- Email: [email protected]
Stronger endocrine responses after brief psychological stress in women at familial risk of breast cancer
Psychoneuroendocrinology 28 (2003) 584–593 www.elsevier.com/locate/psyneuen
Stronger endocrine responses after brief psychological stress in women at familial risk of breast cancer S.M. Gold a, S.G. Zakowski b, H.B. Valdimarsdottir a, D.H. Bovbjerg a,∗ a
b
Biobehavioral Medicine Program, Derald H. Ruttenberg Cancer Center, Mount Sinai School of Medicine, 1425 Madison Avenue, Box 1130, New York, NY 10029, USA Department of Psychology, Finch University of Health Sciences, The Chicago Medical School, 3333 Green Bay Road North, Chicago, IL 60064, USA Received 26 November 2001; received in revised form 21 May 2002; accepted 22 May 2002
Abstract Recent research has linked exposure to chronic stress to altered acute stress responses and suggests a sensitizing effect of chronic stress leading to a stronger endocrine and cardiovascular response to acute stressors. Substantial evidence indicates that familial breast cancer risk is a chronic life stressor with higher levels of self reported distress. In this study, we investigated whether the endocrine response to a brief psychological stressor was stronger for women at familial risk for breast cancer. Thirty-six women at normal risk of breast cancer (FR⫺ Stress Group) and 17 women at familial risk (FR+ Stress Group) underwent a brief psychological laboratory stress test (speech task and mental arithmetic) over a 15 min period. Thirty women at normal risk not subjected to the stressful task served as controls (FR⫺ Control Group). Plasma epinephrine, norepinephrine and cortisol were measured at baseline, directly after the stress test (15 min) and at 30 min and 45 min post baseline. Heart rate data confirmed the effectiveness of the stress regimen. While there were no significant baseline group differences in the endocrine parameters, the response curves for the familial risk group revealed stronger epinephrine and cortisol reactivity to the stress test, as confirmed by significant group by time interactions. Norepinephrine levels showed a similar pattern, but results did not reach significance. These findings are in line with previous research documenting the facilitating effects of chronic stressors on acute stress response in animals and humans and provide the first
∗
Corresponding author. Tel.: +1-212-659-5562; fax: +1-212-849-2566. E-mail address: [email protected] (D.H. Bovbjerg).
0306-4530/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0306-4530(02)00046-X
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
585
evidence in the literature of a heightened endocrine reactivity to acute psychological stress in women at familial risk of breast cancer. The heightened endocrine reactivity to the experimental tasks seen here suggests that these women may experience stronger responses to stressors in their daily lives. According to the recently proposed concept of allostatic load, repeated overly strong stress responses may cumulatively have negative health implications. 2003 Elsevier Science Ltd. All rights reserved. Keywords: Familial breast cancer risk; Allostatic load; Experimental stress; Cortisol; Catecholamines; Recovery
1. Introduction Animal models have compellingly demonstrated that prior exposure to a chronic stressor significantly elevates reactivity to a novel acute stressor in both the hypothalamo–pituitary–adrenal (HPA) axis (Bhatnagar and Dallman, 1998; Weinstock et al., 1998; Bhatnagar et al., 1995; Young et al., 1990; Ottenweller et al., 1989) and the sympathetic pathway (McCarty et al., 1988; Konorska et al., 1989; Weinstock et al., 1998). Research in humans has been scant but has yielded results consistent with a higher neuroendocrine response in some studies, though not all. For example, increased epinephrine peak levels after a brief psychological stressor among healthy volunteers with chronic life stress have been reported (Pike et al., 1997). Recent reports in older women found that having a history of chronic stress, such as caregiving, was associated with increased cortisol reactivity to a brief psychological stressor, although cardiovascular parameters, as well as catecholamine responses did not differ from controls (Cacioppo et al., 2000). One chronic stressor that has received little attention with respect to possible effects on acute neuroendocrine reactivity is being at risk of developing a life-threatening disease (e.g., cancer). Having a family history of breast cancer is the single most powerful predictor that a woman will develop the disease at some point in her lifetime. Consistent with that threat, numerous reports have found significantly elevated levels of psychological distress (e.g., intrusive thoughts) in healthy women with a first-degree relative diagnosed with breast cancer, identifying family history of breast cancer as a potent chronic stressor (see Bovbjerg and Valdimarsdottir, 2001 for review). For example, in an early study of 140 women with one or more firstdegree relatives with breast cancer, Lerman et al. (1993) found levels of intrusive thoughts that were comparable to those seen in clinical populations, with 30% of the sample reporting that intrusive thoughts interfered with their daily lives. More recently, Baider et al. (1999) found that 53% of 230 women with family histories of breast cancer experienced distress reaching clinical levels. Another recent study (Neise et al., 2001) reported that 67% of 129 women with family histories of breast cancer evidenced medium or intense psychological strain, with 9 out of 10 women describing negative effects that impaired their mental and physical well-being. We have recently reported that women at familial risk of developing breast cancer
586
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
show a stronger acute stress response as indicated by: self-reported distress, elevated heart rate, increased natural killer cell activity (NKCA) and NK cell numbers immediately post stressor compared to women at normal risk (Valdimarsdottir et al., 2002). In this paper, we report for the first time the cortisol and catecholamine response to a brief laboratory stressor (speech task and mental arithmetic) in women at familial risk of breast cancer. We further examined the recovery curve after stressor termination.
2. Methods 2.1. Subjects From a sample of 96 women participating in the study, 83 women had complete data in requisite cardiovascular and endocrine measures and were thus included in the analyses. Participants were eligible if they were between 25 and 50 years of age, premenopausal, had not taken birth control pills in the past two months, were not currently pregnant and at least six months after their last pregnancy, did not report a personal history of chronic disease (including any neurological or psychiatric disease) or neoplasm, were not currently taking any drug known to influence immune or endocrine function and did not have relatives currently in active cancer treatment. The mean age of participating women was 35.1 years, the majority were white (75%), well educated (93% college graduates), and employed outside the home (81%). Based on their self-reported family histories of breast cancer, participants were classified as being at increased risk of developing breast cancer (FR+, n=17) or normal risk (FR⫺, n=66) following the algorithm of Claus et al. (1994). The FR⫺ group was randomly assigned to either the stress (n=36) or control condition (n=30) while all FR+ subjects underwent the acute psychological stressor. The groups did not differ significantly in any of the background variables: age (FH⫺ control 33.6 years±9.4; FH⫺ stress 35.8 years±8.2; FH+ stress 35.9 years±7.4; ANOVA F(2, 77)=0.67; df=2; p=0.52), height (FH⫺ control 163.6 cm±7.5; FH⫺ stress 165.8 cm±5.9; FH+ stress 164.8 cm±7.3; ANOVA F(2, 76)=0.80; p=0.45), weight (FH⫺ control 59.2 kg±6.9; FH⫺ stress 64.9 kg±11.7; FH+ stress 61.4 kg±8.8; ANOVA F(2, 74)=2.7; p=0.08), body mass index (BMI) (FH⫺ control 22.3±2.9; FH⫺ stress 23.7±3.8; FH+ stress 22.6±3.4; ANOVA F(2, 72)=1.29; p=0.29), or days since last period (FH⫺ control 12.2 days±11.6; FH⫺ stress 16.4 days±21.2; FH+ stress 15.5 days±14.3; ANOVA F(2, 75)=0.51; p=0.60). Fisher’s Exact Test two-sided tests for categorical data with small cell sizes showed that there were no significant group differences in the distribution of women having a BMI greater than 30, i.e. WHO obesity class I (FH⫺ control one subject; FH⫺ stress three subjects; FH+ stress one subject; Fisher’s Exact Test two-sided significance p=0.65). Chi squared tests revealed no differences on demographic variables: employment (chi square=10.4; p=0.74), level of education (chi square=3.0; p=0.80), family income (chi square=12.2; p=0.28), marital status (chi square=8.6; p=0.20), or ethnicity (chi square=11.7; p=0.17) distributions also did not differ significantly. Thus, none of
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
587
these variables was statistically controlled for in the following analyses. As there was a trend for a group difference in body weight, all the analyses described below were also conducted with weight included as a covariate and an identical pattern of significant results was seen. 2.2. Procedure The exact procedure of the study has been described in detail elsewhere (see Valdimarsdottir et al., 2002). Briefly, potential participants were screened in a telephone interview and an introductory session was scheduled to sign consent forms and allow the women to adapt to the setting in which the experiment would take place. The experimental sessions were scheduled between 0800 h and 1000 h. The session was rescheduled if the women reported any symptoms indicative of a cold or flu in the previous three days. Subjects were asked to refrain from drinking more than one cup of coffee on the morning of the experiment and from consuming more than two alcoholic drinks the night before the session. They were further told not to take medications during the 24 hours before the experiment. Heart rate (HR) was assessed prior to, and throughout the test at 2 min intervals. Subjective distress levels were assessed before and after the experimental manipulation using visual analogue scales. Blood samples were collected at baseline and at several time points after the stress test (at 15 min, 30 min and 45 min post baseline) for cortisol, norepinephrine and epinephrine measurement. An iv-catheter was inserted in a vein of a non-dominant arm and was kept patent with saline during the experimental session. The participants rested for 20 min to recover from possible effects of the catheter insertion procedure per se. The stress test was a modified version of the Trier Social Stress Test (Kirschbaum et al., 1993), comprising a 5 min speech task in which participants were asked to imagine they had been caught for a traffic violation and had to defend themselves at the traffic court. They were allowed two minutes for preparation and three minutes to deliver the speech. Speeches were delivered in front of a video camera, and the women were told that their performance would later be rated by experts for content and style. This was followed by a 5 min mental arithmetic task with harassment, during which participants were asked to add numbers out loud at the pace set by an audiotape. The control task consisted of two 5 min sessions of a non-stressful reading assignment, in which the women were told they could read the material provided by the experimenter at their leisure and would not be tested later on. All participants were provided a modest reimbursement for their time and effort. 2.3. Endocrine assays Epinephrine and norepinephrine plasma concentrations were measured using commercial radioenzymatic assays (Amersham Pharmacia Biotech, Inc., Piscataway, NJ). Plasma cortisol levels were assayed with a commercial radioimmunoassay kit (Diagnostic Products Corporation, Los Angeles, CA, 90045).
588
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
2.4. Statistical analysis To examine endocrine responses induced by the stress test, delta values (change compared to baseline) were computed for all post stressor time points, thus reducing the impact of well known interindividual variability in basal levels. It should be noted that confirmatory analyses of raw data revealed an identical pattern of significant interactions as that seen with delta values. Endocrine acute responses and recovery curves throughout the 30 min post-stress period were examined using repeated measures ANOVA with group as the between-subjects factor (FR⫺ control vs. FR⫺ stress vs. FR+ stress) and changes from baseline as the within-subject factor. To reduce Type I error, Sidak-adjusted between groups post hoc tests were computed for each time point separately, to explore the source of significant interaction terms. Linear regression approaches were used to explore the predictive value of changes in heart rate and self-reported distress on endocrine responses. All analyses were conducted with statistical software (SPSS 10.0), with a p value of less than 0.05 considered significant. 2.5. Effectiveness of the stressor Consistent with our previous findings with a subsample (n=65) of the participants (Valdimarsdottir et al., 2002), results in the present study showed significant group differences in both self-reported distress responses (ANOVA F(2, 79)=63.0; p⬍0.001), and HR reactivity during the stress test (ANOVA F(2, 79)=23.89; p⬍0.001). Post-hoc tests showed significant HR differences between all three groups, with a significantly stronger stress response in FH+ subjects compared to FH⫺ subjects, while both stress groups significantly differed from unstressed controls. Exploratory analyses of relations among these dependent variables, indicated that within the two stress groups, cortisol acute response was significantly predicted by acute response in HR (standardized beta=0.41; p=0.01) while acute self-report response was not predictive (standardized beta=0.11; p=0.41). The same was seen in acute epinephrine response (HR standardized beta=0.32; p=0.03; self-report standardized beta=⫺0.03; p=0.83).
3. Results No significant baseline differences between groups were seen in plasma cortisol (FH⫺ control 15.4 µg/dl±11.4; FH⫺ stress 13.1 µg/dl±4.7; FH+ stress 11.6 µg/dl±4.9; ANOVA F(2, 80)=1.3; p=0.27), epinephrine (FH⫺ control 112.6 pg/ml±109.8; FH⫺ stress 82.7 pg/ml±101.1; FH+ stress 66.8 pg/ml±64.4; ANOVA F(2, 80)=1.4; p=0.26), or norepinephrine (FH⫺ control 331.3 pg/ml±224.7; FH⫺ stress 356.2 pg/ml±283.3; FH+ stress 344.2 pg/ml±136.9; ANOVA F(2, 80)=0.09; p=0.92). A marked acute stress response in epinephrine was observed in both stressed groups, which began to resolve at 30 min post baseline in FR⫺ but not in FR+
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
589
women (see Fig. 1). Multivariate tests in repeated measures ANOVA showed a significant Group main effect (F(2, 80)=5.20; p=0.01), no main effect for Time (F(3, 78)=0.46, p=0.70), but a significant Group×Time interaction (F(6, 158)=2.64, p=0.02). At 15 min post baseline, epinephrine was significantly higher in both stress groups compared to the Control Group. At 30 min, only the FR+ Stress Group remained significantly higher than controls (see Fig. 1). Although the pattern of responses seen in Fig. 1 is suggestive of reductions in epinephrine levels in the control group, single group ANOVAs revealed significant Time effects only for the FR+ stress group (F(3, 14)=4.29; p=0.02), while there were no significant Time effects for either of the other two groups (FR⫺ control F(3, 27)=1.92; p=0.15; FR⫺ stress F(3, 33)=0.93; p=0.44). Although responses in plasma norepinephrine levels showed a similar pattern, no significant Group (F(2, 80)=0.01, p=0.99), Time (F(3, 78)=1.10, p=0.36), or interaction effects (F(6, 158)=1.28, p=0.27) were observed. Women at familial risk for breast cancer showed an increase in plasma cortisol levels after the stress test while women at normal risk in the stress group as well as the control condition declined (see Fig. 2). Multivariate tests showed significant Time (F(3, 78)=16.57, p⬍0.001) and Group (F(2, 80)=3.82, p=0.03) main effects as well as a significant Group×Time interaction (F(6, 158)=2.29, p=0.04) for cortisol. Cortisol levels of the FR+ Stress Group remained significantly elevated above mean
Fig. 1. Epinephrine response to an acute psychological stressor in healthy women at familial risk (stress FR+, n=17) and normal risk for breast cancer (stress FR⫺, n=36) compared to non-stressed women at normal risk (control FR⫺, n=30). Asterisks indicate significant post-hoc difference (Sidak adjustment) to unstressed control group. See text for absolute initial values.
590
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
Fig. 2. Cortisol response to an acute psychological stressor in healthy women at familial risk (stress FR+, n=17) and normal risk for breast cancer (stress FR⫺, n=36) compared to non-stressed women at normal risk (control FR⫺, n=30). Asterisk indicates significant post-hoc difference (Sidak adjustment) to unstressed control group. See text for absolute initial values.
values for the Control Group until 30 min post baseline while the FR⫺ Stress Group did not differ from control subjects at any time (see Fig. 2).
4. Discussion To date, little research has investigated the effects of chronic stressors on responsiveness of endocrine parameters to acute psychological stress in humans. In this paper we report an alteration in the endocrine response to an acute stressor in women at familial risk of developing breast cancer, compared to women at normal risk. Stronger responses were seen in neuroendocrine activity of the major stress systems, the hypothalamo–pituitary–adrenal axis (plasma cortisol) and the sympathetic pathway (plasma epinephrine). Consistent with the literature (e.g., Cacioppo et al., 1995), acute endocrine response (pre vs. post stressor) in cortisol and epinephrine was significantly predicted by change in HR but not in self-report. This further underlines the relevance of endocrine outcome measures above and beyond self-report stress assessments. As familial risk of breast cancer is a chronic stressor (see recent review by Bovbjerg and Valdimarsdottir, 2001), these findings are in line with reports from animal models showing the facilitating effect of chronic stressors on acute stress hormone responses to brief mild stressors. While numerous studies have shown that
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
591
animals exposed to chronic stressors exhibit enhanced HPA and sympathetic responses to novel, acute stressors (Bhatnagar and Dallman, 1998; Weinstock et al., 1998; Bhatnagar et al., 1995; Young et al., 1990; Ottenweller et al., 1989; McCarty et al., 1988; Konorska et al., 1989), the underlying mechanisms for this phenomenon are not yet known. One proposal is that prior stressors leave a central facilitation trace that upon exposure to a novel stressor, overcomes negative feedback effects typical for these systems (Bhatnagar and Dallman, 1998). Studies of acute endocrine stress reactivity in chronically stressed subjects are rare and findings have been somewhat inconsistent in humans, perhaps reflecting differences in the types of chronic stressors, the type of acute stress, the timing of endocrine assessment, or the subject sample used. In line with chronic stressor effects in spouses of Alzheimer’s patients (Cacioppo et al., 2000), we found a significantly stronger endocrine response to an acute stress test in healthy women at familial risk of breast cancer. It is tempting to speculate that unremitting stressors such as caring for a demented partner, or being at risk to develop a life-threatening disease may be more likely to show evidence of an impact on the acute reactivity of physiological stress systems than circumscribed stressors. Consistent with this view, no alterations in endocrine stress reactivity to acute challenge were found in unemployed men and women (Ockenfels et al., 1995), or in male high school teachers with high job strain (Benschop et al., 1994). Another explanation for the mixed findings in the literature may be that our sample and the cited study from Cacioppo and colleagues are the only ones comprised exclusively of female participants, since both animal experiments (Rhodes and Rubin, 1999) and human studies (Young, 1998) have suggested that females may show stronger acute endocrine stress responses during chronic stress. However, Pike et al. (1997) found an epinephrine stress response significantly altered by chronic stress in a study exclusively enrolling men. The fact that cortisol levels declined over time in stressed as well as unstressed FR⫺ women may be attributed to the typical circadian pattern during the morning hours in which this experiment was carried out. While women without familial risk of breast cancer did not evidence an increased cortisol response following the relatively mild psychological stress test, significant increases were seen in women at familial risk despite the underlying steep circadian decline. It is thus conceivable that a smaller response in FR⫺ women was masked by decreasing basal levels in cortisol. Given the kinetics of the endocrine stress response (see Sapolsky et al., 2000), it is also conceivable that earlier assessment of catecholamines (e.g. at 5 min into the stressor) would have revealed even more pronounced group differences. Future research should look earlier in the stress response and do so during the afternoon, when circadian decline is less steep. In summary, our results underline the profound effects that having a familial risk of breast cancer can have on acute endocrine responses to stressors in healthy women. The health consequences of such heightened neuroendocrine reactivity to acute stress are not yet known. It has recently been theorized that while acute stress reactions are essential for adaptation, repeated heightened physiological responses to acute
592
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
stressors can exact an allostatic load on an individual, resulting in negative health consequences (McEwen, 1998). From this perspective, the present results raise the possibility that women with a family history of breast cancer may have increased health risks as a result of increased reactivity.
Acknowledgements This investigation was supported in part by Grant M01RR00047 from the General Clinical Research Centers Program of the Division of Research Resources, National Institutes of Health. The study was also sponsored in part by grants from the Department of Defense (#DAMD J-4139 and J-4164) and the National Cancer Institute (#R01 CA72457). We are required to indicate that the content of the information contained in this report does not necessarily reflect the position or policy of the United States Government. We would also like to acknowledge the helpful comments of the anonymous reviewers.
References Baider, L., Ever-Hadani, P., Kaplan De-Nour, A., 1999. Psychological distress in healthy women with familial breast cancer: like mother, like daughter? Int. J. Psychiatry Med. 29, 411–420. Benschop, R.J., Brosschot, J.F., Godaert, G.L., De Smet, M.B., Geenen, R., Olff, M., Heijnen, C.J., Ballieux, R.E., 1994. Chronic stress affects immunologic but not cardiovascular responsiveness to acute psychological stress in humans. Am. J. Physiology 266, R75–R80. Bhatnagar, S., Mitchell, J.B., Betito, K., Boksa, P., Meaney, M.J., 1995. Effects of chronic intermittent cold stress on pituitary adrenocortical and sympathetic adrenomedullar function. Physiol. Behav 57 (4), 633–639. Bhatnagar, S., Dallman, M., 1998. Neuroanatomical basis for facilitation of hypothalamic–pituitary–adrenal responses to a novel stressor after chronic stress. Neuroscience 84, 1025–1039. Bovbjerg, D.H., Valdimarsdottir, H.B., 2001. Interventions for healthy individuals at familial risk for cancer. In: Baum, A., Anderson, B.L. (Eds.), Psychosocial interventions for cancer. American Psychological Association, Washington, DC, pp. 305–320. Cacioppo, J.T., Malarkey, W.B., Kiecolt-Glaser, J.K., Uchino, B.N., Sgoutas-Emch, S.A., Sheridan, J.F., Berntson, G.G., Glaser, R., 1995. Heterogeneity in neuroendocrine and immune responses to brief psychological stressors as a function of autonomic cardiac activation. Psychosom. Med 57 (2), 154–164. Cacioppo, J.T., Burleson, M.H., Poehlmann, K.M., Malarkey, W.B., Kiecolt-Glaser, J.K., Berntson, G.G., Uchino, B.N., Glaser, R., 2000. Autonomic and neuroendocrine responses to mild psychological stressors: effects of chronic stress on older women. Ann. Behav. Med 22, 140–148. Claus, E.B., Risch, N., Thompson, W.D., 1994. Autosomal dominant inheritance of early-onset breast cancer. Cancer 73, 643–651. Kirschbaum, C., Pirke, K.M., Hellhammer, D.H., 1993. The “Trier Social Stress Test” a tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology 28, 76–81. Konorska, M., Stewart, R.E., McCarty, R., 1989. Sensitization of sympathetic–adrenal medullary responses to a novel stressor in chronically stressed laboratory rats. Physiol. Behav 46, 129–135. Lerman, C., Daly, M., Sands, C., Balshem, A., Lustbader, E., Heggan, T., Goldstein, L., James, J., Engstrom, P., 1993. Mammography adherence and psychological distress among women at risk for breast cancer. J. Natl. Cancer Inst. 85, 1074–1080.
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
593
McCarty, R., Horwatt, K., Konarska, M., 1988. Chronic stress and sympathetic–adrenal medullary responsiveness. Soc. Sci. Med 26, 333–341. McEwen, B.S., 1998. Protective and damaging effects of stress mediators. N. Engl. J. Med. 338, 171–179. Neise, C., Rauchfuss, M., Paepke, S., Beier, K., Lichtenegger, W., 2001. Risk perception and psychological strain in women with a family history of breast cancer. Onkologie 24, 470–475. Ockenfels, M.C., Porter, L., Smyth, J., Kirschbaum, C., Hellhammer, D.H., Stone, A.A., 1995. Effects of chronic stress associated with unemployment on salivary cortisol: overall cortisol levels, diurnal rhythm, and acute stress reactivity. Psychosom. Med 57, 460–467. Ottenweller, J.E., Natleson, B.H., Pitman, D.L., Drastal, S.D., 1989. Adrenocortical and behavioral responses to repeated stressors: towards an animal model of chronic stress and stress-related mental illness. Biol. Psychiatry 26, 829–841. Pike, J.L., Smith, T.L., Hauger, R.L., Nicassio, P.M., Patterson, T.L., McClintick, J., Costlow, C., Irwin, M.R., 1997. Chronic life stress alters sympathetic, neuroendocrine, and immune responsivity to an acute psychological stressor in humans. Psychosom. Med 59, 447–457. Rhodes, M.E., Rubin, R.T., 1999. Functional sex differences (“sexual diergism”) of central nervous system cholinergic systems, vasopressin, and hypothalamic–pituitary–adrenal axis activity in mammals: a selective review. Brain Res. Brain Res. Rev 30 (2), 135–152. Sapolsky, R.M., Romero, L.M., Munck, A.U., 2000. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocrine Rev 21 (1), 55–89. Valdimarsdottir, H.B., Zakowski, S.G., Gerin, W., Mamakos, J., Pickering, T., Bovbjerg, D.H., 2002. Heightened psychobiological reactivity to laboratory stressors in healthy women at familial risk for breast cancer. J. Behav. Med. 25, 51–65. Weinstock, M., Poltrev, T., Schorer-Apelbaum, D., Men, D., McCarty, R., 1998. Effect of prenatal stress on plasma corticosterone and catecholamines in response to footshock in rats. Physiol. Behav 64 (4), 439–444. Young, E.A., 1998. Sex differences and the HPA axis: implications for psychiatric disease. J. Gend. Specif. Med. 1 (1), 21–27. Young, E.A., Akana, S., Dallman, M.F., 1990. Decreased sensitivity to glucocorticoid fast feedback in chronically stressed rats. Neuroendocrinology 51, 536–542.
Stronger endocrine responses after brief psychological stress in women at familial risk of breast cancer S.M. Gold a, S.G. Zakowski b, H.B. Valdimarsdottir a, D.H. Bovbjerg a,∗ a
b
Biobehavioral Medicine Program, Derald H. Ruttenberg Cancer Center, Mount Sinai School of Medicine, 1425 Madison Avenue, Box 1130, New York, NY 10029, USA Department of Psychology, Finch University of Health Sciences, The Chicago Medical School, 3333 Green Bay Road North, Chicago, IL 60064, USA Received 26 November 2001; received in revised form 21 May 2002; accepted 22 May 2002
Abstract Recent research has linked exposure to chronic stress to altered acute stress responses and suggests a sensitizing effect of chronic stress leading to a stronger endocrine and cardiovascular response to acute stressors. Substantial evidence indicates that familial breast cancer risk is a chronic life stressor with higher levels of self reported distress. In this study, we investigated whether the endocrine response to a brief psychological stressor was stronger for women at familial risk for breast cancer. Thirty-six women at normal risk of breast cancer (FR⫺ Stress Group) and 17 women at familial risk (FR+ Stress Group) underwent a brief psychological laboratory stress test (speech task and mental arithmetic) over a 15 min period. Thirty women at normal risk not subjected to the stressful task served as controls (FR⫺ Control Group). Plasma epinephrine, norepinephrine and cortisol were measured at baseline, directly after the stress test (15 min) and at 30 min and 45 min post baseline. Heart rate data confirmed the effectiveness of the stress regimen. While there were no significant baseline group differences in the endocrine parameters, the response curves for the familial risk group revealed stronger epinephrine and cortisol reactivity to the stress test, as confirmed by significant group by time interactions. Norepinephrine levels showed a similar pattern, but results did not reach significance. These findings are in line with previous research documenting the facilitating effects of chronic stressors on acute stress response in animals and humans and provide the first
∗
Corresponding author. Tel.: +1-212-659-5562; fax: +1-212-849-2566. E-mail address: [email protected] (D.H. Bovbjerg).
0306-4530/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0306-4530(02)00046-X
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
585
evidence in the literature of a heightened endocrine reactivity to acute psychological stress in women at familial risk of breast cancer. The heightened endocrine reactivity to the experimental tasks seen here suggests that these women may experience stronger responses to stressors in their daily lives. According to the recently proposed concept of allostatic load, repeated overly strong stress responses may cumulatively have negative health implications. 2003 Elsevier Science Ltd. All rights reserved. Keywords: Familial breast cancer risk; Allostatic load; Experimental stress; Cortisol; Catecholamines; Recovery
1. Introduction Animal models have compellingly demonstrated that prior exposure to a chronic stressor significantly elevates reactivity to a novel acute stressor in both the hypothalamo–pituitary–adrenal (HPA) axis (Bhatnagar and Dallman, 1998; Weinstock et al., 1998; Bhatnagar et al., 1995; Young et al., 1990; Ottenweller et al., 1989) and the sympathetic pathway (McCarty et al., 1988; Konorska et al., 1989; Weinstock et al., 1998). Research in humans has been scant but has yielded results consistent with a higher neuroendocrine response in some studies, though not all. For example, increased epinephrine peak levels after a brief psychological stressor among healthy volunteers with chronic life stress have been reported (Pike et al., 1997). Recent reports in older women found that having a history of chronic stress, such as caregiving, was associated with increased cortisol reactivity to a brief psychological stressor, although cardiovascular parameters, as well as catecholamine responses did not differ from controls (Cacioppo et al., 2000). One chronic stressor that has received little attention with respect to possible effects on acute neuroendocrine reactivity is being at risk of developing a life-threatening disease (e.g., cancer). Having a family history of breast cancer is the single most powerful predictor that a woman will develop the disease at some point in her lifetime. Consistent with that threat, numerous reports have found significantly elevated levels of psychological distress (e.g., intrusive thoughts) in healthy women with a first-degree relative diagnosed with breast cancer, identifying family history of breast cancer as a potent chronic stressor (see Bovbjerg and Valdimarsdottir, 2001 for review). For example, in an early study of 140 women with one or more firstdegree relatives with breast cancer, Lerman et al. (1993) found levels of intrusive thoughts that were comparable to those seen in clinical populations, with 30% of the sample reporting that intrusive thoughts interfered with their daily lives. More recently, Baider et al. (1999) found that 53% of 230 women with family histories of breast cancer experienced distress reaching clinical levels. Another recent study (Neise et al., 2001) reported that 67% of 129 women with family histories of breast cancer evidenced medium or intense psychological strain, with 9 out of 10 women describing negative effects that impaired their mental and physical well-being. We have recently reported that women at familial risk of developing breast cancer
586
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
show a stronger acute stress response as indicated by: self-reported distress, elevated heart rate, increased natural killer cell activity (NKCA) and NK cell numbers immediately post stressor compared to women at normal risk (Valdimarsdottir et al., 2002). In this paper, we report for the first time the cortisol and catecholamine response to a brief laboratory stressor (speech task and mental arithmetic) in women at familial risk of breast cancer. We further examined the recovery curve after stressor termination.
2. Methods 2.1. Subjects From a sample of 96 women participating in the study, 83 women had complete data in requisite cardiovascular and endocrine measures and were thus included in the analyses. Participants were eligible if they were between 25 and 50 years of age, premenopausal, had not taken birth control pills in the past two months, were not currently pregnant and at least six months after their last pregnancy, did not report a personal history of chronic disease (including any neurological or psychiatric disease) or neoplasm, were not currently taking any drug known to influence immune or endocrine function and did not have relatives currently in active cancer treatment. The mean age of participating women was 35.1 years, the majority were white (75%), well educated (93% college graduates), and employed outside the home (81%). Based on their self-reported family histories of breast cancer, participants were classified as being at increased risk of developing breast cancer (FR+, n=17) or normal risk (FR⫺, n=66) following the algorithm of Claus et al. (1994). The FR⫺ group was randomly assigned to either the stress (n=36) or control condition (n=30) while all FR+ subjects underwent the acute psychological stressor. The groups did not differ significantly in any of the background variables: age (FH⫺ control 33.6 years±9.4; FH⫺ stress 35.8 years±8.2; FH+ stress 35.9 years±7.4; ANOVA F(2, 77)=0.67; df=2; p=0.52), height (FH⫺ control 163.6 cm±7.5; FH⫺ stress 165.8 cm±5.9; FH+ stress 164.8 cm±7.3; ANOVA F(2, 76)=0.80; p=0.45), weight (FH⫺ control 59.2 kg±6.9; FH⫺ stress 64.9 kg±11.7; FH+ stress 61.4 kg±8.8; ANOVA F(2, 74)=2.7; p=0.08), body mass index (BMI) (FH⫺ control 22.3±2.9; FH⫺ stress 23.7±3.8; FH+ stress 22.6±3.4; ANOVA F(2, 72)=1.29; p=0.29), or days since last period (FH⫺ control 12.2 days±11.6; FH⫺ stress 16.4 days±21.2; FH+ stress 15.5 days±14.3; ANOVA F(2, 75)=0.51; p=0.60). Fisher’s Exact Test two-sided tests for categorical data with small cell sizes showed that there were no significant group differences in the distribution of women having a BMI greater than 30, i.e. WHO obesity class I (FH⫺ control one subject; FH⫺ stress three subjects; FH+ stress one subject; Fisher’s Exact Test two-sided significance p=0.65). Chi squared tests revealed no differences on demographic variables: employment (chi square=10.4; p=0.74), level of education (chi square=3.0; p=0.80), family income (chi square=12.2; p=0.28), marital status (chi square=8.6; p=0.20), or ethnicity (chi square=11.7; p=0.17) distributions also did not differ significantly. Thus, none of
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
587
these variables was statistically controlled for in the following analyses. As there was a trend for a group difference in body weight, all the analyses described below were also conducted with weight included as a covariate and an identical pattern of significant results was seen. 2.2. Procedure The exact procedure of the study has been described in detail elsewhere (see Valdimarsdottir et al., 2002). Briefly, potential participants were screened in a telephone interview and an introductory session was scheduled to sign consent forms and allow the women to adapt to the setting in which the experiment would take place. The experimental sessions were scheduled between 0800 h and 1000 h. The session was rescheduled if the women reported any symptoms indicative of a cold or flu in the previous three days. Subjects were asked to refrain from drinking more than one cup of coffee on the morning of the experiment and from consuming more than two alcoholic drinks the night before the session. They were further told not to take medications during the 24 hours before the experiment. Heart rate (HR) was assessed prior to, and throughout the test at 2 min intervals. Subjective distress levels were assessed before and after the experimental manipulation using visual analogue scales. Blood samples were collected at baseline and at several time points after the stress test (at 15 min, 30 min and 45 min post baseline) for cortisol, norepinephrine and epinephrine measurement. An iv-catheter was inserted in a vein of a non-dominant arm and was kept patent with saline during the experimental session. The participants rested for 20 min to recover from possible effects of the catheter insertion procedure per se. The stress test was a modified version of the Trier Social Stress Test (Kirschbaum et al., 1993), comprising a 5 min speech task in which participants were asked to imagine they had been caught for a traffic violation and had to defend themselves at the traffic court. They were allowed two minutes for preparation and three minutes to deliver the speech. Speeches were delivered in front of a video camera, and the women were told that their performance would later be rated by experts for content and style. This was followed by a 5 min mental arithmetic task with harassment, during which participants were asked to add numbers out loud at the pace set by an audiotape. The control task consisted of two 5 min sessions of a non-stressful reading assignment, in which the women were told they could read the material provided by the experimenter at their leisure and would not be tested later on. All participants were provided a modest reimbursement for their time and effort. 2.3. Endocrine assays Epinephrine and norepinephrine plasma concentrations were measured using commercial radioenzymatic assays (Amersham Pharmacia Biotech, Inc., Piscataway, NJ). Plasma cortisol levels were assayed with a commercial radioimmunoassay kit (Diagnostic Products Corporation, Los Angeles, CA, 90045).
588
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
2.4. Statistical analysis To examine endocrine responses induced by the stress test, delta values (change compared to baseline) were computed for all post stressor time points, thus reducing the impact of well known interindividual variability in basal levels. It should be noted that confirmatory analyses of raw data revealed an identical pattern of significant interactions as that seen with delta values. Endocrine acute responses and recovery curves throughout the 30 min post-stress period were examined using repeated measures ANOVA with group as the between-subjects factor (FR⫺ control vs. FR⫺ stress vs. FR+ stress) and changes from baseline as the within-subject factor. To reduce Type I error, Sidak-adjusted between groups post hoc tests were computed for each time point separately, to explore the source of significant interaction terms. Linear regression approaches were used to explore the predictive value of changes in heart rate and self-reported distress on endocrine responses. All analyses were conducted with statistical software (SPSS 10.0), with a p value of less than 0.05 considered significant. 2.5. Effectiveness of the stressor Consistent with our previous findings with a subsample (n=65) of the participants (Valdimarsdottir et al., 2002), results in the present study showed significant group differences in both self-reported distress responses (ANOVA F(2, 79)=63.0; p⬍0.001), and HR reactivity during the stress test (ANOVA F(2, 79)=23.89; p⬍0.001). Post-hoc tests showed significant HR differences between all three groups, with a significantly stronger stress response in FH+ subjects compared to FH⫺ subjects, while both stress groups significantly differed from unstressed controls. Exploratory analyses of relations among these dependent variables, indicated that within the two stress groups, cortisol acute response was significantly predicted by acute response in HR (standardized beta=0.41; p=0.01) while acute self-report response was not predictive (standardized beta=0.11; p=0.41). The same was seen in acute epinephrine response (HR standardized beta=0.32; p=0.03; self-report standardized beta=⫺0.03; p=0.83).
3. Results No significant baseline differences between groups were seen in plasma cortisol (FH⫺ control 15.4 µg/dl±11.4; FH⫺ stress 13.1 µg/dl±4.7; FH+ stress 11.6 µg/dl±4.9; ANOVA F(2, 80)=1.3; p=0.27), epinephrine (FH⫺ control 112.6 pg/ml±109.8; FH⫺ stress 82.7 pg/ml±101.1; FH+ stress 66.8 pg/ml±64.4; ANOVA F(2, 80)=1.4; p=0.26), or norepinephrine (FH⫺ control 331.3 pg/ml±224.7; FH⫺ stress 356.2 pg/ml±283.3; FH+ stress 344.2 pg/ml±136.9; ANOVA F(2, 80)=0.09; p=0.92). A marked acute stress response in epinephrine was observed in both stressed groups, which began to resolve at 30 min post baseline in FR⫺ but not in FR+
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
589
women (see Fig. 1). Multivariate tests in repeated measures ANOVA showed a significant Group main effect (F(2, 80)=5.20; p=0.01), no main effect for Time (F(3, 78)=0.46, p=0.70), but a significant Group×Time interaction (F(6, 158)=2.64, p=0.02). At 15 min post baseline, epinephrine was significantly higher in both stress groups compared to the Control Group. At 30 min, only the FR+ Stress Group remained significantly higher than controls (see Fig. 1). Although the pattern of responses seen in Fig. 1 is suggestive of reductions in epinephrine levels in the control group, single group ANOVAs revealed significant Time effects only for the FR+ stress group (F(3, 14)=4.29; p=0.02), while there were no significant Time effects for either of the other two groups (FR⫺ control F(3, 27)=1.92; p=0.15; FR⫺ stress F(3, 33)=0.93; p=0.44). Although responses in plasma norepinephrine levels showed a similar pattern, no significant Group (F(2, 80)=0.01, p=0.99), Time (F(3, 78)=1.10, p=0.36), or interaction effects (F(6, 158)=1.28, p=0.27) were observed. Women at familial risk for breast cancer showed an increase in plasma cortisol levels after the stress test while women at normal risk in the stress group as well as the control condition declined (see Fig. 2). Multivariate tests showed significant Time (F(3, 78)=16.57, p⬍0.001) and Group (F(2, 80)=3.82, p=0.03) main effects as well as a significant Group×Time interaction (F(6, 158)=2.29, p=0.04) for cortisol. Cortisol levels of the FR+ Stress Group remained significantly elevated above mean
Fig. 1. Epinephrine response to an acute psychological stressor in healthy women at familial risk (stress FR+, n=17) and normal risk for breast cancer (stress FR⫺, n=36) compared to non-stressed women at normal risk (control FR⫺, n=30). Asterisks indicate significant post-hoc difference (Sidak adjustment) to unstressed control group. See text for absolute initial values.
590
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
Fig. 2. Cortisol response to an acute psychological stressor in healthy women at familial risk (stress FR+, n=17) and normal risk for breast cancer (stress FR⫺, n=36) compared to non-stressed women at normal risk (control FR⫺, n=30). Asterisk indicates significant post-hoc difference (Sidak adjustment) to unstressed control group. See text for absolute initial values.
values for the Control Group until 30 min post baseline while the FR⫺ Stress Group did not differ from control subjects at any time (see Fig. 2).
4. Discussion To date, little research has investigated the effects of chronic stressors on responsiveness of endocrine parameters to acute psychological stress in humans. In this paper we report an alteration in the endocrine response to an acute stressor in women at familial risk of developing breast cancer, compared to women at normal risk. Stronger responses were seen in neuroendocrine activity of the major stress systems, the hypothalamo–pituitary–adrenal axis (plasma cortisol) and the sympathetic pathway (plasma epinephrine). Consistent with the literature (e.g., Cacioppo et al., 1995), acute endocrine response (pre vs. post stressor) in cortisol and epinephrine was significantly predicted by change in HR but not in self-report. This further underlines the relevance of endocrine outcome measures above and beyond self-report stress assessments. As familial risk of breast cancer is a chronic stressor (see recent review by Bovbjerg and Valdimarsdottir, 2001), these findings are in line with reports from animal models showing the facilitating effect of chronic stressors on acute stress hormone responses to brief mild stressors. While numerous studies have shown that
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
591
animals exposed to chronic stressors exhibit enhanced HPA and sympathetic responses to novel, acute stressors (Bhatnagar and Dallman, 1998; Weinstock et al., 1998; Bhatnagar et al., 1995; Young et al., 1990; Ottenweller et al., 1989; McCarty et al., 1988; Konorska et al., 1989), the underlying mechanisms for this phenomenon are not yet known. One proposal is that prior stressors leave a central facilitation trace that upon exposure to a novel stressor, overcomes negative feedback effects typical for these systems (Bhatnagar and Dallman, 1998). Studies of acute endocrine stress reactivity in chronically stressed subjects are rare and findings have been somewhat inconsistent in humans, perhaps reflecting differences in the types of chronic stressors, the type of acute stress, the timing of endocrine assessment, or the subject sample used. In line with chronic stressor effects in spouses of Alzheimer’s patients (Cacioppo et al., 2000), we found a significantly stronger endocrine response to an acute stress test in healthy women at familial risk of breast cancer. It is tempting to speculate that unremitting stressors such as caring for a demented partner, or being at risk to develop a life-threatening disease may be more likely to show evidence of an impact on the acute reactivity of physiological stress systems than circumscribed stressors. Consistent with this view, no alterations in endocrine stress reactivity to acute challenge were found in unemployed men and women (Ockenfels et al., 1995), or in male high school teachers with high job strain (Benschop et al., 1994). Another explanation for the mixed findings in the literature may be that our sample and the cited study from Cacioppo and colleagues are the only ones comprised exclusively of female participants, since both animal experiments (Rhodes and Rubin, 1999) and human studies (Young, 1998) have suggested that females may show stronger acute endocrine stress responses during chronic stress. However, Pike et al. (1997) found an epinephrine stress response significantly altered by chronic stress in a study exclusively enrolling men. The fact that cortisol levels declined over time in stressed as well as unstressed FR⫺ women may be attributed to the typical circadian pattern during the morning hours in which this experiment was carried out. While women without familial risk of breast cancer did not evidence an increased cortisol response following the relatively mild psychological stress test, significant increases were seen in women at familial risk despite the underlying steep circadian decline. It is thus conceivable that a smaller response in FR⫺ women was masked by decreasing basal levels in cortisol. Given the kinetics of the endocrine stress response (see Sapolsky et al., 2000), it is also conceivable that earlier assessment of catecholamines (e.g. at 5 min into the stressor) would have revealed even more pronounced group differences. Future research should look earlier in the stress response and do so during the afternoon, when circadian decline is less steep. In summary, our results underline the profound effects that having a familial risk of breast cancer can have on acute endocrine responses to stressors in healthy women. The health consequences of such heightened neuroendocrine reactivity to acute stress are not yet known. It has recently been theorized that while acute stress reactions are essential for adaptation, repeated heightened physiological responses to acute
592
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
stressors can exact an allostatic load on an individual, resulting in negative health consequences (McEwen, 1998). From this perspective, the present results raise the possibility that women with a family history of breast cancer may have increased health risks as a result of increased reactivity.
Acknowledgements This investigation was supported in part by Grant M01RR00047 from the General Clinical Research Centers Program of the Division of Research Resources, National Institutes of Health. The study was also sponsored in part by grants from the Department of Defense (#DAMD J-4139 and J-4164) and the National Cancer Institute (#R01 CA72457). We are required to indicate that the content of the information contained in this report does not necessarily reflect the position or policy of the United States Government. We would also like to acknowledge the helpful comments of the anonymous reviewers.
References Baider, L., Ever-Hadani, P., Kaplan De-Nour, A., 1999. Psychological distress in healthy women with familial breast cancer: like mother, like daughter? Int. J. Psychiatry Med. 29, 411–420. Benschop, R.J., Brosschot, J.F., Godaert, G.L., De Smet, M.B., Geenen, R., Olff, M., Heijnen, C.J., Ballieux, R.E., 1994. Chronic stress affects immunologic but not cardiovascular responsiveness to acute psychological stress in humans. Am. J. Physiology 266, R75–R80. Bhatnagar, S., Mitchell, J.B., Betito, K., Boksa, P., Meaney, M.J., 1995. Effects of chronic intermittent cold stress on pituitary adrenocortical and sympathetic adrenomedullar function. Physiol. Behav 57 (4), 633–639. Bhatnagar, S., Dallman, M., 1998. Neuroanatomical basis for facilitation of hypothalamic–pituitary–adrenal responses to a novel stressor after chronic stress. Neuroscience 84, 1025–1039. Bovbjerg, D.H., Valdimarsdottir, H.B., 2001. Interventions for healthy individuals at familial risk for cancer. In: Baum, A., Anderson, B.L. (Eds.), Psychosocial interventions for cancer. American Psychological Association, Washington, DC, pp. 305–320. Cacioppo, J.T., Malarkey, W.B., Kiecolt-Glaser, J.K., Uchino, B.N., Sgoutas-Emch, S.A., Sheridan, J.F., Berntson, G.G., Glaser, R., 1995. Heterogeneity in neuroendocrine and immune responses to brief psychological stressors as a function of autonomic cardiac activation. Psychosom. Med 57 (2), 154–164. Cacioppo, J.T., Burleson, M.H., Poehlmann, K.M., Malarkey, W.B., Kiecolt-Glaser, J.K., Berntson, G.G., Uchino, B.N., Glaser, R., 2000. Autonomic and neuroendocrine responses to mild psychological stressors: effects of chronic stress on older women. Ann. Behav. Med 22, 140–148. Claus, E.B., Risch, N., Thompson, W.D., 1994. Autosomal dominant inheritance of early-onset breast cancer. Cancer 73, 643–651. Kirschbaum, C., Pirke, K.M., Hellhammer, D.H., 1993. The “Trier Social Stress Test” a tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology 28, 76–81. Konorska, M., Stewart, R.E., McCarty, R., 1989. Sensitization of sympathetic–adrenal medullary responses to a novel stressor in chronically stressed laboratory rats. Physiol. Behav 46, 129–135. Lerman, C., Daly, M., Sands, C., Balshem, A., Lustbader, E., Heggan, T., Goldstein, L., James, J., Engstrom, P., 1993. Mammography adherence and psychological distress among women at risk for breast cancer. J. Natl. Cancer Inst. 85, 1074–1080.
S.M. Gold et al. / Psychoneuroendocrinology 28 (2003) 584–593
593
McCarty, R., Horwatt, K., Konarska, M., 1988. Chronic stress and sympathetic–adrenal medullary responsiveness. Soc. Sci. Med 26, 333–341. McEwen, B.S., 1998. Protective and damaging effects of stress mediators. N. Engl. J. Med. 338, 171–179. Neise, C., Rauchfuss, M., Paepke, S., Beier, K., Lichtenegger, W., 2001. Risk perception and psychological strain in women with a family history of breast cancer. Onkologie 24, 470–475. Ockenfels, M.C., Porter, L., Smyth, J., Kirschbaum, C., Hellhammer, D.H., Stone, A.A., 1995. Effects of chronic stress associated with unemployment on salivary cortisol: overall cortisol levels, diurnal rhythm, and acute stress reactivity. Psychosom. Med 57, 460–467. Ottenweller, J.E., Natleson, B.H., Pitman, D.L., Drastal, S.D., 1989. Adrenocortical and behavioral responses to repeated stressors: towards an animal model of chronic stress and stress-related mental illness. Biol. Psychiatry 26, 829–841. Pike, J.L., Smith, T.L., Hauger, R.L., Nicassio, P.M., Patterson, T.L., McClintick, J., Costlow, C., Irwin, M.R., 1997. Chronic life stress alters sympathetic, neuroendocrine, and immune responsivity to an acute psychological stressor in humans. Psychosom. Med 59, 447–457. Rhodes, M.E., Rubin, R.T., 1999. Functional sex differences (“sexual diergism”) of central nervous system cholinergic systems, vasopressin, and hypothalamic–pituitary–adrenal axis activity in mammals: a selective review. Brain Res. Brain Res. Rev 30 (2), 135–152. Sapolsky, R.M., Romero, L.M., Munck, A.U., 2000. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocrine Rev 21 (1), 55–89. Valdimarsdottir, H.B., Zakowski, S.G., Gerin, W., Mamakos, J., Pickering, T., Bovbjerg, D.H., 2002. Heightened psychobiological reactivity to laboratory stressors in healthy women at familial risk for breast cancer. J. Behav. Med. 25, 51–65. Weinstock, M., Poltrev, T., Schorer-Apelbaum, D., Men, D., McCarty, R., 1998. Effect of prenatal stress on plasma corticosterone and catecholamines in response to footshock in rats. Physiol. Behav 64 (4), 439–444. Young, E.A., 1998. Sex differences and the HPA axis: implications for psychiatric disease. J. Gend. Specif. Med. 1 (1), 21–27. Young, E.A., Akana, S., Dallman, M.F., 1990. Decreased sensitivity to glucocorticoid fast feedback in chronically stressed rats. Neuroendocrinology 51, 536–542.