The findings on basal cortisol secretion demonstrate that the magnitude of the CAR was similar across groups. In contrast, both CD subgroups showed increased mean cortisol levels compared with control subjects, largely because of higher evening levels. All groups showed significant declines in cortisol levels across the day, and cortisol slope values did not differ significantly between groups. Adolescents with EO-CD exhibited reduced cortisol and cardiovascular responses to psychosocial stress relative to control subjects. Furthermore, the mean cortisol response to stress was smaller in EO-CD participants compared with control subjects, even if only those participants who exhibited a measurable cortisol response to stress were considered. Participants with AO-CD displayed a strikingly similar pattern of cortisol and cardiovascular hyporeactivity to that observed in those with EO-CD. There was also a trend toward lower cortisol reactivity in AO-CD responders (despite the loss of power incurred by considering only those with measurable responses). The size of the peak cortisol response observed was also less than half that seen in responders from the control group. As such, there was no evidence for differences in basal cortisol secretion and physiologic reactivity to stress between AO-CD and EO-CD participants.
Differences in emotional responses did not explain the cortisol and cardiovascular hyporeactivity findings because increases in self-reported negative affect during stress followed a similar course across groups. This discrepancy between subjective and physiologic changes suggests poorer coordination between emotional and physiologic arousal in both CD subtypes (10,16
Our basal cortisol findings, showing an intact CAR and cortisol diurnal rhythm but elevated 11 am
and evening cortisol levels in both forms of CD, are consistent with previous results in population-based samples (11,39
). They are, however, at variance with studies showing lower cortisol levels in clinic-referred CD samples (2,6
). The latter studies obtained single-point cortisol measurements or may have been confounded with acute stress due to venipuncture. An earlier study reported that prepubertal boys with disruptive behavior disorders had similar CAR magnitudes compared with control subjects, using the AUCI
value for cortisol (40
). However, in contrast to our results, that study observed reduced absolute cortisol secretion over the first hour after waking in CD/ODD participants relative to control subjects.
Our findings of cortisol and cardiovascular hyporeactivity during stress in CD are consistent with previous work in clinic-referred children with ODD (3,10
) and delinquent prepubertal adolescents (16
). They are also in agreement with results in high-risk children of fathers with substance dependence (41,42
). These data indicate that, unlike reduced basal cortisol, cortisol and cardiovascular hyporeactivity are associated with CD in both clinic-referred and population-based samples. This has clinical implications because cortisol hyporeactivity during stress is associated with poor treatment outcome (18
), and some treatments for CD encourage patients to become aware of their physiologic reactions as triggers for anger states. Remaining questions relate to the physiologic origin of the cortisol hyporeactivity observed in CD and its implications for psychological functioning. Future studies should examine whether cortisol hyporeactivity precedes CD onset using longitudinal designs.
Our findings of blunted cortisol and cardiovascular responses to stress in AO-CD and EO-CD participants relative to control subjects contradicts the developmental taxonomic theory, which implies that such neurobiological differences should be unique to EO-CD. Physiologic hyporeactivity during stress could reflect a latent trait that increases vulnerability to CD, whereas age of CD onset may be moderated by psychosocial factors (e.g., differences in parental supervision, exposure to antisocial models). Alternatively, it may be unnecessary to invoke a latent trait in either subgroup: rather, both CD subgroups may have experienced increased social adversity during development (e.g., maltreatment), or, because of heightened risk-taking behaviors, they may place themselves in stressful situations more frequently than other adolescents (leading to habituation to stressors).
Two limitations are noted. First, aside from spit diaries, no measures were in place to ensure compliance with the saliva collection protocol. This is problematic because poor adherence may be expected in adolescents with CD and may have resulted in elevated evening cortisol in the CD subgroups. The robust cortisol response to awakening and diurnal rhythm observed in all three groups nevertheless suggests that, in most cases, participants followed the collection protocol with acceptable accuracy. Furthermore, increased basal cortisol levels were observed in both CD subgroups at 11 am when under experimental supervision.
Second, although we obtained information from multiple informants and enquired about the age of onset of each CD and ODD symptom, the study relied on retrospective accounts. As such, the findings require replication in a longitudinal design, including follow-up to the point of remission.
In summary, this study demonstrated that cortisol and cardiovascular responses to an ecologically valid psychosocial stressor were reduced in adolescents with both early-onset and adolescence-onset CD. These findings were not explained by differences in subjective responses to the stressor, suggesting a discrepancy between mood changes and physiologic reactivity in CD. The basal cortisol data showed a normal diurnal rhythm and CAR in participants with CD, suggesting intact HPA axis function. Finally, contrary to predictions, we observed elevated 11 am and evening cortisol in both CD subgroups.