سطح کورتیزول در واکنش به شروع مدرسه در کودکان در معرض خطر هراس اجتماعی

Background Research on depression has identified hyperactivity of the HPA axis as a potential contributory factor to the intergenerational transmission of affective symptoms. This has not yet been examined in the context of social phobia. The current study compared HPA axis activity in response to a universal social stressor (starting school) in children of 2 groups of women: one with social phobia and one with no history of anxiety (comparison group). To determine specificity of effects of maternal social phobia, a third group of children were also examined whose mothers had generalised anxiety disorder (GAD). Method Children provided salivary cortisol samples in the morning, afternoon and at bedtime across 3 time-blocks surrounding the school start: a month before starting school (baseline), the first week at school (stress response), and the end of the first school term (stress recovery). Child behavioural inhibition at 14 months was assessed to explore the influence of early temperament on later stress responses. Results All children displayed an elevation in morning and afternoon cortisol from baseline during the first week at school, which remained elevated until the end of the first term. Children in the social phobia group, however, also displayed an equivalent elevation in bedtime cortisol, which was not observed for comparison children or for children of mothers with GAD. Children in the social phobia group who were classified as ‘inhibited’ at 14 months displayed significantly higher afternoon cortisol levels overall. Summary A persistent stress response to school in the morning and afternoon is typical for all children, but children of mothers with social phobia also display atypical elevations in evening cortisol levels when at school – signalling longer-term disruption of the circadian rhythm in HPA axis activity. This is the first study to report HPA axis disruption in children at increased risk of developing social phobia. Future research should determine whether this represents a pathway for symptom development, taking early temperament into account.

Social phobia is a disabling, chronic, anxiety disorder, characterised by excessive and persistent fear of scrutiny and negative evaluation, and marked restrictions in social functioning. Social phobia is one of the most common psychiatric disorders, with lifetime and 12 month prevalence rates of around 13% and 6%, respectively (Furmark, 2002). Onset is typically during adolescence (e.g. Schneier et al., 1992, Verhulst et al., 1997 and Grant et al., 2005), although precursors are commonly seen earlier, and these can have a profound deleterious impact on child socio-emotional development (Beidel et al., 1999). As with other affective disorders, social phobia aggregates in families, and clinical studies demonstrate specificity of transmission (i.e. first degree relatives of individuals with social phobia are at increased risk of social phobia, but not other anxiety disorders) (e.g. Maier et al., 1993, Fyer et al., 1995, Stein et al., 1998 and Merikangas et al., 2003). Such specificity has been supported in both top-down and bottom-up studies. For example, Mancini et al. (1996) found that 23% of children of mothers with social phobia had the disorder; a four-fold increase over child base-rate. Similarly, elevated rates of life-time social phobia, but no other affective disorder, were reported in mothers of shy children, compared to mothers of children with another form of disturbance (e.g. fearfulness) or no disturbance at all (Cooper and Eke, 1999). A genetic contribution to symptom development has been identified, but is modest, with an estimated heritability of around 10% (Nelson et al., 2000). This implies that additional mechanisms are important in the transmission of symptoms, and it appears that both child temperament and parenting style also play important aetiological roles (see review by Murray et al., 2009). Recent research in the context of depression suggests that the involvement of physiological systems, particularly those implicated in the response to stress, such as the hypothalamic pituitary adrenal (HPA) axis, should also be considered in the transmission of affective disorders. The HPA axis is an essential part of the organism's homeostatic system, and one of the most reactive of all neuroendocrine systems. The mechanism through which the axis operates is characterised by the activation of a cascade of biochemical events, which begins with the release of corticotropin-releasing hormone (CRH) from the hypothalamus, and culminates with the release of glucocorticoids (primarily cortisol) from the adrenal gland (Antoni, 1986 and Plotsky, 1991). HPA axis activity is one of the many bodily functions to display a distinct daily or diurnal rhythm. Cortisol levels are at their highest at awakening and then decline over the day, reaching near zero in the evening and remaining low until the later hours of sleep, when levels begin to increase (Clow et al., 2004). Imposed on this rhythm is a response to awakening, consisting of a 50–160% increase in the first 30–40 min after eyes open, the ‘cortisol awakening response’ (CAR) (Fries et al., 2009); and this, in turn, fluctuates in anticipation of the stressors or challenges of the day. While the release of glucocorticoids is critical for enabling adaptation to perturbations in homeostasis caused by both real and anticipated increases in demand, i.e., stress ( Selye, 1973), chronically, or frequently elevated cortisol levels (e.g. raised baseline levels or poorly regulated stress responses) are associated with adverse effects on health, including the development of mental health problems, particularly affective disorders ( Goodyer et al., 2001, Goodyer et al., 2003, Goodyer et al., 2009, Goodyer et al., 2010 and Harris et al., 2000). A proportion of depressed adults are reliably reported to display hyperactivity of the HPA axis, as demonstrated by elevated baseline cortisol levels (particularly in the evening), hypertrophy of the adrenal gland, raised cerebral-spinal levels of CRH, and reduced negative feedback in response to administration of synthetic glucocorticoids, such as dexamethasone ( Arato et al., 1989, Heuser et al., 1994, Rubin et al., 1995 and Carroll et al., 2007). Furthermore, a recent meta-analysis confirms that increased HPA axis activity is also a feature of depression in both children and adolescents ( Lopez-Duran et al., 2009). And this finding extends to include children at risk for depression as a result of their mothers having a history of the disorder ( Halligan et al., 2004 and Dougherty et al., 2009). With regard to anxiety, elevated basal cortisol levels have been consistently reported in individuals with panic disorder (e.g. Wedekind et al., 2000), and, to a lesser extent, in individuals with OCD (Gustafsson et al., 2008) and GAD (e.g. Mantella et al., 2008). Furthermore, research suggests that comorbid anxiety may explain some of the HPA axis hyperactivity effects commonly observed in depression (Young et al., 2004). Research in the context of social phobia is more limited and findings have not always been consistent. However, some studies have reported elevated cortisol levels in response to stress in individuals with social phobia compared to controls (Furlan et al., 2001, Condren et al., 2002 and Roelofs et al., 2009). For example, Roelofs et al. (2009) found that individuals with social phobia displayed significantly elevated stress-induced cortisol levels (i.e. post-stressor samples) following the Trier Social Stress Test than those with PTSD and a healthy comparison group. Moreover, stress-induced cortisol levels in the social phobia group were found to be positively correlated with an objective measure of social avoidance behaviour to angry faces manifested during an affect evaluation task, providing the first evidence of a link between HPA axis hyperactivity and levels of social avoidance behaviour in the context of a social phobia diagnosis. There is more data on HPA axis functioning in relation to child temperamental traits thought to be associated with the development of social phobia, i.e., behavioural inhibition, or BI (Kagan, 1994). Inhibited/socially wary infants have been shown to display the highest cortisol levels under both stressed (e.g. Kagan et al., 1987) and non-stressed (e.g. Schmidt et al., 1997) conditions. Studies involving children undergoing normative stressors, such as entering novel peer-group settings (e.g. starting school), have been particularly fruitful in demonstrating these cortisol-temperament relationships (e.g. Gunnar et al., 1997, Tout et al., 1998, Bruce et al., 2002, Watamura et al., 2003 and Turner-Cobb et al., 2008). While such social events are generally considered positive for children, in terms of stimulating the development of social capacities, the social and emotional resources required for their navigation are considerable, and may cause significant stress. It appears that an initial stress response (i.e. rise in cortisol levels) in the morning and afternoon, but not in the evening, is normative for children when entering these settings, and tends to reduce as the environment becomes familiar (e.g. Turner-Cobb et al., 2008, Watamura et al., 2009 and Sumner et al., 2010). However, a number of studies have shown that children with internalising temperamental traits (e.g. teacher-rated shyness) are particularly prone to mounting persistent stress responses in these settings, i.e. displaying continued elevations in cortisol, even when the surroundings are no longer unfamiliar (e.g. Gunnar et al., 1997 and Tarullo et al., in press). Others report that elevations in evening cortisol levels, in particular, are observed for shy children on school days as compared to weekend days; more extroverted children, by contrast, may show elevations in morning cortisol levels that drop to the equivalent of baseline levels by the end of the day ( Bruce et al., 2002 and Turner-Cobb et al., 2008). Interestingly, where apparently conflicting results have been observed, i.e. with the highest cortisol levels being reported for the most extroverted children, these have been explained by the type of interactions these children display (i.e. with higher levels of bold, risk-taking, problematic behaviour), and if these social interactions are taken into account, the relationship between shyness and high cortisol is upheld ( Gunnar et al., 2003). Two studies have reported a predictive association between elevated cortisol levels and the development of social anxiety. Smider et al. (2002) measured children's average afternoon cortisol levels at 4.5 years, and found higher levels to be predictive of teacher-reported withdrawal (including social wariness) a year and a half later. The same relationship was found using maternal ratings of withdrawal, but for girls only. Importantly, these predictive effects held even when controlling for behaviour assessed at the time the cortisol samples were provided (Smider et al., 2002). In the second study, greater increases in cortisol levels across an 18-month period predicted greater general and social anxiety in 9–14-year-old girls (Schiefelbein and Susman, 2006). If HPA axis hyperactivity can contribute to the development of affective psychopathology, it could well be important in the intergenerational transmission of affective disorders (including social phobia) from mother to child. To substantiate this notion, however, systematic evidence supporting an association between maternal psychopathology and child HPA axis hyperactivity is required. Ashman et al. (2002) found that 7-year-old children of depressed mothers, who demonstrated internalising symptoms themselves, displayed higher basal cortisol levels than children of non-depressed controls; importantly, maternal depression in the child's first 2 years was the most significant predictor of these elevations. Similarly, in a study investigating the effects of maternal stress on children, maternal depression in infancy was identified as the strongest predictor of future increased cortisol levels in preschoolers also exposed to concurrent maternal depression (Essex et al., 2002). In another study, 13-year olds exposed to postnatal depression had higher and more variable 8 AM cortisol levels than non-exposed adolescents, and this relationship held when controlling for concurrent child and maternal symptoms, as well as other potential confounds (Halligan et al., 2004). Notably, when directly observed mother–child interactions were examined, maternal withdrawal in infancy (but not at five years) predicted the 13-year morning cortisol elevations (Murray et al., 2010). In turn, these higher morning cortisol levels mediated the effects of maternal postnatal depression on offspring depressive symptoms at age 16, even when controlling for depression at 13 years (Halligan et al., 2007), supporting the idea that HPA axis hyperactivity constitutes a potential process through which the transmission of familial risk for depression occurs. Comparable intergenerational studies in the context of maternal anxiety disorders are less numerous, and, indeed, absent in the case of social phobia. Nonetheless, one study (Brennan et al., 2008) did take into account maternal co-morbid anxiety disorders while examining the link between child HPA axis activity and maternal depression, and found that, whereas maternal depression was associated with increased offspring baseline mean cortisol levels, depression with co-morbid anxiety was more strongly associated with increased child cortisol reactivity to stress. This suggests the effects of the two disorders on offspring HPA axis activity may not be equivalent, with maternal anxiety showing stronger associations with child elevated responses to stress, rather than the baseline effects commonly observed in depression. This is consistent with the finding that 14–15 month old children of mothers with panic disorder demonstrated elevated cortisol levels in a stressful laboratory setting, but not at home ( Warren et al., 2003). Overall, the evidence supports an association between HPA axis hyperactivity and the development of affective disorders (including social anxiety), and suggests that maternal affective disorder has a putative causal association with offspring HPA axis hyperactivity. This indicates that there may be intergenerational transmission of social phobia mediated in part by dysregulated cortisol activity. To our knowledge, no studies have yet attempted to characterise HPA axis activity in a group of off-spring at increased risk for social phobia by virtue of a maternal diagnosis of this disorder. The aim of the current study was to characterise HPA axis activity in response to the social stressor of starting school in 4–5-year-old children of mothers with social phobia and a similarly aged comparison group of children of mothers with no history of anxiety (this comparison group was therefore considered to be at low risk of developing social phobia). We tested the hypothesis that offspring of social phobic mothers would demonstrate cortisol hyperactivity in response to the stressor compared to offspring of mothers with no lifetime history of anxiety disorders. This was done in the context of a wider ongoing longitudinal study on the intergenerational transmission of social phobia, as part of which child BI had been assessed at 14 months (Murray et al., 2007 and Murray et al., 2008). In the current study, the mothers and their children, then aged 4–5 years, were reassessed (i) one month before starting school (BSS) (ii) across the first week at school (FWS), and (iii) at the end of the first school term (EoT), with salivary cortisol levels sampled over several days, across the course of the day, for each of these three time blocks. Starting school is an ideal setting for assessing individual differences in HPA axis activity in response to social challenge, being a normative, yet socially demanding, event. Based on previous findings (e.g. Gunnar et al., 1997), we expected all children to show some initial cortisol response, i.e. an elevation in cortisol levels, but that this would be more pronounced and persistent (i.e. extending to the end of the first term) for those in the social phobia group, thus indicating poorer habituation to the social stressor. Furthermore, we expected, on the basis of previous findings (Bruce et al., 2002), that this would apply particularly to evening/bedtime cortisol levels. Following on from this, when investigating these effects, it was considered important to take child sex and behavioural inhibition into account, given that both have been shown to affect cortisol ( Kagan et al., 1987 and Rosmalen et al., 2005). Finally, given that an important recent issue to emerge in the anxiety disorder literature is the notion of specificity of transmission and the idea that different processes might be involved in the intergenerational transmission of different anxiety disorders (Murray et al., 2009), as a final step we included in our study a further anxious group, namely, children whose mothers had generalised anxiety disorder (GAD). This enabled us to determine whether any differences observed between children of mothers with social phobia and children of comparison mothers were specific to a maternal diagnosis of social phobia, rather than a feature of maternal anxiety more generally. Given the predominantly social nature of the challenge involved in starting school, and that previous research in relation to GAD highlights elevated basal cortisol rather than elevated stress-responses ( Mantella et al., 2008), we expected our indices of poorer HPA axis habituation to occur more in the social phobia than in the GAD group offspring.