescent Physical and Mental Health

Integrative Statement

Background: Obesity remains a public health concern for youth. Child maltreatment increases risk for obesity in some studies, but the effect is conditional on age or maltreatment type. One potential mechanism whereby maltreatment may affect obesity is through the impact on cortisol reactivity. While there is consistent evidence that maltreatment affects cortisol reactivity, and that cortisol is linked with BMI, no studies have prospectively followed maltreated individuals across adolescence to examine how stress physiology affects BMI trajectories. Consideration of maltreatment type may reconcile the disparate findings regarding links between maltreatment and obesity. Therefore, this study examined whether the association between maltreatment type (sexual, physical, emotional, and neglect) and BMI trajectories was conditional on cortisol reactivity among a sample of adolescents referred to child welfare for maltreatment and a comparison group from the same neighborhoods.
Methods: Data came from the first four waves of an ongoing longitudinal study examining the effects of maltreatment on adolescent development. At Time 1 the sample was comprised of 454 adolescents aged 9-13 years (212 females). Time 2, 3, and 4 interviews occurred approximately 1, 2.5, and 4.5 years after Time 1. At Time 4, 73% of the original sample were retained. Sex-specific BMI percentiles were calculated at each time point with measured height and weight. Cortisol reactivity at Time 1 was calculated as Area Under the Curve. Multilevel modeling with random intercepts and slopes was used to model BMI percentile trajectories arrayed by age, separately for boys and girls. Unconditional and conditional linear and quadratic growth models of BMI percentile across development were calculated and time-invariant covariates ethnicity/race, pubertal stage, and depression and anxiety symptoms were included in the models. Lastly, interaction terms were included to test whether cortisol reactivity at Time 1 moderated the association between maltreatment type and BMI trajectories.
Results: In girls, cortisol reactivity moderated the association between maltreatment type and quadratic change in BMI for girls with sexual abuse compared to girls with physical abuse (.90, 95% CI .05 to 1.74, p < .05) and comparison girls (.74, 95% CI .01 to 1.47, p < .05). At low levels of cortisol, girls with sexual abuse had a steeper quadratic increase in BMI compared to comparison girls (-.38, 95% CI -.65 to -.10, p < .01). At high levels of cortisol, girls with sexual abuse did not differ from girls with physical abuse (-.09, 95% CI -.69 to .51; p > .05) or comparison girls (.02, 95% CI -.57 to .62, p > .05) in quadratic change in BMI. In boys, cortisol reactivity did not moderate the association between maltreatment type and BMI growth.
Conclusions: Girls with sexual abuse had a faster rate of growth in BMI across late adolescence than comparison girls but only at low levels of cortisol reactivity. It appears that sexual abuse may be associated with a pattern of hypo-reactivity that increases risk for obesity in girls. This finding can be used by clinicians and interventionists to identify girls at greatest risk for obesity following sexual abuse experiences.

Authors

• Sonya Negriff, Kaiser Permanente Southern California
  Presenting Author

• Melissa Peckins, University of Michigan
  Non-Presenting Author

• Janet U. Schneiderman, University of Southern California
  Non-Presenting Author

• Elana B. Gordis, University at Albany, SUNY
  Non-Presenting Author

• Elizabeth J. Susman, The Pennsylvania State University
  Non-Presenting Author