Children’s Aggression and DSM-III-R Symptoms Predicted by Parent Psychopathology, Parenting Practices, Cortisol, and Ses

Abstract

A handful of studies over the past 10 years have examined whether cortisol (a hormone which in humans is associated with stress and catabolic processes) is related to behavioral measures of aggression or conduct problems. In reviewing this small literature, McBurnett and Pfiffner (McBurnett & Pfiffner, in press) reached the tentative conclusion that low cortisol may be associated only with the most severe, chronic, and aggressive subtype of disruptive behavior. The basis for this hypothesis was that three studies found no significant association between cortisol and diagnostic status when the experimental groups were based on a diagnosis of a disruptive behavior disorder (Kruesi, Schmidt, Donnelly, Euthymia & others, 1989) or diagnosis of Conduct Disorder (CD), (Scerbo & Kolko, 1994) (Targum, Clarkson, Magac-Harris, Marshall & Skwerer, 1990), and that significant findings only emerged from designs in which diagnostic groups were more finely subtyped. In a sample of adults, cortisol was abnormally low only in prisoners who had both Antisocial Personality Disorder and a history of habitual violence (Virkkunen, 1985). When that sample was reanalyzed using childhood characteristics as a grouping variable, only prisoners with a developmental history of Undersocialized Aggressive CD were found to have abnormally low cortisol. This suggested that low cortisol was to be found only among groups who were chronically violent and psychopathic. In a clinic sample of boys, low cortisol was not found to be characteristic of the diagnosis of CD, but cortisol was found to be significantly low in boys who had CD and no anxiety disorder when compared to boys with CD who also had an anxiety disorder (McBurnett et al., 1991). Analyses of follow-up data on these same subjects suggested that cortisol was low in boys with DSM-IV Childhood Onset CD when compared to those with Adolescent Onset CD (McBurnett & Pfiffner, in press) and that low cortisol was associated only with overt, aggressive CD symptoms and not with covert CD symptoms (McBurnett, Lahey, Capasso & Loeber, 1996). These findings are consistent with the interpretation that low cortisol is characteristic only of those boys with CD who have a chronic, aggressive, low anxiety (psychopathic?) form of the disorder, similar to the interpretation of the Virkkunen (1985) study. We were interested in finding out whether this biological variable (low cortisol) was associated with chronic CD independently of known psychosocial correlates of CD, or whether the biological and the psychosocial correlates accounted for essentially the same variance in the development of chronic CD.