Developmental traumatology part I: biological stress systems∗

Abstract

Background: This investigation examined the relationship between trauma, psychiatric symptoms and urinary free cortisol (UFC) and catecholamine (epinephrine [EPI], norepinephrine [NE], dopamine [DA]) excretion in prepubertal children with posttraumatic stress disorder (PTSD) secondary to past child maltreatment experiences (n = 18), compared to non-traumatized children with overanxious disorder (OAD) (n = 10) and healthy controls (n = 24).

Methods: Subjects underwent comprehensive psychiatric and clinical assessments and 24 hour urine collection for measurements of UFC and urinary catecholamine excretion. Biological and clinical measures were compared using analyses of variance.

Results: Maltreated subjects with PTSD excreted significantly greater concentrations of urinary DA and NE over 24 hours than OAD and control subjects and greater concentrations of 24 hour UFC than control subjects. Post hoc analysis revealed that maltreated subjects with PTSD excreted significantly greater concentrations of urinary EPI than OAD subjects. Childhood PTSD was associated with greater co-morbid psychopathology including depressive and dissociative symptoms, lower global assessment of functioning, and increased incidents of lifetime suicidal ideation and attempts. Urinary catecholamine and UFC concentrations showed positive correlations with duration of the PTSD trauma and severity of PTSD symptoms.

Conclusions: These data suggest that maltreatment experiences are associated with alterations of biological stress systems in maltreated children with PTSD. An improved psychobiological understanding of trauma in childhood may eventually lead to better treatments of childhood PTSD.

Introduction

The diagnosis of posttraumatic stress disorder (PTSD) is made after a person experiences one or more overwhelming traumatic event(s) and reacts with fear or disorganized behavior; followed by complaints of three clusters of categorical symptoms for at least one month: 1) intrusive re-experiencing of the trauma(s), 2) persistent avoidance of stimuli associated with the trauma, and 3) persistent symptoms of increased physiological arousal (Frances 1994). The clinical picture of PTSD in children is similar to that of adults (for review see Pynoos and Eth [1985] or De Bellis [1997]) with the exception of children less than age 4 years where more objective criteria based on observable behaviors are warranted (Scheeringa et al 1995). Children seem to be less resilient to trauma than adults. Results from a recent meta-analysis suggest that children and adolescents are about 1.5 times more likely to be diagnosed with PTSD, once traumatized, than their adult counterparts (Fletcher 1996).

Maltreatment of children is defined as neglect, physical abuse, sexual abuse, and emotional maltreatment (that includes verbal threats to the child and witnessing domestic violence). It is a serious public health problem, and may be one of the most common causes of interpersonal traumas and of PTSD in children and adolescents (De Bellis 1997). In 1996, the incidence rate of alleged maltreatment cases were reported to be 44 out of every 1000 children (U.S. Department of Health and Human Services 1998). Maltreatment in childhood is both a cause and a risk factor for PTSD (De Bellis and Putnam 1994). PTSD occurs in 42%–90% of individuals exposed to sexual abuse (McLeer et al 1994), 50%–100% among those witnessing domestic violence (Pynoos and Nader 1989), and 11%–50% Pelcovitz et al 1994, Green 1985 among physically abused children. Famularo et al (1988) showed that 62 out of 156 (39.7%) children, who were removed from their parent’s physical and legal custody secondary to maltreatment, had PTSD within the initial disclosure period. Of those children who disclosed sexual abuse, 63% had a diagnosis of PTSD. Seventeen of the 52 (32.7%) re-examined from the original sample of 62 continued to meet PTSD criteria, although 67.3% no longer met full PTSD criteria at 2-year follow-up (Famularo et al 1996). Subjects with chronic PTSD also had much co-morbidity. Thus, the chronic stress of maltreatment experiences in childhood, especially sexual abuse, is a risk factor for acute and chronic PTSD as well as other negative developmental consequences De Bellis 1997, National Research Council 1993.

Several neurotransmitter and neuroendocrine systems are activated during acute stress (reviewed by Charney et al 1993). Traumatic stress may have negative effects on the development of these systems (De Bellis and Putnam 1994). There is little research on the neurobiological effects of trauma and PTSD in developing children. Studies of the neurobiological effects of overwhelming stress in animal models and of the psychobiology of adult PTSD provide our only comparative models. To date, most investigators have focused on two of the body’s major stress systems, the catecholamine system (the locus ceruleus-norepinephrine [NE]/sympathetic nervous system [SNS]) and the hypothalamic-pituitary-adrenal (HPA) axis.

Animal studies show that traumatic stress activates the locus ceruleus, the major catecholamine (specifically NE) containing nucleus in the brain (Simson and Weiss 1988) and the SNS leading to the biologic changes of the “fight-or-flight reaction” (Aston-Jones et al 1991). Direct and indirect effects of this activation include increases in catecholamine turnover in the brain, the SNS, and adrenal medulla leading to increases in heart rate, blood pressure, metabolic rate, alertness, and in the circulating catecholamines (epinephrine [EPI], NE, and dopamine [DA]) (for review see De Bellis and Putnam 1994). During stress, the brain’s hypothalamic corticotropin-releasing hormone (CRH) is released. CRH activates the HPA axis by stimulating the pituitary to secrete adrenocorticotropin (ACTH). These events, in turn, promote cortisol release from the adrenal gland, stimulate the SNS, and centrally cause behavioral activation and intense arousal (Chrousos and Gold 1992). The locus ceruleus also indirectly stimulates the HPA axis via connections through the limbic system (for review see De Bellis and Putnam 1994). Cortisol, via negative feedback inhibition on the hypothalamus, pituitary, and other brain structures (hippocampus), suppresses the HPA axis leading to restoration of basal cortisol levels (homeostasis). In animal models, activation of the catecholamine system and CRH results in behaviors consistent with anxiety and hypervigilance, the core symptoms of PTSD in humans.

In adult PTSD, it is hypothesized that the catecholamine system and HPA axis responses to stress become maladaptive, causing long-term negative consequences (reviewed by Charney et al 1993). Results from adult combat-related PTSD studies suggest increased sensitivity of the catecholamine system evident under experimental conditions of stress or challenge (for review see Southwick et al 1998). These findings include increased heart rate, systolic blood pressure, skin conductance, and other SNS responses to adrenergic or traumatic reminder challenge (reviewed by Charney et al 1993, Pittman 1993, Southwick et al 1998), and decreased sleep latency and efficiency (Ross et al 1989) in adult PTSD compared to healthy combat or non-combat controls. Although most baseline studies of single or multiple time point plasma catecholamines found no significant differences between adult PTSD and controls (Southwick et al 1995), elevated 24-hour urinary catecholamine excretions were seen in 3 of 5 studies (for review see Southwick et al 1995). Single time point measures of catecholamines and cortisol may not provide an accurate measure of baseline functioning because of circadian influences. Further, the stress of a single-stick venipuncture may result in elevations of cortisol and catecholamine concentration, obscuring any baseline differences. Thus, in adult PTSD, elevated 24-hour urinary excretion of catecholamines provides evidence of an increase in baseline functioning of the catecholamine system.

Unlike the increased sensitivity of the catecholamine system to stress seen in adult PTSD, baseline and challenge studies show that the HPA axis functions in a more complicated manner (Southwick et al 1998). In adult combat-related PTSD, elevated levels of central CRH were found (Bremner et al 1997). Infusion studies of metyrapone, that blocks the conversion of 11-deoxycortisol to cortisol and allows for the direct measure of pituitary release of ACTH, suggested that there is down-regulation of anterior pituitary CRH receptors presumably secondary to elevated central CRH and enhanced negative feedback inhibition of the pituitary for cortisol (Yehuda et al 1996). Further evidence for enhanced negative feedback inhibition includes findings of increased number of glucocorticoid receptors on lymphocytes presumably secondary to decreased circulating cortisol, suppression of cortisol with low dose dexamethasone, and lower 24-hour urinary free cortisol (UFC) concentrations in 3 of 4 studies of adult combat related PTSD compared with controls Mason et al 1986, Yehuda et al 1991, Yehuda et al 1992. Low urinary cortisol was also found in one study of male and female adults with PTSD who survived the Holocaust as children and adolescents compared to survivors without PTSD (Yehuda et al 1995). In two other studies, 24-hour UFC concentrations were higher in male combat veterans with PTSD compared to combat veterans without PTSD (Pittman and Orr 1990) and in women with PTSD secondary to childhood sexual abuse compared to women abused as children without PTSD and healthy non-abused control women (Lemieux and Coe 1995). These discrepant findings may be related to the confounding effects of assay methodology, differences in body weight between groups, current life stressors (Kaufman et al 1997b), an independent contribution of other stress systems on the HPA axis (Chrousos and Gold 1992), or as a reflection of chronic HPA adaptation axis long after trauma exposure.

Because there are few psychobiological studies on childhood PTSD, we will review this area broadly, also focusing on studies of traumatized children with anxiety or depressive disorders. Findings of elevated baseline 24-hour urinary concentrations of catecholamines from the few investigations in traumatized children to date suggest that the psychobiology of child maltreatment may be similar to that of combat-related adult PTSD (De Bellis and Putnam 1994) because there is evidence of an increase in baseline functioning of the catecholamine system in these children. For example, in a pilot study, we showed that sexually abused girls, 58% of whom had histories of severely depressed mood with suicidal behavior (but only one of whom had PTSD), exhibited significantly greater 24-hour urinary concentrations of catecholamines and their metabolites compared with demographically matched non-abused controls (De Bellis et al 1994b). Noradrenergic function as measured by 24-hour urinary catecholamine excretion has been found to be high in male, but not female, children with severe clinical depression having a history of parental neglect (Queiroz et al 1991). Perry (1994) found decreased platelet adrenergic receptors and increased heart rate after orthostatic challenge in physically and sexually abused children with PTSD, suggesting an enhancement of SNS tone in childhood PTSD. Further support for an increase in baseline functioning of the catecholamine system in childhood PTSD is provided by two separate, open-label treatment trials of clonidine, a central alpha₂-adrenergic partial agonist, and propranolol, a beta-adrenergic antagonist, both of which dampen catecholamine transmission. Clonidine treatment was associated with general clinical improvement, and decreases in the arousal cluster of PTSD symptoms and in basal heart rate (Perry 1994), while propranolol treatment was associated with decreases in aggressive behaviors and insomnia (Famularo et al 1988).

The few investigations to date on the HPA axis and childhood trauma have led to somewhat discrepant results. When examined as a function of the subjects being studied after a duration of time elapsed since trauma exposure, results may fall into a predictable pattern of response; elevated central CRH and resultant hypersecretion of cortisol is seen initially, and enhanced negative feedback inhibition of the pituitary for cortisol leading to lower 24-hour UFC findings is seen as a long term, and possibly a developmental (post-pubertal) consequence, of trauma (De Bellis et al 1994a). For example, in two separate studies, maltreated young children with a diagnosis of major depression failed to show the expected diurnal decrease in cortisol secretion from morning to afternoon Hart et al 1996, Kaufman 1991. Maltreated prepubertal depressed children undergoing current psychosocial adversity have increased human CRH induced ACTH response, but normal cortisol secretion compared to depressed children with prior histories of maltreatment, depressed non-abused children, and healthy children (Kaufman et al 1997b). In an NIMH longitudinal study, augmented mean morning serial plasma cortisol levels were found in sexually abused girls recruited within six months of disclosure compared with non-abused sociodemographically matched controls, suggesting morning hypersecretion of cortisol secretion in the former (Putnam et al 1991). We reported attenuated plasma ACTH responses to ovine CRH in a subsample of these sexually abused girls studied several years after disclosure (De Bellis et al 1994a); the majority of these children had histories of severely depressed mood with suicidal behavior, but only one of whom had a diagnosis of PTSD. The abused girls exhibited reduced evening basal, ovine CRH-stimulated, and time integrated total plasma ACTH concentrations compared with matched control subjects. Plasma total and free cortisol responses to ovine CRH stimulation did not differ between the two groups. Twenty-four-hour UFC excretion were nonsignificantly greater in the abused group. These results show that sexually abused girls manifest a dysregulatory disorder of the HPA axis, associated with hyporesponsiveness of the pituitary to exogenous CRH and normal overall cortisol secretion to CRH challenge. Central CRH hypersecretion may have occurred in these sexually abused girls during the emotional and physical stress associated with the traumatic sexual assault(s). CRH hypersecretion may have led to an adaptive down regulation of CRH receptors in the anterior pituitary, similar to the mechanism suggested in adult PTSD (Bremner et al 1997). Another study also found blunted ACTH responses, but normal cortisol secretion to human CRH challenge, in children with major depression and a past history of abuse compared to healthy controls (Kaufman et al 1993). Armenian adolescents who lived close to the epicenter of the 1988 earthquake and experienced a significant direct threat to life had greater PTSD and co-morbid depressive symptoms, lower baseline mean salivary cortisol levels, and greater afternoon suppression of cortisol by dexamethasone, five years after exposure, compared to Armenian adolescents who lived 20 miles from the epicenter (Goenjian et al 1996). The results of this latter study are similar to the HPA findings in adult PTSD. Thus, these studies show that elevated secretion of ACTH or cortisol is seen initially, and enhanced negative feedback inhibition of the pituitary for cortisol is seen as a long term, and possibly a developmental consequence, of trauma.

The objective of this investigation was to non-invasively examine biological stress response systems in maltreated children with PTSD and to compare these measures to two non-traumatized groups, anxious children with DSM-III-R overanxious disorder (OAD) and healthy controls. We were not only interested in studying maltreated children with PTSD from past traumas, but also children who were not currently experiencing overwhelming stress, to ascertain if traumatic experiences were associated with long lasting biological stress system changes during development. We recruited non-traumatized OAD children to disentangle PTSD from other non-trauma related anxiety disorders. Because this investigation is cross sectional, it is difficult to separate out the effects of heterogeneous sources of trauma and other confounding factors, e.g., poverty, substance abuse, low educational levels, poor parenting skills, and legal and social service entanglements (De Bellis and Putnam 1994). In the emerging field of developmental traumatology, measures of trauma (type, age of onset, and duration of trauma) as well as other mediating factors such as social support and demographic measures are regarded as independent variables and behavioral, cognitive, emotional, and biological measures as dependent variables. Because PTSD in maltreated children is hypothesized to be associated with global deficits in behavioral, cognitive, and emotional functioning, and alterations of biological stress systems, a cross sectional study showing an association between abuse and the dependent variables is the first scientific step in evaluating these issues. Although cross sectional investigations do not establish cause-effect relationships, they are undertaken to generate likely hypotheses that can be tested in more expensive prospective longitudinal studies of child abuse. Baseline activity of the HPA axis and the catecholamine system can be measured by sampling concentrations of UFC and the catecholamines, NE, EPI, DA and their metabolites in body fluids. In humans, cortisol and essentially all catecholamines and their metabolites are excreted into urine and have a 24-hour diurnal rhythm Maas et al 1987, Chrousos and Gold 1992. Timed measures of 24-hour UFC and catecholamine concentrations reflect daily baseline functional activity. Therefore, we hypothesized that prepubertal children with PTSD will manifest increased 24-hour UFC and urinary catecholamines concentrations when compared to the non-traumatized groups. We also hypothesized that children with PTSD will have greater anxiety as well as co-morbid depressive and externalizing symptoms than OAD and control subjects and that clinical symptoms will positively correlate with 24-hour UFC and urinary catecholamine concentrations.