نقش سوء در دوران کودکی در واکنش به HPA محور در اختلال اضطراب اجتماعی: یک مطالعه مقدماتی

Abstract Background Studies on depression have found that childhood abuse (CA) is associated with a persistent sensitization of the hypothalamic–pituitary–adrenal (HPA)-axis to stress in adulthood. So far, it is unknown whether this HPA-axis sensitization is specific to depression, or whether this is a more general outcome associated with CA in patients with mood and anxiety disorders. The aim of this study was to investigate whether CA is associated with enhanced cortisol reactivity to psychosocial stress in Social Anxiety Disorder (SAD). Methods Salivary cortisol levels before, during, and after exposure to psychosocial stress (i.e., Trier Social Stress Task, TSST) in SAD patients with a history of childhood abuse (SAD + CA, n = 9) were compared to cortisol levels in SAD patients without a history of childhood abuse (SAD − CA, n = 9), patients with PTSD related to childhood abuse (n = 16), and healthy controls without a history of childhood abuse (n = 16). Results Analyses showed that the SAD + CA group had a strongly increased cortisol reactivity (mean peak: 17.5 ± 1.9 nmol/l) compared to SAD − CA (mean peak: 9.0 ± 1.1 nmol/l), PTSD (mean peak: 9.0 ± 1.1 nmol/l) and healthy controls (mean peak: 9.6 ± 1.4 nmol/l), whereas baseline cortisol levels did not differ. The enhanced increase in the SAD + CA group was not explained by stronger anxiety in response to the TSST. Conclusions Consistent with the findings in depression, these results show for the first time that childhood abuse is also associated with strongly increased cortisol reactivity in SAD. When replicated in a larger sample, these findings may have important implications for the treatment of SAD.

. Introduction Social Anxiety Disorder (SAD) is the most common anxiety disorder, which is characterized by persistent fear and avoidance of social situations (Stein and Stein, 2008 and Mannuzza et al., 1995). The causes and pathogenesis of SAD are not well elucidated. Besides inheritable traits, several studies reported that childhood abuse (CA), and particular emotional abuse, is an important risk factor for the development of SAD (Gibb et al., 2007 and Heim and Nemeroff, 2001). Research in animals found that chronic exposure to early adverse events, such as maternal deprivation, may have a lasting impact on the neurobiology of the stress response, particularly on the stress-regulating hypothalamic–pituitary–adrenal (HPA) axis (see Heim and Nemeroff, 2001, Kaufman et al., 2000 and Sanchez, 2006, for reviews). Epigenetic regulation of hippocampal glucocorticoid receptor expression may mediate the effects of early life experiences on adult behavior (Weaver et al., 2004, Weaver et al., 2007 and McGowan et al., 2009). Despite the importance for our understanding of how childhood abuse might lead to psychopathology, only a few studies investigated the impact of early adverse experiences on HPA reactivity (Heim et al., 2000, Bremner et al., 2003 and Elzinga et al., 2003). Consistent with the animal studies, these studies found that CA is associated with increased HPA-axis sensitivity to stress in women with MDD (Heim et al., 2000), and to a smaller extent in Post-traumatic Stress Disorder (PTSD) (Bremner et al., 2003 and Elzinga et al., 2003). In healthy subjects, in contrast, CA has been related to hypo-responsiveness of the HPA-axis to psychosocial stress ( Elzinga et al., 2008 and Carpenter et al., 2007). An important question that needs to be elucidated is whether HPA-axis sensitization is specific for MDD (and PTSD), or whether it is a more general outcome of CA in patients with mood or anxiety disorders, including patients with SAD. So far, a number of studies investigated cortisol reactivity to psychosocial stress in SAD, with mixed results (Condren et al., 2002, Furlan et al., 2001, Levin et al., 1993 and Martel et al., 1999). Recently, we reported increased cortisol responses to the Trier Social Stress Test (TSST) in SAD compared to PTSD and healthy controls (Roelofs et al., 2009). This was in line with one other study reporting increased cortisol responses to psychological stress in SAD (Condren et al., 2002), but other investigations failed to find group differences (Furlan et al., 2001, Levin et al., 1993 and Martel et al., 1999). Besides several methodological differences, CA might play an important role in these (inconsistent) findings, as none of these studies included information about the presence of CA. The aim of the present retrospective study is to investigate HPA reactivity to a psychosocial stressor (TSST) in patients with SAD who report a history of (emotional, physical, or sexual) CA vs patients with SAD who do not report experiences of abuse during childhood. To investigate the specificity of the relation between abuse and HPA reactivity in SAD, we also included patients with PTSD related to CA and healthy participants without a history of CA.

Results 3.1. Cortisol reactivity In the repeated measures ANOVA a main effect of Time was found, due to a significant increase in cortisol levels after the TSST (F(9,414) = 10.98, p < 0.0001, η2 = 0.19). Also a main effect of Group was found (F(3,46) = 4.03, p < 0.05, η2 = 0.21), which was specified by a significant Group × Time interaction (F(27,414) = 2.62, p < 0.05, η2 = 0.15) (see Fig. 1). Follow-up univariate ANOVA analyses at the separate time points showed that this was due to the fact that SAD + CA patients had higher cortisol levels compared to the other three groups after stress induction (at +25 min (F(3,49) = 3.51, p < 0.05); +40 min (F(3,49) = 5.35, p < 0.005); +50 min (F(3,49) = 5.21, p < 0.005), and +65 min (F(3,49) = 4.46, p < 0.01)), but not at baseline (from −65 to +0 min, all ps >0.2), nor immediately after the TSST (at +15 min: p = 0.17). At all time points from +25 to +60 min the SAD + CA group had significantly higher cortisol levels than the SAD − CA, PTSD and HC group (all ps <0.01), whereas the SAD − CA, PTSD and HC did not differ (all ps >0.5). When controlling for Gender, Medication and TSST-modifications as covariates, the main effect of Group (F(3,43) = 5.27, p < 0.005, η2 = 0.27) and the Group × Time interaction (F(27,387) = 2.37, p < 0.05, η2 = 0.14) remained significant. The TSST-modification was a significant covariate which had a main effect (F(1,43) = 7.38, p < 0.01, η2 = 0.15), indicating that participants with modified (i.e., milder) versions of the TSST had lower cortisol levels (see also Roelofs et al., 2009). Gender did not show a main effect, but interacted with time (F(9,387) = 3.36, p < 0.05, η2 = 0.07), due to larger cortisol increases after the TSST in men than in women. Use of medication was not a significant covariate (F(1,43) = 1.85, n.s.). Finally, to control for depression rates, the individual BDI-II total-scores were entered as an additional covariate into the model. This did not affect the findings either: main effect of Group (F(3,42) = 5.79, p < 0.005, η2 = 0.29) and Group × Time effect (F(27,378) = 2.31, p < 0.05, η2 = 0.14). Moreover, depression was not a significant covariate (F(1,42) = 1.89, n.s.), nor did depression interact with Time (F(9,387) = 0.43, n.s.). The same was true when controlling for social anxiety severity as measured with the SPAI (main effect of Group (F(3,42) = 3.77, p < 0.05, η2 = 0.21) and Group × Time effect (F(27,378) = 2.41, p < 0.05, η2 = 0.15)). Social anxiety was not a significant covariate (F(1,42) = 0.05, n.s.), nor did it interact with Time (F(9,387) = 1.72, n.s.).