Psychiatric comorbidity of childhood obesity

Abstract

The onset of psychiatric symptoms and disorders is relatively common in childhood, occurring among youths across the weight spectrum. However, available research suggests that certain psychiatric comorbidities are more prevalent in obese children and adolescents than in healthy weight youths. First, we review research on disordered eating, including evidence to suggest that loss of control eating is associated with weight gain and obesity in youths, as well as poor outcome in family-based treatment of paediatric obesity. Second, we highlight evidence on the relationship between depression and obesity, especially in girls. Third, we present data on attention deficit hyperactivity disorder (ADHD), particularly the symptoms of impulsivity and inattention, and childhood obesity. We also consider that some medical conditions and psychotropic medications contribute to weight gain and obesity in children and adolescents. Throughout the review, we emphasize that psychiatric comorbidity may be a cause or consequence of childhood obesity, or they may share common aetiological factors.

Psychiatric comorbidity of childhood obesity

As children progress through stages of physical and emotional development, the relationship between psychiatric symptoms and body weight may evolve over time, and may vary by gender and minority status. Although the rapid increases in childhood obesity seen in the 1980s and 1990s in the USA appear to have levelled off, the prevalence remains high. Results from the National Health and Nutrition Examination Survey (NHANES) in 2009–2010 (Ogden et al., 2012) indicate that the overall prevalence of obesity in children and adolescents remains stable at approximately 17%, although increases in obesity prevalence may be continuing among male adolescents. NHANES has also consistently reported significant differences in obesity by race/ethnicity, with higher prevalence among black and Hispanic youths as compared to non-Hispanic white youths.

Many psychiatric disorders begin in childhood. Results from the NHANES in 2001–2004 (Merikangas et al., 2010a) documented 12-month prevalence estimates of Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV; American Psychiatric Association, 2000) disorders of 8.6% for attention deficit/hyperactivity disorder (ADHD), 3.7% for mood disorders, 2.1% for conduct disorder, 0.7% for panic disorder or generalized anxiety disorder, and 0.1% for eating disorders among youths aged 8 to 15 years. There were differences in psychiatric disorders by gender, with 2.1 times greater prevalence of ADHD in boys than girls, and 2 times greater rates of mood disorders in girls than boys. Additionally, the National Comorbidity Survey–Adolescent Supplement indicated that approximately 22% of US adolescents meet lifetime criteria for a mental disorder with severe impairment (Merikangas et al., 2010b). Age of onset was earliest for anxiety (6 years), followed by behaviour (11 years), mood (13 years), and substance use disorders (15 years). Certainly, rates of clinically significant symptoms are more common than threshold DSM-IV disorders.

Although the co-occurrence of psychiatric disorders and obesity is to be expected based on the prevalence of each problem, there is evidence of higher rates of certain psychiatric disorders and symptoms in obese youths than in healthy weight youths. We review the literature on the psychiatric comorbidity of childhood obesity with an additional objective of identifying important gaps in the literature and suggesting future directions. We note when data have been collected from community samples, as well as when research has been conducted with clinical samples seeking obesity or mental health treatment, as ascertainment biases may affect generalizability of results. Herein, we focus on the most common psychiatric comorbidities of childhood obesity including disordered eating, depression, and ADHD.

Disordered eating in obese children and adolescents

Obesity, defined as excess adiposity, is a heterogeneous condition of complex and varying aetiology. Nevertheless, energy intake in excess of energy expenditure is inevitably involved in weight gain, making aberrant eating a focus of interest. Binge eating, defined in the DSM-IV as consumption of an objectively large amount of food accompanied by a sense of loss of control (LOC) over eating during the episode, is a common form of disordered eating among obese individuals. However, it has proven difficult to quantify a large amount of food for growing children and adolescents. Therefore, LOC eating, defined as the subjective experience of uncontrolled consumption of an amount of food that may or may not be objectively larger than most would eat given the circumstances, has been utilized in lieu of binge eating to identify aberrant eating behaviour in children (Marcus & Kalarchian, 2003). LOC eating has been reported by 2–10% of children and adolescents among community samples (Wolkoff et al., 2011).

LOC eating has been associated with other eating disorder symptoms in children. Tanofsky-Kraff et al. (2004) interviewed non-treatment-seeking children aged 6 to 13 years and established that LOC eating is more prevalent in overweight children and associated with more maladaptive cognitions related to eating, shape and weight. An ecological momentary assessment study in youths aged 8 to 13 years also showed that LOC eating episodes were preceded and followed by disordered cognitions about food/eating and body image (Hilbert et al., 2009). Thus LOC is hypothesized to be a precursor of the development of partial or full syndrome DSM-IV eating disorders.

Data also suggest that LOC eating may predispose children to weight gain and obesity (Tanofsky-Kraff et al., 2006, 2009c). The mechanisms are not fully understood, but a preference for palatable foods during LOC episodes may place children at risk for excess energy intake long-term. Although total energy intake has not differed between LOC eating and other types of eating episodes in youths, children's descriptions of the foods consumed during LOC eating indicate a higher percentage of calories from dessert and snack-type foods (Theim et al., 2007). Data from laboratory feeding studies also indicate that youths with LOC consume a greater percentage of energy intake from carbohydrate and less from protein, largely due to consumption of palatable sweet foods, than youths without LOC (Tanofsky-Kraff et al., 2009b; Wolkoff et al., 2011). These findings appear consistent with animal models in which “binge eating prone” rats have been characterized by heightened motivation for palatable food, as measured by the voluntary tolerance of increasing levels of foot shock delivered after the retrieval of the food (Oswald et al., 2011). Furthermore, evidence suggests that a genetic vulnerability to both LOC eating and selection of energy-dense palatable foods may increase risk for obesity. For example, children and adolescents with 1 or 2 FTO rs9939609 obesity-risk alleles report more frequent LOC eating episodes and select foods higher in fat at a buffet meal (Tanofsky-Kraff et al., 2009a).

Parental and environmental factors have been shown to play a role in development and maintenance of eating and weight problems in children. For example, mothers’ binge eating has been indirectly related to children's adiposity through children's LOC eating, suggesting that mothers and children may share similar aberrant eating patterns (Zocca et al., 2011). Additionally, maladaptive patterns of family functioning during mealtimes have been associated with LOC eating in children (Czaja et al., 2011). There has also been some preliminary support for an interpersonal pathway in which social problems lead to negative affect which, in turn, precipitates binge eating episodes among non-treatment seeking youths (Elliott et al., 2010).

Although dieting is strongly associated with binge eating in adults, and efforts to limit caloric intake are regarded as an important factor in the onset and maintenance of binge eating in cognitive behavioural models, dieting has not been consistently linked to LOC eating in youths. Morgan et al. (2002) found that 33% of a community-based sample of overweight children 6–10 years of age reported LOC eating. Consistent with other findings, children with LOC had significantly higher body mass index (BMI) and more adiposity than those who did not. After controlling for BMI, children with LOC reported more anxiety, depressive symptoms, negative moods, and body dissatisfaction than children with no LOC; they also were more likely to endorse that weight and shape played an important role in self-evaluation. However, of particular note, there were no differences between children who did and did not report LOC in dieting and distress about overeating. Thus, dietary restriction may not be a key factor in the aetiology of LOC eating in overweight children. Rather, dieting may be a consequence of aberrant eating as an effort to gain control over eating.

Available data suggest that youths with binge eating behaviours present for obesity treatment with distinctive needs. Among youths aged 10–16 years seeking residential treatment for obesity, binge episodes were reported by 36.5%, with 6% reporting two or more episodes per week (Decaluwe et al., 2002). Obese children and adolescents with and without binge eating did not differ in degree of adiposity, dietary restraint or depressive symptoms, but binge eaters reported more eating, weight and shape concerns, as well as lower self-esteem. Data from our research group also have suggested that child binge eating is related to poor outcome in a family-based treatment of paediatric obesity (Wildes et al., 2010). Among 192 severely obese children aged 8–12 years who were randomized to family-based treatment or usual care, 22 children (11.5%) endorsed binge eating at baseline. Children with binge eating were younger, had more depressive, anxiety and eating disorder symptoms, as well as lower self-esteem, than those without binge eating. Moreover, among children receiving the lifestyle intervention, those with binge eating showed a 2.6% increase in the percentage overweight at the completion of acute intervention, as compared to an 8.5% decrease among children without binge eating. These results suggest that targeting LOC may be necessary to improve childhood obesity treatment outcomes.

In summary, disordered eating in obese children and adolescents, particularly LOC, is relatively common. LOC eating is associated with excess adiposity and psychopathology, as well as poorer response to treatment of childhood obesity. However, surprisingly little is known about aberrant eating in minority youths, despite elevated rates of obesity in black and Hispanic children and adolescents (Ogden et al., 2012). In a recent investigation, data from 16,978 children from 47 Massachusetts middle schools participating in the Healthy Choices overweight prevention study were used to estimate the odds of disordered weight control behaviours (DWCB; defined as vomiting or use of laxatives or diet pills in the past month to control weight) in minority youths compared with their white peers (Austin et al., 2011). Among white youths, 2.7% of girls and 2.3% of boys reported DWCB. However, the odds of DWCB were elevated 2–10 times in most racial and ethnic groups relative to whites. Disparities were attenuated, but persisted, after controlling for multiple individual- and school-level factors. More research is needed to better understand the pathogenesis and psychopathology associated with a broad range of disordered eating behaviours including LOC eating in diverse samples of boys and girls.

Depression and anxiety in obese children and adolescents

Research suggests there is a reciprocal link between depression and obesity such that obesity increases the risk for depression among adults, but also depression predicts the development of obesity (Luppino et al., 2010). Thus, psychiatric disorders and symptoms may be a consequence of being obese in a culture that stigmatizes obesity. Alternatively, psychiatric disorders may contribute to the development of obesity in vulnerable individuals. For example, atypical depression is characterized by increased appetite and overeating, along with hypersomnia and inactivity, and thus recurrent episodes could contribute to weight gain over time. Obesity and depression may have shared neurobiological underpinnings and/or environmental risk factors such as adverse childhood experiences.

Available studies suggest greater psychopathology among obese than non-obese youths. For example, in a small sample of adolescents in the USA, more than half of obese participants had a DSM-IV diagnosis (most frequently major depressive disorder), which was higher than a non-obese comparison group (Erermis et al., 2004). Similarly, in a Dutch population-based study of adolescents who responded to a classroom-based Internet questionnaire, obese boys and girls were more likely to report suicidal thoughts and attempts, and to be classified as psychologically unhealthy than were normal weight students (van Wijnen et al., 2010).

Conversely, studies also suggest weight gain and obesity are common among depressed youths. A prospective cohort study of adolescents in grades 7 through 12 who completed an interview for the National Longitudinal Study of Adolescent Health indicated that depressed adolescents are at increased risk for the onset and persistence of obesity during adolescence (Goodman & Whitaker, 2002). Moreover, a systematic literature review (Leim et al., 2008) has documented a positive association between depressive symptoms at age 6 to 19 years and overweight status up to 15 years later. Obesity also appears to be common among children and adolescents seeking psychiatric treatment. For example, in a tele-psychiatry clinic, approximately 55% of child and adolescents seen for consultation were overweight or obese, with the most common diagnoses of depression and bipolar disorder (Marks et al., 2009). Although the causal mechanisms are not fully understood, they likely involve a complex interplay of biological, psychological and social factors.

Similar to findings from adult samples (Kendler et al., 1996), paediatric data suggest girls may be particularly vulnerable to comorbid mood and eating problems. For example, in a study of childhood to adulthood weight change as part of the Children in the Community Study, depression and anxiety disorders were associated with higher subsequent female BMI z-scores, but not male (Anderson et al., 2006). Additionally, female obese adolescents were at increased risk for the development of depression or anxiety disorders (Anderson et al., 2007). Thus there may be a behavioural phenotype characterized by aberrant eating, depression and obesity among girls.

Better characterization of the intricate interplay among mood, eating and body weight may inform efforts to prevent and treat childhood obesity. For example, it appears that one pathway between depression and obesity in girls is mediated by binge eating. This is supported by a longitudinal study of adolescent girls that verified that binge eating predicted onset of obesity (Stice et al., 2002). Additionally, depressive symptoms emerged along with a number of other psychosocial and biological risk factors for binge eating. Depressive symptoms interacted with other risk factors, suggesting qualitatively different pathways to binge eating, with certain subgroups at increased risk. Results appear consistent with a model in which depression in girls may increase risk for aberrant eating behaviour, which in turn may contribute to the onset or exacerbation of overweight or obesity. Accordingly, early identification and treatment for mood disturbance may mitigate disordered eating and weight gain in vulnerable youths.

Attention deficit hyperactivity disorder (ADHD) in obese children and adolescents

Obesity has been linked to ADHD, a psychiatric disorder marked by symptoms of inattention, hyperactivity or impulsivity. A review of the literature (Cortese et al., 2008) concluded that samples of obese children referred for obesity treatment present with a higher than expected prevalence of ADHD. Additionally, although the data were limited, all studies included indicated that subjects with ADHD were heavier than expected.

Data from community samples also have documented associations between ADHD and obesity. For example, a study of a representative US sample of 62,887 youths aged 5 to 17 years participating in the National Survey of Children's Health, examined the prevalence of overweight in relation to both attention deficit disorder (ADD)/ADHD and medication use (Waring & Lapane, 2008). After adjustment for age, gender, race/ethnicity, socio-economic status, and depression/anxiety, children and adolescents with ADHD not currently using medication had ∼1.5 times the odds of being overweight, and those currently medicated had ∼1.6 times the odds of being underweight, as compared to children and adolescents without ADHD. Moreover, children and adolescents with ADHD were ∼9 times as likely to report depression or anxiety. Additionally, a cross-sectional, nationally representative German sample of 2,863 parents and their children aged 11 to 17 found associations between overweight and ADHD (Erhart et al., 2012). Results indicated that the prevalence of ADHD was significantly higher for overweight or obese (7%) than normal weight (3.5%) and underweight (4.9%) children. Controlling for age, gender, and socio-economic status, overweight or obese children were twice as likely to have an ADHD diagnosis. Conversely, after adjusting for salient covariates, children with ADHD had an odds ratio of 1.9 for overweight/obesity status. Thus, a child with ADHD may be at risk for becoming overweight, and an overweight child is at risk for a diagnosis of ADHD.

Cross-sectional data have yielded information on the potential mechanism underlying the comorbidity of ADHD and problems with eating and weight. In a French clinical sample of 99 severely obese adolescents aged 12 to 17 years (Cortese et al., 2007), bulimic behaviours were significantly associated with ADHD symptoms after controlling for depressive and anxiety symptoms; the degree of overweight, pubertal stage, age and gender were not associated with ADHD symptoms. Moreover, on the Conners Parent Rating Scale, there was a significant association between bulimic behaviours and ADHD index score (which measures symptoms of inattention, impulsivity and hyperactivity taken together) and a lack of association with the hyperactivity-impulsivity subscale (which contains only one item on impulsivity). Thus, Cortese et al. (2007) suggest that the association between ADHD symptoms and bulimic behaviours may be accounted for by impulsivity and inattention rather than hyperactivity. Screening obese children for symptoms of ADHD and impulsive eating may identify a group who would benefit from behavioural self-management interventions or stimulant treatment. Similarly, screening children with ADHD for obesity and impulsive eating behaviours may prove useful. Longitudinal studies are needed to better characterize which dimensions of ADHD (inattention, hyperactivity or impulsivity) drive the relationships with eating and body weight.

Chronic health conditions and obesity

Numerous chronic health conditions are associated with obesity in children. For example, in a study of children with a wide array of chronic conditions using data from the National Survey of Children's Health (Chen et al., 2010), youths aged 10 to17 with certain conditions were at increased risk for obesity compared to children without chronic illnesses. Specifically, the prevalence of obesity was 19.7% among children with asthma, 18.4% among children with a hearing/vision condition, 19.3% among those with a learning disability, 23.4% for youths with autism, and 18.9% among children with ADHD, as compared to 12.2% among those without a condition. Differences in the prevalence of obesity persisted after adjustment for age, sex, race/ethnicity, family income, family structure, parental education, and region, suggesting that the findings were not attributable to these factors.

Data further suggest that children with obesity and non-obesity related chronic health conditions differ in psychiatric comorbidity. In a study using Medicaid claims data from Florida, 35% of youths aged 5 to 18 years with chronic obesity-related health conditions had a psychiatric diagnosis (Janicke et al., 2008). Moreover, after controlling for age, gender and race, youths with type 2 diabetes, metabolic syndrome, and dyslipidaemia had higher rates of International Classification of Disease, ninth revision (ICD-9; Health Care Financing Administration, 1989) psychiatric diagnosis than children with comparison conditions including cystic fibrosis, sickle cell disease, and juvenile rheumatoid arthritis.

More work is needed to explicate how certain chronic conditions and disease processes may lead to increased energy intake and decreased energy expenditure in children. For example, some children with ADHD may have difficulty with self-regulation and impulsive eating, and children with autism frequently engage in selective eating behaviours that can contribute to excess calorie consumption. Other conditions, such as asthma, can pose limitations on physical activity. Additionally, many chronic conditions are managed with medications that can lead to long-term weight gain and obesity-related health problems.

Psychiatric medications and obesity

Certain medications can promote or exacerbate childhood obesity, in particular psychotropic medications. A growing body of evidence indicates that youths are especially vulnerable to antipsychotic medication induced weight gain and cardiometabolic side effects. In a prospective study (Correll et al., 2009), first time use of second generation antipsychotics was associated with significant weight gain and variable metabolic changes across four different medications. Youths aged 4–19 years (N = 338) with mood spectrum (47.8%), schizophrenia spectrum (22.1%), and disruptive or aggressive behaviour spectrum (22.1%) disorders were enrolled within a week of beginning medication. After 10.8 weeks of treatment, weight increased by 8.5 kg with olanzapine, 6.1 kg with quetiapine, 5.3 kg with risperidone, and 4.4 kg with aripiprazole, as compared to minimal weight change in an untreated comparison group. More than half of children initiating one of the four medications gained more than 7% of their initial body weight within a relatively short period, and overall 10 to 36% of patients transitioned to overweight or obese status. Findings are particularly salient as these drugs are relatively widely prescribed, including off-label use, in children. Youths prescribed psychotropic medications may benefit from careful consideration of medication side effect profiles, frequent medical monitoring, and intervention to prevent or manage weight gain.

Conclusion

Childhood obesity is marked by significant psychiatric comorbidity. Available data suggest that obesity is related to numerous disorders and symptoms in youths, most notably LOC eating, depression, and ADHD. It is not clear whether psychiatric comorbidity is a cause or consequence of obesity, or whether common factors promote the expression of both obesity and psychiatric disturbance in susceptible children. Current research suggests that aberrant eating may characterize a particular obese phenotype, particularly in girls, and partially explain the overlap between psychiatric disturbances and obesity. Hypothesis-driven longitudinal studies of children of diverse backgrounds may help youths who are at risk for adverse health outcomes. Targeting intervention efforts may improve both the physical and mental health of children and prevent the persistence of health problems into adulthood.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.