Associations among overeating, overweight, and attention deficit/hyperactivity disorder: A structural equation modelling approach

Abstract

Background

Some recent studies have reported strong links between obesity and ADHD in adults; however, to date, little work has focussed on possible behavioural mechanisms that could account for this association.

Method

This study used structural equation modelling (SEM) in a sample of healthy adult women to test the predictions that ADHD symptoms predict aspects of overeating, including binge eating and emotionally-induced eating, which in turn are positively correlated with Body Mass Index.

Results

The SEM produced a non-significant chi-square and both the measurement model and the structural model fit the data very well.

Conclusions

Plausible mechanisms are discussed to help explain how the symptomatology of ADHD could foster different forms of overeating.

Introduction

Attention-Deficit/Hyperactivity Disorder (ADHD) was originally seen as a psychiatric disturbance of childhood since the developmentally inappropriate signs of distractability, impulsiveness, and inattention typically appear before the age of seven. However, we now recognize that in a large percentage of cases these symptoms persist into adulthood (Barkley et al., 2002, Kordon & Kahl, 2004), prompting the re-conceptualization of ADHD as a “life span disorder” (McGough & Barkley, 2004).

Recently, some intriguing—albeit limited—research has reported strong links between ADHD and obesity. For example, two studies found a substantially higher than expected prevalence of ADHD in adults receiving treatment for obesity; and an even greater occurrence (close to half the sample) in those with Class III obesity (BMI > 40) (Altfas, 2002, Fleming & Levy, 2002). The ADHD–obese also had more clinic visits and longer treatment duration than their non-ADHD counterparts. In the only study that assessed body size in children with ADHD¹, the prevalence of overweight and obesity was significantly greater than in the age-matched population (Holtkamp et al., 2004). At first glance, these latter results are rather counterintuitive since a principal characteristic of ADHD in children is physical restlessness and hyperactivity.

Interestingly, the enhanced novelty-seeking, the aversion to delayed reward, and the diminished sensitivity to negative feedback that are characteristic of adolescents and adults with ADHD are remarkably reminiscent of the immature behaviour of young children. In the normal process of maturation to adulthood, activity in brain reward centres, which are central to approach behaviour, tend to diminish while activity in the amygdala, which is central to avoidant behaviour, increases (Ernst, 2004). Any interruption or deviance in the trajectory of these developmental changes seems to confer risk for disorders like ADHD where individuals show a poor capacity for processing the appropriate motivational value of stimuli (see Johansen, Aase, Meyer & Sagvolden, 2002).

Indeed, individuals with ADHD respond quite differently to rewards and punishments compared to their healthy counterparts. For example, they show diminished activation in the hippocampus when engaged in decision-making tasks that involve weighing the pros and cons between small immediate rewards and larger future rewards (Ernst et al., 2003). One interpretation of these differences is a weaker ability to code the reinforcing properties of stimuli and/or a defective memory of the incentive value of rewarding stimuli in ADHD. Those with ADHD also tend to respond more readily to immediate sensory stimuli than to engage in complex processing when required to make choices that involve a dilemma and that direct behaviour. Of relevance to the present discussion is our recent finding (Davis, Levitan, Muglia, Kennedy, & Bewell, 2004b) that overweight and obese women also showed decision-making deficits (on a validated computer task) similar to those reported in previous studies with substance dependent individuals (e.g. Bechara & Damasio, 2002, Bechara et al., 2002).

The most popular drug treatments for ADHD (e.g. methylphenidate) enhance brain dopamine in the mesocorticolimbic pathway—a key neurochemical substrate of reward (Spencer et al., 2005). The reinforcement-related aspect of these stimulant treatments, plus a host of molecular genetic (see Bobb, Castellanos, Addington & Rapoport, 2005) and neuroimaging studies (see Sergeant, Geurts, & Oosterlaan, 2002) have given rise to a dopamine-dysfunction explanation for the aetiology of ADHD. Hypo-dopaminergic brain activity creates what some have termed a Reward Deficiency Syndrome [RDS] (Blum & Noble, 2001)—a condition which occurs when there is an inadequacy in the “brain reward cascade”, and is characterized by anhedonia and diminished motivation. Causal factors include sub-optimal genetic variants of the dopamine system, chronic psychomotor stimulant drug use, and/or a host of environmental toxins (Johansen et al., 2002). In essence, a person with this condition needs a regular ‘dopamine fix’ to feel good (Blum & Braveman, 2001). Consequently, an under-activation of brain reward circuits could facilitate various self-medicating behaviours as a way for the individual to compensate temporarily for this deficit. The relatively high co-morbidity between ADHD and substance dependence (e.g. Latimer et al., 2004, Saules et al., 2003, Smith et al., 2002) tends to support the RDS view of ADHD.

Recent evidence suggests that natural rewards such as eating are actually among the most commonly used methods of self-medication (see Davis, Strachan, & Berkson, 2004a). Food, especially when it is calorically dense and highly palatable, is a reliable activator of dopamine pathways (Salamone et al., 2003, Wang et al., 2004). Sweet and fatty foods also have the potential for abuse and dependence when used beyond basic energy requirements (Avena et al., 2005, Wellman, 2005). Moreover, as a form of mood enhancement, food is eminently satisfactory; it is easily available, relatively inexpensive, and legal.

To date, the few studies of obesity and ADHD have been limited to estimates of prevalence in clinical samples, and have not focused on behaviours which might link ADHD symptoms to increased body size. One obvious place to search for mechanisms is by examining whether ADHD contributes to the obesity risk profile because it fosters a tendency to overeat. However, ADHD is associated with high rates of psychiatric co-morbidity, especially in adults—e.g. major depression, bipolar disorder, and generalized anxiety disorder (see Faraone, 2005). Therefore, it would be difficult in a clinical sample to disentangle which symptoms and which disorders provide the causal links between ADHD and obesity.

One strategy that avoids the potential confounds of clinical research is to examine path associations among ADHD symptoms, aspects of overeating, and body weight in healthy participants from the general population. This approach is predicated on the assumption that personality factors and symptoms of disorder are best conceptualized dimensionally and occur with normal variation in the general population (see Claridge & Davis, 2003). In the present study we used structural equation modelling (SEM) to test our predictions that ADHD symptoms predict overeating behaviours, which in turn predict overweight and obesity (see Fig. 1). The first two constructs in the path diagram were modelled as multi-factorial latent variables. ADHD had two observed measures: retrospective recall of childhood ADHD symptoms and current impulsivity traits. Since a diagnosis of adult ADHD requires a longstanding history of symptoms in childhood, dating back at least to the age of 7 years, this assessment is central to adult ADHD symptomatology. The measured observations for overeating comprised binge eating, emotionally-driven eating, and eating prompted by external stimuli rather than hunger. All these aspects of overeating have been implicated in the risk profile for obesity (e.g. Delahanty et al., 2002, Lluch et al., 2000, Wansink et al., 2005).