A longitudinal examination of childhood maltreatment and adolescent obesity: Results from the National Longitudinal Study of Adolescent Health (AddHealth) Study

Abstract

Objectives

We sought to explore the association between childhood maltreatment (e.g., neglect, physical and sexual abuse) and longitudinal growth trajectories of body mass index (BMI) from adolescence to young adulthood.

Methods

We used latent curve modeling to examine data from the National Longitudinal Study of Adolescent Health (N = 8,471), pooling years 1995–2008. Relative weight was indicated by BMI.

Results

Children who experienced neglect had a faster average rate of BMI growth over time compared to children who experienced no childhood maltreatment. In addition, the co-occurrence of childhood neglect and physical abuse was related to increased levels of BMI at baseline, even after adjustment for 14 separate child and parental demographic and psychosocial characteristics. Parental obesity and education, child birth place, and family income were also associated with both baseline BMI and the rate of change in BMI over time.

Conclusion

Childhood neglect plays a role in the development of adolescent obesity. Future research should examine potential mediators or moderators of the longitudinal relation between childhood neglect and adolescent and young adult obesity including impulsivity, depression, and compulsive eating behavior.

Introduction

Childhood maltreatment (CM) increases an individual's risk for a variety of physical and mental health problems throughout the lifespan, including cardiovascular disease (Batten, Aslan, Maciejewski, & Mazure, 2004), lowered immunity (Shirtcliff, Coe, & Pollak, 2009), impaired brain development (Choi et al., 2009, De Bellis and Thomas, 2003), physical complaints (e.g., allergies, diabetes, arthritis, asthma, bronchitis, high blood pressure, and ulcers; Rich-Edwards et al., 2010, Springer et al., 2007), neurobehavioral and cognitive symptoms (e.g., attention deficit, poor executive functioning; Beers & De Bellis, 2002) and emotional/behavioral problems (Shin, 2005, Widom et al., 2007). Recent studies have also reported that CM may be associated with obesity. Specifically, these studies found a relationship between CM and obesity during adulthood in clinical settings (Felitti, 1993, Grilo and Masheb, 2001, Grilo et al., 2005, Gustafson et al., 2006), health insurance data (Anda et al., 2006, Bonomi et al., 2008, Williamson et al., 2002), and community samples (Alvarez et al., 2007, Fuemmeler et al., 2009, Rohde et al., 2008). In addition, the importance of studying the effects of CM on obesity is highlighted by the alarming prevalence of obesity in the US Data from 2008 indicating that 33.8% of adults were obese and an additional 34.3% were overweight, representing considerable increases from a few decades ago (Flegal, Carroll, Ogden, & Curtin, 2010).

While previous studies suggest that CM may contribute to an increased risk of adult obesity, few studies have examined the association between CM and obesity in adolescence. Examination of potential predictors of obesity during adolescence is particularly important, as obesity at this age is highly predictive of adult obesity and the serious morbidity and mortality that attend to it (Whitaker, Pepe, Seidel, & Dietz, 1997). Maltreated children suffering from a variety of physiological and psychological challenges as sequelae of CM may be also vulnerable to obesity and overweight problems during adolescence (Gustafson & Sarwer, 2004). Furthermore, a few existing studies on the associations between CM and later obesity have been limited by their cross-sectional design. As a result, while an emerging number of studies have examined the relationship between CM and adult obesity, less seems to have been done to explore the longitudinal relationship between CM and body mass index (kg/m²; BMI) from adolescence to young adulthood. If CM has adverse effects on adolescent obesity, it is possible that consequences of CM persist into young adulthood through increased BMI scores in adolescence. Identification of intraindividual variation in adolescent BMI and its relation to CM is critical to better understand how CM may operate on course of BMI from adolescence to young adulthood.

Given the deleterious effects of CM on public health and the increased risk of adult obesity among victims of CM, the present study investigated the long-term effects of CM on BMI trajectories from adolescence to young adulthood. In specific, we conducted latent curve modeling (LCM) analyses of CM and adolescent and young adult BMI using a nationally representative sample in the US LCM is a commonly used growth modeling approach, which allows each individual in the sample to have distinct patterns of change over time, and provides an estimate of the average initial level and growth rate of adolescent BMI (Bollen & Curran, 2006). While few studies have examined the longitudinal relationship between CM and adolescent BMI, understanding the extent to which CM predicts initial status and growth rates of BMI will contribute to the development of secondary prevention and intervention efforts targeting CM victims at risk for adolescent and adult obesity. Furthermore, our study adds two notable contributions to the existing literature. First, considering the relationship between CM and BMI, we controlled for a large number of potential confounders. Most significantly, our analyses controlled for the number of obese parents each adolescent had, which research has consistently demonstrated to be associated with an increased likelihood of a child being obese in adolescence and adulthood (Whitaker et al., 1997). Furthermore, in order to assess whether CM posed an environmental risk that increases susceptibility for higher BMI among children who are already at risk of being overweight or obese, we included an interaction between CM and number of obese parents in the model testing. Finally, emerging studies also suggested that frequency (number of times each maltreatment episode occurred) of CM may be equally important to presence of CM in understanding the effect of CM on adolescent obesity (Litrownik et al., 2005, Manly et al., 1994). The present study attempted to examine how frequency of CM was associated with change in BMI over time in a separate model testing.