Cortisol levels in children with Attention-Deficit/Hyperactivity Disorder

Abstract

Regulation of the Hypothalamus–Pituitary–Adrenal axis (HPA-axis) and its end product cortisol differs among persons with certain psychiatric disorders when compared with controls. Some reports concern Attention-Deficit/Hyperactivity Disorder (ADHD) but findings are inconclusive. In this study we collected four saliva samples during a regular weekday in children, 6–17 years old, with ADHD (n = 201) and non-affected comparisons (n = 221). Saliva cortisol was measured with radioimmunoassay technique. Clinical data were collected for diagnostic information. Subtypes and severity of symptoms were determined using parental rating scales. Children with ADHD had lower saliva cortisol levels than comparisons at waking up Median = 9.1 versus 12.7 nmol/L (p < .001), 30 min later Median = 15.8 versus 20.1 nmol/L (p < .001) and before going to bed Median = 0.8 versus 1.0 nmol/L (p = .015). No difference was found for the afternoon sample. When the study group was split into three different age groups similar results were found only for children above 10 years of age. Subtype of ADHD or co-occurring symptoms did not affect the cortisol levels. Degree of severity of ADHD symptoms was not associated with cortisol levels in the study group, other than a weak negative correlation between the afternoon sample and hyperactivity symptoms. The low cortisol levels in children with ADHD may indicate a dysregulation of the HPA-axis, for instance a down-regulation or a phase delay of the diurnal curve. The low levels may be related to the under-arousal possibly underlying several of the core symptoms of ADHD.

Introduction

Regulation of the Hypothalamus–Pituitary–Adrenal axis (HPA-axis) and its end product cortisol differs among persons with certain psychiatric disorders when compared with controls (Tsigos and Chrousos, 2002). For instance, depression is associated with high levels of cortisol, possibly as an expression of impaired glucocorticoid-mediated feedback inhibition (Pariante and Lightman, 2008). A hyperactivation of the HPA axis has been reported not only for depression but also for obsessive–compulsive disorder (Gustafsson et al., 2008), panic disorder (Wedekind et al., 2000; Abelson et al., 2007) and anorexia nervosa (Lo Sauro et al., 2008). A decreased functioning of the HPA-axis has been associated with atypical depression (Tsigos and Chrousos, 2002), chronic fatigue syndrome (Roberts et al., 2004; Papadopoulos and Cleare, 2011; Tak et al., 2011), posttraumatic stress disorder (PTSD) (Yehuda, 1998; Heim et al., 2000) and fibromyalgia (Griep et al., 1998; Riva et al., 2010). Several reports concern Attention-Deficit/Hyperactivity Disorder (ADHD), a neurobehavioral developmental disorder with three subgroups characterized by predominantly symptoms of inattention (ADHD-I), hyperactivity/impulsiveness (ADHD-HI) or both types, “combined” (ADHD-C).

However, the results are not conclusive; both hyper- and hypo-functioning have been reported as well as no differences in comparisons with children without ADHD-symptoms (for a short review, see below). Hypothetically, an association between a down-regulated HPA axis and ADHD fits with theories that regard ADHD as a consequence of under-arousal (Fairchild, 2010). One such theory is the optimal stimulation theory (Zentall and Zentall, 1983) which suggests that the manifestations of ADHD can be seen as expressions of increased activity which aims at promoting arousal, e.g. through shifts in attention, talking, seeking stimulation, risk-taking or aggressive behavior (Zentall, 2005).

The secretion of cortisol follows a diurnal cycle, characterized by high levels on awakening, a further increase during the morning and a gradual decrease over the day until midnight (Tsigos and Chrousos, 2002). Abnormalities of the HPA axis in children with ADHD were originally suggested in 1993 by Kaneko et al. who found that a normal diurnal cortisol rhythm (defined as maximum levels in the morning and minimum levels in the evening) was less frequent in children with ADHD than in (adult) controls. They also reported that the children with ADHD less frequently than the controls responded with suppression when exposed to dexamethasone, a synthetic steroid with capacity of suppressing the secretion of ACTH (dexamethasone suppression test [DST]) (Slater et al., 1962). Since then studies of cortisol have focused either on a dysfunctional circadian rhythm or an impaired reactivity after DST or exposure to an experimental stressor. Further, studies have focused on cortisol levels in relation to severity of ADHD symptoms, different diagnostic subtypes or co-morbidity.

Several authors have reported an association between low cortisol levels and either hyperactivity (Kaneko et al., 1993; Blomqvist et al., 2007; Ma et al., 2011) or ADHD symptoms in general (Scerbo and Kolko, 1994; King et al., 1998). Others have reported low cortisol levels in children with co-morbid oppositional defiant disorder (ODD) (van Goozen et al., 1998; Kariyawasam et al., 2002; Freitag et al., 2009) or disruptive behavior (Hastings et al., 2009). In contrast several studies have failed to demonstrate any relation between cortisol levels and ADHD in general (Pesonen et al., 2011; Wang et al., 2011), for attention (Sondeijker et al., 2007), for hyperactivity (Hatzinger et al., 2007) and for ADHD with aggressive behavior (Schulz et al., 1997). Studies on stress reactivity display a similar ambiguity. A blunted response has been associated with ADHD (King et al., 1998; McCarthy et al., 2011), hyperactivity (Hong et al., 2003; van West et al., 2009), inattentive problems (Randazzo et al., 2008; Pesonen et al., 2011) and ODD co-morbidity (Snoek et al., 2004; Hastings et al., 2009). Contrarily, others have not found a decreased cortisol response in children with ADHD in general (Jansen et al., 1999; White and Mulligan, 2005; Christiansen et al., 2010) or with hyperactivity (Hatzinger et al., 2007).

In summary, findings reported on ADHD and cortisol levels are inconclusive. This may reflect methodological limitations such as small samples (only three studies with n > 100), varying – and not rarely vague – diagnostic routines, comparison groups with unknown “contamination” of individuals with ADHD and unsatisfactory control of sampling times. Furthermore, only a few studies have captured the morning cortisol levels by collecting samples when awakening and soon thereafter. Thus, in the present study we aimed at investigating the diurnal levels of saliva cortisol in school aged children (6–17 years of age) with ADHD and age matched healthy comparisons also taking severity, subgroup and co-occurring symptoms into consideration.